UNIVERSITY OF CALIFORNIA 
 
 ANDREW 
 
 SMITH 
 
 HALLIDIE: 
 
SHOP AND ROAD TESTING 
 
 OF 
 
 DYNAMOS AND MOTORS 
 
 A PRACTICAL MANUAL FOR THE TESTING 
 FLOOR, THE CAR BARN AND THE ROAD. 
 
 BY 
 
 EUGENE C. PARHAM, M. E. 
 
 1 1 
 
 Formerly Electrician Steel Motor Co., Johnstown, Pa. 
 AND 
 
 JOHN C. SHEDD, PH. D. 
 
 Professor of Physics, Colorado College, formerly Fellow and 
 
 Instructor in the University of Wisconsin and 
 
 Professor of Physics at Marietta College. 
 
 Experience may bt a dear tck*ol y but it it the best. 
 
 NEW YORK 
 
 McGRAW PUBLISHING CO. 
 1901 
 

 COPYRIGHT, 1901, 
 
 ELECTRICAL WORLD AND ENGINEER, 
 (INCORPORATED.) 
 
PREFACE. 
 
 THE following chapters are the outcome of a need felt 
 by the authors, during their experience on the testing floor 
 and road, for a manual that would cover this field of work 
 in such a way as to be a help alike to the student fresh 
 from the theoretical side of the subject but unacquainted 
 with shop details, and to the so-called practical man, who 
 is largely self-taught, as to the theory of the machines he 
 handles. 
 
 On the purely theoretical side such works as those of 
 S. P. Thompson and D. C. Jackson leave little to be 
 desired, while from the purely practical side a multitude 
 of so-called practical books flood the market; but works 
 that are at once correct from the standpoint of the 
 theorist, and yet valuable to the strictly practical man, 
 have not been easy to find. This need is felt by the 
 college graduate as he steps from the plane of the 
 laboratory and lecture room to the hard-pan of the com- 
 mercial testing floor. It is also felt by a multitude of 
 station managers and engineers who find the hit-and-miss 
 methods of testing far from satisfactory. 
 
 The present work has a two-fold object: I. To give 
 a complete theory of the commercial testing floor, so far 
 
 iii 
 
 
 UNIVERSITY 
 
IV PREFACE. 
 
 as it relates to direct current machines, and of the 
 multitudinous applications of theory to practice. II. To 
 meet the growing demand on the part of operating com- 
 panies for a manual that shall enable them to do their 
 own repair work and consequent testing. 
 
 Part I. is devoted to such fundamental and preliminary 
 conceptions as are needed to help those unacquainted 
 with the general theory. It may be omitted by many 
 readers. 
 
 Part II. treats of instrumental testing. The treatment 
 of the ammeter, voltmeter, and galvanometer is mathe- 
 matically simple and seeks to give the physical con- 
 ception embodied in the formulae. 
 
 Part III. takes up in detail the tests of dynamos and 
 motors. Special attention has been given to the many 
 difficulties that confront the tester, and all examples and 
 illustrations are drawn from personal experience. The 
 chapter on Compounding is specially full. It is hoped 
 that the chapter on Grounds on the Line may be of 
 service to the lighting station and street railway 
 operator. 
 
 Care has been observed not to stray from the paths of 
 actual practice on the one hand, and, on the other, not to 
 offer inadequate or incorrect explanations. 
 
 From the nature of the case mathematical treatment 
 has been simplified to the last degree, and even the 
 graphical method but little used. 
 
 The International electrical units have been adopted and 
 the following abbreviations used: Electromotive force, 
 
PREFACE. V 
 
 E. M. F.; counter electromotive force, C. E. M. F.; 
 alternating current electromotive force, A. C. E. J/. P.; 
 current, /, /; resistance, R, r; other abbreviations that 
 have been used are self-explanatory. 
 
 The authors take pleasure in acknowledging the 
 courtesy and help they have received from several of 
 the manufacturing companies and from old associates 
 
 and friends. 
 
 EUGENE C. P ARM AM, 
 
 JOHN C. SHEDD. 
 OCTOBER, 1897. 
 
PREFACE TO SECOND EDITION. 
 
 In meeting the demand for a second edition of this 
 work, the authors have taken the opportunity of cor- 
 recting such typographical errors as have come to their 
 attention. They have also extended the scope of the 
 book so as to include the field of street-car equipment 
 and operation. In this the endeavor has been to be 
 strictly practical in the selection and treatment of the 
 subjects discussed. At the same lime it is the aim of the 
 writers to aid the reader in gaining a comprehensive grasp 
 of the principles involved in the problems discussed, and 
 to thereby enable him to successfully meet the numerous 
 difficulties which the man on the front platform con- 
 tinually encounters. 
 
 It is believed that Part IV will be found a valuable 
 addition to the book, and that the high standard aimed 
 at in the balance of the work is fully maintained. 
 
 E. C. PARHAM, 
 JOHN C. SHEDD. 
 
CONTENTS. 
 
 INTRODUCTION. 
 CHAPTER I. 
 
 ELEMENTS OF THE DYNAMO. 
 
 PAGE 
 
 Laws of Energy ; Units of Electricity and Magnetism ; 
 Permeability ; the Magnetic Circuit ; the Ampere, Volt, 
 Ohm ; Induction ; Magnetomotive Force ; Electromag- 
 netic Induction ; Elements of the Dynamo ; Methods 
 of Excitation ; Series, Shunt, and Compound-Wound 
 Machines ; Losses in the Dynamo ; Cross and Back 
 Induction ; Efficiency, 3 
 
 CHAPTER II. 
 
 ELEMENTS OF THE MOTOR. 
 
 The Electric Motor ; Counter Electromotive Force ; Elec- 
 trical and Commercial Efficiency ; Maximum Activity ; 
 Maximum Efficiency ; Torque ; Speed Regulation ; 
 Classification of Motors ; Series, Shunt, and Compound- 
 Wound Motors ; Direction of Rotation ; Motors and 
 Dynamos, 36 
 
 vii 
 
Vlll CONTENTS. 
 
 THE TESTING AND USE OF INSTRUMENTS. 
 
 CHAPTER III. 
 
 OHM'S LAW. 
 
 PAGE 
 
 General Discussion of Ohm's Law ; Resistance ; Conduc- 
 tivity ; Fall of Potential ; Various Expressions for Ohm's 
 Law ; Joule's Law ; Inductive Circuits ; Further Defi- 
 nition of Resistance, 59 
 
 CHAPTER IV. 
 
 MEASUREMENT OF CURRENT, 
 
 Measurement of Current by Copper Voltameter ; Standard- 
 izing Ammeters by the Voltameter ; Fundamental 
 Principles of the Galvanometer ; Determination and 
 Explanation of Galvanometer Constants ; Theory and 
 Application of Shunts ; Construction of Standard Shunt ; 
 Its Use with Ammeters ; Multiple Circuits ; Siemens 
 Dynamometer ; Wattmeter, 70 
 
 CHAPTER V. 
 
 MEASUREMENT OF ELECTROMOTIVE FORCE. 
 
 Definition of E. M. F. and P. D.; Standards of E. M. F.; 
 General Discussions of Batteries ; Daniell's Standard 
 Cell ; Leclanche Cell ; the Galvanometer, its Construc- 
 tion, Adjustment, and Mounting ; Measurement of 
 E. M. F. by the Galvanometer ; Proportion Lines ; Pro- 
 portion Boxes ; Wiring the Galvanometer ; Clark Cell ; 
 Calibrating a Voltmeter ; Slide Wire Method of Measur- 
 ing E. M. F.; Multipliers, 112 
 
CONTENTS. IX 
 
 CHAPTER VI. 
 
 MEASUREMENT OF RESISTANCE. 
 
 PAGE 
 
 Resistance in General ; Measurement of Low Resistance ; 
 Method of Comparison of Potentials ; Locating Armature 
 Faults ; Locating Faulty Section ; Bar to Bar Test ; 
 Grounds on Armatures ; Measurement of Moderate 
 Resistances ; " Vienna" Method ; Theory of Wheatstone 
 Bridge, Slide Bridge, Box Bridge, Portable Bridge, 
 Their Application ; Very Low Resistance ; Sir W.Thom- 
 son's Bridge ; Remarks on Bridge Work ; Differential 
 Galvanometer, Its Theory and Practice, . . .145 
 
 CHAPTER VII. 
 
 MEASUREMENT OK INSULATION. 
 
 Insulation Measurement with Galvanometer and Shunt Box ; 
 Testing Glass Insulators ; Marine and Underground 
 Cables ; Telephone Cables ; Electrometer Method ; 
 Armature and Field Insulation ; General Remarks on 
 Insulation Work ; Alternating Current Test for High 
 Tension Service ; Liquid Resistances ; Battery Resist- 
 ances, Water Rheostats ; Slide Bridge and Telephone 
 Method ; Temperature Coefficient of Resistance ; Specific 
 Resistance ; General Remarks on Instruments, . .183 
 
 TESTING OF DYNAMOS AND MOTORS. 
 CHAPTER VIII. 
 
 THE SERIES MACHINE. 
 
 General Discussion ; Troubles Incident to Self-Exciting 
 Machines ; Tests for Open and Short Circuits in Field 
 and Armature ; Tests for Grounded Armature ; Run- 
 ning Series Machines in Multiple; Regulation in Arc 
 Dynamos ; the Thomson-Houston Arc Dynamo ; Brush 
 Arc Dynamo ; the Westinghouse Arc Dynamo, . . 231 
 
X CONTENTS. 
 
 CHAPTER IX. 
 
 SHUNT AND COMPOUND-WOUND MACHINE. 
 
 PAGE 
 
 Shunt Machine ; General Considerations ; E. M. F. Regula- 
 tion ; Field Resistance ; Efficiency ; Rheostat Regula- 
 tion ; Effect of Temperature ; Limits of Regulation ; 
 Rheostats Parallel and Series Running ; Putting a 
 Machine into Circuit; Direction of Rotation upon Rever- 
 sal of Shunt and Series Machines C. W. Machines; Con- 
 nections for Parallel Working ; Introducing and Taking 
 a Machine from Service ; Principles of Compounding ; 
 Over-Compounding; Compounding Volts per Revolution; 
 Compounding a Shunt Machine without Instruments, . 279 
 
 CHAPTER X. 
 
 THE COMPOUND-WOUND MACHINE GENERAL TESTS. 
 
 The Equalizing Bar ; Multiple and Series Running ; Test I. 
 Test of an Eight-Volt Twenty-Five-Ampere Shunt 
 Machine ; Test II. Motor-Generator Test with Engine 
 as Loss-Supplier ; Test III. Motor-Generator Test with 
 Lamp-Bank; Test IV. Motor Generator Test with three 
 Machines ; Test V. Motor-Generator Test, Machines of 
 Different-Current Capacity ; Test VI. Motor-Generator 
 Test, Same as Test V. with Lamp-Bank ; Test VII. 
 Motor-Generator Test, Same as Test V. with Motor ; 
 Test VIII. Motor-Generator Test, Machines of Different 
 E. M. F. and Current Capacity ; Test IX. Test of Single 
 Machine, with Water Rheostat 331 
 
 CHAPTER XL 
 
 COMPOUNDNG. 
 
 Test X. Compounding Test with Full Details, to Run 
 Under Full Load Two 500 Volt 500 K.W. Railway Gener- 
 ators and to Supply the Loss from an Auxiliary Dynamo 
 of the Same Voltage ; Test XI. Compounding Test, 
 Three Machines Run in Series, 378 
 
CONTENTS. XI 
 
 CHAPTER XII. 
 
 MISCELLANEOUS TESTS. 
 
 PAGE 
 
 Test XII. Core-loss Test ; Test XIII. Eddy Current Test ; 
 Test XIV. Saturation Test; Test XV. Distribution Test; 
 Test XVI. Brush Test; Test XVII. Efficiency Test, . 419 
 
 CHAPTER XIII. 
 
 GROUNDS ON THE LINE. 
 
 The Ground Detector ; Use of Voltmeter in Measuring Leak- 
 age Resistance; Use of the Ammeter in Measuring Resist- 
 ance of a Fault ; Use of Voltmeter for the Same ; Ground 
 Detector on Alternating Current Circuits ; Picou's Con- 
 densor Method for Alternating Current Circuits ; Stanley 
 Static Ground Detector ; Methods of Locating Grounds ; 
 Magneto Method ; Underground Circuits and Fall of 
 Potential Method ; First Bridge Method ; Second Bridge 
 Method ; Burning out Faults 439 
 
 CHAPTER XIV. 
 
 MOTOR TESTING. 
 
 Classification of Motors ; Cross-connecting of Armatures and 
 Commutators ; Starting Boxes for Shunt and Series 
 Motors ; Tendency of Series Motors to Race under no 
 Load ; Speed under Sudden Removal of Load ; Test 
 XVIII. Motor Generator Test with Street Car Motors; 
 Test XIX. Testing Series Machines on Water Box ; Test 
 XX. Efficiency Test ; Test XXL Efficiency Test with 
 Prony Brake ; Testing Series Machines Rigidly Con- 
 nected ; The Compound- Wound Motor, in General ; Motor 
 Work in Shops, Speed, Voltage, Load ; Small Motors ; 
 Motors Run in Series ; Stray Facts ; Dynamos and Motor 
 Rating 468 
 
Xll CONTENTS. 
 
 CHAPTER XV. 
 
 INSTALLATION CAR EQUIPMENT TESTS. 
 
 General Remarks. The Trolley Stand; Setting up, Inspecting; 
 Testing of Overhead Switches and Circuit Breakers; Test- 
 ing Fuse Boxes ; Testing and Selecting Fuses ; The 
 Lightning Arrester; Kicking Coil Requirements, Faults, 
 Testing; Starting Coil Use and Abuse of, Faults, Testing; 
 Shunt and Loop Action and Object of; Symptoms of 
 Disorder; Controllers; Type K2, Connections, Care of; 
 Motors 50 H.P. Modern Motor, Unique Features of. . 510 
 
 CHAPTER XVI. 
 
 CAR EQUIPMENT TESTS. 
 
 The Test Circuit; Construction of a Test Lamp Bank; Test- 
 ing a Starting Coil for Open Circuits, Short Circuits, 
 Grounds; Testing a Controller for Open Circuits, Short 
 Circuits, Grounds; Testing a Motor for Open Circuit in 
 Fields and Armature, for Polarity; Marking Motor Leads; 
 Setting the Gear; Mounting Motors, General Suggestions as 
 to Care and Precautions; Testing the Gears, Symptoms of 
 Disorder and their Causes; Testing Connections by Start- 
 ing Up, Faults Located by Symptoms; Grounds and Short 
 Circuits; Location of by Test Lines; Voltmeter Tests; 
 Baked or Short Circuited Fields; Reversed Coils; Test with 
 Nail or Compass; Grounds on Armature, Peculiar Actions 
 Resulting; Open Circuit in Armature; Motors used as 
 Electric Brakes; Bucking of Motors, ..... 558 
 
 APPENDIX. 
 
 Table I. Data for Copper Wire, 613 
 
 " II. Temperature Coefficients for Copper, . . 614 
 
 III. Data for Galvanized Iron Wire, . . . .615 
 
 " IV. Current Capacity for Iron Wire, . . . 616 
 
 " V. Fusing Effects of Currents, . . . . .617 
 
 "* VI. Fusing Effects of Currents, ... . 618 
 
 " VII. Test Sheet for Compounding, . ' . . . 619 
 
 " VIII. Test Sheet for Compounding 620 
 
SHOP AND ROAD TESTING OF 
 DYNAMOS AND MOTORS 
 
 INTRODUCTION 
 
 CHAPTER I. 
 
 ELEMENTS OF THE DYNAMO. 
 
 WHEN told for the first time that more energy is 
 required to rotate a metal disc between magnet poles 
 than away from them, the beginner's impulse is to ques- 
 tion the statement; if further told that the disc will 
 grow warm under continued rotation, his interest is 
 aroused; and the further assertion that with a brush on 
 the axle and another on the rim, the wheel can be made 
 to generate an electric current, calls forth his demand 
 for experimental proof. 
 
 To locate the true seat of energy in a dynamo or 
 motor is a puzzling problem unsolved in the minds of 
 many electrical and mechanical artisans. The intelligent 
 workman observes the belt tugging away at the pulley, 
 and, falling into the popular misconception of the case, 
 he imposes the burden upon every sort of friction, save 
 that of the magnetic field, to which it is due, and any 
 explanation involving conductors, lines of force, and 
 
4 TESTING OF DYNAMOS AND MOTORS. 
 
 kindred terms, evokes a look of bewilderment. Nor is 
 this strange. Science has as yet given no fundamental 
 answer to the problem of the metal disc, and the 
 transference of energy without the intervention of 
 bodies visibly touching each other, is and probably 
 always will be, a source of wonder to the wisest. 
 
 In the above simple experiment is embodied the 
 problem of the dynamo that of converting the mechani- 
 cal energy of the engine or water wheel into the energy 
 of an electric current. 
 
 Let the brushes of the disc be connected to the poles 
 of a battery and the conditions of the problem are 
 reversed; the wheel begins to turn and the electrical 
 energy of the battery is transformed into energy of 
 mechanical motion. The machine is now an electric 
 motor. From this experiment of Barlow and Faraday 
 dates the beginning of all practical applications of 
 electrical science. 
 
 In taking up any scientific study it is necessary to 
 first give elementary definitions of terms and phrases 
 ordinarily familiar, and to acquaint ourselves with the 
 underlying laws of the science. 
 
 There are two things which can be said to exist in the 
 physical universe, viz., Matter and Energy: Matter, as 
 the embodiment of all tangible things, and Energy as we 
 can appreciate it in all motion. The ordinary idea of 
 matter is so far correct as to require no further 
 definition. Energy is also easily recognized and may 
 be variously defined. The flying bullet has energy and 
 the faster the motion the more the energy. Energy, 
 then, is due to motion and can be measured in terms of it. 
 Again, a moving body strikes a second body at rest, and 
 
ELEMENTS OF THE DYNAMO. 5 
 
 the latter, if free to move, does so, while the former 
 loses part of its motion. The second body now has 
 energy in virtue of its motion, and we observe that 
 energy can be transferred from one body to another. 
 A third fact is gained when we observe that if a moving 
 body is stopped by an obstacle, as a hammer by the 
 anvil, both bodies become heated, and the greater the 
 energy of the hammer the greater the heat produced; 
 and further, the amount of heat produced is an exact 
 measure of the energy expended, or, as we say, of the 
 work done. This brings us to the laws of energy upon 
 which all engineering is based, and a thorough under- 
 standing of which would save many an industrious 
 inventor days of useless toil. The first law is that 
 energy is indestructible and uncreatable by human 
 agency. By this law, the Law of the Conservation of 
 Energy, whenever energy seems to disappear at one 
 point we must look for it elsewhere; nor rest until we 
 find its exact equivalent. The second law is that energy, 
 while indestructible, may be transformed into many 
 different forms. This Law of Transformability enables 
 us to account for the apparent disappearance of energy. 
 In the experiment above, to turn the uninfluenced disc 
 requires only the small amount of energy necessary to 
 overcome the friction of the air and axle boxes, and, 
 neglecting the former, the slight expenditure is accounted 
 for by the heating of the latter. Upon placing the mag- 
 net over the disc, as in Fig. i, more energy is required to 
 keep the disc in motion, more energy is consumed, and 
 the added amount is accounted for by the presence in 
 the disc of a current, between which and the magnet 
 a force of attraction exists; moreover, by the first law 
 
6 TESTING OF DYNAMOS AND MOTORS. 
 
 the amount of energy lodged in the current will depend 
 upon that expended in turning the disc, over and above 
 that l6st in frictional heat. 
 
 In all measurements some unit or standard must be 
 selected in terms of which quantity maybe expressed; 
 
 for example, the foot, the 
 pound, the second, the 
 horse-power, all are units. 
 There are two systems of 
 units: the Absolute and the 
 Practical. In the absolute 
 system all quantities are ex- 
 pressed in terms of the units 
 F of Mass, Length, and Time. 
 
 Thus, the unit of length is 
 
 the foot, and the distance from the earth to the moon 
 is 1,267,200,000 feet, and the velocity of a moving train 
 may be 88 feet per second expressed in the absolute 
 system; but it is much more convenient to say that the 
 moon is 240,000 miles away and that the train moves 60 
 miles an hour! There thus arises a set of units of 
 convenient multiples of the corresponding absolute 
 units, constituting the system of Practical Units. The 
 practical units with which we shall have to do, are 
 those of magnetism and electricity, but before taking 
 them up alone, we will see in what way they depend 
 upon the absolute units of Length, Mass, and Time. The 
 unit of length, the centimetre, is about 2/5 of an inch; 
 that of mass, the gram, about 2/1000 of a pound; and 
 that of time, the second; and together are known as the 
 C. G. S. system of measurement. The unit of Force is 
 the Dyne, and it has been agreed to define it as that force 
 
ELEMENTS OF THE DYNAMO. 7 
 
 which, acting upon unit mass, for unit time, shall increase 
 its velocity by i centimetre per second, /. <., if upon a mass 
 of i gram at rest, a force of i dyne be allowed to act 
 for i second, at the end of the second the gram weight 
 would be moving at the rate of i centimetre per sec- 
 ond.; at the end of the next second the velocity would 
 be i centimetre per second greater, and so on for each 
 succeeding second, so the Acceleration is said to be i 
 centimetre per second. At Paris the acceleration of 
 gravity is 981 centimetres per second; the force of grav- 
 ity is therefore 981 dynes, and i dyne is 1/981 of the 
 force of gravity at Paris. 
 
 The more important magnetic units are: 
 
 i. Unit of Pole Strength. A pole of unit strength is 
 one which repels with unit force an equal pole placed at 
 unit distance. Since at unit distance a unit pole repels 
 a unit pole with a force of i dyne, a pole of strength M 
 would repel unit pole with a force of M dynes, for each 
 unit of J/ exerts upon the unit pole a force of i dyne, 
 and there are M units in J/. Likewise a pole of strength 
 M would repel a pole of strength J/', J/' times as strongly 
 as it would unit pole, so we can say that at unit distance 
 two poles J/, J/' repel each other with a force of M J/' 
 dynes, or/ = M x M . Now if the distance between the 
 two poles is divided by 2, the force /becomes four times 
 as strong; if by 3, nine times as strong, and so on; while if 
 the distance is doubled / becomes but 1/4 as great; if 
 trebled, 1/9 as great, etc. By whatever amount the dis- 
 tance, </, is affected, the force, /, is affected by an amount 
 proportional to the reciprocal of d*\ in other words, the 
 rate at which /is affected by moving apart the poles in 
 question, can be measured by multiplying by itself the 
 
8 TESTING OF DYNAMOS AND MOTORS. 
 
 reciprocal of the rate at which d is affected. Hence the 
 law. The force between two poles varies directly as the 
 product of the pole strengths, and inversely as the square 
 of the distance; or, 
 
 MM' 
 
 where the -- sign corresponds to a force of attraction 
 and the -f- sign to one of repulsion. 
 
 2. Magnetic Field. If a free pole could be placed in a 
 magnetic field, the pole would move in a definite direc- 
 tion and with a force depending upon the strength of the 
 field. The force with which a unit pole is urged along is 
 a measure of the field strength, and that is unit field in 
 which unit pole is acted upon with unit force. It is to 
 Faraday, who has been well called 
 the " Prince of Experimenters," that 
 we owe the development of this sub- 
 ject. First, in order to form a mental 
 picture of a magnetic field, he consid- 
 ered it to be filled with imaginary lines, 
 which everywhere coincided in direc- 
 FIG. 2. tion with the path of the free pole; 
 
 these he called lines of force. He 
 next mapped out a uniform field of unit strength, by 
 considering that the lines were everywhere evenly dis- 
 tributed, one to each unit of cross-section, in a plane 
 normal to the lines. Thus each line in any field repre- 
 sents unit force, and the number of lines per unit of 
 cross-section represents the strength of the field. Fara- 
 day next assigned to the lines certain properties which 
 enabled him to explain many magnetic and electro- 
 magnetic phenomena, (i) He considered that lines of 
 
ELEMENTS OF THE DYNAMO. ') 
 
 force tend to shorten in length, hence the attraction 
 between magnets. (2) Lines going in the same direction 
 repel, in opposite directions attract each other, hence 
 N poles repel N poles, but attract C poles. (3) If a 
 conductor be made to cut lines, they resist its motion 
 just as rubber bands would. The result of forcing a con- 
 ductor across lines is to produce in it an E. M. F., and 
 this phenomenon is called Induction. (4) If a conductor 
 carrying a current be placed in a magnetic field, the field 
 acts upon that due to the current, and the conductor 
 moves in a direction depending upon the direction of the 
 current and the lines of the field; by this motion, how- 
 ever, the conductor is made to cut lines of force, and, 
 therefore, becomes the seat 
 
 of an induced E. M. F. By -^==j=. -=--^^^^^='-^^^^- 
 the law of the conservation 
 of energy, given above, the 
 
 induced E. M. F. is opposed FIG. 3. 
 
 to that which urges the 
 
 existing current, and constitutes the all important 
 
 C. E. M. F. (counter electromotive force) so much 
 
 heard of in connection with motors. 
 
 A magnetic pole produces about it a magnetic field, 
 aYid since unit pole exerts unit force at unit distance, the 
 field at unit distance must be of unit intensity; /. ^., 
 there must be one line of force in each square centimetre 
 of sectional area. There will, therefore, from unit pole 
 emanate as many lines as there are square centimetres in 
 the surface of the circumscribing sphere of unit radius 
 (see Fig. 2). Now the number of square centimetres in 
 the surface of any sphere is 
 
 4 X -- X r>, 
 
IO TESTING OF DYNAMOS AND MOTORS. 
 
 where C, d, and r are respectively the circumference, 
 diameter, and radius of the sphere, all expressed 
 in centimetres. The quotient of any circle's circum- 
 ference by its diameter is the number 3.1416, and is 
 designated by the Greek character, TT, called pi. The 
 area over any sphere is, then, 4 n r* and where r i, as 
 in the present case, it is simply 4 n X i 4 TT, which 
 expresses the number of lines due to unit pole. The 
 number of lines due to a pole of strength M is 4 n M. 
 
 3. Permeability. If in a magnetic field there be placed 
 a bar of iron, steel, or other magnetic metal, at the ends 
 of the bar the field is much stronger than elsewhere (see 
 Fig. 3), showing that the lines of force converge and 
 flow through the metal in preference to the air. This 
 property of magnetic "conductivity" is of great impor- 
 tance, and is possessed by iron and its modifications, by 
 nickel and cobalt. Iron and steel, however, so far sur- 
 pass nickel and cobalt in this respect as to be the only 
 magnetic metals used in electromagnetic machinery. 
 Magnetic conductivity, to distinguish it from electric 
 conductivity, is called Permeability, and it may present 
 itself more clearly as follows: If a current be passed 
 through a coil of wire the coil becomes an electromagnet, 
 and through the space inside pass a certain number of 
 lines of force, which we will call 3C. If the internal air 
 space be now filled by an iron core, the coil's strength as 
 a magnet is greatly increased; not that the iron has 
 created any lines of force, but it has coaxed many of the 
 lines which otherwise would form little circles about the 
 separate wires, to be gathered into it and pass through 
 the entire length of the coil to manifest themselves 
 at its ends. 
 
ELEMENTS OF THE DYNAMO. II 
 
 As far, then, as concerns the number of lines passing 
 through a given space, iron has in effect a multiplying 
 power, and this power, measured by dividing the num- 
 ber of lines which pass through the iron, by the number 
 of lines which pass through the air space when the 
 iron, is removed (the current and the number of turns 
 used being the same in both cases), is called the Pcrinc- 
 ability, and is designated by the Greek letter //, mu. 
 
 If JC is the number of lines which the given ampere- 
 turns will send through air, and (B is the number after the 
 iron is inserted, then the permeability, 
 
 (B 
 
 -- jc 
 
 and this is the formula usually given in books. ($> and JC 
 are respectively the number of lines per square centi- 
 metre in iron and in air; but where the same cross- 
 section is considered in both cases-, as above, the formula 
 is still true if we regard (B and JC as the total lines in 
 the given section under the two conditions. 
 
 The permeability of air remains always the same, /. <r., 
 the number of lines of force through air always increases 
 in the same measure as the magnetizing force producing 
 them. 
 
 The following table gives the relation between JC, (B 
 and p in iron. As JC increases (B does not increase with 
 the same rapidity, and hence /< decreases. 
 
 When (B has reached such a value that any increase in 
 JC produces comparatively little increase in (B, the iron is 
 said to be saturated. For wrought iron this is about 
 JC = 105, (B = 17,000. In cast iron the saturation point 
 is much lower, JC = 188, (B = 10,000. Above theso 
 values it is unprofitable to force the magnetization. 
 
12 TESTING OF DYNAMOS AND MOTORS. 
 
 TABLE No. i (SQUARE CENTIMETRE UNITS).* 
 
 ANNEALED WROUGHT IRON. 
 
 CAST IRON. 
 
 3C 
 
 (B 
 
 n 
 
 3C 
 
 (B 
 
 fj. 
 
 1.66 
 
 5,000 
 
 3,000 
 
 5 
 
 4,000 
 
 800 
 
 4 
 
 9,000 
 
 2,250 
 
 IO 
 
 5,000 
 
 500 
 
 5 
 
 10,000 
 
 2,000 
 
 21.5 
 
 6,000 
 
 279 
 
 6.5 
 
 11,000 
 
 1.692 
 
 42 
 
 7,000 
 
 133 
 
 8-5 
 
 12,000 
 
 1,412 
 
 80 
 
 8,000 
 
 IOO 
 
 12 
 
 13,000 
 
 1,083 
 
 127 
 
 9,000 
 
 71 
 
 17 
 
 14,000 
 
 823 
 
 1 88 
 
 10,000 
 
 53 
 
 28.5 
 
 15,000 
 
 526 
 
 292 
 
 11,000 
 
 37 
 
 50 
 
 16,000 
 
 320 
 
 
 
 
 105 
 
 17,000 
 
 161 
 
 
 
 
 2OO 
 
 18,000 
 
 90 
 
 
 
 
 350 
 
 19,000 
 
 54 
 
 
 
 
 666 
 
 20,000 
 
 30 
 
 
 
 
 It can be seen why wrought iron is to be preferred to 
 cast iron for magnetic purposes, and particularly when 
 the cross-section of the field cores is to be a minimum. 
 The best grades of cast steel compare very favorably 
 with the best wrought iron, in regard to permeability. 
 Some special grades of steel have the strength and 
 permeability of wrought iron, and, being handled as 
 easily as cast iron, can be used where a forging would be 
 out of the question. Steel castings are being increas- 
 ingly used in dynamo and motor construction, and are 
 
 * See S. P. Thompson's Electromagnet, p. 67. 
 
 NOTE. The statement that 3C in the iron is the field strength in the 
 air is not rigidly true. Whenever a piece of iron or steel is placed in a 
 magnetic field, the metal itself becomes a magnet of such polarity as to 
 oppose the existing field 3C- Hence the actual strength of JC i n tne 
 iron is slightly less than in air. 
 
ELEMENTS OF THE DYNAMO. 13 
 
 especially well adapted to frames of intricate shape. 
 With some manufacturers cast-iron frames are cast 
 over laminated wrought-iron pole-heads; with others 
 wrought-iron pole-heads are bolted onto the cast-iron 
 frame. 
 
 The terms ampere, volt, and ohm are familiar to 
 most of us, as they occur in Ohm's law. If a wire be 
 connected to the terminals of a cell a current flows. 
 This flow is not instantaneous, but involves time, and this 
 fact is implicitly stated when we say a wire has resist- 
 ance. By increasing the length of a wire, we increase 
 its resistance; l;y increasing the cross-section of the 
 wire, its resistance is decreased in a like measure: 
 whence we can say that the resistance of any wire varies 
 directly as its length and inversely as its cross-section, 
 or since its cross-section is .7854 </', or } n - </*, 
 inversely as the square of its diameter d; i. e., if the 
 length is doubled, so is the resistance, but if the diameter 
 is doubled, the cross-section is quadrupled, and the 
 resistance is but one quarter as great. With all solid 
 conductors, except carbon, the resistance rises when the 
 temperature does, but of carbon and liquids, the reverse 
 is true. Striking examples of these exceptions are met 
 with in using incandescent lamps and water boxes for 
 loading dynamos under test. On either of these devices 
 the current will creep above its initial value, but the 
 effect is most marked on the water box. 
 
 The practical unit of resistance is the Ohm and its 
 symbol (R or r. As defined by the Electrical Congress 
 held in Chicago in 1893, it is as follows: "The unit of 
 resistance shall be known as the International Ohm, and is 
 reoresented by the resistance offered to an unvarying 
 
14 TESTING OF DYNAMOS AND MOTORS. 
 
 electric current by a column of mercury, at the tempera- 
 ture of melting ice, 14.4521 grams in mass, of constant 
 cross-section and of a length of 106.3 centimetres." 
 The absolute unit and its relation to the practical unit 
 will be given later. 
 
 The quantity of electricity which passes through a 
 given conductor depends upon the rate and duration 
 of flow. The quantity per second is the rate of flow 
 and indicates the Current Strength. The international 
 unit is based upton the experimental fact that a given 
 quantity of electricity passing through a solution of a sil- 
 ver salt, decomposes the salt and deposits a fixed amount 
 of metallic silver upon the negative plate. The unit of 
 current is the Ampere, and is that current which deposits 
 silver at the rate of .0001118 of a gram per second. Its 
 symbol is / or /, but C is very commonly used. 
 
 Since all conductors have resistance, force is required to 
 urge a current through them. Whatever causes or tends 
 to cause a flow of electric current is called an Electromotive 
 Force. The term electrical pressure is also used. On 
 a conductor through which a current flows, successive 
 points are said to vary in potential, and according to a 
 well-defined law. The potential falls in the direction of 
 flow, and is a minimum at the negative pole of the source. 
 Two points may be said -to be at different potentials, and 
 the degree of difference is called the " drop " of poten- 
 tial between the points. The practical unit of poten- 
 tial difference corresponding to the ohm and ampere is 
 the international Volt, and is officially defined as the 
 electromotive force that, steadily applied to a conductor 
 whose resistance is i ohm, will produce a current of 
 i ampere. Its symbol is E or e. 
 
ELEMENTS OF THE DYNAMO. 15 
 
 Ohm's law, to be analyzed later, gives the relation 
 which exists between /, E, and R, in any part of a cir- 
 cuit, and in words states that the rate of transfer of 
 electricity in any part of a circuit is directly proportional 
 to the difference of potential between the points taken, 
 and is inversely proportional to the resistance between 
 these points. Since E = I A\ the "drop" or loss of 
 potential along a line can be found if /and /?are known. 
 
 The distinction between current, /, and quantity of elec- 
 tricity (denoted by 0, should be clearly grasped. The 
 clerk who said that his company would furnish lighting 
 at one cent per ampere was confusing current with quan- 
 tity, and was as much in error as the small boy would be, 
 should he, in answer to the question as to how/V/r he had 
 walked, reply,*' Four miles an hour." The unit of current, 
 the ampere, we have defined; the unit of quantity is the 
 Coulomb; and is the quantity of electricity which passes any 
 given cross-section of a conductor, when a current of i 
 ampere flows for i second, 1/2 ampere for 2 seconds, 
 or 2 amperes flow for 1/2 second, and so on. The prac- 
 tical unit of quantity is the Ampere-Hour, and since the 
 hour contains 3,600 seconds, i ampere-hour is equiva- 
 lent to 3,600 coulombs. 
 
 Electrical power, or rate of doing work, depends upon 
 7 and , and is equal to their product. This may be 
 illustrated by the analogy of water pressure. The head 
 of water corresponds to /, and the pressure in the pipe 
 to E, and just as the power of the water in the pipe 
 equals the head of water x pressure, so electrical power 
 = E X /. Since E I R the expression for watts be- 
 comes watts = / X I R = / a R. The unit of electrical 
 power is the Watt (or Kilowatt =. 1,000 watts), and its 
 
l6 TESTING OF DYNAMOS AND MOTORS. 
 
 symbol, P, although IV is frequently used. Energy is 
 capacity for doing work, and is measured in the same 
 units. The amount of electrical energy consumed, or 
 work done, depends upon time and power; i. e., upcn the 
 number of seconds and the work done in one second. 
 The unit of power, or working rate, we have seen to 
 be the watt; the unit of time, the second; so the unit 
 of work must be (rate X time) watts x seconds, or, as 
 we may say, Watt- Seconds. The unit most used is larger 
 than this, and is the Watt-Hour. The watt second is 
 called the Joule, and i watt-hour = 3,600 joules. The 
 same distinction that has been made between the ampere 
 and the coulomb must be observed between the joule 
 and the watt. The watt is the unit of power, or is the 
 work done per unit of time, while the joule is a given 
 amount of work irrespective of the time it may take 
 to do it. 
 
 Joules Work done = Power x Time = Watts X Sec- 
 onds. The mechanical unit of work is the work done in 
 lifting i pound through i foot against gravity, and when 
 work is done at such a rate that 550 pounds are lifted 
 through i foot in i second the rate is i Horse Power, 
 which is the mechanical term analogous to watt. Now, 
 the amount of work done is the same whether 550 
 pounds are lifted i foot, or 1,100 pounds 1/2 foot, and 
 it does not matter whether the work is done in i second 
 or i hour; but to do work at the rate of i HP, the 
 equivalent of 550 pounds must be lifted through i foot in 
 i second of time. The difference between work and 
 power, watt-hours and watts, cannot be too carefully 
 noted, as any confusion is a fruitful source of error. 
 
 The values of the practical units of E, 7, and R have 
 
ELEMENTS OF THE DYNAMO. 17 
 
 been given. In establishing the units of the absolute 
 system, upon which the practical system is based, use is 
 made of the relation existing between the electric current 
 and magnetism. As early as 1819 Oersted observed that 
 a compass needle in the neighborhood of a wire carrying 
 current was influenced by it, tending to take up a posi- 
 tion at right angles to the wire. He observed also that 
 this influence was increased by using a coil of many turns; 
 if a current be made to flow through ten turns, its effect 
 upon the needle placed at the centre of the coil is ten 
 times that of one turn. Upon this fact is based the defi- 
 nition of current, as follows: Con- 
 sider a unit pole at the centre of a 
 coil of a single turn whose radius 
 is unity (see Fig. 4). The length 
 of the coil is 3.1416 x diameter 
 n d = 2 n r, expressed in centi- 
 metres. Since we assume that /= i, 
 
 i* IG, 
 
 the length of wire is 2 TC centimetres. 
 If there now be sent around the coil a current of such 
 strength as to exert upon unit pole at the centre a force 
 of i dyne for each centimetre of wire in the coil, the 
 current is said to have unit value. There are 2 n units of 
 length in the wire, so that the force exerted by the whole 
 turn is 2 n dynes. This current is the absolute unit: 
 the practical unit or International Atnpfre is i/io of this 
 current. 
 
 Before Oersted's time magnetic fields were produced by 
 means of natural magnets only, or steel magnets charged 
 from them. His experiment led to the discovery that 
 a piece of iron or steel could be magnetized by means of 
 an electric current, and to a much higher degree than by 
 
i8 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 the old method. The piece to be magnetized is placed 
 within a coil of wire, and a heavy current sent through 
 the coil. By means of Faraday's conception 01 lines of 
 force the effect is readily understood. He supposed that 
 surrounding every electric current were lines of force 
 disposed in concentric circles, as in Fig. 5. With the 
 current passing down through the plane 
 of the paper, the lines pass around the 
 wire clockwise, as indicated by a com- 
 pass needle. If instead of one there 
 be many turns, the total number of 
 lines offeree is proportionately greater. 
 The effect of introducing an iron core 
 into the coil, we have seen above. At 
 either end of such an electromagnet the field is very 
 strong (see Fig 6). Fig. 7 gives a still better form: here 
 the magnetic circuit is all of iron, except the small air 
 space, N S. The relation between the magnetizing 
 force = current X turns = ampere turns, (usually 
 designated as S i ) and the resulting number of lines, 
 3C, due to the coil in air, is as follows: 
 
 FIG. 5. 
 
 ae = 
 
 10 
 
 x = 
 
 L 
 
 x 
 
 L 
 
 where 3C is the number of lines of force per square centi- 
 metre, or strength of field due to the ampere-turns Si, ex- 
 clusive of the influence of the iron, and L is the average 
 length of the magnetic circuit. The cross-section of the 
 circuit does not enter this expression because the path is 
 air (before the ring is inserted), and since the permeability 
 of air is unity, increasing the cross-section does not in- 
 crease the capacity for lines of force, but simply decreases 
 
ELEMENTS OF THE DYNAMO. 
 
 their density. As soon as iron is introduced into the coil 
 the multiplying power of the iron becomes an important 
 factor and interests us in 
 (B, the field strength in the 
 iron, rather than in JC, that 
 in air. To find (B, we must ^: 
 know JC, or find it from the ^~ 
 above formula. For exam- 
 
 ple, if 5 and 7 centimetres ^^ 
 
 are respectively the inside FIG."^" 
 
 and outside diameters of 
 
 the ring, the average diameter will be 6 centimetres and 
 the average path, Z, for lines of force 6 n centimetres, 
 or 18.8 centimetres = Z. If the current (/') = 5 am- 
 peres, and the number of turns (S) = 50, the expression 
 for 3C becomes, 
 
 _ 1.26 X 50 X 5 l6 7 
 18.8 
 
 Now in many cases to get (B from 3C we would have 
 to know the permeability (//) of the metal in question, 
 but since all samples of good wrought 
 iron are very similar in magnetic prop- 
 erties, we can assume that Table I. gives 
 close enough results, and from it derive 
 the value of (B corresponding to any 
 given value of JC. 
 
 Thus for 3C = 16.7, we see (B = 13,900. 
 Where the magnetic circuit is not all 
 FJG 7 iron, but is part air, part cast iron and 
 
 part wrought iron, as in some actual 
 machines, the 3C necessary to force the required number 
 of lines through each part is determined, and the sum of 
 
20 TESTING OF DYNAMOS AND MOTORS. 
 
 these 3C's is the Magnetomotive Force which must be 
 provided to get the desired number of lines through the 
 magnetic circuit. The magnetic circuit cannot be treated 
 as a whole because the permeability of its parts is not the 
 same. The product 6" / designates, as we know, Ampere- 
 Turns, and to secure a given magnetic effect it makes no 
 difference, except in cost of manufacture, whether i be 
 composed of many turns and small current, or of few 
 turns and large current, provided that in any case the 
 product 6* i is not changed. One ampere flowing around 
 an iron core ten times has the same magnetic effect as ten 
 amperes flowing around once. In actual designing there 
 are considerations which dictate what arrangement of 
 current and turns shall be used to obtain a desired mag- 
 netization. 
 
 If a piece of iron or steel be placed in a coil through 
 which current is made to gradually vary from zero 
 to a maximum, and back to zero, it is found that upon 
 removing the current the magnetization does not fall to 
 zero, but remains to a marked degree. This property 
 which iron and steel have of retaining magnetism is 
 called retcntivity) and the magnetism retained is known 
 as residual magnetism. Soft iron is more readily mag- 
 netized than steel, but parts with its magnetism more 
 readily also. Or we may say it is more permeable than 
 cast iron or hard steel, but is not so retentive. It is by 
 virtue of their residual magnetism that dynamo fields (to 
 not ordinarily require an initial charge each time they are 
 used. 
 
 Following on the track of Oersted, Ampere, in 1821, 
 advanced the following principles, which he proved ex- 
 perimentally: i. Parallel currents, if flowing in the same 
 
ELEMENTS OF THE DYNAMO. 21 
 
 direction, attract, and if flowing in opposite directions, 
 repel, each other. The explanation lies in the reaction of 
 the magnetic fields surrounding each current (see Figs. 8 
 and 9), and it follows from Faraday's conception of lines 
 of force oppositely directed attracting each other, that 
 lines of force cause an attraction between conductors. 
 It must be borne in mind, however, that the currents 
 exert the attractive force and not the conductors; the 
 latter are simply dragged along with the currents. 
 
 In 1831 Faraday performed an experiment which sealed 
 the doom of all chemical sources of electricity as applied 
 to light and power. He discovered Electromagnetic Indue- 
 tion. If a conductor forming part of a closed gal- 
 vanometer circuit be made to cut the lines of force of a 
 magnetic field, the conductor becomes the seat of a cur- 
 rent as indicated by the swing of the galvanometer needle. 
 The current is generated only while the conductor is in 
 motion, and its direction depends upon that of the motion. 
 It is evident that the cause of the current is the motion 
 of the conductor, and the sole condition necessary is that 
 the conductor cut lines of force. Even if the circuit be 
 open there is still present the force tending to produce a 
 current. This force, then, fulfills the definition of an 
 electromotive force in that it causes or tends to cause a 
 flow of electric current. The following statement is then 
 true : There is an E. M. F. present in every conductor cut- 
 ting lines of force. Furthermore, the value of this 
 E. M. F. will depend upon the number of lines cut per 
 unit of time, and this, In turn, upon the length of the 
 conductor, the field strength, and the rate of motion. 
 
 Let the circuit be closed through the unit resistance 
 already given, and let the conductor velocity be raised 
 
22 TESTING OF DYNAMOS AND MOTORS. 
 
 until a current of one ampere flows, the E. M. F. will 
 now be one volt. The volt may now be defined in terms 
 of the number of lines of force cut per second, and is in 
 fact the E. M. F. generated by a conductor which cuts 
 100,000,000 lines per second. Observe that the length 
 of the conductor (its resistance being supposed zero), 
 does not enter, for according as it is long or short the 
 velocity is correspondingly less or greater, the area 
 swept through and the number of lines cut remaining the 
 same. The absolute or C. G. S. unit of E. M. F. is that 
 produced in a conductor cutting one line of force per 
 second. Having now the absolute units of current and 
 E. M. F., Ohm's law gives the corresponding absolute 
 unit of resistance, 
 
 and is that resistance through which unit E. M. F. will 
 urge unit current. 
 
 Moving a conductor in a magnetic field is one way of 
 obtaining induction. In this method, the conductor is 
 the active agent and cuts the lines of force. A second 
 method provides a moving field, and here the lines of 
 force are active and cut the conductor. If two coils be 
 placed side by side, and a variable or alternating current 
 be sent through one of them, there will be produced a 
 variable field growing stronger and extending further out 
 as the current strengthens, and closing in on itself as 
 the current weakens. This variation in field strength 
 causes lines of force to move and cut the conductors of 
 the second coil, with a resulting current therein if its 
 circuit be closed. 
 
ELEMENTS OF THE DYNAMO. 
 
 FIG. 8. 
 
 the 
 
 The cutting of one conductor by lines of force due to 
 current in another conductor constitutes what is called 
 Mutual Induction, and forms 
 the underlying principle up- 
 on which is based all trans- 
 former construction. 
 
 The lines-of force also cut 
 the first coil, tending to pro- 
 duce in it a current opposed 
 to that already flowing; this 
 
 effect is known as self-induction^ and to it is due 
 spark seen upon breaking a field circuit. 
 
 Passing from the abstract theory to its more practical 
 application, the experiment of the disc can now be readily 
 explained. 
 
 Let us consider the disc as made up of conductors 
 radiating from the axle outward. It is seen at once that 
 we have here conductors cutting lines of force, and as 
 the conductors pass under the brush (Fig. i) the circuit 
 is closed and a flow of current results. 
 The maximum current would be obtained by using a 
 
 brush long enough 
 to include all the 
 conductors in the 
 field, or what would 
 be the same in ef- 
 fect, have several 
 brushes arranged 
 FIG. 9. around the periph- 
 
 ery of the disc and 
 
 connect them together. It may be noted that while this 
 explains the action of the wheel, it does not answer the 
 
2 4 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 Y' 
 
 FIG. 
 
 fundamental question as to why an E. M. F. is produced 
 when a conductor moves in a magnetic field. This ques- 
 tion has never been answered. 
 
 We have now all the elements of 
 a dynamo, and need only to assemble 
 them. In the first place there must 
 be a magnetic field. This the field 
 coils and frame provide, for the fields 
 are simply powerful electromagnets 
 whose lines of force flow around the 
 frame and across the path of the ar- 
 mature wires. To secure high per- 
 meability the cores C, C', Fig. 10, are 
 made of wrought iron or soft steel; 
 the yoke Y Y' is preferably of the 
 same material. P and P' t if of cast iron, are of relatively 
 larger cross-section, for since cast iron is less permeable, 
 there must be more square centimetres for the given 
 number of lines to pass through than in the wrought iron 
 cores. In passing from a medium of high to one of lower 
 permeability, there is a tendency to thronle the lines of 
 force at the joint; to minimize this 
 effect, some makers flange the good 
 conductor, as in Fig. IT ; others prefer 
 to set it in as in Fig. 12. In either 
 case the idea is to increase the cross- 
 section of the joint. To decrease 
 magnetic leakage and other losses in- 
 herent in iron, it is customary to build 
 the armature core of soft sheet iron, 
 bility of the metal part of 
 
 FIG. ii. 
 
 The permea- 
 the circuit depends upon 
 the degree of saturation to which the metal is worked. 
 
ELEMENTS OF THE DYNAMO. 25 
 
 Under the working conditions of present-day machines, 
 the number of lines per square centimetre cross-section, 
 /. e. , the induction (B, is about 16,000 in wrought iron, corre- 
 sponding to a permeability of about 320. This means 
 that the permeability of the metal 
 part of the magnetic circuit, when it 
 is carrying the necessary number of 
 lines, is 320 times that of air. By far 
 the greater portion of the magnetiz- 
 ing force, JC, is employed in forcing 
 
 the lines across the air gap. When 
 
 , . FIG. 12. 
 
 05 = 16,000 in wrought iron, as above, 
 
 the same ampere-turns would send the same magnetism 
 through 320 inches of iron as through i inch of air. 
 
 The next requirement is a set of conductors to cut the 
 lines of force of the field. The armature fulfills this 
 requirement. On the surface of an iron cylinder are 
 placed conductors which cut the lines twice in every 
 revolution of the cylinder. Let a, b, c y d. Fig. 13, be a 
 coil revolving in a magnetic field; all the wires in a, b y 
 cut upward through the field, and their several E. M. 
 Fs. are in the same direction from a to b. The side 
 c, d) however, cuts downward, and its resulting E. M. 
 Fs. are in the direction c to </, opposed to those in </, b, 
 as regards a point in space, but concurring as regards the 
 total circuit a, b, c, d; i. c., all the conductors of the coil 
 are in series and the total E. M. F. is the sum of the 
 separate E. M. Fs. in each wire. When the plane of the 
 coil is vertical, the motion of the conductors is parallel to 
 the direction of the lines, and the conductors slide along 
 rather than cut the lines. The E. M. F. is therefore 
 zero when the plane of the coil is vertical. In the hori- 
 
26 TESTING OF DYNAMOS AND MOTORS. 
 
 zontal position the number of lines cut is a maximum, 
 and the E. M. F. is also a maximum. The E. M. F. is 
 reversed in each side of the coil as it crosses the vertical 
 plane, for here one side ceases to cut up and commences 
 to cut down, and the other side -vice versa. 
 
 If the coil be a closed circuit it will carry a current 
 which is half the time in one direction and half in the 
 other, constituting what is called an alternating current. 
 
 If instead of closing the armature circuit by bringing 
 its -f- and ends together, we bring them out to col- 
 lector rings to which are attached also the terminals of an 
 outside circuit, there will be in this circuit an alternating 
 E. M. F., the number of single alternations per second 
 being the same as the number of poles multiplied by 
 the number of revolutions per second. A set of such 
 coils properly disposed and connected to collector rings 
 on the shaft constitute the ordinary alternating current 
 machine. If it is desired to have the line E. M. F. 
 direct, /. e., always in the same direction, there must be 
 used a special device called a commutator. In Fig. 13, 
 suppose that during the first half revolution the line 
 E. M. F. and current are in the direction of the arrows: 
 unless something is done to prevent it the E and / will 
 
ELEMENTS OF THE DYNAMO. 2"J 
 
 reverse in the second half of the revolution. If, how- 
 ever, as the coil passes through the zero position the 
 terminals of the coil be reversed with respect to the 
 brushes, this will keep the end of the coil which corre- 
 sponds to the right side of the armature always in contact 
 with brush i, and the same side of the line; and the end 
 of the coil which corresponds to the left side in contact 
 with brush 2, thus keeping the E. M. F. on the line 
 always in the same direction. Effecting this change is, 
 as the name implies, the duty of the commutator, and 
 the fundamental distinction between the direct and the 
 alternating current machine is that the former has a 
 commutator, while the latter has not. 
 
 On direct current machines, brushes are set with due 
 regard to the neutral or non-sparking points, and occupy 
 a position midway between pole pieces, unless fur 
 mechanical reasons the armature connecting wires are 
 given a lead so as to bring the brushes elsewhere. On 
 bipolars these points are diametrically opposite, but on 
 multipolars the angle to be included between brushes 
 equals the quotient obtained by dividing 360, the cir- 
 cumference of the armature, by the number of poles. 
 On an alternator, the brushes are placed in any position 
 easy of access. 
 
 The principle of the production of an E. M. F. by con- 
 ductors cutting lines by force was known twenty years 
 before the construction of a machine that could be fairly 
 said to resemble the modern dynamo. In 1856 Siemens 
 of Berlin patented his shuttle-wound drum armature. 
 The winding consisted of a single coil connected to a two- 
 part commutator. Its action can be followed by the aid 
 of Fig. 14. We learned above that when the conductors 
 
28 TESTING OF DYNAMOS AND MOTORS. 
 
 slide along the lines of force they give rise to no E. M. F. , 
 or the E. M. F. is o, and this value corresponds to posi- 
 tion a in Fig. 14. Now, as the conductor leaves the 
 neutral field, it becomes the seat of an E. M. F. which 
 increases and becomes greatest when the conductors cut 
 the lines at right angles, and this condition is represented 
 by point b. With each revolution the E. M. F. rises to 
 a maximum, falls to zero, rises to an opposite maximum, 
 when the coil cuts the lines upward, and again falls to zero. 
 In Fig. 14 let a e represent the time of one revolution. 
 
 6 
 
 Then one-quarter of the length a e will represent one- 
 quarter revolution. Starting at a with the coil at the 
 neutral line, the E. M. F. rises to a positive maximum b 
 in one-quarter revolution, falls to zero at <r, rises to a 
 negative maximum at d (i. e., E. M. F. is reversed), and 
 again returns to zero after one complete revolution. Fig. 
 15 shows the effect of using the commutator. The part 
 of the curve c d e has been reversed so that both maxima 
 are of the same polarity, in the outside circuit, and now 
 occur at b and d. The internal armature changes remain 
 as before, but the character of the line E. M. F. has been 
 changed from alternating to pulsating. 
 
 The defect of the Siemens machine is that its E. M. F. 
 
ELEMENTS OF THE DYNAMO. 29 
 
 is too unsteady. But if a second coil be placed at right 
 angles to the first, it will have its greatest E. M. F. when 
 that of the first coil is zero, the two coils being always 
 just a quarter revolution apart. These facts are shown 
 in the dotted line of Fig. 15. The E. M. F. in each coil 
 still fluctuates between zero and a maximum, but the line 
 E. M. F. does not fall below ^, though the fluctuation 
 is still marked. The resultant E. M. F. is shown by the 
 curve Ab B b' etc. Each additional coil narrows the 
 range of variation, so that although each coil has its rise 
 
 FIG. 15. 
 
 and fall of E. M. F., the E. M. F. at the brushes is prac- 
 tically constant. 
 
 E. M. F. is dependent upon speed, number of armature 
 conductors, and field strength. Increase any of these 
 and the E. M. F. is increased. Constancy of speed de- 
 pends upon the engine (and perhaps belt or clutch), and 
 can be closely regulated from no load to full load. The 
 number of conductors on a given armature is fixed. 
 The production and maintenance of the magnetic field is 
 the remaining factor. There are four distinct methods 
 of excitation, and a fifth, which is a combination of two 
 others, i. Excitation from permanent magnets; 2. so- 
 called separate excitation; 3. series; 4. shunt; 5. com- 
 pound (C. IV.) i. f.j part series, part shunt excitation. 
 
30 TESTING OF DYNAMOS AND MOTORS. 
 
 The first method has its only present practical appli- 
 cation in the familiar magneto machine, provided with a 
 steel horseshoe magnet. The second method consists in 
 connecting the field winding to some external source of 
 supply batteries, or to another machine. Batteries are 
 out of the question, and to presuppose another machine 
 is to again raise the question of excitation. Separate 
 excitation is adapted to testing-room practice on account 
 
 of its steadiness and indepen- 
 
 /^\ ^--^-^ dence. It also meets the re- 
 
 \^s UUU(j quirements of special cases 
 
 outside, and is most seen 
 where machines feed into the 
 same ^ bars; here all fields 
 FIG. 16. connect directly to the bars, 
 
 and each machine may be re- 
 garded as separately excited till its own switch is closed. 
 The third method is shown in Fig. 16. Armature, 
 field, and line are in series, and the total current passes 
 through the exciting coils. But few turns are therefore 
 needed to produce the required ampere-turns, and the 
 coils are of heavy wire -to minimize losses incidental to 
 all wires carrying current. 
 
 Shunt machines form the fourth type. As the name 
 implies, the field -winding is a shunt to something, and 
 in fact does shunt the line. On isolated machines the 
 field connection is made below the line switch, so that 
 upon opening the latter the field is not broken. On 
 starting up, a slight E. M. F. exists, due to the residual 
 field, and a small current therefore flows around the field 
 coils, increasing their magnetization; this in turn raises 
 the E. M. F., till finally it reaches its normal value for the 
 
 \ 
 
ELEMENTS OF THE DYNAMO. 31 
 
 given speed. The final field current equals the E. M. F. 
 at the terminals, divided by the resistance of the field 
 circuit; anything which lowers the E. M. F. lowers the 
 field ampere-turns, this reacting again on the E. M. F. 
 We have learned that current /, passing through a 
 resistance, A*, expends energy at the rate of P R watts 
 per second, and that the loss of potential is I R. This 
 loss of potential takes place in the armature as well as in 
 the rest of the circuit. As the load increases the arma- 
 ture loss increases, the terminal voltage decreases, and 
 with it the field current, whose weakening causes a 
 further decrease in E. M. F. At full load the voltage is 
 then lowest, at no load it is highest, while on a series 
 machine the reverse is true. 
 
 The compound-wound dynamo is the shunt and series 
 machine in one; on open circuit it is shunt excited, 
 while on full load the series winding has its greatest 
 effect. Compound- wound machines have two methods 
 of connecting, viz.: long shunt and short shunt. In 
 the long-shunt connection, as given in Fig. 17, the 
 shunt winding is connected 
 to the machine's terminals 
 and hence outside the series 
 windings. When the line 
 switch is open the E. M. F. 
 is that due to the shunt field 
 plus a small amount due to 
 the series field. In the short- _. 
 
 shunt connections of Fig. 18 
 
 the shunt winding excludes the series, and the latter is 
 entirely idle on open circuit. The two types require 
 slightly different calculations in designing, but are identi- 
 
 W} 
 
32 TESTING OF DYNAMOS AND MOTORS. 
 
 cal in action. If the coils are properly proportioned the 
 gain in the series field, as the load goes on, balances the 
 loss in the shunt field, and from no load up to the point 
 ," / o^~> 1 where the field cores become 
 
 saturated, the terminal E. M. F. 
 remains constant. Any further 
 increase in load will lower the 
 E. M. F., because not only does 
 armature reaction set in, but the 
 increased current raises the 
 / 7? loss, without appreciably 
 strengthening the already saturated field. 
 
 There are three sources of loss in a dynamo or motor, 
 viz., mechanical, electrical, and magnetic. The first 
 comprises the frictional losses in the bearings, brushes, 
 and^elts. In the second are found the I*R losses of 
 armature and fields; their value is easily calculable, and 
 in well-designed machines is not large. The third source 
 of loss is in the magnetic circuit, and the component 
 factors are not so easily separated. In general, magnetic 
 losses are due to a change in the quantity or direction of 
 magnetic flow setting up molecular friction or inducing 
 currents; in either case the product is heat, and rep- 
 resents lost energy. Thus the field current fluctuates 
 within narrow limits, and the iron accordingly gains or 
 loses magnetization, molecular friction ensues, and heat 
 is produced. In the pole-heads this action is strongest, 
 and is most marked in machines having lug armatures. 
 The lugs are concentrators of lines of force, and sweep 
 across the polar faces very much the same as one might 
 do with a brush. The effect can even be detected with 
 the hand by feeling the polar horns on a machine in ser- 
 
ELEMENTS OF THE DYNAMO. 
 
 33 
 
 FIG. 19. 
 
 vice. On a dynamo the leading horns, while on a motor 
 the trailing horns, become heated. The most serious 
 magnetic loss is due to eddy currents in the armature core 
 and pole-heads. Considering the armature to be made up 
 of concentric cylinders of metal, as indicated in Fig. 19, 
 it is virtually composed of layers of con- 
 ductors on closed circuit, revolving in a 
 powerful magnetic field. Were the ar- 
 mature core solid, in it would flow a cur- 
 rent giving rise to much heating of the 
 core itself. Fortunately this trouble 
 can be avoided by building up the ar- 
 mature of thin sheet iron plates. The 
 resistance of so many joints prevents any serious flow of 
 current. _ t Another magnetic loss is due to a modification 
 of the magnetic field by the armature. The armature is 
 a huge electromagnet with its poles at the neutral line, 
 and therefore with its poles approxi- 
 mately at right angles to those of the 
 field. The armature tends to send its 
 lines of force up and down like the 
 line a b, Fig. 20, while the field 
 would send its own straight across, 
 like the line c d. Either alone would 
 have its way, but the resultant of the 
 forces is in the direction of the dotted 
 line f /, and this marks the diameter on which are set 
 the brushes of a bipolar machine. This armature effect is 
 known as cross-induction, and its effect upon the position of 
 the neutral point is plainly seen in the necessity of shifting 
 the brushes on most machines. Now the shifting of the 
 neutral line is in such a direction that like poles of field 
 
 F|G 20 
 
34 TESTING OF DYNAMOS AND MOTORS. 
 
 and armature are brought nearer together, and the arma- 
 ture ampere-turns strive to force a flow of magnetism 
 through the same magnetic circuit as the field ampere- 
 turns, but in the opposite direction. Hence the arma- 
 ture exerts a demagnetizing effect, with the result that 
 the field current must be increased to make it up, and 
 this means an additional P R loss in the field wind- 
 ing. This effect is called back induction, and in machine 
 design allowance is made for it by providing additional 
 field winding. The lamination relieves the magnetic 
 troubles as far as they can be relieved. In alternators 
 the pole-heads are laminated as well as the armature 
 cores. 
 
 The extent to which these many losses are done away 
 with determines the Efficiency of the machine. If we could 
 get out of a machine all the work which we put into it, 
 we could say the machine had an efficiency of one hun- 
 dred per cent. But such a thing is impossible, as in any 
 machine part of the power given to it is always wasted in 
 friction, and in electrical machines we must add to this 
 electrical and magnetic losses. Of the energy given to 
 an electrical machine we can reclaim only a fraction, 
 and the value of this fraction expresses the efficiency. 
 The term efficiency is accepted in either of two senses 
 Electrical Efficiency or Commercial Efficiency. The 
 electrical efficiency is gotten by dividing the electrical 
 energy available by the total electrical energy generated 
 by the machine, and takes into account only the 7 2 R 
 losses. The commercial efficiency is had by dividing the 
 total energy gotten out of the machine by the total energy 
 put into it, and includes all the various losses. Calling 
 L and 7 L the E. M. F. and current respectively of the 
 
ELEMENTS OF THE DYNAMO. 35 
 
 line, and r and 7 T the total E. M. F. and current of the 
 machine, the expression for electrical efficiency is 
 
 Elect. E. = ^ L/I - 
 
 or watts produced divided by watts taken out. The special 
 form which this expression takes depends upon the type 
 of the machine. The formula for commercial efficiency is 
 
 Com. Eff. = y* , 
 
 where w is the energy taken out electrically and' \V 
 is the energy put in mechanically by the steam engine. 
 To compare them they must be reduced to the same 
 units. 
 
CHAPTER II. 
 
 ELEMENTS OF THE MOTOR. 
 
 IF the terminals of a dynamo be connected to a 
 source of current, the machine becomes a motor, trans- 
 forming the electrical energy drawn from the line into 
 mechanical energy at the pulley. Motors are classi- 
 fied in the same manner as are dynamos, and the same 
 general principles of dynamo design and regulation 
 enable us to anticipate and explain motor action under 
 like conditions. The fundamental fact which makes 
 electric motors possible, is the following: If a conductor 
 carrying a current be placed in a magnetic field, it will be 
 impelled in such a direction that the E. M. F. induced in it by 
 this motion will oppose the flow of current already existing 
 in the conductor. This law states that there will be 
 motion, and also gives the direction of this motion. The 
 conditions to be fulfilled are then, a magnetic field, con- 
 ductors in this field, carrying a current and free to 
 move. The same machine that is used as a dynamo fills 
 these requirements of the motor. Let us take a machine 
 with separately excited or permanent fields, and investi- 
 gate its action. If the brushes be connected to a battery 
 a current flows through the armature. Let N and 6" 
 (Fig. 21) be the fields, and A, the armature. With 
 fields the same as when the machine is used as a 
 dynamo, let the positive brush (*. e., the positive 
 
ELEMENTS OF THE MOTOR. 
 
 37 
 
 dynamo brush) be connected to the positive pole of the 
 battery. The direction of the current will then be the 
 reverse to what it was in the dynamo, and the direction 
 of rotation of the armature the same as that of the 
 dynamo, for the E. M. F. thus induced will tend to set 
 up a current opposed 
 to the existing one. 
 If, then, a machine be 
 so connected as to 
 have its fields of a 
 given polarity, and its 
 brushes of constant 
 polarity, /. f. , the 
 same brush always 
 positive on the dy- Fig. 21. 
 
 namo, and when on 
 
 a motor connected to the positive side of the circuit 
 the direction of rotation is the same whatever be the 
 machine's nature. In other words, if in the two cases 
 the field current flow in the same direction while the 
 armature curfent flows in opposite directions, the direc- 
 tion of rotation will, in the motor, be the same as in the 
 dynamo. This explains why shunt machines run in the 
 same, but series machines in opposite directions, when 
 their nature is changed but the connections left the 
 same. 
 
 The E. M. F. to which we have referred as oppos- 
 ing the existing flow of armature current, is called the 
 Counter E. M. F. y and is of fundamental importance in 
 motor theory. It is a source of confusion, unless properly 
 understood, and some misguided inventors have vainly 
 endeavored to devise motors without any C. E. M. F., 
 
38 TESTING OF DYNAMOS AND MOTORS. 
 
 supposing its presence to be a detriment to the machine. 
 A very little consideration shows that whether a help or 
 a hindrance, its presence is unavoidable, and an arma- 
 ture wound to have no C. E. M. F., would have no power 
 of motion; for, to deprive an armature of its counter is 
 to deprive it of its ability to generate voltage when run 
 as a dynamo in a magnetic field, its C. E. M. F. being 
 nothing other than the dynamo property of a motor. To 
 construct such an armature its conductors must be so 
 wound that any tendency of one conductor to generate an 
 E. M. F. is met by the tendency of some other conductor 
 with which it is in series to generate an equal but oppo- 
 site E. M. F. To run such an armature as a motor would 
 be impossible, because the flow of current through it 
 would be so disposed that one-half the conductors would 
 attract a pole piece, the other half repel it; and the 
 equal but opposite tendencies to rotate the armature 
 would neutralize each other and there would be no 
 motion at all. In every conductor moving in a magnetic 
 field an E. M. F. is produced; that this E. M. F. must 
 oppose the motor current, can be easily see"n, for suppose 
 it to assist the current: as the speed and induced E. M. F. 
 increased, the current and, with it, the power would 
 increase also. That is to say, the motor would con- 
 tribute to its own driving power, and would do the same 
 work at the pulley with less and less demand upon the 
 line, and finally, the line supply might be. dispensed with 
 altogether. This is absurd, and theory as well as experi- 
 ence indicates that the induced E. M. F. will oppose the 
 impressed or line E. M. F. Furthermore, as we shall 
 learn, the higher this C. E. M. F. becomes, the greater 
 is the efficiency of the motor. 
 
ELEMENTS OF THE MOTOR. 39 
 
 The power given to the motor equals the product of 
 the armature current and the impressed E. M. F. or, 
 
 Watts consumed = E / a . 
 
 This is divided into two parts: that which is wasted as 
 heat in the armature, and that which is transformed into 
 mechanical power, and does work at the pulley. We 
 may, then, write: ll'atts consumed -- watts transformed 
 -f- watts wasted in heat. The heat waste is due to mag- 
 netic and electric losses, but until we take up commercial 
 efficiency the latter only will be considered, and these in 
 the armature alone. The electric loss we know to equal 
 /"*' / = / x the E. M. F. necessary to urge the given cur- 
 rent through the ohmic resistance t\ of the armature, 
 and since the total watts consumed = E / a , the watts 
 transformed must equal the difference between the two, 
 or, in other words, is the product of the current by that 
 part of the impressed E. M. F. not expended in overcom- 
 ing ohmic resistance. We may then write, 
 
 E / a = watts transformed -\- / 9 a r tt , (1) 
 
 whence, 
 
 Watts transformed = Et\ /* a r a = / a (E i a r a ), (2) 
 Now, 
 
 E - e 
 
 '=-7' 
 
 where e, is the C. E. M. F. ; clearing of fractions we get 
 / a r a = E e, and substituting this in the expression for 
 watts transformed, we have, 
 
 Watts transformed = / a (E E + e) = / a *, (3) 
 which shows that (barring other losses) the watts 
 transformed into mechanical power equal the product 
 of the current and the C. E. M. F. The equation 
 E e ; a r u shows how far the C. E. M. F. falls below 
 
TESTING OF DYNAMOS AND MOTORS. 
 
 the impressed, and that the difference is smaller the lower 
 the armature resistance. If it were possible to construct 
 an armature of zero resistance, e would equal , and for 
 all loads the electrical efficiency would be 100 per cent. 
 The expression for the electrical efficiency is 
 
 watts transformed e z a e ,^. 
 
 ^ e ~ watts consumed ~ ~E7 & ~ ~E* 
 
 showing the efficiency to increase as the C. E. M. F. 
 
 approaches the impressed. Two questions here arise 
 
 (a) With what current does a motor do most work in 
 
 d 
 
 FIG. 22. 
 
 least time? (b) With what current does it work most 
 efficiently? The first may be solved as follows: Sub- 
 stituting W for watts transformed, in equation (2), we 
 have, 
 
 whence, 
 
 E 
 
 (5) 
 
 and the question is, for what value of i a will Wbe a max- 
 imum. Let us represent the relation between / a and W 
 by a diagram, Fig. 22, in which the values of Fare laid 
 off on the vertical scale, and the corresponding values 
 
ELEMENTS OF THE MOTOR. 4! 
 
 of / a on the horizontal scale. If in equation (5) we 
 begin at zero, and give gradually increasing values to 
 //', and, with the knowledge of E and ;- a , work out the 
 corresponding values of / a , the points on 4, and \V, will, 
 if projected as indicated by the dotted lines, give a series 
 of intersections whose path will be the curve a c b. For 
 every value of W there will be two values of / a . "\Vhen 
 IV = o, equation (5) becomes, 
 
 and the two values of / a are, 
 
 2 E _ F^ 
 /n ~ 2 r. - rl 
 and 
 
 E- E 
 
 'n = - - = O. 
 
 The o value corresponds to open circuit, and is repre- 
 sented by the point a\ the value 
 
 E_ 
 r* 
 
 means that the armature is blocked so that it cannot 
 turn, and is represented by the point b. At c the two 
 values of JFare identical, and here we find the maximum 
 value for W. This condition is fulfilled when 
 
 and from this we get by equation (5) 
 
 E 
 
42 TESTING OF DYNAMOS AND MOTORS. 
 
 or one-half its maximum value. But 
 
 E-e 
 
 ' = > 
 
 therefore 
 
 _E_ _ E-e 
 
 2~^ " r * ' 
 
 and e = 1/2 E, so that the expression for electrical effi- 
 ciency becomes 
 
 1/2 E f p\ 
 
 V* = J -- = 5^ (6) 
 
 That is to say, a motor works fastest at an electrical 
 efficiency of 50^. As a rule motors are designed for 
 a much higher efficiency than 50 $, and to run them at so 
 low a figure might require excessive current overload. 
 For a long time this efficiency of maximum activity was 
 mistaken for the highest attainable efficiency, and it was 
 declared that motor efficiency could not exceed this 
 value. This is manifestly an error, as electrical efficien- 
 cies of 90 f c and 95 % are commonly attained in motors 
 of to-day. On the other hand, as the efficiency in- 
 creases, work per minute decreases, and to meet the 
 demands of traffic it is found more profitable to run at 
 higher outputs per minute and lower efficiencies (80 $ 
 to 90 #). 
 
 Thus far we have had to do with electrical efficiency 
 only, and have confined our attention to/ 2 ./? losses. The 
 commercial efficiency takes into account frictional and 
 other losses, and is of more practical importance. It can 
 be found as follows: Ascertain at what speed the arma- 
 ture turns when loaded, and call the current flowing, i & . 
 Next determine the current, *' a , necessary to turn the 
 
ELEMENTS OF THE MOTOR. 43 
 
 armature free and at the same speed. With load, the 
 current doing useful work is, / a /"' a , and the expression 
 for commercial efficiency becomes, 
 
 - 
 
 VC - 
 
 Remembering that, 
 
 E -e 
 
 where r a is the armature resistance, we get, 
 
 E - e ., 
 
 < >\ ~ l a _ e E - e* - c r a / \ 
 7/c - E ~~E^rr~ E*-tE 
 
 e (E-e - r a /' a ) 
 
 E (E - c) 
 The commercial efficiency for maximum activity, is, 
 
 The maximum commercial efficiency is a more complex 
 question, as it involves/", and r a , as well as the C. E. M. F. 
 If in equation (7), r a and E be given, the value of / a , cor- 
 responding to maximum efficiency, can be found.* The 
 condition is that 
 
 E - e = ^^T* (8) 
 
 or that _ 
 
 e = E - 
 
 To illustrate: a motor has an impressed E. M. F. (E) 
 
 * For a full discussion, see Kapp's Electrical Transmission of Energy, 
 4th ed., pages 155 to 157. 
 
44 TESTING OF DYNAMOS AND MOTORS. 
 
 of no volts. Running free, the armature takes (/' a ) 5 
 amperes. Armature resistance (r a ) = .35 ohm. What 
 is the maximum commercial efficiency? From equa- 
 tion (8) 
 
 no e ^.35 X no x 5 
 whence, e 96 volts = C. E. M. F. 
 Now, 
 
 E - e 14 
 
 / a = - = = 40 amperes. 
 
 'a -35 
 
 Therefore the maximum electrical efficiency is 
 c 96 
 
 and maximum commercial efficiency, by equation (7), is 
 
 /7 C = 76.6 %. 
 While the HP available at the pulley is, 
 
 HP = 5.1. 
 
 The reason the efficiency is so low is that the armature 
 resistance is so high; let r a be lowered to .2 ohm, and 
 we have 
 
 HP = 7.1. 
 
 Torque. It is a familiar fact that between unlike poles 
 of neighboring magnets a force of attraction exists. If 
 the magnets be crossed at right angles to each other, and 
 one of them be free to rotate, it will do so until the 
 unlike poles are as near together as possible. The 
 relation between the armature and fields of a motor is 
 similar to this. The armature is an electromagnet with 
 
ELEMENTS OF THE MOTOR. 45 
 
 its poles midway between those of the field, resulting 
 in an attraction of unlike poles of armature and field. 
 As soon as a coil reaches a position where it would natu- 
 rally remain at rest, its current is reversed, its former 
 relative polarity restored, and continuous rotation ensues. 
 The force with which the armature tends to turn is meas- 
 ured by the product of the pole strengths of field and 
 armature, and will vary as this product varies. If we can 
 assume the field poles to remain constant, the turning 
 force will vary as the armature current varies. This all- 
 important turning force is called Torque, and its value 
 has been mathematically proven to be, 
 
 where AT is the total number of lines of force crossing the 
 air gap, C is the number of conductors on the armature, 
 and / a equals the current in each conductor or one-half the 
 current in the armature. In so many words torque or 
 twisting force is nothing more than a force acting with a 
 leverage in the same way as a man, with the force of his 
 own weight, can, with a crowbar as a lever, move a weight 
 many times greater than his own. In the case of an 
 armature the force acting is the force of attraction 
 between the armature current and the lines of force of 
 the field, and the leverage with which it acts is the dis- 
 tance from the centre of the armature core to the con- 
 ductors. If the force of attraction is measured by the 
 product of field lines and armature current, it is obvious 
 that increasing either of these will increase the turning 
 power of the armature : this the above formula tells us, 
 for if N or C or / a is increased, the value of the frac- 
 
46 TESTING OF DYNAMOS AND MOTORS. 
 
 tion is also. The work put into an armature can 
 be expressed as equal to the speed of the arma- 
 ture multiplied by its torque, or W 6.28 x n X T. 
 If the amount of work, IV, is kept the same, as the speed 
 rises the torque falls, and, could the armature attain the 
 impossibility of having e =. E, there would be no current, 
 hence no torque; there would be no work put into the 
 motor and none taken out. When the current is first 
 turned into a motor the torque is greatest, for torque 
 depends upon the current, and just before starting, when 
 the C. E. M. F. is zero, the current is greatest for a 
 given mechanical load. When running free the arma- 
 ture speed is highest, and e approaches, but never reaches, 
 E. One sometimes hears the statement that the C. E. 
 M. F. could only equal the impressed were the speed 
 infinite: this is an error, for with any field strength an 
 infinite armature speed would produce an infinite E.M.F. 
 The formula for the E. M. F. of a dynamo is E = N C 
 n -4- 100,000,000. Since C. E. M. F. is only the dynamo 
 E. M. F. of a motor for that speed and excitation, this 
 expression is good for the C. E. M. F. of a motor, and 
 leaving off the 100,000,000, which serves to reduce 
 C. G. S. units to volts, we can write as the C. E. M. F. 
 of any motor e N C n. Dividing both sides of this 
 equation by the same term does not alter its value, so we 
 divide by N C to get n on one side alone ; 
 
 leaving 
 
 n = 
 
ELEMENTS OF THE MOTOR. 47 
 
 an expression for speed, which does not contain torque, 
 showing the torque to be independent of the speed. 
 Again, this formula shows that if N is kept the same, 
 the field a series field, or separately excited, the denomi- 
 nator of the fraction becomes a fixed number, or, 
 as we say, a constant, and the value of the fraction can 
 only be varied by varying the numerator e. In other 
 words Con any given armature is a fixed number; if we 
 fix N, then n depends upon e and varies with it in direct 
 proportion, /*. <-., the speed of any motor armature is pro- 
 portional to its C. E. M. F., and if the armature and field 
 resistances are low, ;/ is also proportional to /:, the 
 impressed or line voltage. If these resistances are high, 
 the loss of voltage through them cannot be neglected, 
 and n cannot be said to vary directly as E y but directly 
 as E less the internal loss, which difference is e, the 
 C. E. M. F. In other words if with an impressed 
 E. M. F., E 500, we get a speed ;/ = 200, a voltage of 
 1,000 will not give a speed of 400, and to know what speed 
 it will give we must assume same current value at which 
 both tests are to be run, and know the value of internal re- 
 sistance. Let / = 30 amperes, and let r 10 ohms (an 
 exaggerated case). In the first test, then, the internal 
 loss of voltage = / r = 10 X 30 = 300, and the speed of 
 200 revolutions is really due to 200 volts, or i revolution 
 per volt. In the second case the current, /, and re- 
 sistance, r, are the same as before, so the volts lost, = 
 i r, must be the same, or 300. This leaves 700 volts to 
 cause motion, and at i revolution per volt this would 
 correspond to a speed of 700 revolutions. If we assume 
 r = $y$ ohms, the loss in both tests is 100 volts, and rais- 
 ing E from 500 to 1,000 raises n from 200 to 450, while if 
 
48 TESTING OF DYNAMOS AND MOTORS. 
 
 r = i ohm, n will increase only from 200 to 412, or very 
 nearly proportionally to E. Now in the formula 
 
 anything which increases the denominator of the fraction 
 decreases the value of the fraction, /. ^., decreases n; this 
 shows us, then, that if we increase N or C, keeping e the 
 same, we will decrease ;z, and vice versa. Theory here 
 predicts what is every day practiced, viz. : that to speed 
 up a motor we weaken its field. The physical explanation 
 of this fact is, that in weakening the field the armature 
 cannot generate as high a C. E. M. F., hence a larger 
 current flows through the armature and its speed rises. 
 That weakening the field on a motor will raise the speed 
 is true of all recent types of motor, but it is a remarkable 
 fact that where a motor runs at: an efficiency of less than 
 50 $, strengthening the field will raise the speed, 
 and weakening the field lower the speed. The explana- 
 tion is as follows: For a motor to run at normal load 
 with an electric efficiency of less than 50 $, means 
 that the internal resistance must be so high, that of the 
 total E. M. F. applied to the motor terminals over half 
 is wasted in urging the current through this resistance, 
 and less than half is left to cause motion. To run a 
 motor at a certain speed requires a certain amount of 
 work, and to run it above or below this speed requires 
 more or less work respectively; now the effect of weaken- 
 ing the field is to lower the C. E. M. F. and to raise the 
 current through the armature, and this ought to raise the 
 speed, but if the resistance is high, the increased current 
 causes so large a loss in voltage that although the motor 
 draws more energy from the line than before, there is less 
 
ELEMENTS OF THE MOTOR. 49 
 
 available for motion, so that with decreased energy the 
 motor must run slower. An example will help us. Let 
 / = 30, and r = 10, and at a given field strength let there 
 be a given speed. , as before, = 500. Volts lost = 
 i r 30 x 10 = 300. Useful volts = 500 300 = 200. 
 If useful E = 200, and / = 30, the watts expended in 
 motion E (useful) x / ' = 200 x 30 = 6,000 watts. 
 Now let the field be weakened till / = 35, then, volts 
 lost = / r = 35 x 10 = 350. Useful volts = 150, and 
 useful energy = useful x t = 150 x 35 = 5 2 5 as 
 against 6,000, so the speed must fall. If r = i, then, 
 lost and useful volts at 30 amperes are 30 and 470 respec- 
 tively, and at 35 amperes, 35 and 465, and the useful 
 energy at 30 amperes is 14,100 watts, while that at 35 
 amperes is 16,275. So the effect in one case is to decrease 
 the useful energy and in the other to increase it, the 
 speed varying accordingly. Upon this fact depends the 
 action of the Thomson wattmeter. 
 
 The question of speed regulation is important in 
 motor work. Viewed from this standpoint there are 
 two general classes of motors, (i) traction and (2) sta- 
 tionary motors. In the first class a large initial torque is 
 required and variable speed under easy control. The sec- 
 ond class in almost all cases calls for constant speed under 
 all loads. Another classification regards the nature of 
 the circuit upon which the motor is to be used : (i) motors 
 on constant potential (C. P.) mains, (2) constant current 
 motors. In the first all the motors are in parallel, /. e., 
 the total current flowing divides among the motor circuits, 
 and each circuit, supplied with the full line voltage, is 
 independent of the other circuits; whereas with constant 
 current motors there is but one circuit, and the total cur- 
 
5<D TESTING OF DYNAMOS AND MOTORS. 
 
 rent passes successively out of one motor into another, s<? 
 that any unusual action on the part of one may affect all. 
 
 The usual classification is based upon the style of field 
 excitation: (i) separately excited, (2) series, (3) shunt, 
 (4) compound. The series motor is represented in street- 
 car work; the shunt motor in stationary work, where 
 small speed variations are unimportant. Compound- 
 wound motors are used where automatic speed regula- 
 tion is desired under variations of load. The behavior 
 of each type may be briefly outlined as follows: 
 
 Separately excited motors have a field independent of 
 the armature current. As the load increases so does the 
 armature current, and with it the torque, while the speed 
 gradually falls off. As the armature reaction increases it 
 may be necessary to give the brushes a negative or back- 
 ward lead, thus increasing the demagnetizing effect upon 
 the field. This weakening of the field tends to increase 
 the speed, so that such a motor with strong armature re- 
 action will regulate fairly well for speed within certain 
 limits. Whether it is necessary or not to shift the 
 brushes, the demagnetizing effect of the armature in- 
 creases as its current does, for it radiates more lines of 
 force, and care must be taken that the armature does not 
 overpower the fields and reverse the direction of rota- 
 tion. Warning of this is given by destructive sparking 
 and an abnormal increase in the current. 
 
 The series motor is adapted to railway 'sevice where it 
 is often necessary to start a heavy car on a grade. On 
 closing the circuit the field and armature currents rise 
 immediately to a value controlled by the resistance in 
 circuit, and the torque is very great. Sometimes this 
 resistance is in a separate coil alone, and sometimes the 
 
ELEMENTS OF THE MOTOR. 51 
 
 fields are in sections whose resistance, when in series, is 
 almost sufficient to limit the initial current to a safe 
 value. With fields in series their ampere-turns are 
 greatest and provide the torque so much needed at 
 starting. As the motor acquires speed the current 
 decreases, on account of the C. E. M. F., and with it 
 the torque, to which it remains very nearly propor- 
 tional. Speed regulation is obtained either by throw- 
 ing the field coils from simple series to different 
 combinations of series ~ncl multiple, as in the old Edison 
 system, or by cutting out or shunting part of the field 
 ampere-turns as in more recent types. In the method 
 of commutatecl fields a special switch is used, by the 
 turning of whose handle the different combinations are 
 effected. In the more recent methods of control, the 
 car is started with motors in series, and by steps they are 
 thrown into multiple, thereby increasing not only the 
 speed control, but economy in service a point to be 
 considered later. 
 
 Shunt motors have a wide scope of work, being very 
 generally used for isolated motor work and for shop 
 service on constant potential lighting mains. The fields 
 are so connected that they are charged before the arma- 
 ture circuit is closed. These precautions are necessary 
 because the field in shunt with the armature is of high 
 resistance and self-induction, so that not only is the 
 armature virtually a short circuit across it, but even 
 under the best conditions its induction would prevent its 
 acquiring full strength immediately. Were no precau- 
 tion taken, the switch closed across field and armature 
 at the same time, and the machine left to pick up a 
 C. E. M. F. as best it could, trouble would ensue before 
 
52 TESTING OF DYNAMOS AND MOTORS. 
 
 the rush of current could be checked. Even when the 
 fields are connected above the switch so as to give the 
 required field strength before closing the armature cir- 
 cuit, it is necessary to close the armature circuit through 
 a resistance, and then to cut the resistance out after the 
 speed is up. 
 
 Variations of load on a shunt motor fed from constant 
 potential mains produce slight speed variations, the 
 regulation being analogous to regulation for E. M. F. 
 in a shunt dynamo run at constant speed. The action is 
 as follows: the work depends upon the product of the 
 speed and torque, and the torque depends upon the product 
 of the pole strengths of armature and field. The arma- 
 ture pole strength depends upon the current, and in 
 order to increase this the C. E. M. F. must be lowered; 
 this requires a decrease in either speed or field strength. 
 As a matter of fact both effects take place. The added 
 load on the pulley slows up the armature, the C. E. M. F. 
 falls, and a larger current flows. With the increased cur- 
 rent the armature reaction is increased, and the weaken- 
 ing of the field raises the speed. The decrease in speed, 
 however, is not entirely compensated for, and the motor 
 gradually slows down. Still another effect which pre- 
 vents compensation from being complete, is that the lost 
 voltage in the armature increases, thus decreasing the 
 E. M. F. effective in producing rotation. Armatures of 
 no resistance would, aside from their own reaction, regu- 
 late for constant speed on a motor just as they would 
 for constant E. M. F. on a dynamo. Hand regulation 
 may be resorted to, either in raising the impressed 
 E. M. F., or in weakening the fields. 
 
 This gradual weakening of the field as load increases 
 
ELEMENTS OF THE MOTOR. 53 
 
 is automatically attained in the differentially connected 
 compound-wound motor. In this, the shunt winding is 
 made strong enough to always control the polarity of the 
 field, and the series winding, instead of assisting the 
 shunt, as in a compound-wound generator, opposes it. 
 The windings are so proportioned that the weakening 
 effect of the series coils just compensates for the fall in 
 speed due to increased load. By this means is secured 
 a constant speed for wide variation of load. The regula- 
 tion is perfect, however, only when the difference of 
 potential across the mains is that for which the motor 
 is designed. Such a motor must be started in the same 
 manner as a shunt motor, otherwise the strong initial 
 series field would overpower the shunt field and reverse 
 the direction of rotation. Another device which would 
 give the motor the advantage of great starting power 
 would be to temporarily reverse the series field while 
 starting, making it agree with the shunt, and after full 
 speed is attained and the armature current falls to its 
 normal value, to successively short circuit and reverse 
 the series coil, restoring it to its normal condition of 
 opposing the shunt winding. 
 
 The efficiency of the compound-wound motor is lower 
 than that of the compound-wound generator or any other 
 generator, for more energy is expended in the field to 
 produce a given magnetization, some of the lines of force 
 being always neutralized. 
 
 There are various mechanical devices for speed regula- 
 tion, none of which, however, approach the differential 
 compound-wound motor in point of merit. We have 
 seen that with increasing load, motor brushes require 
 greater and greater backward or negative lead, as 
 
54 TESTING OF DYNAMOS AND MOTORS. 
 
 opposed to the forward or positive lead of generator 
 brushes. This is so because the lead depends upon the 
 armature reaction. If the motor field-polarity and 
 direction of rotation concur with that of the dynamo, the 
 direction of the motor armature current is opposite to 
 that of the dynamo. This reverses the armature polar- 
 ity, and with it the field distortion, making the neutral 
 line recede instead of advance, and necessitating the 
 backward lead. Hence the trailing pole corners are 
 strengthened by the distortion, become the seat of eddy 
 currents, and grow warm. Again, in the motor the arma- 
 ture and whatever resisting force is attached to it is 
 pulled around by the conductors; while in a generator 
 the engine must turn the armature and pull the conduc- 
 tors around against the drag of the magnetic field. 
 
 Continuous current generators can in all cases be used 
 as motors, though some are better adapted than others. 
 Small motors cannot always be profitably used as self- 
 exciting generators, because, while in a motor the field is 
 independent of the armature, and the exciting energy 
 comes from without, and whatever amount may be 
 needed to overcome the reluctance of the magnetic 
 circuit is furnished over and above what is needed for 
 the armature, in a generator the exciting energy must be 
 supplied by the armature, and if the reluctance be high, 
 the energy demanded by the field may sxceed the total 
 output of the machine. In small motors the air gap is 
 very wide relative to the size of the machine, so that the 
 magnetic reluctance is large. 
 
 To have the same rotation, the relation of armature 
 and field polarity must in the motor be the reverse of 
 that in the generator. A series motor and generator run 
 
ELEMENTS OF THE MOTOR. 55 
 
 in opposite directions for given connections, and to have 
 them turn in the same direction, either the field or arma- 
 ture terminals must be reversed when the nature of the 
 machine is reversed. The shunt dynamo and motor run in 
 the same direction for given connections, while the direc- 
 tion of rotation of a compound-wound motor will depend 
 upon the relative strength of series and shunt windings. 
 These points will be explained elsewhere. The ease with 
 which a shunt machine may be reversed is illustrated 
 where several are in multiple on the same bus-bars. If 
 for any reason the voltage on one machine falls far below 
 that on the others, it will not only lose its own load, but 
 will absorb energy from the others and run as a motor, 
 and since the rotation is the same there will be no violent 
 demonstration. The brushes will spark some, the oppo- 
 site side of the belt will become slack, and the ammeter 
 needle will go to zero and will reverse, or rise again, 
 according as it is of the Weston or Edison type. In any 
 case, whether the machine runs as a motor or not can be 
 ascertained by strengthening its field and watching the 
 ammeter. If the current rises, the machine is a dynamo, 
 but if it falls, a motor. In the latter case the field can 
 be gradually strengthened till the machine loses its load 
 as a motor, when the ammeter will register zero, and if 
 the field is further strengthened the needle will rise, 
 showing the machine to be now acquiring load as a gen- 
 erator. It is partly on account of this liability to rever- 
 sal that shunt dynamos have been superseded by 
 compound machines on street railway service, where the 
 load fluctuations are very violent and between wide 
 limits. 
 
THE TESTING AND USE OF INSTRUMENTS 
 
THE TESTING AND USE OF INSTRUMENTS. 
 
 CHAPTER III. 
 
 OHM'S LAW. 
 OHM'S law in its original form is 
 
 where 7 is the current flowing in the conductor, E the 
 potential difference at the terminals, and R the con- 
 ductor's resistance. To better comprehend all that this 
 law, in its various forms, means, let us sec upon what 
 foundation it rests. In words it can be stated thus: /// 
 any circuit containing only ohmic resistance, the ralue 
 of the current in amperes is expressed by the fraction 
 whose numerator is the fall of potential throughout the 
 circuit, and whose denominator is the resistance therein. 
 In formulating such a law we seek for an expression 
 which shall give the relation between the current, volt- 
 age, and resistance of a circuit, and at the time of Ohm's 
 investigation it is likely that he did not know but that 
 there were other factors than E and R to modify the 
 current value. Current first attracts the attention of 
 the investigator, because it is the current's actions that 
 are manifest to us: It does the work, or is the "vehicle 
 of energy." When we see a coil, around which a current 
 
 59 
 
60 TESTING OF DYNAMOS AND MOTORS. 
 
 flows, suck up a piece of iron, the presence of a current 
 suggests itself to us rather than the voltage to which 
 it is due, for we can more readily form an idea of 
 current flow than of the voltage causing it. This may 
 have influenced Ohm in expressing his law as 
 
 /- 
 instead of 
 
 = -j- or E = I R, 
 
 neither of which latter forms conveys the same physical 
 meaning as the original. Having satisfied ourselves as 
 to a definition of current, we next investigate voltage, 
 the agent causing this current, and ascertain the quanti- 
 tative relation existing between the two if E is increased 
 what change takes place in the value of /? If E is 
 decreased, how is /affected? Ohm did not know but that 
 doubling E might treble or square /, until he experi- 
 mentally proved that any change in E caused a like 
 change in I; if E were doubled / was doubled; if E 
 were halved, so was /, and this fact is stated by saying 
 that / varies directly as E. 
 
 The law now is in reality complete for a given con- 
 ductor, for we have found the relation existing between 
 E and /. It was early observed, however, that the value 
 of / for a given E in conductors of similar dimensions, 
 but of different substances, is not the same; nor for 
 various sized conductors of the same substance. This 
 modifying property which a conductor has by virtue of 
 its form and nature, is termed resistance, and its study has 
 been fruitful of valuable results. In experimenting Ohm 
 found that by increasing the length or decreasing the 
 
OHM'S LAW. 6 1 
 
 cross-section of a conductor, the current was decreased, 
 and vice versa. This sounds reasonable, because in in- 
 creasing the length, the given E. M. F. has a longer path 
 through which to urge the current of electricity, and in 
 decreasing the cross-section the path becomes smaller. In 
 either case the E. M. F. meets more obstruction in forcing 
 the electricity through, and therefore cannot do so at so 
 fast a rate. This is what is meant by saying that the con- 
 ductivity of a wire decreases or the resistance increases 
 with the length, the reverse being true for changes in 
 the cross-sections. Now, were the dimensions all that 
 influence the resistance of a conductor, we might formu- 
 late a law showing the relation between these properties, 
 and containing only these, but it is an established fact 
 that a conductor's resistance is modified by its nature or 
 the substance of which the conductor is composed; /'. e. t 
 the dimensions in all cases being the same, a silver wire 
 will have less resistance than a copper one; an iron wire 
 more than a copper one, the difference perhaps being 
 due to difference in the molecular structure of the 
 bodies. 
 
 To talk intelligently about this property, known as spe- 
 cific conductivity, we must have some standard to which we 
 can refer. The standard generally accepted is pure 
 copper. If a pure copper wire of given dimensions has 
 a certain capacity for carrying current, and another wire 
 of similar dimensions but of different substance has 
 twice the capacity of the first, the specific conductivity 
 of the latter is 2. This 2 is fixed for the given substance, 
 being dependent solely upon the material of the sub- 
 stance, and hence it is a constant. For the same metal 
 referred to copper this constant is of course always the 
 
62 TESTING OF DYNAMOS AND MOTORS. 
 
 same; but for different substances, or for the same 
 substance in a different degree of purity, it varies, 
 and is commonly designated as K. When K occurs in 
 a formula we know that it represents a number which we 
 must either get from tables compiled from the experi- 
 ments of others, or determine it ourselves. 
 
 Now that we know all the modifying factors, we must 
 determine the exact extent to which resistance depends 
 upon each. The relation of R to /and E is contained 
 in the definition of resistance and in Ohm's law, as 
 already given in the form, current value varies directly as 
 the E. M. F.; i. e., when the E. M. F increases, the cur- 
 rent value increases at the same rate. The relation sought 
 is : the current value varies inversely as the resistance of 
 the conductor ', i. e., when the resistance increases, the cur- 
 rent value decreases at the same rate. To make this 
 plainer, conductivity is defined as the property of a wire 
 to carry current, and resistance as the property of check- 
 ing back current and preventing its rising to an infinite 
 value for a finite E. M. F. Anything that increases the 
 resistance, then, decreases the conductivity, and this 
 fact is expressed by saying that conductivity is the recip- 
 rocal of resistance; or, mathematically, 
 
 Cond. 
 
 Res.' 
 
 where it can be seen that increasing the resistance 
 increases the fraction's denominator, thereby decreasing 
 the fraction's value and the conductivity to which it is 
 equal. 
 
 Let us now suppose a current to be flowing in a con- 
 ductor of a certain resistance; let this resistance be 
 
OHM'S LAW. 63 
 
 doubled. What do we mean by doubling R ? Since R 
 is the power of checking back /, and this power is doubled, 
 we simply mean that R is increased until / is halved. 
 When /falls to half its first value R is doubled. This 
 takes place when either the length, Z, of the conductor 
 is doubled or its cross-section, A, halved. So we may 
 write 
 
 R A fC 
 ~L 
 
 where K is the specific conductivity above explained. 
 Bearing in mind this definition of R, Ohm found that 
 in every case the number expressing the current strength 
 was equal to the number obtained by dividing the "drop " 
 between two points by the resistance between them, all 
 being expressed in proper units. In practical units, it 
 is as follows: To find the amperes flowing in a circuit, 
 divide the difference of potential, as indicated by a volt- 
 meter placed between any two points, by the ohms resist- 
 ance between those points, or 
 
 if E 10, and R 2, then 
 
 E 10 
 / = -g - 5 amperes. 
 
 To retain Ohm's law in this, its original form, we think 
 first of the current, /, as that with which we have to do 
 primarily; next of E. M. F., E, or that which causes the 
 current flow, and last of resistance, R, as that which 
 hinders it. We remember that the value of / is ex- 
 pressed by a fraction, and to increase the value of a 
 fraction we increase the numerator or decrease the 
 
64 TESTING OF DYNAMOS AND MOTORS. 
 
 denominator, so that to increase / we increase E or 
 decrease R, since E is the numerator and R the 
 denominator. The law of the circuit is then expressed 
 by the fraction. 
 
 Ohm's law in this, its simplest form, is found to hold 
 good only on circuits where there is no source of 
 C. E. M. F. Thus, if two equal batteries be connected 
 in opposition, / is not equal to 
 
 but is o. So also if there is a running motor in circuit, 
 the current value depends much more upon the C. E. 
 M. F. of the motor than upon its internal resistance, 
 which is made as low as possible. Another example of 
 C. E. M. F. is seen in the arc of an arc lamp. The pri- 
 mary of a transformer, or any device wherein self-induc- 
 tion takes place, is an example, for self-induction always 
 gives rise to a C. E. M. F. In many cases the self-induc- 
 tion lasts but a fraction of a second, and / soon rises to 
 the value called for by Ohm's law a shunt field winding 
 for example. Since self-induction takes place in a coil of 
 any kind, a circuit to be practically free from it must 
 consist of straight conductors. In some cases it is nec- 
 essary to neutralize this self-induction by means of an 
 equal and opposite induction, as in resistance boxes and 
 in the resistance coils of a Wheatstone bridge, where the 
 wire is doubled in the middle and then wound back on 
 itself. The effect is to have the current pass through 
 half the coil right-handed, and through the other half 
 left-handed, with the result that the opposing tendencies 
 balance each other and there is no magnetic effect. 
 
OHM'S LAW. 65 
 
 We lay stress on the fact that Ohm's law in its simple 
 form holds good only in non-inductive circuits. On a 
 circuit containing an electromagnet, the law applies only 
 after the current has become steady. On a circuit con- 
 taining a running motor, the expression of the law must 
 be modified. In a circuit containing only ohmic resist- 
 ance, if we double the voltage we double the current, 
 but where there is a running motor whose mechanical 
 load is kept constant, the effect of doubling the voltage 
 is to increase the motor's speed, its current remaining 
 almost the same. If the voltage of the dynamo is E, and 
 the C. E. M. F. of the motor is e, that part of the 
 dynamo E. M. F. which drops through the ohmic resist- 
 ance of this circuit is E c, or the difference between 
 the impressed and C. E. M. F. But we shall see later 
 that this drop due to ohmic resistance is / R, where / is 
 the current and R the ohmic resistance of the circuit; 
 these two values, being identical, equal each other, and 
 we have E e = I R, whence, as we shall see presently, 
 
 calling E t, E', we get 
 
 the simpler form. The impressed voltage equals the 
 drop due to resistance plus the motor's C. E. M F.,or 
 E e + I R\ therefore e E I R, or to get the C. E. 
 M. F. of a motor, subtract the drop due to internal 
 resistance of the motor from the impressed E. M. F., 
 as measured by a voltmeter at the motor terminals. 
 On a series motor this internal drop = / (r f -f- r a ), where 
 
66 TESTING OF DYNAMOS AND MOTORS. 
 
 r f and r a are the resistances of field and armature 
 respectively. For a shunt machine the drop across field 
 and armature circuits is the same, and can be figured 
 from either thus, internal drop = 7 a r M where 7 a is the 
 armature current. On compound-wound machines with 
 the long shunt connection, the drop is 7 a (r a -f- ^ s ) where 
 r s is the series field resistance. In the case of the short 
 shunt (see Fig. 18) the series field current exceeds the 
 armature current since it includes that of the shunt 
 field. The series field drop is therefore slightly increased, 
 though not appreciably when 7 a is large. 
 
 When the circuit contains an induction coil, the action 
 upon applying voltage is as follows: Upon closing the 
 circuit a small current passes through; it throws out 
 lines of force which cut neighboring conductors and set 
 up in them an E. M. F. opposed to the impressed E. M. 
 F., and whose magnitude depends upon the rate of cur- 
 rent variation, the number of turns in the coil, and upon 
 the magnetic quality of the core. When 7 reaches its 
 final value the magnetic field is fully established, the lines 
 of force become stationary, the C. E. M. F. becomes o, 
 and 7 then, and only then, equals 
 
 E 
 
 It can now be more plainly seen that under conditions 
 where the current strength varies through such coils, the 
 ohmic resistance is not the only hindrance to be over- 
 come, and that Ohm's law as first expressed will no 
 longer give the true relation between 7 and E. The 
 original expression must then be changed in problems 
 which involve alternating currents, where the greater 
 
OHM'S LAW. 67 
 
 part of the effective resistance is due to self-induction. 
 Calling r the ohmic resistance of the circuit, and r { that 
 due to the induction (sometimes called spurious resist- 
 ance), then we may write 
 
 J ~ > 
 
 >'o + >'i 
 
 but even here it must be remembered that r { has not 
 a constant value, but varies from o to a maximum for 
 every alternation of the current. There may be taken, 
 however, a mean value which shall represent its effective 
 value. 
 
 To revert to the consideration of the formula 
 
 we here assume that E and R are given, and that we wish 
 to find /. If this form can be retained in mind there is 
 little difficulty in deriving from it the expression for E 
 when / and R are known, and for R when E and /are 
 known. / can be written 
 
 for dividing a quantity by unity does not alter its value. 
 The law can then be written 
 
 /_ E 
 i " R' 
 
 which is but another way of writing the proportion, 
 / : i ; ; E : R. Multiplying means together and ex- 
 tremes together, we get, E x i = / X R, or E = I R, 
 which gives E when / and R are known. 
 
 Next, take the form E X i = / X R\ from a rule in 
 arithmetic, if we have two products equal to each other, 
 
68 TESTING OF DYNAMOS AND MOTORS. 
 
 one can be taken as the means, the other as the extremes of 
 a proportion, hence, R : i ; ; E : 7, which can be written 
 
 3T> 77 
 
 J\ & 
 
 I 7 
 
 or, simply, 
 
 _ E 
 
 which gives R when 7 and E are known. 
 
 The same result can be gotten by a method of common- 
 sense analysis. We start with 
 
 which is to say, if we divide E into R parts, each part 
 will be equal to /; therefore 
 
 / = one Itth of E = ~ of E = ^ X E. 
 K K 
 
 If one ^?th of E equals /, R 7?ths, or all of E, = R times 
 as much as one R\.\\, or R times /; i. e., E R 7, or the 
 drop of potential between any two points of a circuit 
 equals the product of the current by the resistance in- 
 cluded. Again, from 
 
 we have E = 7 R as above, or / times R equals E. If / 
 times R = E, 
 
 i E 
 
 R one 7th of E = j x E = -j , 
 
 or the resistance between any two points of a circuit 
 equals the drop of potential between those points divided 
 by the current flowing. 
 
 Resistance has been spoken of as a sort of obstruction 
 
OHM'S LAW. 69 
 
 to the current, very much as we speak of friction in a 
 water pipe, or that caused by the pebbles in a river bed. 
 There are many points of similarity, and perhaps the 
 term electrical friction could be correctly used. We will 
 close this discussion of Ohm's law by looking at it fron 
 another standpoint: 
 
 The fact or law of the transformation of energy from 
 one form into another has already been spoken of. It is 
 also true, so far as can be determined, that a material 
 agent is always active in such transformations. It has 
 been observed that when current flows in a conductor a 
 certain amount of heat is developed. This shows that a 
 certain amount of electrical energy is being transformed 
 into the energy of heat. Careful experiment shows that 
 the amount of electrical energy so transformed equals 
 7 2 /^, where / is the current flowing, and R the resist- 
 ance of the portion of the circuit investigated. This law 
 is due to Joule, and is called by his name. To illustrate 
 the law, if / = 12 amperes, and ft 10 ohms, the 
 energy lost in heat = 12* x 10 1,440 watts. (We 
 say energy lost because ordinarily it is unavailable for 
 any useful purpose.) If R be doubled the energy lost = 
 i2 a x 20 = 2,880 watts. If R be halved the energy lost 
 = i2 a x 5 = 7 20- watts, and so on. The energy trans- 
 formed is, then, directly proportional to R, and R may be 
 called the conductor 's property, by virtue of which it can trans- 
 form electrical energy into heat energy. This property would 
 naturally have different values in different substances, 
 i. e., each substance has its own specific resistance, etc., 
 and resistance may be regarded as a measure of one of 
 the fundamental properties of conductors. 
 
CHAPTER IV. 
 
 MEASUREMENT OF CURRENT. 
 
 IN studying the factors involved in an electric circuit, 
 no better order can be selected than that in which they 
 occur in Ohm's law. We shall then take up in succession, 
 Current Strength, Electromotive Force, and Resistance, 
 and consider in a common-sense way the methods em- 
 ployed in their measurement along the most practical 
 lines of industrial work. 
 
 To do quantitative work of any kind, there must be 
 chosen proper units in terms of which operations and 
 results can be expressed. The universally adopted units 
 of current strength, E. M. F., and resistance are the 
 ampere, volt, and ohm, which have already been briefly 
 defined. The practical basis upon which the unit of cur- 
 rent strength is founded is the amount of metal which 
 the current will deposit from a solution of the metal per 
 unit of time. When a current is passed through a solution 
 of a salt of any metal, the salt is decomposed and the 
 metal deposited upon the negative plate or cathode. The 
 amount deposited depends upon the metal, the current 
 strength, and in lesser measure upon the nature of the 
 solution. That current which deposits from a solution 
 of silver nitrate .001118 grams of metallic silver, or from 
 a solution of copper sulphate .000329 grams metallic 
 
 70 
 
MEASUREMENT OF CURRENT. 7 1 
 
 copper, per second, is of unit strength, and is called the 
 ampere. 
 
 The deposition of metals is one of the most popular 
 and most reliable methods of measuring current strength, 
 and is used a great deal in calibrating and standardizing 
 the higher grades of galvanometers, ammeters, and 
 shunts. In every-day work it is customary to employ 
 the method of copper deposition, as it requires the use of 
 neither expensive apparatus nor of chemically pure silver, 
 which is hard to get. 
 
 The apparatus required for the copper method consists 
 of a pair of scales, a timepiece, copper 
 plates fitted to a wooden support which 
 holds them in place in a bath of copper 
 sulphate, and the means for cleaning and 
 drying the plates. Fig. 23 shows a very 
 simple and cheaply made voltameter. V is 
 
 a glass vessel containing the solution of 
 copper sulphate (blue vitriol). T is a 
 wooden cover provided with holes to receive the wires 
 connected with the plates P. The plates are drawn up- 
 snug into slots to prevent turning, and are joined to the 
 circuit by the connectors C, which are of the ordinary 
 type. It is usual to make the anode heavier than the 
 cathode, for deposition is carried on at its expense. 
 
 The chemical action in the voltameter is simple. The 
 current entering at the anode causes the copper of this 
 plate to pass into solution, and a progressive interchange 
 of copper in the molecules of the solution between the 
 two plates presumably takes place, and metallic copper 
 is deposited on the cathode where the current leaves the 
 voltameter. It would naturally be expected that the 
 
72 TESTING OF DYNAMOS AND MOTORS. 
 
 gain of the one would exactly equal the loss of the other, 
 and this is true, excepting that the formation of by-prod- 
 ucts in the solution makes the anode's loss exceed the 
 cathode's gain, and it is for this reason that the cathode's 
 gain in weight is alone used in estimating current value. 
 
 In preparing the voltameter for use, the copper plates 
 (called electrodes) must first be cleaned, dried, and 
 weighed. In cleaning, all dirt and oxides, shown by 
 discoloration, are removed with a scratch brush; the 
 plates are then put through a series of baths, whose order 
 and composition are variously given by different writers. 
 The following commendable course is given by Stewart 
 and Gee in their Laboratory Manual : 
 
 Bath No. i. Alkaline liquid for cleansing copper: i 
 part, by weight, of caustic soda; 10 parts, by weight, 
 of water. 
 
 Bath No. 2. Acid solution : i part, by volume, strong 
 sulphuric acid; 10 parts, by volume, water. 
 
 Bath No. 3. Dipping liquid : equal volumes commer- 
 cial nitric acid (or that left from battery fluid) and water. 
 
 Bath No. 4. Brightening liquid : 100 parts, by volume, 
 strong nitric acid ; i part, by volume, strong hydro- 
 chloric acid. 
 
 Enough of these solutions should be prepared to com- 
 pletely cover the copper plates. No. i being placed in 
 a porcelain evaporating dish, the others in glass beakers. 
 The plate is first washed under the tap, rubbing the 
 plate well with a rag. It is then boiled in (alkaline) 
 liquid No. i. This will cause discoloration, due to 
 oxidation. From this point on, the plate should not be 
 touched with the fingers. After boiling until the discol- 
 oration is pronounced, raise the plate with a lifter and 
 
MEASUREMENT OF CURRENT. 73 
 
 wash under the tap; then remove it and place it in liquid 
 No. 2 long enough to enable the acid to dissolve the dark 
 colored oxide. Wash again with water (distilled water 
 preferred), and place in liquid No. 3 for about fifteen 
 seconds, after which it is washed again and dipped for 
 a few seconds in liquid No. 4. As this liquid is very 
 strong, the plate should be washed quickly and thor- 
 oughly in distilled water. Should it now fail to present 
 a bright clean surface, the process must be repeated. 
 Once bright, the plates are kept in a dilute solution of 
 copper sulphate until desired for use. 
 
 The solution in the depositing bath is made by dis- 
 solving 100 grams of copper sulphate in 500 cubic centi- 
 metres of water (or 17^ ounces in one pint). Before 
 weighing, the plates are thoroughly dried either by 
 placing them in a vessel containing chloride of lime, 
 which in a short time will absorb all moisture, or by 
 heating, care being taken that the plates are not 
 touched by the fingers or otherwise soiled. To prevent 
 rusting, and thereby to save much work on future 
 occasions, the plates, when not in use, are kept in a box 
 containing a dish of chloride of lime. 
 
 Before proceeding with a measurement we will build 
 up the formula to be used. We have defined the ampere 
 as that current which will deposit .000329 gram of 
 copper per second; as it is impracticable to deal with 
 a current that flows for but i second, it is allowed to 
 flow for an hour or more, when the deposition which 
 takes place in i second is gotten by dividing the total 
 deposition by the number of seconds. This value of 
 deposition, made by the unknown current in i second 
 divided by .000329, gives the value of the unknown cur- 
 
74 TESTING OF DYNAMOS AND MOTORS. 
 
 rent expressed in amperes. If / is the current to be 
 measured, K, the electro-chemical equivalent of copper 
 (/. e., the .000329 gram deposited by i ampere in i 
 second), and W, the weight of the copper deposited by 
 current / in / seconds, then W K 1 1 weight depos- 
 ited by i ampere in i second X the number of amperes, /, 
 X the number of seconds, /. From W = Kit, we get, 
 
 or the unknown current equals the total weight deposited 
 divided by the weight which i ampere will deposit in the 
 allotted time. 
 
 To conduct a test the voltameter is connected up so 
 that the current enters at the heavier plate, anode, 
 and leaves at the lighter one, or cathode. A current 
 indicator and a variable resistance are placed in series 
 with the bath. The object of the test is to measure the 
 current flowing through the voltameter, with a view, 
 generally, of standardizing some ammeter or galvanom- 
 eter. The current indicator need not be calibrated, 
 but by means of the variable resistance the needle is 
 kept at one spot throughout the test. Preliminary to the 
 test it is well to use a pair of little test plates to insure 
 that deposition takes place in the right direction, also it is 
 well to secure approximately the current to be used, with 
 the voltameter short-circuited, and thereby avoid the 
 error incident to adjustment. All being ready, the switch 
 is closed and the time carefully noted. It is not well for 
 the current to exceed a density of .08 ampere per square 
 centimetre (.5 ampere for square inch), as the deposi- 
 tion then goes on too rapidly, and the copper granu- 
 
MEASUREMENT OF CURRENT. 
 
 75 
 
 lates. and falls to the bottom as a sort of "mud," or 
 black powder. 
 
 The form of voltameter just 'described is capable 
 of giving results with an error of from i to 4 #, 
 according to the experimenter's skill. When more 
 accurate work is desired, the spiral coil form may 
 be used. This has the advantage that it is easier 
 to manipulate, and has a 
 normal error of but 0.3 %. 
 The following description is 
 taken from a paper read by 
 Professor H. J. Ryan before 
 the American Institute of 
 Electrical Engineers, May, 
 1889: Two copper wires are 
 selected whose size and 
 length are determined by 
 the current to be measured. 
 They are first cleaned by 
 fastening one end in a vise, FIG. 24. 
 
 and carefully sand-papering 
 
 the surface. The wire is then coiled on a cylinder, care 
 being taken that it does not touch the hands at any time. 
 The cathode is made into a smaller coil than the anode, 
 hence is lighter. Fig. 24 shows the final appearance 
 when mounted. On block A, a glass vessel , is placed; 
 T and 7", are two binding posts serving as terminals and 
 also holding supports, c t , <-,, to which are attached the 
 two coils hanging one inside the other, but not touching. 
 Before use, the coils are chemically cleaned as fol- 
 lows: After polishing, the gain coil is washed by plung- 
 ing it into a jar of water containing a little sulphuric acid. 
 
76 TESTING OF DYNAMOS AND MOTORS. 
 
 It is then rolled on filter or blotting paper to remove all 
 but a film of water. The coil is then dipped in 95 % 
 alcohol, removed, and the excess of alcohol allowed 
 to drip back into the jar. By again rolling the coil on 
 clean filter or blotting paper, nothing but a mere film 
 of alcohol remains, and that is thoroughly evaporated 
 in a few moments, leaving the coil entirely dry. Coils 
 that have become corroded can be rapidly cleaned by 
 plunging into a mixture of 100 parts strong nitric acid, 
 i part hydrochloric acid, and then proceeding as 
 already directed. At the end of the deposit the gain 
 coils are immediately removed and plunged first into 
 clean water, then into the acidulated water, from which 
 they are dried by means of the alcohol. When dry they 
 are ready to be weighed. The copper sulphate, water, 
 and acid need not necessarily be chemically pure. The 
 density of the voltameter solution should not be less than 
 1. 1, nor more than 1. 18, referred to water. A voltameter 
 of the above type suitable for measuring currents up to 
 4 amperes has a gain coil of No. 16 B. & S. wire, 2 1/2 
 metres (8.2 feet) long. For heavier currents two or 
 more voltameters can be placed in parallel. 
 
 The voltameter cannot be used to measure alternat- 
 ing currents, for any tendency to deposit on one plate is 
 met at the next alternation by a counter tendency to 
 deposit on the other. Nor is the method applicable to 
 industrial work where the current value must be read at 
 a glance. It is, however, a most valuable method in a 
 laboratory or testing room, for calibrating instruments 
 which in turn are to be used as standards. 
 
 To standardize an ammeter it is connected in series 
 with the voltameter, as shown in Fig. 25, where B is the 
 
MEASUREMENT OF CURRENT. 77 
 
 source of current, G the meter to be standardized, R a 
 variable resistance, K a key for closing the circuit, and 
 A' a key for short-circuiting the voltameter V. In 
 making the test, V is first short- 
 circuited and R adjusted till the 
 reading on G is a little above the 
 desired value, and then A' l is 
 opened. This precaution reduces 
 the error due to a variation in the 
 current. The test plates, above 
 referred to, can be profitably used, Fi<;. 25. 
 
 and may indicate the current to be 
 
 flowing the wrong way ; if necessary, the battery leads 
 can be reversed, and this may necessitate reversing G's 
 leads to right its deflection. Everything adjusted, A\ 
 is opened, A' closed, and the time noted. 
 
 If the instrument under calibration is one with a needle 
 and scale, such as the Weston type, several points on the 
 scale can be determined and the intermediate values 
 interpolated, or a curve can be plotted giving true current 
 values and their corresponding scale readings. If G is a 
 tangent galvanometer, it is usual to determine what 
 current causes a deflection of 45, and the current 
 value corresponding to any other deflection can be 
 figured from the formula / = K tan a, where a is the 
 deflection, /the current to be found, and A' that value 
 of / giving a 45 deflection. The formula / = A" tan 
 a is the equation of the tangent galvanometer, and in 
 words may be stated thus: The current passing through 
 a tangent galvanometer is equal to the tangent of the 
 angle of deflection multiplied by a constant. The angle, 
 a, is read on the instrument itself, its tangent is gotten 
 
78 TESTING OF DYNAMOS AND MOTORS. 
 
 from a table of tangents. The constant, K, which has 
 been defined, is called the constant of the galvanometer, 
 and is different for different instruments. The factors 
 upon which it depends are the number of turns of wire in 
 the coils, the coil's average diameter, and the horizontal 
 component of the earth's magnetic field. Why this is true 
 can in a measure be gathered from the following brief 
 outline of galvanometer principle placed here for those 
 readers to whom it may not be familiar. Every magnet has 
 lines of force, or a field, and any magnetic needle placed 
 near the magnet will take up a position from which it will 
 resist being turned. The magnetic field, then, has a 
 directing force over the needle, and holds it in a position 
 to give the longest metal path to the lines of force. 
 Now, the earth is a huge magnet, having lines of force 
 with definite direction, and these lines exert a directing 
 or guiding force on all magnetic materials. If a needle 
 be suspended or pivoted in air, it takes up a position 
 depending upon the direction of the earth's lines at that 
 place; and on galvanometers having only the earth's field 
 as a directing force, the frame of the instrument must be 
 turned until the earth's magnetic influence holds the 
 needle over the zero mark. If a magnet be brought 
 near the needle, it will leave its zero position and take 
 up a new on.e where the directing force and opposing 
 deflecting force balance each other. If the magnet be 
 removed, the needle ought to resume its original posi- 
 tion. Now replace the magnet by an electromagnet, 
 i. ^., wind around the needle turns of wire capable of 
 carrying a current, and we have a galvanometer. Next 
 send a current through the coils and the needle deflects; 
 increase the current and the deflection increases. 
 
MEASUREMENT OF CURRENT. 79 
 
 Weaken the directing force by opposing a magnet to the 
 earth's field, and the deflection increases. Decrease the 
 diameter of the coils, bringing the current nearer to 
 the needle, and the deflection becomes greater; increase 
 this diameter and the reverse is true. Anything which 
 increases the directing force or decreases the deflecting 
 force, lessens the deflection, and, rife versa, to decrease 
 the directing force or increase the deflecting force 
 increases the deflection. 
 
 The earth's lines of force do not run parallel to the 
 surface except at the equator, while at the poles they 
 run perpendicular to the surface. At places between the 
 poles and equator the lines run at an angle to the earth's 
 surface, and therefore have two effects upon a suspended 
 needle : one is a vertical pull which tries to make the 
 needle dip ; the other, which is the one we have 
 most to do with, acts upon the needle horizontally, and 
 is known as the horizontal component of the earth's 
 magnetism. 
 
 Since the earth's field has a slight cyclic and, indeed, 
 daily variation, A", in the strictest sense, does not remain 
 constant; but as it is influenced vastly more by the 
 proximity of masses of iron or steel, it may be regarded 
 as constant so far as the earth is concerned. The value 
 of K is found as follows: in the above vqltametric test 
 adjust the needle to a deflection of 45, and by means of 
 the voltameter determine what current causes this deflec- 
 tion. The tangent of 45 = i, hence the above expres- 
 sion becomes, / = K x i = K ; but / is the current 
 which gives a deflection of 45 ; therefore K has the same 
 value as /. For any other deflection I and K have 
 different values. While this form of the experiment is 
 
80 TESTING OF DYNAMOS AND MOTORS. 
 
 simplest and best, because the needle is most sensitive in 
 this part of the scale, and tangent 45 = i, any other 
 angle can be used in determining K. 
 
 A serious deviation from the above law indicates either 
 an imperfection in the instrument, the presence of some 
 outside influence, or lack of adjustment of the coils to the 
 magnetic meridian, /. e. f direction in which the needle 
 points when influenced only by the earth. This latter 
 adjustment is secured by moving the instrument until the 
 same deflection on opposite sides of o is gotten upon 
 reversing the galvanometer current. If the galvanom- 
 eter is of the reflecting type, /. e. t with mirror and 
 telescope, or lamp, and scale, it is not feasible to obtain 
 a deflection of 45, so that the deflection is not read in 
 degrees at all, but in terms of the scale divisions. The 
 value of one scale division depends upon the distance of 
 the scale from the mirror. The farther off the scale, the 
 greater the number of divisions through which a given 
 deflecting force will cause the needle to throw the ray of 
 light. This distance, then, should be fixed once for all. 
 Another factor of great importance in determining the 
 value of a division is the damping or directing magnet, 
 which is free to turn or slide up and down on a rod above 
 the instrument. The directing force of this magnet can 
 be made so strong as to render the needle independent of 
 outside disturbing fields. 
 
 If these points remain fixed, their absolute distance need 
 not be known, it being customary to determine what cur- 
 rent causes a deflection of 100 or 200 scale divisions. With 
 the scale at least 3 feet from the mirror, we can assume, 
 without sensible error, that the current is proportional to 
 the deflection as read on the scale, so that with the cur- 
 
MEASUREMENT OF CURRENT. 
 
 8l 
 
 rent value of one point determined, all others can be 
 figured out. Thus, suppose the current for 100 divisions 
 is .05 = 5/100 ampere, what is the current for 45 
 divisions? The proportion is .05 : x j | 100 ; 45, 
 whence jc = .0225 ampere. Or, if 100 divisions be due 
 to .05 ampere, i division is due to i/ioo of 5/100 
 ampere = 5/10,000, and 45 divisions would be due to 45 
 times this, or 225/10,000 = .0225 ampere. 
 
 Most reflecting galvanometers are extremely sensitive, 
 and but a suggestion of current will throw the ray off the 
 scale. By shunting the instru- 
 ment, however, it may be used 
 to measure currents many 
 thousands of times its own ca- 
 pacity. For practical meas- 
 urements of current it is usual 
 to set up these instruments to 
 read from a reliable standard 
 shunt of known resistance. 
 
 The theory of the shunt is briefly as follows: Let r t r t 
 be the resistances of the two branches included between 
 A and B of Fig. 26. The fall of potential from A to B is 
 the same along either branch. Let /be the total current 
 and /', and /, the currents in r l and r 3 , respectively. 
 / = t\ -f- /, and, from Ohm's law, the drop from 
 A to B t\ r v But it also = /, r s . Therefore, 
 /, r t = / a r a , and we have an equality of two products, 
 whence from arithmetic we get, t\ : / a ' ; ;- a : r t . That is 
 to say, the currents in the two branches are to each 
 other, inversely, as the resistances of these branches. 
 For example, let r g be the resistance of the branch, 
 including the galvanometer, G, Fig. 27, and r B that of 
 
 FK;. 26. 
 
82 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 shunt box, S, then, t' B ' t g '.' r g : r a , whence taking prod- 
 uct of means and extremes and solving for z g , we get 
 
 If 
 
 and 
 
 r g an r s are known, and f g is measured, / 8 is also 
 determined, and since / = i g + / s , the current in the 
 external circuit is known. Since 
 
 to get z'g, we need not know the absolute value of either 
 r 9 or r gt but only their ratio, or the value of the fraction 
 
 Furthermore, it is convenient to have this fraction a 
 simple one, such a i/io, i/ioo, i/iooo. This makes the 
 shunt current 10, 100, or 1,000 times as great as that read 
 on the galvanometer, and the total current 10 -f i, 100 -f- 
 i, or 1,000 -f- i times the galvanometer current. Now 
 this, while vastly better than any haphazard value, can 
 be improved upon. Thus, instead of i/io, i/ioo, etc., 
 put 1/9, 1/99, J /999- The shunt 
 current is now, 9, 99, and 999 
 times (according to shunt used) 
 as great as the galvanometer cur- 
 rent, while the total, or line, 
 current is 9 -f i = 10 times, 99 -f- 
 i = TOO times, or 999 -f- i = 1,000 
 FIG. 27. times, the galvanometer current. 
 
 With this arrangement the value 
 
 read on the scale is i/io, i/ioo, or 1/1,000 of the total 
 current, /, and G becomes direct reading. 
 
MEASUREMENT OF CURRENT. 83 
 
 In accordance with this theory, the finer instruments 
 are provided with a shunt box, which has resistances 
 bearing the ratios 1/9, 1/99 and 1/999 to that of the 
 galvanometer. Each box can be used only with its own 
 galvanometer, or one similar in all respects, and, in very 
 fine work, only at temperatures approximating that at 
 which it was standardized. 
 
 In the calibration of a sensitive galvanometer, intended 
 for current measurements, the constant is determined by 
 use of shunt and voltameter. It 
 can be then used on any direct 
 current circuit, and with any 
 shunt box whose resistance is 
 known. The connections for 
 the test are shown in Fig. 28, 
 where G is a galvanometer of Fli; 2g 
 
 known or easily determined re- 
 sistance, S, the shunt, also of known resistance, in multi- 
 ple with G\ V is the voltameter, A', its cut-out switch, 
 B a battery or other current source whose voltage can 
 be varied within wide limits; R is a variable resistance 
 for further adjustment, r and r resistances in series 
 with G. On high resistance galvanometers intended for 
 measuring current, voltage, and resistance, r, r is gen- 
 erally divided into two boxes, one of which has a constant 
 value of 75,000 or 100,000 ohms, the other a variable one 
 of 10,000 to 15,000. 
 
 Suppose our scale has 200 divisions, and it is desired to 
 calibrate it throughout. Call G's resistance 1,000 ohms: 
 and r's 105,000, making a total of 106,000 ohms in G's 
 circuit. Let the shunt resistance, S, be.i ohm (i/io). 
 Now close A'and K' and adjust R until 's needle deflect? 
 
84 TESTING OF DYNAMOS AND MOTORS. 
 
 200. To avoid exceeding the current carrying capacity 
 of S, it should be protected by a fuse or short-circuiting 
 switch; and it is well to have in circuit a current indicator 
 as a guide to preliminary adjustment. If it is impossible 
 to get a deflection of 200 by means of -/?, r may be 
 decreased or the resistance of the shunt increased. Cir- 
 cumstances dictate the position of the directing magnet, 
 and where the instrument is set up amid disturbing 
 external influences, it is not well to move it far from 
 the needle. The proper deflection secured, K v is opened, 
 time noted, R quickly readjusted, if necessary, and the 
 deflection of 200 maintained for about an hour. The 
 current value is determined by the method of weighing 
 as already explained. Suppose this, /, to have been 20 
 amperes. The problem now is to find the value of G's 
 current. We have as data, / = 20 amperes; resistance 
 of galvanometer circuit 106,000 ohms; resistance of 
 S = . i ohm. From the theory of shunts already out- 
 lined, we get the proportion, i e : 4 * * . i : 106,000, whence 
 
 . i i 
 
 8 io6,ooo s 1,060,000 
 
 also, ig + i s = I 20 amperes, or i a = 20 f e 
 Substituting this value of 4 in the expression, 
 
 i 
 
 Iff ^^ la 7 J 
 
 I,O6O,OOO 
 
 we have 
 
 . . . i_ _ 20 fg 
 
 * g ~ ' **'i,o6o,ooo " 1,060,000' 
 
 Multiplying the quantities on both sides of the equality 
 sign by 1,060,000 and we get, 1,060,000 i e 20 i & 
 
MEASUREMENT OF CURRENT. 85 
 
 or 1,060,001 / = 20, and /',,= - amperes 
 
 1,060,001 
 
 through the galvanometer circuit. Expressed decimally, 
 this is .0000189 ampere, or that current which, passing 
 through G, under existing circumstances, causes a deflec- 
 tion of 200 scale divisions, and is commonly accepted as 
 the instrumental constant of the galvanometer. If there 
 are no variable external influences (such as heavy ma- 
 chines under test, passing cars or trains, traveling cranes, 
 elevators, etc.), and the instrument is carefully set up 
 with scale at a proper distance, we are safe in assuming 
 that if .0000189 ampere deflects the needle 200 divisions, 
 1/200 part of it will cause a deflection of i division. 
 This would be .000000094 ampere, and more properly 
 constitutes the constant of the instrument. To elimi- 
 nate errors of adjustment and observation it is well to 
 experimentally fix at least three points on each side of o. 
 
 In setting up a galvanometer calibrated elsewhere, and 
 with the same shunt, S, it is unnecessary to repeat the 
 entire calibration. Before removal, a portable instru- 
 ment, such as a Weston milliammeter, can be placed in 
 circuit and the indication corresponding to several points 
 on the galvanometer scale marked on the scale of the 
 Weston instrument. Then upon resetting the galvan- 
 ometer the current can be adjusted by this indicator, 
 and the deflection on 's scale adjusted to the proper 
 value by the controlling magnet. 
 
 We found .000000094 ampere in the galvanometer cir- 
 cuit to cause a deflection, of i division, and this we called 
 later the instrumental constant. With the . i ohm shunt a 
 deflection of 200 corresponded to a current of 20 amperes 
 in the outside circuit; this makes a deflection of i division 
 
86 TESTING OF DYNAMOS AND MOTORS. 
 
 correspond to an external current of .1 ampere, and this 
 we call the working constant. For example, a deflection 
 of 47 indicates a current of 47X-i = 4.7 amperes. 
 
 From the instrumental constant can be determined the 
 working constant of any other shunt, without using the 
 voltameter. Suppose we use a .25 ohm shunt: with 
 the same current as before, the ray is thrown quite off 
 the scale, and the deflection must be reduced to be read- 
 able. This done, let us inquire what line current a de- 
 flection of 200 indicates. From the theory of the shunt 
 given above, we have, 
 
 galvanometer current : shunt current ', ', shunt resist- 
 ance : galvanometer resistance, or, i g : i s ' ; .25 : -106,000, 
 whence 
 
 ^t? ^8 X ' "2 ~ - 
 
 IO6,OOO 
 
 For a deflection of 200 we have seen that t' g = .0000189 
 ampere, and substituting this in the expression for / g , 
 we get, 
 
 .0000189 = - ! , 
 424,000 
 
 or / 3 = .0000189 X 424,000 = 8.0136 amperes, the shunt 
 current; this added to t g = .0000189 ampere, gives the 
 line current. Practically speaking, the following propor- 
 tion holds good, 20 : 8 ; ; .25 : .i. Showing that when 
 the resistance in the galvanometer circuit is so great that 
 the galvanometer current can be neglected, we may say 
 that for the same deflection, with two different shunts, 
 the line currents are inversely as the resistance of the 
 shunts. Thus with a shunt of .03 ohm a deflection of 
 200 indicates what current? Here we have 20 : x \\ 
 .03 : .1, or .03.* = 2, whence x = 66.66 amperes. 
 
MEASUREMENT OF CURRENT. 87 
 
 More exactly it is 66.69, which is approximately the same. 
 With such a galvanometer and its circuit resistance 
 undisturbed, it is a simple matter to use any shunt of 
 known resistance. 
 
 Another way of altering the range of readings, is to 
 leave S the same, and vary the galvanometer circuit 
 resistance. Let us suppose that this resistance is to be 
 reduced. Since more current will now flow in G, for a 
 given line current, the deflection will be greater. The 
 following is a simple way to proceed: Adjust the line 
 current so that the deflection shall be, say, 50 divisions, 
 showing the line current to be 5 amperes. Then cut out 
 resistance from G's circuit till the deflection is 100. A 
 deflection of 100 now corresponds to 5, and 200 to 10, 
 amperes. This assumes the line current to have been 
 kept constant throughout. We can, however, verify 
 results by checking up. Since the galvanometer current 
 can be neglected, the drop through the shunt is . i x 5 .5 
 volts. If we are to produce a deflection of 100, the galvan- 
 ometer current required for this, is, .000000094 x 100 = 
 .0000094 ampere, .000000094 ampere being the current 
 to deflect the ray i division. Since a potential difference 
 of .5 volt is applied to '8 circuit, we have by Ohm's law, 
 
 .0000094 = : - , 
 
 where x is the galvanometer circuit resistance under the 
 new conditions. Solving for .r, we find x = 53, 191 ohms, 
 and the first resistance, 106,000 ohms, must be reduced to 
 this value. If experiment and theory fail to check up very 
 closely, it indicates either an error in the determination of 
 the galvanometer constant, or inaccuracy in the resistance 
 
88 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 boxes. Here again we see the same rough proportionality 
 as before : halving the galvanometer circuit resistance Jias 
 doubled the deflection due to a given current. A third 
 way of increasing the reading range is to draw out the 
 scale till its o occupies the TOO or 200 point. On a 
 properly adjusted instrument with correct constant, this 
 introduces no error. 
 
 Thus far we have assumed the controlling magnet to 
 occupy a constant position, a condition hard to secure. 
 
 Fig. 29 gives connections for 
 a method of setting up a 
 galvanometer independent of 
 all local conditions, and one 
 popular in actual practice. G 
 is a galvanometer with a di- 
 recting magnet. 7? is a vari- 
 able, and ' a constant resist- 
 ance. B is a standard cell whose voltage is accurately 
 known, and K a key. Resistance, r, completes the 
 battery circuit, and the galvanometer reads the drop of 
 potential off a known fraction of r. This resistance, r, 
 must be so high as to produce no perceptible effect upon 
 the cell's terminal voltage when the circuit is closed. 
 Let r be 10,000 ohms, and the E. M. F. of the cell, a 
 Daniell, i volt. If the resistance of the galvanometer 
 circuit is high, it can be placed as a shunt across any 
 part of the box without affecting the resistance of the 
 battery circuit. The drop of potential through any por- 
 tion of the box is proportional to the resistance of this 
 portion; this enables us to apply to the galvanometer cir- 
 cuit any fraction of the cell's E. M. F., by simply taking 
 the drop from that same fraction of r. Thus let the 
 
MEASUREMENT OF CURRENT. 89 
 
 resistance included be 1,000 ohms ( i/io of 10,000), or 
 i/io of the box; then the potential difference applied to 
 's circuit is i/io volt. Should we do away with a high 
 resistance galvanometer, and with ft and ft', and suppose 
 the galvanometer circuit resistance to be but rooohms, it 
 will, when shunted across any part of /-, perceptibly lower 
 the resistance between the points included, and the pre- 
 vious condition no longer exists : the current is increased, 
 and the potential difference at the galvanometer terminals 
 no longer bears the same ratio to the total voltage of the 
 cell that the included resistance of the box does to the 
 total resistance. With a high resistance galvanometer cir- 
 cuit, of 10,000 ohms and upward, the effect upon the circuit 
 can be neglected. In the connections considered above, 
 the galvanometer resistance was 1,000 ohms, and ft and ft' 
 respectively 5,000 and 4,000 ohms, making the total gal- 
 vanometer circuit resistance 10, ooo ohms. With a current 
 of .0000188 ampere (corresponding to a deflection of 200), 
 the potential difference through the circuit is by Ohm's 
 law, E I R, = .0000188 x 10,000 = .188 volt. This 
 may be called the working voltage constant. The 
 instrumental voltage constant is .000000094 x 1,000 = 
 .000094 volt. That is to say, if .000094 volt be applied 
 to the galvanometer terminals, a deflection of one division 
 will result. We now place the galvanometer circuit across 
 2/10 of the box, and have .2 volt applied to this circuit. 
 This should produce a deflection of 213 divisions. For if 
 200 divisions are due to.i88 volt, i division is due to 
 1/200 of .188 volt, or .188/200 = .00094 volt, and if 
 .00094 volt causes a deflection of i, .2 volt will cause a 
 deflection = .2 -j- .00094 = 212.7, practically 213 divi- 
 sions. Having .2 volt on the galvanometer circuit, it only 
 
90 TESTING OF DYNAMOS AND MOTORS. 
 
 remains to adjust, by means of the controlling magnet, the 
 deflection to 213. This must be done by raising or lower- 
 ering the magnet, and not by rotating it. To test the 
 adjustment, reverse the current through the galvanom- 
 eter circuit: the deflections on opposite sides of o should 
 be equal. 
 
 The galvanometer adjusted, any standard shunt can be 
 used, and the scale indications calculated as in the last 
 case. For example, with a .01 ohm shunt we have the 
 proportion .01 : 10,000 ; ; .0000188 : x, where x is the 
 current in the shunt. Whence, by arithmetic, x = 18.8 
 amperes. A deflection of 200, then, corresponds to a 
 current of 18.8 amperes, 100 divisions to 9.4 amperes. 
 
 If nothing is known regarding the galvanometer save 
 its resistance, the method just described can be used as 
 follows, to set it up and calibrate it, and without recourse 
 to a voltameter. The galvanometer circuit is placed 
 across a part of the box r, say i/io of it, and the con- 
 trolling magnet adjusted to give a certain deflection, say 
 100. With the resistance of the galvanometer circuit 
 known, its current is found by Ohm's law, and from this 
 current both the voltage and current constant are deter- 
 mined. If it is desired not to disturb the magnet, the 
 resistance of R' can be varied until the deflection 
 sought is gotten. This is allowable only so long as the 
 resistance of the galvanometer circuit is not reduced too 
 much, and to avoid this condition it is customary to 
 Hiave R so high as to avoid error, even though R' be all 
 cut out. 
 
 The range for which a galvanometer is adjusted to 
 read current should not be so great as to make the 
 lower readings inaccurate. As a guard against this, the 
 
MEASUREMENT OF CURRENT. 91 
 
 galvanometer must be so adjusted that the maximum cur- 
 rent in the shunt corresponds to the highest possible 
 deflection. If under this condition the lower readings 
 are still liable to error of observation, it is best to divide 
 the readings into two parts either using two shunts cr 
 two adjustments of the galvanometer. In setting up the 
 instrument it is desirable to know the current range as 
 well as resistance of at least one shunt, so that some 
 estimate can be made of the potential difference to which 
 the galvanometer is to be subjected. Suppose shunt No. 
 i to have a resistance of .001 ohm, and a carrying capac- 
 ity of 100 amperes: 100 amperes x .001 ohm i volt, 
 which is to be impressed upon the galvanometer circuit 
 at full load. Fora minimum load of i ampere, the poten- 
 tial difference impressed will be i x .001 .= .001 volt. 
 The conditions must be such that . i volt will give a deflec- 
 tion of say 300 scale divisions. This can be secured by 
 shifting the magnet and varying the resistance in A' 1 , 
 while the galvanometer includes 1,000 ohms in r. One 
 ampere will now give a deflection of 3 divisions, or i/ioo 
 of what 100 amperes give. Having made the scale direct- 
 reading for one shunt, the readings for the other shunt 
 are a matter of calculation. Thus, if the resistance of 
 No. 2 be half that of No. i, and of twice its current 
 capacity, it will require the full 200 amperes to give the 
 full deflection of 300 divisions. This gives us i 1/2 
 divisions per ampere, and this shunt can be used for 
 currents between 100 and 200 amperes with the same 
 degree of accuracy as No. i can for those between i and 
 100 amperes. 
 
 Standard shunts are furnished with a guarantee of 
 their resistance at a given temperature, and tables show- 
 
92 TESTING OF DYNAMOS AND MOTORS. 
 
 ing their resistance at other temperatures can be had. 
 In ordering, the desired current carrying capacity should 
 be specified. On the other hand, it is not difficult to 
 make a shunt that will serve all ordinary purposes. The 
 precision of the work will depend largely upon the 
 sensitiveness of the galvanometer, for if it is not very 
 sensitive, minute variations in resistances will not affect 
 the deflection. Let the range of the proposed shunt be 
 
 1 to 100 amperes, and let the galvanometer scale contain 
 200 divisions. We first determine what voltage must be 
 used to cause this deflection of 200. Assume it to be . i 
 volt. The resistance of the shunt must then be .001 
 ohm, for (Ohm's law) 
 
 J? = . = .1 -+- 100 = .001. 
 
 The next step is to select conductors of proper resistance 
 and of capacity to carry the current without undue heat- 
 ing. The wire table (see appendix) shows No. 10 B. & 
 S. copper wire to carry 40 amperes with little heating, and 
 to measure i ohm per 1,000 feet. One foot then measures 
 .001 ohm; five wires in multiple, and each i foot long, 
 measure 1/5 of .001 ohm = .0002 ohm, and will carry 200 
 amperes. Five wires in multiple, and 5 feet long, measure 
 .0002 x 5 = .001 ohm, and will also carry 200 amperes. 
 This then, furnishes the desired combination. Allowing 
 
 2 inches for the soldering, the wires are cut 5 feet 2 
 inches long, and are drawn into holes bored into copper 
 lugs, shown at A B in Fig. 30. The mortised joints are 
 then well sweated with solder, and the whole mounted on 
 a wooden base. On the lugs are two sets of binding 
 posts; one set to serve as the main circuit connection, 
 the other for the terminals of the galvanometer circuit, 
 
MEASUREMENT OF CURRENT. 93 
 
 and care must be taken that these terminals do not 
 make a bad joint in the main circuit, for this would be 
 equivalent to increasing 
 the shunt's resistance. 
 In testing the shunt by 
 the standard cell and re- 
 sistance box, there should I- K.. 30. 
 be provided a 3-way 
 
 switch, by which the galvanometer can be shifted alter- 
 nately to the standard cell and to the shunt. Through- 
 out this test, as in all others, readings should be taken 
 on both sides of o. 
 
 In ammeter calibration it is usual to have one man 
 look after the galvanometer, while an assistant varies the 
 current strength as required and marks the deflection of 
 the needle if the scale is ungraduated, or reads out the 
 deflection on a graduated scale. The galvanometer man 
 makes a record, in parallel columns, of the galvanometer 
 and corresponding ammeter readings, and these can be 
 plotted as a curve for future reference. If the variation 
 exceeds 20 or 25 % the instrument should be returned 
 to the maker for correction. 
 
 In graduating new ammeters they are all joined in series. 
 The current is adjusted to the proper value by the galva- 
 nometer, and this is marked under the needle of the meter ; 
 the next highest current is then secured and the marking 
 repeated, and so on to the limit of the scale. If of differ- 
 ent current capacities, switches must be provided to cut 
 out each instrument when its limit is reached. Ammeters 
 of very different capacity may have scale cards of the 
 same size, the divisions of the higher reading meter being 
 close t< gather and more in number. If a i5o-ampere card 
 
94 TESTING OF DYNAMOS AND MOTORS. 
 
 have every tenth graduation elongated and numbered, 
 it can be used on a i5-amperemeter without further 
 change, the smaller divisions now indicating tenths of an 
 ampere. On meters used for large currents it is custo- 
 mary for the makers to furnish a shunt to be used with 
 it. The ammeter terminals go to the shunt which is in 
 the main circuit, and the card is scaled off as before. 
 Although but a small current goes through the meter its 
 needle indicates the main circuit current. In the finer 
 Weston meters the shunt is contained within, but on the 
 station instruments the shunt is a separate device, the 
 meter itself appearing more like a voltmeter. An 
 observer may be puzzled by the presence alone of two 
 small flexible cords on an instrument registering perhaps 
 2,000 or 3,000 amperes, the shunt in most cases not being 
 in sight. Where commercial ammeters are turned out 
 in quantities, the galvanometer is replaced by a standard- 
 ized meter which is recalibrated at regular intervals by 
 the galvanometer and shunt. By means of this second- 
 ary standard, calibrating can be done rapidly. Readings 
 above i ampere are taken in steps ascending and de- 
 scending, the two sets of readings differing more or less 
 according as iron does or does not enter into the working 
 part of the instrument. The descending readings will 
 be higher than the ascending ones for the same current, 
 on account of the residual magnetization of the iron. The 
 descending readings are influenced not only by the cur- 
 rent, but by the fact that the iron is more strongly 
 magnetized at the time they are taken. If the current is 
 kept constant for some time, and the instrument tapped 
 with the finger, the needle will return to the ascending 
 deflection for that current value. The mechanical fric- 
 
MEASUREMENT OF CURRENT. 95 
 
 tion of moving parts also has its influence upon the 
 ascending and descending readings, and in the rougher 
 forms of instrument the needle often fails to return to 
 o after the circuit is opened. 
 
 in instruments depending for action upon springs, 
 permanent magnets, electromagnets without iron cores, 
 and the heating and consequent expansion of wires, 
 errors due to residual magnetism are absent. In the 
 well-known Weston instruments, the directing force is a 
 permanent magnet, two delicate springs the restraining 
 force, and a coil, free to turn on jeweled bearings, the 
 deflecting force. The permanent magnet produces a 
 strong magnetic field and protects the needle from 
 external disturbing fields. When handled with care, they 
 should show no variation even after months of use. 
 
 The Ayrton and Perry instruments are examples of 
 the electromagnetic type whose deflecting and directing 
 forces both depend upon the working current. The 
 Cardew instruments depend upon the elongation of a 
 wire when heated by a current. As the wire expands it 
 actuates a roller to which is attached the index. 
 
 Ammeters of any sort, like galvanometers, may have 
 their capacity increased by means of a shunt. The 
 resistance of this shunt depends upon the multiplying 
 power it is desired to give the meter, and is measured by 
 the quotient 
 
 where R 8 is the shunt resistance, R g that of the galvan- 
 ometer, and n the desired multiplying power, or the num- 
 ber of times the meter's capacity is to be multiplied. That 
 
9 6 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 this is so can be seen from the following: We have 
 learned elsewhere that the ratio of the currents in the 
 branches of a divided circuit is the inverse ratio of the 
 resistance of those branches. If in Fig. 31 i\ * 2 , indi- 
 
 cate the currents, r t J\ the 
 resistances of the branches, 
 
 then t\ : * a :: 
 may write it, 
 
 or we 
 
 FIG. 31 
 
 Let the main current be 
 40 amperes, and suppose we 
 
 wish 30 amperes to go through the shunt and 10 through 
 
 the meter; then, 
 
 *', 10 /, r n 
 
 therefore, 
 
 30 
 
 r i\j i ' i 'S v -' i 
 
 - = ~ and ^ : ^ = 3, and r, = 3 r, , 
 
 where r } = galvanometer resistance, and r^ that of the 
 shunt. If 10 amperes is the meter or galvanometer read- 
 ing, when 40 amperes are in the line, the multiplier is 4. 
 Taking the formula above, 
 
 n i 
 
 and substituting for R K its value 3 R^ we get 
 
 dividing through by ^ 8 we have 
 
 n i 
 whence, n i = 3 and n = 4, as above. 
 
MEASUREMENT OF CURRENT. 97 
 
 Take now a general case. What must ^ 8 be to have 
 i/;/ of the line current go through the meter G? For this 
 purpose the combined conductivity of galvanometer and 
 shunt must be known. To get a wire's conductivity, we 
 divide i by the wire's resistance, and in this case 
 
 -'- and ~ 
 *g X t 
 
 are the respective conductivities of the galvanometer or 
 meter, and of the shunt, and 
 
 
 their combined conductivity. The line current is the 
 sum of the branch currents, and the current in any 
 branch divided by the total or line current equals the con- 
 ductivity of the branch divided by the total conductivity, 
 hence, if / t is the line current, and i/n of this goes 
 through the ammeter, the following proportion is true: 
 
 / / -L * 
 
 J t <.. -p 
 
 or 
 
 1* _L 
 
 / t : n 
 
 whence, R^ -f- R % = n R 9 \ subtracting Jt, from both 
 sides, we have, R 9 = n Jt^ K n or R 9 = R t (n i) and 
 
 as above. If n is to equal 10, then ^ B = 1/9 It v etc 
 We must distinguish between the condition that a certain 
 
98 TESTING OF DYNAMOS AND MOTORS. 
 
 fraction of the whole current shall pass through the 
 ammeter, and the condition that the shunt current shall 
 be a certain number of times greater than that in the 
 meter; in the above case where n 10, the main current 
 is 10 times, and the shunt current 9 times, that in the 
 meter. 
 
 The effect of shunting an instrument is to decrease the 
 total resistance of the circuit and thereby increase the 
 total current. In ammeter work this effect is negligible, 
 but in galvanometer work a compensating resistance is 
 put in circuit. If the galvanometer resistance is R%, and 
 that of the shunt R^ their multiple resistance is 
 
 and RZ 
 
 is the resistance of the compensating coil. This coil is 
 used only in very delicate work, for ordinarily the 
 resistance in circuit is so high that any small variation 
 can be neglected. 
 
 To shunt an ammeter, and increase its capacity by any 
 desired amount, is a very simple operation, and can be 
 done without previous calculation. Suppose we have a 
 loo-ampere ammeter, and we wish to use it for reading as 
 high as 500 amperes. The meter must carry one-fifth 
 and the shunt four-fifths the total current. The wire to 
 be used in shunting must carry 400 amperes without 
 heating. In Fig. 32, A and B are the terminals to the 
 meter, A, to which the shunt, S, is to be attached. One 
 end of 6" is fixed at A, and the other end is left free to 
 slide through a clamp at B. Before connecting 6* to B, 
 adjust the current at 100 amperes, then insert S, and 
 draw it through until the reading falls to 20. If this is 
 
MEASUREMENT OF CURRENT. 99 
 
 done quickly we can assume that the current has not 
 
 varied during the operation. Otherwise we must place 
 
 between S and // a 
 
 switch, which can be 
 
 repeatedly opened, to 
 
 insure that the deflec- 
 
 tion is always 100 
 
 when the whole cur- 
 
 rent flows through 
 
 A ', and that it is 20 when S is in circuit. To get the 
 
 line current we have now only to multiply the meter 
 
 current by 5. The strong point in favor of this most 
 
 practical method is that no resistance values need be 
 
 known. 
 
 If circumstances call for a predetermination of the 
 shunt resistance, that of A must be known. Knowing 
 this, and the proposed relation of S's current to A"s, S 
 can be figured. Suppose that with a 2oo-amperemeter, 
 whose resistance is .002 ohm, we are to measure a current 
 of 500 amperes. S must, then, carry 300 amperes, and 
 the relation of A's current to S's, is 
 
 7 A > _ 200 _ 2 
 
 ~ 3 ' 
 
 The ratio of the respective resistances must then, be 
 3 : 2 or 
 
 whence the proportion, 7 A ': / ; * R n : R^ and substi- 
 tuting the known value, we have, 2 : 3 ' R % : .002, or 
 <, = .00133 ohm. All resistance must be measured 
 
100 TESTING OF DYNAMOS AND MOTORS. 
 
 from post to post, so that the joints may be common to 
 both paths, otherwise an error may be introduced. 
 
 In place of a shunt, a second ammeter may be used 
 in multiple with the first. In such a case, although the 
 carrying capacity of the meters in multiple may exceed 
 the current to be read, it does not follow that each will 
 take its proper share; one needle may swing off its scale 
 A and the other read 
 
 low. In order that 
 the meters may read 
 proportional parts of 
 the total current, 
 their resistances 
 must be inversely 
 proportional to their 
 current capacities. 
 For example in Fig. 33 let A and A l be two meters 
 of 100 and 300 amperes capacity respectively, and 
 the current to be read 400 amperes. If A's resistance is 
 three times ^4/s, each will read its full capacity, and the 
 sum will be 400. If A's resistance is twice A^s, A will 
 take 133 amperes, throwing its needle off the scale, while 
 A l will take but 266 amperes, two-thirds of its capacity. 
 To rectify matters a resistance R must be put in series 
 with A l and of such value as to establish the proper rela- 
 tion between the two readings. This can be done either 
 experimentally or by calculation. If A measures .004 
 ohm, and A l .002 ohm, the resistance necessary to put in 
 with A is .002 ohm, for since A's capacity is but one-third 
 A^s, its resistance must be three times A^'s to have the 
 current divide proportionately. If A's is .004, and we 
 add R = .002, we get .006 (= 3 X .002), which is what 
 
MEASUREMENT OF CURRENT. IOI 
 
 we wish. By proportion we have, 100 : 300 ; ; .002 : 
 (.v -|- .004) where x is the value of A'. Multiplying 
 together the means and the extremes we get 
 
 (x -p- .004) 100 = 300 x .002. 
 100 x -f- -4 = -6, and .v = .002 ohm. 
 
 If the adjustment is made experimentally with the meters 
 in circuit, R is varied till ,7's reading is three times that 
 on A^ and this ratio should hold on all parts of the scale. 
 Laws demonstrated for two wires or meters in multiple 
 hold for any number in multiple. Suppose it is desired to 
 replace one wire carrying 10 amperes by three wires which 
 shall carry 7, 2, and i amperes respectively, what must be 
 their resistances? Calling the conductivity of the com- 
 bined wires 10, that of the separate wires must be 7/10, 
 2/10, and i/io respectively. Their resistances must be 
 in the same ratio as the reciprocals of the conductivities, 
 or as 10/7-: 10/2 : 10/1, which is 1.4 : 5 : 10. No par- 
 ticular values can be given as answers to the problem 
 thus stated, since any three resistances which are to each 
 other as 1.4.: 5 : 10, will cause the current to divide in 
 the proper ratio, but if we modify the proposition by 
 assuming the resistance of the original wire to be known, 
 the resistance of each branch can be determined. Thus, 
 if we assume the resistance of the single wire to be 2 
 ohms, then with a current of 10 amperes, the potential 
 difference across their common junction will be (/ x R = 
 10 x 2) 20 volts. Since the fall of potential along each 
 of the branches is to be 20 volts, the resistance of each 
 will be 20 divided by the respective currents. Thus the 
 7-ampere branch must have a resistance of 20/7 ohms, = 
 2.86 ohms; the 2-ampere branch, 20/2 = 10 ohms, and 
 the last 20/1 = 20 ohms. 
 
102 TESTING OF DYNAMOS AND MOTORS. 
 
 The same problem may arise in a different way. Sup- 
 pose we are given three conductors of 2, 3, and 4 ohms re- 
 spectively; when connected in multiple on the line, how 
 will the current divide among them? Suppose the line 
 current to be 10 amperes, with the resistances 2, 3, and 4, 
 the conductivities will be 1/2, 1/3, and 1/4 respectively, 
 and their combined conductivity 1/2 -|- 1/3 -f- 1/4 = 
 13/12; the multiple resistance, being the reciprocal of 
 this, is 12/13 f an ohm. With a current of 10 amperes 
 the potential difference across their junction is 12/13 X 
 10 = 9.23 volts. The respective currents will now be, 
 
 i i = 2^3 _ 4 6j- am peres; / 2 9Ji3 3.07, 
 
 and / = ?-' 5 == 2.31 amperes, 
 4 
 
 and their sum 9.99 amperes, very nearly 10 amperes, the 
 current conditioned. This is as convenient a method as 
 any for solving problems of this character, though gen- 
 eral results may be derived from abstract theory and 
 formulae obtained, in which we need only to substitute the 
 values given. Thus for the multiple resistance of three 
 wires the formula is, 
 
 R = - 
 
 * + 'I 'I + ', ^3 
 
 where r l9 r^ and r 3 are the respective resistances of the 
 three wires. Substituting above, we get, 
 
 R = - * X 3 X 4 = as above< 
 
 2 X3 + 3X4 + 2 X4 13 
 
 Circumstances sometimes arise where the only reliable 
 
MEASURING OF CURRENT. 
 
 103 
 
 ammeter must be sent away for repairs, and some tempo- 
 rary expedient must he resorted to for replacing it. The 
 most practicable plan is to use a voltmeter in connection 
 with a known resistance. In anticipation of this, a piece 
 of copper conductor should be placed in circuit with the 
 ammeter and the drop across it taken for different cur- 
 rent values. Where the voltmeter is too high reading, 
 its calibrating coil can be used. The method is not to be 
 recommended for extensive testing rooms, where the 
 range of current to be read is large, nor is it capable of 
 great accuracy, but for a single machine it suffices to 
 keep the current constant 
 during a heat test, and is 
 especially well adapted to 
 testing series machines, 
 where the series coil can be 
 used as the'standard resist- 
 ance. Fig. 34 shows the 
 connections for using a 
 single voltmeter to read 
 both current and voltage in a street railway motor test. 
 V is the voltmeter, C, one terminal of the calibrating coil, 
 its other terminal being in common with the high voltage 
 terminal at S, T is the trolley, G the ground, F the 
 motor field, A the armature, and R the resistance by 
 means of which we are to read the current. When the 
 full line connection is made, the high voltage coil is in 
 action and the instrument reads the voltage of the line. 
 With the dotted connection, the calibrating coil gives the 
 drop on R. Care must be taken in making connections 
 that the calibrating coil is not subjected to the full line 
 voltage. 
 
 Fir.. 34. 
 
104 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 The methods of current measurement so far given have 
 been those by copper deposition, by the shunted galvanom- 
 eter, the ammeter, and the voltmeter, indirect, but none 
 of these have a universal application. Deposition cannot 
 be used at all for alternating currents, while galvanometers 
 
 and meters must have 
 special features to fit 
 them for this work. 
 Instruments depend- 
 ing upon the heating 
 and consequent ex- 
 pansion of wires may 
 be used; also those 
 depending upon the 
 attraction of a solen- 
 oid upon a piece of 
 soft iron free to move 
 in a magnetic field. 
 The Cardew instru- 
 ments are types of the 
 former, the Edison, 
 Westinghouse and 
 Thomson-Houston of 
 the latter. 
 
 The ideal instrument for use in alternating work is 
 the Siemens dynamometer. The following description 
 is adapted from Ayrton's ''Practical Electricity." The 
 instrument belongs to the class of> spring control meters 
 and its principle is essentially that of the electric motor, 
 one coil corresponding to the field, the other to the 
 armature, the two coils being generally connected in 
 series as are the fields and armature of a series motor. 
 The effect of reversing either or both of the coils is the 
 
 FIG. 35. 
 
MEASUREMENT OF CURRENT. 
 
 same as that seen in a motor under similar treatment. 
 The dynamometer is shown in perspective in Fig. 35 and 
 symbolically in Fig. 36. It consists of a fixed coil 
 A B C D and a movable coil E F G, the latter being 
 frequently made of a single stiff wire. Connection is 
 made to the movable coil by two mercury cups into 
 which its terminals dip. The suspension is by means of 
 a silk thread and a 
 delicate spiral spring 
 resists the turning 
 force due to the cur- 
 rent. A pointer M 
 shows the angle 
 through which the 
 spring is turned by 
 the milled head T, in 
 bringing the coil back 
 to zero. The scale is 
 marked in degrees, or 
 in 400 equal divisions. 
 The turning force of 
 the spring is propor- 
 tional to the angle 
 through which it is 
 turned. The force 
 exerted on the coil 
 is proportional to the 
 product of the cur- 
 
 FIG. 36. 
 
 rents carried by them, so that when the coils are in series 
 the square of the current flowing is proportional to the 
 angle D through which M is turned. 
 
 In symbols, 7 a = A'x D, that is, the square of the cur- 
 rent = the number of divisions D through which m is 
 
106 TESTING OF DYNAMOS AND MOTORS. 
 
 turned x by a constant. Extracting the square root of 
 both sides we get, / = \/A' D. Now as A' is a number 
 we can take its square root, thereby placing it outside 
 the radical sign, where we will call it A; our formula 
 then reads, I A y ' D, where A is that current which 
 will cause a deflection of one division. 
 
 The advantages of this instrument are (i) it is a zero 
 instrument, /'. e. , the coils at the time of taking a reading 
 always occupy the same relative position, so that all 
 observations are made under the same conditions; (2) it 
 is adapted to alternating current work since the currents 
 in both coils agree in phase, thus producing a resultant 
 action that is always in the same direction; (3) no iron 
 being employed in the construction, ascending and 
 descending readings agree and no averaging is necessary. 
 The disadvantages are: (i) the moving coil must be 
 brought to zero before a reading can be taken, so that 
 rapid variations of current are not indicated; the current 
 must remain steady for a time before a reading can be 
 taken; (2) it is not dead-beat, i. e., the pointer does not 
 quickly come to rest, so that readings cannot be taken rap- 
 idly; (3) its readings are influenced by neighboring 
 magnetic fields,' and when the dynamometer current is 
 considerable the earth's field exerts an influence, so that 
 the instrument must be set up with the plane of the sus- 
 pended coil at right angles to the magnetic meridian; (4) it 
 is not portable in the ordinary sense of the word; (5) the 
 coils are exposed and liable to injury and the swinging 
 coil is influenced by air currents; (6) the scale being grad- 
 uated in degrees or arbitrary divisions the instrument is 
 not direct reading; (7) it does not indicate the current's 
 direction, which is necessary in some work. 
 
MEASUREMENT OF CURRENT. 107 
 
 In conclusion it may be said that the Siemens dynamom- 
 eter is a valuable standard instrument when used under 
 fixed conditions, as in a laboratory, but there are other 
 instruments better adapted to portable work. Tests, 
 however, could be cited where the dynamometer has 
 given good service under very unfavorable conditions. 
 In a test on a three-phase alternator a Siemens dynamom- 
 eter was used to check up the readings of three ammeters. 
 In addition to the alternators ten direct current machines 
 were used as auxiliaries, and in the midst of these was the 
 table of instruments. The dynamometer readings com- 
 pared very favorably with those of a newly calibrated 
 ammeter. 
 
 In reading small currents great care must be taken 
 to eliminate sources of error liable to modify results, 
 as the error might in such cases amount to a consider- 
 able portion of the whole current. With the currents 
 usually employed in machine testing the results may 
 be relied upon even under very unfavorable conditions. 
 Accompanying every instrument is a table giving the 
 current value corresponding to each graduation of the 
 torsion head; this table is determined by the formula 
 I A <J D, where A is the same for all parts of the scale, 
 and can be experimentally found by means of a standard 
 ammeter in series with the dynamometer to determine the 
 value of I : then 
 
 The table can thus be checked up. It is well to stan- 
 dardize the dynamometer from time to time if it is carried 
 around very much or is in constant use. 
 
 It may sometimes be desirable to connect the two coils 
 
I08 TESTING OF DYNAMOS AND MOTORS. 
 
 in parallel instead of in series. In such a case the cur- 
 rent divides between the two coils in the inverse ratio of 
 their resistances, and the turning force exerted upon the 
 coil is proportional to the product of the two currents. 
 The formula then becomes / A' ^/ D ', where A has 
 a new value different from that with the coils in series. 
 The deflection JD' is now due to two factors which have 
 different values also. By placing resistance in series 
 with one of the coils its current may be varied and its 
 deflecting influence altered. The current in the other 
 coil is also altered and its deflecting power affected. 
 When the sum of the two currents is constant, the total 
 deflecting force as represented by their product will be a 
 maximum when the currents are equal. Thus if the total 
 currents be 10 amperes it may divide in various ways as 
 given in the table and the maximum effect is seen to be 
 when /j = 1 2 = 5 amperes. When this is the case the 
 indication on the dial will be the same as if the current 
 of either coil were flowing through the two in series; but 
 with coils in series the current in either would be the 
 total current, while when in multiple, under the above 
 condition, it is but half; and the indication on the dial 
 must be multiplied by 2 to get the true value of the cur- 
 rent flowing. 
 
 7 in No. i. 
 
 The table shows that when /, = / a the instrument is 
 most sensitive, /'. e. t gives the largest deflection for a 
 
 7 in No. 2. 
 
 Tot. 7. 
 
 Deflec Force. 
 
 2 
 
 10 
 
 8 X 2 = 16 
 
 4 
 
 10 
 
 6 X 4 = 24 
 
 5 
 
 10 
 
 5 X 5 = 25 
 
 6 
 
 10 
 
 4 X 6 = 24 
 
 8 
 
 10 
 
 2 X 8 = 16 
 
MEASUREMENT OF CURRENT. 109 
 
 given current. The advantage gained by the multiple 
 connection is to increase the capacity of the dyna- 
 mometer without increasing the current through its 
 coils. With an alternating current the multiple connec- 
 tion is not to be recommended, as the difference in the 
 self induction and capacity of the two coils may throw 
 the two currents out of phase and thereby render their 
 deflecting effect less than the product of the two currents. 
 
 Another instrument adapted equally well to direct or 
 alternating current work is the Thomson or Kelvin 
 balance, which has the advantage of being more portable. 
 An objection to its use save in special work i*s its narrow 
 range. As it belongs rather to the class of high grade 
 laboratory instruments than to commercial types, the 
 reader is referred to the various laboratory manuals for 
 its description (Carhart or Gray). 
 
 We will close our consideration of current measure- 
 ment by touchingon the theory of the electro-dynamom- 
 eter used as a measurer of work, /. e., the wattmeter. 
 We have as an expression of the watts consumed in any 
 portion of a circuit, E X / where E is the potential 
 difference between the ends of the circuit and / the cur- 
 rent flowing. Now if the heavy coil of an electro-dyna- 
 mometer be placed in the main circuit, the deflecting 
 force due to this coil will be proportional to the line 
 current /. If the other coil be wound for high resist- 
 ance, and placed across the mains or across that portion 
 of the circuit under consideration, its defecting force 
 will be due to , and proportional to it. The resultant 
 deflecting force must then be proportional to E X /, the 
 watts. 
 
 For testing-room work the dynamometer is an invalua- 
 
110 TESTING OF DYNAMOS AND MOTORS. 
 
 ble instrument. For commercial purposes, however, a 
 wattmeter is necessary that will make a continuous 
 record of the energy consumed on a circuit, precisely as 
 a gasmeter records the consumption of gas. 
 
 This need is filled by the various recording wattmeters, 
 of which the Thomson recording meter will serve as a 
 type. The instrument consists of a simple skeleton 
 motor having no iron in either armature or field. The 
 fields, which consist of two coils one on either side of 
 the armature, are connected in series with the lamps or 
 motor whose absorption of energy they are to measure, 
 all the current in use passing through them. The arma- 
 ture, which is of what is familiarly known as the "Sie- 
 mens-drum " class, having a small silver commutator and 
 brushes, with silver contact pieces, is in shunt across the 
 line, like a lamp. It has a considerable resistance of 
 fine wire in series with it, and thus forms a "pressure 
 coil " whose current varies with the voltage. 
 
 Now as the field varies with the current, and the arma- 
 ture with the E. M. F., the speed of the motor will vary 
 directly with the product of the current and the E. M. F., 
 thus measuring the watts passing through the circuit. 
 
 In order to secure accuracy the influence of friction 
 must be compensated for. This is accomplished by 
 taking off the armature circuit beyond the fields, thus 
 securing a slight constant field independent of that caused 
 by the armature current though the field coils. On very 
 low loads, where friction is a strong factor, this constant 
 field forms a considerable part of the total field, while on 
 medium or full load both the friction and the constant 
 field are but small factors, and may be neglected. 
 
 A device is also adopted to cause the armature to rotate 
 
MEASUREMENT OF CURRENT. Ill 
 
 at a low speed, thus reducing the wear on the instru- 
 ment. This is very neatly accomplished by placing on 
 the armature shaft a thin disc of copper, which rotates 
 between the poles of three permanent magnets. The 
 fields of these magnets generate eddy currents in the 
 disc, and the work thus done forms a drag on the motor 
 proportional to the speed. 
 
 The indications of the meter are shown in watt-hours 
 on the dial, whose train of wheels engages with a worm 
 on the shaft, thus connecting it with the revolutions of 
 the armature. The calibration of the meter is accom- 
 plished by moving the poles of the permanent magnets 
 toward or away from the periphery or centre of the disc, 
 thus regulating its speed, until it is synchronous with the 
 standard instrument. 
 
CHAPTER V. 
 
 MEASUREMENT OF ELECTROMOTIVE FORCE. 
 
 LEAVING the subject of current and its measurement, 
 we pass to the study of that influence, electromotive 
 force, which sets the electricity in motion and produces 
 the current. An E. M. F. has been defined as that which 
 causes or tends to cause a current flow. The word force 
 is not here used in its mechanical meaning, for electricity 
 is not matter in the usual sense. An E. M. F. is analo- 
 gous to water pressure and potential difference (P. D.) is a 
 term corresponding to difference of level. P. D. is inde- 
 pendent of the current or resistance in a circuit just as 
 the difference of level between the two ends of an inclined 
 pipe is independent of the water flow, or of the obstruc- 
 tion offered. P. D. thus defined is the total E. M. F. 
 of a circuit, and cannot be limited to mean the fall of 
 potential between any two points; as, for example, the 
 terminals of a battery. If a battery develops a P. D. of 2 
 volts, measured on open circuit by a static voltmeter, 
 there will still be 2 volts on closed circuit, but there 
 will now be a loss of potential through the battery's 
 internal resistance and the terminal P. D. will be less 
 than the total E. M. F. The static voltmeter, never clos- 
 ing the circuit, actually measures the battery's total P. D. 
 An ordinary voltmeter practically realizes this condition 
 in that its resistance is very high, and the current so 
 small that the internal loss of potential can be neglected. 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 113 
 
 Looked at a little differently, the battery and meter are 
 in series and the total E. M. F. distributes itself accord- 
 ing to the resistance: the meter resistance is so great 
 compared to that of the battery that the drop through 
 the battery can be neglected. 
 
 U'hile the total E. M. F. is independent of resistance or 
 current, the ''drop" between any two points of a closed 
 circuit depends upon both. This has its hydrostatic 
 analogy. If through a pipe of uniform section a stream 
 of water flow, the fall or decrease of pressure will be 
 uniform, as would be shown by inserting pressure gauges 
 at regular intervals' along its length. If an obstruction 
 be placed in the pipe, as, for example, a lot of sponges, 
 the greatest fall of pressure takes place from one side 
 to the other of this obstruction, which, if so effectual as 
 to entirely stop the current flow, will cause all gauges on 
 the pump side to register full pressure and those beyond 
 the obstruction no pressure at all. Under a natural cur- 
 rent flow, the gauge nearest the pressure side of the 
 pump registers highest and each succeeding gauge a less 
 amount. So in an electric circuit, if composed of a uni- 
 form homogeneous conductor, the P. D. per unit of length 
 of circuit will be everywhere the same; but if of variable 
 section or composition the greatest drop will take place 
 through the unit length of highest resistance. If at any 
 point the circuit be opened the resistance here becomes in- 
 finite, and across it will take place the total P. D. of the 
 source. These facts are expressed quantitatively in Ohm's 
 law, E IR, where E is the drop across a portion of the 
 circuit of resistance, R, with a current, /, flowing. Since 
 /is the same for the whole circuit the drop, , is seen to 
 be everywhere proportional to R. If R = o, indicating 
 
114 TESTING OF DYNAMOS AND MOTORS. 
 
 the two points to be coincident, E = o. If 7? increases 
 more and more E does likewise and reaches a maximum 
 when R = oo, /. e., when circuit is opened. In this 
 case, / = o. When instead of absolute open circuit we 
 have a voltmeter across the break, there flows a small 
 current. Upon. this depends the utility of voltmeters, 
 test-lamps, magnetos, etc., for locating an open circuit. 
 The test consists in successively placing the terminals 
 across different parts of the circuit; when the instru- 
 ment includes the break, the circuit is completed and 
 the lamp, bell, or needle shows it. In such testing it is 
 necessary to have the resistance of the test circuit large 
 to obviate abnormal flows of current through the devices. 
 It may seem unnecessary to say that a magneto must 
 not be placed across a break in a live wire, nor must we 
 expect a voltmeter or lamp circuit to give results across 
 a break in a dead one, unless the test lines themselves 
 include a source of voltage. The voltage in use dictates 
 the meter or the number of lamps that should be in series. 
 With new men there seems to be a weakness due to 
 thoughtlessness, not ignorance, for placing a low voltage 
 meter across a high voltage circuit, or putting a single 
 lamp across several hundred volts, or using a live test 
 circuit to locate a fault in some other live circuit. 
 , The commonly used sources of E. M. F. are primary 
 batteries, storage batteries, and dynamos. In commer- 
 cial testing we have most to do with primary batteries 
 and dynamos. Of batteries, the Daniell's and Leclanche 
 are most met with, the former in closed, the latter in 
 open circuit work. In closed circuit work the battery 
 carries a current continuously, but in open circuit work, 
 only when the circuit is closed by the button or switch. 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 
 
 IT 5 
 
 
 t 
 
 
 h 
 
 
 
 
 p 
 
 E 
 
 
 1 
 
 
 
 i 
 
 , 
 
 _WOODEH 
 COVER 
 
 GLASS VESSEL 
 
 FOflOUS POTS 
 
 FIG. 37. 
 
 The Daniell cell is a double fluid one, having a zinc plate 
 clipping into zinc sulphate and a copper plate dipping 
 into copper sulphate while a porous partition separates 
 the two fluids; more commonly each plate occupies a 
 porous cup containing the proper fluid, and both are 
 placed in a vessel of zinc sulphate. Fig. 37 shows a 
 cell much used as a 
 
 standard in testing ^ ^b COV.ECTORS 
 
 rooms. In a glass 
 vessel porous cups are 
 placed and a wooden 
 cover serves the dou- 
 ble purpose of pro- 
 tecting the solution 
 and holding the plates 
 in place. Three 
 points recommend the 
 
 cell as a standard of E. M. F., viz., (i) It has but 
 little tendency to polarize; (2) its voltage is nearly 
 unity; (3) its E. M. F. is little modified by temperature 
 variations. To understand what is meant by polariza- 
 tion of a cell, we must know that as soon as the circuit 
 is closed and chemical action begins, hydrogen gas is 
 set free at the zinc plate, and if allowed to accumulate 
 sets up an E. M. F. opposed to that of the cell. If the 
 current used is small, the gas is taken up chemically by 
 the sulphate liquid, thus keeping the plate free. A well- 
 prepared Daniell cell is, if carefully handled, free from 
 polarization, and the constancy is even improved by 
 allowing the cell, when not in use, to send a small cur- 
 rent through a high resistance, thereby preventing the 
 mixing of the copper and zinc sulphates by diffusion. 
 
Il6 TESTING OF DYNAMOS AND MOTORS. 
 
 A single cell or dynamo or any number of cells or 
 dynamos permanently arranged may be regarded as a 
 single cell or dynamo of increased voltage or current 
 capacity, or both, and will yield the largest current on 
 short circuit: /. e. , any given arrangement of cells or 
 dynamos, constituting a battery, will give the most cur- 
 rent when the external or outside resistance, R o, and 
 the internal resistance, r, alone limits the current value. 
 For 
 
 7=_A_ 
 
 and if R o, corresponding to short-circuiting the 
 battery terminals, then 
 
 must have its greatest value because r will go into E 
 more times than R -\- r will. Assuming no secondary 
 or armature reactions to exist, cells and dynamos do the 
 most work when short circuited, because 1 is then greatest 
 and E remains the same, and work per second = / E. 
 But none of this work is useful; it is wasted in internal 
 heating and is a total loss. Since efficiency, 
 
 useful work 
 
 total work 
 
 7; is here o for there is no useful work, though the heat- 
 ing of the cell shows that much work is being wasted. 
 If from o, the external resistance be gradually increased, 
 the amount of work done in the external circuit, and with 
 it the efficiency, increases. This external or useful work 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 117 
 
 becomes a maximum when r = R and here ;/ = 50 ^. 
 Work per second, as we have seen, EL If / is 
 the same throughout, the work done in different parts of 
 the circuit must depend upon the distribution of E. 
 Now ^distributes itself according to the resistance, and 
 since R = ; each must have a drop equal to 
 
 E 
 
 and each is the seat of work per second in amount 
 equal to 
 
 '*? 
 
 i. f., out of a total work per second = / ., one-half is 
 useful and // is, therefore, 50 %. Further, if by vary- 
 ing R, the efficiency be made above or below 50 % the 
 useful work per second will decrease: this is because 
 a given change of relation between r and R has a 
 greater per cent, effect upon the distribution of work 
 between them than it has upon the total work done. 
 Ex. : Let E = 2, R = i, r = i. Then 
 
 , 2 
 
 / = - = i ampere. 
 
 The total work done per second is S = 2X1 = 2 
 watts. The useful work is 
 
 '- X I = i X 1 = 1 watt. 
 Now let R be reduced to 1/2 ohm. Then, 
 
 r 2 24 
 
 / = - T- = j- = ampere, 
 i + 1/2 3/2 3 
 
Il8 TESTING OF DYNAMOS AND MOTORS. 
 
 and the total work per second is 
 
 4 8 
 
 - X 2 = = 22/3 watts, 
 
 <3 o 
 
 and the useful work 
 
 248 
 = - - X = watt, 
 339 
 
 as opposed to i watt before the change; showing that 
 although the total work has been increased 33 % the 
 useful work has been reduced n $, and the total 
 effect has been to not only decrease the useful out- 
 put per second but to decrease rj as well. Let us next 
 suppose R to be increased to i 1/2 ohm, everything else 
 being the same as before. 
 
 2 224 
 
 / = - 7- = . = r = - - ampere, 
 
 I + I 1/2 2/2 + 3/2 5/2 5 
 
 the current: 4/5 x 2 = 8 '5 = i 3/5 watt, the total work 
 per second: 3/2 x 4/5 = 12/10 = 6/5 = drop across R 9 
 and 6/5 x 4/5 = 24/25 watt, the useful work per 
 second. So we see the effect in toto has been to decrease 
 the total work, increase the relative amount of useful 
 work, and hence the efficiency. That is to say that 
 although there is not as much work done, of that which 
 is done a larger per cent, is useful. 
 
 We must carefully distinguish between the problem of 
 finding under what condition a given cell or battery 
 generates maximum current, and the problem, how must 
 a given number of cells be arranged to give maximum 
 current through a fixed external resistance. The answer 
 to the first is R o to the second, R = r. In the first 
 case r is fixed, R is not; in the second case the reverse 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 119 
 
 is true. In other words we might take a number of cells 
 and arrange them in a certain way, and whatever way 
 that happened to be, the battery would give greatest 
 current on short circuit; but unless by chance we struck 
 the right combination, we could rearrange the cells to 
 give a greater current by fulfilling the condition A' = r. 
 The greater R compared to /-, the greater the efficiency 
 at which the battery works, and when R becomes very 
 great ?; is very nearly 100 , but we cannot say that 
 the efficiency is greatest when R infinity, correspond- 
 ing to open circuit, for then / is o, and although the 
 battery is not taking in or wasting any work, neither 
 is it giving out any useful work, and the expression for 
 efficiency 
 
 useful work 
 
 
 
 total work 
 becomes 
 
 an indeterminate quantity. 
 
 A cell's internal resistance depends upon the density 
 of the solutions, the size of the plates, their distance 
 apart, and the thickness and composition of the porous 
 pots. On the other hand the E. M. F. is independent 
 of dimensions and depends solely upon the nature, purity, 
 and density of the acids, and on the nature and purity of 
 the plates. The use of impure metals gives rise to local 
 action detrimental to the E. M. F. and durability of the 
 cell. To insure homogeneity in the plates the zinc 
 should be amalgamated, and the copper electroplated. 
 Before amalgamation the zinc must be thoroughly 
 cleansed with scratch brush and buffer, then dipped for 
 
120 TESTING OF DYNAMOS AND MOTORS. 
 
 a moment in sulphuric acid. Next pour mercury into a 
 flat tray with dilute sulphuric acid, and with a rag wash 
 acid and mercury together on to the zinc, until its surface 
 is uniformly bright. Where dirt spots exist the mercury 
 and zinc will not unite. After amalgamation wash in 
 running water to remove all trace of acid. The copper 
 plate is prepared after the manner of the voltameter 
 plates already given. The influence of solution density 
 and condition of plates is as follows: 
 
 Increase in density of copper sulphate increases 
 the E. M. F. 
 
 Increase in density of zinc sulphate decreases the 
 E. M. F.* 
 
 Oxidation in zinc surface decreases the E. M. F. 
 
 Oxidation in copper surface increases the E. M. F. 
 
 The presence of oxide is to be particularly guarded 
 against by electroplating immediately before using the 
 cell, otherwise an error of as much as 2 % may result. 
 Fleming recommends the following as standard solutions: 
 
 1. Copper sulphate saturated at 15 C. density 1.2. 
 
 2. Zinc sulphate saturated at 15 C. density 1.2. 
 This combination will with clean plates give an E. M. F. 
 of very nearly 1.102 volt. A second receipt is: 
 
 1. Copper sulphate saturated at 15 C. density i.i. 
 
 2. Zinc sulphate saturated at 15 C. density 1.4. 
 Giving an E. M. F. of 1.072 volt. If the cell is allowed 
 to stand an hour or so its voltage will rise about .003 
 volt, while the presence of any impurity on the plates 
 will lower it 2 or 3 <f c . 
 
 Says Professor Carhart of this cell: "The many pre- 
 cautions required to insure a normal E. M. F. in a 
 
 * Carhart's Primary Batteries, pp. 100, 103. 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 121 
 
 standard Daniell cell, on every occasion of its use, are 
 more than an offset to a negligible temperature coeffi- 
 cient in comparison with that of a Clark cell." Inas- 
 much, however, as the Daniell cell can be constructed in 
 any testing room while the standard Clark cell can only 
 be procured by purchase, the former will no doubt hold 
 its own as a standard easily made and easily replaced. 
 
 Where the Daniell is used as a standard it is well to 
 always make up two cells, in order that they may be 
 checked up by comparison with each other. If one causes 
 the galvanometer to show a deflection of 1 10 divisions, the 
 other, under similar conditions, should do the same; with 
 a series connection and the cells in opposition there 
 should be no deflection. 
 
 For open circuit work such as bells and signals and in 
 some zero methods of testing, the Leclanche cell is much 
 used. It is not suited for general testing on account of 
 its rapid polarization when giving a current. To appre- 
 ciate this effect it is only necessary to press a bell button 
 for a few moments, when the bell will finally cease to ring. 
 The cell will however regain its former K. M. F. if allowed 
 to stand for a while, because the hydrogen gas accumu- 
 lated disappears, and with it, its back E. M. F. The 
 structure of this cell is as follows: In a glass jar contain- 
 ing a solution of sal ammoniac is placed a zinc rod con- 
 stituting one terminal of the cell. The other terminal, 
 a carbon plate, is in a porous cup, and is surrounded by a 
 mixture of powdered carbon and manganese dioxide; the 
 office of the porous cup is to hold the dioxide and carbon 
 closely about the carbon plate. The liquid diffuses 
 through the cup and the cell is ready for use. The object 
 of the dioxide is to oxidize the hydrogen as fast as it is 
 
122 TESTING OF DYNAiMOS AND MOTORS. 
 
 liberated and thus to free the cell from polarization. Ex- 
 cepting for small currents, the action of the manganese 
 dioxide is not rapid enough to do away with all polariza- 
 tion, but a cell soon recovers itself when allowed to stand 
 idle. A milky color of the solution indicates that the 
 salammoniac needs renewal, and it is best to entirely 
 replace the old solution with new.* 
 
 Having briefly spoken of the sources of E. M. F. we 
 may now pass to its measurement. The galvanometer is 
 the backbone of all instrumental work. In the testing 
 room it is desirable to have the instrument arranged to 
 measure current, potential ' difference, and resistance. 
 To fulfill these varied requirements, no galvanometer is 
 better adapted than a high resistance reflecting Thomson 
 galvanometer provided with suitable shunts, directing mag- 
 net and scale. Such an instrument is qualified to read the 
 drop on field'or armature, to give insulation resistance from 
 zero to millions of ohms, or to indicate current values be- 
 tween limitsfixed onlyby the shunt in use. The requisites 
 which a good galvanometer should fulfill are as follows: 
 
 (a) An astatic magnetic system' of small weight. 
 
 (b) A variable magnetic control. 
 
 (c) Four coils of nearly equal resistance and effect. 
 
 (d) High insulation and resistance. 
 
 The first is secured by constructing the needle as 
 follows f : a No. 15 aluminum wire from 2 1/2" to 3" long is 
 flattened at both ends, and to it, at a distance apart equal 
 to \he distance between the centres of the coils, mica 
 
 * The subject of Primary Batteries is very ably treated in a little work 
 by Professor Carhart, and the more ambitious reader is referred to this 
 work for further information. 
 
 f Carhart's Elect. Meas., p. 145. 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 123 
 
 discs are attached, by means of beeswax. Pieces of small 
 piano wire 3/8" to 1/2" lung are permanently magnet- 
 ized and are placed on the back of the discs, as shown in 
 Fig. 38, with their poles opposing each other. 
 Finally the needle is suspended by a silk' fibre, 
 and by drawing a magnet over the system it is 
 made perfectly astatic (/. e. , so that the yVand ^V 
 poles exactly balance). This condition is indi- 
 cated by the needle finally setting itself E and 
 IT instead 'of N and S. To understand this 
 action we must look more fully into the mean- 
 ing and advantages of an astatic combination. 
 In an astatic combination one needle has its 
 TV 7 , the other its 6" pole pointing in the same 
 direction. The result is that any tendency 
 of one needle to place itself in regard to the 
 earth's field is opposed by the counter ten- 
 dency of the other needle to which it is rigidly 
 attached. The earth's field then having an 
 equal but opposite influence on the two needles, has no 
 influence on the system as a whole, other than to cause 
 them to take up the - and }V position; in other words, 
 this peculiar behavior is due to the fact that the needles 
 though of equal and opposite strength are not perfectly 
 aligned, and the resultant effect of the earth's field sets 
 them east and west.* A perfectly astatic balance is seldom 
 attained and is seldom needed in the more practical lines 
 of work. A perfect balance deprives the needles of all 
 directing force and makes the system unsteady. To 
 complete the suspended part a small round mirror is 
 attached with wax to one of the mica discs. 
 
 * See Ayrton's Prac. Elect., p. 282, note. 
 
1*4 TESTING OF DYNAMOS AND MOTORS. 
 
 The directing magnet, which stands directly over the 
 coils and can be either rotated or moved up and down, 
 secures the (b] condition of variable magnetic control. 
 The object of this is two-fold: i. To render the system 
 independent of the earth's field, or of other disturbing 
 fields. 2. To change the galvanometer's sensibility, /. <?., 
 to alter the deflection which a given current causes, and 
 thereby varying the galvanometer's reading range. 
 
 In the testing rooms of large electrical works where 
 moving masses of iron and charged fields of varying 
 intensity act as disturbing influences it is customary to 
 shield the galvanometer with several soft iron rings sur- 
 rounding the coils in the horizontal plane of the needles. 
 Sometimes the instrument is put in an iron box provided 
 with a hole through which the mirror reflects the ray. 
 The iron gathers to itself any wandering lines of force 
 and frees the needle from their influence. (It is for this 
 reason that iron cases protect watches from magnetism.) 
 With an unprotected galvanometer, care must be taken 
 that masses of iron in the immediate neighborhood of the 
 galvanometer are not disturbed during the test, nor must 
 the directing magnet be touched. For if the needle is 
 adjusted to o under influences which are afterward 
 changed the adjustment is destroyed, and error intro- 
 duced. Where the disturbing influences are very strong, 
 as in commercial testing rooms, it is found best to have 
 the controlling magnet well down over the needle; this 
 also lessens the error likely to arise from want of adjust- 
 ment in the earth's meridian. Another precaution to be 
 taken is to be sure that the suspension fibre is not under 
 torsion when the ray of light is at o; on delicate instru- 
 ments this torsion should be negligible since the maker 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 125 
 
 is supposed to eliminate it as far as possible, by using 
 only the finest of silk, or a quartz fibre. The presence of 
 error due to torsion can only be proven ultimately by 
 making the needle as sensitive as possible, adjusting<-Aw//y 
 in the earth's meridian and noting if the deflections on 
 opposite sides of o are equal; provided other sources of 
 error may be assumed to have been eliminated, the 
 inequality may be charged to torsion and, if appreciable, 
 the suspension changed. 
 
 To facilitate adjustment in the earth's field, the mag- 
 netic meridian should be determined as accurately as 
 possible, and laid off on the testing table; this can be 
 done by allowing a compass needle to come to rest, sight- 
 ing along its A* and 6" poles, and projecting its direction 
 line on the table: in doing this external magnetic influ- 
 ences of local character should be removed, and the 
 position of all magnetic bodies be that in which it is 
 intended the work shall afterward be carried on. A 
 second -A 7 ' and S line must now be laid off at a distance 
 from the first equal to the intended distance between 
 needle and scale. These lines are each laid off by means 
 of the compass, and if both observations have been cor- 
 rect the lines will be parallel. An /sand /F line must 
 now be drawn and the galvanometer needle placed over 
 one intersection, the zero of the scale over the other. 
 In mapping out the lines the tester must bear in mind 
 that the aluminum pointer on most home-made compasses 
 is at right angles to the needle itself. It may happen 
 that the mirror is not mounted exactly parallel to the 
 needles and the scale may have to be shifted or the lamp 
 moved to bring the ray to o. In all galvanometer work 
 vibrations of the needle due to jarring greatly hinders 
 
126 TESTING OF DYNAMOS AND MOTORS. 
 
 rapid readings. The best results are obtained by plac- 
 ing the instrument on a stone pier built up from the 
 ground, and entirely free from the building. If this is 
 impracticable there should be set into the wall of the 
 building a shelf long enough to accommodate the galvano- 
 meter and scale. The wall chosen should have its length 
 in the direction of the meridian, so that the shelf may be 
 long and narrow. With regard to the instrument itself 
 it is necessary that the needle should be at the centre 
 of the coil; if : t is not, not only will the coil's influence 
 vary for different positions of the needle, but a more 
 evident annoyance will be a swinging motion of the needle 
 each time the circuit is closed, due to the needle's being 
 more powerfully attracted to the nearer side of the coil. 
 Proper adjustment is secured by making the fibre the 
 right length and then accurately leveling the instrument 
 by means of screws provided for the purpose. On high 
 grade galvanometers is found a plumb line attachment. 
 On other instruments the eye alone is used to centre the 
 needle. In regard to the (<r) and (d) requisites above 
 cited: they are in a degree akin to each other. On in- 
 struments having symmetrically disposed coils, these 
 must be electrically balanced so that their effects will be 
 the same; and for a given size bobbin and a given size 
 wire, condition (c) is fulfilled if the electrical resistances 
 are the same, for we may then be reasonably sure that all 
 spools have the same number of turns. Specification (d) 
 for high resistance is a technical way of saying that 
 the instrument must be wound with fine wire so as 
 to get on a great many turns, and thereby make it very 
 sensitive: for, to give it a high resistance merely for its 
 current limiting effect, would be much less economical 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 127 
 
 than to introduce resistance outside of the coils. High 
 insulation is necessary to the successful operation of 
 any instrument. Finally, there should be adequate 
 protection from dust and moisture. The former is 
 secured by keeping the galvanometer under cover; the 
 latter, by placing near it a box of chloride of lime or an 
 open vessel of sulphuric acid, either of which will absorb 
 moisture. The presence of dust not only invites a short- 
 circuit in its more manifest forms, but by absorbing 
 moisture, and affording a leakage path, affects the value 
 of the shunt and thereby introduces error. 
 
 Wires leading to and from the galvanometer should be 
 twisted together, so as to neutralize each other's mag- 
 netic effect upon the needle. These wires should also be 
 secured to a stout insulator before passing to the instru- 
 ment, thus securing it against injury should the lines 
 receive an accidental jerk. Convenient to the operator's 
 right hand should be a reversing switch, and a key for 
 closing the galvanometer circuit. It is very convenient 
 to have the two combined in one device. 
 
 The knack of taking readings rapidly on both sides of 
 o, is acquired only by practice in the use of the key. 
 Thus if the ray be at o and the key be closed, the ray 
 swings immediately and the needle's inertia carries the 
 ray far beyond the true deflection; if the key is kept 
 closed some time will elapse before the ray comes to 
 rest. If, however, as soon as the ray is well started up 
 the scale the key be opened, the needle stops almost 
 immediately and the ray begins to descend; at this 
 instant the key may be again closed: by repeatedly 
 opening and closing the key at the right instant the ray 
 is soon brought to a stand. The same tactics may be 
 
128 TESTING OF DYNAMOS AND MOTORS. 
 
 pursued in returning the ray to o, and in taking reversed 
 readings. 
 
 As a precaution against danger in handling, as well as 
 against error, it is well, where high voltages are used, to 
 cover all keys and switch handles with hard rubber, and to 
 mount the tester's stool on insulators. Since it is unsafe 
 to introduce high voltage into the galvanometer circuit, it 
 is customary when measuring such voltages to use pro- 
 portion lines. Proportion lines are a device by which 
 a known fraction of the external voltage is applied to the 
 galvanometer circuit. In using this device the operator 
 is still liable to shocks incidental to a grounded circuit, 
 but is not liable to the full voltage by getting his hand 
 across the key when it is open. Again, by using these 
 lines the higher voltages are read without the necessity 
 of placing in the galvanometer circuit very expensive 
 resistance boxes to reduce the current to a readable deflec* 
 tion. It is true that, outside high resistances must be 
 used, but here it is necessary to limit the current to per- 
 haps 30 amperes, while in the galvanometer circuit it must 
 not rise above say 30/10,000 of an ampere. The propor- 
 tion lines consist of a resistance box of such a value and 
 capacity as to carry without undue heating the current 
 sent through it by the voltage to be measured. Across 
 a known fraction of this resistance are connected lines 
 to the galvanometer which measures the P. D. included. 
 The total voltage is to the P. D. measured as the totai 
 resistance is to that included by the lines or V t : V Q \ * 
 R : r, whence 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 
 
 129 
 
 where V g is the P. D. measured by the galvanometer. In 
 direct current work where potential differences exceed- 
 ing 700 or 800 volts are seldom met with, this method 
 can be used throughout the whole range of commercial 
 work; higher voltages, usually found on arc machines, 
 are handled by increasing the resistance, R t . Nor will 
 the heating of the box, R t , affect the result if it is 
 made throughout of the same material so arranged as 
 to insure uniform ventilation. As R t increases, due to 
 rise of tempera- 
 ture, r does like- 
 wise, and at the 
 same rate, and re- 
 mains the same 
 fraction of R t as 
 when cold. If, 
 however, one part 
 
 FIG. 39. 
 
 be of copper and the other of German silver, or any other 
 metal than copper, the device will give correct results 
 only for the temperature at which it was set up, for all 
 metals differ in the amount of their resistance variation 
 for a given change in temperature. 
 
 It is practicable to arrange a single proportion box 
 for reading over a wide range of voltage, but it is better 
 to have two or more boxes standardized between their 
 limits. Fig. 39 shows the connections for a box 
 intended to read any voltage up to 2,000. For reading 
 up to 125 switch A"" is closed and the other switches 
 remain open. For 250, K" is closed, the rest being open, 
 etc. Since the resistance, off which the galvanometer G 
 takes the drop, remains constant, it becomes a smaller 
 and smaller fraction of R t each time R t is increased, so 
 
130 TESTING OF DYNAMOS AND MOTORS. 
 
 that a new constant must be used with each switch. Sup- 
 pose, for example, that the resistance cut in by K'" is 
 1,500 ohms and that the galvanometer reads the drop from 
 15 ohms. Suppose also, that by means of the standard 
 cell the galvanometer has been set up to read 100 divisions 
 for a P. D. of i volt at the terminals of its circuit. Now 
 if across 1,500 ohms there is a drop of 125 volts, across 15 
 ohms, which is i/ioo of 1,500 ohms, there will be a drop 
 of i/ioo of 125 volts = 1.25 volt; further since i volt will 
 deflect the ray 100 divisions, 1.25 volt will deflect it 
 1.25 x ioo divisions = 125 divisions. That is to say the 
 scale is direct reading since for every division deflection 
 there is i volt applied to J? t . From proportion, 125 : 
 V t \ \ 1,500 : 15; or V g 1.25 volt as before. Also i 
 volt : 1.25 volt \\ ioo divisions : x. Whence x = 
 125 divisions as explained above. Next suppose the 
 voltage under measurement to be 250 and suppose that 
 K' 1 cuts in an additional 1,500 ohms. The galvanom- 
 eter circuit has not been disturbed and we now have, 
 250 : V g \\ 3,000 : 15; or V g 1.25 the same as 
 before, because although we have doubled V t we have 
 also doubled R t , and since the resulting current is the 
 same so also must the drop across the given resistance, r, 
 be, and hence the deflection 125. But we know (by 
 mean of a voltmeter, if necessary) that V t = 250; i 
 division therefore means 2 volts on the line and the scale 
 is no longer direct reading, but its indication must be 
 multiplied by 2; 2 is the constant for the 250 volt 
 lines. In delicate measurements on such a box care must 
 be taken not to introduce error by using the 125 volt 
 lines a long while and suddenly switching in the 250 volt 
 lines; for the galvanometer then reads the drop from 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 
 
 a combined cold and warm total resistance, which is not 
 what it may ordinarily be supposed to be. It is desir- 
 ble to have r (15 ohms in the above case), a sub-multiple 
 by 100, 200, 300, etc., of the successive values of R. 
 Thus if r 15, R x == 1,500; R^ = 3,000; R^ -- 4,5, 
 etc. Then if the galvanometer is set up to read 100 
 divisions per volt the scale reading need only be multi- 
 plied by i, 2, 3, etc., to give the line voltage. 
 
 A neat way to secure a direct reading scale is to have 
 several scales ruled one above the other on the same 
 
 card, so mounted as to slide 
 
 up and down so as to bring 
 the ray on a level with the 
 scale desired to be used. 
 Each scale should be num- 
 bered to read volts direct, 
 and, to avoid confusion in 
 reading, colored inks can be 
 used in graduating the several scales. This is illustrated 
 in Fig. 40, whose scale was prepared for the proportion 
 box described above and in which the ratio of r to R 
 was successively i/ioo, 1/200, and 1/300. 
 
 One proportion box can be easily calibrated by means 
 of another; thus suppose it is desired to calibrate a 1,000 
 volt box by means of a 125 volt box. First, by means 
 of the 125 volt box, adjust the line voltage to 100 volts; 
 then apply this known voltage of 100 to the terminals of 
 the 1,000 volt box, and observe the deflection; we can 
 now do one of two things: either rroject the 100 volt 
 scale graduation of the lower scale onto the high volt 
 scale and there mark it 100, or, when practicable, we can, 
 by shifting the connections Df the galvanometer leads, 
 
 FIG. 40. 
 
132 TESTING OF DYNAMOS AND MOTORS. 
 
 readjust the resistance r included by the galvanometer, 
 till a deflection of 10 is obtained on the 125 volt scale. 
 This is then projected to the upper scale and marked 100. 
 The latter method is preferable as the readings in the last 
 case have simple ratios to each other. Next secure a 
 reading of 200 on the 125 volt box, and apply the 200 
 volts as before to the 1,000 volt box. The higher scale 
 reading should now be twice its original value, and any 
 departure indicates an error, most probably in observa- 
 tion. Having secured two points, the rest of the scale 
 can be divided into equidistant graduations. The value 
 of r, once determined, should not be changed even 
 though R is varied. It is well to have r composed of 
 double rather than single wire, for should a single wire 
 burn off or break, the galvanometer would be thrown 
 into the main circuit, instead of being a shunt to it, there- 
 by inviting injury to the instrument. 
 
 In using high resistance boxes care must be taken that 
 they are not exposed to moisture, for the adjustment of 
 a dry day may prove to be useless on a wet one, as the 
 resistance becomes practically short-circuited through 
 leakage paths, and it may be impossible to reduce the 
 deflection to a readable amount. If comparison with the 
 standard cell shows such a condition to exist, the boxes 
 may be subjected to the mild heat of a couple of incan- 
 descent lamps until under normal conditions the ray shows 
 the proper deflection for a standard voltage. 
 
 When using a galvanometer the tester must remove from 
 his person all magnetic articles such as knives, watches, 
 keys, etc. A not infrequent source of annoyance often 
 overlooked is the shifting of iron window weights, and 
 with high grade laboratory instruments the magnetic 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 133 
 
 disturbances coincident with auroral displays render 
 observations impossible. These precautions should be 
 particularly observed when the instrument is adjusted to 
 its maximum sensitiveness, as it is for delicate work. 
 Ordinarily the directing force of the earth's field is weak 
 enough to allow sufficiently large deflections, but if neces- 
 sary it can be weakened to any degree by placing the 
 directing magnet with its N and S poles opposing the 
 earth's field, and at the proper distance, determined by 
 trial, above the instrument. If the magnet on reversal 
 be placed too near the needle, it will not only neutralize 
 the earth's field but establish one of reversed polarity, a 
 condition indicated by the needles turning halfway round 
 in obedience to the magnet's attraction. It may happen 
 that the magnet is too strong even when at its highest 
 point on the supporting rod; in such a case, either the rod 
 must be lengthened, a weaker magnet used, or a second 
 magnet so disposed as to help the earth's field. 
 
 The wiring of a galvanometer service is a field for judg- 
 ment and care; on the switchboard are brought out the 
 terminals of all circuits used in measuring, and in a cen- 
 tral position should be found the galvanometer terminals. 
 All live wires should be placed on one side of the gal- 
 vanometer terminals, all dead wires on the other; the 
 galvanometer terminals being understood to include the 
 galvanometer circuit complete resistances, switches, etc. 
 Potential lines should never be used for measuring 
 resistance nor resistance lines as potential lines. Failure 
 to observe this rule will result, sooner or later, in injury 
 to the galvanometer or tester. In plugging in proportion 
 lines be careful that the lines including r, and not R, go to 
 the galvanometer; confusion will result not only in pos- 
 
134 TESTING OF DYNAMOS AND MOTORS. 
 
 sibly burning out r, but in subjecting the galvanometer to 
 the total line voltage with probably fatal results. As a 
 certain preventive of confusion the plugs from the R 
 terminals can be made so that they cannot go into the gal- 
 vanometer connection plate. Galvanometer and switch- 
 board lines should be placed on porcelain insulators to 
 avoid leakage, which will otherwise seriously modify re- 
 sults when measuring insulation. Rubber-handled flexible 
 cables should be used in connecting the galvanometer to 
 the various shop lines. In making switchboard connec- 
 M N tions, plug in the live 
 
 r~~^~]-*-/^yr~~{ r~ n-^/^Y r~~l blocks last; in breaking 
 \^ i#'\ y(r/ */ E connections, unplug 
 
 ^"\ .^^^ .^ the live blocks first; 
 
 ^ ' JUUUUL, ^* tnen should a cable 
 
 drop from the hand 
 
 FlG - 4I - there is little danger 
 
 of shock or short circuit. For similar reasons, it is well, 
 in making new connections, to clear the switchboard of 
 those used in a previous test. And in changing connec- 
 tions never plug a live wire or a grounded wire into one side 
 of a circuit unless its other side is free. Thus suppose 
 (Fig. 41) dynamo N to be sending current through resist- 
 ance R; suppose further that in order to accommodate a 
 fellow tester who is in a hurry, we are to transfer R to 
 dynamo M. Suppose, as is too frequently really the case, 
 N. to be accidentally grounded at g and M at g'; the 
 result of connecting B and C or A and D would be to 
 ground both sides of N in the first case and M in the 
 second, causing a dead short circuit. The safest way to 
 avoid trouble is to disconnect N entirely before connect- 
 ing M : however, if it so happens that the tester gets the 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 135 
 
 grounded sides together there will be no demonstration. 
 The above frequently arises in the hasty and apparently 
 complicated wiring of large testing rooms. 
 
 We are now ready to set up the galvanometer to read 
 voltage. All shunts are removed and the magnet 
 secured midway on its post. The galvanometer is then 
 placed in position at the intersection of one of the A" S 
 and the E IV lines already laid out. It is also leveled, 
 and with the eye aligned approximately in the meridian. 
 The scale and lamp are set in place and a clear ray 
 secured. We next adjust the coils to the meridian; this 
 is done by removing the magnet and allowing the ray to 
 come to rest under the earth's influence alone. The scale 
 is then moved till the ray rests on o. The galvanom- 
 eter is now slightly rotated on a vertical axis until the 
 deflection, due to a small fraction of a cell's voltage, is 
 the same on both sides of o. This low voltage can be 
 obtained by placing a piece of wire across the cell's termi- 
 nals and taking the drop from an inch or more of it. 
 The magnet is now replaced and the ray brought to zero 
 under its influence, by means of the worm and screw 
 generally supplied for that purpose. 
 
 The next step is to determine what deflection a known 
 voltage at the galvanometer circuit terminals will pro- 
 duce. It is always best to make the deflection 100 
 divisions per i volt or .1 volt or .01 volt, etc., accord- 
 ing to the sensitiveness of the galvanometer; the read- 
 ings are thus much simplified. For ordinary commercial 
 testing the standard Daniell cell is used, but for more 
 responsible work it is customary to at least compare the 
 home-made standard with some guaranteed cell such as 
 the Clark or Carhart-Clark standard cell. These cells 
 
136 TESTING OF DYNAMOS AND MOTORS. 
 
 are made up in lacquered brass and ebony finish, and 
 are provided with a thermometer to facilitate accurate 
 temperature corrections. Their temperature correction 
 is negative; /. ., the E. M. F. falls as the temperature 
 rises. Another standard cell is the Calomel, invented 
 by Von Helmholtz. The E. M. F. of this cell increases 
 as the temperature rises, making its temperature cor- 
 rection positive. This makes possible a combination of 
 Clark and Calomel cells free from temperature correction. 
 Such sets of cells self-correcting between certain tem- 
 peratures can be secured from the manufacturers. It is 
 fatal for such cells to be short-circuited, because the 
 current decomposes the paste of which the cell is made. 
 To preclude such a mishap there is placed inside, and 
 in permanent connection with the cell, a high resistance 
 coil. This coil does not appreciably affect the cell's 
 terminal voltage, because as a rule the cell is used only 
 in zero methods, and when it does send a current it is 
 through such very high resistance that that within the 
 cell can be neglected. 
 
 The Clark cell has an E. M. F. of 1.434 volt at 15 C, 
 and this E. M. F. decreases .00115 volt per degree C. 
 rise of temperature up to 30 C. and increases at the 
 same rate below 15 C. The cell should never be used 
 at temperatures below 10 C. or above 30 C. The 
 temperature coefftciency of the Carhart-Clark cell is 
 about y 2 that of the Clark, while that of the Calomel 
 is -j- .000094 volt per degree C. above 15 C. up to 
 30 C. : it is about -fa of that of the Clark cell; t. e., the 
 Clark cell's E. M. F. decreases .00115 vo ^ P er degree 
 rise while the Calomel cell increases .00115 volt -=- n = 
 .000104 P er v lt degree under similar conditions. From 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 137 
 
 this it follows that n Calomel cells in series with i 
 Clark will give a self-correcting battery. 
 
 For most forms of industrial testing a Daniell cell 
 gives satisfaction. Where a finer standard is available 
 the E. M. F. of the Daniell should be compared with it, 
 and then used as a secondary standard. Assuming this 
 to be the case, and the Daniell's E. M. F. to be 1.07 
 volt, we now place the galvanometer with its resistance 
 boxes, containing 100,000 ohms, constant, and 10,000 
 ohms, variable, in series, and no shunt; the directing mag- 
 
 s 
 
 FIG. 42. 
 
 net and resistances are adjusted until a deflection of 107 
 is obtained. This will mean i division per .01 volt, and 
 i volt per 100 divisions. With a scale of 250 divisions on 
 either side of o, anything over 2.5 volts will throw the 
 ray off the scale. For voltages higher than this the 
 galvanometer must be shunted or the proportion lines 
 used. On high voltages the shunt itself is liable to 
 injury unless the extra resistance in circuit with the 
 galvanometer at least equals that given above. At all 
 events be certain that the shunt shunts only the gal- 
 vanometer and not the boxes also, for this would place 
 the shunt across the line and a burn-out would follow. 
 Provided galvanometer and shunt are both in series with 
 a high resistance as in Fig. 42, the danger of short cir- 
 cuit is removed and the shunt may be safely used. With 
 the 1/99 shunt a voltage of 100 would give a deflection of 
 
138 TESTING OF DYNAMOS AND MOTORS. 
 
 100, because with no shunt i volt gives 100 divisions; with 
 the 1/99 shunt only i/ioo of the current goes through 
 the galvanometer and i volt will give but i/ioo of the 
 deflection that it gives without the shunt, or i/ioo of 100 
 divisions = i division. Therefore 100 volts would give 
 a deflection 100 times that due to i volt or 100 divisions; 
 or the galvanometer is direct reading and using the 1/99 
 shunt is equivalent to using the 125 volt proportion lines. 
 With the shunts any voltage can be read up to 2,500, 
 using the 1/999 Shunt. By using more than one shunt 
 at a time, not a general practice, this range is somewhat 
 increased. With a combination of shunts and propor- 
 tion boxes the carrying capacity of R would alone 
 limit the range of the galvanometer. 
 
 One of the most useful applications of the above " set- 
 ting up " with a Daniell cell is in calibrating commercial 
 voltmeters; the voltmeters are arranged on a table at 
 a safe distance from each other, and by means of their 
 push buttons successively connected in multiple with the 
 galvanometer circuit. Note that we say galvanometer cir- 
 cuit, because, were we using proportion boxes, the meters 
 would be placed in multiple not with G but with R, the 
 resistance to which the line voltage is applied. A read- 
 ing is adjusted on the galvanometer, and its value marked 
 on the scales of the meters. All the meters are success- 
 ively graduated for each galvanometer reading until 
 their entire scale is calibrated. As fast as the ranges of 
 the lower reading meters. are attained the instruments 
 are cut out of circuit. Instruments under calibration 
 should be slightly tapped with the finger to release the 
 needle should it be inclined to stick. In some types of 
 voltmeter it is unnecessary to use the full voltage to 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 139 
 
 calibrate the scale, there being provided what is called 
 
 a calibration coil. Its principle is seen in Fig 43. 
 
 A B C 'is the movable coil carrying the needle. In 
 
 series with this coil is the 
 
 usual high resistance R; 
 
 coming clown to binding 
 
 post No. 2 is a lead which 
 
 includes but a fraction, ;-, 
 
 of the total resistance R. 
 
 The calibrating coil is that 
 
 part of R included between FlG - 43> 
 
 posts No. 2 and No. 3. From No. i to No. 3 is the work- 
 
 ing coil. The principle is that of the proportion box. 
 
 r is so related to R, that when a specified fraction of 
 
 the voltage which applied to i and 3 gives a certain 
 
 deflection, is applied to 2 and 3, the same deflection 
 
 obtains. This means that in either case the current 
 
 through A B C is the same and hence the deflection is 
 
 the same. Let i be this current and let E and e be the 
 
 voltages respectively necessary to apply to R and r to 
 
 produce /; then 
 
 i = -f- = - r : R\\c : r 
 K r 
 
 whence E \ e \\ R : r. Hence if r = R -r- 10, then 
 must e E -f- 10, and by applying this lower voltage to 
 the calibrating coil, r, the whole scale can be marked off. 
 Since all measurements of E. M. F. with the galvan- 
 ometer depend upon the accuracy of the standard cell, 
 great care should be used to preserve this accuracy. 
 The best methods are on this account zero methods which 
 
140 TESTING OF DYNAMOS AND MOTORS. 
 
 involve no current flow, and which are free from tor- 
 sional and other disturbances incidental to deflectioa 
 methods. Fig. 44 illustrates a zero method due to Pog- 
 gendorf. B generates the E, M. F. to be measured 
 
 and it sends a small current 
 through RR. Fis a stand- 
 ard cell in series with which 
 is a low resistance galvanom- 
 eter G, a variable high resist- 
 FlG - 44 ' ance r, and switch, K. One 
 
 terminal of the galvanometer circuit is fixed at R, and 
 the other is free to slide along R R . Fis opposed ia 
 polarity to B. Resistance R R is an exposed German- 
 silver wire of known length and uniform cross-section, 
 mounted on a board graduated in inches or cms. 
 The cross-section is uniform if the P. D. per unit of 
 length is everywhere the same. Should it be not the 
 same, the resistance of each unit length must be found 
 and marked on the underlying scale. With r in circuit 
 P is moved along R R till ^Tcan be closed without affect- 
 ing G. r can now be gradually cut out and a finer ad- 
 justment secured. The condition then existing is this: 
 From R to P is a certain P. D. which by trial is 
 made equal to that of V. Since the drop of potential is 
 proportional to the resistance, the drop from RtoP bears, 
 the same relation to the total drop, as resistance R P 
 does to the total resistance. But the total drop is the 
 voltage of B, and the total resistance, neglecting that 
 of B, is R R'. Therefore 
 
 Voltage R to P : Voltage of B \ \ Resistance RP : 
 Resistance R R' or E y : E*\\ Res. R P : Res. R R'\ 
 whence 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 14! 
 
 Res. R R' 
 * ~ Res. R P 
 
 If the cross-section of R R is uniform, the resistance of 
 each of the parts into which R R may be divided will be pro- 
 portional to its length; therefore ZiV '. E^\\ R P" \ R R n 
 and 
 
 B = v x 
 
 This zero method of comparison is used to calibrate 
 high grade voltmeters. Any number of meters can be 
 put in multiple with R R. j9's voltage is gradually 
 raised, measured at each step, and marked on the dials 
 of the meters under test. This device has another very 
 useful testing-room application, nor should it be confined 
 to testing-room practice. If we know F's voltage and 
 R ^"s resistance, or if of uniform section, R J?"s length, 
 />'s position for any value of B can be calculated from 
 the formula, 
 
 s E. M. F. = x F's voltage; 
 
 or 
 
 If v = i and R R' = 100 where must P be for E^ = 50 
 volts ? 
 
 RP = ~XR' = X ioo = 2 inches 
 E* 5 
 
 or cms., as the case may be. In this way the whole 
 distance can be laid off, and any desired impressed 
 
142 TESTING OF DYNAMOS AND MOTORS. 
 
 E. M. F. can be secured by placing P on the proper 
 point and varying B till balance obtains. The chief 
 liability to error lays in the cell's temperature variation. 
 In working with any slide wire device care must be 
 taken that the slide P does not cut or scrape the wire 
 and thereby introduce error by diminishing R fi"s cross- 
 section. To avoid having an inconvenient length of 
 wire in R R' a resistance box is often used in connection 
 with it and sometimes entirely replaces it. The objec- 
 tion to a box is that R R' 
 
 H -^ K 
 
 and R P must be ex- 
 pressed in ohms and not 
 in lengths, whereas the 
 delicacy of the method 
 FIG. 45. rests largely on the 
 
 fact that a -length can 
 
 be more easily and accurately measured than a resist- 
 ance: furthermore the temperature of the exposed 
 wire will probably be lower than that of the box; 
 this, however, is of minor importance as the current 
 should not be large enough to raise the temperature of 
 any part of the circuit. When boxes alone are used, 
 as delicate a balance cannot be secured since points 
 between plugs are not considered. Fig. 45 gives con- 
 nections for the box method. Here, instead of the slide 
 wire, two boxes are used, and balance effected by vary- 
 ing the resistance in each. The total resistance of both 
 boxes becomes R R' and that of r, R P. The same 
 equations hold. To facilitate calculation it is con- 
 venient to keep r l -f r a constant; this is done by plug- 
 ging as much resistance into one box as is taken out of 
 the other. There is no limit to the voltage readable 
 
MEASUREMENT OF ELECTROMOTIVE FORCE. 143 
 
 by this method, the single restriction being that A' ' 
 be always sufficient to prevent E^ from generating too 
 large a current. In dealing with very high voltages 
 "multipliers" are used. The same idea is utilized in 
 the so-called "multipliers" supplied (by request) with 
 Weston voltmeters, to increase their reading range. 
 These multipliers consist of a resistance box with a 
 known simple ratio to the meter's resistance and is 
 placed in series with it. To get the impressed voltage 
 the deflection must be multiplied by a number, known as 
 the "constant," stamped in a conspicuous place on the 
 case. Since it is desirable that the constant be a simple 
 number the ratio of resistances must be a simple one; 
 a box suitable as a multiplier for one meter will not be 
 conveniently so for most others, the constant only 
 holding good for meters of the same resistance. The 
 same number is therefore stamped on meter and multi- 
 plier. Let R v be the voltmeter's resistance and R m 
 that of the multiplier. If ^ v = R m the same drop will 
 take place across each, and we know the meter indi- 
 cates 1/2 the impressed voltage, so that the constant is 2. 
 If R m = 2 R v , or 
 
 the drop across the meter is 1/2 that across the multi- 
 plier and 1/3 the total drop and the constant is 3. In 
 this case, 
 
 2_R L _X I \ 
 
 Jtv)- 
 
 and this is the formula used in determining the constant 
 
144 TESTING OF DYNAMOS AND MOTORS. 
 
 when the multiplier is made. The general expression is 
 
 = n 
 
 where n is the constant, and the impressed E. M. F. 
 equals n times the indicated voltage. 
 
 If there is no standard multiplier convenient any box 
 of known resistance can be used if the meter resistance 
 is known. The constant n will not in this case be a 
 simple number, unless it so happens. 
 
 The above principles have been applied in producing 
 an instrument which can be used both as an ammeter 
 and a voltmeter. A paper read before the American 
 Institute of Electrical Engineers, March 17, 1894, by 
 E. G. Willyoung, describes such an instrument. 
 
CHAPTER VI. 
 
 MEASUREMENT OF RESISTANCE. 
 
 HAVING outlined the methods of current and E. M. F. 
 measurement, we pass to that of resistance, the Last of 
 the three factors found in Ohm's law. Resistance is that 
 which checks back, or hinders the flow of current through 
 .a conductor. Its analogue in mechanics, or rather 
 hydraulics, is friction. The presence of resistance is 
 looked upon as an unavoidable evil, and generally such is 
 the case; its absence, however, would be sometimes incon- 
 venient, for resistance alone limits the current which an 
 E. M. F. shall send, and should a circuit's resistance ever 
 become zero, the smallest E. M. F. conceivable could 
 send through it a current of infinite value. Take a given 
 circuit and suppose its resistance to be gradually 
 decreased. The current produced by a given E. M. F. 
 is 
 
 and as R grows smaller and smaller the fraction 
 
 R 
 
 grows larger and larger until, when R becomes zero, 
 
 7- 
 o 
 
 145 
 
 
 
146 TESTING OF DYNAMOS AND MOTORS. 
 
 or infinity, written, oo . As soon as a current flows in a 
 wire lines of force spread out from the wire, and in doing 
 so cut the circuit at different points, generating therein 
 an E. M. F. opposed to that which urges the current. 
 The impressed E. M. F. has then to oppose, and do work 
 against this opposing tendency, until when the current 
 reaches a steady value, the lines of force cease to move, 
 there is no more self-induction, and the field is steady. 
 
 We see then that even had wires no ohmic resistance it 
 would require work to bring a current up from zero to 
 any given value. Were all conductors without resistance 
 a current once set up would continue forever: there 
 being nothing to stop it. 
 
 There is an interesting application of this idea in the 
 theory of magnetism. If we assume a molecule of mag- 
 netic substance to be of zero resistance, and further 
 assume that by some means a current has been set up 
 around the molecule, this current would continue for- 
 ever. The molecule would then be a permanent electro- 
 magnet. This is one of the theories advanced to explain 
 magnetism. 
 
 The study of resistances falls under two heads : i. The 
 resistance of conductors. 2. The resistance of non- 
 conductors or insulators. The first is exemplified in the 
 resistances of armatures, coils of different sorts, rheo- 
 stats, leads, trolley wires, feeders, etc. ; the second in 
 the insulation power of such substances as mica, paper, 
 rubber, glass, oils, air, etc. Conductors are either solid 
 or liquid. For measuring resistances the instruments 
 used are those already described. The galvanometer, or 
 its equivalent, the battery and the standard resistance, 
 and in some insulation measurements the condenser and 
 
MEASUREMENT OF RESISTANCE. 
 
 '47 
 
 electrometer. All measurements are based on methods 
 dependent directly or indirectly upon Ohm's Law; the 
 conditions being so arranged that the unknown resist- 
 ance is balanced against a known resistance. 
 
 For low resistances, the method of the *' Comparison 
 of Potentials" is much used, and it 
 depends upon the fact that when an 
 E. M. F. is applied to a circuit this 
 E. M. F. in its fall from the positive 
 to the negative terminal distributes 
 itself according to the resistance it 
 meets; where there is the greatest 
 resistance, the greatest P. D. exists; 
 if two or more parts of a circuit are 
 of the same resistance then must 
 the P. I), across them be the same 
 if their currents are the same. To insure that the cur- 
 rent in both shall be the same, the standard resistance 
 and that to be measured are placed in series as in Fig. 46, 
 where B is the source, of current, R a standard resist- 
 ance, and .r, that to be measured. G, is a high resistance 
 galvanometer, S, its shunt, and A, a device for connect- 
 ing G to read the P. D. across R or ..r at will. A con- 
 sists of a hard wood block 6" X 4*, with six 3/4" holes 
 |]=nrj=r) bored half through its face. Let in from the 
 ^ edge and attached to copper lugs on the bottom 
 FIG. 47. Q f tnese holes are Xo. 8 or No. 10 soft copper 
 wires provided with connectors or carried to binding posts. 
 The holes are then filled three-quarters full of mercury. A 
 rocking block is used for cross-connecting and is made as 
 follows: Two pieces of Xo. 10 wire are bent as shown in 
 Fig. 47, with the middle leg longer than the end ones; the 
 
148 TESTING OF DYNAMOS AND MOTORS. 
 
 legs are at such a distance apart as to fit into the three 
 sets of holes. The two pieces are fastened to a small 
 block so that they move together. Since the middle 
 legs are longest, the block will rock upon them, and 
 according as the block is depressed on the right or 
 left, will the galvanometer be placed across x or ft 
 respectively. 
 
 The test consists in sending the same current through 
 R and x and then comparing the deflections due to their 
 respective resistances. Thus, call D l the deflection due 
 to R and Z> a that due to x. 
 
 Then Z>, : Z> 2 ; ; R : x, and 
 
 x - * 
 - 
 
 Now since Z> 2 -=- D l is simply a ratio, it is not necessary 
 to know the actual voltage value of a deflection, but it 
 is necessary that the deflection be proportional to the 
 voltage applied: /. e., we must be certain that if i volt 
 causes a deflection of 20 divisions, 2 volts will cause a 
 deflection of 40. The method is a reliable one and is 
 universally used for measuring fields, armatures, shunts, 
 and transformer coils, in fact any low resistance. The 
 galvanometer must be a reliable one, such as a Thomson 
 mirror galvanometer, and R must be an accurate standard. 
 The method is adapted to locating crosses and open cir- 
 cuits in armatures. In winding and connecting armatures 
 it is possible that a section may get a turn too much or too 
 little; a drop of solder or a metal filing may lie across 
 neighboring commutator bars, or turns of wire; rough 
 handling of insulation may have resulted in the crossing of 
 two wires, or the whole armature may have been wound 
 
MEASUREMENT OF RESISTANCE. 
 
 149 
 
 an inferior quality or wrong size of wire. Under any 
 of these conditions, or those of an open circuit or loose 
 connection, the galvanometer deflection will show dis- 
 crepancies when compared with those taken on a similar 
 armature known to be sound. 
 
 The connections for the test are shown in Fig. 48, 
 where A is the armature, B the source of current 
 (usually several storage cells), R the* 
 standard resistance, r a variable re- 
 sistance to assist in regulating the 
 current 7, and G the galvanometer. 
 By means of the galvanometer and 
 standard R we first get, by varying 
 /, a deflection which experience has 
 taught us is convenient to compare 
 with that caused by the resistance 
 of A. 
 
 On two-pole machines the service 
 
 lines are held to opposite commutator bars as shown in 
 Fig. 48. On four-pole machines these lines must include 
 one-fourth the circumference of the commutator, since 
 this is the distance from the -f- to - - brushes when the 
 machine is running, while on machines of n poles the 
 portion to be included is 
 
 i 
 n 
 
 of the circumference. Care must be taken that the galva- 
 nometer lines make good contact on the commutator. 
 The latter must be free from dirt, oil, and shellac, and 
 the lines had best make contact with the commutator 
 itself rather than with the brushes, thus preventing the 
 
TESTING OF DYNAMOS AND MOTORS. 
 
 resistance of brush contact raising the indicated resist- 
 ance of the armature. In practice there is a frame so con- 
 structed as to hold four brushes, two for the main current, 
 and two smaller ones for drop lines. To throw the galva- 
 nometer alternately to R and 
 A we may use either a three- 
 way switch or the device of 
 Fig. 47. Taking the drop on 
 J? and A the latter can be 
 figured as above. By thus 
 measuring the resistance be- 
 tween opposite bars and 
 
 knowing what the resistance should be, or by simply 
 comparing the deflections with the known proper deflec- 
 tion, any error in winding or insulation or loose con- 
 nection can be detected. For example, in Fig. 49 
 suppose an open circuit to exist in coil No. 7. That 
 half of the armature will carry no current and the resist- 
 ance between terminals being doubled, there now being 
 but one path from brush to brush instead of two, the 
 deflection will be much higher than it ought to be. If the 
 break is in the winding itself the de- 
 flection will be abnormal all round the 
 commutator; but if in one of the con- 
 necting wires as at a in Fig. 50, the 
 needle will show no abnormal deflec- 
 tion except when this bar is under the 
 brush, for at all other positions the 
 break has no influence. Should the 
 the wrong number of turns per section, or the wire be a 
 little hard or above or below gauge, the fact will be shown 
 by a uniformly "off" deflection too high if the effect 
 
 FIG. 50. 
 winder get on 
 
MEASUREMENT OF RESISTANCE. 151 
 
 has been to increase the armature resistance, too low if it 
 has been diminished. 
 
 To locate a faulty section, be it a cross in the coil or 
 commutator, a wrong number of turns, a loose connec- 
 tion or open circuit, the test is somewhat modified. The 
 main current enters and leaves the armature at the same 
 points as before, but the 'galvanometer lines touch on 
 adjacent bars and therefore include but a single section, 
 and the drop on each section is taken. This is known 
 as the bar to bar test, and is very effective. All arma- 
 tures should be subjected to this test before the binding 
 wires are put on, as frequently the fault can be easily re- 
 moved and a burn-out avoided. In passing from bar to bar 
 it will soon be seen what the normal deflection is going 
 to be, and any marked deviation indicates a defect. A 
 "high" deflection indicates a loose connection or open 
 circuit, or possibly a turn too much in the section. In 
 case of open circuit the galvanometer will show no deflec- 
 tion until its lines span the bars which include the break, 
 that is if the break is in the winding itself. Where the 
 winding itself is not brought down to the commutator, 
 but is continuous and has leads tapped on, the galvanom- 
 eter will show a deflection until one of its lines rests on 
 the bar to which the broken lead runs, when the deflec- 
 tion will be zero. A "low" deflection indicates a cross 
 or short circuit more or less serious. This, we have 
 seen, can be caused by filings, solder, or wires touching 
 through abraded insulation. 
 
 The "bar to bar " test, when the galvanometer lines 
 include only one section, is open to the objection that 
 it will not detect crosses between adjacent sections, 
 but this objection can be eliminated by including be- 
 
TESTING OF DYNAMOS AND MOTORS. 
 
 tween the leads two sections instead of one. Other 
 things being unaltered the normal deflection should be 
 twice that due to a single section. If in any case it 
 is not, the sections can be tested individually; if each 
 is all right in itself the trouble must be between them. 
 The test can be further modified so as not only to 
 apply the galvanometer leads to adjacent bars but the 
 current leads also. The objection to this arrangement 
 is that on ring, or other armatures where the connecting 
 leads are tapped on to the main wind- 
 ing, the test would not reveal a broken 
 lead without opening the main circuit. 
 On armatures wound with copper bars 
 whose outer surface constitutes the 
 commutator, a fault can be located to 
 within an inch or so. The method is 
 also valuable for locating grounds in 
 an armature. The connections are 
 shown in Fig. 51. As in the measure- 
 ments of armature resistance the current enters and leaves 
 at the same points as when the armature is in actual ser- 
 vice. The machine of Fig. 51 is a bipolar, so that the 
 current enters and leaves at opposite points of the com- 
 mutator. B is the source of current, G is a galvanometer, 
 one of whose terminals is attached to the armature shaft, 
 the other being free to move along from bar to bar on 
 the commutator. Suppose a ground to exist at a; i. e., 
 through some defect in insulation the wire touches the 
 iron core. Ordinarily the battery current enters the 
 armature at M, flows around the two halves and leaves at 
 TV 7 , the single ground at a having no effect whatever. 
 If now the free galvanometer line be touched to any part 
 
 FIG. 51. 
 
MEASUREMENT OF RESISTANCE. 153 
 
 of the commutator the latter becomes grounded in two 
 places through the defect and through the galvanom- 
 eter. . The galvanometer now, being a shunt to part of 
 the armature, carries current, and indicates the fact by a 
 deflection. The galvanometer really gets the drop on 
 the armature wire included between the fault and the 
 galvanometer lead which touches the commutator. Now 
 if the free galvanometer lead be moved around the com- 
 mutator it brings the two grounds nearer together, so 
 that the galvanometer gets less drop, and the deflection 
 grows less and less, until when the lead reaches a, both 
 galvanometer leads are at the same potential, and the 
 deflection becomes o. When the free lead, ^, passes a 
 the deflection rises again but with reversed direction.* 
 This is taken advantage of where the galvanometer is 
 not very sensitive, for then equal deflections can be 
 gotten on opposite sides of o, and the fault located 
 halfway between these points. Having thus located the 
 ground the current brushes M N should be shifted and 
 the ground again located. The two determinations 
 should coincide. This precaution is necessary, because 
 between M and N there are two paths and the 
 drop of potential through both is the same. For every 
 point in path M a N there is a point in Me N at the 
 same potential; so there is a point c corresponding to 
 point a, where the fault exists, and this point c will give a 
 zero deflection also. When M and N are shifted the 
 point c shifts but a does not. There will be a new point 
 at the same potential as a, but a being common to both 
 tests is the point sought. 
 
 A second method of measuring moderate resistances is 
 called the " Vienna Method," and is especially applicable 
 
154 TESTING OF DYNAMOS AND MOTORS. 
 
 where the portion of circuit to be measured is in service, 
 as in the case of burning lamps or fields on running 
 
 machines. Fig. 52 gives con- 
 
 1||||| 1 A I ^ r- 1 nections for measuring a 
 
 burning lamp. The source 
 of current, B, ammeter, A, 
 and lamp, Z, are in series, 
 
 Lf 
 
 FlG while voltmeter, V, reads the 
 
 potential difference across the 
 
 lamp. If 7 be the current and E the P. D. then by 
 Ohm's law the resistance is 
 
 .. 
 
 Strictly speaking there is an error due to the fact that 
 V's presence in shunt with Z, lowers the resistance 
 between the lamp's terminals, for the resistance of Z 
 and Fin multiple must be lower than that of Z alone, 
 and hence the P. D. across the two is less than it would 
 be across Z. On the other hand, decreasing the circuit 
 resistance increases Z, so, on the whole, considering the 
 law, 
 
 we see that decreasing E and increasing Z lowers the 
 indicated value of Z's resistance. In all but most par- 
 ticular work the results obtained are satisfactory, for 
 since the meter resistance is from 15,000 to 75,000 ohms 
 it shunts but little current from the lamp resistance of 
 250 ohms. If it is desirable to allow for this, the correc- 
 tion can be made in either of two ways. F's current can 
 be calculated from K's resistance and indicated voltage, 
 
MEASUREMENT OF RESISTANCE. 155 
 
 and this value subtracted from the amperemeter to 
 obtain the true value of /. E divided by this cor- 
 rected value of / gives Z's re- B 
 sistance. Or the connections 
 may be changed to those of 
 Fig- 53> where V reads the 
 voltage across both A and Z, 
 while A reads only Z's current. 
 Knowing A's resistance and cur- 
 rent the drop through it (/ ;- A ) can be figured and sub- 
 tracted from F's indication to get the true drop across Z. 
 The choice between the connections of Figs. 52 and 53 
 depends upon how accurately A's and F's resistances are 
 known. Absolute freedom from these corrections can be 
 secured by using an eleetrometer for reading the voltage. 
 The electrometer's action depends upon the fact that two 
 objects which are statically charged at different poten- 
 tials, attract each other: if one is free to move it will do 
 so. An electrometer has two metal plates opposite each 
 other, one of which carries an indicator and is free to 
 move. One plate is attached to the one point, the other 
 plate to the other, of the two points, between which the 
 P. D. is to be measured. By means of a voltmeter and its 
 multiplier the electrometer is calibrated; thereafter it can 
 be used to 'measure voltage and has the advantage that 
 since no current passes through it does not disturb the 
 conditions under which it is desired to be used. They are, 
 however, only used for measuring very high voltages. 
 The main advantage of the foregoing method over that 
 of comparison of potentials is that no standard shunt is 
 required. On the other hand the comparison method 
 does not require a galvanometer calibrated to read volt- 
 
'56 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 eoo 
 
 FIG. 54. 
 
 age. The choice between the two would then depend 
 upon what standard is available. The following- method 
 is an adaptation of the above and requires no calculat- 
 ing. It is useful in checking up 
 resistances that are almost certain 
 to be right, or in measuring where 
 some margin is allowed on both sides 
 of a given value. In Fig. 54 A and 
 B are the -j- and terminals re- 
 spectively of, say, a 500 volt circuit, 
 R is a standard resistance across 
 which V reads the drop, and at r is 
 placed the resistance to be tested. 
 The first step is to place at r a re- 
 sistance known to be right; next adjust the voltage be- 
 tween A and B at 500: close the circuit and note F's 
 deflection; (call it 200). Any time thereafter if we adjust 
 E to 500 and get 200 across R we may be certain that r 
 is what it ought to be. The advantage of putting V 
 across R instead of r is that in most cases it saves 
 changing the volt lines. 
 
 The next resistance measuring method to be considered 
 is the " Bridge Method." The 
 bridge is an instrument invalu- 
 able to an electrical workshop, 
 testing room, laboratory, or line- 
 man's " shanty." It has a range 
 unexcelled by any other resist- 
 ance measuring device. Refer- 
 ence to Fig. 55 will aid in 
 understanding the underlying principle. C is a cell; 
 G, a galvanometer, none of whose constants need be 
 
MEASUREMENT OF RESISTANCE. 157 
 
 known, since it is to simply indicate the presence of 
 current without measuring it. One galvanometer ter- 
 minal is fastened at JV, the other is free to move. 
 From A to B the current has two paths A N B, 
 and A M B; the drop of potential between A and 
 />, being a definite quantity, is necessarily the same by 
 both paths. For every point in A N B, then, there is 
 in A Af B a point at the same potential, and a galvanom- 
 eter joining these points will give no deflection, for no 
 current will flow between points of the same potential. 
 Under this condition we have what is called a balance, 
 and r l : r y ; ; r : .- 4 where /-,, r a , ;-.,, and ;- 4 are respectively 
 the resistances of A JV, B A 7 , A M, and B M. Knowing 
 any three of these quantities the fourth can be found. To 
 see this more clearly, suppose the two paths to be com- 
 posed of the same size wire and that N has been secured 
 at the middle point of A N B. Then, if the wire is 
 assumed to be of uniform cross-section, the only point to 
 balance A r will be J/, the middle point of A MB. Under 
 this condition of balance A N = B N and A M = B M; 
 therefore 
 
 AN _ 
 
 and also 
 
 A M 
 TTM~ 
 
 whence 
 
 B~N ~ ~B~M* 
 
 What we have proven to be true in this special case is true 
 in all cases. Again, if t\ be the current in A N B, and 
 
158 TESTING OF DYNAMOS AND MOTORS. 
 
 / 2 that in A MB, the P. D. between A and B is, by 
 Ohm's law, t\ r l -)- t\ r 2 ; also it is /, r b -j- / 2 r 4 , hence 
 t\ t\ -j- *, ^ 2 *a *"g + * 2 r 4- Now at balance, the drop 
 from ^4 to N is the same as that from A to M ' ; also 
 drop j? N = 'drop B M, or i\ r } = i z r s and i l r 2 / 2 r 4 ; 
 whence from proportion we have / 2 : ^ ; ; r l : r s and 
 *a : z i r a : r 4> therefore ^ : r a ; ; r s : r 4 . And this is the 
 entire theory of the bridge. If we put r l = a; r 2 = b; 
 r 3 r\ r 4 = x (i. <?., just use letters that are most popu- 
 lar) we have a : b \ \ r : x, or 
 
 whence it can be seen that it is not necessary to know 
 the absolute value of a's and <'s resistance, but only their 
 ratio; and this multiplied by r gives x. 
 
 In all forms of bridges the arms a and b are called 
 proportion arms. In some bridges r remains constant 
 while the ratio 
 
 b_ 
 
 a 
 
 is varied, but in most bridges r alone is varied. 
 
 The first type is known as the " slide wire 
 bridge " and is shown in Fig. 56. Between two heavy 
 copper bars A B a German-silver wire of uniform 
 cross-section is stretched and its terminals soldered. 
 Under the wire a metre rod or scale is laid, carefully 
 graduated into 1,000 divisions. If the wire is of uniform 
 cross-section its resistance need not be known, otherwise 
 it must be calibrated in ohms throughout its length. 
 Suitable terminals are provided to receive the known 
 
MEASUREMENT OF RESISTANCE. 
 
 *59 
 
 resistance, ^t*, and the unknown, X. With connections as 
 in the figure, M is moved along A B till a balance obtains. 
 Then calling the resistance of A M a and B M b, 
 we have a. : b \ \ r : A', and 
 
 ; . 
 
 If the wire A B is of uniform cross-section, we can, in- 
 stead of using A M's and B J/'s resistances, use their 
 lengths, which are in the 
 same proportion, whence 
 R is the only resistance 
 which need be known. 
 If R = X balance will 
 obtain when M rests at 
 500 (t. *., in the middle 
 of the scale). Accord- 
 ing as R is greater or 
 
 1 
 
 M 
 
 J B 
 
 Hill 
 
 
 FIG. 56. 
 
 less than X the balance potjU will be on one side or the 
 other of the 500 mark. Let R = 2 ohms and A' = 3 
 ohms : then balance will be when M is at such a point that 
 
 a 2 ' 
 
 /. e., A B 1,000 divisions must be divided into two 
 parts which shall be to each other as 3 : 2, therefore 
 a = 400 and b = 600. Now suppose R = 5 ohms and 
 suppose a balance is gotten when a = 413 and b = 587; 
 then, 
 
 X =- x JR = 
 a 
 
 587 
 
 X 5 = 
 
i6o 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 whence X 7. i ohms. One must understand clearly 
 which length is a and which b. The range of accurate 
 
 measurement depends 
 upon R's value ; the 
 accuracy is greatest 
 x when R = X, and the 
 probable error in- 
 creases as M recedes 
 from the middle of the 
 scale. In using such 
 a bridge care must be 
 taken that the slid- 
 ing contact neither 
 
 scrapes nor nicks the wire, thereby reducing its cross- 
 section. The bridge can be bought for less than ten 
 dollars and made for about half this sum. 
 
 In the second class of bridge the arms a and ^, known 
 as proportion arms, are first selected to suit conditions, 
 and remain of fixed value, while R is varied to get a 
 balance. Fig. 57 illustrates a commercial form of such 
 a bridge. Its accuracy like that of all bridges depends 
 upon the ratio 
 
 b 
 
 FIG. 57. 
 
 used, upon the voltage of the cells, the sensitiveness of 
 the galvanometer, and upon the care with which the 
 instrument is kept and used. A B and B C are the 
 proportion arms, and consist of three coils, of 10, 
 100, 1,000 ohms respectively. By means of these coils 
 we can employ for 
 
 b 
 
MEASUREMENT OF RESISTANCE. l6l 
 
 the ratios 
 
 I I I 10 100 
 
 - - > - or , - 
 
 L 10 IOO I I 
 
 according as A" is large or small, if we use the same coil 
 of both arms, then 
 
 b i 
 
 a i 
 
 and at balance R X; nor can we measure a resistance 
 exceeding that in J?, if equal arms are used. 
 
 If* : JL A' :=-*-; if s il,A'=io*; 
 
 a 10 10 a i 
 
 , b i R . c b 100 
 
 if - = - --, A = - ; also if = - , A = 100 R. 
 a 100 100 a i 
 
 From which we see that with equal arms, the largest 
 resistance possible to be measured is the total resistance 
 of R\ the smallest is that of the lowest coil in R. With 
 the ratio 1/100 the largest measurable value of 
 
 A" = Total R X ioo, 
 
 while X can be as low as T/IOO part of the lowest coil in 
 R, and so on. To facilitate fine adjustments of R re- 
 sistances from i ohm to the full capacity of the box can 
 be introduced by withdrawing those plugs which short 
 circuit the coils required. The range of measurement is 
 from i/ioo of the value of the lowest coil to ioo times 
 the total value of R. Suppose the lowest coil to be 
 .1 ohm and the total R to be 11,110 ohms. The range 
 is then .001 ohm to i, 111,000 ohms, or over a megohm. 
 
162 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 There are provided two keys for the battery and galva- 
 nometer circuits respectively. In many boxes the keys 
 are placed one over the other, so that a single pres- 
 sure closes first the battery circuit, then that of the 
 galvanometer. The object of closing the battery circuit 
 
 FIG. 58. 
 
 first is to allow extra currents due to self-induction 
 to die away, and leave everything in a state of equili- 
 brium. If the X arm be a field winding or other coil 
 having self-induction the current will not work its way 
 through this arm as quickly as it will through the others, 
 with the consequence that the needle deflects upon clos- 
 ing the key, even though there may be a perfect balance 
 so far as concerns ohmic resistance. By closing the 
 battery key first, and allowing the current to reach a 
 steady value, this objectionable feature is eliminated. 
 Figs. 58, 59, and 60, give the perspective view, plan, and 
 connections of the latest form of portable bridge gotten 
 out by Queen & Co. The instrument is of a high class 
 of workmanship, is convenient, accurate and merits the 
 
MEASUREMENT OF RESISTANCE. 
 
 163 
 
 reputation it has won. Its theoretical range is from .001 
 ohm to 11,110,000 ohms, but for the higher resistances 
 higher external voltages must be used. One feature which 
 commends the bridge to commercial testing room use is 
 its dead beat galvanometer. Readings can be quickly 
 taken and are not influenced by external magnetic fields. 
 Mounted in a handsome mahogony box, 8*^" x 524* X 5", 
 and weighing but 6 pounds, it forms an invaluable 
 addition to a testing set. Fig. 59 gives the connection 
 scheme and Fig. 60 the application of this scheme to the 
 
 FIG. 59. 
 
 FIG. 60. 
 
 bridge proper. Let A be the left bridge, arm; B, right 
 arm; R, resistance in rheostat arm; A', resistance to be 
 measured. When A plugs ^and X plugs B, 
 
 When A plugs X and R plugs B, 
 
 X A A 
 
 - = -,orX = - J X. 
 
 To measure resistance, first connect the terminals of the 
 resistance to be measured to the two binding posts seen 
 
164 TESTING OF DYNAMOS AND MOTORS. 
 
 at the lower left hand corner of the set, Fig. 58, and 
 make an estimate of X's approximate value; suppose it 
 to be in the neighborhood of 400 ohms; plug A to R and 
 X to B and unplug the 100 ohm coil in both A and B, 
 also the 500 ohm coil in R. Put on one cell of the 
 battery and close the key for an instant. (It is well to 
 start with small battery power so that no harm results 
 in case X's estimated value is far from correct.) If the 
 needle swings to -)-, R is too high ; if to , R is too low. 
 By altering R, and if necessary increasing the battery 
 power, a balance can be effected and X determined to 
 within i ohm. If it is desired to be more accurate 
 make A i ohm and B 10 ohms, so that 
 
 R must then be made about 10 times the approximate 
 value of X, Upon securing a balance the error will be 
 but i/io as great as in the first adjustment. Still greate*- 
 accuracy can be gotten by using the ratio 1/100, and a 
 value of R about 100 times the estimated value of X. If 
 the resistances to be measured are higher than the total 
 resistance of the rheostat, plug A to X and B to R. 
 Suppose X is less than 100 times the total resistance of 
 R, say 880,000 ohms: unplug TO, in A, 1,000 in J3, and 
 8,800 in R ; balance and proceed as in the two cases 
 above. 
 
 To use the set as an ordinary resistance box without the 
 galvanometer and battery, arrange the reversing blocks, 
 A, R) B, X, as in preceding cases and make connections 
 to the binding posts. It is immaterial upon which 
 diagonal the plugs in the reversing blocks lie. To 
 
THOMSON'S 
 SLIDE BRIDGE 
 
 FOR MEASURING VERY 
 LOW RESISTANCES 
 
 EXAMPLE: 
 R =.000322 
 
 a 1000 
 6-=ioo 
 
 X=. 000322 X .1 
 =.0000322 OHMS 
 NOTE: 
 
 SEE MUNROE A JAMIE8ON, 
 7TH ED. PAGE 08 
 
 FlG. 62. 
 
l66 TESTING OF DYNAMOS AND MOTORS. 
 
 exclude bridge arms and use rheostat only, connect A or 
 B to R and X. 
 
 To measure very low resistances such as are met with 
 in cable work, the error due to contacts must be elimi- 
 nated, so that the usual form of bridge is by no means 
 satisfactory. For this work Sir William Thomson has 
 devised a form of Wheatstone bridge in which this error 
 does not occur. Fig. 61 illustrates the principle while 
 Fig. 62 shows the connections for the bridge itself. B 
 is a battery, and G a galvanometer of great sensitive- 
 ness; R, a standard resistance, and X, that to be 
 measured, are soldered together. Resistances a, b, c, and 
 </are so great as to render contact resistance negligible. 
 K is a key for closing J?'s circuit and K' a key for clos- 
 ing G's. The condition of balance as indicated by the 
 galvanometer, is a i b\\c I d \\ JR '. X\ or 
 
 x=L R . 
 
 a 
 
 This can be shown as follows : suppose R and X to be 
 absent; S' coincides with 6 1 and T' with T, and we have 
 an ordinary Wheatstone bridge where a : b \ \ c : d or 
 a d = be. Now introduce R and X, varying R till a bal- 
 ance obtains ; then, a : b\ \ R + c : X -\- d, or a X -j- a d 
 b R + b c; but b c a d. Therefore a X = b R> and 
 
 Y b /? 
 X. = jft 
 
 a 
 
 as above. The manner of varying R is shown in Fig. 62, 
 and is done very gradually by means of a lever. The 
 lettering is similar to that of Fig. 61. The manner of 
 using this bridge does not differ from that of the ordinary 
 
MEASUREMENT OF RESISTANCE. 167 
 
 type. Its accuracy depends upon the accuracy, prima- 
 rily of 7?, and also of a and b. An example will help us. 
 Let a 1,000, b = 100, R = .000322; then 
 
 b 100 
 
 X = R - = .000322 - - = .0000322. 
 a 1,000 
 
 The Wheatstone bridge is by far the most universally 
 used instrument, and its thorough understanding is of 
 indispensable value. We therefore feel justified in adding 
 the following general remarks 
 incidental to bridge practice. In 
 using any bridge the first step to 
 take is to insure that all connec- 
 tions are correct and that all 
 circuits are intact. With equal 
 resistances in the proportion 
 
 arms take out the 10 ohm plug 
 
 FIG. 63. 
 in the rheostat: holding apart 
 
 the test lines going to X, press successively the battery 
 and galvanometer circuit keys. In the typical diagram 
 of Fig. 63, A and B are the proportion arms, R the 
 rheostat, and X the unknown resistance. When the test 
 lines leading to X are held apart, the X arm is open cir- 
 cuited, its resistance is infinite, and current from arm B 
 therefore flows across through the galvanometer causing 
 a decided deflection, due to the excess of resistance of X 
 over^; for X equals oo while R equals 10. Say the de- 
 flection is to the right. Now short circuit the test lines 
 by holding them together: upon pressing the keys there 
 will be a deflection to the left, for R now equals 10 
 while X, being short circuited, equals o. If both deflec- 
 tions are in the same direction the indications are that 
 
1 68 TESTING OF DYNAMOS AND MOTORS. 
 
 there is either an open or a short circuit in the test lines. 
 If the line is broken holding the lines apart will not alter 
 conditions, nor will crossing them if they are already 
 crossed elsewhere. It is barely possible that the bridge 
 may have some internal disorder. An open or short cir- 
 cuit in A will give the same symptoms as if in X, so also 
 with B and R. But such a trouble can be quickly located 
 by measuring the same resistance with different ratios. 
 It is possible one line wire may be broken inside the 
 insulation or the two wires may be mashed together. 
 Under either condition it will be impossible to get a 
 balance; for in one case there is not enough resistance in 
 R to balance when X equals oo, and in the second case, 
 none small enough to balance when X equals o. If while 
 securing a balance the needle give a sudden lurch for some 
 unaccountable reason, it is probably due to a loose con- 
 nection. If successive measurements of the same resist- 
 ances show discrepancies in results, it is likely due 
 either to contact troubles or to the presence of dust 
 between the rheostat blocks. A common source of 
 annoyance is in screwing the plugs in just tight enough 
 to trap a layer of dust in between them and the 
 blocks. Failure to get any deflection on the galvanom- 
 eter may be due to open circuit either in its own or 
 the battery circuit. In the first place a battery lead 
 may have become corroded or the contact tips of K or 
 K' may be oxidized and refuse to carry current; care- 
 ful inspection will generally reveal troubles of this 
 nature : but if it is necessary to test out, first disconnect 
 the galvanometer and replace it by a piece of wire; do 
 the same with the battery. This leaves the circuits as 
 complete as before and we have a galvanometer and 
 
MEASUREMENT OF RESISTANCE. 169 
 
 battery to test with. J<;in one side of G to one side of the 
 battery. From the two remaining terminals bring two 
 small w-ires to serve as test lines. Touch them together; 
 ifcV deflects, G and Care all right. If G does not deflect 
 it may be due to a poor contact somewhere; be certain 
 that this is not so. There may be a loose connection 
 inside the galvanometer. If there are two cells of 
 battery they may be joined in opposition. If only one 
 cell, its plates may be touching inside. The needle if 
 suspended may have a broken fibre, if pivoted, may be 
 stuck: the ingoing and return galvanometer wires may 
 be crossed so as to cut out the coil. Once assured that 
 the test circuit is all right test across the bridge keys; G 
 should deflect, showing the circuits to be all right. If it 
 does not a metallic wedge can be shoved into the key of 
 the defective circuit so as to complete it there. The 
 test lines are then made to span different parts of the 
 circuit till G deflects, and we know that the break is 
 repaired. 
 
 We have seen that with a ratio 
 
 I 10 100 
 
 or or 
 i 10 100 
 
 no greater resistance can be measured than is contained 
 in ft. When using this ratio, the resistance can be got- 
 ten only to a first approximation unless it happens to be 
 an exact number of ohms within the range of R. More 
 frequently a point is reached where the insertion of i 
 ohm in R causes a deflection one way, and its withdrawal 
 a deflection the other, showing the value of X to lie 
 between the two values of R. 
 
 Suppose 9 and 10 ohms cause deflections respectively 
 
I7O TESTING OF DYNAMOS AND MOTORS. 
 
 to right and left: if the two deflections are of the same 
 magnitude it indicates that X lies midway between 9 and 
 10 ohms, and according as one or the other resistance 
 gives the greater deflection the value of X will be the 
 more removed from its value. To obtain a closer approx- 
 imation the ratio 
 
 10 100 
 -or_, 
 
 in some cases, can be used. The larger the current 
 around the coils of any galvanometer of given sensibility 
 the larger will be its deflection. In measuring high 
 resistances, on a bridge, in order to increase the deflec- 
 tion for a given difference between R and X, a higher 
 voltage is used. As long as high resistances are being 
 measured, that necessary to balance them is also high, 
 and each proportion arm having a high resistance in 
 series with it is less liable to damage from the increased 
 voltage: but when measuring low resistances the i 
 ohm or 10 ohm coil of the proportion arm is liable to 
 injury, and the voltage must be lowered to suit the new 
 conditions. 
 
 On some bridges the plugs are drawn out to introduce 
 resistance while in others they must be inserted. In 
 the first case every coil is short circuited by a plug which 
 must be drawn out to let the coil into circuit. In the 
 second case it is only by inserting a plug that the circuit 
 through the coil is completed as seen in Fig. 65. The coil 
 in this case is continuous, and is tapped off at intervals to 
 plugs which are to be plugged to a bar running the full 
 length of the coil. The latter has the advantage that 
 fewer plugs are used, and hence there are fewer contacts 
 
MEASUREMENT OF RESISTANCE. 17 1 
 
 to introduce error. On many bridges the proportion 
 arms are identical, but on others one arm may have 
 i, 10, 100 ohm coils, and the other 10, 100, 1,000 ohm coils. 
 According as one ratio or another is used the error intro- 
 duced in results by an error of i ohm in the adjustment 
 of 7?, varies. Using a ratio 
 
 i 
 
 an error of i ohm in fi causes an error of i ohm in A". 
 That is, if through lack of care in adjustment a balance is 
 taken at 9 when it really calls for 10, A''s value will be just i 
 
 FIG. 64. FIG. 65. 
 
 ohm out. Such an error can be caused by undue haste in 
 adjusting 7?, by dust in the rheostat, or by not having the 
 needle at the zero of torsion. Some suspended galvanom- 
 eter needles intended for bridge work have their side play 
 limited by two non-magnetic stops placed on both sides of 
 the needle, and the normal position of the needle is midway 
 between these stops. If the directive force of the magnetic 
 field holds the needle over against one of these stops, and 
 if instead of restoring the needle to its proper position by 
 turning the instrument it be done by turning the torsion 
 head, the needle then holds its position against the 
 twisting moment of the fibre. Under such a restraining 
 force the sensibility of the needle is much diminished. 
 
172 TESTING OF DYNAMOS AND MOTORS. 
 
 The writers have preferred to do away with the stops on 
 either side the needle, to allow it to find its own position of 
 rest each time, and then to give it a directive force about 
 this point by means of a magnet laid somewhere near it. 
 Using the ratio 
 
 10 
 i 
 
 an error of i ohm in R causes a difference of .1 ohm in 
 X\ with the ratio 
 
 100 
 
 i 
 
 the error in X is .01 ohm, and with the ratio 
 
 1,000 
 i 
 
 .001 ohm. With a ratio 
 
 10 
 i 
 
 suppose a balance to have been gotten with R = 25. 
 Then, 10 : i ; \ 25 : X, and X = 2.5 ohms. Should the 
 balance properly have been at 24, corresponding to a 
 resistance of 2.4 ohms, the error would be 2.5 2.4 = . i 
 ohm. To measure greater resistances than that in R, 
 the ratios 
 
 10 10 i 
 
 100 1,000 I,OOO 
 
 are used. Results are not as accurate as when using the 
 inverse of these ratios, for in the first case an error of 
 i ohm in R means an error of 10 ohms in X. In the last 
 case a difference of i in R causes a difference of 1,000 
 in JVT< These ratios are used mainly for measuring resist- 
 ance of insulation where the error may be very large, and 
 still be but a fractional per cent, of the total resistance. 
 
MEASUREMENT OF RESISTANCE. 173 
 
 In all low resistance work the resistance of the test 
 lines must be taken into consideration, and in some work 
 that of the bridge rheostat arm. To measure the test 
 lines, clamp their further ends together and if possible 
 get a balance with the ratio 1,000 to 10. Suppose this 
 to be impossible and that the insertion or withdrawal of 
 i ohm causes respectively a -j- and deflection. Sup- 
 pose the galvanometer dial to be graduated and that 
 with R =. 90 a -f- deflection of 27 obtains, and that 
 for R 91 a deflection of 17. A balance at R 90 
 would correspond to A' = .9 ohm: and a balance at 
 R = 91 to A" = .91 ohm, a difference of .01 ohm. 
 This difference of .01 ohm causes a deflection of 27 -f 
 17 = 44 divisions: one division would be caused by 
 
 .01 
 
 = .000227 ohm. 
 44 
 
 Since 90 ohms causes a deflection of -f 27 such a resist- 
 ance must be added to 90 as will bring the needle 
 back 27 divisions to zero, corresponding to a balance; 
 .000227 ohm will move the needle i division; to move 
 it back 27 divisions, .000227 x 27 = .006129 ohm must 
 be added to 90, making A", at the balance point, = 
 .90006. This method of ascertaining the balance point 
 without actually effecting it, is known as the method 
 of interpolation. To enable finer measurements to 
 be actually made a variable low resistance can be con- 
 nected in series with R. This resistance usually replaces 
 the oo plug. To get the resistance of a bridge rheostat 
 arm, substitute for the test lines a piece of bare copper 
 wire whose dimensions can be easily determined. The 
 amount of this wire included between the "AT" binding 
 
174 TESTING OF DYNAMOS AND MOTORS. 
 
 posts can be varied till it just balances the resistance of 
 the arm. Its own resistance is then determined either 
 from a wire table or by calculation based upon its 
 dimensions and specific resistance: a constant to be 
 considered later. In ordinary bridge work, the rheostat 
 arm resistance can be neglected in comparison with 
 that of X. 
 
 Another very satisfactory method of measuring low 
 resistances is by means of the differential galvanometer. 
 This is a galvanometer having two coils, which, when 
 connected either in series, or in parallel and in opposi- 
 tion, should have equal and opposite effects upon the 
 needle which, of course, then remains at rest. The 
 differential action depends upon the magnetic opposition 
 of the coils when an equal current is sent through both 
 but in opposite directions. If the coils are identical as 
 regards number of turns, and in position as regards the 
 needle, complete balance is secured when the coils are in 
 series, because the current through both is necessarily the 
 same. For a balance with the coils in parallel the resist- 
 ances must be the same, otherwise the same current will 
 not flow through both. To test for symmetry of position 
 connect the coils in opposed series, and apply a cell; upon 
 pressing the key a deflection to one side or the other 
 indicates one coil to be magnetically stronger than the 
 other. A controlling magnet should be used to bring 
 the needle back to zero, or an adjusting coil, sometimes 
 provided, brought into use. Some such method of 
 compensation is necessary, for any lack in the coil itself 
 cannot be conveniently remedied. The equality of 
 resistance is tested by connecting the coils in opposed 
 parallel. If the resistances are equal the currents will 
 
MEASUREMENT OF RESISTANCE. 
 
 '75 
 
 be the same, and on a magnetically balanced instrument 
 the deflection will be zero. This is seldom the case, how- 
 ever, and compensation is secured by placing a resistance 
 box in series with the coil of lower resistance and adding 
 resistance till the needle returns to zero. Having deter- 
 mined the value of this added resistance a special coil 
 can be made, and permanently attached to the base of 
 the galvanometer. If this coil is non-inductively wound, 
 that is, so wound that there are as many turns from right 
 to left as from left to right, it will in no way disturb 
 the magnetic balance. The non-inductive winding is 
 secured by doubling the wire in the middle and winding 
 it on double, as is done with resistance box coils. 
 
 There are two schemes of connections which can be 
 used in measuring resistance with the differential galvan- 
 ometer. Fig. 66 shows the method best adapted to 
 medium or low resistances. 
 The coils are here joined 
 i n opposed series; t h e u n k now n 
 resistance, A', is placed as a 
 shunt across one coil while 
 the variable resistance, R, is 
 across the other. If the 
 galvanometer is balanced to 
 begin with, it will be bal- 
 anced when R = X\ for under 
 
 this condition the shunting power of each is the same, 
 and equal but opposite currents still pass through coils 
 c and c '. The delicacy of this method depends upon the 
 sensitiveness of the galvanometer and upon the accuracy 
 of R. The test consists in varying R till the needle 
 returns to zero. The method is due to Heaviside. It 
 
 FIG. 66. 
 
176 TESTING OF DYNAMOS AND MOTORS. 
 
 sometimes happens that, just as on a bridge, an exact 
 balance cannot be obtained, because J?'s units are too 
 large. In this case we calculate the balance point by 
 interpolation. Observe the lowest possible 4- deflection, 
 say it is 10 divisions. Next, put one unit into R or 
 take one out, as the case may be, to get the lowest 
 possible deflection; say it is 25 divisions. For the -j- 
 deflection supposed? = 50 ohms and for the deflection 
 R 51 ohms. Then the total deflection 25 -f- 10 = 35 
 divisions is caused by a difference of i ohm in R; 1/35 
 ohm will cause a deflection of i division, so, in order to 
 bring the needle back to zero, 10/35 = - 2 & ohm, which is 
 to be added to the 50 ohms, making 50. 28ohms the point of 
 balance, and the value of X. r, the resistance to be added 
 to the smaller value of R in any case, can be got from 
 the proportion. 7?'s unit : r \ \ d -|- d' : d, where d is 
 the -f- deflection and d -\-d' the total deflection caused 
 by a difference of i unit in R. 
 
 For measuring high resistances the coils should be con- 
 nected in parallel, X being in series with one and R with 
 the other. Fig. 67 shows the plan of connections. Under 
 these conditions, too, a balance obtains when R = X. 
 Where perfect balance cannot be obtained, interpolation 
 is resorted to. It is well to introduce the box r into the 
 battery circuit to cut the current down if necessary. 
 This method can be given a wide range of measurement 
 by introducing the shunts S l and S^ which should be 
 either standard shunts, or accurate resistance boxes with 
 a total resistance about equal to that of the coil which 
 it is to shunt. As the two galvanometer coils are 
 alike so will the shunts be. Better still, have S t and S^ 
 so designed with respect to the coils as to give the 
 
MEASUREMENT OF RESISTANCE. 
 
 standard ratios of 1/9, 1/99 and 1/999. If this method is 
 adopted it is well to adjust the coils' resistance to some 
 even figure, and to order the shunt boxes accordingly. 
 In Fig. 67 it can be seen that current from B passes 
 around c from left 
 to right but around 
 (' from right to 
 left. If the galvan- 
 ometer coils are 
 electrically and 
 magnetically bal- 
 anced no deflec- 
 tion will be gotten 
 when, with R and 
 
 A" cut out, the key FIG. 67. 
 
 is pressed. Placing 
 
 X in the circuit with c> balance is destroyed but can be 
 restored by introducing into r"s circuit a resistance 
 equal to that of X. A perfect balance established, R = 
 X. To eliminate any error likely to arise from want of 
 a proper balance between the coils themselves, it is only 
 necessary to substitute for X a second standard box and 
 find what resistance, R' just balances R as already found 
 If R' = R, the measurement is correct: if R' does not 
 equal R it still equals AT, for it has replaced X under the 
 same conditions. 
 
 The smallest resistance to be had in R is perhaps 
 i ohm; the largest 10,000 ohms. According as A" is 
 greater than 10,000 or less than i ohm shunts S l and 
 S^ respectively are used. For example, suppose it is 
 desired to measure an armature whose resistance is 
 .001 ohm. It cannot be measured without shunting the 
 
178 TESTING OF DYNAMOS AND MOTORS. 
 
 galvanometer, for there is in R no resistance small 
 enough to balance it on equal terms. X being much 
 less than R a large current flows through coil c and 
 causes a deflection; before the needle can be brought 
 to zero, the current in c must be made the same as that 
 in c', and this can be done either by reducing R till it 
 equals X, or by shunting coil c so that of the total 
 current which flows through the circuit only an amount 
 equal to that in c' flows through c, the rest going 
 through shunt S 9 . The shunt to be used is dictated by 
 the degree of accuracy desired. . Using the 1/99 shunt 
 the removal of i ohm from R has the same effect as 
 i/ioo of an ohm would have were no shunt used; and 
 with the 1/999 shunt, the same as 1/1,000 ohm would 
 have without a shunt. For suppose that with no shunt, 
 the removal of i ohm from R caused a decrease of i am- 
 pere in the current flowing through circuit r M N B. 
 The whole variation takes place in coil c'. If only 1/1,000 
 of the current in r c X B flowed around c, a variation of 
 i ampere in r X B would cause a variation of approxi- 
 mately but 1/1,000 of i ampere in coil c. This condition 
 is more properly explained but less easily understood by 
 considering- not the change which takes place in c when 
 R is varied, but the change which takes place in c', 
 because if we suppose a constant potential to be main- 
 tained between points M and N t variation in R will cause 
 no change in <r's current. However, using the 1/999 
 shunt S^ the effect of <r's coil is but 1/1,000 of what it 
 would be were no shunt used, so the indicated value of 
 X at time of balance must be divided by 1,000. The 
 object attained then by using S 9 is to reduce the effect of 
 each ohm in R, so that R can be used for measuring re- 
 
MEASUREMENT OF RESISTANCE. 179 
 
 sistances outside its own range, and with greater accuracy 
 those within its range. If A" is many times greater than 
 R the shunt must be used in c' circuit; then instead of 
 dividing 7?'s value at balance by the shunt power used we 
 multiply it. Suppose X to be 102,000 ohms. Without a 
 shunt, R would have to contain 102,000 ohms to balance 
 this, but by using 5, =1/999 we g et a balance with R 
 102 ohms. In general, calling 6" the resistance of the 
 shunt, G that of the galvanometer, R the standard, and A" 
 the unknown; when A" > R (is greater than R) and c 9 
 is shunted, X R X the shunting power of S } . When 
 X < R (X is less than R} and c is shunted, A' = R -^ 
 shunting power of 5 2 . Here R means the resistance 
 with which a balance is gotten. More briefly expressed, 
 If 
 
 A > R, c' shunted, A = i -- ~- c X 
 
 Cr 
 
 If 
 
 s 
 
 X < R, c shunted, X = .-$ X R, 
 
 & +\ 
 
 where 
 
 _S_ 
 
 G + S 
 
 is the multiplying power of the shunt and is gotten by 
 inverting the fraction denoting the part of the whole 
 current which passes through the galvanometer coil. In 
 its most general form the formula giving the value of X 
 in terms of R, G, G' S S' is 
 
 ' + 
 * = ** 
 
 1+3, 
 
 where and G' are the coil resistances and 6" S' that of 
 
l8o TESTING OF DYNAMOS AND MOTORS. 
 
 their respective shunts. If G and G' are of the same re- 
 sistance then G G and 
 
 If 5 = oo, that is if no shunt is used on G, 
 y- R - s ' R 
 
 / "" /~* i o/ " 
 Cr Cr j o 
 
 ^~S' 
 Further if 
 
 G -\- S' 1,000' 1,000 ' 
 
 If no shunts are used both S and S' oo and X = R. 
 To investigate it a little more mathematically: if G is 
 the resistance of the galvanometer coil, and S that of its 
 shunt, their resistance in multiple is 
 
 G S 
 
 and the resistance of the circuit is 
 
 G S 
 
 <^ + - 
 If E\s the voltage of the battery, then 
 
 I= E+(R + 
 
 Also we know that 
 
 where 7 g is the current in the galvanometer coil. Now 
 it is this coil's current which interests us: 
 
 We therefore solve for f e as follows: 
 
MEASUREMENT OF RESISTANCE. l8l 
 
 (r r a- .<n F s -|- G 
 
 , ' r 
 
 __ + 
 therefore, 
 
 (g + 5) 
 
 + C) + 5 S + ^ " 
 SE 
 
 (G + .S 1 ) + G S' 
 Reducing still further and dividing by S, we get, 
 
 '- 
 
 
 The meaning of this last equation is this: the value of 
 the current in the shunted coil is the same as it would 
 be were there no shunt used, and if instead, J?'s value 
 were increased to 
 
 S 
 The effect of the shunt then is to multiply^ by 
 
 s 
 
 which is the multiplying value of the shunt. For example, 
 suppose the 1/99 shunt to be used and R= 5,249. What 
 is AT? Here R is to be multiplied by 100, therefore X = 
 5,249 x ioo = 524,900 ohms. If the 1/999 shunt were 
 used and R to balance were 9,847 ohms, X would be 
 9,847 x 1,000 = 9,847,000 ohms. When X is smaller 
 than any unit in R, S, is used and the formula is 
 
l82 TESTING OF DYNAMOS AND MOTORS. 
 
 x = *^s- 
 
 Thus suppose the 1/99 shunt is used and R = 45 ohms; 
 then 
 
 X = = .45 ohm. 
 100 
 
 Again, with the 1/999 shunt and 
 
 R = 13, X = 41 = -013 ohm. 
 
 We have thus increased the range of the method so that it 
 is useful anywhere from 1/1,000 of the smallest unit in R 
 to 1,000 times the total value of R. 
 
 Except for the calculations involved in reducing results, 
 the use of an ordinary resistance box as a shunt is prefer- 
 able to the standard shunts, as it admits of finer adjust- 
 ments. For example, suppose G = 500 ohms, -5" = 315 
 and R = 9,950. What is X? 
 
 3 r 5 5 
 
 x = x R = 6 \;r - x 9,950 
 
 *J 3*5 
 
 __ 815 X 9,950 = 25)744 ohms. 
 3i5 
 
 If a 1/99 shunt had been used the balance value of R 
 would have been 257 and the value of X found to be 
 25,700 ohms, or 44 ohms too small. Had the 1/9 shunt 
 been used, R would have balanced at 2,574, correspond- 
 ing to X = 25,740 or 4 ohms too small. So we see that 
 although a resistance box requires more work than a 
 standard shunt, it also gives greater accuracy. 
 
CHAPTER VII. 
 
 MEASUREMENT OF INSULATION. 
 
 WHEATSTONE'S bridge and the differential galvanom- 
 eter fulfill the requirements for a wide range of measure- 
 ments, but they become, less and less accurate as this- 
 range is increased. For very low resistances the " Com- 
 parison" method is much the best, and for resistances 
 reaching up into megohms, there are better methods- 
 which we will now investigate. 
 
 Fig. 68 shows the connections for a method much used: 
 in testing insulation resistance: A is a dynamo; G, a 
 galvanometer, with shunt 5; R y a resistance box of 
 100,000 ohms or more ; r is a standard resistance, and 
 X, the resistance to be measured. R is necessary, not 
 only to set up the galvanometer, but to protect it in 
 cases where X is found to be defective. For a test of 
 this sort to be thoroughly reliable, a high voltage, (125 
 to 500 volts) must be used, and it would not do to apply 
 this directly to the galvanometer. Switch, K, makes it 
 possible to throw either X or r into circuit at will. 
 The test consists in first noting the deflection with X in 
 circuit: it is well to start with the 1/999 shunt in; and 
 if the deflection is zero or small, successively put in 
 the 1/99 an< 3 1/9 shunts, and finally remove the shunt 
 entirely. If the deflection is now zero the insulation is 
 
 183 
 
184 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 WWvVJ 
 y 
 
 FIG. 68. 
 
 above the maximum range of the galvanometer, for the 
 given voltage, and may be considered perfect. When 
 
 changing shunt, care 
 must be taken that K is 
 open, otherwise the gal- 
 vanometer may get a 
 current that will twist 
 the fibre several turns, 
 or even twist the needle 
 off. If a deflection is 
 obtained, its value is 
 noted; this we will call 
 
 dj.. Next throw r into circuit, and if it is possible, 
 get a readable deflection with the same shunt as used 
 with X. In this case the following proportion holds: 
 X : r ; ; 4 : </ x ; whence 
 
 X A r. 
 
 Suppose r = 100,000 ohms; d^ = 50 divisions, and d Y = 
 240 divisions. Then 
 
 240 
 
 ~5^ 
 
 If a shunt is used with either X or r the formula must 
 be modified accordingly. Suppose the 1/999 shunt is 
 used with r, and the 1/9 shunt is used when X is in cir- 
 cuit, then 
 
 r 
 
 X = . - r = 
 
 X 100,000 = 4,800,000 ohms. 
 
 The range of measurement then extends from 
 r 
 
 1,000 
 
 to 1,000 r, 
 
MEASUREMENT OF INSULATION. 185 
 
 or from 100 ohms to 100,000,000 ohms where r = 100,000 
 ohms. Another method of regulating the deflection is 
 to keep a constant shunt on the galvanometer and vary 
 the voltage, E, of dynamo A. In this case E is varied 
 till the same deflection is obtained from A' and r, when, 
 A" : r \ ; E^. : E r \ and 
 
 Suppose x 500; E v - 50 and r 100,000; then 
 
 soo 
 X -" - X 100,000 = 10,000,000 ohms. 
 
 5 
 
 A simple and inexpensive method of varying E is shown 
 in Fig. 69. The total voltage of A is applied to the high 
 resistance wire, R R '. One 
 galvanometer circuit termi- 
 nal is secured at R and 
 the other is fastened to the 
 movable contact, c. The 
 voltage in the galvanometer 
 circuit is varied by sliding c 
 along the wire. The con- 
 ditions to be filled by the 
 wire R R' as regards uni- 
 
 formity of cross-section, etc., are the same as those 
 discussed under the slide wire bridge. The total length 
 of R R should be divided into as many equal parts as 
 there are volts on A\ then any desired voltage below 
 this can be gotten by taking the proper proportion 
 of R R"s length. Where R R' is very long, 100 feet or 
 more, it must be mounted in some manner that will 
 make every inch of it easily accessible. To do this it 
 can either be wound back and forth on wooden pins 
 
l86 TESTING OF DYNAMOS AND MOTORS. 
 
 driven into a board, or it can be wound in a spiral on 
 a wooden drum with brush contacts at either end. In 
 both arrangements R R' should be graduated in volts 
 and the values marked at intervals on the board or drum 
 just beneath the wire. If there are 500 volts on A, the 
 galvanometer circuit will be subjected to the full 500 if 
 the free end c rests on R '; to 250 volts, if c rests mid- 
 way between R R\ and so on, the fraction of the total 
 voltage being the same as the fraction of the total length 
 of R R' included. For this reason it is more convenient 
 to mark R R' in fractions of its own length, then the 
 marking is good whatever may be the voltage of dynamo 
 A. It may not always be convenient to conduct the 
 test by varying E, in which case recourse must be had to 
 the method of variable shunts described above. 
 
 To be strictly accurate the resistance of generator, gal- 
 vanometer and R must be considered. When X is in cir- 
 cuit (see Fig. 68) its resistance maybe considered so high 
 as to require no shunt to get a readable deflection; in this 
 case the deflection is inversely proportional to the cir- 
 cuit resistance, and if A's voltage is , we may write 
 
 , 
 
 A + G + R + X ' 
 
 When r replaces X, a shunt generally has to be used, 
 because r is so much less than X that the current sent 
 through it by E throws the ray off the scale. The effect 
 of this shunt, as has been shown, is to virtually increase 
 the circuit resistance 
 
 S+ G 
 ~S~ 
 
 times, so that we may write, 
 
MEASUREMENT OF INSULATION. 187 
 
 E 
 
 '.= 
 
 From these values of d^ and d l we get by division, 
 E E 
 
 {, ' A+G+R+X 
 
 O 
 
 : 
 
 A + G + R + X 
 
 _ ^^ (A + r + R) + G . 
 
 A + G + R + X 
 Clearing of fractions and dividing by d^ we have 
 
 Whence solving for X, we get 
 
 where A = generator resistance, and 1?, r, and S, that 
 of the constant box, standard resistance, galvanometer 
 and shunt respectively. X is the unknown resistance. 
 When r and X are large A and G may be neglected and 
 the formula becomes, 
 
 If X and r are alone taken into account, then 
 
l88 TESTING OF DYNAMOS AND MOTORS. 
 
 d^ S -(- G 
 
 which reduces it to the case already discussed. If we 
 use the 1/999 shunt, 
 
 S+ G 
 
 and 
 
 X- * 
 X - t 
 
 = 1,000 
 
 1,000 r 
 
 These methods are used in the commercial testing of 
 glass and porcelain insulators and marine cables. To 
 
 FIG. 70. 
 
 test insulators they are arranged along a shelf containing 
 holes to receive them. Each one is then filled with water 
 to within an inch of the top, and the tank is filled to the 
 same level. They are then allowed to stand several 
 days to give the water every chance to creep into any 
 crevices which might exist. Fig. 70 shows the connec- 
 tions for the test which then follows. The tank, T, con- 
 tains the insulators, #, a, a. The test lines are fastened 
 one to the iron tank and the other to a long ebonite rod 
 so as to be movable. By means of the two-way switch, K, 
 the galvanometer may be thrown on the lines L L' or on 
 the standard high resistance box R. The other appur- 
 
MEASUREMENT OF INSULATION. 189 
 
 tenances are the same as those already noted. The test 
 consists in throwing the galvanometer alternately upon 
 Z Z', the test lines, one of which touches the water 
 through the tank case, the other of which touches the 
 water in the insulator, and upon the box R. The deflec- 
 tions obtained are then compared. Let </ f be the deflec- 
 tion due to R, and d^ that due to A r , the insulation 
 resistance. If the 1/999 shunt be used with R. we 
 have 
 
 X ~- '' 4~ \ ( B + r + R ) S \ G + G \ - (G + B -f r) 
 ( & ) 
 
 or with sufficient accuracy 
 
 A' 1,000 j- R. 
 
 Throughout the test, the lines Z Z' should be exchanged 
 to detect any insulation defect in the testing system. 
 Before and after the test, line Z should be held in air and 
 K closed. Any deflection shows faulty line insulation, 
 which should be remedied. It is, however, impossible to 
 maintain perfect insulation in the pressure of such high 
 voltage; perfection on a dry day often proving faulty oi\ 
 a wet one. All wiring and devices should be supported 
 on rubber cleats and legs, and these should be carefully 
 dusted daily. Besides that caused by surface leakage due 
 to moisture, there will be, on closing the key, a momen- 
 tary deflection due to the line's capacity; the line takes a 
 static charge, the amount of which depends upon the 
 potential of the battery and the line's capacity. 
 
 The effects of capacity depend on the well-known fact 
 that if we connect two points which are at different 
 potentials, a flow of electricity takes place between them 
 
IpO TESTING OF DYNAMOS AND MOTORS. 
 
 until the points are at the same potential. If some 
 agent is employed to keep the points' potentials dif- 
 ferent, at the same rate that the current equalizes them, 
 we have a continuous flow of current, as in an ordinary 
 active lighting or power circuit where the external flow 
 of current represents the -(- and brushes' effort to 
 equalize their potentials, while the armature is the agent 
 which maintains the potential difference. Now if a lot of 
 dead wiring be tapped on to one brush of a live dynamo 
 this wiring will acquire the potential of that brush, and 
 in doing so causes a momentary flow of current, whose 
 strength and duration will depend upon how much elec- 
 tricity it takes to raise the wires to the same potential as 
 the rest of the system. 
 
 In the above case all the wires on one side of the 
 key are attached to the battery or dynamo, and are at a 
 certain potential: wires on the other side of the key are 
 dead or at zero potential, except in so far as they may 
 preserve a charge from their last communication with the 
 live side. As soon as the key is closed there is an 
 equalization of potential, which sends a momentary cur- 
 rent through the galvanometer and causes a deflection. 
 This deflection, however, should quickly subside, whereas 
 a deflection due to leakage will persist and be of constant 
 value, and must be subtracted from the deflections 
 obtained from the box R, and from the insulation X, in 
 order to get the true deflection due to these resistances 
 alone. In the above test the insulators are first rapidly 
 tested by tapping the stalk with the terminal L. Those 
 measuring below a certain value are removed, and the 
 balance are then connected in groups of 10 or 100 by con- 
 necting all the stalks together. The insulation resist- 
 
MEASUREMENT OF INSULATION. 191 
 
 ance of these groups is then taken accurately. The pres- 
 ence of a poor insulator in a group will bring down the 
 average of the group; on the other hand a lot of good in- 
 sulators may boost a moderately poor one through : hence 
 the preliminary test. The surface leakage of the insulators 
 taken together further lowers the apparent value of the 
 insulation. This, however, is the condition under which 
 the insulators are used, and hence is the test which they 
 should be called upon to stand. The average insulation 
 per insulator is required to be above 250,000 megohms, 
 and anything below 200,000 megohms is rejected. If 
 100 insulators are in a group and they are assumed to be 
 of equal insulating power, their total resistance con- 
 nected in parallel will be i/ioo of that of any single 
 insulator. If 200,000 megohms is the minimum for a 
 single insulator, 200,000 -r- 100 = 2,000 megohms is the 
 minimum for the group. 
 
 In measuring the insulation of marine cables and 
 cables for underground work, the same method is 
 resorted to. The cable is coiled in a tank 6 or 7 feet 
 deep, and its ends are brought out and supported in mid- 
 air by means of ebonite rings. Water is then run into the 
 tank and the cable left there for several days. Each end 
 is then stripped of its insulation and lead, or other 
 sheathing, for a few inches back, so as to increase the 
 length of the surface between the lead and the copper. 
 To further decrease surface leakage the ends are pencil- 
 shaped, and coated with hot paraffine before being tested. 
 The longer a cable, the less will be its total insulating 
 power. Contracts therefore require not that the whole 
 cable have a certain insulation resistance, but that 
 it have a certain insulation resistance per foot. Since 
 
192 TESTING OF DYNAMOS AND MOTORS. 
 
 the total insulating power diminishes with the length 
 of the cable (since every foot adds a path for leakage), 
 the average insulation per foot is gotten by multiplying 
 the total insulation by the number of feet, 
 
 or, Insulation of the whole cable = L nsulation P e . 
 
 Length in feet 
 
 Hence, as above, insulation per foot = total insulation x 
 number of feet. Cables are in general of three types: 
 single conductor cables, double and triple conductor ca- 
 bles, and telephone cables. In the first the insulation is 
 measured from the conductor to the lead; in the second, 
 between conductors, which are concentric, and from the 
 outer one to ground; in the third, between each of the 52 
 to 104 pairs of conductors. That is, each wire is tested 
 with every other. They are all then twisted together 
 and tested to ground. In telephone cables the insulation 
 need not be very high from wire to wire, since the 
 voltage at which they will work is not high. The test is 
 more for the purpose of insuring that the cable be free 
 from crosses. Even in case of a cross the defective wire 
 is simply ignored; it is on this account that several extra 
 wires are put in each cable, so that a 100 wire cable has 
 104 wires. The insulation to ground is tested at a lower 
 voltage than in high tension cables. 
 
 High tension cables are subjected to a further test : the 
 cable is put into a tank, and the secondary of a step up 
 transformer connected to the cable and the tank. The 
 voltage is then raised to a value several times above that 
 to which the cable will be subjected in practice. If it 
 stands this test the voltage is lowered to a point some- 
 what exceeding that at which the cable is to work, and it 
 
MEASUREMENT OF INSULATION. 193 
 
 is allowed to stand for an hour. At the end of this time 
 the insulation is measured by the above method. 
 
 The method also admits of the locating of a fault. To 
 do this it is only necessary to leave one terminal of the 
 testing lines on one end of the cable the other remaining 
 on the tank. The cable is then slowly drawn from the 
 tank. As soon as the faulty spot leaves the water the 
 insulation improves, as shown on the galvanometer: after 
 verifying the results by repeating the test, the cable is 
 cut off at the faulty point provided the portion left is 
 long enough to be of any use. 
 
 Another method, well adapted to the measurement of 
 high resistance insulation, and equally applicable to 
 underground or overland lines, depends upon the elec- 
 trometer, explained above. The method of using it in this 
 test is as follows and is illustrated in Fig. 71. The elec- 
 trometer, , has one side connected to tank T, and the 
 other to cable, C. The cable is first connected to a 
 source of E. M. F., say a 500 volt line, and is given a static 
 charge the value of which depends upon the cable's 
 capacity. The electrometer is then switched onto the 
 cable. The plates now acquiring the potential of C 
 and T respectively, attract each other and cause a 
 deflection which keeps its initial value so long as the 
 cable's charge remains the same. The charge will 
 remain constant if C's insulation is perfect: if not, leak- 
 age takes place, the deflection becomes smaller and 
 smaller, and becomes zero in the course of half an hour 
 or more, according to the value of the insulation. In 
 line work the method is not as good as those already 
 described, for an electrometer has not the range of use of 
 a galvanometer, and is hence a relatively more expensive 
 
194 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 instrument. We do not, therefore, give the equations 
 involved in the method. 
 
 In measuring comparatively low insulations, such as 
 are found in armatures, commutators, and fields, a Weston 
 voltmeter can be used to advantage. It is accurate, por- 
 table, needs no setting up, and can be used with any voltage 
 within the range of its scale. Fig. 72 gives the connec- 
 
 FIG. 71. 
 
 tions. B is the source of E. M. F.; V, a 500 volt voltmeter; 
 A, an armature whose insulation from copper to iron is to 
 be measured, and K, a switch by means of which A can 
 be cut out and the voltmeter constant taken. Here the 
 meter resistance takes the place of that of the galvanom- 
 eter and boxes. The test consists in first closing K and 
 observing F's deflection. This deflection will be the full 
 voltage of B and is called the "constant." If B is a 500 
 volt dynamo, the constant will be the full scale reading, 
 or 500, and is due to the current which 500 volts will send 
 through the resistance of the meter. Suppose this resist- 
 ance to be 80,000 ohms, = R v . If upon introducing the 
 insulation resistance, X, into circuit, and opening K, the 
 deflection drops to 250, just one-half, we know that 
 the current through Fhas been halved, and that since 
 
MEASUREMENT OF INSULATION. 195 
 
 the voltage has been unchanged, the circuit resistance 
 has been doubled and is now 160,000 ohms. Taking from 
 this F's resistance of 80,000 
 ohms, we have left 80,000 
 ohms as that of X. If d^ the 
 deflection with X in circuit, 
 falls to one-quarter of </,, we 
 know X -j- R v has become 4 
 X 80,000 ohms = 320,000 FJ G. 72. 
 
 ohms, and that A' is 320,000 
 
 -80,000 :: 240,000 ohms. In any case inserting X 
 makes d l as many times d^ as the resistance of X -{- -# T 
 is times that of J? v alone. In other words to get the 
 value of AT -f 7? v , multiply ^ by 
 
 This gives 
 
 if we take from this the meter resistance ^ T , we have left 
 that of the insulation, or 
 
 which is the same as 
 X 
 
 -(ft-) 
 
 and is the working formula. Using the same letters^ 
 we will derive this formula a little differently. Since 
 d l and d^ are proportional to the currents which cause 
 
196 TESTING OF DYNAMOS AND MOTORS. 
 
 them, we can in this case substitute them for these cur- 
 rents and write, 
 
 /r* 
 
 < = 7 
 
 and 
 
 dividing the first by the second, we get, 
 
 A ^y + ^ A , _*" ,X 
 
 <*,'' , : R, + ^ v = " ^ ' 
 
 or 
 
 ~~7~ I ~\ TT > 
 
 ' 
 
 whence 
 
 TT /? I ' I 
 
 Thus, let ^ v = 80,000 ohms, as before; d l = 500, and 
 </ a 3. Then 
 
 '^(i-')-*-.^-)- 
 
 / S oo 3\ 
 80,000 I - ~ I == 80,000 x 165 = 13,200,000 ohms, 
 
 or 13.2 megohms. 
 
 If R v = 8,000; d i = 200, and d^ = 20; 
 
 / 200 \ 
 Jf = 8,000 I i J = 8,000 X 9 72,000 ohms, 
 
 a very low insulation. 
 
 * More correctly d\ = -77- >i', but the constant may be omitted with- 
 
 out error. 
 
MEASUREMENT OF INSULATION. 197 
 
 Now suppose^ = 80,000; d l 500, and </ a = 500 also: 
 then 
 
 A' -- 80,000 | : - i J = 80,000 (i -- i) 
 
 = 80,000 X o = o, 
 
 which shows that A* is no insulation at all: /. e., that the 
 copper actually touches the iron. It is now customary to 
 put a voltmeter at the disposal of the workman, so that he 
 can test his work in its different stages of construction, and 
 thereby detect a fault as soon as it occurs, and thus to 
 avoid having it returned after its completion. This saves 
 time, labor, and material. It is then necessary to tell the 
 field or armature winder, or commutator builder what de- 
 flection will pass his work. Suppose R v 80,000, E = 500, 
 and it is deemed safe to pass any work whose insulation 
 hot is 1,000,000 ohms or i megohm. What is the maxi- 
 mum deflection (</ a ) allowable? The value of X must not 
 be below 1,000,000 : this is to say X -|- R^ must not be 
 less than 1,080,000 ohms. We saw above that X -f- R* is 
 always as many times R v as d l is times d 9 . How many 
 times R v is X -\- R v l X 1,000,000; R v 80,000; A' -f- 
 R^ 1,080,000 
 
 ^4- ^v _ 1,080,000 _ 
 
 R 80,000 
 
 d^ then, can be 13^ times d^ d^ must be 1/13^ of d l = 
 I / I 3 I /4 of 5 = i/V- of 500 = fa of 500 = 37. A 
 deflection of 37 with X in circuit means, then, an insula- 
 tion resistance of i megohm. 37 is the maximum allow- 
 able deflection when E = 500. If E = 1,000, d 9 may be 
 
190 TESTING OF DYNAMOS AND MOTORS. 
 
 74; if E = 125, d^ must be no more than 9^ divisions. 
 To get this result more direct take the formula 
 
 ?H 
 
 X = 
 
 and substitute for the letters their numerical values, and 
 we get 
 
 1000000 
 
 /5 \ 
 = 80,000 i i 1 
 
 dividing both sides by 10,000, to simplify, we get 
 
 100 
 
 or 100 d^ = 8 (500 </ 3 ) : 100 </ 3 = 4,000 8 d^ whence 
 1 08 d^ 4,000 and 
 
 as before. Any piece of insulation letting a deflection of 
 over 37 take place through it would be rejected. In prac- 
 tice, it is customary to dispense with switch A" and to get 
 d^ by holding the test lines together. One line is then 
 held to the shaft and the other on to the commutator if the 
 armature is connected up. If the commutator alone is to 
 be tested one line goes to shell and the other is passed 
 from bar to bar: any bar giving a deflection of over 37 
 is marked, afterward examined, and if necessary replaced. 
 The next test is that of insulation between bars. This 
 need not be high, since under working conditions the 
 difference of potential between adjacent bars is only a 
 fraction of the machine's E. M. F. The connections are 
 
MEASUREMENT OF INSULATION. Ip9 
 
 given in Fig. 73. Test lines T T' are first crossed and 
 */, obtained. They are then touched to the same bar; if 
 d^ is the same it shows that there is no oil or shellac on 
 the commutator to influence the readings. The lines are 
 now gotten such a distance 
 apart as to span one mica 
 body, without touching the 
 same bar at the same time. 
 To start with, one mica body 
 is proven to be all right and 
 is then short circuited with a Fui. 73. 
 
 lead-pencil mark. The lines 
 
 are now held stationary, the commutator slowly turned 
 and the deflection noted each time a body is spanned by 
 the lines. When the pencil mark is reached after making 
 a complete revolution the needle will indicate a short 
 circuit. The lead mark makes it unnecessary to watch 
 the commutator, so that the attention can be centered on 
 the needle. When a " low body " is spanned, the needle 
 shows it by a deflection whose magnitude depends upon 
 the nature of the trouble: A burr on the " ear "of a 
 bar, or a filing touching both bars will give the full deflec- 
 tion. A burr may be found by inspection but a filing 
 may be so imbedded as to give great trouble. If the 
 fault is not too serious it can often be burnt out with 
 current from a lamp circuit. Holding one terminal on 
 one bar, the other terminal is tapped on the next bar; 
 if the test lines are held on, matters may be made worse, 
 for when the cross gives way an arc may be set up and 
 the insulation charred, or a bead of copper deposited 
 between the bars. If the contact is too "good " to burn 
 out the commutator must be dismounted, and the faulty 
 
20O TESTING OF DYNAMOS AND MOTORS. 
 
 insulation replaced. This means that the nut and washer 
 must be removed, several bars taken out, new insulation 
 put in, the parts reassembled, the bars gauged again to 
 see that they lay straight, the commutator turned down 
 and baked again, and then retested. This will impress 
 upon the mind of the most skeptical the advisability of 
 having the builder test his work in course of construction 
 with something less delusive than a magneto. 
 
 Where a commutator shows a uniform high deflection 
 all around, it indicates moisture, and must be baked more. 
 On repaired commutators this symptom is due sometimes 
 to the soldering acid forming a bridge between bars or to 
 the insulating bodies having become carbonized. In the 
 first case the mica bodies must be well scraped between 
 the " ears " where the acid is apt to have crept. In the 
 second case, the chances are that carbonization has taken 
 place where the commutator is exposed to the joint 
 action of the paraffine and sparking of the brushes, and a 
 light cut off the commutator will generally, though not 
 always, remove the difficulty. Where a commutator tests 
 perfect, with the exception of 20 or 30 bars, it indicates 
 moisture; shellac may have accumulated in a space be- 
 tween the bars and shell, and there being so much of it, 
 it would require longer to dry it out. The most difficult 
 trouble to locate is where it is due to poor quality of the 
 insulating material itself. This fault can only be located 
 by the negative method of proving every other source of 
 trouble to be absent. The manufacturers themselves are 
 fooled sometimes, and warrant goods that prove to be 
 worthless. Inferior grades of tape, shellac, varnish, and 
 rnica sometimes give no end of trouble, and are mislead- 
 ing, because they pass the test when new, but do not stand 
 
MEASUREMENT OF INSULATION. *2Ol 
 
 service. Another source of trouble, sometimes unavoid- 
 able, is the practice of having the winding room and 
 commutator department exposed to the iron-filings and 
 flying dust of machine shops and emery wheels. The 
 writers know of several instances where much trouble 
 was traced to such surroundings. 
 
 This method of measuring insulation by means of a 
 Weston voltmeter is being adopted everywhere, and it 
 would be a great convenience if the makers would furnish 
 with each instrument a chart or curve, giving the insula- 
 tion corresponding to each deflection when a specified 
 voltage is used. This is a matter of interest to all, and 
 not only to those unfamiliar with formulae. 
 
 It must be remembered that the insulation value ob- 
 tained by any of the above methods is by no means a 
 constant definite quantity, but varies according to the 
 voltage used in the test. Insulation, perfect under a 
 test of 500 volts, would not be so when subjected to r,ooo. 
 This is because the higher voltage not only has a greater 
 ability to break down insulation, but by increasing the 
 leakage, the instrument is affected by a current that passes 
 over and not through the insulation. It is very impor- 
 tant that the voltage used in the test should at least 
 equal that at which the machine is to work. The writers' 
 experience has been that commutators can be built to 
 stand the 500 volt test from bar to shell without disturb- 
 ing the needle at all, and from bar to bar with only 
 a slight wiggle of the needle. If an insulation from bar 
 to shell of i megohm can be considered enough, that 
 from bar to bar should be at least one-half as much. On 
 armatures, especially of the toothed type, and on fields 
 it is much more difficult to perfect the insulation on 
 
202 TESTING OF DYNAMOS AND MOTORS. 
 
 account of the increased surface exposed to the iron. 
 On a smooth-core armature, only the bottom layer of 
 wire is exposed to the core, but on toothed types every 
 wire is next to either the bottom or side of a slot. As an 
 example of the effect of increased surface, we cite the 
 instance of an unconnected armature, each of whose coils 
 tested perfect alone; when they were all connected 
 together by means of a piece of No. 18 bare copper wire 
 wound around the ends, the voltmeter showed a deflec- 
 tion of 20; upon cutting the No. 18 wire so as to divide the 
 armature into two insulated halves, each half gave a de- 
 flection of 10 ; each quarter of 5, and so on. At present 
 it is customary to perfect the insulation of each part 
 before it goes into the machine. The completed machine, 
 after being subjected to a severe running test, is again 
 tested for insulation; this should be taken about ten 
 minutes after the machine is shut down, for while run- 
 ning, fanning of the air keeps the armature surface cooler 
 than the interior; and insulation should be tested only 
 when everything is at its uniform maximum temperature. 
 This fact accounts for what may seem strange, that many 
 armatures burn out five or ten minutes after being shut 
 down, and it also explains how an armature hastily tested 
 and pronounced sound, sometimes shows up a cross or 
 ground when it is attempted to put it into service out on 
 the road. It has always been the writers' custom, where 
 proper testing instruments were wanting, to run a 
 machine its regular test, allow it to cool off, and run it 
 again for a short while. 
 
 As it is often convenient to use a galvanometer for 
 conducting the above tests we will give some suggestions 
 in regard to galvanometer practice. What can be said 
 
MEASUREMENT OF INSULATION. 
 
 203 
 
 of either voltmeter or galvanometer, can with few excep- 
 tions be said of both. All lines in any way connected 
 with an insulation test must be run on porcelain or hard 
 rubber insulators to minimize leakage, and to prevent 
 grounds. The presence of such faults in testing systems 
 are fruitful sources of trouble, resulting in deflections of 
 unwonted value, sometimes in blowing fuses, and some- 
 times in twisting the needle from its suspension. 
 
 Fig. 74 illustrates the connection of the above voltmeter 
 test. K is a reversing switch to galvanometer G. Jt is 
 a constant resist- 
 ance; A is the ar- 
 mature whose insu- 
 lation to shaft is to 
 be measured. Sup- 
 pose the shaft to 
 be resting on iron 
 horses and hence to 
 
 be in communication with the ground at e. In the event 
 of a second ground occurring on the line its effect would 
 depend upon its location. If a fault is at e lt current from B 
 will pass direct from 7 1 , down through ^, up through e lt and 
 back to B. B having both terminals directly to ground at 
 e and e lt the galvanometer and box, 1?, are both cut out. 
 If 's voltage is high, as is usually the case in insulation 
 testing, some part of the line must give way, and for this 
 purpose a fuse should be introduced at B. If B is a few 
 cells of battery the only effect is a low or zero deflection 
 on the galvanometer for both positions of K. If a fault 
 occur at e^ It's resistance and that of A's insulation are 
 cut out, leaving the galvanometer in circuit with only 
 its fuse to protect it from injury. In such a case, the 
 
204 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 proper step is to interchange the lines T and T' ; i. e., 
 the one on the shaft goes to the commutator and vice 
 versa : the effect of this is to place both grounds on the 
 same side of B y and if the accidental ground is at <? v , R 
 is cut out, leaving the armature alone in circuit with the 
 galvanometer. If X, the insulation resistance, is high 
 compared to R, the deflection will not be appreciably 
 affected as long as X is in circuit; but if 7"and T' should 
 happen to touch, B would short circuit through G and 
 perhaps injure it. 
 
 Next suppose the ground to be at e z \ the deflection in 
 this case would be the same as if the test lines were 
 
 held together, for the 
 armature insulation, 
 X, is cut out ; by 
 interchanging T and 
 T', ground e and e z 
 are placed on the 
 same pole of B, and 
 the deflection returns. 
 to its proper value. 
 
 It is thus seen that the question of insulation in 
 a testing system is important. The arrangement of 
 Fig. 75 is an improvement over that of Fig. 74. In 
 this arrangement, the insulation of G and R can be 
 easily made perfect, and a ground on the lines either at 
 e l9 e^ or ? 3 cannot cut out R and harm G. For the given 
 position of the lines a ground at e t will, as before, short 
 circuit B and cause a deflection lower than the ^con- 
 stant." If B is a high resistance cell and the constant 
 gotten by holding T and T' together is small, the 
 difference in deflection caused by such a ground is apt 
 
 e, 
 
 FIG. 75. 
 
MEASUREMENT OF INSULATION. 
 
 205. 
 
 to pass unnoticed unless the ground makes a dead short 
 circuit, in which case the galvanometer circuit is sub- 
 jected only to the drop which takes place through the 
 fault. Short circuiting a source of high voltage will 
 hardly pass unnoticed, as more or less violent demonstra- 
 tions attend such a fact. The proper proceeding in the 
 above case is, as before, to interchange the test lines. 
 A ground at e n will have no effect with the connections 
 given, unless the lines get interchanged, when the case of 
 a ground at <? 3 is re- 
 peated. A ground at 
 c y cuts out A and 
 gives the " constant " 
 deflection. Inter 
 changing T and T' 
 removes the trouble. 
 Fig. 76 gives a plan for 
 
 over that of Fii 
 
 . 
 
 further improvement 
 
 . 
 
 Ik-re all 
 
 connections save the test lines are included in a compact 
 system which can be well insulated and permanently 
 mounted. K now reverses only T'and T' and does not 
 affect 's current. If it is desired to reverse 6"s cur- 
 rent, as is often the case where we wish to get a deflec- 
 tion on both sides of zero in order to use the average, the 
 reversal can be done by a second switch to which the gal- 
 vanometer lines run, as in any of the foregoing diagrams. 
 The effect of a ground on T or T' is to cut out A, which 
 is indicated by the deflection being the constant. Such a 
 fault is quickly detected by interchanging the test lines; 
 hence the advantage of having K as in Fig. 76, where 
 reversal of T and T' is controlled by the galvanometer 
 operator. Otherwise there must be some signal for tell- 
 
2O6 TESTING OF DYNAMOS AND MOTORS. 
 
 ing the " lineman " to interchange T and T'. If both 
 test lines are grounded or otherwise crossed, a deflection 
 will obtain even when T and T' are held apart. Where 
 a man is experienced in galvanometer work he is not apt to 
 reject sound insulation which tests apparently low through 
 the influence of a line fault, especially where high voltage 
 is used, for he is always more or less disagreeably reminded 
 of the presence of the fault when any part of his person 
 comes in contact with a key, switch, or other device 
 which is on the ungrounded side of the circuit. On 
 a test circuit where several cells of battery are used 
 as the source of E. M. F., the operator gets no shock 
 even if he touches the two leads from the battery; 
 with the result that if one side of the circuit is already 
 grounded, as at e^ in Fig. 75, and he should stand 
 on the wet floor and touch the key, K, on the upper 
 side he would produce a ground virtually at e t , cut 
 out A and make it appear low. Personal contact 
 with any part of a low voltage testing circuit will' 
 introduce error in any of the foregoing tests. The 
 operator must then keep clear of exposed parts of the 
 circuit, and the " lineman " must not only keep his hands 
 off the test line terminals but must be very careful that 
 the points do not rest on dirt, lacquer, or grease, and 
 that they are not burnt to an oxide on the end. Where 
 high voltages are used these errors due to personal con- 
 tact are not so apt to occur. Grounds at <? 4 and e b in 
 Fig. 76 should be of very rare occurrence, and are readily 
 located. The above cited complications which arise in 
 galvanometer practice as a result of defective insulation 
 are representative, and most others are greater or lesser 
 modifications of them. A little experience on a faulty 
 
MEASUREMENT OF INSULATION. 207 
 
 testing circuit will soon impress the tester with the 
 importance of good insulation. 
 
 The tendency of the day seems to be toward the use of 
 high voltages, because to transmit a given amount of 
 energy to any distance, the line loss is much less where 
 the energy is transmitted as high voltage and low current, 
 than where low voltage and high current are employed; 
 for instance suppose we wish to transmit one horse 
 power (i HP) a distance of one mile over a line composed 
 of No. 10 B. & S. copper wire. What would the loss be 
 in sending i ampere current at 746 volts, and what would 
 it be with 746 amperes at i volt, the voltage in both 
 cases being that at the user's end of the line ? i Mile = 
 5,280 feet. The resistance of 1,000 feet No. 10 B. & 
 S. copper wire is i ohm: the resistance of 5,280 feet is 
 5.28 times i ohm = 5.28 ohms. Since there is a return 
 wire the line is really 2 miles long and its resistance 2 x 
 5.28 = 10.56 ohms. Now from Ohm's law, E = I R, 
 we have E i x 10.56 = 10.56 volts as that necessary 
 to- send the current of i ampere through the line; the 
 loss = El = 10.56 x i 10.56 watts. In the second 
 case / = 746 and R = 10.56, and E, the drop consumed 
 in sending the current, is I R 746 x 10.56 8,877 
 volts, making the enormous loss of 10.56 x 8,877 = 
 90,000 watts approximately. This is selected as an 
 exaggerated case to bring out the difference boldly. A 
 No. 10 B. & S. wire will not carry 746 amperes. 
 
 Large electrical companies have a very perfect system 
 of testing insulation intended to work with high voltages, 
 and in this system the tests above cited are usually 
 preliminaries. Before subjecting the insulation to an 
 alternating voltage test of from 1,000 to 5,000 volts (some- 
 
208 TESTING OF DYNAMOS AND MOTORS. 
 
 times much higher than this), it is tested by galvanometer 
 or voltmeter at 500 volts, or less, to insure the absence of 
 moisture and other immediate causes of short circuit. 
 Fig. 77 gives the connections for a high voltage test. 
 In this particular test the voltage from the generator was 
 1,050 and was brought to transformer, T, which reduced 
 it to 50 volts. In cases where a 50 volt alternating cur- 
 rent lighting circuit is available, T can be dispensed with. 
 J3 is a combination switch and fuse box by means of 
 which all communication with the live source can be cut 
 off. T's 50 volt secondary passes to switch K. C is 
 a circuit breaker, which flies out when poor insulation 
 permits the current to exceed a certain value. A is an 
 amperemeter, and from here passes one side of the sec- 
 ondary circuit of the stationary transformer, 7", to one 
 side of the primary of the variable transformer, T' 
 The other side of T's secondary runs direct from switch 
 K to the remaining primary terminal of T'. Now the 
 primary of T' slides in and out of the secondary by 
 means of the windlass at P, and it is in this manner 
 that the secondary voltage is varied from 100 to 5,000 
 volts. The theory being that when the primary is inside 
 the secondary all of the latter's coils are cut by the 
 primary's lines of force, whereas when outside, only a 
 few of the lines reach it and its voltage is correspondingly 
 low. To the terminals of T "s secondary the test lines 
 are attached and are applied to the insulation to be meas- 
 ured. In this case, however, the test lines are not held 
 on by hand, but are fastened on and the circuit closed and 
 opened at K. Before K is closed, T"s primary must 
 be at zero as indicated by the arrow. Closing K, P 
 is turned so as to run up the primary into the secondary 
 
 
2IO TESTING OF DYNAMOS AND MOTORS. 
 
 until the insulation under test either gives way or stands 
 the full test. Windlass P had best be acted upon by an op- 
 posing weight, which will automatically return the coil to 
 the zero position should the operator carelessly leave it in 
 the position of high voltage. V is a voltmeter attached 
 across the secondary of 7", and registers 50 volts at time of 
 test. An electrometer is used (not shown in Fig. 77) for 
 registering the voltage on T"s secondary, and it is this 
 that the tester watches when testing. As soon as switches 
 B and K are closed, ammeter A registers the current 
 which the 50 volts send against the self-induction of T' 's 
 primary. If the insulation under test is perfect, no cur- 
 rent flows through T"s secondary, there is no reaction 
 on its primary, and A's current does not increase; but if 
 the insulation at X breaks down current flows in T"s 
 secondary, its lines of force spread out and cut the 
 primary turns, reduce the induction, and let the 50 volts 
 send a larger current through the primary circuit. If 
 the breakdown is very bad, C flies out and opens the 
 circuit. In such a test all apparatus should be well pro- 
 tected from passers-by. 
 
 It is not always convenient to bring the electrical 
 parts, armatures, fields, transformers and the like, to a 
 testing room to be tested, so to obviate this inconvenience 
 some of the leading manufacturers have adopted the fol- 
 lowing very flexible plan. In Fig. 78, P is the primary 
 coil of a transformer which, being connected to a plug, can 
 be inserted in any lamp socket on the shop alternating 
 current lighting mains. The secondary coil has several 
 leads brought out, and by means of these different sec- 
 ondary potentials can be gotten. By using the two end 
 leads the voltage is raised to 3,000: leads i and 3 give 
 
MEASUREMENT OF INSULATION. 
 
 211 
 
 2,000; 3 and 2, 1,000; and soon, in as many steps between 
 the limits as there are extra leads. In this case, no 
 voltage indicator is needed on the secondary coil because 
 the transformer is usually designed for a particular pri- 
 mary voltage, that of the shop, and this assumed, the 
 various secondary voltages can be marked under their 
 respective leads on the box. Small testing sets can be 
 
 
 :> 1 
 
 1000 
 
 < ^ 30( 
 
 
 
 1 
 
 
 
 FIG. 78. 
 
 carried by hand, but the larger ones of greater range, 
 containing coils immersed in oil, are heavy, and must be 
 pushed from shop to shop on a truck. 
 
 The question of liquid resistance arises in connection 
 with batteries and water rheostats. In very accurate 
 work it is necessary to know the battery resistance. If 
 the resistance of a column of any liquid is known and 
 taken as a unit, it is easy to estimate the dimensions of a 
 box to fulfill any conditions of rheostat service. 
 
 The Wheatstone bridge affords a ready means of apply- 
 ing a battery to the measurement of its own resistance. 
 Fig. 79 shows the connections for such a test. Here the 
 battery occupies the bridge arm in which we usually find 
 the unknown X : but in this test the 1 battery is not only 
 X but is the source of E. M. F. as well. The method 
 differs a little from the ordinary bridge method, in that 
 balance is effected with a constant deflection rather than 
 
212 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 a zero deflection. The arms P and Q are the proportion 
 
 .arms and their values once chosen are not changed. The 
 
 galvanometer circuit being 
 always closed there is a cur- 
 rent and deflection, R is al- 
 tered till upon closing K the 
 deflection is not changed in 
 value. Under this condi- 
 tion closing K does not cause 
 a redistribution of poten- 
 tial over the system, and 
 hence does not disturb the 
 
 galvanometer, and we have the proportion/ 5 : Q ; \ R : X, 
 
 whence 
 
 X = Q- x R. 
 
 One objection to the method is that any polarization of 
 the battery will raise its resistance; on the other hand this 
 has the advantage of being the condition of actual ser- 
 vice. The error due to polarization can be eliminated 
 either by using a sensitive galvanometer so that only a 
 small current passes or by employing other methods. 
 When the battery resistance is very small, such as that 
 of a storage cell, the following method is recommended. 
 A Weston amperemeter and voltmeter serve the pur- 
 pose in the connections of Fig. 80, where V is a volt- 
 meter of such high resistance as to have practically no 
 shunting power, whence its current can be neglected and 
 the reading of A called /?'s current. R is a resistance. 
 Call E, /?'s open circuit E. M. F., and E' its terminal 
 E. M. F. as indicated on V, when the current in A is /. 
 
MEASUREMENT OF INSULATION. 
 
 : ; 
 
 Now when current flows the E. M. F. of B is divided 
 into two parts that which drops through the external 
 resistance and which it is unnecessary to calculate 
 because Vindicates it; and that which drops through r, 
 the battery resistance, and equals by Ohm's law / r. 
 Then E = E' -j- / / whence E E' = I r and 
 
 The test consists in taking E, then closing K to get E' 
 and I ; this method is adapted to industrial work.* 
 
 Liquid resistance can be found by confining the liquid 
 in a glass tube having at its ends metal electrodes, one of 
 which can be moved along the tube to vary the length of 
 the liquid column. In Fig. 8r E is the movable elec- 
 
 1 ' 
 
 1 ir- 
 
 
 L^_l 
 
 
 $71 
 
 L$E 
 
 
 FIG. So. 
 
 FIG. 81. 
 
 trode, G a low reading ammeter, and R a resistance 
 adjusted to give a convenient current value. 's position 
 being noted, A' is closed and 7 quickly read on G. The 
 reading must be made quickly to avoid the error of 
 polarization. E is than moved along a few inches and a 
 new reading taken. I is now less than before; now, by 
 means of R restore 7 to its former value. If .ffand E' 
 have been pushed nearer together, in order to reduce 7 
 to its first value we must introduce in R a resistance 
 * For other methods of measuring battery resistance see some standard 
 treatise, such as that of Carhart and Patterson. 
 
214 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 equal to that of the column of liquid included between 
 the first and second position of E. Knowing the dimen- 
 sions of this column we can estimate the resistance of 
 any column of the same liquid. It is well to start with 
 E at about the middle of the tube and to draw it out, 
 
 FIG. 82. 
 
 thus increasing the resistance, using smaller currents 
 and minimizing the effects of polarization. 
 
 Water rheostats, or water boxes as they are commonly 
 called, are made upon this principle. In place of the 
 tube is a long trough as in Fig. 82, or a box or barrel, as 
 in Fig. 83, which should rest upon insulators, but seldom 
 does. The electrodes are iron plates, one of which is free 
 to slide in guides as in Fig. 82, or free to move up and 
 down as in Fig. 83. The resistance of the rheostat is 
 varied either by varying the dimensions of the liquid 
 column (moving the plates or putting in or taking out 
 some water) or by improving its conductivity or impair- 
 ing it (putting in salt or soda to increase conductivity; 
 substituting for some of the briny solution pure water to 
 decrease it). The water box finds a place in the testing 
 of various electrical machines, especially in providing 
 converters and high tension alternating current machines 
 with non-inductive loads equivalent to a lamp load 
 , without the intervention of transformers. The water 
 
MEASUREMENT OF INSULATION. 
 
 215 
 
 box finds a more industrial application in connection 
 
 with the speed regulation of turbines and is growing in 
 
 favor as a starting box for 
 
 stationary motors, or as a 
 
 permanent resistance in cir- 
 
 cuit with a low voltage motor 
 
 running from high voltage 
 
 mains. 
 
 The foregoing methods of 
 testing liquid resistances are 
 open to the objection that 
 polarization must more or less 
 modify results. The follow- 
 ing method is recommended 
 as free from this objection. 
 In Fig. 84 A is an ordinary 
 slide bridge with a standard 
 non-inductive resistance, j?, 
 
 and the liquid contained in tube, L. B is a source of 
 alternating current, such as is furnished by an ordinary 
 lighting circuit, and Z"a telephone receiver used in place of 
 a galvanometer. As the contact S is moved along the wire 
 the buzzing in the telephone becomes louder or fainter 
 as the case may be. When S reaches a point at the same 
 potential as Af t the sound ceases, or at least becomes a 
 minimum. Under this condition we have: 
 
 P: 
 
 Other forms of Wheatstone bridge are not so well suited 
 to this test, on account of the coils' self-induction. 
 which, however small, cannot be said to be zero. 
 
TESTING OF DYNAMOS AND MOTORS. 
 
 The method is adapted to the determination of a sub- 
 stance's specific resistance, which we have learned is the 
 
 resistance of a piece 
 of any substance us 
 compared with the 
 resistance of a piece 
 of similar dimensions 
 of another substance 
 taken as a standard. 
 The standard is pure 
 copper. If a piece of 
 FIG. 84. pure copper measures 
 
 so much and a similar 
 
 piece of some other substance measures twice as much, 
 the specific resistance of the latter substance is said 
 to be 2. 
 
 In speaking of a conductor's resistance, we tacitly in- 
 volve not only the conductor's dimensions and specific re- 
 sistance, but we must know the temperature at which the 
 measurement has been made: hence in compiling tables of 
 resistances it is customary to give resistance values at 
 some accepted standard temperature. This standard is 
 usually taken as 15.5 C. or 60 F., the average yearly 
 temperature near London, England. Carbon excepted, 
 all solid conductors rise in resistance as the temperature 
 rises, but with carbon and liquids the reverse is true. To 
 find the amount of this effect, it is necessary to measure 
 the resistance of the conductor at several temperatures 
 and to divide the change of resistance by the change 
 of temperature. The result is ohms per degree, and 
 is called the temperature coefficient. Very careful 
 experiments have been made, particularly by Dr. 
 
MEASUREMENT OF INSULATION. 2iy 
 
 Mathiessen, upon copper, showing that the resistance of 
 a pure copper conductor changes i for each 2.58 C. 
 change in temperature. Thus suppose the air to be at 
 i5.5 k> C and that at this temperature a machine's shunt 
 winding measures 10 ohms. What will it measure at 
 77.5 C. ? The change of temperature is 77.5 15.5 
 = 62. i <? of 10 ohms is .1 ohm. The resistance 
 increases as many times . i ohm as 2.58 is contained in 62, 
 and 
 
 62 
 --.- = 24 times; 
 
 and 24 x .1 ohm = 2.4 ohms increase in resistance, 10 
 ohms -j- 2.4 ohnm = 12.4 ohms, the resistance at 77.5 C. 
 We have here assumed the air to beat standard tempera- 
 ture. Take a more likely case. Let R be 13 ohms at 
 21. 5 C. What will it be at 15. 5 C.? Here 
 
 ^'-'5:5: 
 
 2-5** 
 
 i % of 13 = .13 and 2.3 x .13 = .299 ohm change in 
 resistance. Now since the final temperature is less than 
 the initial, the value of R is less, so we must subtract 
 the correction from the initial value, and we have, 13 
 .299 =12.7 ohms at standard temperature. In these 
 examples the change in temperature is given and the 
 change in R estimated. It is more often that we 
 must figure the final temperature from the rise in R, 
 obtained by measuring before and after heating. If a 
 field's resistance is 27.2 ohms at air temperature of 30 C. , 
 and the resistance is 31.4 ohms at the end of a two-hour 
 test; What is the final temperature? Since R rises i $ 
 
2l8 TESTING OF DYNAMOS AND MOTORS. 
 
 for every 2.58 rise in T, T will increase as many times 
 2.58, as the rise in R is times i % of the initial value of R. 
 Initial R = 27.2 ohms. Final R 31.4 ohms. 31.4 - 
 27.2 = 4.2 ohms rise in R. i % of 27.2 = .272 and 4.2 -=- 
 .272 = 15.4 times; 15.4 X 2.58 = 39.7 rise in T. 30 + 
 39.7 = 69. 7 final temperature. These examples show that 
 for temperatures below the standard the correction is 
 negative, /". <?., must be subtracted from the resistance 
 value at the standard temperature, while for temperatures 
 above the standard the correction is positive and is to be 
 added. These facts are expressed in the following for- 
 mula: R' = R -J- R(t 15. 5 C.) k, where R is the resis- 
 tance after test, R the initial resistance, / the temperature 
 at which the resistance is desired, and k the percentage 
 change in R for i C. In this case, considering copper, k 
 = .33 of i %. 
 
 To build the above formula up from first principles we 
 proceed as follows: Since R represents the final value 
 and,/? the initial value of the resistance, R(t 15.5 C.) k 
 must represent the increase in resistance, and the formula 
 must read, the final resistance = the initial resistance 
 -\- the increase in resistance. Further, let us analyze 
 the expression which denotes the increase, namely 
 R(t 15. 5 C.) k. Here / 15. 5 gives us the number 
 of degrees rise in temperature. Also /is the percentage 
 of R which a rise of i in temperature causes R to rise. R 
 X k is then the actual increase in ohms which i causes 
 and this X by the degrees (/ 15.5) rise, gives (/ 
 15.5 C.) R k; whence, R = R -f R (t 15. 5 C.) k, or 
 R R [!_!_(/_ I5 . 5 o c.)k}. If /is greater than 15.5, R' 
 is greater than R. If / is less than 15.5 the expression 
 R (t 15.5 C.) is negative and must be subtracted 
 
MEASUREMENT OF INSULATION. 219 
 
 from R to give A'. We now wish to know the value of 
 k. For copper it is practically = 
 
 .01 
 
 - = .0018 , 
 2.58 
 
 so the above formula becomes R' = R [i -{- (* I 5-S^ C.) 
 .0038]. 
 
 The temperature coefficient of alloys is generally less 
 than that of its component metals. This decrease 
 in the coefficient suggests the possibility of finding an 
 alloy with a zero coefficient; this is nearly attained in 
 manganin an alloy 12 $ manganese, 84 % copper, 4 % 
 nickel. The following table shows its value: 
 
 MANGANIN.* 
 
 RANGE OF TEMP. MEAN TEMP. COIL. 
 
 10 to 20 -[-.000025 
 
 2O " 30 -J-.OOOOI4 
 
 30 " 35 4- .000004 
 
 35 " 40 4~ -000003 
 
 40 " 45 ... -j- .000001 
 
 45 " 50 -- .000001 
 
 50 " 55 --.000002 
 
 55 " 60 -- .000004 
 
 60 " 65 .000005 
 
 It is here seen that the coefficient is positive up to 
 45 and then becomes negative; the average coefficient 
 for the whole range is -j- .0000 1 and for a range of from 
 30 to 65 the mean value is zero. 
 
 The question of temperature coefficient is important 
 in dealing with resistances and batteries, for where 
 
 * Dr. Lindeck, Proc. International Elect. Congress, 1893, p. 165. 
 
220 TESTING OF DYNAMOS AND MOTORS. 
 
 accuracy is desired it is now customary to calculate final 
 temperatures from the increase in 7?, instead of using a 
 thermometer, for the latter, being in contact with only 
 the surface layers of the winding, registers less than the 
 true value. It is possible, however, to perform a series 
 of experiments which shall show how the true tempera- 
 ture compares with that indicated by the thermometer. 
 This consists in carefully measuring a resistance at 
 atmospheric temperature and at the same time noting 
 this temperature on the thermometer. Next, send a 
 small current through the coil and allow it to get as hot 
 as that current will make it; now put the thermometer 
 on and get a temperature while the current still flows; at 
 the same time measuring R so that from the rise in 7? the 
 temperature can also be figured. Next, increase the 
 current 5 % or 10 % and allow the coil to reach its maximum 
 temperature for that current. Read the thermometer 
 and place opposite its indication the true temperature as 
 figured from the rise in R. Repeat these observations a 
 dozen or more times and there will result two rows of 
 figures the thermometer temperatures and the corre- 
 sponding true temperatures, which we will call respectively 
 a and b. To be most useful these values should be 
 plotted in a curve, where a is laid off horizontally 
 and b vertically. By establishing several points by means 
 of the above experiment we find the law of variation, and 
 are justified in assuming that the law which holds for 
 those points holds for any points in between; hence the 
 points may be connected by a curve. To find what value 
 of b corresponds to an undetermined value of a, draw from 
 the a point in question a perpendicular line to cut the 
 curve at c\ from c draw a horizontal to the b line: the 
 
MEASUREMENT OF INSULATION. 221 
 
 intersection with this line is the value of sought. Such 
 a curve must be plotted for each style of coil which the 
 tester has ordinarily to do with, because the relation 
 between thermometer readings and the true temperature 
 is not the same on, say, a field coil, of an open motor, 
 as it is on a closed one. In the above experiment there 
 must be in circuit with the coil under test, a current in- 
 dicator with which to keep the current constant in the 
 successive steps. The current may be regulated by 
 means of an auxiliary resistance in series with the coil, or 
 in multiple with it, provided the ammeter is cut in with 
 the coil alone, or better still by varying the voltage on 
 the dynamo which supplies the current. To tell when the 
 maximum temperature corresponding to any given cur- 
 rent is reached, we depend upon the fact that when the 
 temperature ceases to rise, so does R, for R's increase is 
 due to increase in temperature. Now if a voltmeter be 
 put permanently across R, its deflection will always be 
 proportional to R, and when R rises, so also will the de- 
 flection. When j? ceases to rise the deflection becomes 
 constant, and we know that the temperature is at its 
 highest value. It is well to wait five minutes after the 
 deflection seems to be constant to be certain of it: also 
 to be sure that the current is at its proper value. 
 
 We know a conductor's resistance to depend upon its 
 length and cross-section, and we have just learned that 
 it also depends upon temperature and specific resistance. 
 Specific resistance is dependent upon quality, and is of 
 the same nature as specific gravity, in that both express a 
 substance's property as compared with the same property 
 of some other substance taken as a standard. The stand- 
 ard of specific gravity is a cubic centimetre, or cubic inch, 
 
222 TESTING OF DYNAMOS AND MOTORS. 
 
 of water at its greatest density. If a cubic inch of another 
 substance weighs twice as much as a cubic inch of water 
 the specific gravity of the substance is 2. The standard 
 of specific resistance is a cube of pure copper at standard 
 temperature. If the resistance of a similar cube of an- 
 other substance measures twice as much as that of the 
 copper one, that substance's specific resistance is 2. It 
 is impracticable to measure accurately the resistance 
 of a cube and moreover very inconvenient to make one. 
 If, however, we know the resistance and dimensions of 
 any regularly shaped conductor, the resistance of an inch 
 cube can be deduced. Suppose the resistance of 1,000 
 feet of No. 10 B. & S. copper wire is i ohm. Its cross- 
 section is .0082 square inch. It will take as many such 
 wires to make up a cross-section of i square inch as .0082 
 is contained in unity, or 
 
 - = 121.95. 
 .0082 
 
 The resistance of 1,000 feet of such a cable is then 
 
 of i ohm or .0082 ohm: 
 
 121.95 
 
 the resistance of i foot is 1/1,000 of this or .0000082 
 ohm, and of i inch 1/12 of .0000082 or .00000068 ohm: N 
 /. <?., the resistance of i cubic inch of the same copper as 
 the wire is, is. 00000068 ohm. If the copper is pure 
 copper, then the actual resistance of the standard is this. 
 The appended table gives the specific resistance of the 
 various metals. The formula for finding the resistance 
 of any conductor is: 
 
 R = TV^^T X Spec. Res. X .00000068, 
 Cross-Sec. 
 
 and the result is expressed in ohms. 
 
MEASUREMENT OF INSULATION. 
 TABLE OF SPECIFIC RESISTANCE.* 
 
 223 
 
 SUBSTANCE. 
 
 RESISTANCE OF 
 I CUBIC INCH. 
 
 RELATIVE 
 RESISTANCE. 
 
 METALS AT O C. 
 Copper (annealed) 
 
 IN MICROHMS. 
 
 I OO 
 
 Copper (hard) 
 
 A,a 
 
 I OIQ 
 
 Silver (annealed) . 
 
 U4J 
 
 Q- 2A 
 
 Silver (hard) . . ... 
 
 oy* 
 
 6J7 
 
 I 086 
 
 Gold (annealed) 
 
 8 10 
 
 i -;i6 
 
 Aluminum (annealed) 
 
 I I J7 
 
 I 8^2 
 
 Platinum (annealed) 
 
 3efjs 
 
 5 882 
 
 Iron (annealed) 
 
 j 2C 
 
 6 2; 
 
 Lead (pressed) 
 
 7 728 
 
 12 05 
 
 Mercury 
 
 
 62 so 
 
 Carbon (graphite) . 
 
 / ' 3 
 
 2CQO 
 
 ALLOYS. 
 
 8.240 
 
 I3.I6 
 
 Copper 60 %, Zinc 26 , Nickel 14 <. 
 Platinum-Silver. . . 
 
 Q 4 C 
 
 1C -ifi 
 
 Platinum 67 %, Silver 33 %. 
 Man^an in 
 
 v-o 
 
 _Q _ 
 
 
 Copper 84 , Nickel 12 , Manganese 4 %. 
 
 LIQUIDS AT 18 C. 
 Pure Water 
 
 18.7 
 
 IN OHMS. 
 
 30.30 
 
 Dilut. Sulphuric Acid 5 % 
 
 1020 ) 
 
 More than 
 
 INSULATORS. 
 
 Glass at 20 C 
 
 I.9I j" 
 IN MEGOHMS. 
 6 X IO 8 } 
 
 1,000,000 
 
 Glass at 200 C 
 
 g 
 
 More than 
 
 Guttapercha at 24 C 
 
 177 X io \ 
 
 1,000,000,000 
 
 * See S. P. Thompson, Elementary Elec. and Mag., p. 404. Also 
 Electrical Dictionary, Houston. 
 
224 TESTING OF DYNAMOS AND MOTORS. 
 
 Now 5 to build up this formula: If the conductor is i 
 inch long, i square inch cross-section, and we deal with 
 pure copper whose specific resistance is i, the formula 
 becomes: 
 
 R X i X .00000068 = .00000068 ohm, 
 
 which is the original expression for the resistance of a 
 cubic inch of pure copper. If the cube is L times as long, 
 the cross-section being kept the same, the resistance 
 becomes L times as great and we have R = L X 
 .00000068 ohm. If the cross-section becomes A times as 
 great, R is diminished as many times as A will go into 
 the original value and we now get 
 
 _ L X .00000068 
 R-- j- 
 
 If we go further and use a metal whose specific resistance is 
 not i, but is S, the formula becomes in its general form, 
 
 X .00000068 , L 
 
 R = ~ - = --X .00000068 ohm. 
 
 A A 
 
 Or in words, the resistance of any conductor varies directly 
 as the length, directly as the specific resistance, and inversely as 
 the cross-section ; and = 
 
 -^ X 68 X io- 8 . 
 
 As an example of the formula's usefulness, we will 
 apply it to the designing of a water box, which shall carry 
 50 amperes at 500 volts. To keep the boiling and 
 evaporation within proper limits, the box should not 
 carry more than 1/2 ampere per square inch cross- 
 
MEASUREMENT OF INSULATION. 225 
 
 section. Fifty amperes will therefore require 100 square 
 inches cross-section. Now since 
 
 F t^oo 
 
 / = or 50 = R --, R -- 10 ohms, 
 
 which must be the resistance of the box. We therefore 
 
 have, 
 
 A' x S X .00000068 ohm or L = --. 
 
 A 6 X .00000068 
 
 Knowing 5 for any liquid, L is readily determined. 
 
 In closing the discussion of current, E. M. F., and 
 resistance measurements, we will consider in a general 
 way what instruments are best suited to different classes 
 of work. A high resistance Thomson reflecting gal- 
 vanometer, provided with a shunt box and directing mag- 
 net, can be adapted to any sort of work, not excepting 
 ballistic work, if the needle has a damping vane at- 
 tached. The Thomson galvanometer is, however, such 
 an expensive and delicate instrument, that its use should 
 be confined to accurate work, and a lower grade article 
 used where it is possible. In current measurement, 
 where the current passes through the galvanometer, the 
 latter should be of low resistance. To use a high resist- 
 ance instrument the main current passes through a shunt, 
 and the instrument takes a known fraction of it. In 
 measuring a potential difference across a resistance, the 
 galvanometer resistance must be high compared to that 
 across which it is placed, so as not to appreciably lower 
 the resistance between the two points to which its ter- 
 minals are attached. In Fig. 85 suppose R to be an un- 
 known resistance in series with battery, B. A is an 
 ammeter and G a galvanometer, to read the potential 
 
226 TESTING OF DYNAMOS AND MOTORS. 
 
 difference between A and C. If G is of low resistance, 
 placing it in multiple with R will lower the resist- 
 ance between A and C, and G's reading will be valueless 
 
 because it is not due to 
 
 I f I I P r~^~| nn the drop on R alone. 
 
 The greater R compared 
 with G, the greater the 
 error caused by intro- 
 FIG. 85. during G. Thus a gal- 
 
 vanometer of 5,000 ohms 
 
 resistance would be a'comparative short circuit across a 
 resistance of 200,000 ohms, and hence not suitable in 
 such a case. We can draw the conclusion, then, that a 
 high resistant galvanometer is adapted to reading poten- 
 tial differences across a low resistance and that a low re- 
 sistance galvanometer is not adapted to reading potential 
 differences at all. Again, the high resistance galvanom- 
 eter must be used if great sensibility is required. A 
 low resistance instrument has only a few turns of wire 
 on it, and hence produces a comparatively weak magnetic 
 field; it is more useful in qualitative than in quantitative 
 work, hence its use in the Wheatstone bridge where it 
 merely indicates the polarity of a deflection; the same 
 is true of any zero method, although in the most accurate 
 zero tests the high grade instrument is frequently used. 
 
 When a galvanometer is to be put in the main circuit 
 the character of instrument is determined by that of the 
 circuit. On a high resistance circuit such as one includ- 
 ing a piece of insulation under test, although the galva- 
 nometer may be of high resistance it will not materially 
 affect the result, for its resistance is small when com- 
 pared to that of the insulation; also the greater sensitive- 
 
MEASUREMENT OF INSULATION. 227 
 
 ness of the instrument is an advantage in this work. For 
 series use in a low resistance circuit the low resistance 
 galvanometer or an ammeter is to be used. On power 
 and lighting circuits the instruments are permanently 
 connected; ammeters should be of as low, and voltmeters 
 of as high resistance as possible, so as to consume only 
 the smallest amount of energy. Ammeters are but low 
 resistance galvanometers; voltmeters are high resistance 
 galvanometers. The ammeter's resistance is made low 
 so that the drop through it shall be small. The volt- 
 meter's resistance is made high so that it shall take but a 
 small current. A station full of poorly designed meters 
 is a source of considerable loss. 
 
 The broad and general rule for all measuring instru- 
 ments is that they must be such as to in no way disturb, 
 or modify the existing conditions of the circuit in which 
 they are to be introduced and used. 
 
TESTING OF DYNAMOS AND MOTORS 
 
TESTING OF DYNAMOS AND MOTORS. 
 CHAPTER VIII. 
 
 THE SERIES MACHINE. 
 
 THE simplest type of self-exciting dynamo is the 
 series wound, a diagram of which is given in Fig. 86. A 
 is the armature, F the field coils, and L L lamps or other 
 load. The distinguishing characteristic of this type is 
 the interdependence of all parts of the circuit. There is 
 but one circuit, and breaking any part of it whatever 
 throws the machine out of action. As a motor, it has 
 several features which make it very valuable in street 
 railway, or other work where frequent interruptions 
 of the circuit are incidental to the service. In doing 
 heavy duty, suddenly imposed, there flows through the 
 circuit an abnormal current, which if prolonged might 
 do harm, but since the field current grows at the same 
 rate as that of the armature, there develops a powerful 
 C. E. M. F., which checks the rise of the current. This 
 and other features, combined with the fact that there 
 is but one circuit to control, makes the series motor an 
 ideal one for street-car work. 
 
 The series dynamo is adapted to arc lamp, or other 
 service where a constant current must be maintained 
 through a variable external resistance. The field winding 
 
 331 
 
232 TESTING OF DYNAMOS AND MOTORS. 
 
 being included in the main circuit, the armature does not 
 generate until the circuit is closed, and then the readiness 
 with which it does so depends upon the external resist- 
 ance, the speed, and in lesser measure upon the strength 
 of the residual field, and the re- 
 luctance of the magnetic circuit. 
 For a given external resistance 
 the initial flow of current will de- 
 pend upon the speed and residual 
 field: if the reluctance is small, 
 the field will respond promptly, 
 and in any case builds up until 
 the current reaches such a value 
 
 that the internal "drop" balances the gain in E. M. 
 F. due to increase of field. As the speed rises so 
 does the E. M. F. , and with it the current, hence also the 
 field strength: finally the field cores become saturated, 
 being usually of lower permeability than the armature, 
 and above this point the field strength remains practically 
 constant. For every particular value of external resist- 
 ance there is a speed corresponding to field saturation, 
 and below this speed the machine will not work satisfac- 
 torily; this is called the critical speed, and below it the 
 field is too light, fluctuating with every slight variation of 
 load. So, also, to every speed there corresponds a criti- 
 cal resistance above which the current is too low to suffi- 
 ciently excite the fields. The higher the speed, the 
 higher the resistance through which a series dynamo will 
 excite itself, and conversely, the lower the resistance, the 
 lower the critical sj^eed. A point to be noted in connec- 
 tion with the critical speed is, that a series dynamo will 
 maintain an established field with a smaller current than 
 
THE SERIES MACHINE. 233 
 
 it takes to produce that field. Thus, if a high resistance 
 be hi circuit while the armature runs at constant speed, 
 and the resistance be decreased till the field " picks up," 
 the machine will hold its field even when the resistance is 
 again increased: however, when the field once begins to 
 fall it does so very rapidly, a process assisted by vibra- 
 tion of floor and machine. On the other hand the field 
 builds up with equal rapidity, and in putting on a load 
 care must be had lest an excessive P>. M. F. be developed 
 and trouble ensue. This point is well illustrated in a bar 
 commutator arc light machine. When working, each arc 
 has a C. K. M. F. of about 50 volts, which is absent when 
 the load is first thrown on; through the very low re- 
 sistance at start the initial E. M. F. sends a large current, 
 over saturating the fields, with a twofold result: first 
 the neutral line is advanced, and if the brushes are not 
 quickly brought forward, vicious sparking results; sec- 
 ond, there is an excessive E. M. F., which, availing itself of 
 the sparking, may flash around the commutator. Spark- 
 ing assists the flashing by volatilizing the copper or car- 
 bon surfaces, thereby providing a conducting vapor for 
 the arc. This flashing, so common to arc machines, can 
 be caused by a sudden variation of external resistance. 
 On full load the voltage may be 2,500 or 3,000; if the 
 line be now broken, the neutral point is thrown back 
 away from the brushes and sparking is set up, resulting, 
 generally, in flashing from brush to brush. Fortunately 
 the arc machine has an inherent factor of safety in its 
 armature reaction, assisted by comparatively easily sat- 
 urated field cores. If the speed be kept constant, at 
 some value above the critical value, and the resistance be 
 gradually decreased, the current will rise, but the field 
 
234 TESTING OF DYNAMOS AND MOTORS. 
 
 will remain almost the same, for after saturation 
 the magnetism does not respond so readily to variations 
 of current. As already learned, the internal drop increases 
 with the current, and hence the terminal E. M. F. de- 
 creases: furthermore shifting the brushes forward to 
 meet the requirements of the neutral line, brings like 
 poles of armature and field closer together, and places the 
 former in a position to exercise greater demagnetizing 
 influence on the latter, thereby weakening the field and 
 reducing the E. M. F. so much that at over load it is less 
 than at quarter, or half load, when the reaction is not 
 so great. Every self-exciting dynamo should be able to 
 pick up its field readily when the armature runs at its 
 rated speed: and in the case of the series dynamo with 
 the added condition that the circuit resistance is not too 
 high. In a series dynamo failure to pick up indicates a 
 fault in either the machine or the line. Whether it is 
 internal or external can be determined by holding a 
 short wire across the terminals of the line, so as to cut 
 out all resistance save that of armature and field coils. 
 If the machine now generates, the fault is in the line; 
 otherwise, it is internal. 
 
 Among the more common troubles occurring to self- 
 exciting dynamos are the following: (i) Loss of 
 residual field; (2) wrong connection of armature or 
 field; (3) open, or (4) short circuit in armature or field; 
 (5) high resistance of brush contact, due to oxidation 
 of brushes, or to shellac on the commutator; (6) the 
 speed may be below the critical value. To these may be 
 added as less common: (7) loosening of a field core or 
 pole piece; (8) armature core below standard diameter; 
 (9) pole pieces bored above standard diameter. The 
 
THE SERIES MACHINE. 235 
 
 effect of the last three failings is to increase the magnetic 
 reluctance, thereby reducing the E. M. F. of the machine 
 and impairing its ability to pick up a field. As a motor 
 the C. E. M. F. would be reduced and the last three faults 
 would cause an increase in speed. On street railway 
 motors parts frequently work loose as a result of pounding. 
 
 The best order to be observed in testing for faults 
 depends upon the circumstances of each particular case, 
 so only general directions can be given here. When 
 a machine in active service suddenly refuses to gener- 
 ate, the trouble may be due to a burn out, to loss 
 of residual field, to open field circuit; or on a series 
 machine to some interruption on the line, e. g., a loose 
 lamp connection or a wire broken by swinging in the 
 wind but held up by the insulation. If a burn out 
 occurs while the machine runs the symptoms will leave 
 little doubt as to the cause, but experience shows that as 
 many machioes burn out stan'ding still, as do running. 
 
 Recharging will restore a lost residual field. Generally, 
 in course of construction, the fields acquire under the file 
 and hammer sufficient magnetism to obviate the neces- 
 sity of charging, but this is not always the case, and 
 strange to say, machines in service from day to day have 
 been known to not only lose their residual field, but to 
 even acquire one of reversed polarity. Since it is impor- 
 tant that a machine's polarity should remain always the 
 same, recharging is resorted to when necessary. This is 
 done by connecting the fields in series with a machine 
 known to be in working order. When the field or arma- 
 ture terminals have been disturbed in any way, and 
 the machine will not generate, either set of terminals, but 
 not both, should be reversed if there is any uncertainty 
 
236 TESTING OF DYNAMOS AND MOTORS. 
 
 as to connections. It is easy to confuse connections in 
 reconnecting a machine dismounted for repair, and unless 
 the proper relation exists between armature and field con- 
 nections, the machine will not generate : for example, sup- 
 pose the two halves of the field winding to be bucking 
 each other: there will result no field at all, for any 
 tendency which one spool exerts toward a definite 
 polarity is neutralized by the counter tendency of the 
 second spool. For securing proper connection every 
 dynamo has its rule which must be observed to obtain 
 satisfaction. A rule familiar to those experienced with 
 Edison machines is as follows: " Connect the right-hand 
 magnet to the lower brush," for the shunt winding, and 
 the "right-hand magnet to the upper brush," for the 
 series. The difference is due to the fact that on one 
 winding the inside ends, on the other the outside ends, 
 are brought out for connection. Where no rule is given, 
 and former experience gives no clew, direct experiment 
 is resorted to, and the connections shifted till the right 
 combination i"s secured. The proper shunt connections 
 once secured, it is easy to get the series connections right, 
 as will be shown under generator testing. 
 
 It sometimes happens that a dynamo when first started 
 shows a small E. M. F. due to the residual field, but on 
 closing the field circuit the E. M. F. falls to zero, and the 
 machine refuses to generate at all. Such action indicates 
 a wrong connection of field or armature, and can be ex- 
 plained as follows: Suppose the dynamo to be properly 
 connected and to be generating; this involves the follow- 
 ing conditions: that the field current is in such a direc- 
 tion as to produce a magnetic field which shall enable the 
 armature conductors, cutting this field, to generate an 
 
THE SERIES MACHINE. 237 
 
 E. M. F. that in turn shall reinforce the initial arma- 
 ture current due to the residual field. Now, without 
 disturbing anything else, let the field terminals be re- 
 versed. For the sake of clearness we will suppose 
 that there remains a residual fiekl, due to the current last 
 flowing: under this condition the pole pieces are of the 
 same polarity as when the machine was properly con- 
 nected, and was generating: since the lines of force due 
 to the residual field are in the same direction as when the 
 armature generated. The small current now in the arma- 
 ture, and due to the residual field, will be in the same 
 direction as it was then, but the field connections being 
 reversed, the current now flows around the spools in 
 such a direction as to neutralize the residual magnet- 
 ism. The slight magnetizing force due to the residual 
 field now opposes this field and soon reduces it to zero, 
 thus totally depriving the machine of all ability to pick 
 up. Nor can a reverse field, even if established by re- 
 charging, be maintained; for assuming a reversed residual 
 to be provided, the lines of force have changed direction, 
 the armature current does likewise, and previous con- 
 ditions being re-established the residual field is again 
 destroyed. Confusion of connections is a common source 
 of failure to generate, but there are other causes less 
 common but similar in effect. Among these are errors 
 in winding and connecting. In rewinding an armature 
 the new coils should be wound on in the same way that the 
 old ones were. Other things being the same, the armature 
 polarity depends upon the manner of winding and con- 
 necting. If in the original armature, the winder turned 
 the core over and from himself, in rewinding the same 
 rule must be observed. In connecting, the leads may be 
 
238 TESTING OF DYNAMOS AND MOTORS. 
 
 brought around oppositely to what they were originally 
 or they may be brought around one bar too far, any of 
 which mistakes reverses the armature polarity, and 
 changes its relation to the field. Under these circum- 
 stances, the attendant may be very much surprised to 
 find that his dynamo refuses to generate under apparently 
 the same conditions that existed before rewinding. In 
 such a case, a reversal either of direction of rotation, or of 
 field or arntature terminals, will restore the dynamo to 
 working order. It is not always convenient to reverse 
 the direction of rotation, as it involves either reversing 
 the engine or turning the dynamo end for end; while the 
 brushes, if of copper, must also be changed, and this may 
 necessitate altering the brush holders. On arc machines 
 provided with regulating devices the latter are often con- 
 structed and adjusted for a given direction of rotation. 
 On such a machine let us suppose the regulation to be 
 effected by means of the brushes, and also suppose the 
 direction of rotation to be changed. If for increase of 
 current the mechanism throws the brushes/tfr^w^/giving 
 them a positive lead, with the reversed rotation, this 
 would now be a negative lead instead of positive; while 
 for decreasing current drawing the brushes back is to give 
 them a positive lead. In either case the sparking is 
 increased instead of diminished, and in the latter case 
 flashing is apt to ensue. It is therefore necessary to 
 preserve the direction of rotation undisturbed. 
 
 If one of the above reversals is found necessary, it is 
 customary to reverse the field connections, as they are 
 near together and a neat job can be done. Having 
 determined that failure to generate is not due to wrong 
 connection some other source of trouble must be sought 
 
THE SERIES MACHINE. 239 
 
 There are many tests for open and short circuit in field 
 or armature. If while the fields are separately excited, 
 the armature be rotated, a short circuit in the fields will 
 probably be indicated by too low voltage at the brushes, 
 since part of the field coil is inactive, and the armature 
 is unable to produce the voltage called for by the 
 known voltage applied to the fields. This test is not 
 infallible, and an existing short circuit might escape 
 detection, unless the field circuit contains an ammeter, 
 and the field current is maintained at normal value. 
 Thus, suppose there are two field coils of equal mag- 
 netizing power and resistance, and that the short circuit 
 is such as to entirely cut out one coil; for a given 
 E. M. F. at the field terminals, the current will now be 
 twice its normal value, and hence the field ampere-turns 
 exactly the same as if no short circuit existed : the field 
 strength will therefore be unaltered. If the fields under 
 test are those of a constant current machine, and are 
 connected in series with a second constant current 
 machine, or in any case if the field current is kept con- 
 stant, the test becomes a decisive one. Open circuit 
 would, in a series machine, be indicated by the absence 
 of all current. Open circuit in the fields can in all cases 
 be tested for with an ordinary bell and battery, a mag- , 
 neto, or a test lamp circuit. Fig. 87 gives the connections 
 for a bell and battery. B is the battery of two cells, 
 and at M a bell. T T' are test lines. The whole outfit 
 can be put into a small box, and T T', when not in use, 
 are coiled around a cleat on the cover. Fig. 88 gives 
 the arrangement of a test lamp circuit. A and B are 
 light or power mains, L l Z 2 Z 3 are lamps in series, the 
 number depending upon the voltage in use: these lamps 
 
240 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 are inserted in keyless sockets mounted on a board. 
 T 7 1 ', are test lines, and just before use should be held 
 together to insure that the lamp circuit is intact: they 
 are then held across the field terminals, when the light- 
 ing of the lamps will indicate the fields to be a closed 
 
 5 
 
 
 V-^\> 
 
 r . 
 
 T 
 FIG. 8; 
 
 
 circuit. If the test circuit is taken from a street rail- 
 way line, one must bear in mind that one side of it is 
 ''grounded," and should the machine under test be on 
 the ground, some of the results might be delusive. If the 
 normal field resistance is known, short or open circuit 
 can be readily detected by measuring it; if it proves to 
 be high or infinitely great, a partial or complete open 
 circuit exists; while a low value indicates short circuit. 
 If the fields are in sections or on separate spools, they 
 should be disconnected, tested individually, and the 
 faulty one removed. In cases of necessity, and where a 
 machine runs alone, the faulty spool can be disconnected 
 and left till a more convenient time for repairing it. 
 
 This test can be applied to armatures when their brush 
 to brush or bar to bar resistance is known. In general 
 it is more difficult to locate armature troubles, unless gal- 
 vanometer tests are resorted to, because there is, through 
 the armature, a double path fromibrush to brush, so that 
 
THE SERIES MACHINE. 24! 
 
 although a wire may be broken in one path, the circuit is 
 complete through the other. If the armature be run in sep- 
 arately excited fields it will spark at the brushes in case of 
 either an open or short circuit, because the armature is 
 electrically unbalanced, and the neutral point shifts con- 
 stantly. A short circuit can often be located by remov- 
 ing the brushes, and running the armature at its usual 
 speed in the fully excited field. The field must be 
 separately excited, for with a defective armature the 
 machine will not excite itself even though the brushes be 
 restored. Now the resistance of a whole armature is 
 not high, and the resistance of the short circuited coil 
 being very low, the electromotive force generated by the 
 coil produces sufficient current to heat the wire and 
 ultimately to burn it out. Open or short circuit can also 
 be detected by separately exciting, as above, then turn- 
 ing the armature slowly, and watching the needle of a 
 voltmeter placed across the brushes. In case of open 
 circuit, the needle will fluctuate between zero and some 
 definite value, depending upon the field strength and the 
 speed. If a short circuit is present, the needle will not 
 drop to zero, but will fluctuate between two definite 
 values. The above methods were well illustrated in 
 the following case, of an armature which refused to move 
 when connected as a motor; it was then run as a sepa- 
 rately excited generator, and the voltmeter read zero 
 across the brushes, but the armature became very warm 
 after a ten-minute run. The heating, however, was not 
 local, but uniform over the whole armature. The com- 
 mutator was then disconnected and tested, and the 
 insulation between bars found to be defective, thus short 
 circuiting every coil upon itself. Connected as a motor 
 
 OF THE 
 
 ((UNIVERSITY J 
 
242 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 the current entering at one brush passed directly through 
 the carbonized insulation and the commutator bars to the 
 other brush, without entering the coils at all; when run 
 as a generator, each coil, short circuited on itself, was 
 heated, but since there was no E. M. F. additive from 
 brush to brush the voltmeter gave no sign. 
 
 The same method in a modified form can be used to 
 detect a ground in an armature, the effectiveness of the 
 test depending upon the fact that 
 in a defective armature the E. 
 M. F. generated is greater when 
 the faulty coil is in one part of the 
 field than when in another. The 
 voltmeter will indicate this fluctua- 
 tion if the speed is not too high. 
 When an armature becomes inter- 
 nally grounded it is customary 
 'to burn out the fault if possible 
 so that it may be more readily 
 seen in course of stripping; this is a practice adhered 
 to even in the case of a short circuit, but the writers 
 do not approve of the practice unqualifiedly in that 
 it often- proves to be an injustice to the winder. 
 Apparent difficulty in an armature can often be traced 
 to a defect in the commutator, by disconnecting and 
 testing it, and such a fault can certainly not be laid 
 on the winding. Fig. 89 gives connections for detecting 
 and burning out a ground. The armature, A, is grounded 
 at G' (/. e., the copper wire touches the iron core, or as 
 railway men say : "The machine has an iron circuit "). 
 A second ground is created by connecting one brush to 
 the frame. If the machine be now run as a separately 
 
 FIG. 89. 
 
THE SERIES MACHINE. 243 
 
 excited generator and from a motor of moderate capacity, 
 the armature will alternately start and stop once in every 
 revolution. This "bucking" is not so marked if the 
 driving power is a larger unit. When the fault, G', is in 
 the position of the figure, there are two short circuits 
 present, one through G ', i, E, 4, 6, G, the other through 
 ', 2, E, 4, 6, G. Two parts of the winding are in mul- 
 tiple, and through the low resistance of the short circuit 
 a large current flows, loading the 
 dynamo so that the motor cannot 
 run it. As soon as the dynamo 
 stops, its current, and hence load, 
 becomes zero, and the motor thus 
 relieved starts again, only to re- 
 peat the same operation. The 
 point of maximum speed corre- 
 sponds to the position where 
 the fault passes under the 
 grounded brush holder, for it is 
 
 here the two grounds become one, and have no more 
 influence than a single ground has on any otherwise 
 insulated circuit. As G' is carried past the grounded 
 brush the E. M. F. in the grounded circuit rises and cur- 
 rent once more flows. Fig. 90 shows a method of find- 
 ing open circuit: Suppose a break to be at A; it is 
 desired to determine between what two bars it is located. 
 Open a circuit at a second point, as at B, by disconnect- 
 ing one of the leads. There are now two parts of the 
 winding insulated from each other. Next place one test 
 line on bar No. i and move the other around the commu- 
 tator till the bell ceases to ring, which will be as soon as 
 the sliding contact rests on bar No. 10, thus showing the 
 
244 TESTING OF DYNAMOS AND MOTORS. 
 
 fault to be between Nos. 9 and 10. To verify the result, 
 place one line on No. n and slide the other back and 
 forth over Nos. 9 and 10: the bell will cease ringing upon 
 reaching bar 9. Enough of the head can now be removed 
 to examine the damaged coil. Trouble of this nature 
 
 does not generally occur in 
 the more removed parts of 
 the winding, and is often due 
 to the melting of a connect- 
 ing wire. Poorly " sweated " 
 connections are a fruitful 
 source of trouble: a poor 
 connection will always heat 
 and may reach a temperature 
 sufficient to melt the solder and release the connector 
 from the ear or cup of the commutator bar, thereby 
 causing an open circuit. Such trouble is more often due 
 to mismanagement than to any defect in the machine 
 itself. 
 
 We have learned that a short circuit in the field wind- 
 ing impairs a machine's ability to excite itself and in 
 increased measure as the short circuited portion is 
 greater. This fact is utilized on series dynamos for 
 removing them from service. In Fig. 91 A' is the line 
 switch and K' one which short circuits the field, F. 
 When the dynamo is in service K is closed and K 1 open. 
 To remove the load K' is closed. K is generally left 
 closed. 
 
 Whether failure to generate is or is not due to low 
 speed can be ascertained by taking the speed of the 
 armature or that of the countershaft which drives it, and 
 multiplying this by the ratio of the diameter of the 
 
THE SERIES MACHINE. 245 
 
 countershaft pulley to that of the armature pulley. 
 Assuming that the engine and countershaft speed are 
 correct, that of the armature is governed by the size of 
 its pulley, and can be altered by using a pulley of greater 
 or less diameter; this step, however, should not be taken 
 unless the designed speed of the armature is known, for 
 many armatures will not run at full load without sparking, 
 if the speed be much below or much above its rated value. 
 As a rule, in order to secure sparkless running the field 
 magnets are made so powerful, and the shunt and series 
 coils (on a compound-wound machine) so proportioned that 
 even on full load the armature causes but slight distortion 
 of the field, and the position of the neutral points, hence of 
 the brushes, is practically unaltered. In poorly designed 
 constant potential machines, and in series machines in- 
 tended to regulate for constant current, this proportioning 
 of armature and field is not observed, so that the position of 
 the neutral points depends upon the load; changing the 
 load shifts the position of the neutral points, and unless 
 the brushes are shifted accordingly, sparking ensues. 
 The effect of changing the armature speed is, for any 
 given load, to change the above relation between armature 
 and field, and hence to cause a shifting of the neutral line, 
 with resulting sparking. For a given position of the 
 brushes sparking can be avoided only by strengthening 
 or weakening the field enough to restore the neutral line 
 to its position under the brushes. This is an effect seen 
 not only on badly designed constant potential machines, 
 but also on well designed ones if the armature is run 
 much below its rated speed. To have the field strong 
 enough to control the position of the neutral point there 
 must be the designed voltage at the field circuit terminals: 
 
246 TESTING OF DYNAMOS AND MOTORS. 
 
 to lower the speed is to lower this voltage, and with it 
 the field current, so the field magnetizing effect is lower, 
 while that of the armature remains constant. Raising 
 the speed strengthens the field, but raises the E. M. F. 
 of each coil as it is short circuited by the brush, and 
 enables it to generate current enough to cause sparking. 
 This is avoided in ordinary running by using carbon 
 brushes of comparatively high resistance. 
 
 Series machines are widely used on arc light circuits, 
 and when run together are generally connected in series. 
 When so connected they give no trouble-in load regula- 
 tion, each unit supplying its share of the total voltage, 
 the same current passing through all. The total watts 
 generated equals the total E. M. F. X current = E I = 
 (e -f- ^, + ^ 2 + ^ 3 +, etc.) 7, where e, e lt etc., are the 
 E. M Fs. of the individual machines, or W = (e -\- e l -\- 
 e i + e 3 -K etc - ) ^ anc * we see that the load on each machine 
 is directly proportional to its E. M. F. Machines of 
 different capacities can therefore be run together without 
 overloading any, provided / does not exceed the current 
 capacity of the smallest machine. The advantage of the 
 series connection is that many lamps can be placed on 
 one circuit, thereby saving copper and other expenses. 
 Series machines can be run in multiple, provided one of 
 the following precautions is taken: (i) the machines with 
 rigid connection between the shafts (direct coupled), 
 must be started up with both line switches in; or (2) the 
 fields must be separately excited, when of course they cease 
 to be series machines; or (3) an equalizer must be used 
 to regulate the sharing of the load, and to prevent any 
 dynamo from reversing and running as a motor. The equal- 
 izer is especially used with compound-wound machines, 
 
THE SERIES MACHINE. 247 
 
 and its consideration is therefore deferred. If the arma- 
 ture shafts are rigidly connected the necessity of start- 
 ing with both line switches in is easily seen; for since 
 a series machines is devoid of field until the external 
 circuit is closed, throwing in its line switch, when other 
 machines are in service, would result in a short circuit 
 through its armature. This would run the machine in 
 question as a motor and in the opposite direction, since 
 series motor and dynamo reverse rotation for given con- 
 nections. The rigid connection would prevent actual 
 reversal of rotation, but unless a clutch slipped, or a cir- 
 cuit breaker worked very promptly, something serious 
 would happen. With separate excitation the machine's 
 E. M. F. can be made equal and opposite to that on the 
 line, and its switch closed. Other things being equal, 
 the load which each machine takes depends upon its own 
 E. M. F., and this in turn upon the field strength. With 
 separate excitation regulation is obtained by means of 
 a rheostat, and automatic devices depending upon the 
 armature reaction are generally dispensed with. In 
 practice the only circumstances under which arc machines 
 are run in multiple, is where the lamps used require a 
 current double that of either machine. The machines 
 cannot be expected to regulate very closely, unless they 
 are magnetically and electrically balanced throughout 
 their load range. As a rule the lamps do not give 
 satisfaction. 
 
 Series dynamos can be designed to regulate for a 
 fairly constant potential between certain load limits. 
 We have learned that below a certain current value the 
 fields are insufficiently excited, and up to this point the 
 voltage fluctuates. If this critical point be brought near 
 
248 TESTING OF DYNAMOS AND MOTORS. 
 
 the saturation point of the iron, the voltage due to 
 increased field strength as load goes on may increase 
 at the same rate as the drop in the line increases, thus 
 giving constant potential at the lamps or other load 
 across the mains. The limits of regulation, however, are 
 narrow, and for light and heavy loads, poor. In attempt- 
 ing to run several machines in multiple there arise com- 
 plications which practically exclude series dynamos from 
 this class of work, now covered by shunt- and compound- 
 wound dynamos. On the other hand, while arc lamps 
 are generally run from series machines, fhey are often 
 operated from the same dynamo that supplies incan- 
 descent lamps, or trolley system, at constant potential. 
 In this case as many lamps are placed in series as can be 
 worked from the given voltage, and a German silver 
 resistance is placed in circuit with them to avert the 
 short circuit, which otherwise would occur were all the 
 carbons to come together at once. This added resist- 
 ance also improves the regulation of the lamps, which 
 are unsteady when connected directly across the mains. 
 The effect of the resistance is to cushion the current 
 fluctuations by narrowing the limits between which it 
 can fluctuate. Thus suppose two arc lamps are in series 
 across 125 volt mains. Let the total lamp resistance 
 when the carbons are together be i ohm, and let the 
 extra resistance be 4 ohms, making the total circuit 
 resistance 5 ohms. Without the extra resistance the 
 current could fluctuate between o and 125 amperes, while 
 with it the limits would be o and 25 amperes, the zero 
 value corresponding to open circuit, due to the carbons 
 drawing too far apart. On modern lamps such action is 
 hardly probable, as each lamp is provided with a device 
 
THE SERIES MACHINE. 249 
 
 for cutting out the lamp when the arc reaches a certain 
 length. 
 
 The problem of regulation on arc machines is solved 
 in one of three ways. In the first there is a mechanism 
 operated either by a solenoid around which the working 
 current flows, or by a magnetic vane, which moves under 
 the influence of the external field magnetism; in both 
 cases a system of levers is made to operate the rocker arm 
 to which are attached the brushes. According as the 
 load increases or decreases, the brushes are moved 
 forward or backward. Such a regulator governed by a 
 solenoid can be made very sensitive, as is illustrated in 
 the Thomson-Houston arc machine. The magnetic vane 
 was used on the Edison arc machine, now relegated to 
 history. The second method of regulation depends upon 
 a solenoid for its action, but the function of the magnet 
 is to press together with varying force, according to load, 
 a number of carbon plates which constitute a shunt to 
 the field winding. If the number of lamps in circuit is 
 increased, and the current begins to fall off, the plunger 
 of the solenoid relaxes its pull on the lever, which con- 
 trols the pressure on the plates, and the conductivity of 
 the contacts is thus impaired : less current passes through 
 the shunt and more through the field coil, and the field 
 becoming stronger the machine's E. M. F. is increased 
 sufficiently to provide for the additional lamps. This 
 type of regulation is found only on the Brush arc light 
 machine, and gives great satisfaction. It is the simplest 
 electrical regulator in use. Since the solenoid is oper- 
 ated by the main current it responds very promptly to 
 load variations, the carbon plates forming a high non- 
 inductive resistance. 
 
250 TESTING OF DYNAMOS AND MOTORS. 
 
 Other less well known makers use one or the other 
 of these principles. 
 
 The third and simplest mode of regulation requires the 
 use of no mechanism whatsoever, and depends upon the 
 reaction between armature and fields. As has beeri 
 pointed out in the preceding chapter, all series dynamos 
 depend in a measure upon armature reaction for regula- 
 tion, and in this type of machine the design is such that 
 the regulation is close from no load to full load. 
 
 On account of the demagnetizing effect of the arma- 
 ture, an arc machine can be short circuited without 
 injury, such action resulting in removal of the load, 
 accompanied by some sparking. This method might be 
 used for removing the load ordinarily, but short circuit- 
 ing the fields alone accomplishes the same object and 
 without sparking, besides rendering the machine abso- 
 lutely inactive. To short circuit the armature brings its 
 reaction to a maximum, and this might reverse the field 
 polarity, thereby causing the arc lamps to burn upside 
 down. This armature reaction gives rise to peculiar 
 behavior at times, and the complications which it causes 
 in certain tests will be considered later. 
 
 As over 90^ of the world's arc lighting is done by 
 Thomson-Houston and Brush arc machines, a more par- 
 ticular study of these machines will be in order. 
 
 Thomson- Houston Arc Dynamo. Professor Sylvanus P. 
 Thompson says of this machine: " Its spherical arma- 
 ture is unique among armatures; its cup-shaped field 
 magnets are unique among field magnets; its three- 
 part commutator is unique among commutators." The 
 armature is of the open coil type, but is wound in a 
 peculiar way. The inside ends of the three coils are 
 
THE SERIES MACHINE. 25! 
 
 soldered together, while the outside ends connect each 
 to one commutator bar. In the earlier machines 
 the armatures were hand wound as follows: first, half 
 of the first coil was put on; then half of the second 
 coil; then the whole of the third coil, next, the second 
 half of the second coil, and finally the first coil is 
 finished. By this means the average distance of each 
 coil from the core is made the same. In the more recent 
 
 F 
 
 FIG. 92. 
 
 types the coils are lathe wound and are laid onto the 
 core, which has a removable section. This cheapens the 
 cost of production, and greatly facilitates their repair. 
 The operation and regulation of the machine can be 
 seen in Fig. 92. The three coils of the armature are 
 indicated by the figures, i, 2, 3. The current leaves the 
 armature, A, by the brushes B\ B^ passes through the 
 field F^ through the regulating apparatus and out on 
 the line, returning to the machine through field, F^ and 
 brushes B\ B v The regulation of the machine is effected 
 by means of the magnets M and r, while 5 is a high 
 resistance carbon shunt. The magnet c is the controller 
 and is adjusted by means of a spring, so that the rated 
 
252 TESTING OF DYNAMOS AND MOTORS. 
 
 lamp current will just raise the cores, and separate K, 
 If A" is closed, the current in J/, which is the regulator 
 magnet operating the lever arm and brushes, becomes 
 weaker and the core in M drops by gravity. Let us sup- 
 pose that the current in c, and hence in the lamps, is too 
 small: the cores in c drop, K closes, the current in M 
 becomes smaller, the regulator core drops, and brushes 
 B^ and B^ are shifted backward while B\ B\ are shifted 
 forward. This drawing together of each pair of brushes 
 shortens the period of short circuit of each coil, lengthens 
 the time that two coils are in series, hence, raises the 
 E. M. F. and with it the line current. If <r's current gets 
 too large, the reverse action takes place; K opens, M 
 rises, and the pairs of brushes separate, the period of 
 short circuit is increased, and the voltage lowered. 
 The resistance 6" takes up the spark when A" is opened, 
 and a dash pot eases off the motion of M. The current 
 obtained from the machine is always in the same direc- 
 tion, but fluctuates between zero and its maximum value 
 no less than thirteen times * per revolution. These 
 fluctuations are too rapid to have any perceptible effect 
 on the working of the lamps, but undoubtedly produce 
 induction effects on neighboring telephone circuits. 
 
 All direct current series dynamos spark somewhat upon 
 starting, because any regulating device requires time to 
 take hold. Ordinarily the brushes should not spark to 
 any extent under load, or even overload. In open coil 
 armatures the sparking is greater than in closed coil 
 ones, because in them each coil is alternately cut out and 
 cut into circuit during every revolution. In the T.-H. 
 
 * See paper by M. E. Thompson, read before the A. I. E. E., May 
 21, 1891. 
 
THE SERIES MACHINE. 253 
 
 machine this effect is increased on account of the large 
 angle covered by each coil; the coil being often cut out 
 of circuit while generating a strong E. M. F. The 
 method by which the sparking is controlled, and even 
 made use of, is very ingenious. In the first place we 
 must note that during part of each revolution each coil 
 is short circuited by each pair of brushes, /. e., when two 
 commutator segments touch two brushes which are con- 
 nected, the included coil short circuits through the 
 brushes. Since this short circuiting takes place in a by 
 no means weak field, the current in the coil is consider- 
 able, and so also the danger of breaking down the insula- 
 tion, if the current is too suddenly broken. First then 
 it is necessary to guard against a too sudden breaking of 
 the current, and this is done by the spark which acts 
 precisely as the carbon shunt S in the controller. The 
 machine, however, is generating a very high E. M. F., 
 and were this spark allowed to persist unmodified it 
 would heat the air between the commutator bars, and 
 cause flashing from brush to brush. Furthermore, in 
 order that the commutator may run smoothly and last 
 long, it must be oiled, and this adds to the danger of 
 flashing. The air blast was introduced to suppress the 
 spark after it has served to free the coil of its current of 
 self-induction, and to introduce a thin layer of cold air 
 in place of the heated air due to the spark. This air blast 
 makes it possible to run a machine to as high as 4,500 
 or 5,000 volts, where without it 500 volts would cause 
 flashing. The value of the air blast is most keenly felt 
 when it gets out of order. With each machine a brush 
 gauge is sent, by means of which the four brushes may be 
 given the proper lead and correct relative positions. 
 
254 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 The brushes constitute the "cut-out," so called because 
 they cut out each coil a part of every revolution. Care 
 must be taken that the angle of the brush holder is cor- 
 rect, and that the distance of each brush holder is exactly 
 the same, as measured on the straight edge of the gauge. 
 General directions for brush setting are as follows par- 
 
 FIG. 93. 
 
 ticular directions accompany each machine: Set the 
 brushes in an exactly straight position; turn the com- 
 mutator in the direction of rotation until there is just 
 contact between brush J? 4 and the end of the segment^', 
 (see Fig. 93). In this position the tip of brush J3 l should 
 hang over the edge, C, of the segment A" one sixty- 
 fourth (1/64) of an inch. Now turn the commutator till 
 the brush B^ just makes contact with the point C on 
 segment A" '; the brush B^ should hang over the follow- 
 ing segment 1/64 inch. Having set the brushes properly, 
 next see that the air blast is adjusted; to do this, the 
 regulator arm must be raised and fixed at its extreme 
 height; then the jet is put in, and the point brought to 
 
THE SERIES MACHINE. 255 
 
 within about 1/32 inch in front of the brush, and the 
 same distance above the face of the commutator. In 
 small machines the air blast is sometimes omitted, and 
 when tliis is the case the spark should be about 1/16" 
 for full load. With the air blast the spark is increased 
 to 1/8" and on large machines should be 3/16". The 
 proper working of a T.-H. arc machine depends upon 
 the lead of the commutator, and upon the proper adjust- 
 ment of air blast and "cut-out" securing the correct 
 length of spark. Often, important conclusions can be 
 drawn from the character of the spark, so that the fol- 
 lowing points should be carefully noted : 
 
 Possible causes of excessive flashing at the commuta- 
 tor: (i) The air blast may be loose on the back plate; 
 (2) Air screens may be stopped with dirt; (3) Air blast 
 "wings " may be in end for end top edge should be in 
 front; (4) Air blast jet may be jammed; (5) Air blast 
 jet may be set too far from the ends of the brushes; (6) 
 Air blast pins may have been changed in position and 
 be incorrectly adjusted; (7) Air blast jet may not be 
 properly set with respect to the commutator slots; (8) 
 The commutator may be set wrong; the mark on the 
 commutator and that on the machine should line up; 
 (9) The "cut-out" may be too strong; (10) The regula- 
 tor yoke connections may be sticky; (n) There may be 
 a poor contact in the controller; (12) The dash pot may 
 be either too stiff or too weak; (13) There may be too 
 much oil used about the blower, or the oil may be of 
 inferior quality; (14) Too frequent wiping of commuta- 
 tor may have accumulated lint, which, by carbonizing, 
 short circuits the commutator; (15) The machine may 
 be overloaded, or the belt may be slipping; (16) There 
 
256 TESTING OF DYNAMOS AND MOTORS. 
 
 may be either a short or open circuit in the armature or 
 field; (17) There may be a portion of the circuit which 
 is periodically cut out by contact to ground, as by wires 
 swinging in the wind; (18) There may be bad or dirty 
 controller contacts which keep the regulator cut in almost 
 constantly. Many of the above cases, especially those 
 due to the presence of dirt, should never occur, and 
 some of the remaining ones can be quickly eliminated. 
 A contact or ground in a single coil of the armature 
 makes a sputtering, broken sparking at the commutator, 
 the current surges, rising high, then falling low. This 
 surging may also be caused by a lack of sensitiveness in 
 the controller magnet, for this makes the regulator rise 
 above then fall below the normal points of variation. 
 
 The machine will generate a heavy current if by 
 grounds or other short circuit part of the turns on the 
 regulator magnet are cut out; this weakens the magnet 
 and allows the regulator to fall too low. On the other 
 hand a short circuit in the field coil, cutting out part of 
 the turns, weakens the field's intensity, and lowers the 
 capacity of the machine. Two grounds, one on each 
 end of the field coil, deprives the machine of generating 
 power. If the machine fails to carry the standard num- 
 ber of lights, the failure may be due to a short circuited 
 armature. To test for this, turn off the lights and run 
 for a short time on dead short circuit, watching the am- 
 meter to see that the current is not 'too high; then if 
 there is no short circuit the regulator armature should 
 be up to within r/i6" of the bottom of the regulator. If 
 there is a short circuit the regulator armature will hang 
 down from 1/4 to 1/2 inch. A short circuit in any 
 armature can be detected by running it in separately 
 
HIE SERIES MACHINE. 257 
 
 excited fields, and on open circuit; upon holding a piece 
 of iron in the air gap near one pole a vibrating puil can be 
 felt if the armature has a cross in it. The vibration is 
 unmistakeable, and is due to the fact that the faulty 
 armature is a stronger magnet when the cross is in one 
 part of the field than in another, thus giving rise to an 
 alternately stronger and less strong attraction for the 
 bit of iron. If there is a cross the armature will, if run 
 long enough, grow hot and ultimately burn out the fault. 
 In recent types of armature the short circuited section 
 alone will heat and may be thus located, but in the old 
 types this is not possible. 
 
 Short circuit in the field can be roughly guessed at by 
 noting which pole pulls strongest on a piece of iron; or 
 better, the resistance of each field can be measured 
 directly or calculated by use of Ohm's law. 
 
 A ground between armature and shaft can be detected 
 by connecting each brush alternately to the frame by 
 means of a "jumper" (/'. e., short wire). Flashing at 
 the commutator results if there is a ground. The use of 
 the jumper will also cause flashing if there is a ground 
 between regulator and frame. These grounds may be 
 too undeveloped to show up through a magneto, and 
 should be subjected, as above, to the full voltage of the 
 machine. At all events they should be located and 
 removed, as it is hazardous to run so high a potential on 
 a grounded circuit supposed to be insulated. Should a 
 machine that has been in good working order begin to 
 flash, first look over the* sliding connections on the 
 yokes; on old machines these become weakened and 
 dirty; next clean out the blower and air screens, and see 
 that the blower is properly set; look over the controller 
 
258 TESTING OF DYNAMOS AND MOTORS. 
 
 contacts. If the machine flashes when being loaded 
 and the spark becomes very short, and the current runs 
 low, there is a poor, /. <?., high resistance, connection on 
 the right lead to the controller, whereby the regulator is 
 not short circuited when the controller contacts touch. 
 If the regulator drops clear down, either the regulator 
 has a short circuit or the controller contacts touch per- 
 manently. 
 
 A very long spark indicates that the commutator has 
 been wrongly set, and has a long positive lead; the 
 regulator also drops to the bottom in such a case. On 
 the other hand if the commutator works loose and slips 
 back, making a negative lead, the regulator arm is brought 
 up and flashing ensues. The lead ot the armature is very 
 important. If the fields are very strong a long positive 
 lead must be given, otherwise the regulator will drop, 
 the current will be low, and the spark long and heavy. 
 In rare cases a rheostat is placed in shunt across the 
 fields to weaken them if it becomes necessary. In such 
 a case care mus.t be taken that the box is placed across 
 the fields and not across the armature. To make mis- 
 takes less likely, the field connections must be so 
 changed as to place both spools on the same side of the 
 armature, otherwise the shunt box must be in two parts 
 to avoid shunting the armature. The cut-out is of equal 
 importance with the lead, and should be so set that the 
 current will not be more than 4$ or 5$ over the rated 
 value. If the cut-out is set very strong, and the lead at 
 the same time be very weak or negative, the regulator 
 will come up hard, and the current will be high; as the 
 load goes on, flashing is likely to ensue. In general, 
 where flashing is due to wrong cut-out and lead, as 
 
THE SERIES MACHINE. 
 
 259 
 
 Q 
 
 1< 
 
 i 
 s 
 
 nj 
 * J L 
 
 
 
 
 - 
 
 
 ^i 
 
 _ _ - 
 
 J I 
 
260 TESTING OF DYNAMOS AND MOTORS. 
 
 indicated by surging and easy flashing, the cut out 
 should be weakened and the lead lengthened. It is 
 interesting to note that by setting the commutator back 
 90, the T.-H. machine can be run as a series motor, 
 though at a very low efficiency. 
 
 Enough has been said to show that a general experi- 
 ence with other machines does not make one thoroughly 
 at home with a T.-H. arc machine; and that an intimate 
 knowledge of the latter will greatly facilitate caring for 
 it. When understood, and intelligently handled, the 
 machine easily holds, on its own merits, its recognized 
 place among arc dynamos. 
 
 The Brush Arc Dynamo is the pioneer of arc machines. 
 The first one had its bobbins wound on a solid armature 
 core, and heated to a wasteful degree. This fault was 
 soon eliminated. In structure this excellent machine 
 differs very materially from other types. In the first 
 place the field poles confront either side of the arma- 
 ture, and hence the magnetic lines pass through the 
 armature parallel, not perpendicular, to the shaft. The 
 armature core is therefore laminated, not with discs, but 
 with concentric sheets formed by winding up band iron. 
 The space between coils is occupied by projections 
 built up of sheet iron, and serving the double purpose of 
 holding the coils in position and improving the magnetic 
 circuit. By means of Figs. 94 and 95 we see that the 
 field spools are in series with each other and the arma- 
 ture. To facilitate assembling, the brass tips of each 
 magnet are lettered, those of like letters connecting 
 together. A unique feature of the field structure is that 
 between each coil and the iron core is a sheet of copper 
 soldered to form a sheath, which, by itself becoming the 
 
THE SERIES MACHINE. 
 
262 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 seat of induced currents, due to the fluctuating field cur- 
 rent, protects the iron core from these effects. 
 
 The second .point in which the Brush machine differs 
 from others, follows from the first. Since the field poles 
 face the armature's side, here are to be found the active 
 wires, and not on the outer face. The armature is, there- 
 fore, made as short as is practicable, and somewhat 
 
 A 3 
 
 DIAGRAM OF 
 ARMATURE CONNECTIONS 
 
 FOR 
 NO. 1 & 1 1 ARC DYNAMO 
 
 FIG. 96. 
 
 resembles a large disc. Machines as now made have 
 eight, sixteen, or twenty-four armature bobbins, divided 
 into sets of four coils. The coils of each set are in 
 series, their final leads are connected to adjacent com- 
 mutator bars, and alternate in position and action with 
 those of the other sets. The connections for two sets 
 of bobbins are shown in Fig. 96. Each set is indepen- 
 dent of the others, is effective only when its bars pass 
 under the brushes, so that the machine is virtually two 
 machines with open coil armatures. The armature coils 
 are set symmetrically not only with respect to each other 
 
THE SERIES MACHINE. 263 
 
 but to the magnetic field. All four bobbins are in the 
 maximum field together, and in zero field together. 
 The E. M. F. in each set therefore rises to a maximum, 
 then falls to zero. Fig. 97 shows the character of the 
 E. M. F. and current curves of an eight-coil machine. 
 Curve I is due to one set of coils, and II to the other 
 set 45 behind the first. The heavy line denotes the 
 resulting E. M. F. at the brushes. In the larger 
 
 II 
 FIG. 97. 
 
 machines additional sets of coils are used, making the 
 E. M. F. steadier, and giving a larger current capacity 
 or a higher E. M. F., as the design may require. In the 
 twenty-four coil machine shown in Fig. 96 we have prac- 
 tically three dynamos of eight coils each, in parallel; 
 this is shown by the three commutators which are set 
 each 1 1 % behind each other, corresponding to the rela- 
 tive position of their respective armature coils. In the 
 figure only eight of the twenty-four coils are shown. 
 A^ A lt A^ A % are in series and their leads connect to bars 
 A^ and A % . In Figs. 94 and 95 it can be seen that the 
 brushes are set 90 apart, instead of 180, as on many 
 machines, indicating that the commutator is cross con- 
 nected, /. e, , opposite bars are metallically connected, 
 thus obviating the necessity of using over one pair of 
 brushes per commutator. 
 
264 TESTING OF DYNAMOS AND MOTORS. 
 
 This compound commutator is the third distinguishing 
 characteristic of the machine; it is air insulated, and built 
 upon wooden blocks. The bars are long enough to keep 
 each set of coils under the brush during the period of 
 strongest action. The brush is broad enough to cover 
 two bars side by side. These bars overlap so that as one 
 set of coils dies down and is cut out, the other set is 
 brought into service. Similar action takes place in the 
 other commutators, so that while the current in each set 
 of coils is periodically broken, the line current is not, and 
 its fluctuations are confined to narrow limits, though it 
 is by no means as steady as in the ordinary closed qoil 
 armature of many section's. 
 
 We next turn our attention to the automatic regulator 
 shown in Fig. 98. The principle of the regulator has 
 already been given, viz.: that of placing a variable shunt 
 across the fields, thereby controlling their magnetization. 
 The duty of the regulator is twofold: First, it must keep 
 the current constant by means of the shunt; second, it 
 must keep the brushes in such a position as to give a 
 spark of proper length to take up the self-induction of 
 each coil as it is cut out of circuit, and must not let the 
 spark get long enough to cause flashing. The regulator 
 connections are shown in Fig. 100. The box bar sweeps 
 across a set of contacts, throwing more or less resistance 
 in shunt with the field magnet circuit, and at the same 
 time rocking the brushes so that the spark is kept con- 
 stant; a small belt driven by the armature shaft runs the 
 worm P t which in turn operates the worm wheel Q. 
 Attached to Q is shaft ^?, on which are mounted clutches 
 S and S'. 7"and T' are the armatures for the respective 
 clutches, and to these are rigidly connected two small 
 
THE SERIES MACHINE. 
 
 26 5 
 
 I 5 
 
266 TESTING OF DYNAMOS AND MOTORS. 
 
 bevel gears, X and X', both of which engage the small 
 bevel gear W. W\s fast to shaft K, on which is mounted 
 the spur pinion Z, engaging sector JV. The contact arm, 
 or box bar A, is also attached to shaft Y. When the cur- 
 rent increases above its normal value, a certain amount 
 passes through clutch S, which immediately seizes its 
 armature X, and the contact arm sweeps over the con- 
 tacts, cutting out resistance in the shunt, weakening the 
 
 field, and restoring the current 
 to its normal value. At the 
 same time the brushes are 
 shifted backward. In case 
 the current falls below its 
 proper value, clutch S' acts, 
 the shunting power is decreased, the field thereby strength- 
 ened, and the brushes moved forward to meet the condi- 
 tion. This regulator is then a mechanical device kept in 
 continuous motion and electrically controlled by means of 
 what is called the "wall controller." 
 
 The controller, Fig. 98, consists of two relays (electro- 
 magnets A' and Y) and a small German silver resistance. 
 The main current enters the controller at the right-hand 
 top binding post, passes around the two relays and 
 through the German silver resistance, and so on out at 
 the upper left-hand binding post. At the point A the 
 circuit is connected to the yokes of the relays, and in 
 metallic connection with this are two silver contacts, B 
 and C\ opposite these are two platinum contacts, B' and 
 C ', which connect to posts JF and E. In case the current 
 rises above its normal value, relay X raises its armature 
 and makes B touch B'. If the current falls too low, 
 relay K drops its armature, letting C touch C', and cur- 
 
THE SERIES MACHINE. 
 
 267 
 
 rent passes down through E. In order to reduce the 
 sparking at the make and break contacts, a small spool of 
 
 FIG. 100. 
 
 German silver wire is connected across the breaking 
 point, to take up the secondary current due to the self- 
 induction in the clutch. The principle of its action can 
 
268 TESTING OF DYNAMOS AND MOTORS. 
 
 be seen by aid of Fig. 99, which shows the connection 
 on the field of a street railway generator. We have seen 
 that when we break a circuit containing an electro- 
 magnet, the induced E. M. F. is high and causes a long 
 spark, besides straining the insulation. In Fig. 99 F is 
 the generator field, L some lamps whose number had best 
 be determined experimentally, K is a switch, and, when 
 the dynamos are working, occupies position i. If it is 
 desired to break the generator field, K is thrown over to 
 position 2, which includes the lamps, whose resistance is 
 so high that the machine cannot support its field through 
 them. In the regulator case before us, the lamps are re- 
 placed by the high resistance German silver wire, which 
 instead of being placed in with a switch is in all the time, 
 and when B B 1 or CC' opens, the induced current is 
 given the alternative of an air gap or this German silver 
 resistance. The method of connecting the wall control- 
 ler and automatic regulator to the machine is shown in 
 Fig. 94. Facing the commutator the left-hand binding 
 post, as indicated, must always be positive, and the wire 
 from this post should run to the line direct. On return- 
 ing to the station the wire passes through the ammeter 
 and thence to the right-hand post A ' of the wall con- 
 troller, through the controller out at B\ and to the right- 
 hand upper post B on the dynamo; from here it passes 
 through the field and back to the lower post on the same 
 side of the dynamo, and from here on through the arma- 
 ture. The lower binding posts, F' and ', on the wall 
 controller, connect respectively to F and E on the 
 clutches S and S'. From the clutches the current passes 
 to the centre ring of the governor. In Fig. 100 we see 
 that this ring connects to the small binding post./)', 
 
THE SERIES MACHINE. 269 
 
 which in turn connects to the small upper binding post 
 D on the right-hand side of the machine. The connec- 
 tion between binding post E on S' is of use only when 
 the machine is overloaded. For example, suppose the 
 current to fall below 9.6 amperes, the right-hand relay of 
 the wall controller acts and sends current through the 
 right-hand clutch S, 1 which tends to rotate its own arma- 
 ture; all resistance in the shunt rheostat is cut in and 
 the brushes are back as far as they can go. To relieve 
 clutch S'of its current, thereby taking off P, a strain 
 which would soon wear it out, the clutch S', in this 
 extreme position of the contact arm A, is short circuited. 
 In all other positions of A the resistance box is in shunt 
 with the field through binding post C', which connects 
 to lower post Con the machine, while D' connects to D, 
 thereby throwing the resistance and field in shunt. 
 
 Before starting up a machine see that its rocker arm 
 moves freely; a little thin oil can be used here to advan- 
 tage. Care should be taken to have no cramping where 
 the teeth of the small spur pinion Z engage the teeth of 
 the sector JV (see Fig. 100). A small amount of oil 
 should be placed in the oil chamber surrounding the 
 worm to insure its lubrication. It is well to test the re- 
 lays in the wall controller before throwing on the load. 
 To do this, rock the brushes well forward and see that 
 the contact arm A is on the extreme right. Connect the 
 main binding post A (Fig. 94) of the dynamo to post A^ 
 of the ammeter; disconnect E' and F' on the wall con- 
 troller, then start the machine and adjust the right-hand 
 relay so that it will drop its armature when the current 
 falls below 9.4 amperes, and the left-hand relay so it 
 will raise its armature when the current exceeds 9.8 
 
270 TESTING OF DYNAMOS AND MOTORS. 
 
 amperes. Reconnect E and F, close the field armature 
 switch, rock the brushes back and the machine is ready 
 for load. The spark must then be adjusted to proper 
 length by loosening clamp L (Fig. 100), and rocking the 
 brushes backward or forward as may be necessary; then 
 reset the clamp and the regulation will be automatic for 
 all load variations. It may be necessary to make a final 
 adjustment of the controller relays after the load is 
 actually on the machine. 
 
 In this machine as in all others, failure to operate 
 satisfactorily is frequently due not to any fault in the 
 machine itself, but to inefficient knowledge of details 
 as to connections, switches, commutators, regulators, 
 brushes, and bearings. Some practical points will, then, 
 be in place. To aid in distinguishing between the inside 
 and outside field spool terminals, the inside ones are 
 painted red, and are also tagged before leaving the shop. 
 The field switch should be mounted on the same side of 
 the machine as the rocker arm. In handling the brushes, 
 always close the small switch located just above the com- 
 mutator, short circuiting the armature, otherwise the 
 attendant is liable to get a severe shock, even though 
 the field switch be closed. To "kill" the machine in- 
 stantaneously, as in case of accident, close the armature 
 switch. Under ordinary circumstances it is best to close 
 the field switch first. In starting up it is immaterial 
 which is opened first. 
 
 To set the brushes a sheet metal gauge is used. This 
 gauge is shipped with the machine. Sometimes after the 
 brushes are uniformly adjusted it will be found that the 
 spark varies on the different commutators. This is 
 owing to lack of uniformity in cross-section of different 
 
THE SERIES MACHINE. 271 
 
 parts of the armature and to local variations in its 
 magnetic quality. This is unavoidable, and in such cases 
 the brush should be lengthened or shortened from 1/16" 
 to 1/32" to suit the circumstances. Under ordinary cir- 
 cumstances i/8" is the proper length of spark, but where 
 large fluctuations of load are likely to occur, as for 
 example where hotels, concert halls, skating rinks, etc., 
 are in circuit, it is well to carry a 3/16" or 1/4" spark, 
 because where variations are wide and sudden a great 
 many lines of force are put in motion and the self-induc- 
 tion is heavy. All series machines, and especially those 
 which depend to any degree on armature reaction for 
 regulation, give lower and lower efficiencies as the load is 
 decreased. This is so on the Brush machine, because the 
 fewer the lamps in circuit the less is the voltage needed, 
 and hence the more the current which must pass through 
 the shunt and be wasted. On the Thomson-Houston 
 machine running light, the coils are short circuited for so 
 long a time during each revolution, that they may be in- 
 active, as far as concerns the external load, as long or 
 longer than they are active. The local current in the 
 short circuited coils gives rise to heat, and impose 
 additional drag on the engine without making any return 
 for it. On either of these arc machines the iron losses 
 are greater under light load, because the fluctuation 
 and hence movement of lines of force is greater. If the 
 machine supplies, say i lamp, whose resistance equals 
 that of the machine, and if we suppose the resistance of 
 the lamp when burning to be 5 ohms, the resistance of the 
 line will be 10 ohms; and if for any reason the i lamp cuts 
 itself out, the line resistance suffers a variation of 50^, 
 whereas if there are 9 lamps in circuit, i lamp is but i/io 
 
272 TESTING OF DYNAMOS AND MOTORS. 
 
 of the line resistance, and its being cut out makes a dif- 
 ference of but 10$. On those machines whose regulation 
 depends on armature reaction (such as the Edison and 
 Westinghouse direct current arc machines), it is inefficient 
 to run on light loads, because then the armature reaction 
 is greatest, and the field lines of force are created only 
 to be neutralized by those of the armature, whereas 
 at full load the armature has little positive influence, 
 and all the lines of force generated are used. It is 
 safe to say, then, that no arc machine should run on 
 light load any longer than can be helped. To revert to 
 the Brush machine in particular: its commutator should 
 have the best care, for upon it satisfactory working 
 largely depends. It should be kept perfectly clean and 
 only enough oil used to prevent cutting; this should be 
 mineral oil, and should be applied with a piece of felt or 
 leather. To insure safety to the attendant it is abso- 
 lutely necessary that the felt or leather be applied from 
 a dry stick, piece of fibre, or other insulator at least twelve 
 inches long. Machines whose E. M. F. range from 3,000 
 to 6,000 volts are dangerous traps, unless respected, and 
 any mistake is apt to be a fatal one. The commutator 
 bars are built upon hard wood blocks (generally lignum 
 vitse), and it is sometimes necessary to renew them; as 
 even well seasoned wood is apt to be a little damp, new 
 blocks may occasion more or less trouble at first. To 
 prevent flashing, the spark can be lengthened slightly for 
 a few hours, till the wood is dried out. 
 
 The troubles to which the machine is liable are in 
 general the same as those already spoken of, and the 
 same general precautions and methods of testing are to 
 be followed. The first sign of a defective armature 
 
THE SERIES MACHINE. 273 
 
 bobbin is seen on the commutator; the spark will grow 
 short and may disappear at one of the commutator sec- 
 tions, and there may be some flashing. If the machine is 
 badly needed, the brush can be cut a little back and the 
 machine may run a long time if not overloaded. Upon 
 shutting down, the short circuited or defective coil can 
 be located by the fact that it will be overheated. As 
 there is considerable space between bobbins, an accumu- 
 lation of dust sometimes starts a ground between bobbin 
 and core. To prevent this a stiff brush and bellows 
 should be used every day or so. Dust and oil must not 
 be allowed to gather on the armature leads where they 
 enter the shaft and where they come out on the commu- 
 tator end* oil is doubly harmful because it not only holds 
 the dust, but rots the insulation. As a rule the commu- 
 tator segments wear out first, and at the tips, where the 
 arc holds. Their life can be prolonged by taking them off 
 and bending up the tips. The tip can then be filed down 
 to a circle and replaced ; next cut out a block of wood to 
 fit the commutator; cover the inside of the block with 
 sand-paper (not emery), and press it against the com- 
 mutator while the armature turns at about half speed. 
 In order to line all segments up to the circumference of 
 the same circle it may be necessary to slightly raise a 
 segment; to do this slack up the screws and pack under 
 the segment with sheet fibre or other insulator, and then 
 proceed with the sand-papering. During this operation 
 the leads should be carefully wrapped to prevent copper- 
 dust from entering the shaft hole. The commutator 
 should be then carefully wiped off. The Brush regulating 
 apparatus is perhaps not so compact as that of the Thom- 
 son-Houston regulator, nor is it as purely an electrical 
 
274 TESTING OF DYNAMOS AND MOTORS. 
 
 device; nevertheless, with due attention it gives excellent 
 satisfaction. The dial, if kept in good order, will for the 
 little extra attention, make ample returns, but if neglected, 
 will require the assistance of the attendant to regulate the 
 lamps, which will either flame or hiss. It is a good idea 
 to mount all the station dials on a piece of wood about 
 i 1/2 X 3 inches square, and long enough to run the whole 
 length of the dial. The upper strip should be hinged 
 to the wall frame or switchboard. This facilitates 
 removal of carbons for inspection and cleaning. In 
 replacing it should be seen to that the carbon piles work 
 freely between the dividing slates, and that they make 
 good contact across from one pile to the other. 
 
 There are in use other arc machines and regulators, 
 which are gradually gaining a foothold, and which might 
 well be described here, but they are only adaptations and 
 modifications of principles above dwelt upon, and any- 
 one who can handle and understand the Thomson-Hous- 
 ton and Brush machines will be at a loss with no other. 
 
 We therefore turn to the last type of arc machine to 
 be here considered. This is the new Westinghouse 
 direct current generator. Though the machine differs 
 very radically in construction from former types, it 
 belongs to the open-coil continuous current class, which 
 has special advantages for arc work. It delivers a cur- 
 rent which is pulsating, but always in the same direction, 
 /. <?., the current is not constant in value, varying 
 between narrow limits, and the brushes remain always of 
 the same polarity. It is upon the fact that the current 
 pulsates that the machine's automatic regulative ability 
 depends. The fields are separately excited at no volts, 
 thus avoiding the introduction of high voltage into the 
 
THE SERIES MACHINE. 275 
 
 field circuit, and securing a field current of constant 
 strength. The main current has a steady mean value, 
 and the pulsations keep the mechanisms of the lamps in 
 a constant state of mild vibration, thereby making them 
 less liable to become slow in action or to stick. In gen- 
 eral appearance the machine resembles the Westinghouse 
 alternator: consisting of a circular cast iron yoke parted 
 horizontally and having inwardly projecting poles. The 
 armature is of the toothed type, having especially heavily 
 insulated lathe wound bobbins. Between the teeth are 
 wedge-shaped iron lugs to be driven in after the bobbins 
 are in place. Each coil of the armature consists of two 
 of these bobbins, which are put on one at a time, and then 
 taped together on the ends. There are eight armature 
 coils and but six field spools and poles; this peculiarity 
 is accounted for by the fact that from the nature of the 
 commutator two coils are always idle, leaving an active 
 coil to each pole. The strong point claimed for the 
 machine is its method of automatic regulation without 
 regulating devices of any kind. For regulation it depends 
 solely upon the armature reaction about which we have 
 spoken above; the principles here involved can be better 
 understood if we look for a moment at the reactive 
 effect of an alternator armature. In the field winding of 
 an alternator the direction of the current does not 
 change, but that of the armature flows first one way and 
 then the other, alternating as each pole piece is passed. 
 It can be seen then that an armature conductor opposite 
 a pole piece may assist the field coil and strengthen its 
 field, or may oppose the coil and exert a weakening effect, 
 according as the conductor's current concurs with, or 
 opposes the magnetizing effect of the current in the coil. 
 
276 TESTING OF DYNAMOS AND MOTORS. 
 
 Whether there is opposition or concurrence depends 
 upon the direction of the current for a given position of 
 the conductor. Assume for simplicity that our con- 
 ductor lays midway between two pole pieces, and that 
 when the armature carries its full load, the conductor 
 current reaches its greatest value in this position. The 
 conductor being midway between two poles exerts 
 counter influences on the rear and forward poles, and its 
 resultant influence is zero. If lamps are in series and 
 several be turned off, the effect is to momentarily increase 
 the current, and also the self-induction of the circuit. 
 Now if the self-induction of a circuit is increased the 
 current does not reach its maximum value as soon after 
 the E. M. F. reaches its maximum value as it otherwise 
 would, so that in this particular case the lagging current 
 no longer reaches its maximum value when the conductor 
 is midway between pole pieces, but does so beyond this 
 point, and being then nearer the approaching pole is in a 
 position to exert a stronger demagnetizing influence on 
 this pole. The result is to reduce the E. M. F. and 
 restore the current to its proper value. When the 
 machine is short circuited, the self-induction and result- 
 ing lag become so great that the conductor has its maxi- 
 mum demagnetizing ability when it is immediately in 
 front of a pole piece. As seen above a pulsating current 
 partakes of the nature of an alternating current in that it 
 fluctuates between limits, but differs from the latter in 
 that there is no change of sign. The current in each 
 coil starts from zero, goes to a plus maximum and 
 returns to zero, to repeat the same cycle. The amount 
 of reaction between any armature conductor and a pole 
 piece depends upon the conductor's position, its current 
 
THE SERIES MACHINE. 
 
 277 
 
 strength, and in part upon whether this current is rising 
 or falling in value. At full load the current in any con- 
 ductor may be at its least value when the conductor is in 
 the best position for demagnetizing the field; and is at 
 
 TO 
 
 LAMPS 
 
 FIG. loi. 
 
 its highest value when best disposed to reinforce the 
 field. On short circuit the armature current reaches its 
 maximum value when the conductor is opposite a pole 
 piece of similar sign. In these machines, as in alternators, 
 the shape and disposition of the pole pieces have been 
 experimentally determined. The machine in question 
 can be better understood by looking at Fig. 101. On 
 the machine itself the two commutators occupy posi- 
 tions side by side on the shaft, while in the figure the 
 outside commutator represents the one next the arma- 
 ture, the inside small one the one next the end of the 
 
278 TESTING OF DYNAMOS AND MOTORS. 
 
 shaft. Notice that a straight line through the centre of 
 the shaft and the centre of a bar on one commutator, 
 passes through the centre of the insulation on the other, 
 showing the commutators to be on the shaft, one a little 
 ahead of the other. Besides the mica insulation, there 
 are two air gaps between each segment, one on each side 
 of the mica. Points numbered the same are in metallic 
 communication, thus 8, the inside end of a bottom sec- 
 tion, goes to binding post 8, on the end of the shaft, 
 and this connects to three bars marked 8, on the little 
 or outside'-commutator. The field circuit is completed 
 through a device operated by the plunger of a solenoid 
 in the main circuit. If the outside circuit opens, the- 
 plunger falls and opens the field. 
 
CHAPTER IX. 
 
 TESTING OF DYNAMOS AND MOTORS, SHUNT AND COM- 
 POUND MACHINES. 
 
 THE shunt dynamo differs from the series dynamo in 
 that the field circuit is independent of the external or 
 line circuit, and its E. M. F. is highest when the line 
 switch is open. .With the fields connected below the 
 switch, the dynamo generates on open circuit only such 
 current as the existing E. M. F. of the machine can urge 
 through the combined resistance of the field and its 
 rheostat. The amount of current which will pass through 
 any resistance depends upon the potential difference at 
 its terminals, and conversely, the current which any 
 given potential difference will send through a circuit de- 
 pends upon the resistance of the circuit; and, further, 
 we know that any applied E. M. F. distributes itself 
 around a circuit according to the disposition of the re- 
 sistance. Wherever the greatest resistance is found there 
 also is the greatest potential difference per unit of length 
 of circuit; and when a circuit is broken, the resistance of 
 the air gap at the break is so great compared with the 
 resistance anywhere else that the entire potential differ- 
 ence takes place across the break. On the other hand, 
 if between any two points a second wire be placed, as 
 between points A and B in the circuit of Fig. 102, the 
 effect is to decrease the resistance between these points, 
 
 279 
 
280 TESTING OF DYNAMOS AND MOTORS. 
 
 since there are now two paths instead of one, and thereby 
 to decrease their difference of potential. The original 
 path A D B being now subjected to a less potential dif- 
 ference, does not get as much current as it did before 
 being shunted. If, however / by any means (increasing 
 the E. M. F. applied to the circuit) the potential differ- 
 ence between A and B be raised 
 to its initial value, path A D B 
 will take the same current that it 
 had when alone in circuit, and 
 
 D path A c B. a current depending 
 
 FIG. 102. 
 
 upon its resistance. 1 he total 
 
 current has been increased. We may say, then, that 
 the current taken by any branch is independent of 
 that flowing in any other branches, provided the poten- 
 tial difference at the common junction is maintained 
 the same, and the current which each branch takes 
 depends upon its own resistance. If the potential dif- 
 erence between A and B is not regulated, then the 
 effect of introducing an additional path is to divert the 
 current from existing paths, and in a degree depending 
 upon its conductivity. Both conditions can be practically 
 illustrated by means of Fig. 103, where A is the armature, 
 /"the field, and A" and Z, respectively, the line switch and 
 lamp load of a shunt dynamo. The smaller lettering is 
 identical with that of Fig. 102. The armature A applies 
 its E. M. F. at a and b, the common terminals of field 
 and line. When K is open A and F are in series, and all 
 the current A generates goes through F. The amount of 
 this current depends upon A's E. M. F. and upon jF's re- 
 sistance. A's E. M. F. depends upon its speed, its num- 
 ber of conductors, and the number of lines of force they 
 
SHUNT AND COMPOUND MACHINES. 281 
 
 cut. The number of lines of force depends in turn upon 
 the field current. With A and F alone in circuit, A's re- 
 sistance being very low, the total potential drop takes 
 place through F's high-resistance windings, and can be 
 measured between a and b. Potential measured under 
 this condition is called the " open circuit " potential of 
 the dynamo. When K is closed, resistance between a 
 and b becomes less, and the armature current increases, so 
 that there is a redistribution of potential. We know that 
 the drop between a and b is the machine's useful E. M. F., 
 and that the drop through the armature resistance is a 
 dead loss, and equals the product of armature resistance 
 and current, or = 7/' a , where /is A's current in amperes, 
 and / its resistance in ohms. The final effect then of 
 closing K is to decrease the circuit resistance and in- 
 crease the current. The increased current causes a 
 greater drop to take place through the armature, and 
 hence decreases the E. M. F. available at the line termi- 
 nals; but since the field winding is energized by the 
 potential difference between these points, when this 
 grows less, the field gets less current, the magnetism 
 decreases, and the voltage does too. To determine 
 what part of the voltage decrease is due to increased 
 armature drop, and what part is due to decreasing the field 
 current, proceed as follows : Get the dynamo up to 
 voltage, throw on a certain number of lamps, and observe 
 the loss of potential between a and b. This loss will be 
 due to the combined effect of the above two causes; next 
 separately excite the field, bring the armature up to 
 voltage, and again throw on the lamps.' The loss is now 
 due entirely to increase of armature drop, consequent 
 upon increasing the current. To be strictly correct, the 
 
282 TESTING OF DYNAMOS AND MOTORS. 
 
 armature should carry the same current in both cases. In 
 Fig. 103 the armature is the undivided part of the circuit, 
 and the branches F and L the divided part. If we call / 
 the current in the undivided part, and i\ and i\ that in the 
 branches, then/ a -f- / a = /, and t\ : / 2 ; ; ;- a : r^ where r a and 
 fj are the respective resistances of branches. If we call 
 
 r & the armature resist- 
 ance, then, since from 
 Ohm's law the drop is in 
 any case the product of 
 current by resistance, 
 we have the armature's 
 " lost volts " = 7 r a , and 
 in designing armatures 
 allowance must be made 
 for this loss, which must be calculated when the armature 
 is hot, not cold, for as more lamps are thrown across the 
 mains, the external resistance is further decreased, a 
 larger flow of current takes place through the armature, 
 giving rise to heat, and robbing the fields of more mag- 
 netizing power. A shunt dynamo, with an armature of 
 zero resistance, would be self-regulating for constant 
 potential at the brushes, for without resistance there 
 could be no loss of potential in the armature, and hence 
 the total E. M. F. would always be the same as that at 
 the brushes. This ideal state is approached by mak- 
 ing the armature resistance as low as possible. A shunt 
 dynamo is theoretically adapted to automatically regulate 
 for constant potential at the brushes, when the lamps or 
 other devices are in series. As the number of lamps 
 increases so does the circuit resistance, and with it the 
 potential at the field circuit terminals, and hence the ma- 
 
SHUNT AND COMPOUND MACHINES. 283 
 
 chine's ability to supply an increased voltage; but where 
 devices are in series they all take the same current, so 
 that the line voltage must be very high to supply any 
 considerable amount of energy. A shunt winding which, 
 when subjected to very high voltage, would give the nec- 
 essary magnetizing power, and at the same time be of 
 sufficiently high resistance to shunt an economical frac- 
 tion of the armature current, would be very expensive to 
 make. The cost of insulated wire per pound increases 
 very rapidly as the diameter decreases. A pound of No. 
 5, double cotton-covered B. & $., copper wire sells for 
 14 1/2 cents a pound, of No. 10 at 17 1/2 cents, No. 20 
 at 28 cents, etc. As the wire grows smaller its insulation 
 takes up a relatively larger and larger part of the winding 
 space, till when we reach a No. 13 B. & S. the copper and 
 cotton occupy about the same space. This means that 
 in order to get on enough copper the machine must be 
 made larger than its output would call for, and the maker 
 would lose the interest on money uselessly expended. 
 The larger a wire, the smaller is the relative amount of 
 space taken up by insulation. On a series machine 
 the whole armature current flows through the field wind- 
 ing, so that fora given magnetization there need not be as 
 many turns of wire. Accordingly, under the dictates of 
 economy, series machines have been adopted in high 
 voltage constant current, or series working, separate ex- 
 citation in other high-voltage work, and shunt machines, 
 for the most part, in constant-potential multiple work. 
 Separate excitation is the most flexible method, in that 
 it affords a ready means of securing any combination of 
 amperes and turns to make up a given number of ampere 
 turns. For a given number of ampere turns (usually written 
 
284 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 Si} the same amount of work, or watts, is expended in 
 maintaining a field, whether the winding is coarse or fine. 
 Suppose, for instance, that around a spool of 10 turns, 
 measuring i ohm, there flows a current of 10 amperes. 
 If we halve the cross-section of the wire we can get on 
 twice as many turns, and it will require but 5 amperes to 
 give the same Si as before. But in halving the cross- 
 section we double the resistance of every turn of wire, 
 and since we have doubled the number of turns we double 
 again the resistance of the wire on the spool. In other 
 words, we have not only doubled the length, but also 
 halved the cross-section. Tabulating this we get: 
 
 
 Turns. 
 
 Amps. 
 
 Amp. -Turns. 
 
 Res.-Ohms. 
 
 Watts Lost. 
 
 First Case 
 
 10 
 
 10 
 
 IOO 
 
 i 
 
 (/' 2 J?) IOO 
 
 Second Case. . . . 
 
 20 
 
 5 
 
 100 
 
 
 (/ A } ) = ioo 
 
 Now while it makes no difference so far as the magneti- 
 zation is concerned whether the Si are made up of many 
 turns and small current, or vice versa, it does make a dif- 
 ference electrically where a machine is to furnish its own 
 exciting current. _ If in a shunt dynamo the field winding 
 is of the right number of turns but of too low resistance, 
 it will take more current than it ought to, but will not 
 raise the machine's voltage, and hence the output, pro- 
 portionally. Properly designed shunt windings are of 
 such resistance as to let in that current which will mag- 
 netize the iron to best saturation, and any current above 
 this over-saturates the fields, reduces the number of lines 
 of force per ampere, and brings down the efficiency of 
 
SHUNT AND COMPOUND MACHINES. 285 
 
 the machine. Sir William Thomson (Lord Kelvin) has 
 pointed out that no shunt dynamo can run at an efficiency 
 of 90$ unless its field resistance is at least 324 times that 
 of the armature. Without repeating the mathematical 
 work presented by both Sir William Thomson and S. P. 
 Thompson, we will pass directly to the more practical 
 deductions which they have developed as regards the re- 
 sistance of different parts of the circuit, and its influence 
 upon the efficiency. First, for all practical work the effi- 
 ciency (//) is given nearly enough by the formula, 
 
 i 
 
 Second. The most economical value of the external re- 
 sistance to work with is given by the formula A' = 
 V'" s '"a where r a and r & are respectively the field and 
 armature resistances. Suppose we wish a machine to 
 have an efficiency of 90$, what must be the relative re- 
 sistances of armature and field? Substituting in the 
 above formula we get 
 
 9 " 
 
 
 V 7 '-. Vr. 
 
 &. 
 
 Whence clearing of fractions, 
 
 .9 Vr s + i.Sy^ = ^ and 1/7 8 - . 9 V^ 8 = 1.8 W7 
 or .1 Vr s = 1.8 Vr & . Multiplying through by 10 to get 
 rid of the decimals we have ^ - 18 y^ a 
 
 ' >'s = (iS)Va = 324 r a . 
 
286 TESTING OF DYNAMOS AND MOTORS. 
 
 For any other proposed efficiency it is only necessary to 
 substitute and go through a similar operation to get the 
 relation which r s must bear to r & . As a practical ex- 
 ample, let 1\ .206 : what value of r s will secure a 
 maximum efficiency of 95^, and of 90$ respectively? 
 The equation is 
 
 _ i 
 
 ~ 
 which gives 
 
 .-= 
 
 as shown above in the determination of the best relation 
 of r a and r s for ;; = 90^. Clearing this equation of frac- 
 tions we have : t? Vr s -j- 27; yV a = \fr s ; transposing, 
 Vr s 11 \/r a = 2?7\/ r a! factoring on Vr s , Vr a (i ^/) 
 27 7 |/ ; 'a Dividing both sides by i 77 , 
 
 & = -^. 
 
 Kr s 177 
 
 and substituting in this formula for rj and r a , their re- 
 spective values .95 and .206 ohm, we get 
 
 2 X -95 V- 206 1-90 X 4/^06 
 
 or Vr s 38 t/ 206, and r s (squaring both sides) = 1,444 
 X .206 = 297 ohms. For/; = 90^ = .9, r s = 324 X .206 
 = 66.74 ohms. As a practical example of the second 
 case, let r & = 034 ohms, and r s = 13.6 ohms, what is 
 the best value for R ? We have R = V~r*r* 
 = V 13.6 x .034 = .68 ohm. From a consideration of 
 
SHUNT AND COMPOUND MACHINES. 287 
 
 the two problems in Case i, we see that to secure a 5$ 
 increase in efficiency the magnet resistance must be 
 raised from 66 to 297 ohms, and this means a great in- 
 crease in cost of production. Practice compromises be- 
 tween first cost and efficiency. Let us suppose the 
 resistance of a given set of magnets to be 200 ohms. 
 What must r & be to give an efficiency of oo^? Of 95$? 
 Taking the equation A//' S (i ty) = 2// f / tt , and dividing 
 off by 2;;, we have, 
 
 V^. (' -- v) 
 ~ST 
 
 Substitute for /; and r a their respective values, 
 
 <\/200 (i -- .QO) <\/200 X.I I 
 
 Vr & ~- -j-jp - 8 - --V*oox^ 
 
 Whence 
 
 -(*) 
 
 for 90$ efficiency, and for 95$ 
 
 3 
 
 ' ''KXD) = 200 X - - = .6172 ohm, 
 3 2 4 
 
 200 
 
 = . iios ohm. 
 ,444 
 
 Here to secure 5$ increase in efficiency, r a must be low- 
 red to less than one-fourth of its value for 90$ effi- 
 ciency. It is not easy to vary the resistance of an 
 armature intended to do a certain amount of work, 
 unless it so happens that several points in bad design 
 may be changed to right matters. If there is no such 
 outlet the machine must be either redesigned or rated 
 lower. Field resistance should be made as high, and 
 armature resistance as low, as possible. 
 
288 TESTING OF DYNAMOS AND MOTORS. 
 
 As shunt dynamos are not perfectly self-regulating for 
 constant potential, some method of regulation must be 
 used; the universal method is to place a variable resist- 
 ance (rheostat) in the field circuit so that its resistance, 
 and hence field current, can be .varied to meet the de- 
 mand. As we have seen in a shunt machine with closed 
 switch, the current leaving the armature takes two paths; 
 the field winding and external circuit, and the current 
 through each depends upon their relative resistances un- 
 less the potential difference is kept constant, in which 
 case the current in each depends upon its individual re- 
 sistance. If at any given load the mains are found to be 
 at the proper potential difference, it means that the field 
 circuit resistance lets in just field current enough to 
 maintain the proper field strength. If R is increased or 
 diminished, by diminishing or increasing the load, the 
 rheostat must be turned to restore the potential to 
 its former value. As the load increases, or what is the 
 same thing, as R decreases the potential difference at the 
 brushes decreases, and the rheostat resistance must be 
 reduced until the increased field current provides volt- 
 age enough to look after the increased drop in the arma- 
 ture. If the external resistance becomes too low, as in 
 the case of a short circuit on the line, the machine will 
 lose its field, if the suddenly precipitated load does not 
 open a circuit breaker or throw the belt first, and will 
 refuse to "pick up" a field until the short circuit is 
 removed. This is because when a short circuit occurs 
 on the line, the low resistance of the short circuit is in 
 multiple with the high resistance of the field circuit, and 
 prevents the latter from getting any current: or more 
 properly speaking the armature "drop" becomes so 
 
SHUNT AND COMPOUND MACHINES. 289 
 
 great that there is not left across the field terminals 
 potential difference enough to support the field. In such 
 a case the line switch must be opened, and the trouble 
 sought on the machine itself. We specify line switch be- 
 cause in stations where many dynamos run in multiple it 
 is customary to keep the head board switch closed, and 
 to do all testing across the switch on the station switch- 
 board. Where several shunt machines feed into the same 
 " bus " bars, a short circuit on the outside is apt to make 
 them all lose their fields, but in some cases the fault 
 burns out as soon as it is made, while in others special 
 preparations must be made for burning it out. This is 
 treated of elsewhere, as is also the special case of short 
 circuit in the machine itself as a consequence of a broken 
 field wire. 
 
 If a shunt machine has most of its load suddenly 
 removed, by cutting out parts of the service, the volt- 
 age will rise and endanger the lives of the lamps still 
 in circuit. This is because at the larger load the arma- 
 ture drop is the greater, and the field current is adjusted 
 to look after it. When the load is removed the arma- 
 ture drop is much less, leaving a greater potential differ- 
 ence for the external circuit whose resistance has been 
 increased. The immediate effect of raising the voltage 
 at the field terminals is to send a larger current through 
 the winding, thereby again sending the voltage upward. 
 This action is most marked where there is but a single 
 dynamo in service, the reason being that where several 
 dynamos are in multiple, the decrease in load is divided 
 among them all, and the amount per machine is not so 
 large. The remedy is to cut resistance into the field 
 rheostat. This liability to injure lamps is one reason. 
 
290 TESTING OF DYNAMOS AND MOTORS. 
 
 aside from the desire to keep lamps at constant brilliancy, 
 that the attendant watches so closely at that time in the 
 evening when a great many lamps are being turned out. 
 Turning off all the lamps in a public building will very 
 perceptibly affect the voltage of a single dynamo. An 
 attendant soon comes to know the unsteady times of the 
 day and keeps a hand near the rheostat. Where the ser- 
 vice is such that load variations are not sudden or great, 
 an attendant is equal to the occasion, otherwise the shunt 
 machines had better give way to compound-wound 
 machines, whose rheostats are set once for all and re- 
 quire no further attention. 
 
 The extent to which a given change in the rheostat 
 resistance will affect the dynamo's voltage depends upon 
 the relative resistance of the field winding and rheostat, 
 and this relation in turn depends upon their temperatures. 
 It must be borne in mind that the two are usually com- 
 posed not only of different substances, but have very 
 different facilities for radiating heat, so that a given cur- 
 rent will raise their resistances at quite different rates. 
 In winter when the surrounding air may be at 15 C. to 
 20 C., it might be necessary to use all of the rheostat to 
 hold the voltage at its proper value, while in summer 
 with the temperature of the air surrounding field and box 
 at 70 C. or 80 C., the rheostat must be almost cut out, 
 the increased resistance of the field sufficing. For the 
 same reason a rheostat adjustment when the field is cold 
 is of no value when the field * a r loss has raised the field 
 temperature, and with it its resistance. Further, the 
 armature heats under load, and its increased resistance 
 causes a greater internal "drop" through it; and finally, 
 after heating up, the iron or steel parts of the machine 
 
SHUNT AND COMPOUND MACHINES. 29! 
 
 do not carry lines of force so well, or as we say, the 
 reluctance of the magnetic circuit has increased, so that 
 it takes more magnetizing force to produce a given 
 amount of magnetization. All of these causes call for 
 more ampere-turns on the field, to keep up the voltage, 
 and since we cannot in most cases increase the turns, we 
 must use a rheostat to increase the. amperes, and this 
 device is necessary aside from its use in regulating varia- 
 tions due to changes in load. True, minor variations 
 can be cared for by rocking the brushes, forward to raise 
 the voltage and backward to lower it, but since there is 
 always a best position for the brushes, other positions are 
 not so good, and for several reasons the practice is con- 
 demned. One reason is the effect upon the efficiency. 
 The brushes, although apparently not sparking, really do, 
 and hence heat more. One other reason, to be dealt with 
 later, is that the regulation depends largely upon arma- 
 ture reaction, and this means that a certain amount of 
 the field's magnetism, which costs money, is being neu- 
 tralized. 
 
 One must not be deluded into the idea that a rheostat 
 can of itself raise a dynamo's voltage, for when the 
 rheostat is short circuited it might just as well be out of 
 circuit. This is a very delusive error, wanting thought, 
 and when sustained by circumstances is misleading. 
 The writers recall one instance where two young men 
 were using a machine running from a line of heavily 
 loaded shafting, turning below speed because the steam 
 pressure was low. They were imbued with the idea that 
 the voltage was generally brought up by cutting the 
 rheostat out, so not having one in they decided to put 
 one in so they could cut it out. Acting upon the resolu- 
 
292 TESTING OF DYNAMOS AND MOTORS. 
 
 tion they, put one in. In the meanwhile the speed of the 
 shaft had risen 10 revolutions, raising that of the dynamo 
 about 25 revolutions. Of course the voltage was raised 
 and they were satisfied that the rheostat was the respon- 
 sible party. 
 
 In changing the field rheostat resistance on a dynamo 
 two effects obtain, and these are best studied by 
 considering a separately excited machine, for in this 
 case the two can be separated. In Fig. 104, A is an 
 
 FIG. 104. 
 
 armature, and F a field, separately excited from the 
 battery or exciter B. R is a variable resistance rheostat, 
 in series with F, but in no way connected with A. This 
 arrangement typifies a separately excited machine. Now 
 when switch K is open no current flows in F, and what 
 voltage can be measured on A is due solely to the residual 
 magnetism, and we will suppose that by some means or 
 other this is gotten rid of, so that until K is closed A 
 generates no voltage. Next close K, and suppose A 
 generates 500 volts. R is cut out, and the current 
 through F\<*> that which B will send through its resistance. 
 Now increase R until its resistance equals that of F: we 
 have halved the field current because we have doubled 
 the field circuit resistance, the voltage of B having 
 
SHUNT AND COMPOUND MACHINES. 293 
 
 remained the same. In halving /, the exciting or mag- 
 netizing power or ampere-turns, Si, have been halved, 
 and' A is found to generate 250 volts its speed assumed 
 to be the same. If the total resistance of F's circuit be 
 again doubled, the voltage will fall to the neighborhood 
 of 125 volts. We cannot say exactly in any case, because 
 the effect which any given increase in field current has, 
 depends upon how nearly saturated the iron frame is, but 
 since this error will occur in both cases it will not affect 
 our comparison. Better still let us suppose the machine 
 has no magnetic metal in it, and the above figures become 
 true, and we can say that A's voltage is strictly pro- 
 portional to F's current and inversely proportional to 
 (F -|- Ry?> resistance, that is, if the field circuit resist- 
 ance is halved, A's voltage is doubled; if it is doubled, 
 A's voltage is halved. Now suppose B to be removed, 
 and the shunt machine connected to excite itself (the 
 machine would never build up a field through a magnetic 
 circuit having no iron or steel in it, but this will not affect 
 our demonstration). Further suppose R to be such that 
 A gives a voltage of 500: now vary R so the field circuit 
 resistance (F-\- R) is doubled. In the separately excited 
 machine varying R did not affect the exciting voltage, 
 because B was independent of all changes in A's voltage, 
 but in this case, as soon as R is increased the voltage on 
 A decreases as before; but now A's voltage excites F so 
 that not only is the field circuit resistance increased, but 
 the exciting E. M. F. is also decreased: both changes con- 
 spire to reduce A's E. M. F., and if R is increased enough 
 the machine will drop its field entirely. We may say then, 
 that qn a shunt machine a given change in the rheostat 
 has a greater effect than on a separately excited machine. 
 
294 TESTING OF DYNAMOS AND MOTORS. 
 
 The presence of the rheostat is sometimes referred to 
 as objectionable, in that it introduces an additional /V 
 loss into the equation for efficiency. While this is in a 
 certain sense true, it can only be avoided by winding the 
 fields with an unheard-of metal whose resistance is the 
 same for all temperatures, and doing away with all arma- 
 ture drop. If the former could be done there would 
 remain the variation of E. M. F. due to variation of load, 
 and could the latter difficulty be overcome there would 
 still be a great difference between the voltages for hot and 
 cold fields. It must be remembered that for a specified 
 voltage at the terminals there must be a specified current 
 in the field circuit, which must therefore be of a specified 
 resistance. The ampere-turns necessary to maintain the 
 proper potential for any given load is independent of 
 temperature as far as concerns the windings of the field, 
 and depends only upon speed (), the number of armature 
 conductors (C), and the goodness of the magnetic circuit. 
 But, as we have seen, the temperature influences the 
 ampere-turns indirectly by raising the armature resist- 
 ance, and increasing its /V loss. Assuming the machine 
 to be running under constant load, so that the rheostat 
 will not have to regulate load variations, and that 
 the potential at the brushes is to be kept constant, 
 there is a certain field circuit resistance which corre- 
 sponds to a particular value of the ampere-turns, and this 
 resistance is the same whatever the temperature, except 
 in so far as the temperature affects the permeability of 
 iron or steel. It matters not, as far as concerns the 
 dynamo's electric efficiency, whether- this resistance is 
 found in the winding alone, as it often is in hot weather, 
 or in winding and rheostat combined, as in cold weather. 
 
SHUNT AND COMPOUND MACHINES. 295 
 
 The 7V loss is a constant, and whether or not it is 
 divided between field and rheostat is immaterial. How- 
 ever, heating in itself is guarded against in well-designed 
 machines, for excessive heating carbonizes the insulation 
 and ultimately gives rise to serious trouble. Since heat 
 is in any case generated at the rate of 7V watts per 
 second, the temperature will rise rapidly, unless there are 
 provided proper facilities for ventilation. This means 
 practically that the winding must have ample surface for 
 radiation. The temperature limit for ordinary machines 
 may be set at 50 C. above that of the surrounding air, 
 but on inclosed motors of the street railway type, the 
 temperature rises above this, and on machines of special 
 design, as in the Marvin rock drill, where the magnets 
 are mica insulated, the temperature may reach much 
 higher. 
 
 To secure the 50 C. limit there should be 2 1/2 square 
 inches of radiating surface for every watt of energy 
 wasted. Suppose, for instance, a magnet spool of i 
 ohm resistance takes 30 amperes under full load. Watts 
 wasted 7V I x I X /' 30 X 30 x i 900 watts; 
 900 x 2 1/2 square inches = 2,250 square inches of surface 
 required to be exposed to the air. In the Brush machine, 
 2 square inches are allowed the fields, and 9 square inches 
 per watt waste in the armature; in the T.-H. armature 
 i 2/3 square inch is allowed. These figures illustrate 
 the fact that the rule is seldom adhered to; in other 
 words, machines are in most cases allowed to run at a 
 temperature exceeding 50 above the atmosphere. Note 
 that less radiating surface is allowed on armatures than 
 on fields; this is because, as a rule, field wires are wound 
 many layers deeper than armature wires, and besides this 
 
296 TESTING OF DYNAMOS AND MOTORS. 
 
 the armature is fanned by the current of air produced 
 by its own motion. In general, the 2 1/2 square inch rule 
 applies to machines run continuously at full load, and as 
 this condition is not ordinarily required to be fulfilled, 
 the rule is not rigidly adhered to. The main reason for 
 the reduction in radiating surface has been the desire to 
 make the machine lighter and smaller, and it is brisk 
 competition that has forced the figure so low. 
 
 To understand how increasing the radiating surface 
 also increases a machine's weight, let us suppose that we 
 have a field winding that is running too hot. To increase 
 the square inch of radiating surface per watt, we must 
 either arrange the layers of wire to lie less deep, but 
 wider, or we must decrease the number of watts. In the 
 first case the pole piece or magnet core must be lengthened, 
 and with it necessarily every part of the frame parallel 
 to the direction of elongation. In the second case (in- 
 directly the same as the first) a larger wire must be used, 
 and in order to get in the same number of turns as before, 
 more room is required. This gives more surface, and we 
 have not reduced the number of watts wasted, but we have 
 reduced the watts per square inch. The same number 
 of turns of a larger size wire would be of too low a 
 resistance, and would let in too great a current, so a larger 
 resistance must be inserted in the rheostat. In cases 
 where a machine is unavoidably overloaded, artificial 
 cooling must be resorted to, otherwise the machine must 
 be shut down at intervals to cool off. The writers recall 
 one instance of where a street railway waterproof motor 
 attached to the machine shop shafting was kept cool by 
 allowing a stream of water to play on it. In another 
 case reversed fans were placed on either side of a dynamo 
 
SHUNT AND COMPOUND MACHINES. 297 
 
 to draw a strong current of air through it. Shunt wind- 
 ings are so designed that at no load the rheostat is called 
 largely .into play, and as the load goes on, the ampere- 
 turns can be increased to suit the demand by taking 
 resistance out of the rheostat. In winter, and with full 
 load, there is still enough resistance in the box to equal 
 the amount by which the fields will need to be raised in 
 summer. 
 
 Field rheostats are generally made of German silver 
 wire, which rises to a higher temperature for a given 
 current than copper, but its resistance does not rise 
 nearly so much per degree rise in temperature. Since 
 its specific resistance is greater, not so much wire is 
 needed to make up a given resistance, and German silver 
 rheostats are therefore lighter and less bulky than those 
 of iron or copper. The rheostat should be placed where 
 ventilation is good, and, if properly made, will heat but 
 little. Rheostats are shipped with each machine. In 
 adapting a box for use as a rheostat, care must be taken 
 that not only the desired resistance is had, but that the 
 wire is of sufficient current carrying capacity, otherwise 
 the heating will be abnormal, and a connection may be 
 melted off, thereby perhaps opening the field circuit on a 
 machine running in multiple with others. We recall one 
 test in which it was necessary to get along with a rheostat 
 of insufficient current capacity; so we put a similar one 
 in series with it, and kept cutting one out at the same 
 rate as we cut the other in, and in this way relieved first 
 one and then the other. Had the boxes been put in 
 multiple their current capacity would have been doubled, 
 but there would not have been enough resistance. The 
 best arrangement would have been two in series and two 
 
298 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 in multiple. This would give the resistance of one box 
 and the current capacity of two in multiple. But only 
 two boxes were to be had. 
 
 Shunt machines can be run either in series or parallel, 
 but are specially adapted to the latter practice. When 
 
 run in series it is usually 
 on a three-wire lighting or 
 power system. Sometimes 
 two low-voltage machines 
 are connected in series, 
 and the combination run in 
 multiple with a machine of 
 twice the voltage of either 
 as when two 250 volt 
 dynamos in series are run 
 in multiple with a 500 
 
 Flo I05 volt machine. Connections 
 
 are shown in Fig. 105. In 
 
 such a case the current flowing through the two in series 
 must not exceed the capacity of the smaller machine. 
 The current in both being the same, each one's load will 
 be proportional to its voltage, and can be regulated by 
 means of the latter. It is customary to connect the 
 two fields in series across the 500 volt "bus "bars. A 
 single rheostat, if of sufficient range, answers for reg- 
 ulation. 
 
 Where there is a single shunt dynamo on a circuit, it is 
 customary to connect the fields on the armature side 
 of the open switch, or, as it is usually said, " below the 
 switch," so that the field may be excited before the switch 
 is closed. Very often in single dynamo private installa- 
 tions the switch is never opened nor the field rheostat 
 
SHUNT AND COMPOUND MACHINES. 299 
 
 changed from one month's end to the other. The load 
 being put on and taken off by starting up and shutting 
 down the engine. When several machines are run in mul- 
 tiple the fields are connected above the switch, or direct 
 to the " bus" bars, thus in a measure separately exciting 
 the machine, since it has a field whether running or not. 
 As soon as the machine contributes its share of work to 
 the line, it may be regarded as exciting itself. Ordina- 
 rily this arrangement insures that the machine's polarity 
 is such that when the switch is closed the machine will 
 be in multiple and not in series with the machines already 
 in service. However, because a machine's field is charged 
 from the line, it does not necessarily follow that the po- 
 larity may not be wrong, for it is possible that the field 
 connections may be wrong. In such a case the machine 
 and the line will be in series and with no resistance in 
 circuit save that of the armatures a situation by all 
 means to be avoided. Promiscuous line charging does 
 not of course insure proper polarity, and the fact should 
 be tested before closing the line switch. The test is as 
 follows: Connect the fields below the switch, raise the 
 brushes (or where there are many sets of brushes, discon- 
 nect an armature cable) and close the switch, thus 
 charging the fields from the line. Now open the switch. 
 The field circuit is broken, but the poles retain a con- 
 siderable a.nount of residual magnetism. Lower the 
 brushes and observe if the dynamo supports its own 
 field; if so, the polarity is correct, because charging from 
 the line insures proper polarity if the connections are 
 correct, and if they are not correct the machine will not 
 support its field. Assuming that the machine generates, 
 the connection which was below the switch can be per- 
 
300 TESTING OF DYNAMOS AND MOTORS. 
 
 manently secured above it. If it refuses to generate it 
 means that the armature E. M. F. is such as to send a 
 reverse current through the fields opposing the residual 
 field due to charging. In other words, the field leads are 
 connected wrong and must be reversed, and the fields 
 recharged. This point is more fully considered in a later 
 chapter. 
 
 The fact that the fields of shunt machines are independ- 
 ent of the external circuit adapts them for running in 
 multiple on constant potential mains. To run dynamos 
 in multiple successfully it is desirable that they be of the 
 same rated E. M. F., fora machine will not always behave 
 well when run at other than its designed voltage, and 
 careless handling of the brushes or rheostat is apt to 
 reverse the machine, running it as a motor with destruc- 
 tive sparking. The extent of sparking depends upon 
 how much load is on the machine at the time of reversal. 
 For a dynamo, on heavy load, the non-sparking point is 
 well forward. A motor under similar conditions has this 
 point well backward, so that in reversing from dynamo to 
 motor the non-sparking point is shifted, and unless the 
 brushes are shifted also, sparking ensues. Aside from- 
 sparking no bad effects attend a shunt machine's reversal 
 of nature, for it continues to run in the same direction as 
 a motor that it did as a generator. The behavior of 
 series and compound-wound machines under like condi- 
 tions has already been briefly considered. We will now 
 consider the question more fully, and explain the facts 
 that (i) a shunt machine rotates the same way as a motor 
 and as a generator; (2) a series machine rotates in oppo- 
 site directions in the two cases; (3) a compound-wound 
 machine's behavior when changed from dynamo to motor 
 
SHUNT AND COMPOUND MACHINES. 301 
 
 depends upon the relative strength of the shunt and 
 series windings. 
 
 When a machine runs as a dynamo the source of cur- 
 rent is within the armature; when run as a motor cur- 
 rent is supplied from without. Whether a machine will 
 run in the same or opposite direction as motor and 
 dynamo is independent of the direction of the current in 
 the external circuit, and depends solely upon the relation 
 existing between the current direction in the armature 
 and that in the field. On a shunt dynamo it matters not 
 which brush the current flows toward, it must eventu- 
 ally divide between the line and the field circuit. The 
 field circuit shunts the line. If on a shunt dynamo we 
 assume that field and armature currents flow in opposite 
 directions as regards a point in space, they will always do 
 so. Reverse the machine's polarity, if you will, by 
 reversing its residual field. Current through arma- 
 ture and field both reverse direction, and therefore are 
 still opposed as regards a point in space. Now regard 
 the machine as a motor: current from without, whatever 
 its direction, must divide between armature and field. 
 The field circuit now shunts the armature, and it is impos- 
 sible to introduce current from outside and have it go 
 through field and armature in opposite directions as re- 
 gards a point in space, the connections being the same. 
 In a shunt machine, reversing from dynamo to motor, 
 the relative direction of armature and field currents is 
 changed, because the armature current reverses, while 
 the field current does not. This is true, for suppose the 
 machine to be running as a motor; the armature has a C. 
 E. M. F., and this counter tries to send through the arma- 
 ture a current opposed to that urged by the impressed or 
 
/ 
 
 302 TESTING OF DYNAMOS AND MOTORS. 
 
 line E. M. F. If the impressed E. M. F. is removed and 
 the armature kept turning by some other means, the C. E. 
 M. F. will assert itself as the E. M. F. of the machine run- 
 ning as a dynamo, and will dictate the direction of the 
 current. This will be more clear by considering Fig. 106. 
 
 B B* ', M, and F are re- 
 spectively the brushes, 
 armature, and field of 
 a shunt motor. A A' 
 and D are the brushes 
 and armature of a 
 p ia Io6 dynamo sending cur- 
 
 rent in the direction 
 
 indicated by arrows. The current arriving at B' splits, 
 part going through F, in the direction indicated by the 
 arrow, and part up through the armature M. The C. E. 
 M. F., if allowed to assert itself, would send a current 
 downward through J/, which at B' would split, part going 
 through F, and in the same direction as before, and part 
 through the external circuit, but in a direction opposite 
 to that in which it flowed when M was a motor. We can 
 now say that if in changing a machine from dynamo to 
 motor without changing the connections, the relative 
 direction of the field and armature currents is changed, 
 the rotation in both cases will be the same. 
 
 In a series machine armature, field, and line are all in 
 series: whatever reversal of current takes place in one, 
 takes place in all, so the relative directions remain 
 always the same. A series machine must then be run 
 the opposite way as generator from what it is as motor, 
 otherwise it will not generate unless either the field or 
 armature leads are reversed. 
 
SHUNT AND COMPOUND MACHINES. 303 
 
 In the compound-wound machine the two tendencies 
 conflict, the series winding tries to reverse the polarity 
 of the field, the shunt winding to keep it the same. The 
 stronger of the two dictates the field's polarity, and with 
 it the direction of rotation. In a compound-wound 
 machine connected to run as a generator, current leaving 
 the armature flows around shunt and series coils in the 
 same direction, thereby strengthening the field as load 
 goes on. If through careless handling the machine 
 become reversed, the direction of current in the series 
 winding is reversed while that in the shunt is not. The 
 result is that if the windings are very nearly balanced, 
 the field is neutralized, and there is a short circuit. This 
 matter has a bearing upon the operation of compound- 
 wound generators in multiple. Ordinarily, the equalizer 
 prevents any troubles of this sort, but they do occur 
 occasionally in spite of everything, and it is well to be 
 forewarned as to what is likely to happen. Machines on 
 which shunt and series windings are too nearly balanced 
 are apt to be unstable and give trouble. In any event, 
 should such a machine ever reverse and the circuit 
 breaker fail to act, the armature would race and tear 
 itself to pieces. We have thus far had most to do with 
 the shunt dynamo as adapted to use with lamps in 
 multiple, and we have learned that the serious problem 
 of compensating armature drop has been solved by the 
 use of a rheostat. Now this armature loss increases 
 very rapidly as the current increases, and hence any step 
 that would tend to decrease the current used for a given 
 horsepower would decrease also the "lost watts." This 
 can be approximated by having the lamps or other load all 
 in series, and raising the voltage to a high value ; if this is 
 
304 TESTING OF DYNAMOS AND MOTORS. 
 
 done by strengthening the field and running the armature 
 at a higher speed, but holding the maximum current at 
 one half its original value, there is a saving effected in 
 two ways: first, since the current is halved there will be 
 but one-half the original drop in the armature at a given 
 load, and since this drop is distributed among so many 
 lamps in series, its effect will not be so noticeable; second, 
 it costs much less to transmit a given amount of energy 
 at high pressure, because a small current occasions 
 less 7 2 R loss. If, however, the increased voltage is 
 secured by putting more wire on the armature and 
 increasing its resistance proportionately, there is effected 
 only the saving in transmission as far as drop is con- 
 cerned, but there is but half the energy wasted in the 
 armature, because if for a given armature the current 
 be halved, the lost energy is but one-fourth as great; 
 now double the resistance and the loss is doubled for 
 the same current, leaving the loss one-half as great. It 
 is not practicable to raise the armature speed much 
 above what it is on ordinary machines of to-day, and 
 the series system is not in general vogue for constant 
 potential lighting. The nearest and most successful ap- 
 proach to it is the "Edison Municipal Incandescent 
 Lighting System," on which there are long lines of lamps 
 in ^series, and several of these lines in multiple. The 
 voltage most commonly used on such a system is 1,200 
 volts. 
 
 Separately excited dynamos differ from self-excited 
 ones in that no attention need be paid to the field 
 connections relative to the armature: even where a 
 particular polarity is desired, a reversing switch in the 
 field circuit answers the purpose. If there are several 
 
SHUNT AND COMPOUND MACHINES. 305 
 
 such machines running in parallel, and one of them 
 reverses, its direction of rotation remains the same, but 
 this is as far as its behavior is analogous to that of the 
 shunt machine. On a separately excited machine the field 
 lias no connection with the armature, so that when the 
 line current is reversed the armature alone is reversed, 
 and not the field. If the machine is first a dynamo and 
 now a motor, the direction of rotation remains unaltered; 
 but if it is a motor first, then under the new conditions 
 the direction of rotation is reversed. A separately 
 excited machine will then run the same way as motor 
 and dynamo, provided all machine and line connections 
 remain the same and the line polarity is not changed. 
 
 All four types of machine, series, shunt, compound- 
 wound and separately excited, can be run in series or 
 multiple, or in any series multiple combination, provided 
 proper precautions are taken. Machines of the same type 
 run best together, because their iron is apt to be of the 
 same quality, making their saturation curves similar, and 
 this means that for any change in load their voltages would 
 vary in equal measure, thus keeping the load properly 
 distributed. On similar machines the armature resist- 
 ances are nearly the same, and the armature drops will 
 be the same. Again, on series and compound-wound 
 machines it is, for reasons to be considered later, abso- 
 lutely necessary that the series field resistances bear a 
 certain ratio to each other if the machines are to be run 
 with an equalizer. On machines from the same factory 
 this rule is observed. Machines of widely varying cur- 
 rent capacity should not be run in series, nor those differ- 
 ing greatly in voltage in multiple. In series working 
 there should be a separate voltmeter across the brushes of 
 
306 TESTING OF DYNAMOS AND MOTORS. 
 
 each machine as a means of noting the distribution of 
 load. The load is regulated by varying the voltage, since 
 the current is the same in all. This voltage variation is 
 effected in the usual way on a shunt machine, but on 
 a series machine the rheostat is in multiple with the 
 field winding, and in operating it more or less exciting 
 current is diverted from the field. In multiple work the 
 load distribution is indicated by ammeters placed in cir- 
 cuit with each machine. On account of the greater 
 liability to reversals, more precautions are needed in 
 parallel than in series work. One point to be always 
 observed is to never close a switch across which there 
 is a difference of potential, unless the operator under- 
 stands exactly what is to follow. Occasions often arise 
 for quick action, and it may be necessary to throw a 
 machine in without the usual preliminaries, but this is 
 not good practice. 
 
 The easiest way to throw a machine into service where 
 others are running in multiple depends upon the accuracy 
 of a zero reading across the switch. When a voltmeter 
 reads zero between two points, and there is a certainty 
 that the lines are all right and make good contact, it is 
 safe to join these two points, for no current can flow 
 between points of the same potential. All more hurried 
 methods are adaptations of the following: In Fig. 107, 
 let P and N be the line wires or bus bars of a station, A 
 is a shunt dynamo already running, and increase of load 
 requires B to be put on. A being in service its switch, 
 K, is closed, while K' on the idle machine is open. If 
 B is a new machine, or one whose connections have been 
 disturbed, it will be best to first test its polarity. Lift- 
 ing .Z?'s brushes K' is closed, and the fields changed from 
 
SHUNT AND COMPOUND MACHINES. 
 
 307 
 
 the line. Then opening K' and lowering the brushes the 
 machine will pick up a field if the connections are cor- 
 rect. This test being satisfied, one voltmeter terminal is 
 held on the right hand block b' on the headboard of the 
 machine, and the other terminal is touched to the upper 
 switch block on the left. The reading gives the line 
 voltage. Transferring the meter terminal from the 
 
 FIG. 107. 
 
 t 
 
 upper jaw to the lower, we get the machine's voltage, 
 which must be so adjusted by means of the rheostat that 
 it equals that of the line. Whc'n these voltages are equal 
 and opposite the meter will indicate zero when placed 
 across the switch, for the line voltage tries to send cur- 
 rent through it one way, and &s voltage the opposite; 
 the tendencies are equal, so the needle does not move. 
 At this point the switch may be closed, and /?'s field 
 strengthened a little to prevent reversal in case the speed, 
 and hence voltage, of 's armature may chance to fall. 
 The load can now be gradually transferred from A to B 
 by strengthening B's field, and at the same time weaken- 
 
308 TESTING OF DYNAMOS AND MOTORS. 
 
 ing A's. It may happen that when the meter is placed 
 across the switch, instead of reading nearly zero it reads 
 twice the line voltage. This would indicate the machine 
 and line to be in series instead of multiple, and would 
 necessitate recharging. Charging may be avoided by 
 connecting the field above the switch, thus exciting from 
 the line direct. When there is any uncertainty as to 
 connections the above test must be fully made, but where 
 there is absolute certainty of correct polarity it suffices 
 to use a test lamp instead of a voltmeter, or even a 
 marked position on the field rheostat. However, the 
 position of the rheostat varies considerably for different 
 line loads and different temperature of the machine, so 
 that the attendant must exercise considerable judgment. 
 The test lamp is applied in the same way as the volt- 
 meter, and when it burns above the switch with the same 
 brilliancy as below, the switch can be closed, but as an 
 extra precaution the lamp should be tried across the 
 switch. In doing this care must be had, for if the 
 machines should happen to be in series the lamp will be 
 subjected to double its normal voltage and will explode, 
 not only jeopardizing the operator's eyes, but possibly 
 injuring the armature with broken glass. To lessen 
 such liabilities it is well to use two lamps in series. In 
 station practice connections are permanent, and the test 
 across the switch is omitted. Lines are run from all 
 machines to the switchboard, as are also a pair of lines 
 from the bus bars: by means of a key the voltmeter can 
 be connected alternately to the bus bars, and to the 
 terminals of the dynamo to be put on. When the volt- 
 ages are equal the switch is closed. When K' is closed 
 B does not begin to take a load until its field is 
 
SHUNT AND COMPOUND MACHINES. 309 
 
 strengthened, and for this reason other machines in ser- 
 vice support a certain E. M. F. between S and /*, to 
 which Z?'s terminals are attached. B supports the same 
 K. M. F. between S and P in opposition to that already 
 existing; /'being a point in common to both circuits, 
 and the potential on both sides of K' being the same, cur- 
 rent does not flow when K' is closed. Before the switch 
 is closed there is below it the total E. M. F. of /?, while 
 above is the terminal potential of A\ A's total E. M. F. 
 being much greater, B can share the load only when its 
 total E. M. F. exceeds the terminal E. M. F. of A. 
 This condition is secured by strengthening /?'s field or 
 weakening A's. When there are many machines in 
 multiple the latter procedure is impracticable. As />'s 
 field is strengthened /?'s E. M. F. rises, and for a line 
 current of given value the line E. M. F. does also, because 
 the internal resistance of two or more machines in 
 multiple is less than that of a single one; there is, there- 
 fore, less internal drop, and the E. M. F. available at the 
 terminals common to all the machines is greater. 
 
 After adjusting the loads on all the machines at their 
 proper values, it is well to keep an eye on the newly intro- 
 duced machine, until it has become throughly heated, 
 because a$ its armature and fields rise in temperature 
 their resistance increases, with the result that the internal 
 drop becomes greater, and the field takes less current 
 both effects conspiring to lower the terminal E. M. F. and 
 to make the machine shirk its load. To the competent 
 and experienced practical man such simple admonitions 
 may seem uncalled for, but to the plodding beginner 
 these little points become the food of much thought. 
 Another effect of heating is to raise the reluctance of 
 
310 TESTING OF DYNAMOS AND MOTORS. 
 
 the magnetic circuit, thereby weakening the field and 
 further lowering the E. M. F., and when we remember 
 that each machine's load depends upon its E. M. F., it is 
 not hard to realize the importance of these influences. 
 Should any part of the magnetic circuit become impaired, 
 as by a loose or rusty joint, the effect is the same as the 
 above. Such a defect in a dynamo cuts down its E. M. F. 
 and reduces its load; on a motor it cuts down the 
 C. E. M. F. and increases the load. 
 
 The effect of throwing an additional machine in mul- 
 tiple with others is not only to increase the current 
 capacity of the plant, but to also decrease its internal 
 resistance. The lower the external resistance the greater 
 the useful effect of putting on an additional machine, 
 because if the resistance of feeders and mains is too 
 great, the external resistance will be high compared to 
 the internal, and the latter will be but a small part of the 
 total, and lessening this small part has no apppreciable 
 effect. To use a familiar analogy, take the case of a 
 battery working through an external resistance of 1,000- 
 ohms. It would be useless to put any number of cells in 
 multiple on such a line, for the 1,000 ohms is a so much 
 greater resistance than the cell part of the circuit, that 
 any decrease in the latter cannot be detected. 
 
 The problem complementary to that of " putting on " 
 a machine, is " taking off " one. This occurs when the 
 station load falls below the normal capacity of the 
 machines in service. Of course one way to take off a 
 machine is to pull its switch, and this is practiced in 
 emergencies, but ordinarily is open to several objec- 
 tions; in the first place on machines of large current 
 capacity and of sufficient voltage to support a healthy 
 
SHUNT AND COMPOUND MACHINES. ? I I 
 
 arc, the switch blade and jaws soon blister and burn: 
 secondly, on some machines using carbon brushes, and 
 on all machines using copper ones destructive sparking 
 attends a sudden removal of load. Thirdly, where there 
 are but two or three machines working, the sudden 
 removal of one may overload the rest and give rise to 
 belt troubles. 
 
 Another way to take off a machine is to shut down the 
 engine which runs it; this can be done only when a 
 machine is alone on the circuit. Under no circumstances 
 must the steam be cut off from an engine connected to a 
 dynamo running in multiple with other dynamos, driven 
 from a different engine; for the steamless engine will 
 continue to run driven by its dynamo, which has become 
 a motor. If the dynamos are simple shunt machines 
 there may be no fireworks, but if they be compound- 
 wound or series machines, there will be. The following 
 method covers all cases where two or more shunt 
 dynamos run in multiple: where practicable strengthen 
 the fields of the dynamos that are to remain in circuit 
 and weaken the field of the one to be removed, keeping 
 watch over the machine's ammeter to avoid reversal; 
 over the station voltmeter to keep up the voltage; and 
 over the brushes to shift them if necessary. When the 
 current falls nearly to zero the switch must be opened, 
 and then the field circuit broken. In no case should the 
 field be broken first. Where it is not practicable to 
 change the rheostats simultaneously as above, the line 
 voltage may require adjustment afterward, for the effect 
 of removing a machine is to increase the internal resist- 
 ance and the internal drop. The larger the number of 
 machines in multiple the less the disturbing effect of any 
 
312 TESTING OF DYNAMOS AND MOTORS. 
 
 one's removal. In bringing down the load by means of 
 the field rheostat each increase of resistance should be 
 allowed to have its full effect before introducing any 
 more, for magnetization takes an appreciable time to 
 respond to any change in the magnetizing force. Wher- 
 ever any load or speed is regulated with a rheostat, it 
 saves time and trouble to observe this precaution. On 
 the same principle the ammeter needle will lag behind 
 the current; and the current may reach the point of 
 reversal some time before the needle indicates zero, so 
 on this account it is customary to open the switch on the 
 safe side of zero. This is a point which the attendant 
 should understand, for if the machine's E. M. F. is 
 allowed to fall below that of the line, reversal takes 
 place, and the accompanying sparking may injure the 
 commutator as much as several months' ordinary wear. 
 Aside from sparking (which may be very light under 
 favorable circumstances) and the reversal of the ammeter 
 needle (which takes place only on the Weston type of 
 direct current meter) there are other reversal signs to be 
 looked for : (i) motor brushes have a backward lead, if any, 
 under load, and dynamo brushes a forward lead; (2) the 
 tight side of the belt always runs toward the driving 
 pulley, and should* be on the bottom, so that the sag of 
 the slack side shall increase the pulley contact. If then 
 the bottom side of a dynamo belt is seen .to be flapping 
 or sagging, it indicates that its armature is pulling on 
 the belt and is therefore a motor. Where suspicion is 
 aroused the real test lies in strengthening the field and 
 observing the ammeter needle; if the machine is a motor 
 the load will decrease as the field is strengthened. By 
 looking at Fig. 108 it can be seen that when two 
 
SHUNT AND COMPOUND MACHINES. 
 
 3*3 
 
 dynamos run in multiple the E. M. Fs. of the two arma- 
 tures oppose each other, but concur in direction in the 
 
 outside circuit; if the line 
 switch be opened, as at 
 K, there is no common 
 external path, and the op- 
 position of the two E. M. 
 Fs. can be represented 
 by the two arrow heads 
 pointed against each other 
 in circle A B C of Fig. 
 109. If K be closed, and 
 there is not too great a difference between the voltages 
 of the two sources, they will take the common path D L 
 C, If, however, the E. M. F. of one, B say, is much 
 greater than that of the other, it sends backward through 
 A against its weaker E. M. F. a current which runs it 
 as a motor. This takes place when ^'s terminal E. M. F. 
 is greater than A's total 
 E. M. F. The violence of 
 the demonstration which 
 follows a shunt machine's 
 reversal depends upon the 
 discrepancy between the 
 two E. M. Fs. If this is 
 very little, it may not be 
 noticeable that one machine 
 is running, as a motor; on FIG. 109. 
 
 the other hand if the field 
 
 be broken on one machine its armature will have no 
 opposing power at all, and a short circuit follows, with 
 tremendous sparking. 
 
314 TESTING OF DYNAMOS AND MOTORS. 
 
 We have learned that the terminal voltage of a shunt 
 machine falls as the load increases, but that up to a certain 
 limit that of a series machine rises. It was early sug- 
 gested that such a combination of these two machines 
 might be made as would give between certain limits a 
 constant terminal potential. The result of this sugges- 
 tion is embodied in the compound-wound dynamo. 
 Compounding consists, as the name indicates, in placing 
 both shunt and series coils upon the field cores, and in so 
 proportioning their magnetizing effects that any tendency 
 of the shunt field to weaken is met by a counter tendency 
 of the series field to strengthen. Stated thus boldly the 
 problem may appear to be simple; but there are modify- 
 ing factors which render it more complicated than it at 
 first appears. These we will take up in order. In the 
 first place, if a dynamo has been compounded cold it will 
 be under-compounded when hot; and if compounded hot 
 will be over-compounded when first started up. This is 
 due to the fact that as the temperature varies so does the 
 resistance of field and armature, and also the reluctance 
 of the magnetic circuit. Since, however, all machines 
 after running several hours reach a stable condition as 
 regards temperature, these effects may be considered as 
 fixed quantities to be allowed for. If a dynamo has been 
 successfully compounded its terminal potential remains 
 practically constant for all loads, and the shunt winding 
 will be subjected to a constant voltage; the shunt field 
 ampere-turns will then be constant after the temperature 
 becomes so; hence in calculations for compounding, the 
 shunt field is considered constant, and has such a value 
 as will give the required voltage on open circuit when 
 the series coils are inactive. 
 
SHUNT AND COMPOUND MACHINES. 315 
 
 The third modifying factor is the / J? drop through 
 the armature and series winding. The drop, being due 
 to resistance, would vary with temperature, so the resist- 
 ances are always taken at maximum temperature and 
 considered constant. If R is a constant, then the drop 
 or loss of potential is directly proportional to /; /'. e., 
 to the load. If the drop at quarter load is 10 volts, then 
 at full load it will be 40 volts. To compensate for this loss 
 it is convenient to know the loss of voltage per ampere 
 of current in the armature. Thus, suppose that the shunt 
 field on open circuit gives 500 volts, and has 200,000 
 ampere-turns. This gives 200,000 -f- 500 = 400 ampere- 
 turns per volt. Next suppose that for every 10 amperes 
 of armature current the / R loss is i volt; this means 
 i/ioof a volt per ampere, or 40 ampere-turns per ampere. 
 For every ampere in the -armature there must be in the 
 series field 40 ampere-turns, so that for a current of 100 
 amperes the series field must provide 100 x 40 = 4,000 
 ampere-turns, which is just sufficient. 
 
 A machine compounded as above would still be defi- 
 cient, for there is a fourth factor to be allowed for. In 
 Chapter I., on the "Elementary Theory of Dynamos," 
 mention was made of the armature " reaction," or "back 
 induction," and the term back ampere-turns was ex- 
 plained. The measure of this armature reaction was 
 found to be the product of the armature current and the 
 number of armature coils included between the double 
 angle of lead. On well-designed compound-wound 
 machines the lead is the same for all loads, so that the 
 " back ampere turns " are proportional to the load and 
 equal to T b /, where T b is the number of armature coils 
 included in the double angle of lead. Call this 10; 
 
316 TESTING OF DYNAMOS AND MOTORS. 
 
 then the back ampere-turns per ampere is 10. This, 
 added to the 40 necessary to overcome the / R loss, 
 gives 50 as the number of ampere-turns to be furnished 
 by the series windings for each ampere of armature cur- 
 rent. At full load, 100 amperes, the series winding 
 would give 100X50 = 5,000 Si, and 5,000 -^ 100 = 50, the 
 number of series turns. Under these conditions regula- 
 tion should obtain from no load to such load as the design 
 calls for. That there is a maximum limit for regulation 
 naturally suggests itself, and this limit is found as fol- 
 lows: In Chapter I. it was shown that iron, under the 
 influence of a magnetizing force, is magnetized, and its 
 magnetism increases as the magnetizing force increases, 
 until a point is reached where the magnetism is but 
 slightly affected, even when the magnetizing force is 
 greatly increased. At this point the iron is said to be 
 saturated. Next, we note that in a compound-wound 
 machine the object sought is to secure a constant ter- 
 minal potential, and that this is effected by varying the 
 armature E. M. F. Since the armature's speed and 
 number of conductors are constant, raising the E. M. 
 F. is effected by strengthening the magnetic field, 
 and the regulation is perfect only so long as the iron's 
 magnetism increases at the same rate as the magnetizing 
 force. For a while this condition is satisfied, but as soon 
 as the iron approaches the saturation point the condition 
 is departed from more and more, till finally, even by 
 doubling the magnetizing force (the series ampere-turns), 
 the iron responds but little. From this point on the 
 terminal voltage decreases as the current increases, be- 
 cause the voltage loss due to increased reaction and 
 increased / J? drop is greater than that gained by the 
 
SHUNT AND COMPOUND MACHINES. 317 
 
 extra turns. The range of regulation in a machine 
 depends upon the design and upon the grade of iron 
 used in the fields, cast iron having a lower saturation 
 point than wrought iron or steel. Where the question of 
 weight is important, the cross-section of the field is a 
 minimum, so that even on open circuit, when the shunt 
 coils alone are active, the magnetization of the iron is 
 rather high. Under these circumstances the saturation 
 point is soon reached, and full load voltage will probably 
 be slightly lower than that at one-half or three-fourths 
 load. If weight and size are secondary considerations, 
 the field cores can be made of greater cross-section, so that 
 on open circuit the requisite field is secured with a low 
 magnetization value in the iron, and the range of regula- 
 tion is increased. If, in testing a number of similar 
 machines, one gives voltage readings below the average, 
 particularly at full load, it is quite possible that an inferior 
 quality of iron has found its way into the armature or 
 field core. Before drawing such a conclusion, however, 
 all other sources of error must be carefully eliminated. 
 All magnetic joints should be inspected and found tight. 
 The bore of the pole pieces should be calipered as well 
 as the diameter of the armature core. The resistance of 
 armature and both field windings must be taken, and, 
 finally, the whole test should be run over with a new set 
 of instruments and with the men differently disposed. 
 This will eliminate errors of observation and those due 
 to faulty instruments. If a machine's dimensions, the 
 quality of its iron, and all winding data are given, it is 
 possible to calculate very closely the regulation range; 
 but all commercial machines are put through a thorough 
 test, and finally adjusted experimentally. This adjust- 
 ment is later considered in detail. 
 
.318 TESTING OF DYNAMOS AND MOTORS. 
 
 In actual practice the compounding of a dynamo is not 
 as much of an undertaking as it might seem, and 
 has this advantage over the engineer's calculations, that 
 the data is derived from the machine itself, and not from 
 tests on previous machines; especially is this an advan- 
 tage where new types of machine are being perfected. 
 Suppose that we have on hand a shunt dynamo, which it 
 is desired to compound a case very likely to arise out 
 on the road, where, for instance, it is necessary to adapt 
 a shunt machine to respond to the sudden fluctuations of 
 a street railway circuit. There are two methods of pro- 
 cedure, both of which are equally practical and practicable, 
 and give satisfactory results. The following requires no 
 data and is free from calculation: There should be 
 available a spool of stranded copper cable of about the 
 same cross-section as one of the brush-holder cables, 
 also two heavy clamps to serve as temporary terminals 
 for the cable. The method consists in the actual placing 
 of successive turns of wire on the two field spools and 
 connecting them in series with the armature and external 
 line. The machine's open circuit voltage is adjusted to 
 nominal value, and the rheostat thereafter left undis- 
 turbed. The load is now gradually increased by cutting 
 in more lamps or otherwise lessening the external resist- 
 ance, and the terminal voltage watched; if it falls off, the 
 series turns must be increased. If the voltage raises as 
 the load does, the machine is over-compounded, and some 
 turns must be taken off. This process of hit and miss is 
 continued until a satisfactory combination is secured. 
 To make room for the added coils the rope covering on 
 each spool may be removed. A layer of series winding 
 is temporarily placed on each spool in such a way that 
 
SHUNT AND COMPOUND MACHINES. 319 
 
 the current must flow in opposite directions around the 
 two spools. To effect this, the wire on one can be wound 
 from right to left, looked at from above, and on the 
 other from left to right, and the inside ends connected 
 together. Or, if both spools have been wound alike, the 
 inside end of one and the outside end of the other are 
 connected together. The two coils, as a whole, must 
 now be connected in circuit, so that the armature current 
 passes around shunt and series coils, on any spool, in the 
 same direction. 
 
 Having approximately determined the required number 
 of coils a more careful test is made. With the shunt field 
 thoroughly heated, the rheostat is again adjusted to 
 normal voltage and the load put on; if properly com- 
 pounded the open circuit voltage will be maintained 
 throughout the load range. As a preliminary test the 
 temporary series coils should be cut out and the full 
 current load put on, with the rheostat undisturbed from 
 the position of normal voltage. A fall in voltage will of 
 course result, and its amount should be noted. Then 
 the series coils should be cut in and full current again 
 put on. If the voltage is less than before, it indicates 
 the series and shunt coils to be opposed; this will be the 
 symptom if there are but a few series turns on; but if the 
 series coil is complete, its opposition will make it impossi- 
 ble to get on very much of a load. In any such case the 
 connection of the series coil is the one to reverse not 
 that of the shunt coil; reversing the shunt winding 
 deprives the machine of ability to generate. If the read- 
 ing is but little affected, being almost what it was at full 
 load with shunt winding alone, it is indicative that the 
 two series spools oppose each other, and that the fall in 
 
320 TESTING OF DYNAMOS AND MOTORS. 
 
 voltage is due to their resistance. The last condition is 
 where the voltage is nearly that of open circuit, and it 
 indicates that compounding is progressing and only needs 
 exact adjustment; this is done by adding or removing 
 one turn at a time and successively taking readings. 
 After the adjustment of series turns is made the machine 
 is allowed to run on full load long enough to bring up 
 the series coil to its maximum temperature. At the end 
 of the run the voltage will be found to have lowered on 
 account of the rise in series field resistance; and this loss 
 must be compensated by putting on more turns. The 
 cable must be as large as the winding space will permit, 
 for it is important to keep down the loss in the series field 
 itself. A machine compounded in this way, with all its 
 magnetic and electrical losses present, is compounded 
 under working conditions, and can be relied on to hold 
 its own in the future against all losses, whether due to 
 armature reaction, ohmic resistance, or any other cause. 
 After a compound is secured, it is considered good prac- 
 tice to add enough series turns to raise the full load volt- 
 age 10 % above the open circuit value. Thus, if the open 
 circuit voltage be 500 volts, the full load voltage will be 
 550 volts. The object of this over-compounding is to 
 enable the machine to maintain a uniform delivery volt- 
 age at some point whose distance will not entail an 1 R 
 loss of over 50 volts at full load. By knowing the resist- 
 ance of the feed wire or main between this point and the 
 dynamo, the drop can be found and series coils wound 
 on accordingly. If the series field is stronger than is re- 
 quired, its effect is weakened to any desired extent by plac- 
 ing in multiple with the winding a shunt, as shown in Fig. 
 no. Shunted dynamos are most generally used in street 
 
S~\ ^nnnnnnnr 
 
 J A I/ 
 
 A / _/~vw~> 
 
 ^-^ s 
 
 /K L 
 
 SHUNT AND COMPOUND MACHINES. 321 
 
 railway work, where feeders of different lengths require 
 different degrees of over-compounding. A convenient 
 form of shunt, and one much used, is made of German- 
 silver tape. Knowing from previous tests or by calcu- 
 lation the approximate resistance of the shunt, several 
 lengths of tape are cut, clamped together in multiple, and 
 connected across the series 
 field, as shown in the figure 
 at 6". One of the clamps is 
 so arranged that the strips 
 can be easily slipped through, 
 and the shunt thus length- 
 ened or shortened. Having 
 placed the shunt upon the 
 
 machine the load is adjusted and the voltage meas- 
 ured; if too high the shunt is shortened, thus lowering 
 its resistance and diminishing the current in the series 
 coils. If the voltage is too low the shunt is length- 
 ened or, if necessary, an entire strip can be cut out. 
 There are several combinations of strips that will give 
 the proper resistance, but one must be selected that will 
 carry the current without heating too much. The office 
 of the shunt is to adjust the current in the series field 
 winding, and thereby to secure the proper magnetizing 
 effect. Having thus completed the shunt, it is soldered 
 at its ends, and becomes an integral part of the machine. 
 Sometimes, when a whole plant is installed, the final 
 compounding is done after the machines are in place. 
 One point in favor of this practice is that the machines 
 are compounded for exactly the speed at which they are 
 to run, whether this be above or below the rated speed 
 or not. This is important, for a machine will regulate 
 
322 TESTING OF DYNAMOS AND MOTORS. 
 
 perfectly only at the speed at which it is compounded, 
 because its voltage is due not only to the magnetic field 
 but to the speed, so that any change in speed changes 
 the voltage. Furthermore, this change in voltage, due 
 to speed change, reacts upon the shunt field, strengthen- 
 ing or weakening it according as the voltage has been 
 raised or lowered. Under these circumstances rheostat 
 regulation must be used so long as the speed is other 
 than that at which the machine was compounded. 
 Proper speed is, then, of prime importance, and com- 
 pounding in situ has a decided advantage. 
 
 In practice machines are frequently compounded to 
 maintain constant potential at a distant point on the line. 
 A machine thus appointed is said to be over-compounded 
 a certain per cent, above the normal voltage, according 
 to the loss to be allowed for. If adjusted for a 10 % 
 loss, and intended to give 500 volts on open circuit, the 
 full-load terminal voltage must be 555 volts, for the con- 
 dition is that 500 shall be 90 <f of full-load voltage, and 500 
 volts is 90 f of 555 volts. Again, if a 20 f c loss is to be 
 allowed for, the full-load terminal voltage must be 625. 
 When such heavy over-compounding is called for it is 
 customary not only to increase the series field strength, 
 tout to raise the speed. This calls for a new adjustment 
 of the field rheostat. For, in any case, the normal voltage 
 is 500, and it stands to reason since the voltage is due 
 to speed and field, and since 500 volts are due to the 
 cutting of a certain number of lines of force per second, 
 that if the speed is increased the lines must be decreased 
 in order that the number cut per second may be the 
 same, and 500 volts obtain. To decrease the number of 
 lines resistance is put into the field rheostat to cut down 
 
SHUNT AND COMPOUND MACHINES. 323. 
 
 the field current. In contracts where the conditions are 
 previously known, the machines are provided with special 
 windings. Where machines leaving the shop are com- 
 pounded fora particular speed, the shunt should be such 
 as to admit of ready variation in resistance. Some 
 makers use German-silver tape; others adopt a variable 
 shunt board, by means of which a compound can be 
 effected for a 10 % variation in speed. When it is im- 
 practicable to run the speed within this limit, the machine 
 may be recompounded as above, either by using a shunt 
 or ^v adding more series coils, according as the speed is 
 too nigh or too low. When a compound-wound dynamo's 
 speed is increased, the effect upon the two windings is 
 different. The I*R loss in the armature and series field 
 remains the same for given current, but the voltage con- 
 tributed by the series winding varies with the speed. 
 Assuming /to be kept the same and the speed tb be in- 
 creased, the series Si (ampere-turns) remain the same. 
 Any increase of voltage due to cutting of lines of force 
 furnished by the series windings is due to increase of 
 speed alone. \Vith the shunt windings the magnetic 
 effect is augmented by not only the increase of speed 
 directly, but, since the voltage applied to the shunt field 
 is raised, to the increase in field current. At variable 
 speed, then, the series field current remains the same, 
 but that in the shunt field does not. Assuming the open- 
 circuit voltage to be the same, however the speed may 
 vary, the machine will still fail to compound at any other 
 than its proper speed, for according as the speed is high 
 or low the series effect will be too much or too little. 
 
 Often in testing rooms it is found impracticable, 
 for one reason or another, to run a machine at its. 
 
324 TESTING OF DYNAMOS AND MOTORS. 
 
 proper speed. A somewhat curious and not very accu- 
 rate circumlocution is then resorted to. It is called 
 ''compounding volts per speed," or "volts per revolu- 
 tion," and is based upon the assumption that any change 
 of speed will provide a like change in the voltage. Thus, 
 it is assumed that if a machine generates 500 volts at 
 1,200 revolutions per minute, it will give 1,000 volts at 
 2,400 revolutions per minute, and 250 volts at 600. This 
 assumption may be made without serious error if it is 
 experimentally proven that the open-circuit voltage fol- 
 lows this law, but it is not at all likely that it ever does 
 so. It is much more satisfactory to determine just what 
 effect a given speed variation has on the voltage, and to 
 base corrections upon this, for we have just seen that 
 speed variation of voltage on a compound-winding ma- 
 chine follows no such simple law. Compounding volts 
 per revolution is carried out as follows: Suppose we have 
 a roo-volt dynamo whose proper speed is 1,000 revolutions 
 per minute, but upon taking its speed it is found to be 
 running at only 900 revolutions per minute. This is 100 
 revolutions, or 10 % too low. The normal voltage at 
 the rated speed would be 100, but since the speed is but 
 90 </ of its proper value, the voltage is adjusted to 90 <f, 
 of its proper value, and this gives 90 volts. The as- 
 sumption is that if a speed of 900 gives a voltage of 90, a 
 speed of 1,000 would give a voltage of 100. The load is 
 then put on, and the same performance gone through 
 with. 
 
 The foregoing hit-or-miss method of compounding is 
 satisfactory when one has a reel of cable from which, he 
 is at liberty to use what is needed, and has the privilege 
 of returning the rest. Frequently, however, the cable 
 
SHUNT AND COMPOUND MACHINES. 325 
 
 must be ordered and paid for in advance, and it is de- 
 sirable to know about how much is needed, so as to avoid 
 needless expense. Under these conditions- the following 
 procedure is good practice: If a voltmeter is available, 
 adjust the machine to normal voltage, with line switch 
 open and armature at right speed. Should no reliable 
 voltmeter be obtainable, any device that can be used to 
 read voltage may be calibrated from the dynamo's pilot 
 lamp. With the lamp at proper brightness, as far as the 
 eye can judge, the indicator needle's position is marked, 
 the indicator being across the terminals. Once this sin- 
 gle point is determined we can always tell when the 
 voltage is right. Any indicator that will give a readable 
 deflection will serve the purpose, and, if necessary, it can 
 be placed in series with a high resistance to reduce 
 the deflection. If no indicator can be had, the ability 
 of the eye to judge the pilot lamp's brightness must 
 be relied upon. This, while not very accurate, gives 
 fair satisfaction. Knowing the voltage which the 
 machine is giving, and keeping this constant, we know 
 the field current to be constant, and knowing or 
 measuring the resistance of the field, we can, from 
 Ohm's law, get the value of this current. If there is no 
 way of measuring the field resistance, it can be deter- 
 mined approximately, as follows : Remove enough of 
 the rope, etc., to enable the outside diameter of the 
 winding to be measured; remove the keeper and expose 
 the iron field core; the diameter of the core, less 1/4 inch 
 to be allowed for insulation, gives the inside diameter of 
 the winding; this taken from the outside diameter gives 
 the radial depth of the winding. The radial depth divided 
 by the diameter of the wire gives the number of layers; 
 
326 TESTING OF DYNAMOS AND MOTORS. 
 
 the length of the core divided by the wire's diameter 
 gives the number of wires per layer; the number per 
 layer multiplied by the number of layers gives the number 
 of turns on the spool. From the outside and inside diam- 
 ters can be figured the average length of a turn, from the 
 formula 
 
 X = : 
 
 2 
 
 or, 
 
 D -f- </ 
 
 # = =Q - X 7T, 
 
 Where Z>, ^, and x are, respectively, the outside and 
 inside diameters and the average length of a turn. The 
 average length per turn multiplied by the number of 
 turns gives the length of wire on the spool. Taking the 
 diameter of the wire, not including the insulation, and 
 consulting a resistance table, we find the resistance per 
 foot; this multiplied by the number of feet equals the 
 resistance of the wire on one spool, and this multiplied 
 by the number of spools equals the cold resistance of field 
 coils. Next subject the field to its designed voltage till 
 it reaches its maximum temperature. From the rise in 
 temperature get the rise in resistance, and we have the 
 field resistance hot; whence the field current 
 
 E 
 
 where R-^ is the field rheostat resistance. As a last re- 
 sort, R^ can be gotten by substituting for k a wire of 
 
SHUNT AND COMPOUND MACHINES. 327 
 
 known dimensions whose resistance is readily obtained 
 either by calculation or from a table. In lieu of wire, a 
 combination of lamps can be used. Such a combination 
 must be selected as will either bring the lamps to incan- 
 descence or not heat them at all, for there is a great dif- 
 ference between the hot and cold resistance of a filament. 
 Knowing the resistance of one lamp under either of the 
 above conditions, and assuming them to be all alike 
 (rather a bold assumption), the total resistance can be 
 roughly approximated. We have nov, the field current 
 and the number of turns on the field, from which, by 
 multiplying, we get the field ampere-turns at no load and 
 proper speed. Taking care that the speed does not 
 change, full load is now put on, and /the voltage kept up 
 by the rheostat. In varying the rheostat to meet the* 
 requirements of the load the ampere-turns of excitation 
 have been increased, and the amount of this increase is 
 an exact measure of the Si to be furnished by the series 
 windings, and must be measured. To do this, the new 
 value of the field circuit resistance must be determined. 
 Since the field resistance remains the same, the problem 
 is to again find what j? b is. This known, 
 
 From this we get the Si' at full load, and the difference 
 between Si and Si' gives the ampere-turns to be fur- 
 nished by the series coils. To find the number of series 
 coils needed it is only necessary tc divide the series Si 
 by the full load armature current. Thus, suppose the no 
 load Si to be 25,000, and the full load Si to be 28,000, 
 
328 TESTING OF DYNAMOS AND MOTORS. 
 
 also the full-load armature current to be 100 amperes, 
 then the number of series turns is 
 
 28,000 2C.OOO 
 
 or 15 turns on each spool, if there are two spools. Know- 
 ing the amount of space available for winding, the amount 
 of cable necessary is readily figured and ordered accord- 
 ingly, with a little margin for terminals and final experi- 
 mental adjustment. 
 
 The above method is incomplete in that it makes no 
 allowance for the drop through the series coils themselves. 
 The first result is an approximation then, and must be cor- 
 rected by the addition of sufficient turns to care for the 
 series field IR loss as figured from the size and length of 
 the cable used. Each addition of series turns further 
 increases the resistance, so that more turns must be added 
 to compensate for this increase. The process of adding 
 and reading can be carried to the closest approximation. 
 The method may seem rather tentative, but in the par- 
 ticular case it is assumed that no data is held and that 
 there are no facilities for testing. Should it develop that 
 the iron is so saturated by the shunt coils that series coils 
 have too. little effect, a compound can be effected only by 
 running the dynamo at a lower voltage. This lessens the 
 shunt field lines of force and makes more room for the 
 series. 
 
 In calculating the size of cable for use as series 
 coils, the IR loss in them should not exceed 2 $ of the 
 terminal voltage. Knowing the length of cable to be 
 put on, and the maximum allowable resistance, a resist- 
 ance table will give the minimum cross-section. It is 
 
SHUNT AND COMPOUND MACHINES. 329 
 
 well to keep within the limit, so as to allow for any neces- 
 sary additional turns. Besides the desirability of minimiz- 
 ing the series field loss, it may be observed that if too 
 small a wire is used it will be impossible to compound the 
 dynamo at all. This condition obtains when the resist- 
 ance of each added turn causes an additional drop greater 
 than the increase of voltage which its magnetizing effect 
 produces. Each turn must therefore supply more voltage 
 than it itself consumes. 
 
 Series windings on small dynamos are generally 
 stranded copper wire, thus offering more radiating sur- 
 face for given sectional area than solid wire does. On 
 large machines flat copper tape is used, being almost as 
 flexible and more compact than the stranded cable, and 
 giving a larger radiating surface. Where an old shunt 
 machine is compounded, the series turns are wound 
 outside the shunt, but on new machines the series 
 coils are placed underneath: the tape or strip copper 
 being lathe wound between alternate layers of insulation. 
 The reason old machines take the series coils outside is 
 that the shunt coils' presence makes it cheaper and 
 more convenient to do so: new machines take them 
 inside because that in order to keep the IR drop low the 
 cross-section of the series conductor is comparatively 
 great, and the current density, hence the heating, is less 
 than in the shunt field, which is therefore put outside 
 next to the air. Aside from this, a fine wire winding tends 
 to heat more, because the wires lie closer together and 
 lessen the intermediate air spaces. When not wound one 
 over the other, the shunt and series coils are wound side 
 by side, each having full radial depth. This enables 
 either one to be removed without disturbing the other. 
 
330 TESTING OF DYNAMOS AND MOTORS. 
 
 It simplifies connecting to wind shunt and series coils in 
 the same direction on the spool. Before making perma- 
 nent connections, temporary ones can be used for getting 
 the two windings of the same and alternate pole pieces 
 of opposite polarity. In using the pilot lamp as a guide, 
 allowance must be made for the fact that as a rule it 
 burns brighter than the lamps out on the line. 
 
CHAPTER X. 
 
 THE COMPOUND-WOUND DYNAMO. GENERAL TESTS. 
 
 A COMPOUND-WOUND machine, partaking as it does of 
 the nature of a shunt and series machine, possesses the 
 characteristics of both; like a shunt dynamo it maintains 
 its voltage on open circuit, and like a series machine it 
 maintains its field on short circuit, and will give trouble 
 with belts or clutches unless the cut-outs act promptly. 
 As with shunt machines, the independence of its shunt 
 field adapts it for running in multiple with other ma- 
 chines; but the instability of its series field requires the 
 UFC of a special regulating device called an "equalizer." 
 
 The office and action of an " equalizing bar" is un- 
 derstood by comparatively few. Its object is to enable 
 each of several compound-wound or series machines, run- 
 ning in multiple, to take its share of the load, and to 
 make them independent of small speed variations. Its 
 action is as follows: Suppose two compound-wound 
 machines are connected in multiple, and that the series 
 and shunt coils on both dynamos are cumulatively 
 connected: one terminal of each dynamo goes to the 
 negative "bus-bar," as shown in Fig. in; the other 
 terminal, one end of the series winding, goes to the pos- 
 itive "bus-bar." The equalizer runs from the brush 
 to which the series field on one machine is attached to 
 the corresponding brush on the other machine. It 
 
 33 
 
332 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 always joins the two brushes which are at the same po- 
 tential, when the machines run at the same voltage. 
 
 If by mistake the equalizer is run from the positive 
 brush of one dynamo to the negative brush of another, 
 a short circuit results when the switches are closed, be- 
 cause (Fig. in) it brings together the positive and 
 negative sides of the same dynamo. It is necessary, 
 therefore, that the machines should be of such polarity 
 
 l^lVWWW 
 K' i-V 
 
 K 
 
 FIG. in. 
 
 as to make of the same sign all brushes attached to the 
 series fields. This condition can be obtained as follows: 
 a voltmeter is placed across the brushes of one machine 
 in such a way that when the meter circuit is closed the 
 needle deflects in the proper direction; the line con- 
 nected to the brush next to the series field is marked; 
 placing the meter across another machine, with the 
 marked line on the series field brushj the deflection 
 
THE COMPOUND-WOUND DYNAMO. 333 
 
 should be in the same direction as before; if it is not, 
 the polarity of one machine must be reversed before 
 closing the line switch. This can be done by shifting 
 the series field connection to the other side of the 
 armature, in which case the shunt field must be con- 
 nected in "long shunt," otherwise the shunt connections 
 are reversed and the machine will refuse to generate; it 
 is customary, however, to effect reversal by recharging 
 the shunt field from the line. This done, and the volt- 
 ages of machine and line equal but opposite, the meter 
 should read zero between the series field brushes, and it 
 is here that the equalizer is connected. If one machine 
 runs with load and the other without load, the brush to 
 brush voltage of the former will be higher than that of 
 the latter, although their terminal voltages may be the 
 same. The terminal voltages must be nearly the same, 
 because the machines' terminals are points in common 
 on the bus-bars, except in so far as they may be connected 
 to the bars at some little distance apart, and the terminal 
 voltage of two machines wiH differ only by an amount 
 equal to the IR drop in that part of the bus-bar included 
 between them. In other words, the machine or ma- 
 chines already in service practically dictate the terminal 
 voltage. Their brush voltages differ, because: every 
 dynamo under load has an external and an internal cir- 
 cuit; if there are no motors in service the external cir- 
 cuit has only ohmic resistance; but the internal circuit 
 has both ohmic and inductive resistance ohmic resist- 
 ance due to the armature wire, and inductive resistance 
 due to the very force which produces the difference of 
 potential. The two brushes are then the boundary 
 line between the external and internal circuit, and as far 
 
334 TESTING OF DYNAMOS AND MOTORS. 
 
 as concerns one machine, its series field is part of its 
 external circuit. If a voltmeter be placed across a 
 machine's terminals, it reads the drop through the 
 external circuit less the series field 1R loss; but if the 
 meter is placed across the brushes it reads the total 
 external drop, which is therefore greater than the termi- 
 nal voltage. Now, on a machine which is up to voltage, 
 but is taking no load, no current flows; there is no 
 IR loss, hence its brush and terminal voltage are the 
 same. The machine is statically charged to the line's 
 difference of potential, and the entire drop takes place 
 across the generating force within the armature. When 
 one machine is loaded, then, and the other is free, the 
 series field brushes on the two are at different potentials, 
 hence so are the two ends of the equalizer, and when its 
 switch is closed current will flow through it to the field 
 of the free machine: the equalizer places the series fields 
 of the two machines in multiple, and the current through 
 each will depend upon their relative resistance and upon 
 that of the equalizer, whose, resistance is virtually part 
 of the series field resistance of the idle machine. Hence 
 the importance of having the equalizer short and stout. 
 
 In closing an equalizer switch there are produced two 
 effects: The loaded machine's series field is weakened, 
 for part of its current flows through the equalizer; the 
 free machine's series field is strengthened; hence, if the 
 line switch is closed on the free machine just after 
 the equalizer switch, the free machine will take load, and 
 the loaded machine lose some. The process of equaliza- 
 tion continues until the brush to brush potential of both 
 machines is the same, when no current flows in the equal- 
 izer. If from speed variation the voltage of one machine 
 
THE COMPOUND-WOUND DYNAMO. 335 
 
 falls, the potential of that end of the equalizer falls, and a 
 current flows through the equalizer to the series field of 
 the lower potential machine. The field of this machine 
 being thus strengthened, its voltage rises and it takes 
 more load. Equalization takes place promptly, and cur- 
 rent flows through the equalizer first one way and then 
 the other, and sometimes not at all. 
 
 Any attempt to run compound-wound or series machines 
 in multiple without either rigid connection between the 
 armature shafts or an equalizer, will always give trouble. 
 If the speed, and with it the voltage, of one machine falls 
 off, the other machine takes the extra load, and in this 
 way, strengthening its own series field, induces it to 
 further overload itself and unload its companion. On 
 machines overcompounded 10 % or 20 #, the trouble is 
 aggravated. 
 
 There is one condition under which two compound- 
 wound or series machines will run together in multiple 
 without either rigid connection or equalizer, and this is 
 when both belts are free to slip; it is then impossible to 
 overload either machine, for as soon as its load reaches a 
 certain value, the belt slips, the armature slows down, the 
 voltage falls, and the load shifts to the other machine, 
 which in turn performs the same cycle. This is not an 
 accepted mode of regulation, for aside from its inefficiency, 
 it consumes brushes by the sparking. Where there are 
 more than two compound-wound machines running in 
 multiple, the equalizing bar is carried from machine to 
 machine till all are joined together. Each machine thus 
 depends for regulation upon all the rest. 
 
 To introduce a compound-wound machine into circuit 
 with other compound. wound machines running in multi- 
 
336 TESTING OF DYNAMOS AND MOTORS. 
 
 pie, proceed as follows: The shunt field is first charged 
 from the line, and the open-circuit voltage adjusted to 
 the proper value, and opposed to the line voltage; next, 
 close the equalizing switch, thus putting a series field on 
 the machine. The immediate effect is to raise the 
 machine's voltage, and to lower that on the line by an 
 amount depending upon how many machines are in service. 
 The line switch is next closed, and the machine immedi- 
 ately takes its load. The line switch is not closed first, 
 because in stations wh^re the machines are heavily over- 
 compounded the terminal voltage of the loaded machines 
 exceeds the open circuit voltage of the machine to be put 
 on, and closing the line switch first under these condi- 
 tions would result in running the machine as a motor, 
 and probably in the wrong direction. With the equalizing 
 switch in first this cannot occur for the reasons stated 
 above. In practice a three-pole switch is arranged to 
 throw in the equalizer a little ahead of the line switch. 
 With the shunt field across the line it is impossible 
 to permanently reverse a compound-wound machine's 
 polarity, unless the polarity of all becomes reversed, for, 
 if the shunt field is permanently charged from the line, 
 its polarity is fixed so long as the connections are undis- 
 turbed. In all cases where changes or repairs have been 
 made in field or rheostat wiring, the machine's polarity 
 should be tested. The equalizer connections must be 
 good, and the bar of good cross-section, otherwise the 
 equalizer fails in its object to place all series field brushes 
 at the same potential. Bad regulation between com- 
 pound-wound machines has frequently been traced to 
 insufficient copper, or to a loose connection in the 
 equalizer. 
 
THE COMPOUND-WOUND DYNAMO. 
 
 337 
 
 The discussion so far has involved only machines of 
 the same current capacity and of the same general type. 
 Often, however, it is desired to run in multiple machines 
 of equal voltage but different load capacities. The 
 equalizer makes this possible, provided a single condition 
 is satisfied, namely: that the resistance of the series 
 
 FIG. 112. 
 
 fields be made inversely proportional to the intended out- 
 put of the machines, so that for any variation of line 
 resistance the variation of series field / R drop shall be 
 the same. That is to say, if A's output is to be twice 
 ^'s, 's series field resistance must be twice ^'s; other- 
 wise they will not share the load properly. This feature 
 of the compound-wound machine is a series machine 
 characteristic, and is best understood when considered in 
 connection with them. Fig. 112 gives connections for 
 
338 TESTING OF DYNAMOS AND MOTORS. 
 
 two series dynamos running in multiple, with an equalizer 
 between them. A and B are the armatures of the two 
 machines, and S l and S 9 their series fields respectively; K 
 is the equalizer switch and K ', K" those of B and A. 
 A is the machine of larger output, say twice that of JB, 
 and the resistances of S l and S^ we suppose, for illus- 
 tration, to be the same. Neglecting the resistance of the 
 equalizer, the effect of closing K is to place in multiple 
 two fields (S lt S 9 ) of equal resistance; these fields will 
 therefore take equal currents if one machine is running 
 with load at the time K is closed. This current in itself 
 would perhaps be sufficient to injure the series field of 
 the smaller machine even were its line switch never closed, 
 so that the armature could do work. Aside from this, if 
 the two fields have the same number of turns the smaller 
 machine's field will be disproportionately strong, and if 
 smaller machine has more turns, as it is likely to have, 
 this disproportion increases, so that for all loads the 
 smaller machine takes more than its share, while the 
 larger machine takes less, and the equalizer is never idle. 
 To keep the load proportionately shared either of two 
 things must be done: First, The rheostat of the smaller 
 machine must be constantly changed to suit variations of 
 load just as on a* shunt machine; on circuits subjected to 
 sudden and violent changes, such as are found on street 
 railway circuits, this is impracticable, and where it is 
 tried sparking and belt slipping ensues; in any case, we 
 lose the most valuable property of a compound-wound 
 machine, namely, its regulative power, for a compound- 
 wound machine will regulate for constant potential only 
 when its rheostat is adjusted for normal voltage on open 
 circuit, and left in that position, the series winding pro- 
 
THE COMPOUND-WOUND DYNAMO. 339 
 
 viding the additional voltage necessary to care for load 
 losses. Second, The series field of the smaller machine 
 must be weakened, and this is done by putting in series 
 with it a resistance (such a resistance must be inserted 
 beyond the point where the equalizer taps on, so as not 
 to be in circuit when current flows from B to A}. Sup- 
 pose A has a current output of' 300 amperes, and B, of 
 100 amperes, and that A's series field resistance is .002 
 ohm: What should that of 7?'s be? Since /?'s current is 
 one-third of A's, 7?'s field resistance must be three 
 times A's; or 3 x .002 = .006 ohm. If it happens to 
 be only .0042 ohm, we must add an extra resistance of 
 .006 -- .0042 = .0018 ohm, which had preferably be 
 put between the series coils and the positive bus-bar. 
 On the other hand, if 77' s field resistance is too 
 great, or A's too small, the extra resistance must be 
 placed in A's field. In general, calling 7 A , 7 B the respec- 
 tive currents, and J? S1 ^' 8 the respective resistances of the 
 fields of A and B, then must R^ : R\ \ \ 7 B : 7 A , or in 
 words: to have proper regulation the series field resist- 
 ances must be inversely as the maximum currents they 
 are to carry. Multiplying means together and extremes 
 together, we have fi s 7 A = R' g I B , whence 
 
 J? 8 = - *'., and R\ = ' A - R % . 
 
 y A J V 
 
 Take the example analyzed above: 7 A = 300, 7 B = 100; 
 J?s = .002; jR' n ? From the above 
 
 R'*= -T- <R S = X .002 = .3 x .002 = .006 ohm. 
 
 J B IOO 
 
 Suppose .#'s field resistance = .008 ohm: What must 
 A's be? 7 A = 300, 7 B = 100; ' R =.008: ^ 8 = ? 
 
340 TESTING OF DYNAMOS AND MOTORS. 
 
 J? s = -p X ^' s = X .008 = : = .00266 ohm. 
 
 A 30 3 
 
 Since A's measures .002, an extra resistance .00266 
 .002 = .00066 ohm must be added to A's field. This extra 
 resistance consumes a small amount of power which is 
 the price paid for good regulation. In the cases above, 
 suppose the machines' 'voltage to be 500. Then the 
 watts output of each will be: 
 
 A = 500 x 300 = 150,000 watts 
 JS = 500 x ioo = 50,000 watts. 
 
 Watts wasted in A's resistance = .00066 X 3oo 2 = 
 .00066 x 90,000 = 59.4 watts, or about what is con- 
 sumed in a 16 candle-power lamp. 
 
 Compound-wound dynamos and shunt dynamos can 
 be run in multiple with each other if the load does 
 not vary, or if the latter's field is regulated by hand; 
 otherwise, the compound-wound machine will take more 
 than its share of load for all loads above that at which 
 the shunt machine's rheostat has been adjusted. This 
 is because as the load increases the shunt machine's 
 voltage falls off, while the compound-wound machine's 
 -does not. Again, if the shunt machine is adjusted to 
 share the load properly at full load, then if the load 
 decreases, the shunt machine's voltage rises while that 
 of the compound-wound machine falls, if the latter is at 
 all overcompounded, and should the external circuit be 
 opened, leaving the two machines directly connected to 
 each other, it is very possible that the shunt machine will 
 reverse the compound-wound machine, and run it as a 
 motor. The behavior under these circumstances is very 
 peculiar if the two machines are of about the same 
 
THE COMPOUND-WOUND DYNAMO. 341 
 
 capacity: the shunt machine starts off as a generator 
 with nothing on its circuit except the compound-wound 
 machine's field and armature. The current in the com- 
 pound-wound machine has been reversed, and hence the 
 shunt and series fields oppose and neutralize each other, 
 thus destroying all C. E. M. F. that the machine may at 
 first produce. The shunt machine, now acting through 
 a short circuit, drops its field, and with it its generating 
 power. The compound-wound machine now takes a 
 turn as generator, and is in turn short circuited through 
 the shunt machine's armature. By virtue of its series 
 windings the compound-wound machine cannot lose its 
 field, and unless a belt or circuit-breaker gives way, will 
 drive the shunt machine as a motor and at a high rate 
 of speed. Now the shunt machine runs in the same 
 direction, with given connections, whether as generator 
 or motor, and as the speed rises, so does the C. E. M. F., 
 causing the armature current to decrease, and with it the 
 series field strength on the compound-wound machine 
 till the latter is no longer able to maintain its voltage, 
 and is again reversed and run as a motor. The same 
 cycle could be repeated, if circuit, belts, and engines 
 remain intact. 
 
 With the snunt machine separately excited it cannot 
 lose its field at any stage, and the behavior is somewhat 
 modified; at the point where the compound-wound 
 machine is run as a motor, the current in its series field, 
 being reversed, soon overpowers the shunt field and 
 reverses its polarity, when the two machines now run 
 in series as generators, and something has to give way 
 or burn out. Just before the compound-wound machine's 
 polarity is reversed, the C. E. M. F. decreases and 
 
342 TESTING OF DYNAMOS AND MOTORS. 
 
 becomes zero. Until this point is reached it has been 
 running as a shunt motor, and hence in the same direction 
 as the engine; at the moment of the reversal, however, 
 the motor action of the armature reverses the direction 
 of rotation, and at this point a belt is almost sure to 
 break or fly off. Even the loss of a belt does not 
 permanently relieve matters (except in special cases 
 where the belt drives a loss supplier), for the two arma- 
 tures are still electrically connected, and very curious 
 demonstrations sometimes take place, especially with 
 heavy armatures of great inertia. Some of these will 
 be dwelt upon in treating of generator tests. 
 
 Compound-wound machines may be run in series as 
 well as multiple, and then require no equalizer. Like 
 other machines they are placed in series by connecting 
 any two terminals that will enable the machines' combined 
 voltages to be read at the remaining terminals. When 
 any number of machines are in series the same current 
 passes through all, so that any load variations must be 
 due to variations in voltage, and these in turn are due 
 to speed variations incidental to belt slipping and 
 inequalities in driving of isolated engines. The best 
 regulation is secured by placing the shunt fields in series,, 
 and exciting them with the entire voltage of the set; in 
 this way any speed variation in one machine affects the 
 fields of all, so that the effect is distributed. 
 
 The nicety of regulation is lost when compound-wound 
 machine and shunt or separately excited machines are 
 run in series, for while the compound-wound machine 
 may keep the voltage constant at its own terminals, it 
 cannot compensate for the losses in the machines with 
 which it is in series, unless especially overcompounded 
 
THE COMPOUND-WOUND DYNAMO. 343 
 
 to do so. In general, therefore, compound-wound 
 machines are not run in series with other machines 
 except in testing-room practice, and here it is either for 
 special reasons, or through force of circumstances. It is 
 well, however, to know that it is done, for though one 
 may never be called upon to run such a test, it may be 
 convenient to adopt some feature of it. 
 
 A station employing only compound-wound machines 
 may be regarded as a single compound-wound machine 
 of variable output, and may be put to service on the same 
 mains with another similar station, provided an equalizer 
 is placed between the stations precisely as in the case of 
 two machines, for otherwise first one station and then 
 the other will take the greater load. There is, however, 
 a difference in degree, though not in kind, between the 
 case of two stations and that of two dynamos. Running 
 two stations in multiple would be illustrated by dividing 
 the dynamos of one large station into two sets, all those 
 of a set being connected by an equalizer, but each set 
 being kept distinct from the other. The more dynamos 
 there are in each set, the smaller will be the percentage 
 variation of load per machine for any given change in the 
 station output. Since it is the per cent, change which 
 produces a change in the voltage, and the subsequent 
 unbalancing of the load, it follows that the two sets could 
 be run in multiple without an equalizer more satisfacto- 
 rily than two single dynamos could be. Two distant sta- 
 tions would work more smoothly together in proportion 
 as the total output increased, because the greater the 
 output the less percentage of it is any given variation, 
 and hence the less the effect upon each machine. The 
 fewer the dynamos in each set, the more marked the 
 
344 TESTING OF DYNAMOS AND MOTORS. 
 
 change, and a limit is soon reached where even with hand 
 regulation of rheostats it is impossible to keep in the 
 individual machine circuit-breakers. We specify individ- 
 ual machine circuit-breakers because the latter can be 
 disposed in either of two ways: each machine may have 
 a circuit-breaker, or there may be a circuit-breaker set at 
 their combined output, and situated in the bus-bar com- 
 mon to all the machines, in the first case, as soon as a 
 machine for any reason takes undue load, its own breaker 
 responds; in the second case, even if the equalizer is in 
 use, the machines are not protected, and are free to 
 reverse each other should there be any great discrepancy 
 in their voltages, and the external load becomes greatly 
 reduced. This is because where the external load is 
 great the external resistance is comparative!^ low and 
 offers an easy path to the output of the machines in 
 common; but if the load gets small, or, to take an ex- 
 treme but common case, if the load becomes zero, as it 
 does when either the consumers cease to use power, 
 or the station circuit-breaker goes out, there is no 
 common outside circuit for the machine's currents to 
 concur in, and they are left to oppose each other in the 
 local circuits of the machines themselves. This point 
 and its bearing on station practice is considered else- 
 where. The more highly compounded the dynamos 
 are, the more violent the demonstration upon trying to 
 run them without an equalizer, because so much the more 
 rapidly does a machine's voltage run up when the current, 
 and hence the series field strength, increases. However, 
 instances could be cited where two stations have been 
 successfully run in multiple without an equalizer, but in 
 these cases the stations were several miles apart, and the 
 
THE COMPOUND-WOUND DYNAMO. 345 
 
 comparatively high resistance of the overhead work and 
 ground return made the feat practicable by narrowing 
 the limits between which fluctuations could take place. 
 To see this more clearly, suppose the two stations to be 
 10 miles apart, and that the load, contributed to by both 
 in common, is midway between them; let the line resist- 
 ance between the two stations be, say, 2 ohms. Now 
 since any difference in load taken by the two similarly 
 constructed stations is due to a difference in their 
 E. M. Fs., any one station will take its maximum load 
 when this difference in E. M. F. is greatest, and this 
 will be the case when the E. M. F. of the opposing 
 station is zero. Call one station A and the other />', and 
 suppose y/s effective voltage upon the line to become 
 zero. This condition can be obtained in either of two 
 ways: i. By opening ^'s line switch and cutting B out 
 of circuit; 2. In event of the driving power of B be- 
 coming suddenly disabled, as by bursting a steam pipe, 
 throwing a belt, or breaking a shaft or turbine, thereby 
 leaving /?'s dynamos inactive electrically, but directly 
 across the line as a short circuit. In the first case the 
 maximum current which A can send depends upon R, 
 the resistance in the path of the natural load plus the 
 line resistance from A to R (in this case i ohm), and 
 even if R becomes short circuited the current cannot 
 exceed 
 
 600 
 
 - = 600 amperes, 
 
 where 600 is A's E. M. F. and i ohm = R. In the 
 second case, where we suppose an accident to leave B a 
 short circuit across the line, the maximum current which 
 
346 TESTING OF DYNAMOS AND MOTORS. 
 
 A can be called upon to generate is, if we neglect the 
 natural load taken by the devices midway between the 
 two stations, 
 
 E, 600 
 
 = - - = 300 amperes. 
 
 We see, then, the limiting effect which the line resistance 
 has upon the current under these extreme conditions. It 
 is this effect which, as it were, cushions the variations 
 due to working inequalities, and makes parallel working 
 of distant stations practicable without the equalizer. 
 
 All direct current generators of whatever kind may be 
 run in series with each other, any difference in voltage 
 making no difference in this regard. A 500 volt street 
 railway generator could be connected in series with 
 a Daniell cell, to give 501 volts without injury to 
 either. The question of the possibility of series running 
 is a question merely of direction of voltage, so that 
 running two 500 volt machines in series is, in principle, 
 the same if one of them be replaced by a Daniell 
 cell. The practical drawback to utilizing such a com- 
 bination as a source of current is the discrepancy in 
 current capacity between the two devices; it is even 
 unwise to run in series machines differing greatly in 
 current capacity, lest the one of lower capacity be 
 dangerously overloaded. In the effort to gain other 
 ends, this point is apt to be overlooked, especially in 
 testing rooms where so many armatures of the same out- 
 put but of different E. M. Fs. are run in the same 
 separately excited frames. This error is well illustrated 
 in outside practice by the superintendent operating a 
 three-wire lighting plant who returned to the factory 
 
THE COMPOUND-WOUND DYNAMO. 347 
 
 a 125 volt ioo KW armature to be rewound; the one 
 returned to him was wound for 250 volts, but in 
 general appearance was the same as the old one. The 
 state of affairs existing on the line as soon as the 250 
 volt armature running in 125 volt fields joined forces 
 with its 125 volt mate, would have been deplorable had 
 the line switch been thrown in, but a smoking rheostat 
 attracted attention, and excited suspicion, and upon 
 shutting down some quick eye noticed that the new com- 
 mutator had more bars than the old one. 
 
 The practice of running storage batteries and dyna- 
 mos together is not without illustration, especially in 
 European power and lighting station practice, where 
 batteries carry much of the day load, and also the peak 
 of the load when running at full capacity; in this way the 
 machine capacity required in the station is reduced, and 
 a higher average load secured, so that the station runs 
 at greater efficiency. The factors modifying efficiency 
 are considered elsewhere in this book, but as a natural 
 sequence to the above assertion that "the station runs 
 at greater efficiency," we will say it is because, in all 
 devices for converting or transmitting energy, there are 
 certain losses due to radiation of heat, condensation of 
 steam, friction of moving parts, resistance of electrical 
 conductors, and what not. If a station runs without 
 delivering any work to consumers, the internal losses 
 constitute the entire station load, there is no output to 
 make returns, and the station efficiency is zero. If on 
 the other hand the station runs at full load, the losses 
 increase, it is true, but not nearly in the proportion that 
 the useful output which is being paid for does. Again, 
 if a station is equipped to deliver 1,000 horse power to 
 
34^ TESTING OF DYNAMOS AND MOTORS. 
 
 consumers, and it is called upon to deliver only 500 horse 
 power, it means that the station is twice too large, and 
 that half the money invested might better be elsewhere 
 drawing interest. 
 
 The object sought in series running is generally a 
 higher voltage than is obtainable from a single machine, 
 and cases are on record where the voltage on arc cir- 
 cuits has been run up to 6,000 or 7,000 volts. The rea- 
 son such voltages are not attainable on single machines 
 is that the insulation will not stand it, especially that of 
 the commutator. By connecting several machines in 
 series, each machine's insulation is subjected only to the 
 potential difference Between that machine's terminals, 
 unless one side of the line becomes grounded, then the 
 machine nearest the other side the line is subjected to 
 the full voltage of the system, if the frames of all the 
 machines are metallically connected, as they would be on 
 an iron floor. But this is seldom the case, and the result 
 of a ground on any machine is simply to weaken the 
 breaking down insulation of that machine. If with any 
 number of machines in series it is desired to cut one or 
 more out of service, it is easily done by reducing the 
 machine's field to zero. On shunt machines this is done 
 by opening the shunt field circuit by means of the 
 rheostat and a switch; on series machines by short 
 circuiting the series field. In either case there is no 
 danger of the machine's running as a motor or giving any 
 trouble. Where a shunt- and a compound-wound dynamo 
 run in series, and it is desired to remove one from ser- 
 vice, it is best to remove the shunt machine, for although 
 the compound-wound machine's shunt field may be open, 
 the series field still remains, and the dynamo generates 
 
THE COMPOUND-WOUND DYNAMO. 349 
 
 until the series coils are short circuited. In cutting out 
 a machine gradually weaken its field before breaking it, 
 because aside from injury liable to result from the 
 inductive discharge upon suddenly breaking a field, very 
 curious and inconvenient phenomena arise if the circuit 
 contains a shunt- or compound-wound motor of great 
 inertia. These phenomena are detailed in Chapter IX. 
 The reason that weakening the field of a machine in series 
 with others does not make it liable to be run as a motor, 
 is as follows: so long as there is any field of the same 
 polarity as at full load, the direction of the E. M. F. 
 generated by the armature remains unchanged, hence 
 concurs with and contributes to the line E. M. F. As 
 soon as the field is broken the armature may generate a 
 slight E. M. F. in virtue of the residual field, or in virtue 
 of the slight influence which the armature current may 
 have upon the pole pieces. These two effects are 
 opposed and tend to neutralize each other, so neglecting 
 their resultant, we must consider the moving armature 
 to be electrically inert, and to act purely as any other 
 ohmic resistance in circuit, and to consume energy in 
 the form of heat. Since this energy is consumed as 
 heat, it is unavailable for turning the armature as a 
 motor, and is so small an amount that it could not do so 
 were it available. To run the machine as a motor, a cer- 
 tain amount of electrical energy must, within the machine 
 itself, be transformed into the mechanical energy of rota- 
 tion, and to secure this transformation there must be a 
 greater potential difference at the terminals than that 
 due solely to ohmic resistance. The energy expended 
 in any part of a circuit is equal to the product of the cur- 
 rent flowing, by the potential difference at the terminals; 
 
350 TESTING OF DYNAMOS AND MOTORS. 
 
 if this potential difference is entirely due to ohmic resist- 
 ance, the energy is expended entirely as heat: if, how- 
 ever, part of it be due to the presence of a C. E. M. F., 
 a portion (equal to / x C. E. M. F. ) will, under proper 
 conditions, be transformed into mechanical energy. The 
 conditions are that the armature be free to mo.ve in a 
 magnetic field. But, the reader will say, the last condi- 
 tions are fulfilled when the armature runs as a generator; 
 true, it is free to turn and has a field to turn in, but the 
 .limiting condition is this: the machine's polarity must 
 oppose that of the machine or line with which it is run- 
 ning; its E. M. F. then becomes a C. E. M. F., and only 
 then can the machine act as a motor, because, to retro- 
 .spect a little, we can state that according as a machine 
 runs as a dynamo or a motor, so will the relation exist- 
 ing between motion and repulsion vary also. When an 
 engine-driven armature turns in a magnetic field and is 
 free to generate current, there acts between this current 
 and the magnetic field an attractive or repulsive force, as 
 we may choose to call it, which resists the effort to turn 
 the armature, and it is turning the armature around 
 against this force that causes the engine to do work; in 
 this case the repulsion is due to the motion. When 
 a current is sent through an armature standing in a mag- 
 netic field, the force acting between the current's lines 
 of force and those of the field causes the armature to 
 turn; in this case the motion is due to the repulsion, and 
 we have the motor. As soon as the motor armature 
 begins to turn, its conductors cut the field's lines of force 
 and generate an E. M. F. opposed to that which causes 
 the armature to turn, and is hence called a C. E. M. F. 
 (That this E. M. F. is opposed, why it is opposed, and 
 
THE COMPOUND-WOUND DYNAMO. 351 
 
 that an armature constructed to have no C. E. M. F. 
 would be incapable of motion is taken up in another 
 chapter.) Therefore, if by means of a steam engine we 
 cause a machine to generate an E. M. F. which it is 
 desired to utilize as a C. E. M. F., the polarities of 
 the machine in question and that of the line or other 
 machine must be opposed; otherwise the voltages con- 
 cur in direction, and the machine will take up its work as 
 a generator. 
 
 In running dynamos in series it often happens that 
 they are run from different engines, or that one has a 
 smaller belt than is intended for it; under these circum- 
 stances it is possible for one engine to become over- 
 loaded, or for undue tension on the smaller belt to cause 
 it to slip. To afford relief, it is only necessary to 
 weaken the field on the machine in question; since the 
 current is the same in all the machines, the load on each 
 is proportional to its E. M. F. , and hence decreasing this 
 will decrease the load. The ease with which machines 
 of widely varying voltage can be run in series facilitates 
 the testing of dynamos of low voltage but large current 
 capacity. In this case, the two machines should be of 
 about the same current capacity, and the low voltage 
 machine may be run as a motor to share the load with 
 the lamp bank or water box, or it may, as a dynamo in 
 series with the larger machine, furnish a small voltage 
 and help urge the required current through the lamp 
 bank, water box, or motor, as the case may be. 
 
 TEST I. Eight-volt, five-ampere Shunt Machine. Low 
 voltage generators for electroplating purposes give from 
 5 to 25 volts, and from 10 amperes upward; machines 
 which are to be used as " boosters " are also of low voltage 
 
352 TESTING OF DYNAMOS AND MOTORS. 
 
 and high amperage. In any case it is difficult to get a 
 resistance low enough to admit of the low voltage send- 
 ing the required current through it, and at the same time 
 of sufficient capacity to carry the current. In the case 
 of very large current capacities the voltage is even lower, 
 and the test is conducted by stringing copper cables 
 from brush to brush. When neither lamp bank nor 
 rheostat is available, it is customary to utilize a third 
 machine as motor, and to have it return its energy of 
 rotation to the original motive power. 
 (This method of testing, known as 
 "pumping back," is of wide applica- 
 tion, is an important factor in testing 
 room economy, and is considered by 
 itself elsewhere.) In the above test 
 of a low voltage machine we will 
 FIG L ii3 suppose that the machine is to be 
 
 run as a generator in series with 
 an auxiliary dynamo of equal current capacity but 
 of much higher E. M. F., so that their combined 
 E. M. Fs. will be sufficient to urge the required current 
 through the resistance of a lamp bank. Low voltage 
 machines are generally separately excited, so they can 
 readily be rendered active or idle by making or breaking 
 their field. Let us suppose that we have an 8 volt 25 
 ampere machine to run for 2 hours, and that the machine 
 to be used as an auxiliary is a 125 volt 48 ampere 
 machine. As a preliminary, the two machines are run 
 with a light field and gotten in series without closing 
 switch K, of Fig. 113; where A is the auxiliary machine, 
 B the one under test, C and D machines or lines to be 
 used as exciters, and L the lamp bank filled with 85 volt 
 
THE COMPOUND-WOUND DYNAMO. 353 
 
 lamps. It is not absolutely essential that cither A or B 
 be separately excited, for the test has been frequently 
 run with both machines self-excited, when exciters were 
 not available, but all testers have a preference for sepa- 
 rate excitation, because it admits of such easy control, 
 and there is no danger of a machine losing its field by 
 increasing its field circuit resistance. In the diagram 
 F is A's field excited from C, and F\ J3's field excited 
 from D, and containing a switch at K ' . The first step is 
 to get the two machines in series; putting a light field 
 on both machines, read, by means of a \ f oilmeter, the 
 voltage from i to 2, and that from 3 to 4; the voltage 
 from i to 4 should be their sum; if it proves to be their 
 difference it shows the machines to be in opposition, and 
 to place them in series the polarity of one of them must 
 be reversed. In the above case this is readily done by 
 reversing the field of one machine, but if both A and B 
 are self-exciting, it will be necessary to reverse their 
 line terminals; because if a self-exciting dynamo have 
 either its field or armature reversed, it will, for reasons 
 to be seen later, refuse to generate. Reversing the line 
 terminals reverses both field and armature, and preserves 
 their relation to each other. 
 This can be seen from Fig. 114, 
 where if i is put where 2 is, and 
 2 where i is, both A and F are 
 reversed. The next step is to 
 break .Z?'s field, close A", and 
 
 by means of A alone get the required current (25 
 amperes) on the lamp bank, so as to get an approxi- 
 mate idea of where A's rheostat is to rest, and of 
 the condition of the lamp bank, and how it must be 
 
354 TESTING OF DYNAMOS AND MOTORS. 
 
 plugged; next, without disturbing the position of A's 
 rheostat, throw off the load by opening A", and ob- 
 serve the voltage on open circuit; say it goes up to 90 
 volts; we know then that in order to get 25 amperes 
 through Z, plugged as it is, the combined open circuit 
 voltage of the two machines must be 90 volts. This data 
 once secured, the field on A can be weakened a little, K' 
 closed, the voltmeter placed across B, K closed, and A's 
 and 's field rheostats varied simultaneously until 's 
 voltage is 8 and the current is 25 amperes, the conditions 
 sought. 
 
 In case the low volt machine is series-wound and self- 
 exciting no adjustment of its terminal voltage can be 
 made unless it is too high, when an improvised shunt 
 can be used to reduce it; the current is then the only 
 quantity to be watched. In any case, as a last resort, 
 the size of the pulley can be changed to give a speed 
 that will meet the requirements of the voltage. Varia- 
 tions can be gotten within certain limits, by rocking the 
 brushes, forward to raise and backward to lower the 
 voltage; such a variation being due to alteration- in 
 armature reaction. 
 
 In using a high voltage auxiliary machine in connec- 
 tion with a lamp bank, or several banks in series, care 
 must be exercised in plugging the banks, or they may 
 suffer injury. Lamp banks are generally arranged so 
 that they can either be connected in series, and the 
 lamps plugged in multiple, or connected in multiple to 
 plug the lamps in series. In plugging or unplugging 
 banks connected in series, keep as far as possible the 
 same number of burning lamps in each bank, for if lamps 
 are cut out of one bank faster than out of another in 
 
THE COMPOUND-WOUND DYNAMO. 355 
 
 series with it, the resistance of the bank being increased 
 while the current is not proportionately decreased, the 
 fall of potential through the remaining lamps (equal to 
 / R) is increased above normal value, and may burn 
 them out. In other words, by removing lamps from one 
 of a series of banks the current carrying capacity of that 
 part of the circuit becomes too small to carry the cur- 
 rent, and is burned out. As an extreme case, suppose 
 that in one of three banks of TOO lamps each, 50 have 
 been cut out; if we call the resistance of each full bank 
 r, and the total E. M. F. applied , we have when all 
 the banks are full, 
 
 After cutting out half the lamps in one bank its resist- 
 ance becomes 2 r, all the rest remaining the same as 
 before: then, 
 
 2 r + r + r 4 r 
 In the first case each bank gets a potential difference of 
 
 A, 
 3 
 
 because each one's resistance is one-third of the total. 
 In the second case the total is 4 r, and that of the 50- 
 lamp bank 2 r, or one-half the total resistance. This. 
 bank therefore is subjected to one-half E, which is 
 too much for the lamps. For the same reason, all the 
 rows of a bank should be full of lamps. The first indi- 
 cation of missing lamps is when those remaining in the 
 
356 TESTING OF DYNAMOS AND MOTORS. 
 
 row shine out more brightly than the rest. There are 
 many types of lamp bank; one of those most used is 
 known as the three-wire lamp bank. It can be plugged 
 in any series multiple combination within its capacity. 
 
 TEST II. Motor Generator Test, Engine as Loss Supplier. 
 In the whole range of testing room expenence there is 
 no test so instructive as the ''motor-generator" test 
 alluded to above. From a scientific standpoint it is one 
 of the neatest applications of theory to practice that 
 can be found, and admirably illustrates the flexibility 
 of the electrical system of energy transmission. From 
 an economic standpoint, the coal dealer frowns in evi- 
 dence of its effectiveness. The method is an elabora- 
 tion of Drs. J. and E. Hopkison's " Efficiency Test," 
 and though carried out in several ways differing in 
 detail, is in principle as follows: A dynamo belted to an 
 engine is electrically connected to a motor also belted 
 to the engine; the electrical energy of the dynamo goes 
 into the motor, there to be converted into the mechani- 
 cal energy of rotation, which in turn helps the engine to 
 turn the dynamo. If there were no losses due to heat, 
 friction, belt tension, etc., the motor and dynamo, once 
 started, would continue to run each other without ?.ny 
 help from the engine; but we know that the dynamo can 
 deliver only a part of its own energy to the motor, and 
 that the motor can return only a part of what it receives 
 to the dynamo; in both machines, in the engine itself, 
 and in the shafting there is energy lost, and it is this 
 lost energy which the engine or auxiliary motor or 
 dynamo, as the case may be, must supply in order to 
 keep the system turning. 
 
 In large factories where many machines of large out- 
 
THE COMPOUND-WOUND DYNAMO. 
 
 357 
 
 put are tested daily, this most economical method is 
 adopted, and in the following pages we give the test 
 with the various modifications dictated by circumstances. 
 The test in its simplest form is conducted on two elec- 
 trically connected machines belted to an engine, which 
 is to supply the loss. In this case, the same amount of 
 current flows through both machines, and they must be 
 
 FIG. 115. 
 
 of very nearly the same output, or the full current load 
 of one will overload the other. Let the test be of two 
 200 KW 500 volt machines, shunt wound, and belted to 
 a 75 KW engine to supply the loss! In Fig. 115, Pis 
 the engine pulley, P\ P\ are the countershaft pulleys, 
 to which are belted armatures A and B, by means of their 
 respective pulleys, 6" and S" . F and R are A's field 
 and rheostat, F' and R' those of B. G is an ammeter, 
 and K a switch between the two machines. It can be 
 seen that one of A's brushes connects directly to one of 
 It's through cable i, 2, 3, and that the two remaining 
 
358 TESTING OF DYNAMOS AND MOTORS. 
 
 brushes connect through cable 4, 5, 6, including meter 
 G and switch K, so that the circuit through the two 
 machines is i, 2, 3, 4, 5, 6. The field of each machine 
 is connected to its brushes or their equivalent, and each 
 can generate its own field, even though K be open. 
 A's field circuit is i, ^?, F, 6; and B's, 3, R'< F\ 4. 
 After completing connections, everything is inspected to 
 see that pulleys are tight, brushes are set, bearings are 
 oiled, and that connections are proper for the given 
 direction of rotation. The engine is then turned over 
 slowly to facilitate lining up the pulleys, and getting the 
 right belt tension. It is then brought up to speed, and 
 if the armature pulleys have been properly selected they 
 too will run at their rated speed. The next step is to 
 let down one machine's brushes, close its field circuit, 
 get a field, and by means of a voltmeter and rheostat, R, 
 adjust the voltage to 500. We then hold the volt lines 
 across the rheostat terminals, and get what is called 
 the drop on the box. This data tells the engineer if the 
 quality of the iron is what it should be, because if the 
 iron is poor and everything else is all right, it will take 
 more current in the field circuit to give the required 
 voltage; to get more field current the box resistance 
 must be lessened, and this lessens the drop on the box. 
 We now raise .Z?'s brushes and close K\ this allows A to 
 charge 's field, so that when B generates, its voltage 
 will oppose that of A. K\ now opened, and It's brushes 
 lowered, when, if everything is all right, B will pick up 
 and support its own field; its voltage also can now be 
 adjusted to 500, when the following state of affairs will 
 exist : A and B are generating equal but opposite voltages, 
 so that a voltmeter across K will not be deflected; when 
 
THE COMPOUND-WOUND DYNAMO. 359 
 
 this is the case, K may be closed, and no current will 
 flow between A and B, because they have equal but oppo- 
 site tendencies to send current into each other. A'closed, 
 the system is ready for load. One machine, say A (it is 
 immaterial which when testing shunt machines), is selected 
 to be the motor; B, to be the generator. Placing a man 
 at ^Ts brushes, one at B's brushes, and another to tighten 
 the belt, if necessary, as the load goes on, As field is 
 gradually weakened. B, against A's diminished E. M. F., 
 now sends current through A and runs it as a motor; its 
 former E. M. F. becoming its C. E. M. F. As load gees 
 on, A's brushes must be brought backward to stop their 
 sparking, and ^?'s forward. (The effect of careless hand- 
 ling of brushes in a motor-generator test is detailed else- 
 where.) With the voltmeter across />"s terminals, .Z>'s 
 voltage is adjusted, by means of ./?', to 500 volts. The 
 speed is checked up, and the full load drop on the box 
 taken. The machines are then run half the stipulated time 
 of the test, when they are changed over. A becoming 
 the generator and B the motor. To do this, A* field is 
 very gradually strengthened, .Z?'s weakened, A's brushes 
 rocked forward and j's backward, till the ammeter, G t 
 indicates zero; continuing the operations carefully the 
 needle will rise to full load again. 
 
 The first few tests are free from details concerning 
 data, precautions, troubles, symptoms, and remedies, 
 because it is thought best not to burden or side track the 
 reader's attention, till his conception is clear as to their 
 first principles. These matters will be entered into more 
 fully later. 
 
 TEST III. Motor Generator Test with Lamp Bank.^t 
 test just considered comprised two machines of equal volt- 
 
3<5 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 250V 
 
 age and equal current capacity. We now take up a test in- 
 volving two machines of approximately the same current 
 capacity but of very different E. M. Fs. , the ' ' loss " to be 
 supplied by an engine to which both machines are belted. 
 The two machines can of course be run on a lamp bank or 
 water box, but aside from the economic advantage of hav- 
 ing one machine, as a motor, return its energy to the sys- 
 tem, we shall suppose, as is often the case, that the engine 
 
 available is already nearly 
 loaded, or is too small to 
 support either machine 
 run as a dead load on a 
 lamp bank. In this case 
 we will use a lamp bank, 
 in series with the smaller 
 machine, to absorb the 
 excess of voltage of the 
 larger one. Fig. 116 gives 
 
 the connections. A is the, say, 125 volt machine in 
 series with lamp bank Z, filled with no or 50 volt 
 lamps. B is a 250 volt machine, which must be the 
 generator in this case, for we could never, without 
 separately exciting it, get its voltage enough below 
 that of A to use it as a motor. From the diagram 
 it is seen that, as far as connections go, the two 
 machines and the lamp bank are in series; but inas- 
 much as the E. M. Fs. of A and B are opposed in the 
 test, the question arises: Are A and B in series? Good 
 authorities differ on this point. Kapp in his " Electrical 
 Transmission of Energy " refers to a motor and dynamo 
 being in series, and using this as a basis, the matter may 
 as be reasoned out thus: For two dynamos to be in series 
 
THE COMPOUND-WOUND DYNAMO. 361 
 
 their E. M. Fs. must concur, and for two motors their 
 C. E. M. Fs. must concur. In the test under considera- 
 tion, both machines are brought up to full voltage before 
 the circuit between them is closed. They are therefore 
 primarily dynamos, although they generate only enough 
 current to excite their own fields. If, now, we suppose 
 their E. M. Fs. to be equal, so that the voltmeter across 
 K reads zero, when the switch is closed the machines are 
 still dynamos, and are in multiple; if the field of one be 
 weakened and it becomes a motor, its nature is changed 
 and the two machines, as in the above case, are in series. 
 To start the test, the engine is brought up to speed 
 and the speed of the two machines taken to insure that 
 the right pulleys have been selected. The field is now 
 put on B, and its voltage reduced so as not to injure A's 
 fields when charging them. To charge ^'s field, first 
 raise its brushes and close its headboard switch if it 
 has one. Machines do not always have their switches 
 at the time of testing, but if in this case both of them 
 have, one of the switches can take the place of A", the 
 other remaining permanently closed. A's brushes raised, 
 K is closed and L short circuited by means of plug K', 
 plug S, of course, being in to close the circuit. Particu- 
 lar care must be exercised to get the two machines in 
 opposition, and not in series as generators, for in the 
 latter case, should K and 6" be closed and K' open, the 
 lamps would be destroyed, as many a novice can testify. 
 With K, K', and S closed, and A and B in series, a short 
 circuit would follow and the belts would fly off. Having 
 charged A's field, K is opened, K is opened also, and 
 A's brushes lowered. A should now generate its own 
 field. Next draw out S, and close K', S then becomes 
 
362 TESTING OF DYNAMOS AND MOTORS. 
 
 the only point at which the circuit between A and B is 
 broken, and the E. M. F. across it should be the differ- 
 ence between the E. M. Fs. of A and B. Should it be 
 their sum, it shows the machines to be in series, and A's 
 field must be recharged. The writers do not recall an 
 instance of a machine's reversing its polarity immediately 
 after charging, but as a tester sometimes forgets to 
 charge, leaving the machine to pick up as it may, the 
 test across K is a necessary one. Always be careful to 
 open K or S before lowering the brushes, and to remove 
 K before completing the circuit again, for either over- 
 sight will result in a short circuit. 
 
 With the machines of opposing voltage, that on B is 
 brought to 250, that on A to 125, and the bank plugged 
 in. If L contains 50 volt lamps they must be plugged 
 3 in series, and as many in multiple as is necessary to 
 put on the load keeping A's voltage always at 125. If 
 B is a shunt machine, its voltage will fall as the load 
 goes on, and must be brought up by cutting resistance 
 out of R'. 
 
 The situation now is as follows: As soon as L is 
 plugged in, B sends current through it, and through A, 
 .causing A to run as a motor. The high resistance of 
 the lamps takes up the 125 volts difference between the 
 E. M. Fs. of A and B. At first only a small current 
 flows, but it is gradually increased by plugging in more 
 lamps in multiple till the full load is reached. Both A 
 and B are now loaded; A working as a motor and assist- 
 ing the engine, and B driving A and expending the 
 balance of its load on L. If A and B are shunt machines, 
 after the heat test is run, and voltage, current, and speed 
 are right, the drop on R and R is taken, so the engineer 
 
THE COMPOUND-WOUND DYNAMO. 363 
 
 may know if sufficient box range is allowed for the 
 difference due to variation in summer and winter tem- 
 peratures. 
 
 The engine as a loss supplier is not a very flexible 
 means of speed variation, so if it happens that there are 
 no pulleys to give the proper dynamo speed, or that on 
 account of overload, low steam pressure, or what not, 
 the engine speed is below normal, the dynamo E. M. F. 
 will have to be approximated as follow r s: Let us assume 
 that the proper dynamo speed is 1,000 revolutions per 
 minute (r. p. m. ), and that its rated E. M. F. is 250 
 volts. Then, on the assumption that the E. M. F. varies 
 directly as the speed, we have, 
 
 I.OOO 
 
 = 4; 
 
 250 
 
 4 revolutions per volt; /'. e., if 250 volts will require 1,000 
 revolutions, i volt will require 4; or expressed differently, 
 1/4 volt per revolution. Let the actual speed be 800 
 200 too low: for a fall of 200 revolutions we would have 
 a fall of 200 x 1/4 = 50 volts; so that the dynamo in 
 this case would be tested at 200 volts. The assumption 
 upon which this, practice is based is only approximate, 
 and the only true way to apply a correction is to experi- 
 mentally determine what difference in voltage is caused 
 by a difference of one revolution, both on open circuit, 
 and with normal current flowing. 
 
 During the heat run, when no data are being taken, as 
 much as possible of the load should be put onto the 
 motor, because the larger part of this is returned to the 
 system, while that absorbed by L is wasted. In order to 
 increase the motor's share of the load, its C. E. M. F. 
 
364 TESTING OF DYNAMOS AND MOTORS. 
 
 is raised by strengthening the field. The immediate 
 effect of strengthening the motor field is to reduce the 
 total load of the system, but to give a larger proportion 
 to the motor. The current then can be restored to its 
 proper value by plugging in more lamps in multiple. 
 Should the capacity of L be so limited as not to admit 
 of full load with A's voltage at 125, it can be lowered, 
 care being had that the life of the bank is not endan- 
 gered, because as A's C. E. M. F. is lowered so is the 
 amount of impressed voltage which drops across it but 
 the drop across L increases. The reader must note this 
 point: When one machine is running directly back on 
 another, /. e., without the intervention of a lamp bank or 
 other resistance, to increase the load on the motor, its 
 field is weakened, whereas, if a bank intervenes the field 
 must be strengthened. In the first case, the end in view 
 is to increase the total energy of the circuit, and this we 
 do by lowering the motor's C. E. M. F., and thereby 
 decrease the effective resistance in circuit. Were we to 
 strengthen the field it would decrease the total load. 
 In the second case, the end in view is not to increase the 
 total load of the circuit, but to increase it in one part 
 and decrease it in another, and this we do by raising the 
 effective resistance in the motor part of the circuit and 
 lowering it in the lamp bank by plugging in more lamps. 
 In the first case there is increase in the total energy of 
 the circuit, because its resistance is decreased, thereby 
 increasing the current, while the impressed E. M. F. is 
 kept the same. In the second case there is simply a 
 transference of energy from one part of the circuit to 
 another, for while we temporarily decrease the total load 
 by strengthening the motor field, the circuit resistance is 
 
THE COMPOUND-WOUND DYNAMO. 
 
 365 
 
 restored to its former value by means of the lamp bank. 
 This, test is similar somewhat to Test I, in that the two 
 machines are in series with the lamp bank, but differs 
 from it in that here one machine is a motor and the 
 other a generator. 
 
 The heat test on B done, the next step is to change 
 over and test A as a generator. It would be impossible 
 to self-excite B and reduce its voltage sufficiently to 
 admit of the machine's being run as a motor from A, 
 because the introduction of so much resistance in its 
 field circuit would cause it to drop its field; so B 
 is separately excited. This does away with the neces- 
 sity of the bank, and the test becomes the same as 
 Test II. 
 
 TEST IV. Motor-Generator Test, Three Machines. The 
 lamp bank of the above test can be very profitably 
 replaced by a second 125 volt machine, which has the 
 double advantage that twice as much energy is returned 
 to the dynamo, and that three machines may be tested at 
 once, instead of two. In 
 Fig. 117 the belt is not 
 shown, but all three ma- 
 chines are belted to the 
 same engine. (We say 
 this, because it is not in- 
 frequently the practice to 
 
 FIG. 117. 
 
 run the generator from one engine and feed its current 
 into a motor or motors attached to another engine which 
 needs help badly; this does not constitute a " motor 
 generator" test, because there is no work saved, since 
 the machines do not circulate the current between them 
 but expend it in doing work elsewhere. In a "motor 
 
366 TESTING OF DYNAMOS AND MOTORS. 
 
 generator " test, we understand that not only are two or 
 more machines being tested from the output of one, but 
 this with an expenditure of energy which is but a fraction 
 of that one's output; whereas in this case, the entire 
 output of the generator is utilized elsewhere. The prac- 
 tice is of course more economical than running the 
 generator on a lamp bank, because the energy is then 
 wasted.) In Fig. 117, A and A' are the two machines 
 the sum of whose E. M. Fs. equals that of B. F, F' t F" 
 and R, R ', R" are their fields and rheostats respectively. 
 K, as usual, is the connecting switch. With K open, A 
 and A' are gotten in series. The manner of doing this 
 seems to depend on the man that is doing it. One man 
 will let A and A pick up on their residual magnet- 
 ism regardless of polarity, and if necessary reverse the 
 armature cables connecting the two machines; another 
 man will let one machine pick up and charge the other 
 from it by raising the brushes, closing K, and the head- 
 board switches, if there are any, and sending the current 
 around through B. A and A once in series can be 
 regarded as a single 250 volt machine, and be used as 
 such to charge B'<=> fields. All machines excited, and 
 250 volts on both sides of K, the voltmeter should read 
 zero across it. Closing K, the fields on A and A are 
 gradually weakened, the usual attention being paid to 
 the brushes. By having a voltmeter across each machine, 
 the voltage on B can be kept up to 250, and equally dis- 
 tributed between A and A. 
 
 TEST V. Motor Generator Test, Machines of Different 
 Current Capacity. The problem of this test is to "run 
 back " on each other, two machines of the same voltage, 
 but of different current capacities. The test of the 
 
THE COMPOUND-WOUND DYNAMO. 367 
 
 smaller machine is the same as Test II, for there is no 
 danger in the smaller machine's maximum current injur- 
 ing the larger machine. But in testing the larger machine 
 as a generator it is necessary to 
 have an auxiliary motor of the 
 same voltage, or a lamp bank to 
 absorb all current in excess of the 
 smaller machine's full load. Fig. 
 118 gives the connections when 
 a motor is used, and Fig. 119, 
 
 those of a lamp bank. A and B are the machines under 
 test, C the auxiliary machine, A is the generator, B and 
 C the motors. To simplify the diagram, the fields and 
 rheostats are not shown. K is B's switch, A", C's. A, 
 , and C'are machines of the same voltage, say, 125, and 
 are belted to the same engine. Call A's current capac- 
 ity 480 amperes; ^'s, 320 amperes; C's, 160 amperes. 
 C's capacity can be anything exceeding the difference 
 between A's and B's. To start the test, a field is gotten 
 on A. l?s and C's fields are then charged one at a time, 
 by raising their brushes and closing their respective 
 switches. The voltages on all three machines are then 
 K . adjusted to 125 volts, and the volt- 
 
 1 ^ 1 meter should read zero across K 
 
 eO [U and K'. When this is the case, K 
 
 1 ( i J is closed and B's field weakened 
 
 FIG no until B carries a current of 320 
 
 amperes. A's voltage meanwhile 
 
 being kept at 125 volts by means of its rheostat. We 
 now read across K to insure that everything is right, 
 and if it is, close K': next weaken C's field; it will 
 take load, and unless A's voltage is kept up to 125 
 
368 TESTING OF DYNAMOS AND MOTORS. 
 
 volts B will lose some load. So it is best to keep 
 up A's voltage, and gradually, by means of C's brushes 
 and rheostat, bring its load up to 160 amperes. The 
 currents are read from two ammeters preferably placed 
 as indicated in the diagram, where one meter is in 
 the main circuit and reads A's current, while the 
 other is in ^'s circuit; their difference gives C's cur- 
 rent if it is necessary to know it. If neither meter has 
 range enough for the main circuit, they may be placed in 
 the motor circuits where their combined readings give 
 A's current. The situation of the meters must be borne 
 in mind, because if the tester labors under the impression 
 that one of the motor meters is A's meter, both A and 
 the motor will be heavily overloaded. In putting on the 
 load, the man at the brushes must keep them at the non- 
 sparking point, and the tester at the field boxes, which 
 are all placed together, must regulate the load, and when 
 the ammeter needles cease vibrating, indicating that the 
 brushes are at rest, he can adjust the load exactly. If 
 the motor man does not know his business, he is likely to 
 make trouble by either moving the brushes too much 
 at once or by moving them in the wrong direction; in 
 the latter case he may even reverse the nature of 
 the machines. The proper way to handle motor brushes 
 in a ''motor generator" test, is to let the brushes alone 
 until they show signs of sparking, then to let the spark- 
 ing keep a little ahead of the brush movement. The 
 man then knows where he is. If at no load or small load 
 any of the machines begin to spark badly, the indications 
 are that by some injudicious movement of brushes or 
 rheostat the voltage of one of the motors has been raised 
 above that of the dynamo, resulting in a reversal. Un- 
 
THE COMPOUND-WOUND DYNAMO. 369 
 
 der tl;ese circumstances it is best to pull the switches (A"- 
 and A' or A" and A') and begin again. In an emergency 
 the load can be thrown off by opening A' 2 , but one of the 
 other switches should be opened immediately after it, as 
 otherwise B and C run back on each other with possibly 
 some sparking should there happen to be much of a dif- 
 ference in their voltages. In this case the engine will 
 keep the system moving, and we have the conditions of 
 Test II. The smoothest way to shut down the whole 
 system is to shut down the engine, without touching the 
 machines. The quickest way to remove one of the 
 motors from service is to pull its switch, then if its 
 rheostat is not disturbed it can be replaced in service by 
 simply closing the switch again. The method generally 
 adopted is to strengthen the field on the motor to be re- 
 moved, until its load has fallen nearly to zero, then to 
 pull its switch. The removal of one motor will slightly 
 overload the remaining one, unless A's voltage is kept at 
 125 volts. If A's voltage is not kept at 125 volts, the re- 
 moval of one machine will influence the remaining one 
 only so far as A's terminal voltage is influenced by the 
 removal of some of the load. It amounts practically 
 to the same as pulling one motor switch, and in principle 
 is the same as removing any consumer's motor from a 
 power circuit, except that in this case there being but one 
 generator, and the two motors constituting its entire load, 
 the removal of one of them makes a perceptible differ- 
 ence in the line voltage. If A is a perfectly compounded 
 machine, tampering with the load of one of the motors 
 or even pulling the motor switch will not affect the other 
 motor, because the self-regulation of the compound- 
 wound generator keeps the terminal voltage constant. In 
 
370 TESTING OF DYNAMOS AND MOTORS. 
 
 the case of A's being compound-wound, its shunt wind- 
 ing is adjusted once for all (hot), to give normal voltage 
 on open circuit, and thereafter the series winding pre- 
 serves constancy of voltage throughout the range of 
 load. 
 
 When two motors run in multiple from one dynamo, the 
 effective resistance in circuit, /. e., opposition to current 
 flow, is of two kinds: (i) Ohmic resistance of armatures, 
 leads, cables, etc., made as low as possible and negligible 
 in this case; (2) C. E. M. F. of the motors, and this is 
 all important. The current flowing may be expressed by 
 Ohm's Law as follows: 
 
 Imp. E. M. F. - C. E. M. F. E - e 
 
 J - r _ Q- T - _ 
 
 Ohmic Resistance R 
 
 The larger e is, the less difference is there between E 
 and ^ hence the less the value of the fraction 
 
 R > 
 
 which / equals. Therefore, since / decreases as e in- 
 creases, /is inversely proportional to e. 
 
 In the case of two motors in multiple, the effective re- 
 sistance of the circuit, neglecting the ohmic resistance, 
 is the multiple " resistance " of the two C. E. M. Fs. 
 This can be worked out ia precisely the same way as 
 multiple ohmic resistance is. The "resistance" of one 
 motor's C. E. M. F. can be gotton as follows: From the 
 formula 
 
 R 
 
THE COMPOUND- WOUND DYNAMO. 37! 
 
 from which we get e = E / R; i. e., the C. E. M. F. is 
 what is left after we subtract the drop through resistance 
 R from the impressed E. M. F. Now /<", /, and R can 
 be measured with voltmeter and ammeter, whence per- 
 forming the operations indicated by the formula, we get 
 the value of e expressed as a resistance. Getting thus the 
 value of e on both motors at the given load and speed, 
 we combine them according to the law of multiple resist- 
 ances and get the effective " resistance " of the two 
 motors' C. E. M. Fs. in multiple. 
 
 For our present purpose it is only necessary to ob- 
 serve that lowering the C. E. M. F. of either or both 
 machines will increase the load, and that the greater load 
 will go on the motor of lower C. E. M. F. 
 
 To remove the load we can weaken the dynamo field 
 instead of strengthening the motor fields; if, however, 
 A's E. M. F. be at any time brought below the C. E. M. 
 F. of B and C, reversal takes place with the usual accom- 
 paniment. The demonstration is most violent when a 
 field is broken on some machine, thus depriving it of all 
 ability to oppose an inflow of current from the other 
 machines. If A's field gets broken, its E. M. F. falls 
 nearly to zero: B and C become generators, and short 
 circuit through it; if it is compound-wound, it will try to 
 reverse its direction of rotation. If B and Care shunt- 
 wound, they will drop their fields, but not before they lose 
 their belts. If compound-wound, the differential action 
 of their fields (cumulatively connected to run as motors) 
 puts a limit to their current. Should we break a field on 
 one of the motors, say B, A and C, as generators, would 
 short circuit through it, and the very characteristic howl 
 which rends the air when large units are involved, once 
 
37 2 TESTING OF DYNAMOS AND MOTORS. 
 
 heard, is never forgotten. This trouble with reversals 
 is not confined to testing rooms, but is liable to occur 
 wherever dynamos are run in multiple. All field connec- 
 tions should therefore be inspected at regular intervals, 
 for connectors or box wires are liable to rust off or jar 
 loose. 
 
 In the test above, as soon as the test on A as a genera- 
 tor is completed, C's switch can be pulled, A's voltage 
 gradually reduced, B's increased, and when the current 
 is nearly zero, the motor man signaled that reversal is 
 about to take place, so that he can bring ^'s brushes 
 backward and A's forward. The field on A is then weak- 
 ened, and A takes a load as a motor. This change-over 
 without shutting down we have learned to be practicable 
 only when shunt or separately excited machines are in- 
 volved. If any of the machines are compound-wound, the 
 system must be shut down and the series fields reversed. 
 The reasons are these: If on a shunt machine, running 
 as a dynamo, we assume a certain relation to exist 
 between the directions of current flow in field and arma- 
 ture, this relation is changed upon the machine becom- 
 ing a motor. On a series motor the relation remains the 
 same. In other words, a compound-wound machine 
 cumulatively connected and running as a motor, will, if 
 its C. E. M. F. is allowed to exceed the impressed 
 E. M. F. (that is to say, to become a generator, as B 
 does when its C. E. M. F. is made to exceed the E. M. F. 
 of A,), have its series field current reversed, though 
 the shunt field remains the same as before. The two 
 fields are then opposed, and it will be impossible to work 
 on a load, because the windings neutralize each other. 
 
 If at the signal for reversal the motor man should rock 
 
THE COMPOUND-WOUND DYNAMO. 373 
 
 his brushes the wrong way, the load will seem to hesi- 
 tate for an instant, and then will go on with a rush; 
 how much of a rush depending upon the position of 
 the brushes and the type of the machines. If A, B, and 
 C are shunt-wouno^ machines, the result will be to put 
 on some load in the original way, /. e. y before it was 
 attempted to make A the motor. If A, B, and C are 
 compound. wound, the conditions will not be the same, 
 because their series fields having been reversed, their 
 ability to generate in the original way is limited; the 
 result is a large initial flow of current, which is, how- 
 ever, soon checked by the differential action of A's field. 
 More or less sparking attends these reversals, its gravity 
 depending upon the position of the brushes. If either 
 B or C is a compound-wound machine and A is a shunt 
 machine, and some careless move makes the compound- 
 wound machine a motor when it is connected up as a 
 generator, the trouble will be more serious, for as soon as 
 current from A flows through the compound-wound 
 machine, say B, B's series field neutralizes the shunt 
 field, and a short-circuit follows. If B is separately ex- 
 cited matters are even worse, for the machine cannot drop 
 its field, whereas shunt machines sometimes do so before 
 the belt can fly off. Sometimes the compound-wound 
 machine will throw its belt trying to reverse rotation, 
 showing that not only has the series winding neutralized 
 the shunt winding, but overpowered it, thereby revers- 
 ing H's polarity. 
 
 TEST VI. Same as Test V., with Lamp jBan^.This test 
 is practically the same as Test V., excepting that a lamp 
 bank replaces the auxiliary machine C. Fig. 120 gives the 
 connections. A is the 125 volt, 480 ampere machine to be 
 
374 TESTING OF DYNAMOS AND MOTORS. 
 
 run as dynamo ; B, the 1 25 volt, 320 ampere machine to be 
 run as motor; and Z, a 125 volt lamp bank capable of 
 carrying 160 amperes; L and B are in multiple, and A 
 divides its load between them. In 
 plugging in the lamps the motor 
 must be watched and its load regu- 
 lated with its field rheostat. The 
 general precautions to be observed 
 are similar to those of Test V. 
 JT IG I2O TEST VII. Same as Test V., 
 
 with Motor. This test, a very 
 
 natural outcome of Test V., is used also where two 
 dynamos of the same E. M. F., but of different current 
 capacities, are to be tested. A third machine, C, is used 
 as motor to absorb the current from dynamos A and B. 
 The connections are the same as those of Fig. 118, and 
 the test is almost the same. The exception being that 
 when the switches are closed A's field is weakened, and 
 not those of B and C. A is now the auxiliary machine. 
 If either B or C, or both, are compound-wound, care 
 must be had .that the fields are cumulatively connected 
 as dynamos, while A must be cumulatively connected as- 
 motor. 
 
 TEST VIII. Machines of Different E. M. F. and Current 
 Capacity. This is a test that the writers have never been 
 called upon to run, but which is perfectly feasible 
 should occasion demand it. Fig. 121 shows the connec- 
 tions. A and B are machines of widely varying E. M. Fs., 
 also current carrying capacities, and two banks, L and Z', 
 are auxiliaries to the test; A is a 250 volt, 480 ampere 
 machine, B is a 125 volt, 300 ampere machine, Z is a 
 bank in series with B to absorb A's excess of voltage. 
 
THE COMPOUND-WOUND DYNAMO. 375 
 
 and Z', a bank in multiple with A to take its excess of 
 current. A, B, and L are first put on as in Test III. 
 The system then stands at the proper voltage, but with 
 only the current at A's maximum capacity 
 of 300 amperes. The remaining 180 am- 
 peres are put on L' as in Test VI. 
 
 TEST IX. Single Machine. Where 
 there is only one machine to be tested, 
 or where machines are run from engines 
 between which no connections can be 
 made, the motor-generator test is not 
 practicable, and compounding must be 
 done on a lamp bank or water box. The cost of banks 
 of such capacity as some machines would require pre- 
 cludes their use, and water boxes are found to be a 
 very cheap, simple, and satisfactory substitute. Water 
 box compounding is simpler than the motor-generator 
 test, in that it reduces the number of difficulties possible 
 to be encountered, but is not nearly so economical. In 
 the motor-generator test, the energy supplied to the sys- 
 tem is merely that necessary to overcome the mechanical, 
 electrical, and magnetic losses in the two machines, 
 whereas a water box or other ohmic resistance dissipates 
 the entire output of the machine under test. But where 
 there is but a single machine there is no way of return- 
 ing its energy to the system, so that the waste is an 
 unavoidable necessity. A box 4x5x8 feet will carry 
 the full load of a 550 volt, 500 KW generator. Its iron 
 plates, one of which is movable, the other stationary, 
 should have as much cross-section as the size of the box 
 will allow, and flexible cables must be attached to these 
 plates as permanent leads. When a box is used for the 
 
TESTING OF DYNAMOS AND MOTORS. 
 
 first time, it is well to be on the safe side, and pull the 
 plates as far apart as possible, and fill it only half full 
 of water. As a further precaution the voltage on the 
 dynamo can be reduced until we ascertain what the box 
 will do. The box terminals are connected through a 
 switch, K, to the dynamo terminals as in Fig. 122, where 
 A is the dynamo, F, its field, R, its rheostat, and Z, the 
 water box. K is a long break, double-pole spring switch. 
 Having experimentally found out 
 just about what to expect of Z, A's 
 E. M. F. is adjusted to 500, and 
 K closed. The load can then be 
 FIG. 122. worked on by putting a little salt 
 
 between the plates to improve the 
 
 conductivity of the solution, by putting in more water, or 
 pushing the plates nearer together, any of which proced- 
 ures decreases the resistance of the water column. As 
 soon as current begins to flow the series windings begin to 
 act: this raises the terminal E. M. F., which in turn aug- 
 ments the shunt field. The field rheostat once adjusted 
 must not be changed throughout the test. Care must be 
 exercised in using salt, when the plates are near together, 
 as a handful may precipitate a heavy overload. Sal- 
 ammoniac is sometimes used instead of ordinary salt, 
 but the writers object to it because its effect is much 
 greater than that of salt, but is not permanent. It is 
 admirably adapted to temporarily lower the water box 
 resistance when it is desired to make a series machine 
 pick up a field, but ordinarily it will, in inexperienced 
 hands, cause trouble. In any case, as the water heats, 
 its resistance decreases, and the current increases. As a 
 factor of safety, it is well to have a faucet for drawing 
 
THE COMPOUN 7 D-WOUND DYNAMO. 377 
 
 off some of the water should the current, with the plates 
 as far apart as possible, still be too high: the effect of 
 this is to decrease the cross-section of the water column, 
 and thereby to increase its resistance. One point easily 
 overlooked, but which should be borne in mind, is this: 
 if water from the box is allowed to run onto the ground, 
 it will cause a short circuit if there happens to be a 
 ground elsewhere on the system. For this reason the 
 water should be run into a wooden or paper bucket, and 
 then emptied from that. So, also, if decrease in load 
 requires the addition of more water, it must not be made 
 with a metal bucket, as on high voltage machines such 
 carelessness is fraught with danger. At full load the 
 amount of water, the amount of salt, and the distance 
 between the plates should be so adjusted that all load 
 regulations can be effected by means of the movable 
 plate. A little thought facilitates this. 
 
 With full current load on it will probably be found that 
 the voltage is too high: it is not absolutely necessary, 
 but if desirable, the voltage can be reduced by readjust- 
 ing the series field shunt; or better still, the German 
 silver shunt can be put on and temporarily adjusted. Its 
 permanent adjustment is not made until the machine has 
 fully heated after several hours' run, when the load is re- 
 moved, the shunt field rheostat readjusted to give 500 volts, 
 and the load again put on. The process of compounding 
 is next in order. A very complete and detailed account of 
 the compounding test is given in Test X, which has been 
 especially selected with the view of giving the reader a good 
 idea of what is to be watched in a test involving many 
 machines of different degrees of compounding and widely 
 varying outputs. Tests X and XI are typical tests. 
 
CHAPTER XI. 
 
 COMPOUNDING. 
 
 TEST X. To Run Under Full Load, Two 500 Volt, 
 500 KW Street Railway Generators, and to Supply the 
 Loss from an Auxiliary Dynamo of the Same Voltage. The 
 conditions selected are comprehensive and include lia- 
 bility to many peculiar difficulties. 
 
 The main factors are the two machines under test, and 
 the loss supplier as shown in Fig. 123. A and B are 
 
 FIG. 123. 
 
 belted or clutched together, and C, the supplier, is 
 belted to an engine. In the case where the loss is 
 supplied from an engine belted to both machines, the 
 dynamo is driven partly by the engine and partly by 
 the multipolar motor. In the present case the dynamo 
 is driven by the motor alone. In the first case the 
 current in both multipolar machines is the same; in 
 the last case the motor current exceeds that of the 
 dynamo by an amount equal fro what flows through the 
 supplier. Where C is an engine, the speed of the system 
 
COMPOUNDING. 379 
 
 depends upon the speed of the engine and the relative 
 size of the pulleys used, and it is not always practicable 
 to have pulleys of such size as to give just the speed at 
 which the machine compounds. By supplying the loss 
 electrically, almost perfect control of the speed is secured 
 by varying the E. M. F. of the supplier by means of its 
 field rheostat. Here again the great flexibility of the 
 electrical system of power supply is illustrated, and what 
 is true here is true wherever electricity competes with 
 other sources of motive power. 
 
 In cases where the engine driving the loss supplier is 
 small or doing other work, it is customary to have C 
 consist of two dynamos in series, so that should the 
 engine speed for any reason fall off, the required E. M. F. 
 can still be maintained by field regulation on C. We will 
 suppose, then, that C consists of a 500 volt, 100 kilowatt 
 dynamo in series with a 125 volt, 60 kilowatt dynamo. 
 To get the two machines in series, a voltmeter is required. 
 With both fields excited, and any two brushes or terminals 
 of the respective machines joined together, the voltmeter 
 should read, across the unconnected terminals, the sum 
 of the voltages of each machine. Should it read their 
 difference it proves them to be opposed, and one 
 machine's terminals must be reversed, or its field polarity 
 changed. The latter method is generally pursued, and 
 to facilitate the change the low volt machine is sepa- 
 rately excited, a feature whose other advantages will 
 be discussed later. The loss supplier, C, may now 
 be connected to the multipolar motor, B, by means 
 of cables. If the cable connections are made first, 
 and the two suppliers gotten in series afterward, the 
 procedure is a little different. With all brushes down 
 
380 TESTING OF DYNAMOS AND MOTORS. 
 
 and all switches except one closed, the voltmeter is 
 placed across the open switch: should the reading be 
 lower than that on one machine alone, or zero, the indi- 
 cation is that in the first case the voltages are opposed; 
 in the second case they are opposed and equal, or there 
 is an open circuit. An open circuit can be due among 
 other things to failure to press the meter push button, 
 some open switch, raised brushes, or incomplete connec- 
 tion overlooked, shellac on a commutator, lacquer on 
 the points where the voltlines are applied, or as the case 
 often is, the points of the voltlines are so oxidized, from 
 other uses, as to be non-conducting. The machines once 
 in series, and the cable connections made, it is wise to 
 reduce C's voltage, and to put a field on the motor, B, 
 as a safeguard in case a switch should fall to or be closed 
 by mistake. 
 
 During the writers' earlier experience it was customary 
 to connect the motor shunt windings in series multiple in 
 motor-generator tests of large machines, and to sepa- 
 rately excite them to their full voltage a practice based 
 upon the following reasons: i. The magnetizing power 
 of the shunt winding is much increased, the iron of the 
 fields is brought almost to saturation by these windings 
 alone, thus minimizing the effect of the series coils at 
 starting, and lessening the liability to phenomena here- 
 after to be described. 2. With the shunt fields in 
 multiple the load can be worked on more gradually, for 
 the start is made with saturated fields, and as the shunt 
 field current is weakened the iron becomes less saturated 
 and the series coils begin to take firmer hold, but at a 
 rate easily controlled. 3. The strong field admits of an 
 easy start without an excessive current. Altogether 
 
COMPOUNDING. 381 
 
 then, connecting half the spools in multiple simplifies 
 starting. In inexperienced hands the multiple connec- 
 tion had best be adopted, at least to begin with. 
 
 On general principles and for the following reasons 
 it is now the custom to connect in series all fields intended 
 to run in series under actual working conditions: i. On 
 machines of four or more fields, designed to be connected 
 in series, it is unwise to run them in series multiple, be- 
 cause unless great care is taken they are liable to be 
 injured by overheating. 2. Should a coil be defective, 
 it has three or more good ones in series with it acting as a 
 factor of safety. 3. The series connection economizes in 
 the number of field rheostats required to run the test, and 
 simplifies the cutting out of a defective rheostat without 
 shutting down the test: because, with fields in multiple, 
 the applied E. M. F. remaining the same, the total field 
 current is four times as great and the current capacity 
 of the rheostat must be increased accordingly by placing 
 more boxes in multiple; but as this decreases their resist- 
 ance, and hence the scope of variation, it is necessary to 
 put more boxes also in series. To cut out one of a num- 
 ber of boxes in series, it is only necessary to cut out its 
 resistance and cut in the same amount on another box, 
 to keep the field current constant, and then "jump" a 
 piece of wire across the terminals of the defective box. 
 In connecting boxes in series multiple it is better to 
 adopt the plan of Fig. 124 than that of Fig. 125, because 
 in the latter case any fault with one box affects to a 
 greater degree every other box than in the former case, 
 and in case of open circuit in one, every one with which 
 it is in series becomes useless, and throws the entire field 
 current on the remaining series set. 4. With a given 
 
382 TESTING OF DYNAMOS AND MOTORS. 
 
 current in the armature the motor brushes do not require 
 to be brought as far back with the comparatively weak 
 field as with a strong one, and in putting on a load this 
 is a desirable feature. In such a test the load is put 
 on as in other motor-generator tests, by lowering the 
 
 C. E. M. F. of the motor. This is 
 
 _T^ I ^ lowered by either weakening the field 
 
 L-Q HI! LQ ' current, or by giving the brushes a 
 FIG. 124. negative lead, so as to bring the poles 
 
 of the armature in a position to 
 neutralize some of the field magnetism. In the fol- 
 lowing test we will assume the fields to be four 
 in number and connected in series. It is a practice 
 gaining hold to run the armatures of multipolar carbon 
 brush dynamos against the brushes, so that in starting a 
 test the direction of rotation is at least fixed, and the 
 relation between the field and armature polarities must 
 be adjusted accordingly. Now since both series and 
 shunt windings are on the spool, the current must 
 pass around both windings in the same direction, and to 
 insure that this shall be the case, it must be tested out. 
 The test consists in taking the drop first across one 
 winding, and then the other on any given spool. The 
 deflections should be in the same direction in both cases. 
 The series field resistance being very low and having only 
 that current necessary to turn the motor over slowly, 
 the drop across it will be only a fraction of a volt, and 
 the instrument must be delicate to indicate it. On the 
 other hand, as the drop on the shunt spool is consider- 
 able, the galvanometer must be shunted or otherwise 
 protected before being submitted to this drop. Since it 
 is merely the direction and not magnitude of deflection 
 
COMPOUNDING. 383 
 
 sought, it suffices to shunt the galvanometer terminals 
 with, a piece of copper wire. 
 
 In connecting the multipolar dynamo fields the test of 
 correct polarity is made on the shunt field as follows: 
 Attach the two shunt field terminals 
 to the blocks on either side of the 
 machine, as in Fig. 126, and insert 
 the field rheostat wherever it is FIG. 125. 
 
 most convenient to do so; next 
 remove one of the brush holder cables, or draw the 
 brushes in adjacent holders and close K, Fig. 123, to 
 charge A's field from C. A's field circuit contains a low 
 reading ammeter to be used in finding the field and 
 rheostat resistances, and to serve as a check on the cur- 
 rent necessary to give the required ampere-turns for 
 normal voltage. Upon closing K, the ammeter needle 
 must be observed, to see if it moves at all, or in the right 
 direction. In lieu of an ammeter, a slight spark upon 
 opening K indicates the field circuit to be closed. As 
 soon as the field is charged, open K and replace the 
 brush holder cable, and A should generate its own field, 
 if its connections are proper. If it 
 /~ ~\ refuses to generate, its shunt field 
 
 / / \ \ connections must be reversed. 
 ' ^ 
 
 reason f r charging from C 
 is considered elsewhere. The proper 
 FIG. 126. connections made on A, B, and C, 
 
 .Z?'s field excited, brushes all down, 
 and, as a last precaution, all connections inspected, the 
 test is ready to start. There are two methods of start- 
 ing: First, to use a lamp bank as a starting box in series 
 with the motor armature, bringing up the speed by grad- 
 
384 TESTING OF DYNAMOS AND MOTORS. 
 
 ually cutting in the bank and finally short-circuiting the 
 bank when '$ C. E. M. F. is well up; second, to reduce 
 C's E. M. F., close K\ and start the system on low volt- 
 age. The E. M. F. is then gradually raised till full speed 
 is attained. 
 
 The former of these methods would seem to be pref- 
 erable, as there is little likelihood of brush and belt 
 troubles, and C is not called upon to generate excessive 
 current. 
 
 The second method requires some care to avoid com- 
 plications when heavy machines are involved. If C be a 
 shunt machine it must be separately 
 excited, as otherwise it will lose its 
 field when it is attempted to lower its 
 E. M. F. for starting B. If shunt- 
 wound, and used in conjunction with 
 a lamp bank, .Z?'s field must be con- 
 nected beyond the lamp bank, as in 
 Fig. 127, so that there will be the strongest possible field 
 even when the bank is in circuit. In starting with a bank, 
 the number of lamps to be plugged in series depends upon 
 the E. M. F. used, while the number to be plugged in mul- 
 tiple depends upon the current necessary to start the 
 system. When^"', Fig. 123, is first closed, the total drop 
 is through the lamps, as the motor before starting has 
 no C. E. M. F. ; as the motor begins to move, and 
 increases in speed, its C. E. M. F. rises and absorbs part 
 of the impressed E. M. F., when the drop across the 
 lamps becoming less, they grow dimmer. The bank can 
 next be replugged with fewer lamps in series, and more 
 in multiple, and when the glow becomes almost imper- 
 ceptible the bank is short circuited with a plug, because 
 
COMPOUNDING. 385 
 
 at this stage 's C. E. M. F. is sufficient to control the 
 current value. In replugging a bank care must be taken 
 that the speed is each time given a good chance to 
 respond to the change in the bank, otherwise the lamps 
 may be submitted to an unsafe voltage. 
 
 The second or " low volt " method of starting dispenses 
 with the use of a' sometimes complicated and expensive 
 bank, and in experienced hands need give no trouble. 
 As a time saver it is particularly valuable when local 
 troubles with belts, brushes, and bearings necessitate 
 shutting down frequently. In general, the writers' 
 experience has been that a 5oo-volt compound-wound 
 multipolar street railway generator cumulatively con- 
 nected as motor will start smoothly under a pressure of 
 25, 35, 45, 50, or 60 volts, according as the machine is a 
 100, 200, 300, 400, or 500 kilowatt machine. These figures 
 are averages. If with 75 volts across C, B fails to start 
 upon closing A", the wiring must be inspected to see that 
 no error has been made. If there is no spark upon open- 
 ing K\ it indicates an open circuit, which may be due' to 
 an open switch, raised brushes, an absent fuse, or a loose 
 connection. If C consists of two dynamos in series, it is 
 convenient to have the one of higher voltage compound- 
 wound, and its mate separately excited, not only to facili- 
 tate reversing its polarity if necessary, but to render it 
 active or inert according as its E. M. F. is needed or 
 not; besides this, separate excitation makes certain that 
 the field will pick up when it is needed. The separately 
 excited machine should be of large current capacity and 
 low voltage, as compared with its mate, for on starting 
 it has a large armature current, and no field to determine 
 the neutral point. The result is that the pole heads are 
 
386 TESTING OF DYNAMOS AND MOTORS. 
 
 magnetized by induction, and locate the neutral line in 
 a position that the brushes cannot be made to occupy. 
 The result is sparking at the brushes, an effect much 
 aggravated if the current at starting exceeds the rated 
 output of the machine. 
 
 Since low voltage machines carry a larger current than 
 high voltage machines of the same output, the above 
 precaution lessens the danger of overload, and the 
 brushes will probably need little or no attention through- 
 out the test. 
 
 Failure in getting the system started upon closing the 
 switch may be due to the total absence of a motor field 
 or to its weakness, a condition to be detected primarily 
 by holding a nail or iron key to the motor pole pieces 
 before closing K'. If, however, this precaution has been 
 neglected, the indications are apt to be more violent, a 
 belt may fly off and the supplier brushes spark badly if 
 the switch is not promptly opened. As such cases, or 
 similar ones, arise from time to time, it is good practice 
 to use carbon brushes on all machines of sufficient volt- 
 age to admit of a brush of suitable cross-section. "Of 
 sufficient voltage " is specified, because on a machine of 
 large output and low voltage the current would be so 
 large that it would be impracticable to use carbon brushes 
 large enough to carry it. Copper brushes under certain 
 circumstances either melt and run on to the commutator, 
 or the component wires fuse together, rendering the 
 brush unfit for further use. Carbon brushes are used on 
 machines of 250 volts and over, and are especially adapted 
 for use on machines subjected to wide and rapid variations 
 of load, as in street railway service. 
 
 The resistance of a carbon brush is high enough to 
 
COMPOUNDING. 387 
 
 prevent the low voltage of each short circuited coil pass- 
 ing under it from generating a large current, and thereby 
 causing excessive sparking when the brushes are not on 
 the neutral line. On the other hand, on small machines of 
 low voltage and large current output, the resistance of 
 carbon brushes, practicable to use, would cause so great 
 an / R loss as to be very uneconomical. 
 
 For starting the smaller machines the voltage due to 
 the residual field of the loss supplier often suffices : this 
 residual field-can be increased by working up to full field 
 strength, and then gradually breaking the field by means 
 of the boxes. Where a separately excited machine is the 
 loss supplier, the exciter's field may be broken and its 
 own left intact, thus giving the residual field of the 
 supplier the additional magnetism generated by the 
 small current which the residual field of the exciter urges 
 around the supplier's field coils. 
 
 To start a pair of multipolar machines the field of one 
 loss supplier is opened: the line switch of the other 
 is the only break in the motor circuit, and the voltage 
 across it should read, 25, 35, 45, or 60 volts, according 
 as the machines are of 100, 200, 300, or 500 kilo- 
 watts' capacity. Care should be taken that the voltlines 
 make good contact, as otherwise the true voltage of the 
 machine will be higher than that indicated, because the 
 true voltage will be partly diverted by the resistance of 
 the poor contact. As a final precaution, before closing 
 the switch see that no person is where he can be caught 
 by belt or pulley. A man should be stationed at the 
 motor to signal if the direction of rotation is right. 
 Since the motor shunt field is separately excited, if the 
 system has the wrong rotation, it can be righted by- 
 
388 TESTING OF DYNAMOS AND MOTORS. 
 
 reversing the shunt field, for this being generally much 
 stronger than the series field, dictates the direction of 
 rotation even though the latter may oppose it. 
 
 To determine if the two windings assist each other, 
 any of several methods can be pursued. One method is 
 to break the shunt field while the motor is turning over 
 slowly: if the fields are properly connected, the speed 
 will rise; if connections are such as to oppose shunt and 
 series windings, the motor will stop and perhaps start up 
 in the reverse direction, because, if the fields are opposed 
 they tend to turn the armature in opposite directions, 
 but the weaker series field can dictate this direction only 
 when the shunt field is broken. Whether the motor 
 starts up in the reverse direction or not, depends upon 
 whether, with the weakened field and the current flowing, 
 there is torque enough. This test is safe to try only at 
 a very low speed and at a voltage insufficient to throw 
 the belt on the loss supplier, even when the motor stands 
 still. The rush of current upon breaking the shunt 
 field, where the two fields oppose each other, is greater 
 than that at starting, the voltage on the loss supplier 
 being the same in both cases, because it is a fact, 
 that the current which flows at the time of breaking 
 the shunt field is due to the sum of the impressed 
 E. M. F. and what was the C. E. M. F. of the motor, 
 while at starting, it is due simply to the impressed. 
 If, then, upon breaking the shunt field the motor 
 slows down and stops, the series field connections 
 must be reversed. If on the contrary the speed rises, 
 the connections are correct. Another method of testing 
 the connections is to start up with the series windings 
 short circuited: the field, and hence the direction of 
 
COMPOUNDING. 389 
 
 rotation, will then be due to the shunt winding alone; if 
 while the motor is turning over slowly the short circuit 
 be removed, the series winding will take effect, and will, 
 if the connections are proper, reduce the speed, because, 
 if connected to assist the shunt winding, its introduction 
 will strengthen the total field, raise the C. E. M. F., 
 lower the current flowing, and with it the speed. If, 
 however, the connections are such as to oppose the two 
 windings, the effect of introducing the series winding is 
 to weaken the field and raise the speed. In the hands 
 of a careful tester the above methods are satisfactory 
 and are quickly carried out. Ordinarily it is safer to 
 resort to the galvanometer test already given, but if there 
 are no facilities for doing so, the speed test is recom- 
 mended but with one injunction use as slow a speed as 
 possible; otherwise a belt will fly off whether connections 
 are right or wrong. 
 
 A modification of the above tests of connections consists 
 in varying the strength of the shunt field by means of its 
 rheostat and noting its effect upon the speed. If the wind- 
 ings are concurrent the speed will rise as the shunt field 
 is weakened, and will fall when it is strengthened. This 
 test is given because in experienced hands it has served 
 its purpose, but it is interesting to consider the condi- 
 tions under which its results would be very misleading: 
 i. The series winding partly or wholly short circuited: 
 in this case the shunt field would be practically the total 
 field, and the right or wrong series field connection could 
 give no indication; 2. If the shunt field were so much 
 stronger than the series field that the difference between 
 the two were greater than the series field alone, the effect 
 of varying the box between certain limits would be the 
 
390 TESTING OF DYNAMOS AND MOTORS. 
 
 same as if no series field existed. This test is not recom- 
 mended, because its indications can be properly interpreted 
 only when the operator is thoroughly familiar with the ma- 
 chine under test, for any difference in degree of over-com- 
 pounding involved will modify the behavior very much. 
 
 If the machine used to excite the motor fields is used 
 at the same time to excite other fields than the ones in 
 the test, the several circuits should be made independent 
 by placing a switch in each, or by connecting below the 
 exciter switch all those not to be disturbed, and above 
 this switch, those circuits which it may be necessary to 
 to break at intervals. Opening the switch will then 
 break only the latter. 
 
 When a compound-wound motor runs free, its series 
 winding contributes but little to the magnetization, since 
 there is the minimum current in the series coils, but the 
 influence increases with the load, till a point may be 
 reached where the series winding has as great an effect 
 as the shunt, and at this point, if the two windings are 
 in opposition, they will neutralize each other, with a 
 resulting short circuit. Neutralization will also take 
 place, when, for the purpose of increasing the speed, the 
 shunt field is weakened by means of its rheostat; such a 
 short circuit can only take place where the two windings 
 are in opposition hence the importance of getting them 
 right before hand. However, a comparative short cir- 
 cuit follows a sudden great weakening of the field in all 
 cases; as, for instance, the breaking of the shunt field in 
 the above test of connections reduces the motor to a 
 series motor with a much lighter field than when the shunt 
 field acts also. The result is that the reduced C. E. M. F. 
 allows an abnormal current flow, and if the armature is 
 
COMPOUNDING. 391 
 
 heavy, so that its speed responds too slowly to the new 
 conditions, a belt flies off of some machine. 
 
 Immediately upon closing the switch, at starting, the 
 brushes of the supplier, if of copper, must be brought 
 forward, to prevent sparking. The supplier's field must 
 be strengthened or trouble will follow, because, when 
 the switch is first closed, the compound-wound loss sup- 
 plier is virtually short circuited, having in series with it 
 the armature of its companion (if there are two suppliers 
 in series) and the stationary armature of the motor; the 
 initial flow of current is therefore considerable, and the 
 series field heavy. The result of this is that the E. M. F. 
 of the supplier rises rapidly, and as this is the E. M. F. 
 impressed at the motor terminals, the motor speed rises 
 also, and with it its C. E. M. F. Now, since the motor is 
 separately excited, its C. E. M. F. is independent of the im- 
 pressed E. M. F. , except in so far as the speed is influenced 
 by it. The heavy motor armature having once gained 
 headway, will, by virtue of its large inertia, tend to hold 
 its speed, and will not quickly respond to changes in the 
 impressed E. M. F. Initially, the heavy series field on 
 the compound-wound loss supplier enables it to generate 
 a high E. M. F., but as the C. E. M. F. rises, the current 
 in the motor circuit, and hence in the series field of the 
 supplier, falls off rapidly, and may fall so low that the 
 impressed E. M. F. no longer exceeds the C. E. M. F. 
 In other words, the motor armature becomes a generator 
 running by its own momentum, and sends a current back 
 through the loss supplier, running it as a motor. The 
 preventative of such behavior is to strengthen the shunt 
 field of the loss supplier so promptly that reversal can- 
 not take place. 
 
392 TESTING OF DYNAMOS AND MOTORS. 
 
 The action of the system under the above con- 
 dition may, if the motor armature be very heavy, 
 so that it cannot quickly come to rest, become quite 
 complicated, and it will be instructive to follow the 
 cycle through. As soon as the C. E. M. F. exceeds the 
 impressed, thus making the motor a generator feeding 
 the supplier, the current in the series coils of the supplier 
 is reversed, and so is the polarity of its fields. When the 
 shunt field picks up anew (if it does), it does so in accord- 
 ance with the reversed polarity, and gives us the condition 
 of both machines running in series as dynamos, and with 
 no resistance in circuit save the two armatures and two 
 series fields. Both machines are doing work: the loss 
 supplier, being connected to the engine, can do work as 
 long as the belt stays on; but the motor armature, owing 
 its energy solely to momentum, stops very soon. This 
 brings us to a new stage. The loss supplier has a strong 
 series field of reverse polarity to what it had at the start, 
 and hence sends a reverse current through the motor 
 armature; the motor fields, being separately excited, are 
 the same as at starting, therefore as soon as the motor 
 armature comes to rest, since its fields are of the same 
 polarity as before but that of its armature is reversed, it 
 begins to rotate again, but in the opposite direction. If 
 belts, fuses, and circuit-breakers hold intact, this cycle 
 of operations will repeat itself, the original condition 
 of affairs recurring every second reversal. Unless the 
 machines are allowed to reverse themselves a second 
 time, the shunt field of the motor must be reversed before 
 the system will rotate in the proper direction. It must 
 also be remembered that since the two suppliers were in 
 series at the start, now that one of them has been 
 
COMPOUNDING. 393 
 
 reversed the other must also be, so they will again be 
 in series. 
 
 Promptness and care are the preventatives of the 
 above complications, the compound-wound supplier's 
 brushes being brought forward, and its shunt field 
 resistance reduced so as to counterbalance the weaken- 
 ing of the series field. As the motor speed rises and its 
 armature current decreases, the supplier brushes are 
 worked back to their non-sparking point. 
 
 It is practicable and, in some tests requiring nice volt- 
 age regulation, desirable to have the voltages of the two 
 suppliers in opposition. The minimum line voltage is 
 then gotten when the two machines are of equal voltage. 
 In the present case we weaken the field of the low voltage 
 machine when it is desired to raise the impressed E. M. F. 
 The practice of opposing E. M. Fs. as a means of reg- 
 ulation is fast gaining ground now occupied by ordinary 
 ohmic resistances, which waste so much energy in heat. 
 The flexibility of the method is readily seen when we say 
 that with two 500 volt generators, their rheostats, and a 
 reversing switch in one of the circuits, any voltage is 
 obtainable from o to 1000. 
 
 If after full speed is attained it is found to be too high, 
 it can be reduced by lowering C's E. M. F. In doing 
 this great care must be exercised that it is not done too 
 suddenly, and a reversal precipitated. Reversals at high 
 speeds are generally accompanied by flying belts and 
 grand confusion. 
 
 With B running up to speed, but without load, it is 
 time to introduce the low volt supplier into circuit. 
 Since its armature is always in circuit, to make the 
 machine active it is only necessary to complete its 
 
394 TESTING OF DYNAMOS AND MOTORS. 
 
 separately excited field circuit, and gradually increase its 
 field strength, at the same time decreasing that of the 
 compound-wound machine by about the same amount. 
 The voltage of the low volt machine is brought up 
 to the maximum to be used, so it will require no 
 further regulation, and that of the compound-wound 
 machine is reduced till the sum of the two is that 
 required. The compound-wound machine is then used 
 for all further regulation. The compound-wound machine 
 is provided with field rheostats enough to cause it to 
 drop its field, when all resistance is in, and the low volt 
 machine's field must have resistance enough to reduce 
 its E. M. F. to 30 or 40 volts, which will be necessary 
 both when putting the machine into action and withdraw- 
 ing it. In the first place to prevent the introduction of its 
 voltage from causing too great an increase in the load; 
 in the second 'place to avoid the withdrawal of its 
 E. M. F. from causing a reversal. 
 
 We next prepare to load A on B. If A and ^are belt- 
 connected, A turns over when B does; but if clutched 
 together it is customary to put in A's clutch after B is 
 up to speed, an operation which requires some care, 
 because A's heavy armature has considerable inertia and 
 acts as a brake, and may temporarily overload the com- 
 pound-wound supplier sufficiently to throw Its belt. The 
 low volt machine gives no trouble, because, although it 
 runs at full voltage, it is far from its normal current 
 output and is therefore nowhere near its full load. 
 
 The next step is to get the voltage on the motor at ap- 
 proximately 500 volts, at the same time keeping the speed 
 right by means of the resistance boxes in its field. Sup- 
 posing that the voltage is correct, and that the speed is 
 
COMPOUNDING. 395 
 
 low; bringing up the speed by weakening B's field will 
 cause the line current, and hence "lost volts," to increase, 
 and hence also the voltage impressed on B, unless C is 
 a single machine perfectly compounded for the occasion. 
 Voltage and speed must therefore be adjusted simul- 
 taneously. It would seem at first sight that a perfectly 
 compounded dynamo might be useful as a loss supplier; 
 and so it would for either no load or full load, but not for 
 both and intermediate loads, since at no load the voltage 
 must be 500, at full load 555, and at intermediate loads be- 
 tween these limits. The only recourse, then, is to hand 
 regulation. The voltage on B and the speed of A are ad- 
 justed after A is made to generate its own field, because 
 the making of A's field puts more work on B and brings 
 down its speed. To insure that A's polarity shall be 
 that desired, its field is charged from C, by removing a 
 brush holder cable and closing K. By charging^ from 
 C, we can be sure that their polarities shall be opposed, 
 and for this reason: when the brush-holder cable is 
 removed from A, in order to cut its armature out of cir- 
 cuit; and K is closed, A and B may be regarded as two 
 motors connected in multiple on C. The two shunt fields 
 are charged exactly as they would be were the two 
 machines actually running as motors. Now the C. E. 
 M. F. of a motor opposes its impressed E. M. F., and 
 regarding A and B as motors, the impressed E. M. F. 
 being common to both, is the same on both, and since 
 both C. E. M. Fs. are opposed to the impressed E. M. F. 
 they are opposed to each other, when we consider the 
 local circuit of the two machines, as will be seen in Figs. 
 128 and 129. Here Fig. 128 shows the condition while 
 charging, and Fig. 129 the same after the impressed 
 
396 TESTING OF DYNAMOS AND MOTORS. 
 
 E. M. F. has been removed, leaving the two machines 
 opposed in polarity. 
 
 As soon as A's field is excited, and its voltage adjusted 
 to 500, its brushes are brought as far forward as spark- 
 ing will permit: it will then be unnecessary to shift them 
 when the load is put on. With voltage at 500 and speed 
 correct and constant, we are ready for the free data, 
 This consists of voltage and speed readings, current in 
 shunt field, and drop across series coils, all with no load 
 on the machine, and according to blanks provided 
 
 'C.E.M.F. 
 
 Imp. EJd.F 
 
 ^N^ 
 
 .C.E.M.F, 
 FIG. 128. 
 
 for the test. The "free" data taken, A's E. M. F. 
 is 500 volts, and that on B a little less. The volt lines 
 are now placed across K, which is the only open cir- 
 cuit between the two machines, and the voltmeter should 
 register zero, showing that the two sides of the switch 
 are at the same potential. In holding the voltlines 
 the operator must have care lest there be a potential 
 difference of 1,000 volts, which is the case if A and B 
 happen to be in series; should this be the case, A's field 
 must be recharged from C. Inability to get full voltage 
 from A when running up to speed indicates one or more 
 field spools to be wrongly wound or connected, . so that 
 one spool's magnetizing effect neutralizes that of another. 
 This can be caused either by getting the field spools in 
 the frame end for end, or by the winder bringing the 
 leads around one-half turn too far before bringing them 
 
COMPOUNDING. 397 
 
 out: either of which mistakes results in the inside field 
 lead coming out where the outside lead should be. A 
 simple test is to bring a hand compass up to the pole- 
 pieces and observe which pole of the needle is attracted: 
 should three consecutive poles prove to be alike, the mid- 
 dle spool must be reversed, either electrically or mechanic- 
 ally. In lieu of a compass a piece of soft iron will serve 
 as well, being simply held before the poles, and passed 
 freely from one to the other. Where adjacent poles are 
 unlike, as they should be, the piece of iron will follow a 
 natural path from one to the other, presenting opposite 
 ends to adjacent poles. If the poles are alike, the iron 
 will tend to balance midway between the two poles, and 
 present the same end to both. 
 
 We have now arrived at perhaps the most difficult, and 
 certainly the most interesting, part of the test that of 
 putting on the load. With both sides of A", Fig. 123, at the 
 same potential, or with a slight difference in favor of A's 
 side, it can be closed. At this time there should be a 
 man at Cs boxes; one to take speed on A; one at 's 
 boxes; one at B's brushes to put on the load; and last, 
 but by no means least, the man with the voltlines. As 
 soon as K is closed, ^'s field is weakened a little to 
 minimize chances of reversal. 's brushes are then 
 brought slowly backward till about one-quarter load 
 works on. As the load goes on, Cs voltage is raised by 
 means of its rheostat, in order to keep up the speed, 
 cutting out field resistance slowly, and giving each 
 change time to have its full effect. The tachometer, or 
 instantaneous speed indicator, should be checked up 
 with a timepiece and ordinary indicator, and the speed 
 must be kept exactly right, as a difference of 4 or 5,. 
 
398 TESTING OF DYNAMOS AND MOTORS. 
 
 revolutions will sometimes cause an error of 6 or 8 volts. 
 The tachometer should be handled carefully, and with- 
 out leaning on it, as undue pressure causes it to run hot 
 and stick. Speed must always be taken on the dynamo 
 and not on the motor, for although they are belted or 
 clutched together, a difference in pulleys or slipping of 
 belt or clutch would introduce an error. 
 
 If A and .# are heavily over-compounded, it is customary 
 to use a temporary shunt on A's series field to prevent 
 the load from going on too suddenly. When the load 
 reaches quarter value this shunt is slowly worked out, and 
 the load further increases. The removal of the shunt 
 has the following effect: Primarily, part of the load is 
 put on by reducing j9's C. E. M. F. This is accomplished 
 by rocking the motor brushes back so as to bring the 
 armature poles in a position to demagnetize the fields. 
 The effect of this reaction is so great that at full load 
 when the field is weakest the field current is often greatest. 
 On a dynamo the effect of giving the brushes a forward 
 or positive lead is to have the armature reinforce the 
 field and raise the E. M. F., and since C. E. M. F. is 
 the dynamo property of a motor, the effect of rocking 
 the motor brushes forward is to raise its C. E. M. F. and 
 diminish the load. The full load can now be worked on 
 A, and with its speed and load adjusted it is permitted to 
 run 5 hours, or till it heats thoroughly. One man can 
 easily regulate the load and watch the bearings. 
 
 We will now consider the factors entering into the 
 problem of compounding, and the various points lending 
 aid to success. Compounding consists in experimentally 
 adjusting a permanent German silver shunt across the 
 series field of the dynamo, such that when running at full 
 
COMPOUNDING. 399 
 
 load, proper speed, and fully heated, the dynamo shall 
 give a specified E. M. F. At the beginning of the test 
 a variable shunt board is used, for convenience, and 
 it is this board that the regulation shunt replaces when 
 the load has been removed for readjusting the rheostat 
 for 500 volts just before compounding. So far as the 
 shunt winding is concerned, any properly designed 
 machine that compounds cold, /. <?., maintains its speci- 
 fied voltage from no load to full load, will compound hot, 
 as resistance is taken from the rheostat to compensate 
 for the rise of resistance in the winding due to heating. 
 With the series field, however, such is not the case. If a 
 machine be compounded cold with a shunt of certain 
 resistance, it will, when heated and its rheostat read- 
 justed for 500 volts on open circuit, undercompound, 
 and a new shunt will be required. The reason for this 
 is that the same relation between the series field and the 
 shunt no longer exists. In the first place the shunt, being 
 exposed to the air, does not rise very much in tempera- 
 ture; secondly, German silver does not rise in resistance 
 for a given increase in temperature at the same rate as 
 copper does, the rate being much less. The consequence 
 is the shunt does not rise very much in resistance, while 
 the series field does, thereby sending a greater proportion 
 of the armature current through the shunt and weakening 
 the field. To restore them to their former relation the 
 shunt resistance is raised, by cutting out one or more 
 strips of German silver. The only factor uncompensated 
 directly is the increased reluctance of the heated iron; 
 this is small. 
 
 As has been said elsewhere, the present test has been 
 selected to include as many troubles as possible, but it 
 
400 TESTING OF DYNAMOS AND MOTORS. 
 
 must not be supposed that all of them occur in any one 
 test. Few of them should, but the operator must be 
 prepared for any. 
 
 Should the load go on with a rush when the motor 
 man shifts the brushes, it indicates any of the following 
 troubles: (i) Opposition of motor fields; (2) absence 
 of shunt on dynamo series field; (3) excessive over- 
 compounding of the dynamo; (4) shunt on the motor 
 series field probably left there when the machines 
 were changed over; (5) someone dressing a belt without 
 giving notice; from the last cause alone the writers 
 have seen the ammeter needle go steadily but rapidly 
 from quarter load to overload. The effect of the above 
 cases may be briefly analyzed: (i) If j9's windings are 
 opposed, any weakening of the shunt field results in 
 sending current through the series winding in such a 
 direction as to neutralize and further weaken the field, 
 and as the current increases this effect multiplies, and 
 the load goes on too fast to control. To test polarity, 
 the usual galvanometer test is resorted to. (2) The ab- 
 sence of a dynamo shunt permits its E. M. F. to rise too 
 rapidly and precipitates the load. (3) When the dynamo 
 is highly over-compounded, even with the shunt ordi- 
 narily used, the series field effect is too strong, and pre- 
 cipitates the load for the same reasons as given in (2), 
 but to a less degree. (4) Weakening the motor field is 
 the same in effect as strengthening that of the dynamo. 
 (5) Tightening the belt or dressing it prevents slipping 
 and raises the dynamo speed and voltage, thereby in- 
 creasing the load. 
 
 Should the load fail to go on, it may be due to any of 
 the following causes: (i) The dynamo series field may 
 
COMPOUNDING. 4OI 
 
 be too powerfully shunted; (2) the dynamo series and 
 shunt fields may be opposed; (3) the ammeter needle 
 may be " frozen"; (4) a belt or clutch may be slipping, (i) 
 The effect is to rob the series winding of current, and 
 thereby deprive it of ability to force a load against the 
 growing opposition of the motor series winding. Gradu- 
 ally removing the shunt will show if this is the cause. 
 Virtually the same effect obtains if the series field has 
 a loose connection, because nearly all the current then 
 passes through the shunt. In this case, the shunt grows 
 hot. (2) If shunt and series oppose, neutralization takes 
 place as the load increases, and it refuses to exceed a 
 certain value. The series connection must be reversed. 
 (3) In this case the load is really going on, though the 
 ammeter does not record the fact; tapping the meter so 
 as to free the needle is sufficient. (4) A slipping belt 
 generally announces its condition by squeaking, but not 
 always. As an illustration of belt troubles the following 
 very singular performance may be interesting: Two 500 
 kilowatt street railway generators were under test, and 
 quarter load was on, after which all efforts to increase it 
 were futile. Upon weakening the motor field the load 
 decreased instead of increasing. The action suggested 
 that possibly the supposed motor was really a generator, 
 but weakening the other's field disproved this. Finally the 
 trouble was removed by tightening the belt. The behavior 
 is explained as follows : upon weakening the motor field its 
 speed became higher, but contact between pulley and belt 
 became poorer at the higher speed, so that the dynamo 
 speed became actually less. Upon strengthening the 
 motor field the speed fell off, but its belt grip improved, 
 so that the dynamo speed increased and with it the load. 
 
4O2 TESTING OF DYNAMOS AND MOTORS. 
 
 There was naturally a limit to this unusual method 
 of putting on- a load, for with good belt contact, 
 further strengthening of the motor field brought clown 
 the speed on both machines, and also the impressed 
 and counter E. M. Fs., but the former faster than the 
 latter. The reason for this is: (i) The dynamo, being 
 self-exciting, decreasing its speed not only decreases its 
 E. M. F. directly, but does so further by weakening the 
 field which depends upon that E. M. F., an effect absent 
 on the separately excited motor. (2) Since the dynamo's 
 E. M. F. exceeds the motor's counter by an amount equal 
 to the cable " drop," a percentage reduction in the speed 
 causes a greater absolute effect on the dynamo, so that 
 its excess of voltage gradually disappears, and with it 
 its load. 
 
 Should the difficulty in putting on the load not be 
 found among the above more common ones, it may be 
 due to the rocker arm on one machine being carelessly 
 left 90 out of position, which results in reversing the 
 series field of the motor, and in depriving the generator 
 of its ability to generate unless its shunt field has been 
 reversed to suit conditions, in which case the series field 
 must be also reversed. No sane man is apt to leave 
 brushes out of position if allotted reasonable time for his 
 work, because the handle of the rocker arm is a glaring 
 indicator of its position, but when eight or ten men 
 work on the same test trying to do half an hour's work 
 in five minutes, almost anything is liable to occur. 
 
 There are times when the ammeter needle surges 
 back and forth from no load to full load. To locate the 
 cause of such action it is necessary to note simultane- 
 ously the variation of load and speed. If the two rise 
 
COMPOUNDING. 403 
 
 and fall together, the trouble is with the loss supplier, 
 or some belt, clutch, engine, or countershaft connected 
 with it. If the speed falls when the load rises, it indi- 
 cates trouble in the exciter circuit, for we have seen 
 that weakening the motor field increases the load and 
 decreases the speed, and vice versa; whije raising the 
 voltage on the suppliers increases the load and speed at 
 the same time. The proof of a slipping belt is a warm 
 pulley. 
 
 When there is persistent trouble with the speed and 
 careful inspection fails to class it in any of the above 
 cases, suspicion points strongly to the motor, and is con- 
 firmed by the fact that the speed is equally unsteady at 
 no load and full load. If the trouble is with the fields, 
 as is apt to be the case, the speed will rise as the load 
 falls. A rough test is made by means of a piece of soft 
 iron and a compass; the iron showing any marked 
 discrepancy in pole strength and the compass any 
 irregular polarity. If the trouble is with the armature, 
 the indications are apt to be more violent, and give 
 rise to local heating. A good way to determine exactly 
 the seat and nature of its trouble is to reverse the 
 machines and run the defective motor as a dynamo: if 
 the trouble is, as suspected, with the motor, the incon- 
 stancy of speed will no longer exist. Next see if the 
 defective machine will excite its own field, and if not, 
 charge from the motor terminals: the field ammeter will 
 show by the magnitude of the deflection whether the 
 field resistance is normal or not. Lastly take the drop 
 of potential across each spool and locate the faulty one. 
 
 With all difficulties surmounted and the full load on, 
 the machines are allowed to run until thoroughly heated, 
 
404 TESTING OF DYNAMOS AND MOTORS. 
 
 which requires from three to eight hours. The load is 
 then removed, and the dynamo field rheostat readjusted 
 for 500 volts on open circuit and the shunt board replaced 
 by eight or ten strips of German silver tape. The 
 machine is then compounded according to the directions 
 on the test sheet. With full load on and correct speed, 
 the voltage must be that of the specified over-compound, 
 which in this case we assume to be 555 volts. If on the 
 first arrangement of the German silver shunt the full load 
 voltage is too high, the shunt's length can be diminished 
 by drawing it through the clamps, thereby lowering its 
 resistance, if the voltage is too low the cross-section can 
 be lessened by taking out a strip. Care must be take"n 
 that throughout the process the speed and current are 
 kept right. The "up" and "down" readings of voltage 
 are now taken, and the former should be uniformly lower 
 than the latter. If this is not the case, it shows that 
 either there is an error in one set of readings, or that the 
 machine, not having reached its maximum temperature, is 
 still rising. Having passed this final test, the compound 
 is completed, and the change over is in order. The 
 German silver shunt is removed, carefully labeled, and 
 sent to be made up in compact form. The series field 
 resistance is measured, hot and cold, on both machines, 
 and from these data is figured the actual temperature rise. 
 If there is an ammeter in the supplier circuit we know 
 the total motor current, as well as that of the dynamo, 
 and by taking the drop on the respective series fields we 
 can at any time tell their resistances. In making these 
 measurements care must be taken that the fields are not 
 shunted, for the multiple resistance of coils and shunt is 
 much less than that of either alone. Should the drop 
 
COMPOUNDING. 405 
 
 be abnormally high, that across each spool must be 
 taken, and in such a way as to include no joints; the 
 drop across each joint can then be taken, and any loose 
 or lacquered joint located. 
 
 The above data having all been taken, the machines 
 are shut down and changed over, the motor becoming a 
 dynamo, and vice versa. The change consists: (i) In 
 removing the supplier cable next to A", and placing it on 
 the other side of K' (Fig. 123); (2) in exchanging the box 
 lines of the two machines, and making the motor, to be, 
 separately excited and the dynamo self-excited. In most 
 cases the series fields must be reversed, but if A and B 
 are belted together, and one is provided with right-hand 
 the other with left-hand brush holders, or if the ma- 
 chines are clutched together and both are fitted with 
 similar brush holders, then their direction of rotation 
 must be reversed to conform to the rule requiring car- 
 bon brush dynamos to run against the brushes, and the 
 series field connections need not be disturbed except to 
 transfer the fuse from the old motor to the new one. 
 The shunt board must also be transferred to the dynamo 
 to be. Due regard must be had for the fact that machines 
 belted together generally turn in the same direction, 
 viewed from the commutator end, but when clutched 
 together they stand pulley to pulley, and therefore turn 
 oppositely. The change in connections completed, the 
 machines are once more started up, and allowed to run for 
 a few moments before putting on the load. In throwing 
 in a clutch, after it has been snapped, the pressure ring 
 must be relieved, otherwise it will run hot. Any slipping 
 can be detected by taking speed on both machines. The 
 load is put on as already described, and the system 
 
406 TESTING OF DYNAMOS AND MOTORS. 
 
 allowed to run for three or more hours longer; when the 
 second machine is compounded the set is run long 
 enough to complete the ten or twelve hour test as the 
 case may be. 
 
 If after a load has been running smoothly for some 
 hours it begins to work off slowly, the cause is either a 
 slipping belt or clutch, or more likely a movement of one 
 or both rocker arms. This can be detected at a glance 
 by making a chalk mark on pillow block and rocker arm at. 
 the beginning of the test. There is a tendency for the 
 commutator to drag the brushes and rocker arm around 
 with the armature, and no dynamo should be shipped in a 
 condition to admit of this. Testers pay special attention 
 to this detail, for, when running in multiple with others, a 
 machine with such a weakness would be sure to cause 
 trouble. The movement of the rocker arm on either 
 dynamo or motor has the effect of working off the load. 
 On the dynamo it decreases the lead, and hence the 
 E. M. F. On the motor it increases the negative lead, 
 and hence increases the C. E. M. F., both of which 
 actions tend to remove the load. The first symptom of 
 a shifting rocker arm is a decrease in the ammeter read- 
 ing, followed by sparking of the motor brushes, and 
 unless noticed in time, the load works down to zero, 
 at which point the supplier runs both machines as 
 motors, and the dynamo announces the fact by a char- 
 acteristic howl. 
 
 Circumstances sometimes demand the instant removal 
 of the load, by pulling switch K: in such a case the 
 motor brushes should be brought quickly forward to 
 relieve sparking, the motor field strengthened, and the 
 supplier E. M. F. reduced, otherwise the speed will rise 
 
COMPOUNDING. 407 
 
 to the limit set by the motor's separate excitation. The 
 reason is as follows: Since bringing the brushes back to 
 put the load on has lowered the C. E. M. F., the first ten- 
 dency is for the speed to rise, but this is met by the 
 increased load which the over-compounded dynamo 
 immediately puts on the motor, with the result that the 
 speed must be maintained by raising the supplier's 
 E. M. F. The conditions at the time of removing the 
 load then are these: (i) The supplier voltage is much, 
 above the value necessary to keep the system at the 
 given speed, had the dynamo no load; (2) the powerfut 
 series field of the motor under full load has an effect 
 almost negligible as soon as the load is thrown off. The 
 motor brushes are in a position corresponding to high, 
 speed at no load, and may or may not be counter-bal- 
 anced by the excess of shunt field current necessary to- 
 reduce sparking. The effect, then, of removing the load 
 is to remove the .restraining influence, hence the advis- 
 ability of the above precautions. 
 
 A simple way to prevent all danger is to pull K and K' 
 at the same time: this has the disadvantage that it stops 
 the system. Pulling K" without pulling K stops the 
 set, but not without sparking badly, for it leaves A and 
 B to run back on each other in virtue of their inertia, 
 and since A is self-exciting, it will lose its field when the 
 speed gets below the critical value, leaving A to short 
 circuit B, now running as a separately excited generator. 
 
 Pulling the main switch under full load in a motor- 
 generator test is not attended by serious arcing even 
 with a current of 2,000 amperes, and a switch with a 
 3 inch break is safe. Such is not the case on a similar 
 circuit containing only ohmic resistance, for here a switch 
 
408 TESTING OF DYNAMOS AND MOTORS. 
 
 of three times the break would not prevent arcing. The 
 conditions in the two cases are different: in the latter 
 the full voltage of the system is effective in supporting 
 the arc, while in the circuit containing the motor's 
 C. E. M. F., only the excess of the impressed over the 
 counter E. M. F. is so available; and this is merely that 
 necessary to drive the current through the low ohmic 
 resistance of the machines and connecting cables. It is 
 good practice to adhere to a uniform rule in removing a 
 load: the supplier's E. M. F. must be reduced step by step 
 with that of the dynamo, and the motor speed kept con- 
 stant. This is continued until the load is off, and when 
 K is opened everything is nearly adjusted to put back 
 the load if it is desired to do so. 
 
 The test over, the thermometer temperature of arma- 
 tures and fields is taken, and the hot insulation meas- 
 ured. The thermometer is placed on the part to be 
 taken, and is well packed around with cotton waste. The 
 armature requires half an hour or more after stopping in 
 which to reach its maximum temperature, because while 
 running the fanning of the air keeps the surface layer of 
 wire cooler than those inside. The most reliable infor- 
 mation in regard to temperature rise is gotten from the 
 rise in resistance. 
 
 Insulation measurement is taken from series to shunt, 
 series to frame, shunt to frame, and commutator to shaft: 
 we specify shaft to eliminate any possible error due to 
 the film of oil which lies between the shaft and frame. 
 In testing the fields all exterior wires are disconnected, 
 and all wires remaining are cleared from the frame. One 
 test line is then held on one field terminal, the other test 
 line on the frame. If any of the field tests are low, the 
 
COMPOUNDING'. 409 
 
 spools are separated, tested, and the faulty one marked. 
 A low spool or armature is returned to the oven and 
 baked till it passes the test. 
 
 The galvanometer used in testing insulation is a 
 Thomson reflector, with the scale at such a distance as 
 to introduce no appreciable error in assuming that the 
 deflection produced is directly proportional to the 
 E. M. F. applied to the galvanometer. The galvanom- 
 eter in question is calibrated with a Daniell cell of i.i 
 volt, and the resistance in circuit is such that with no 
 shunt 1. 1 volt produces a deflection of 1 10 scale divisions. 
 Each division therefore corresponds to .01 volt. Using 
 the 1/99 shunt, one division corresponds to i volt. 
 
 By combining shunts and proportion lines, the range of 
 voltage possible to be read is from .01 to 500,000 volts. 
 This range has, however, its limit in the carrying ca- 
 pacity of the proportion boxes. All that is seen and 
 handled of the proportion box is two pairs of small flexi- 
 ble cables leading from it; one pair, including the total 
 resistance of the box, goes to the terminals of the 
 machines under test. The other pair includes the de- 
 sired fraction of the total resistance and goes to the 
 galvanometer. Should the lines be confused, resulting 
 in putting the fractional resistance across the source 
 of E. M. F., a fuse may blow unpleasantly near the 
 operator's eyes, or the galvanometer may be injured. 
 
 In testing insulation, the galvanometer, with its resist- 
 ance boxes, charging lines, safety lamps, and test lines 
 are all in series, as shown in Fig. 130. The 1/99 or 1/999 
 shunt is used on the galvanometer according as the 
 charging lines are from the 125 or 500 volt circuit. The 
 test lines are first held together, and the galvanometer 
 
4io 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 deflection noted. This deflection is due to the current 
 which A sends through R, G, in series, and is called 
 the constant. The lines are now held apart, and no 
 
 permanent deflec- 
 tion should obtain 
 if the insulation is 
 everywhere intact. 
 The lines are now 
 held across the in- 
 sulation to be meas- 
 ured, care being 
 FIG. 130. taken that the fin- 
 
 gers do not touch 
 
 the wires, for the body has an average resistance 
 of but 5,000 ohms, and would be a comparative short 
 circuit across the insulation, which measures perhaps 
 megohms. If now the "constant" is, say, 135 volts, 
 corresponding to a deflection of 135 divisions, and 
 the insulation reading is but i division, the circuit 
 resistance is 135 times as great as it is when the "test " 
 is cut out, and the insulation resistance is approximately 
 135 times that of the galvanometer boxes, etc. In this 
 way the insulation resistance is gotten, and none 
 below 1,000,000 ohms (i megohm) is accepted. The 
 following table gives the insulation value of one division, 
 using either the 125 or 500 volt charging lines. All 
 instruments used in these tests are calibrated at regular 
 intervals by comparison with reliable standards. 
 
 SHUNT USED. DEFLECTION. MEGS. @ 125 VOLTS. MEGS. @ 5OO VOLTS. 
 
 o i J 35 54 
 
 i/99 i !3-5 54 
 
 i/999 I i-35 5-4 
 
COMPOUNDING. 411 
 
 TEST XL Compounding; all Machines in Series. Test 
 X required the loss supplying generator to be of the 
 same E. M. F. as the machines under test, though 
 its current capacity might be considerably less, and fur- 
 ther required that it run in multiple with the dynamo. 
 In the present test the current capacity of the supplier 
 must equal or exceed that of the machines under test; 
 and all machines are in series, as far as having the cur- 
 rent pass out of one into the other fulfills the definition 
 of the term. Fig. 131 shows 
 
 the connections for test. K A/H L h 
 
 A is a compound-wound 
 generator; B, a compound- 
 wound machine similar to 
 A (series fields are omitted E 
 
 in figure), to be used as a 
 motor; C is a shunt wound 
 loss supplier. The test 
 would be easier with all F IGi I3I 
 
 shunt fields separately ex- 
 
 cited, but we will suppose all machines to be self- 
 exciting, which will necessitate the use of lamp bank 
 or water box, Z, to be used in starting, because: 
 C, a shunt machine, is used to start the system, and, 
 since it is self-exciting the result of closing K would be 
 to have Close its field, even assuming that it would be 
 possible to lower C's voltage sufficiently to safely close 
 K. C is driven by the engine E; A and B are belted or 
 clutched together. Let us assume, then, that A and B 
 are 500 volt 500 kilowatt compound-wound machines, 
 and that C is a 125 volt machine of the same current 
 capacity. A is cumulatively connected as a dynamo, 
 
412 TESTING OF DYNAMOS AND MOTORS. 
 
 and B the same, as motor, /. <?., in all cases their series 
 and shunt windings must assist each other. It is only 
 necessary to know the connection as dynamo when that 
 as motor is secured by reversing the series field. For 
 reasons already shown, a differentially connected motor 
 will run as a cumulatively connected dynamo and vice 
 versa. On shunt machines, for the same direction of 
 rotation the connections are the same, and separately ex- 
 cited machines follow the same law. At starting there is 
 no field on A, and when the time comes to make its field 
 there must be no doubt as to its polarity, to avoid intro- 
 ducing it into circuit as a motor. A, being self-exciting, 
 will not pick up its field till a certain speed is reached, 
 and at that point is apt to pick up so rapidly as to give 
 trouble unless C's field is promptly strengthened to raise 
 the voltage supplied. It must be borne in mind that as 
 soon as a machine generates voltage it begins to do work 
 if the circuit is closed, and since the machines are all in 
 series and have the same current, the amount of work 
 that each does is proportional to its voltage. The order 
 of starting up is as follows: C is first brought up to full 
 speed by the engine, with K open, to enable C to get as 
 much field as 125 volts will give it, K\*> now closed and L 
 plugged till the system gets well started. With the lamp 
 bank in, the supplier will keep the system in motion at a 
 fair rate of speed, until A begins to pick up a field and 
 do work, when the demand on C becomes so much 
 greater that, unless some of L is cut out so that C can 
 devote its voltage to the motor instead of Z, the system 
 will slow down until it reaches a point where A loses its 
 field, then the load being off, the motor will speed up 
 again, until A once more acquires its field, only to repeat 
 
COMPOUNDING. 4 T 3 
 
 the same cycle. The reason for this action is this: The 
 two machines, A and C, being in series as dynamos, the 
 amount of electrical work each does depends upon 
 the E. M. F. each generates, and this in turn upon their 
 respective field strengths; now C, in its work, is sup- 
 ported by the engine to which it is belted; but A owes 
 its energy of rotation primarily to C, and then retains 
 it by its own inertia. The result is that when A 
 acquires its field, it throws on the two belted machines 
 a load entirely out of proportion to the amount 
 of energy available from C, and unless this amount is 
 increased accordingly, the machines must slow down 
 until the dropping of A's field removes the abnormal 
 load. If C is separately excited, L need not be used: 
 then C"s voltage is reduced to 35 or 40 volts and K 
 closed. To facilitate an easy start, it is well to have B^ 
 field circuit resistance low so it will take as much current 
 as possible, and better still to start the belt by hand, 
 thereby raising the C. E. M. F. of j9's armature and 
 letting more current through its shunt field. If, as in 
 the present case, B is compound-wound, its series wind- 
 ing gives it a good starting torque, and the hand start is 
 unnecessary. So also by using the lamp bank, B, has 
 a strong shunt field even before K is closed. Assuming 
 the system started, whatever method may have been 
 adopted, as soon as a fair speed is attained, the polarity 
 of A is tested with reference to that of C. If A is sepa- 
 rately excited the test is made by closing the shunt field 
 circuit through considerable resistance, and observing 
 either an ammeter in the main circuit, or a voltmeter 
 placed across A and C: if A and Care in series, as they 
 should be, making the field on A will increase the read- 
 
414 TESTING OF DYNAMOS AND MOTORS. 
 
 ing on the voltmeter, also that on the ammeter. Another 
 -indication of the same fact is a sudden falling in the 
 speed, showing that the introduction of A has increased 
 the load, while it has not increased correspondingly A's 
 ability to do the added work. If A and C prove to be 
 opposed, the fact is indicated by a decrease in the volt- 
 meter and ammeter reading: for since the E. M. Fs. are 
 opposed, the voltmeter will register their difference; and 
 since A and B both are then motors, the sum of their 
 C. E. M. Fs. is greater than that of B alone, hence the 
 decrease in current and speed. Thus, we see that the 
 speed decreases in both cases, but for different reasons. 
 If A and Care found not to be in series, and C is sepa- 
 rately excited, its shunt field must be reversed; if self- 
 exciting, A's fields must be charged from C, by tempo- 
 rarily reversing A's shunt winding, raising its brushes, 
 -closing K, when the field of A will become so charged 
 as to leave its residual field in reversed polarity to what 
 it was before. A's connections are now restored, for if 
 left reversed A will not, for reasons to be seen later, 
 generate. Upon opening K, lowering the brushes and 
 again starting up, A's E. M. F. will be found in series 
 with C's. 
 
 The system is now in motion with a slight load on, 
 but with the speed low. The next step is to work on 
 full load and adjust the speed. A and Care generators 
 in series, and running B as a motor. While this is their 
 relation to each other, their relation to the circuit is very 
 different. C, being belted to , maintains its speed inde- 
 pendently of load variations; /. e., unless E itself becomes 
 overloaded. Such is not the case with A. As its load 
 goes on, it must depend upon B to keep up its speed, and 
 
COMPOUNDING. 415 
 
 B in turn depends jointly upon A and C. The first 
 effect of putting a field on A is to throw a load on B, 
 whose speed will fall unless C's E. M. F. is brought up, 
 thus calling upon E for support. The office of C, then, 
 is to supply the additional energy necessary to keep up 
 the speed as the load is increased on A. The load is 
 worked on by slowly strengthening A's field, with an eye 
 on the ammeter, and at the same time strengthening C's 
 field, but weakening B's. After the usual run, A is com- 
 pounded, and the change over made. To shut down the 
 test, A's and C's E. M. Fs. are lowered, and fi's C. E. 
 M. F. raised by strengthening the field. When the load 
 is nearly off, K is opened and the system stopped. 
 
 No free data can be taken in this test, for A's armature 
 always carries a current, hence some load. At the end 
 of the test, however, an independent test can be run, 
 using B as a motor and supplying its voltage from some 
 500 volt machine. Points of particular care in this test 
 for the most part relate to the putting on of the load. 
 When it is desired to make A work, its field circuit resist- 
 ance is slowly worked out, till the field begins to pick up, 
 when the resistance should be promptly worked in 
 again, but not to a point where A will lose its field. 
 The introduction of resistance should be governed by the 
 indication of a voltmeter placed across A's terminals. 
 As soon as the field begins to pick up, the needle will 
 rise rapidly. Resistance should then be used until the 
 needle stops; but if it starts back, more resistance must 
 be cut out, otherwise A will drop its field. C's voltage is 
 increased each time that the load is increased. It must 
 be kept in mind that A's E. M. F. controls the load, and 
 C's the speed. When A's E. M. F. reaches its proper 
 
416 TESTING OF DYNAMOS AND MOTORS. 
 
 value, the rest of the load is put on by weakening .Z?'s 
 field. If A is compound-wound, however, its rheostat is 
 put at as near as can be reckoned its position for normal 
 voltage, free, and the load put on, letting the series 
 winding bring the voltage up to what it may. 
 
 If A and C happen to be opposed in E. M. F., thus 
 making B and A motors in series and running from C, as 
 generator, the inertia of the system will, if A's field picks 
 up too rapidly, enable their combined E. M. Fs. to run C 
 as a motor, with the usual brush display. The flow of 
 current in this case, however, would be limited by the 
 fact that j^'s fields, cumulatively connected as motor, 
 would act differentially as soon as B became a generator. 
 If C is compound-wound the reversed current through 
 its series windings may reverse its polarity, unless its 
 shunt winding is separately excited, when A, B, and C will 
 be generators in series, working on short circuit. Such 
 a reversal is indicated by the ammeter needle falling to 
 zero, and then either rising again, or deflecting to the 
 wrong side, according as the meter is of an alternating cur- 
 rent or a direct current type. There will also be a general 
 sparkingat the brushes. This lasts only until A and B ex- 
 pend their energy of inertia, slow down, A drops its field, 
 and the speed rises once more. If, however, C's voltage 
 is high, and the reversal is so violent as to bring the 
 system to a sudden stop, A and^ will start up as motors, 
 if separately excited. If, however, A and B are both 
 self-exciting, at the instant when they would become 
 motors neither has any shunt field, since the low resist- 
 ance armatures short circuit the shunt windings, and the 
 series fields tend to turn the armatures in opposite direc- 
 tions, with the result that they do not turn at all, but 
 
COMPOUNDING. 417 
 
 stand at a short circuit through C. If C is self-exciting, 
 it loses its field, or belt possibly. If compound-wound the 
 series winding prevents its losing the field, and if its 
 E. M. F. is high enough the belt must go. If C is com- 
 pound-wound self-exciting, and A and B are separately 
 excited, and a too sudden strong increase in A's field 
 causes a reversal, C's polarity is reversed by the reversed 
 current in its series windings, its shunt winding picks up 
 accordingly, and if A and B are not already stopped by 
 the reversal, they will be, because the polarity of their 
 fields is unchanged, while the E. M. F. impressed upon 
 their armatures has been reversed. At the risk of per- 
 haps tiresome repetition, we might enumerate many other 
 manifestations depending upon the type and manner of 
 exciting A, B and C. 
 
 If necessary there is no objection to pulling the line 
 switch under full load, except that the motor brushes 
 may be well back, and the resulting flashing injures the 
 commutator. If C's field gets broken under load, it 
 leaves A and B to run back on each other, precipitating 
 a heavy load, and stopping them. If C is compound- 
 wound the E. M. F. due to the series turns remains in 
 series with that of A, helping to turn .# as motor. If B, 
 however, is separately excited, as is often the case, it 
 still has a field after A's goes (due to fall of speed), and 
 tries to run A and C in the opposite direction as motors. 
 The result is a short circuit which brings A and B to a 
 stand. 
 
 Where a low voltage machine of sufficient current 
 capacity is available, this test is commendable, and is 
 correct practice. Shunt, compound-wound, and sepa- 
 rately excited machines can be run in a motor-generator 
 
418 TESTING OF DYNAMOS AND MOTORS. 
 
 test under almost any conditions, but series machines 
 cannot. The latter require that the loss be supplied, 
 either by an engine, by a machine in series with the load, 
 or by a machine having no electrical connection with 
 the system. Practically, series machines fall under two 
 classes: arc light dynamos and street railway motors. 
 Arc machines are always tested on a lamp load, to realize 
 working conditions. It being impracticable to lay down 
 general rules covering all types of series machine, the 
 three principal types have been separately considered in 
 a previous chapter. Motor testing will be considered 
 in another chapter, and points covering series machine 
 testing will be given there. 
 
CHAPTER XII. 
 
 MISCELLANEOUS TESTS. 
 
 HAVING considered dynamo tests generally adopted in 
 good practice, we will now consider a variety of tests 
 which belong to the experimental stage of development 
 of all machines. These are: (i) Core Loss; (2) Satura- 
 tion; (3) Distribution; (4) Efficiency, Electrical and 
 Mechanical, and (5) incidentally, Hysteresis and Fric- 
 tion. 
 
 TEST XII. Core Loss Test. The test for measuring 
 the work done in turning a naked armature core in an 
 excited field constitutes a core loss test, and this test is. 
 run on every new type of machine, and at intervals on old 
 types. If the loss is thought excessive, a change is made 
 in the quantity, quality, or disposition of the iron used. 
 To better understand what is meant by "core loss," 
 we will recapitulate a little. Wfcen a moving conductor 
 cuts lines of force an E. M. F. is set up, and a current 
 flows if the circuit be a closed one. Now, iron is a con- 
 ductor, and a solid armature body is just as much a closed 
 circuit as is a bare wire properly disposed. The old 
 Siemens solid bodies are in evidence of this. These 
 induction currents are greatly reduced, but not entirely 
 eliminated, by laminating the armature bodies. The 
 next source of loss is "hysteresis," which is a mole- 
 cular opposition to the magnetizing, demagnetizing, andi 
 
42O TESTING OF DYNAMOS AND MOTORS. 
 
 remagnetizing of every part of the core as many times 
 per revolution as there are pairs of poles. This "molec- 
 ular friction," as it were, resists the rapid reversals of 
 polarity and manifests itself as heat. 
 
 The armature core under test is set in a frame and 
 belted to a motor, through which is electrically measured 
 the work done in the system. The instruments required 
 are disposed as follows: A voltmeter across the motor 
 armature and one across its separately excited field; an 
 ammeter in the motor circuit, and one in the field circuit 
 of the naked core under test; a speed indicator or a 
 tachometer. Of the energy given the motor, part is 
 dissipated as I*R losses in motor armature and field, and 
 the rest is expended in turning the two armatures against 
 frictional and other opposing forces. The armature / 2 R 
 loss varies as the field of the core is varied,, but that in 
 the motor field is kept constant throughout the test. 
 This is accomplished by thoroughly heating the fields 
 before the test, and by keeping the applied E. M. F. 
 constant during the test: or better to keep the field cur- 
 rent constant by means of a field rheostat and low read- 
 ing ammeter. The frictional losses are air fanning, 
 motor brushes, bearings, and belt tension. As the speed 
 is kept constant, frictional losses may be regarded as 
 constant also. The power given to the motor = im- 
 pressed E. M. F. X armature current, = gross power 
 consumed. Call the motor A, and the core with its 
 separately excited field, B. To separate the friction 
 losses of A and B, A is run free at the speed which would 
 be necessary to run B at its proper speed, and the power 
 measured. Subtract from this the 7 a R loss in A and 
 A's bearing and brush friction loss is left. With the belt 
 
MISCELLANEOUS TESTS. 421 
 
 on, the speed is adjusted and the power again measured. 
 The difference between the two measurements is due to 
 ^'s bearings and the belt tension. In the last case care 
 must betaken that B has no field, otherwise the apparent 
 loss will be too high. 
 
 The test proper now begins, and consists in putting 
 variable field currents through J?s field, and noting the 
 power consumed by A while running at the proper speed. 
 As v9's field current increases, so do the induction cur- 
 rents, etc., in the core, and A must be supplied with 
 more energy, or the speed will fall. To facilitate speed 
 regulation without disturbing A's field, its armature is 
 supplied from a variable source of E. M. F., as, for 
 example, a dynamo whose field regulation affords a ready 
 means of altering its E. M. F. For this test, this means 
 of regulation has the following advantages: (i) There is 
 less liability of exceeding the motor's current carrying 
 capacity if the field is left strong; (2) the field loss on 
 A being kept constant, it does not enter into the calcula- 
 tions except as a constant, and its effect on the other 
 factors is always the same; (3) by making the initial 
 field strong, the armature current necessary to supply 
 the required work is small, the impressed E. M. F. being 
 made correspondingly great, and a lower reading ammeter 
 can be used, thus lessening the errors of adjustment and 
 observation; (4) there is freedom from sparking; (5) 
 with a constant field, the armature E. M. F. can be used 
 to check up the accuracy of the speed readings. Should 
 there be an abnormal increment in the voltmeter reading 
 at any point, the indication is that there has been error 
 in taking the speed. There is quite a trick in handling a 
 tachometer properly: in the first place, it should not be 
 
422 TESTING OF DYNAMOS AND MOTORS. 
 
 leaned against, because it injures the instrument and adds 
 slightly to the load. It should be held lightly and level. 
 To decrease the pressure necessary to keep it from 
 slipping, its point should be covered with soft tape 
 so as to fit the hole in the shaft, and both hole and 
 point should be well chalked. The tachometer should 
 be level, and tilting it either way lowers its reading, 
 because the instruments generally depend upon the 
 principle of centrifugal motion, and therefore give their 
 true and maximum reading when the axis of rotation is 
 vertical, and the force of gravity properly directed. 
 
 There must be considerable range of voltage available, 
 and, if necessary, two machines in series to supply it. 
 The instrument in the motor field, whether volt- or 
 ammeter, need not be very accurate, as it is only required 
 to indicate its initial deflection. A's armature resistance 
 must be known so its / 2 R loss can be accurately figured 
 for the different current values. It is best to heat A 
 throughout before using, then its armature resistance, 
 hot, can be measured and can be assumed to remain 
 constant throughout the test. The frictional losses 
 having been determined, full field is put on B, the speed 
 adjusted and a power reading taken. Subtracting from 
 this the frictional loss, and the / 3 R loss in A, the core 
 loss remains, 's field current is then decreased by 
 regular steps, till zero is reached, when it is reversed 
 and increased in the same manner. The results should 
 be tabulated in parallel columns, and the core loss deter- 
 mined for each current value. A curve is then plotted, 
 in which the horizontal scale gives the current values, 
 the vertical scale the corresponding core losses. The 
 readings are started with full field on JE>, to insure that 
 
MISCELLANEOUS TESTS. 423 
 
 the facilities at hand are adequate to maintain the speed 
 for all loads without excessively overloading A or exceed- 
 ing the ammeter's range. This test when interrupted 
 cannot be resumed where the readings left off, but must 
 be run over again in order that the influence of the 
 residual field may be uniform throughout. The above 
 method secures a regular variation from full positive field 
 to zero, back to full negative field. If the current is 
 carried in the same steps back to the starting point, we 
 complete a cycle of magnetization. 
 
 TEST XIII. Eddy Current Test. The above test sepa- 
 rates the electrical, magnetic, and mechanical losses, but 
 gives no detailed information in regard to hysteresis, 
 or to eddy currents. To determine these, it is neces- 
 sary to run the completed armature as a separately 
 excited motor at at least two different speeds. The 
 losses due to friction and to hysteresis vary (approxi- 
 mately) directly as the speed; /. <?., if the speed is 
 doubled the losses are doubled. The loss due to 
 foucault or eddy currents varies as the square of the 
 speed; /". e., if the speed is doubled, the loss is quad- 
 rupled. Calling the total loss JF, it is divided as 
 follows: friction W v \ hysteresis F H , eddy currents JF E , 
 and we may write, 
 
 The first two were determined in Test XII, so that 
 they are known quantities. If now, with the same 
 field excitation the voltage is raised till the speed is 
 doubled, we have the following equation: 
 
424 TESTING OF DYNAMOS AND MOTORS. 
 
 Combining this with equation (i), we get, 
 
 W' - 2 W = 2 W^ ............ (3) 
 
 or 
 
 W z =~- W, 
 
 which gives the eddy current loss without knowing the 
 friction loss or hysteresis loss, but knowing their com- 
 bined losses, W. The above expression supposes the 
 second speed to be double the first; this is not a neces- 
 sary condition. Let us suppose the new speed to be K 
 times the old speed, where K is either a whole number 
 or a fraction, Eq. (2) then becomes, 
 
 W = KW W + KW n + K*W m ..... (4) 
 Eq. (i) is W = W w .+ W n + W z : 
 
 multiplying Eq. (i) by K, we get 
 
 KW = KW Y + KW* + KW^ ....... (5) 
 
 subtracting Eq. (5) from Eq. (4) we get 
 
 W - KW = K^W^-KW* ......... (6) 
 
 or 
 
 W - KW- KW*(K - i), 
 or 
 
 w _W'-KW 
 
 W * K(K - i)' 
 
 Then suppose K = 2. Then, 
 
 as above. 
 
MISCELLANEOUS TESTS. 4 2 5 
 
 Let K = 3/2, then 
 
 , FE = iv ' - 3/*jr - 
 
 3/2 (3/2 - i) 3/4 
 
 80" - 12 
 
 In all cases several speeds should be taken and the 
 values for JF K compared. 
 
 In core loss tests run under full load, as they often are, 
 it is desirable to know how much power is consumed in 
 driving the system when the generator field is very weak 
 and there is full current in the armature. Under these 
 conditions the brushes are brought well forward to 
 eliminate the sparking, and the armature reaction is so 
 great that there is liability to reversal. If the machine 
 is shunt-wound, it drops its field when reversal takes 
 place, or even before, because greatly increasing the 
 field circuit resistance has the same effect as greatly 
 decreasing the line resistance. If compound-wound or 
 separately excited, it cannot do this. On a compound- 
 wound machine the field picks up with the reversed 
 polarity. On a separately excited machine the polarity 
 cannot thus be permanently reversed, but is over- 
 powered for the time being. If the above investigation 
 is the only object of the test, it may be more convenient 
 to run the generator back on a motor, and supply the 
 loss from an engine belted to both. In this case the 
 result of a reversal is to throw the two machines in series 
 as generators, with a short circuit through the two 
 armatures. An ammeter in the dynamo self-excited 
 shunt-field circuit will be found to have reversed its 
 
426 TESTING OF DYNAMOS AND MOTORS. 
 
 deflection after the field has picked up to suit the new 
 conditions. 
 
 TEST XIV. Saturation Test. The next test to be 
 considered is the saturation test, so called because 
 it shows when the field cores are magnetically satu- 
 rated. This test is usually, though not necessarily, 
 run in conjunction with the core loss test, and 
 determines the extent to which it is profitable to 
 expend energy in magnetizing the fields of a dynamo. 
 When a machine under test develops a lower E. M. F. 
 than experience with that type leads one to expect, the 
 natural inference is that an inferior quality of iron has 
 found its way into the frame. A saturation curve of this 
 machine, when compared with those of other machines, 
 will show whether this inference be true. The test is 
 run with the assumption that the more common sources 
 of error, connections, loose joints, windings, crosses, etc., 
 have been eliminated. The test consists in running the 
 armature free at proper speed, and in separately excited 
 fields. An ammeter is placed in the field circuit, and a 
 voltmeter across the brushes. The field current is grad- 
 ually increased from zero to a point where for a given 
 increase in field current the increase in E. M. F. is about 
 constant. As the field current is increased step by step, 
 the increase in E. M. -F. is also recorded, and a curve 
 plotted, with field current values as abscissas and E. M. Fs. 
 as ordinates. Suppose five amperes to be the limit to 
 which it is safe to temporarily carry the field current: 
 commencing with zero field current, the E. M. F. due to 
 residual field is read, and then the field current increased 
 in steps of one-fourth ampere, and the E. M. F. read at 
 each step. The E. M. F. will be found to increase rapidly 
 
MISCELLANEOUS TESTS. 427 
 
 at first, and finally by equal amounts. Having reached 
 the upper limit the current is decreased by the same steps 
 until zero is reached, when the current is reversed and 
 the operation repeated, thus completing the cycle. It is 
 important that the field current should not be broken 
 throughout the test, save at the zero readings; otherwise 
 the curve will not be a smooth one, and the test must be 
 repeated. For a similar reason care must be taken not 
 to exceed a current value in passing to it, and if it is ex- 
 ceeded not to return to it, but note its actual value and 
 pass to the next. The reason for this is that as the cur- 
 rent increases the iron tends to hold its magnetization, 
 and if the current be interrupted this retaining power 
 will not immediately resume its former value. In reced- 
 ing to a reading, the reading is likewise influenced by the 
 higher degree of magnetization due to the current receded 
 from. This retaining power is illustrated by the differ- 
 ence between the up and down readings of voltage, the 
 latter of which for the same current are always higher 
 than the former. In this test it is all important that the 
 speed be maintained constant, and that the volt lines have 
 good contact. 
 
 The rapid increase of E. M. F. at first shows the iron 
 to be far from saturation, but as the " knee " of the curve 
 is reached the iron becomes more highly magnetized, 
 and just around the knee is said to be saturated. 
 Beyond this point the increase in magnetization is not at 
 all proportional to the increase in magnetizing current, 
 and it is uneconomical to use a larger current than that 
 corresponding to the point just beyond the knee. 
 When the curve is plotted it will be found that the 
 descending curve lies above the ascending one and does 
 
428 TESTING OF DYNAMOS AND MOTORS. 
 
 not cross the zero of E. M. F. till the current has been 
 reversed. The descending reading of E. M. F., taken 
 when the current is zero, is a measure of the retentive 
 power of the iron, and may be called its temporary field. 
 If left alone for some time this effect disappears, and the 
 residual magnetism resumes its original value. The 
 reversed current necessary to reduce the E. M. F. to 
 zero before it changes sign, is a measure of the iron's 
 coercive force, and is but another way of measuring the 
 retentive power. All instruments should be calibrated 
 immediately before the test, and should be provided with 
 push buttons or cut-out switches to relieve them when 
 not being read, and thus minimize the effect of tempera- 
 ture variation. Instruments should not be near enough 
 to influence each other, and should be removed as much 
 as possible from external magnetic influences. The test 
 for such errors is to turn the instrument around and 
 observe if its own or any other instrument's deflection i& 
 changed. They must be calibrated after test to insure 
 that no jar or exposure has introduced an error. Once 
 a test is started all existing conditions must remain the 
 same; even those which might have been corrected at 
 first must now remain undisturbed. Oil must not be used 
 on the commutator of the dynamo, as its variable resist- 
 ance will cause a marked difference in the voltmeter 
 readings. Very often it is required, after the free satu- 
 ration test above, to run one with variable currents in the 
 armature: by comparing the two the effect of armature 
 reaction is readily seen. Finally, if for any reason the 
 test is shut down before completion, the partial data is 
 of no value, and the whole test must be run over. 
 
 TEST XV .Distribution Test. This test is related to 
 
MISCELLANEOUS TESTS. 429 
 
 Test XIV. in that both have to do with the magnetic field. 
 Its object is to determine the character and strength of 
 the field all around the commutator; to determine the 
 field's variation from point to point, and whether this 
 variation is such as would be expected or could be 
 improved by altering any of the details of construction. 
 The test is run first with the armature free, so as to 
 determine the normal character of the field, then at in- 
 termediate and full load, 
 to determine the law of in- 
 crease and the full effect 
 of the armature reaction. 
 There are two principal 
 methods, either of which 
 may be adopted: i. One 
 terminal of a voltmeter is FIG. 132. 
 
 fastened to one of the 
 
 brushes, the other terminal, attached to an exploring 
 brush, is moved around the commutator, bar by bar, till 
 it reaches a brush of opposite sign. This is known as 
 Mordey's method, and its connections are given in Fig. 
 132. A is the positive, B the negative machine brush, 
 and C the explorer. V is a voltmeter, one of whose 
 terminals is secured to A, the other to C. Starting with 
 A and C on the same bar, C is worked bar by bar toward 
 B. At the start no deflection obtains unless Vis very 
 sensitive, and on some machines the explorer includes 
 several bars before a readable deflection is obtained. As 
 C progresses toward B, the deflection increases, and 
 upon reaching B t should give the full voltage of the 
 machine. When running the test under load, two volt- 
 meters may be used; one, V, used with the explorer, the 
 
430 TESTING OF DYNAMOS AND MOTORS. 
 
 other, V' t being permanently connected across A and B. 
 V will, under load, show a higher maximum deflection 
 than V , because the latter does not include the IR drop 
 due to brush contact. After the "flat" or neutral field 
 is passed, the deflection increases slowly, then more 
 rapidly, and reaches its fastest rate of increase at that 
 point on the commutator which introduces the most 
 active coil. After this point is passed the deflection 
 continues to increase, but at a decreasing rate, till, when C 
 approaches B, the inclusion of an additional bar does 
 not augment the voltage at all. A double scale voltmeter 
 facilitates the reading of the lower variations near the 
 brushes. It is well to move V's fixed terminal to B, and 
 explore in the reverse direction. Upon exploring under 
 load, it will be found that distortion has taken place, and 
 that the field is no longer symmetrical, but is massed 
 toward the forward brush. This distortion is sometimes 
 so very great that it is impossible to find a strictly non- 
 sparking point for the brushes. This means that the 
 lines of force are so crooked that there is no point at 
 which the conductors slide along them without cutting 
 them. On one machine sold on the market, Mordey 
 found that at one point the inclusion of an additional 
 bar lowered the reading, showing that the E. M. F. 
 generated in that part of the field opposed that in the 
 other part. 
 
 To facilitate the movement from bar to bar, a rack or 
 frame is made for C to slide in, and the face of the rack 
 is graduated into division equal to the width between 
 the centres of consecutive bars. By means of a second 
 voltmeter, V' , connected across A and B, the machine's 
 voltage is kept constant during the test. Even with the 
 
MISCELLANEOUS TESTS. 
 
 431 
 
 best of the care, F's needle will vibrate more or less, 
 due to the variable contact resistance of the small 
 brush. 
 
 To obviate this inconvenience, Swinburne's modifica- 
 tion of Mordey's method maybe used. Connections are 
 as in Fig. 133. A and B are the machine's brushes; C, the 
 explorer's; A C B the part 
 of the commutator included 
 between brushes of opposite 
 sign; G is a galvanometer 
 whose constants need not be 
 known; R, a resistance box 
 used to protect G and to 
 vary its sensibility; /and H 
 are resistance boxes, either 
 of which can alone stand 
 the machine's brush voltage; 
 V l and V^ are voltmeters reading the potential differ- 
 ence across / and H respectively, while 1\ is another 
 for maintaining the machine's E. M. F. constant. 
 The adjustment for taking readings is based on the 
 following fact: brush A is positive, /?, negative, so 
 that in passing from A to B, whether we go through 
 the commutator A C B, or the outside circuit AMI 
 H N B, we must pass through a point of zero potential, 
 /. e., in every circuit there is an internal and an exter- 
 nal point which are at the same potential and a galvan- 
 ometer joining two such points would not be deflected. 
 The procedure is as follows : When C is at A, Cand Z>, the 
 two ends of G, are at the same potential, provided there 
 is no resistance in box /, so that no current flows through 
 G. V^ then reads the same as V f As C is moved toward 
 
 FIG. 133 
 
43 2 TESTING OF DYNAMOS AND MOTORS. 
 
 B, G's terminals are at different potentials, and G deflects 
 upon pressing its key. Such resistance is then put in / 
 as will make the drop across it the same as that from A 
 to C. Pressing the key will then no longer cause a 
 deflection, because (7 and .D will -be at the same potential. 
 In this way V^ reads the increasing, F 2 the decreasing, 
 voltage, while their sum equals the reading of V z . 
 
 Care must be taken that the combined resistance of 
 / and H is not so much reduced as to cause an injuri- 
 ous flow of current: the adjustment is effected not by 
 decreasing the total resistance of the two boxes, but 
 by shifting resistance from one to the other, the total 
 remaining the same. 
 
 2. The second method of exploring the magnetic field is 
 due to Professor S. P. Thompson, and consists in attaching 
 
 the voltmeter terminals 
 
 ^^ ^"\^ to two brushes, held far 
 
 A ^ -^ B enough apart to span 
 
 FIG. 134. the insulation between 
 
 adjacent bars. Mov- 
 ing these brushes from bar to bar, the voltage of 
 each single coil is measured, and the sum -of these 
 voltages gives that of the machine. Any marked dis- 
 tortion is thus readily detected. This method has the 
 advantage that it gives directly the result sought, while 
 the other requires more or less calculation. The results 
 of a distribution test can be graphically shown by letting 
 the horizontal scale represent the distance from the 
 positive to the negative brush, and the vertical scale the 
 E. M. F. generated at each point. The curve will then 
 rise to a maximum and fall to zero. Or the vertical 
 scale may represent the total E. M. F. from one brush 
 
MISCELLANEOUS TESTS. 
 
 433 
 
 to the given bar; in this case, the curve will always rise, 
 but more rapidly in some places than in others. Still 
 another way is to represent the armature by a circle, and 
 to draw the curve of E. 
 M. F. around it, consider- 
 ing the circumference of 
 the circle as a base line, A- 
 
 FIG. 135. 
 
 and the vertical scale as 
 
 radial to it. Figs. 134, 
 
 *35 J 36 illustrate these methods. For more extended 
 
 information, the reader is referred to S. P. Thompson's 
 
 Dynamo- Electric Machinery. 
 
 TEST XVI. Brush Test, One other test, simple to 
 execute, and quite necessary, is the brush test, to determine 
 what thickness of brush will, on a given style of machine, 
 spark least: also what cross-section is necessary to carry 
 the required current without undue heating. The brush 
 thickness best adapted to a given machine is determined 
 experimentally in each case. Some of the factors 
 modifying the thickness are: speed, field 
 strength, width of pole pieces and number 
 of turns of wire per armature section, all 
 of which affect more or less the voltage 
 generated in the sections short circuited 
 by the brushes. 
 
 TEST XVII. Efficiency Test. Having 
 considered in turn the several methods 
 of conducting temperature tests, the 
 compounding, core loss, saturation, distribution, and 
 brush test, there follows very naturally the efficiency 
 test, as conducted on machines before they were put on 
 the market. The term efficiency is used in two senses: 
 
434 TESTING OF DYNAMOS AND MOTORS. 
 
 1. The electrical efficiency, a term having to do with the 
 machine considered purely as an electrical device; 
 
 2. Commercial efficiency, having to do with the machine 
 considered as an energy transforming device, and 
 expressed by the fraction obtained by dividing the 
 machine's available output by its actual intake. If an 
 engine delivers 5,000 foot-pounds of work per second to 
 the dynamo's pulley, and only 3,000 foot-pounds are 
 available for use in the working circuit, the commercial 
 efficiency is 
 
 5,000 
 
 a rather low figure. The other 40 % has been consumed 
 in electrical, magnetical, and frictional losses. It is 
 possible, however, to determine exactly how much elec- 
 trical energy is produced in the dynamo, and also what 
 proportion of this is delivered to the line, and what pro- 
 portion is wasted in the machine. The quotient arising 
 from dividing the energy delivered to the line by the 
 total amount produced in the electrical circuit con- 
 stitutes the electrical efficiency. This ignores the iron 
 losses and frictional losses, and is therefore greater than 
 the commercial efficiency. In the above case, suppose 
 4,000 foot-pounds per second of electrical energy is pro- 
 duced, and 3,000 are delivered for use. The electrical 
 efficiency is then 
 
 - 75 *. 
 
 4,000 
 
 It is customary to express electrical energy in electrical 
 units, and since one foot-pound per second = 1.3 watts, 
 the total watts would be, 5,200; watts delivered = 3,900. 
 
MISCELLANEOUS TESTS. 435 
 
 Electrical efficiency - = 75 % 
 
 as before. The commercial efficiency is the item which 
 interests purchasers, for the electrical efficiency is often 
 misleading, inasmuch as it is sometimes advantageous to 
 sacrifice electrical efficiency to gain some other good 
 point more requisite. A direct driven dynamo may have 
 an electrical efficiency of 95 %, and its commercial effi- 
 ciency fall as low as 75$. There is still a third efficiency, 
 which for want of a better name we will call the "coal 
 pile efficiency," and this involves directly the amount 
 of coal used per watt-hour of electric energy available. 
 It is not until this " over-all " or " plant " efficiency is 
 raised to its highest value that electrical stocks will 
 yield dividends commensurate with the theoretical 
 advantages of electrical transmission. 
 
 To return to the question in hand: to determine the 
 commercial efficiency, we must carefully measure the 
 intake and output of the machine. The last divided by 
 the first gives the expression sought. There are several 
 mechanical, and as many electrical, methods of doing this. 
 In testing-room practice it is customary to run the 
 machines as in Test X, where the loss is supplied 
 from a dynamo. Fig. 137 shows the connection of the 
 machines and the disposal of the instruments. A is a 
 dynamo whose efficiency is to be determined; B is a 
 motor on which A is to run back, and whose voltage 
 must be the same as A's-, C is the loss supplier. A and C 
 each have an ammeter: their combined currents is that 
 of B. To obtain greater accuracy, a standard shunt and 
 galvanometer can be conveniently placed for checking 
 
43 6 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 up the readings of both ammeters. A voltmeter is 
 arranged to be thrown across the terminals of any 
 machine, or better still, three voltmeters are used. As 
 a preliminary step determine the internal resistances of 
 A and B. This is done by either blocking the 
 
 pulleys or reversing half the fields on a machine, 
 so that it cannot start, and then sending a current 
 through it by means of C, and reading the drop 
 on the respective terminals. The load is next put 
 on as in Test X, and the speed and voltage of A 
 adjusted. Calling JS, A's E. M. F. , and /its current, its 
 useful output is E x /, which gives the numerator of the 
 fraction expressing either the electrical or commercial 
 efficiency. The next step is to measure the total energy 
 supplied to A mechanically by B. To do this it is 
 necessary to measure the energy delivered to B, and to 
 subtract from this the energy lost in B itself: this 
 necessitates running a core loss test on B, and at the 
 same voltage, current, and speed as it has in the 
 efficiency test, otherwise the magnetic and frictional 
 losses on the two machines cannot be separated. /?'s 
 energy is from two sources, A and C. That furnished 
 by A is A's useful output less the loss in the connecting 
 
MISCELLANEOUS TESTS. 437 
 
 cables this cable loss is the drop in the cables (in this 
 case equal to the difference in readings of VM^ and 
 F J/J multiplied by the current /flowing in them. The 
 energy furnished by C is the voltage registered on V M v 
 X C's current: or \V E' i. Also we have the energy 
 from .4 = ( E cable drop) x/, or JF= (E - d) /. Then 
 W + W = E' i + E I - d I = E ' i + (E - d} I; i. e. t 
 the total energy given to B W -f- IV = E' i -j- 
 (E - d) /, or JF T = E' i -f (E - d) I. From this 
 expression we must deduct whatever work is lost or 
 expended in the motor itself. To follow the steps more 
 clearly we give a form of test sheet with actual figures 
 appended. We assume all resistances and core losses, 
 etc., to have been preliminarily determined. Internal 
 resistances are measured between the machines' termi- 
 nals so as to give the multiple resistance of fields and 
 armature. If compound-wound, the shunt winding is 
 connected "long shunt." All of the machines are 
 self-excited. 
 
 r = A's internal resistance = .015 ohm. 
 
 /= A's current = 1,000 amps. 
 
 r'= B's internal resistance = .015 ohm. 
 / = C's current 200 amps. 
 
 I -\- i B's current = 1,200 amps. 
 E = A's voltage (FJ/J 500 volts. 
 
 E' = ^'s voltage (FJ/J 498 " 
 
 d- Cable drop ( FJ/, - FJ/J = 2 " 
 
 W= A's useful output 
 
 = 500 X 1,000 = 500,000 watts. 
 W' = Energy from C 
 
 = 498 X 200 = 99,600 " 
 
438 TESTING OF DYNAMOS AND MOTORS. 
 
 W -\- W = Total energy furnished 
 
 to B. = 599,600 watts. 
 
 w = Core loss, etc., in B = 30,000 " 
 
 W-\-W w = Energy expended on A = 569,600 " 
 
 A's commercial efficiency, 
 W 
 
 W + W - w 569,600 
 
 The electrical efficiency = useful electrical output 
 divided by total electrical output. / A's current 
 1,000 amperes; r = A's resistance = .015 ohms. /* r 
 = 1,000 X 1,000 .015 = A's internal loss (electrical) 
 15,000 watts. W + 7 2 r = A's total production = 
 515,000 watts. 
 
 W 
 
 the electrical efficiency. 
 
CHAPTER XIII. 
 
 GROUNDS ON THE LINE. 
 
 THE most formidable difficulty with which engineers 
 have to contend is the maintenance of good line insula- 
 tion. Upon the effectiveness of this insulation depends 
 the smooth running of the station, and in a measure its 
 dividends. However carefully the work of installment 
 may have been done, and however efficient may be the 
 daily attendance, grounds and crosses are always liable 
 to occur, and generally do so at most inconvenient times. 
 
 In the present chapter it is the writers' aim to give the 
 more common and efficient methods of detecting faults, 
 and the more practical methods of removing them. To 
 detect a ground is the office of a device known as a 
 " ground-detector." The method of locating and remov- 
 ing a ground depends upon its nature and gravity. 
 
 The presence of a single ground in no way impairs the 
 service of a metallic circuit; if one point of a perfectly 
 insulated circuit is earthed, there is a momentary flow 
 of current which causes a re-distribution of potential 
 throughout the system, and brings the point in question 
 to the earth's potential, whereupon the equalizing current 
 ceases to flow. In many cases the ground detector itself 
 constitutes the harmless ground, and depends for action 
 upon the occurrence of a second ground elsewhere on 
 
 439 
 
440 TESTING OF DYNAMOS AND MOTORS. 
 
 the system. Fig. 138 gives the connections of a common 
 form of detector. A and B are the two mains of a two- 
 wire metallic circuit; Z, Z', two incandescent lamps in 
 series across them. L and L are as nearly as possible 
 of the same resistance, and under 
 ordinary conditions burn at the 
 same brilliancy. From the junc- 
 tion C of the two lamps passes to 
 earth a wire C, S, G, including a 
 low resistance bell. Ordinarily 
 C and G are at the same po- 
 tential, and no current flows between them; if, how- 
 ever, a ground occur, as at G f , C's potential becomes 
 different from that of the earth, a current flows through 
 G C and operates the signal bell, continuing to do so until 
 the detector circuit is opened. The bell rings, and one 
 lamp burns brighter than the other. Which lamp burns 
 brighter depends upon which side of the line contains the 
 fault. If it is on A, L brightens, but if on B, L does so, 
 and L grows dimmer, because the second ground at G' 
 places the fault in multiple with Z', and the latter does 
 not get as much current as it would if no ground existed. 
 If G' is a ''dead ground," Z' is short circuited and Z 
 gets the entire line voltage, because Z's terminals are in 
 multiple with those of G', whose resistance is practically 
 zero, and since the potential difference between A and 
 B distributes itself according to the resistance, no drop 
 will take place between the terminals of Z', and Z is sub- 
 jected to the entire line voltage. For this reason each 
 lamp should be of the full line voltage. 
 
 The effect of an external ground, then, is to place in 
 multiple with one lamp a by-path whose shunting power 
 
GROUNDS ON THE LINE. 441 
 
 depends upon the extent of the fault; a slight ground 
 dimming the lamp with which it is in multiple, a dead 
 ground extinguishing it altogether. 
 
 According as A or B is faulty the current in C, S, G 
 will be toward the earth or from it, and depending upon 
 this fact, a pair of high resistance electromagnets actu- 
 ating a needle over a graduated scale, or a pair of volt- 
 meters, can replace the lamps and indicate the faulty 
 main, also giving a fair idea of how serious the fault is. 
 If the ground wire C, S, G, contain a voltmeter adjusted 
 to read right and left from a zero in the centre of the 
 scale, or if it contain an electromagnet operating a needle 
 free to move either way from a central position corre- 
 sponding to zero current, connections are simplified. 
 The objection to such detectors is that they sometimes 
 give misleading indications. For instance, suppose that 
 ordinarily one lamp burns a little brighter than the other, 
 indicating one main to contain a slight ground: and that 
 upon a certain day the discrepancy disappeared, the 
 lamps then burning at the same brilliancy. This 
 behavior would indicate either of two things, that the 
 existing ground had disappeared or that a similar ground 
 had developed on the other main. In the first case the 
 lamps would burn alike because neither would be shunted 
 by a fault, and in the second case they would burn alike 
 because they would be shunted by faults of equal resist- 
 ance; in other words, a fault on A would tend to send 
 through (7, S, G, a current from the earth, while a similar 
 fault on B would have an equal tendency to send through 
 C, 6", G, a current toward the earth, with the result that 
 no current at all would flow through C, S, G, and the 
 lamps would burn as if no ground existed. To test the 
 
44 2 TESTING OF DYNAMOS AND MOTORS. 
 
 true condition of affairs we must disconnect the ground 
 wire, and by means of a voltmeter or ammeter, one of 
 whose terminals is earthed as at G, Fig. 139, touch alter- 
 nately A and B with the free end. If the two deflections 
 are the same, and abnormally high, it indicates a ground 
 on both mains; if there is no de- 
 flection at all, or if the deflections 
 are equal, but low, it indicates a 
 clear line, except in so far as leakage 
 has its influence. In order that 
 even these indications may be 
 reliable, instruments must be se- 
 lected to suit the conditions. As an aid to this selec- 
 tion the following discussion may not be out of place. 
 
 If, as in Fig. 140, we suppose that we have a telegraph 
 line extending from A to B, and supported on ordinary in- 
 sulators having a resistance of many millions of ohms each, 
 we might expect all the current from the battery at A to 
 pass through the primary of the relay at B, but it is a 
 well-known fact that it does not, hence the necessity of 
 the relay. The reason for this is that although the leak- 
 age over the surface of a single insulator might be very 
 small, when we have a great many insulators in use 
 more leakage paths are placed in multiple, and on long 
 lines so much current leaks from positive to negative by 
 way of the by-paths indicated by the dotted lines, that 
 that which reaches B is insufficient to operate the 
 sounder unless the relay is used, and this leakage effect 
 is most marked in damp weather. Again, if we take a 
 commutator completed, and, by means of one of the 
 voltmeter methods already described, measure the re- 
 sistance of the insulation between each bar and the 
 
GROUNDS ON THE LINE. 
 
 443 
 
 shell, it will be found to be very high. Probably the 
 needle of the voltmeter will just wiggle; if now by means 
 of a fine bare wire passed several times around the com- 
 mutator we connect all the bars together, the insulation 
 to the shell will be much lower, as indicated by the in- 
 creased deflection. Each section of an armature winding 
 will test high to the core, but the completed armature 
 will test much lower, and toothed armatures test lower 
 than smooth ones because more surface of wire is exposed 
 to the core. Take two 500 volt test lines and hold them 
 
 1 
 
 
 
 
 
 FIG. 140. 
 
 on opposite sides of a sheet of asbestos 1/16 inch thick, 
 and with no previous drying a small deflection will 
 obtain; next place the sheet of asbestos on a smooth 
 iron surface and put a second sheet of iron upon it, and 
 test between the iron surfaces. The deflection will be 
 found to have greatly increased, and to a degree depend- 
 ing upon the extent of the surfaces exposed. 
 
 We thus appreciate the fact that many high resistance 
 paths placed in multiple soon afford a path of tolerably 
 low resistance, and it is more easily seen how, on exten- 
 sive networks of wiring supported on insulators, fed from 
 many machines, and regulated from switchboards con- 
 taining many devices of large surface, the insulation to 
 earth may be as low as 50 or 100 ohms. On the other 
 hand, on less complicated systems the insulation to earth 
 
444 TESTING OF DYNAMOS AND MOTORS. 
 
 may run up into thousands of ohms. Between these 
 limits some instruments are better adapted than others 
 for testing insulation during hours of service. We know 
 that the leakage path on any system has a definite 
 resistance, which can be expressed in ohms. Suppose, 
 Fig. 138, that the insulation resistance from A to earth 
 is 50 ohms, and that that from B to earth is also 50 ohms: 
 then the total earth resistance from A to B is 100 ohms, 
 and a line voltage of 100 would cause a leakage current 
 of i ampere, but the total insulation resistance of the 
 system is that of the two leakage paths in multiple, or 25 
 ohms, and this is the value we would get were we to 
 measure the system's insulation after working hours by 
 means of a voltmeter. 
 
 Let us suppose that we have two mains, A and B, Fig. 
 141, and that one of them is free from leakage and per- 
 fectly insulated, but that the other 
 has leakage paths whose aggregate 
 resistance is 56 ohms. So long as 
 B remains perfectly insulated there 
 will be no leakage current, because 
 there is no path through which it 
 FIG. 141. can reach B. Now suppose with 
 
 one terminal earthed, we place the 
 
 other terminal of a 1,000 ohm voltmeter upon B, thus 
 grounding B through the resistance of the voltmeter. 
 There is then a complete circuit, A, G', G, V, B, through 
 the leak and the voltmeter, and the two are in series, so 
 that the line voltage divides between them proportionately 
 to their respective resistances. Since the meter resistance 
 is 20 times that of the leak, 20 times as great a drop will 
 take place across it; /. ^., if the potential difference be- 
 
GROUNDS ON THE LINE. 445 
 
 tween A and B is 105 volts, the meter will show 100 volts, 
 the remaining 5 volts dropping through the leak, whose 
 resistance we will call r. In other words, in this particular 
 case, the drop through each is proportional to its resist- 
 ance. Call the meter resistance R, the line voltage F, 
 and the meter deflection d. Then in any case the drop 
 across r will be V - d, and from the above analysis we 
 have the proportion 
 
 d : V - - d ; ; R : r, 
 whence 
 
 We know or can ascertain R, we get d in the test, and 
 can measure V. The above case is assumed, to simplify 
 reasoning, and is not likely to arise in practice, except on 
 isolated plants. In the every-day case where A and B 
 are both more or less faulty, the tests consists in touch- 
 ing the free end of the voltmeter alternately to one main 
 and then the other, and observing the two deflections. 
 If one meter terminal is fixed, the meter must either read 
 both ways from a zero in the centre, or must be provided 
 with a reversing switch, because moving the free end 
 from one main to the other reverses its current. The 
 insulation resistance in this case is not so easily derived 
 from the value of the deflections, because here the volt- 
 meter not only has a fault in series with it, B G', Fig. 142, 
 but has also a fault in multiple with it, A G" , The reader 
 will very readily conceive of the two faults forming a 
 closed circuit, A, G" , G', B, between the two mainland 
 he might further consider the taking of the drop between 
 
446 TESTING OF DYNAMOS AND MOTORS. 
 
 A and G and between B and G, to be simply dividing 
 into two operations what might be gotten by at once 
 taking the drop between A and B. The two faults do 
 form a closed circuit, and moreover the drop from A to G 
 plus that from G to B does equal that between A and B, 
 as can be proven by using an elec- 
 trometer instead of a voltmeter. An 
 electrometer, being an open circuit 
 instrument, is of infinite resistance, 
 and allows no current flow; whereas 
 the voltmeter, being a closed circuit 
 instrument, does not read the drop on 
 the fault alone, but reads the drop 
 on the combined resistance of itself and the fault in 
 multiple. Suppose, for instance, the voltmeter resistance 
 to be 60,000 ohms, that of each fault to be the same, 
 and the line voltage to be 500. Ordinarily, then, 
 there would flow through the two faults in series a 
 current of 
 
 ^_ I _ - 
 
 1200007 
 
 ampere. 
 
 120,000 
 
 As soon as the voltmeter is connected from either main 
 to earth, it shunts one fault and the combined multiple 
 resistance of the fault and meter is only 30,000 ohms, and 
 the meter registers the drop on this. The total resistance 
 in the test circuit is then 60,000 -f- 30,000 = 90,000 ohms, 
 and the leakage current becomes 
 
 = .0055 ampere, 
 90,000 
 
 which causes across the voltmeter terminals a drop of 
 (/ X R = .0055 X 30,000) = 165 volts. Transferring 
 
 (5 I 
 00007 
 
GROUNDS ON THE LINE. 447 
 
 the free end of the meter to the other main, we get a like 
 drop, making the sum of the readings (d ^ -f- ^ 2 ) 330 volts. 
 But we know the potential difference between A and B 
 to be 500 volts, therefore there is a difference of 170 volts 
 between the sum of the successive readings taken above 
 and the line voltage, and a voltmeter would be poorly 
 adapted to measure the potential difference between two 
 widely separated mains, by taking their drops to earth. 
 An electrometer would serve this purpose. On the 
 other hand, an electrometer would not answer the pur- 
 pose in the insulation measurement, because, without a 
 knowledge of the current flowing through the two faults, 
 it would give no indication of use in determining the 
 absolute fault resistance, and would only indicate the 
 relative resistances of the two faults. When the volt- 
 meter is used, its application in the test, as we have seen, 
 disturbs existing conditions, and it is this disturbing 
 influence, together with its causes, that enables us to 
 derive an expression giving the insulation resistance 
 of either or both mains. From the above course of 
 reasoning the conclusion follows that the lower the 
 resistance of the instrument the greater will be its 
 modifying influence, and if the instrument's resist- 
 ance is negligibly low, it practically short circuits 
 the fault with which it is in multiple, and the leakage 
 current which flows is limited in value only by the 
 resistance of the meter and the fault with which it is 
 in series. Voltmeters of negligible resistance are not 
 obtainable, but if we use an ammeter, the test is reduced 
 to the conditions of Fig. 141, and nearly the same course 
 of reasoning follows. Suppose, for example, that the 
 voltmeter of Fig. 141 is replaced by an ammeter of .001 
 
448 TESTING OF DYNAMOS AND MOTORS. 
 
 ohm resistance, and suppose that with the free end of the 
 meter on B, a current of i ampere flows. Now since 
 the resistance of the meter is so low as to short circuit 
 whatever fault might exist on B, this fault, unless itself 
 of very low resistance, would have no appreciable in- 
 fluence upon the total current flowing, and this current 
 would be 
 
 I- E 
 
 -r+le' 
 
 where E is the line voltage, r the resistance of the fault 
 on A, and R that of the meter. From this expression we 
 get 
 
 ,-*_* 
 
 In dealing with a network about whose resistance of 
 insulation we know nothing, it is well to start by placing 
 in series with the meter such an extra 
 resistance as will limit the current to the 
 meter's capacity, if there happens to be 
 a short circuit on one main when the 
 free end of the meter is placed on the 
 
 i other. This method is only absolutely 
 
 u G Q . 
 
 _, correct when the meter resistance can 
 
 SflG. 143. 
 
 be neglected in comparison to that of the 
 fault with which it is in multiple. 
 
 We will now derive formulae applicable, whatever may 
 be the relation between the ammeter and the fault with 
 which it is in multiple. In Fig. 143, let A and B be two 
 mains between which a potential difference of E volts 
 exists. Let the resistance of the ammeter be r. Call 
 the resistance of the faults on A, x, and that on B, y. 
 
GROUNDS ON THE LINE. 449 
 
 Let d be the ammeter reading from A, and d that from 
 B. Then 
 
 is the quantity by which d must be multiplied in order to 
 express the total leakage current value in the first posi- 
 tion of the meter, and 
 
 is the multiplier in the second position when we get a de- 
 flection d . Now the resistance of r and x in multiple is 
 
 i r x 
 
 r x 
 
 and the resistance of rand y in multiple, is 
 I ry 
 
 Therefore, in the first position of the meter the total 
 test circuit resistance is 
 
 rx 
 
 and in the second position, it is 
 
450 TESTING OF DYNAMOS AND MOTORS. 
 
 Since from Ohm's Law 
 
 '-* 
 
 and since in the first case the expression for current is 
 
 '-^. 
 
 and in the second case 
 
 we have, 
 x + r _ E _ E E (r + x) 
 
 ry + xy rx + ry -+- xy 
 
 r + x 
 dividing both sides of the equation by x -J- r, we get 
 
 d_ _ E 
 
 x r x -\- r y -\- xy ' 
 Whence, 
 
 _ dry . 
 
 E d r d y 
 
 Also, 
 
 ry 
 
 and by the above processes, we find that 
 
 d r x 
 
 E d n r d n x ' 
 
GROUNDS ON THE LINE. 451 
 
 Substituting this value of y in equation (i), we get 
 
 dr -= -j- 
 
 _ E d r d x 
 
 E - dr- 
 
 E - d r - d x 
 
 E - d r - d Q x 
 
 dd rx 
 E dr 
 
 E - d r - d n x 
 
 **-= i. 
 
 dd r x _ 
 -Ed x- Edr+dd r*' 
 
 Clearing of fractions, we have, 
 
 E*x -.Ed rx - Ed x* - Edr x + afe^^^Tz^^i^, 
 
 whence, leaving out canceled quantities, 
 
 E* x - Ed rx - Ed x* - Edrx = o. 
 Dividing through by the common factor Ex, we have 
 
 E d r d x dr Q, 
 transposing, 
 
 E - d r - dr = d x, 
 whence, 
 
 _E-d r-dr_E -r(d +d) 
 d d Q 
 
 and this value of x is the insulation resistance from main 
 A to the earth. If instead of substituting the value of y- 
 
45 2 TESTING OF DYNAMOS AND MOTORS. 
 
 in the equation for x t we had substituted the value of x in 
 equation (2) for 7, the result would have been the insula- 
 tion resistance of B to earth. 
 
 Having the insulation resistance of each side of the 
 line, the insulation of the whole network is the multiple 
 resistance to earth of both sides. 
 
 x= -r(^ + J) (3) 
 
 and 
 
 y = - r -^r- w 
 
 Calling S the total line insulation, we have: 
 i i 
 
 S = 
 
 1 + 1 i 
 
 ^ ~ 
 
 y E d r d r~ E - d r d r 
 
 E d n r - dr ' E d n r d r E d n r d 
 
 E-r(d +d) 
 
 -3C+7- 
 
 A result which can of course be used to obtain directly 
 the line insulation resistance without finding that of each 
 main. The use of a resistance coil in series with the 
 ammeter will not alter the mathematical accuracy of this 
 method; r then stands for the meter resistance plus 
 that of the coil. 
 
 When using a voltmeter to make this test, the analysis 
 is a little different, because as we know nothing of the 
 
GROUNDS ON THE LINE. 453 
 
 currents passing through either the fault or the meter 
 we can find no expression of their relationship; but we 
 can find an expression giving the distribution of the line 
 voltage in the test circuit: /. ^., how much drops through 
 the multiple resistance of the meter and one fault, and 
 how much drops through the fault which is in series with 
 both. Let us use the lettering of Fig. 143, but suppose the 
 ammeter to be replaced by a voltmeter of resistance r. 
 The path of leakage is A, x, G' ', G' t y, B, as before. In 
 thus defining the path of leakage we do not wish to imply 
 that the leakage all takes place between the points indi- 
 cated on the diagram, for as a matter of fact it takes 
 place all along the line, although a specific fault might 
 greatly localize it. Assuming the free end of the volt- 
 meter to be touched to A, the multiple resistance of r 
 and x is 
 
 and the voltmeter indicates the drop across this resist- 
 ance, the remainder taking place across fault y. The 
 total resistance of the test circuit is then 
 
 ~+~ 
 
 when the test line is transferred to main B, the test cir- 
 cuit resistance is 
 
454 TESTING OF DYNAMOS AND MOTORS. 
 
 Since the potential falls from A to B and distributes 
 itself in proportion to the resistance it meets, we have in 
 either case the fact that the drop across the meter is to 
 the drop across the fault in series with the meter as the 
 resistance between the meter terminals is to the resist- 
 ance of that fault, or, expressed as a proportion. 
 
 i.i r -- x r x 
 
 r x d r x r -f- x _ r x 
 
 y ~^^~ y y ~ ry + 
 
 whence clearing of fractions, 
 
 r x (E d) = dry -j- d xy, 
 r x E rxd= dry -\-dxy, 
 rxE rxd dxy = dry, 
 x (r E rd d y) = dr y; 
 
 X = rE-7d-dy 
 
 Transferring the meter to B main, we have, 
 i i 
 
 T i_ r + y ry 
 
 ~~ i .. 
 
 d r . ' y ry r + y r y 
 
 E-d ~ x x x rx+xy 
 
 Clearing of fractions, 
 
 ry(E-d ) = d (rx + xy). 
 Multiplying out, 
 
 ryE ryd = d r x + d xy. 
 
 Transposing, 
 
 ryE r yd .- d Q xy = d r x. 
 
GROUNDS ON THE LINE. 455 
 
 Factoring, 
 
 y (r E r d Q - d x) = d r x. 
 Dividing, 
 
 _ dr x , * 
 
 y '" TE~^r d -d x' 
 
 Substituting this value in equation (6), we have, 
 
 r E r d Q d x 
 
 _ d d rx 
 
 r E r d 
 
 r E r d d x 
 
 - rEd x - r* 
 
 - S 9 - r* Ed - rEd x - r* E d + r*d 
 Clearing of fractions, 
 
 t* E*x - r*Ed x - rEd x* - r* E d x + 
 =^/ y o P* uv. Dividing by r x E, 
 
 r E r d d x r d o, 
 or 
 
 If instead of substituting the value of y in equation 
 (6), we had substituted the value of x in equation (7), the 
 result would have been a value of 
 
456 TESTING OF DYNAMOS AND MOTORS. 
 
 Having these two'values we can as before derive a for- 
 mula for giving at once the insulation of the whole system : 
 
 i_i i 
 
 x v rE rd n rd* rE rd^ rd 
 
 d 
 
 rE rd rd rE rd rd rE rd rd 
 
 rE -rd a -rd _r[E- (4, + *)] 
 d + d d, + d 
 
 Fig. 144 gives the connections for a detector on an 
 alternating circuit system where it is desirable to avoid 
 a permanent ground. A and B are the two mains be- 
 tween which exists an alternating E. M. F., and i and 2 
 are the primaries of two transformers, which are in no 
 way connected to each other, but 
 K i? B either of which can by means of 
 
 sw i tc h K be put to earth at G. L 
 
 and L are lamps in series with the 
 
 respective secondaries, but have 
 no metallic connection with each 
 FIG. 144. other. Suppose a ground to occur on 
 
 B at G '; upon throwing K over to 2, 
 
 there would be no positive indication, because the result 
 would be simply to ground both sides of primary No. 2, 
 one ground at G, the other at G'. Upon putting K to 
 i, however, L will light up, for we now establish a path 
 Ay i, G, G', B by means of which current can pass 
 from A to B, thus exciting primary i, which induces in 
 
GROUNDS ON THE LINE. 457 
 
 its secondary current enough to light L. A fault on A, 
 as at G", would give no positive indication till K touched 
 2, when path B, 2, G, G", A, including primary 2, would 
 be established. The parts of the detector can be so 
 arranged that the lamp nearest the faulty main always 
 lights. 
 
 The alternating system detector described below is 
 due to Picou. Depending for action upon the principle 
 of the condenser, it affords a safe means of continuous 
 indication, involving no metallic connection between 
 either main and the ground. A and B are the 
 mains, Fig. 145; M' and N' are the inside coatings of two- 
 condensers connected to A and B re- 
 spectively; M and N, the outer coat- A , , B 
 ings, are connected together; from their Q-| h-jH f-Q 
 junction, and including a telephone 
 receiver, T t runs a ground wire, G. 
 Under ordinary circumstances, the 'phone 
 emits little if any sound, but upon the 
 occurrence of a ground on either FIG. 145. 
 
 main, a distinct buzzing is heard, and 
 this can be very much intensified by the use of proper 
 devices. The detector can be made to indicate the 
 faulty main by arranging switches to successively 
 remove the condensers from circuit. Suppose a 
 ground to occur at G' on A; if condenser M' M is 
 removed from circuit, the 'phone will continue to buzz, 
 as it is still in circuit with N' JV, one of whose coatings 
 goes directly to main B, the other through the ground 
 to A. If M' M is replaced and N' N removed, T will 
 not buzz, because its terminals are in circuit with a con- 
 denser both of whose coatings are to earth. The fault 
 
458 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 is always found in the main attached to the condenser 
 whose removal fails to stop the buzzing of the 'phone. 
 
 On arc light and other series systems, the two mains 
 are seldom^ together except where they leave and enter 
 the power house. A detector arrangement for such 
 a circuit is shown in Fig. 146, where A is the constant 
 
 current dynamo, L l Z 2 , etc., the 
 lamp load. As it is not safe to 
 permanently earth a high po- 
 tential system; a switch is used 
 to temporarily make this con- 
 nection. It is customary to use 
 a water or gas pipe as the 
 ground, and the ground wire 
 should include a high resistance 
 voltmeter, by means of which 
 can be told the number of 
 
 lamps included between the fault and the detector. If 
 each lamp is adjusted to regulate for a certain current 
 and voltage, say 10 amperes and 50 volts, and with the 
 current right the detector voltmeter should register 300 
 volts, it would indicate the fault to be just beyond the 
 sixth lamp from the detector. If a dead ground exists 
 on an arc lamp circuit, the effect of a second ground is 
 to cut out all lamps included between the two grounds. 
 In the present case the high resistance of the voltmeter 
 circuit obviates this, with the result that the meter reads 
 the drop on those lamps included between the two 
 grounds. Every arc machine is supposed to maintain 
 constant current, whatever the load, so in this case the 
 voltmeter reading depends directly upon the number of 
 lamps included between its terminals. As the meter 
 
 FIG. 146. 
 
GROUNDS ON THE LINE. 
 
 459 
 
 may .be called upon to register very high voltage, it 
 either must be very high reading in itself or must be 
 provided with a multiplier. On arc circuits contain- 
 ing a great many lamps, it is customary to arrange 
 them in sections, so that in case of grounds serious 
 enough to cut out lamps the faulty sections can be 
 readily located. A single external ground will not 
 impair an arc service unless the machine itself be 
 grounded. 
 
 Fig. 147 gives an ingenious form of ground detector 
 recently put on the market by the Stanley Electric 
 Manufacturing Co. of Pittslield, 
 Mass., and is unique in that it 
 not only absorbs no energy, 
 but requires no metallic connec- 
 tion whatever between earth and 
 line. It is a long-known fact 
 that attraction exists between 
 objects at different electrical po- 
 tentials, that if two metallic 
 plates hung from two fine wires 
 attached respectively to the posi- 
 tive and negative poles of any source of E. M. F. 
 be free to move, they will move toward each other, 
 and it is upon this principle that the electrometer 
 is built. The detector in question is a development of 
 this idea, and consists of four fixed metallic vanes, and 
 a movable vane made of aluminum to secure lightness. 
 The movable vane carries a pointer which ordinarily 
 rests at zero. Two of the fixed vanes go to one side of 
 the line to be indicated; the remaining two are con- 
 nected to the opposite side, and the movable vane goes 
 
460 TESTING OF DYNAMOS AND MOTORS. 
 
 to earth. The movable vane is oppositely attracted by 
 each pair of fixed vanes, and ordinarily is at zero on the 
 scale. If a ground occurs, however, the effect is to 
 establish connection between the movable vane and one 
 pair of the fixed ones, and the pointer moves over 
 toward the faulty main. 
 
 The existence of a ground being assured, the method of 
 locating it depends upon the facilities at hand for test- 
 ing, upon the complexity of the wiring, and upon the 
 gravity of the fault. If the fault occur on one of such 
 a network of wires as is found on city incandescent sys- 
 tems, any or all of the following methods can be used to 
 locate it: i. Send an expert lineman out on a tour of 
 inspection. This may sound very commonplace, but it 
 has succeeded in instances while finer methods were in 
 course of preparation. 2. The different branches of the 
 system can be isolated and the faulty one located by 
 means of a bridge, a magneto, or a voltmeter with some 
 insulated source of E. M. F. 3. The fault can be 
 burned out. 
 
 A tour of inspection often reveals some defect in insu- 
 lation a broken insulator, a naked feeder making con- 
 tact with a tin roof or iron bridge or wet trunk of a tree. 
 If the system be underground inspection cannot be very 
 extensive, and one of the other methods must be used. 
 If method No. 2 is resorted to, the engineer, with a 
 map of the system before him, decides which wires can 
 be cut most conveniently. Suppose the system is divided 
 into halves; the test will show in which half the 
 fault lies. The faulty half can be again halved, and the 
 position of the fault found to a further approximation: by 
 further subdivision the limits are further narrowed till 
 
GROUNDS ON THE LINE. 461 
 
 they become small enough to make inspection practicable. 
 On overhead work the magneto is generally satisfactory, 
 but on very extensive systems and on underground 
 work it is not always reliable, because the magneto's 
 current is alternating, and in successively charging and 
 discharging a system of great capacity, sufficient current 
 would flow to ring the bell, thus indicating a ground 
 though none existed. An ordinary magneto turned by 
 hand will give an indicated voltage of from 50 to 100 
 volts, but this can be greatly increased by belting the 
 armature to a hand wheel of larger diameter, or, where 
 practicable, by belting to a shop shafting. Connections 
 for ringing out a fault are given in Fig. 148, where A B 
 is the isolated section of line; a ground exists at G '. 
 M is the magneto, one of whose terminals goes to a gas 
 or water pipe and the other to A B. Upon turning the 
 handle current flows through circuit G, G', B, A^ M, G. 
 Without any ground at G ' the circuit 
 would be incomplete, and the silence 
 of the bell would indicate good insu- 
 lation, assuming the bell to be in good 
 order. To begin with, the leads are 
 held apart and the armature turned to 
 insure that no cross exists in the leads 
 or inside connections. The test is re- 
 peated with the leads held together; the 
 bell should of course ring. The magneto can be re- 
 placed by a bridge or voltmeter. 
 
 On underground work more delicate instruments are 
 generally used. Except at junction boxes, the cables are 
 open to inspection only by digging, and it is a matter of 
 much importance to locate the exact seat of the trouble 
 
462 TESTING OF DYNAMOS AND MOTORS. 
 
 before opening the street. In locating cable grounds the 
 test is much simplified by isolating as far as possible the 
 cable under test. If in cable A D B, of Fig. 149, a fault 
 
 were located apparently at D, it 
 
 would be impossible, from the nature 
 . of the test, to decide if the fault lay 
 
 at D or at some point on the branch 
 ~ B cable E D C; if disconnecting E D C 
 
 removes the fault, then E D C is the 
 (; seat of trouble. The test consists in 
 
 F sending a steady current through A D 
 
 B by means of a dynamo, storage 
 battery, or other steady source of E. M. F., and with a 
 sensitive galvanometer taking successively the drop of 
 potential between A and the earth and B and the earth. 
 The metal sheathing of the cable makes a good earth. 
 The ratio of the two deflections is the ratio of the num- 
 ber of feet included between the fault and the respec- 
 tive ends of the piece of cable under test, so that 
 knowing the total length of the piece enables one to 
 exactly locate the fault. 
 
 In Fig. 150 let A B be a section of cable 150 feet long, 
 and suppose a ground to exist at G. Y is the battery 
 and g the galvanometer. At G, then, the conductor 
 comes in contact with the ground, and at that point 
 assumes the earth's potential. G, G', G", being all 
 grounds, are practically at the same potential, so that 
 placing the galvanometer terminals on A and G', or on 
 B and G", is the same as placing one on G and the 
 other successively on A and B. Throughout the 
 circuit Y, A, Z>, B, F there is a drop of potential, 
 and part of this drop takes place from B to G 
 
GROUNDS ON THE LINE. 463 
 
 and part from G to A. Suppose the galvanometer 
 deflects 60 divisions due to the drop between A and 
 G, and 40 divisions due to the drop between G and B. 
 To locate the fault we must divide the total length (150 
 
 Y 
 
 1 il I 
 
 
 1 H 1 
 A D 
 
 I 
 
 k I 
 
 5 
 
 G' Q 
 FIG. 150. 
 
 Q" 
 
 feet) of the cable into two parts that shall be to each 
 other as 60 is to 40, or as 3 : 2. To get this division 
 arithmetically, suppose the cable to be divided into unit 
 lengths, each unit length containing such a number of 
 feet that 5 = (3 + 2) units shall equal the total length 
 of the cable, and in this case = 150 feet. Each unit 
 length then contains 30 feet. One piece of the cable 
 must contain 3 of these units = 90 feet, and the other 
 part 2 units = 60 feet; /. e., the fault is 90 feet from A 
 and 60 feet from B. 
 
 Another method of locating a fault to within a foot or 
 so depends upon the bridge principle. While no actual 
 bridge is employed, the connections and computations 
 are those of the bridge. The test requires an auxiliary 
 conductor to serve as a return. 
 
 Fig. 151 gives the actual connections for the test, and 
 Fig. 152 a scheme to more readily identify the bridge 
 principle. In Fig. 151 A N S S' is the return cable, and 
 CJ/S'the faulty one. Completing the circuit at the 
 station end is a battery, B, including in circuit a switch, 
 
464 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 K. In multiple with cables A N S S' M C is a piece of 
 cable, A P C, of known length and uniform cross-section, 
 and corresponding to the proportion arms of an ordi- 
 
 M _Js' 
 
 a "6' 
 
 FIG. 151. 
 
 nary bridge. Current leaving battery, B, has choice of 
 two paths, A N S' C and A P C, between which it divides 
 in the inverse ratio of their resistances. M is the fault, 
 and it is required to find its position. P G is a wire 
 making earth, and including galvanometer, g. The un- 
 
 FIG. 152. 
 
 grounded end of g has a contact free to slide along A P C. 
 P is moved along until a position is found, such that 
 when the battery and galvanometer circuits are closed no 
 movement of the needle is observed; under this condi- 
 tion balance is established and 
 
GROUNDS ON THE LINE. 465 
 
 A M : CM :: AP : PC, 
 
 whence it follows from the laws of proportion that 
 AM : A M + CM :: A P : A P + P C. 
 But ,4 M+ CM - A NS C*n& AP + PC=APC; 
 therefore, 
 
 AM : A N S' C :\ A P : A P C 
 and 
 
 Knowing the value of 
 
 A P 
 
 A 
 
 it is necessary to know the length, A N S' C, and their 
 product will be A M, the distance in feet from A to M. 
 Should A N S' C contain two sizes of wire, its total 
 length corresponding to a certain resistance must be 
 expressed in terms of one size wire selected as the 
 standard. 
 
 The following method employs an actual bridge. Fig. 
 153 gives the connections. It will be noticed that the 
 position of the battery is such that the fault resistance 
 has no influence upon the accuracy of the test. A B and 
 B C are, as usual, the proportion arms of the bridge, but 
 are, together, used as a single arm, while R is the variable 
 arm. Calling x the distance from C to the fault, the 
 proportion under the condition of balance is 
 
 x = 
 
 where L is the length of the line, C x M, under test. 
 
466 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 A practice often resorted to in central station work is 
 that of burning out a cross or ground. It is effective in 
 most cases, but sometimes the fault has such great 
 current carrying capacity that the station cannot supply 
 enough current to burn it out. To burn out easily the 
 
 V r " ^_ 
 
 
 
 > QC 
 
 
 ^ FAULT 
 \ \ 
 
 J< / s 
 
 R / I 
 
 
 ft 
 
 \ 
 
 A 
 -\ 
 
 
 
 
 / 
 
 FIG. 153. 
 
 cross must offer some resistance, so that the heat gen- 
 erated by the passage of current will be sufficient to fuse 
 the contact. In order for a ground to give immediate 
 trouble on a metallic circuit, a second one must occur, so 
 that to burn out a single ground it is necessary to estab- 
 lish artificially a second one. When the detector indi- 
 cates a ground which requires immediate removal, the 
 detector ground is replaced by one consisting of a stout 
 cable including a heavy fuse. The full station over- 
 load, if necessary, is then run through the two grounds 
 until either the fault or station fuse, or fuse nearest the 
 fault, gives way. The best way to accomplish this is as 
 follows: Use one dynamo to separately excite all the 
 others, for if the fault is a short circuit, shunt machines 
 in circuit with it will not support a field. A light field is 
 put on all the dynamos, and their armatures must all be 
 
GROUNDS ON THE LINE. 467 
 
 in multiple. To render the station inactive it is only 
 necessary to pull the exciter's switch. Next, put in a fuse 
 that will carry 40 or 50 per cent, overload, and load the 
 station till this fuse, or something else, gives way. When 
 starting, be certain that all machines are of the same 
 polarity, close the exciter-switch, and observe the indi- 
 vidual ammeters to see that no machine is carrying more 
 than its share of the load. If there happen to be any, 
 their loads can be regulated by means of their rheostats. 
 Should the fault give way, the exciter switch must be 
 opened to avoid maintaining any arc which the burning 
 out of the fault might establish. Very often the fault 
 does not give way, but the fuse immediately in circuit 
 with it does, and patrons deprived of service hasten to 
 report the fact. 
 
 The advantage of separate excitation is that the load 
 can be worked on under good control, and need not exceed 
 the value necessary to remove the fault. 
 
CHAPTER XIV. 
 
 MOTOR TESTING. 
 
 ANY machine that can be used as a generator can be 
 used as a motor, not excepting even arc machines, whose 
 qualifications, however, are very unfavorable. 
 
 It is a matter of history that the discovery of the 
 motor preceded that of the generator, and that the prin- 
 ciple of reversibility was first applied to transforming a 
 motor into a generator. 
 
 Motors are classified in the same manner as are dy- 
 namos; according as they are series-wound, shunt-wound, 
 or compound-wound. They may be unipolar, bipolar, 
 or multipolar, according as the number of well-defined 
 poles is one, two, or more. The so-called unipolar 
 machine is really bipolar, one pole being more prominent 
 than the other. The number of poles dictates the num- 
 ber of neutral points, and hence the number of brushes, 
 except in special cases, where the armature is cross-con- 
 nected. Cross-connecting is used as a means of reducing 
 the number of brushes. On a four-pole machine there 
 are, ordinarily, four brushes (one at each neutral point), 
 alternating in polarity, so that opposite brushes are of 
 the same polarity, and can be connected; but adjacent 
 brushes, being of opposite polarity, must be well insu- 
 lated from each other. The result of connecting like 
 brushes is to place the four quarters of the armature 
 
 468 
 
MOTOR TESTING. 469 
 
 winding in multiple, thereby rendering the armature's 
 current capacity four times that of the component wires. 
 
 Now, since opposite brushes are at the same potential, 
 the respective armature wires under these brushes are at 
 the same potential, and, in fact, diametrically opposite 
 wires all around the armature are at. the same potential, 
 and may be profitably connected, thereby doing away 
 with one pair of brushes, providing the remaining pair 
 can carry the total current alone. This device, in sub- 
 stance, constitutes cross-connecting, and consists in 
 bringing down to the same commutator bar the leads 
 from all conductors which are at the same potential. 
 This reduces the number of bars required. Where it is 
 not convenient to cross-connect the armature, opposite 
 commutator bars are connected; this serves the same 
 ends, so far as the brushes are concerned, but does not 
 reduce the number of commutator bars required. As the 
 number of poles on a machine increases, complications 
 arise which lead to the using of the full number of 
 brushes, instead of cross-connecting. On motors, and 
 especially on street-railway motors, the practice of cross- 
 connecting is more generally adopted, on account of the 
 limitations of space and accessibility. On inclosed mo- 
 tors there are but two brushes, and these are easily 
 inspected from above. 
 
 A multipolar armature of the ordinary type, if not 
 cross-connected, must have as many brushes as there are 
 poles. If a smaller number be used the armature's cur- 
 rent capacity is reduced; it becomes electrically unbal- 
 anced and heats. Figs. 154 and 155 will serve to make 
 this more clear. 
 
 In Fig. 154, A is a motor armature running in a four- 
 
470 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 pole field, and has its opposite brushes connected. The 
 arrow-heads indicate the direction of the current from 
 brushes -j- b -\-b to brushes b b. The current enter- 
 ing by the positive brushes traverses one-fourth of the 
 
 FIG. 154. 
 
 armature and then leaves by the negative brushes. The 
 four quadrants are therefore in multiple. 
 
 Let us assume that the current passing under the posi- 
 tive pole, P lt produces rotation in a clockwise direction; 
 then that under the negative poles P 9 , P t , will concur in 
 this effort, for although the current is flowing in the 
 opposite direction, the fields also are reversed, and their 
 combined effect is the same. The current under P 9 
 concurs with that under P lt and thus all four quadrants 
 impel the armature in the same direction. Fig. 154 is 
 the usual connection for a four-pole motor. Fig. 155 
 
MOTOR TESTING. 471 
 
 gives the same machine with one pair of brushes re- 
 moved. In this case the several poles remain the same 
 in polarity as before, but the current under pole P % is 
 reversed, so that now one-quarter of the armature is in 
 multiple with three-quarters, thus resulting in an une- 
 
 FIG. 155. 
 
 qual distribution of current and in wasteful heating. 
 The arrow-heads of Fig. 155 show the direction of the 
 current under these circumstances. 
 
 A motor differs from a dynamo in that it has no criti- 
 cal speed of excitation, its fields depending either upon 
 separate excitation or upon the impressed E. M. F. , and 
 not upon any E. M. F. generated within the armature. 
 
 Series motors are peculiar in that they can have a field 
 only when current flows through the armature, and any 
 increase of armature current results in additional field 
 
472 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 strength and an increased C. E. M. F., thus providing a 
 factor of safety against prolonged short circuits. The 
 fact that armature and fields are in series removes the 
 necessity of a high-resistance starting box, and on some 
 series motors of early date the fields are wound in sec- 
 tions, which, when in series, contain resistance enough 
 to dispense with an extra starting coil. For obtaining 
 the several speeds, the sections are thrown into several 
 
 -f 
 
 FIG. 156. 
 
 combinations and are finally in multiple, the combination 
 of least resistance. 
 
 With shunt motors, the field, being in shunt with the 
 armature, is of high resistance, has great self-induction, 
 and hence takes longer to attain its full value. In 
 starting up a shunt motor a starting box is absolutely 
 necessary to avoid a short circuit, and this box must be 
 put in series with the armature, as shown in Fig. 156, 
 which gives the usual method of connecting in a starting 
 box. 
 
 A further precaution is to connect the fields orf a shunt 
 motor, or the shunt fields of a compound-wound motor 
 above the switch, so that the motor may be given a field 
 before the armature circuit is closed. When the motor 
 is idle, both the box and the switch K, Fig. 157, are 
 
MOTOR TESTING. 
 
 473 
 
 FIG. 157. 
 
 open. To start the motor its field switch (not shown in 
 figure) is first closed. Then the starting box is slowly 
 advanced to its final position. A" closed, and the starting- 
 box circuit once more opened, 
 so that simply by opening K 
 the motor can be shut down. 
 The proper selection of a 
 starting box depends upon the 
 object in view, (i) It maybe 
 desired simply that the initial 
 current flow shall not exceed 
 the motor's capacity; or, (2) 
 that the start shall be so 
 gradual as to avoid jarring the 
 mechanism operated by the 
 motor. For instance, on a 
 car, the first flow of current 
 
 may be considerably below the motor's rated capacity, 
 and yet seriously incommode passengers. To design 
 a resistance to meet the demands of such variable 
 conditions as are encountered in street-railway prac- 
 tice, the engineer must effect a compromise between 
 theory and practice. Theory would have a car start 
 smoothly under all conditions imposed by variations 
 in car weight and grade climbing, and \vould provide a 
 ready means to this eirl Practice, at least as voiced in 
 the demands of most street-railway men, would have a 
 car start on the first notch of the controller under all 
 conditions. This means that a resistance designed to 
 start smoothly a loaded car on a grade will violently jerk 
 an empty car on a level. In fact, starting coils are 
 designed to-day to give an initial current flow equal to 
 
474 TESTING OF DYNAMOS AND MOTORS. 
 
 the maximum output of one motor, or the normal output 
 of both. 
 
 If the carrying capacity alone of the motor is to be 
 considered, the problem resolves itself to a simple applica- 
 tion of Ohm's law, but a consideration of smoothness in 
 starting requires experimental data. Knowing the cur- 
 rent capacity of the motor to be started, the resistance 
 that will just permit this current to flow can be gotten 
 from the equation, 
 
 where E is the impressed voltage, / the allowable cur- 
 rent, and R the resistance value sought. The armature 
 resistance is usually so low as to be negligible, and need 
 not enter the equation. An important consideration is 
 to be sure that the wire making up the resistance has 
 enough cross-section to carry the current called for. 
 The coils remaining longest in circuit are generally made 
 of heavier wire than those that are cut out first. Table 
 IV. of the Appendix gives the current capacity of tinned 
 iron wire under different conditions. 
 
 It must be borne in mind that a box designed for one 
 particular voltage cannot be used safely on lines of much 
 higher voltage, nor will it give the same nicety of control 
 when used on a line of lower voltage; because, in the first 
 case, the excess of voltage is liable to heat the box or 
 overload the motor, and, in the second case, the resist- 
 ance is likely to be too high to admit of a ready 
 start. A box intended for a motor of large output can- 
 not be used with impunity on a small motor, because 
 
MOTOR TESTING. 475 
 
 the starting current of the larger motor would be 
 too much for the smaller one; on the other hand, a 
 box designed for a small motor would not admit 
 sufficient current to start a large one without advanc- 
 ing the handle to a position dangerous to hold for any 
 length of time. 
 
 Starting boxes are preferably placed in multiple with 
 the switch, so that when the resistance is all out the 
 box maybe short circuited by closing the switch, and 
 the box bar returned to the "off" position. Should 
 the box be left in, both it and the switch must be opened 
 when it is desired to stop the motor. 
 
 On shunt machines, the box is used only in start- 
 ing, but on series machines, which race when the load 
 is removed, it is used to control the speed. Occasions 
 sometimes arise for using a street car motor to run the 
 repair shop shafting. In such a case the instability of 
 speed of a series motor is a serious drawback and it can 
 only be overcome by replacing the series coils by an 
 equivalent shunt winding. Suppose that we have such 
 a. motor and that we wish to change its winding. Let 
 the series coils, two in number, consist of 100 
 turns of No. 4 B. & S. wire. Suppose each spool 
 measures .175 ohm. The total resistance then is .35 
 ohm, corresponding to a length of 1,400 feet. If full 
 load current be 50 amperes, full load drop (I J?) is .35 X 
 50 = 17.5 volts, and the watts lost in heat (e 7) is 17.5 X 
 50 = 875 watts. With the fields hot this loss would 
 increase to 1,000 watts. If the line voltage be 500, the 
 total intake of the motor is 500 X 50 25,000 watts, 
 and the field loss hot = 1,000 -^ 25,000 = 4 <f>. Fifty 
 amperes passing around 200 turns of wire give 50 x 200 
 
476 TESTING OF DYNAMOS AND MOTORS. 
 
 = 10,000 ampere-turns as the field coils' magnetizing 
 force at full load. 
 
 The shunt winding must provide this magnetizing 
 force and must not waste more than 4 % of the full 
 load output in doing so. To restrict the loss to 
 1,000 watts, the field current at 500 volts must not 
 exceed 1,000 -^ 500 = 2 amperes. A current of 2 
 amperes at 500 volts calls for a resistance of 500 
 -i- 2 = 250 ohms, or 125 ohms per spool. The re- 
 quired number of ampere-turns is 10,000, and since the 
 amperes are 2 the number of turns is 10,000 -=- 2 = 5,000 
 or 2,500 per spool. Since the ampere-turns are the 
 same, the weight of copper in the two cases will be 
 about the same and will approximate 180 pounds, or 90 
 pounds per spool. Now, to figure the winding space: 
 the sectional area of a No. 4 wire is .0328 square inch, 
 and that of 200 turns is 200 x .0328 = 6.56 square 
 inches. Consulting a wire table we see that 5,000 turns 
 of No. 18 wire gives a cross-section of (5,000 x .0013) 
 6.5 square inches, and that a length sufficient to measure 
 250 ohms (39,000 feet) weighs 195 pounds. This winding 
 then about fulfills the conditions. As a matter of fact, a 
 shunt winding requires more space than a series winding 
 of the same power, because not only is the radiation 
 poorer but a larger per cent, of the winding space is 
 occupied by insulation. 
 
 The series motor's instability under variable load can 
 be explained as follows: we have learned that on self- 
 exciting dynamos a certain speed must be attained be- 
 fore the armature will build -up a field, and it follows 
 that its voltage will be proportional not to the actual 
 speed, but to the actual speed less the critical speed. 
 
MOTOR TESTING. 477 
 
 Thus the voltage at two different speeds have the follow- 
 
 ing relation to each other: 
 
 where N o is the critical speed, and N^ N^ the speeds 
 corresponding to voltages F,, K a respectively. This 
 proportion is true for a considerable range of speed 
 variation, but is departed from as the load increases 
 because of the influence of armature reaction, and it is 
 only by maintaining the armature current constant that 
 this disturbing effect can be eliminated. On separately 
 excited dynamos, which have no critical speed, the 
 voltage is strictly proportional to the speed when run- 
 ning free, and when under load, any change in speed 
 changes the voltage proportionately provided the arma- 
 ture current is kept constant. If the machine is run as 
 a motor, what was true for the dynamo's voltage is true 
 for the motor's speed if its field is kept constant. The 
 speed will be proportional to the voltage and theoretic- 
 ally become zero when the voltage does so, but prac- 
 tically, a little before this, on account of internal 
 resistance, friction, etc. The speed is, then, propor- 
 tional to the impressed voltage less that voltage below 
 which the armature stops turning or below which the 
 armature refuses to turn. Here 
 
 JV : :JV, :: V t - V,:V,- V,, 
 
 where F is the critical voltage. A separately excited 
 motor does not race unless the voltage is abnormally 
 high or the field very weak. 
 
TESTING OF DYNAMOS AND MOTORS. 
 
 The speed which any motor attains is such that the 
 following equation always holds true: 
 
 C. E. M. F. + / ^ ft = Impressed E. M. F. 
 
 Here I /? a is the armature drop. A change in any of 
 these terms will alter the speed; thus an increased field 
 raises the C. E. M. F., and the speed falls until the 
 above balance is restored. The effect of putting on 
 a load is to decrease the speed, lower the C. E. M. F., 
 and let in a larger current and increase / R^ until the 
 balance again obtains. Upon removing a load the 
 retarding force is removed, and the speed leaps upward, 
 -striking a final value dependent upon the field strength. 
 This is why a separately excited motor does not race 
 unless the voltage is very high or the field very weak. 
 A shunt motor's speed will increase about 20 <f, upon 
 the removal of load, but not more because its field is 
 independent of the armature. On the other hand a 
 series motor races badly because the field decreases 
 as the armature current decreases, and the higher the 
 speed the weaker does the field become, until, if un- 
 checked, the armature may wreck itself. 
 
 This peculiarity may be profitably followed out more 
 closely. Let us suppose that the impressed E. M. F. is 
 from a compound-wound generator and remains con- 
 stant, and that the motor carrying a certain load has a 
 terminal E. M. F. of E volts, distributed between its 
 C. E. M. F. and its internal resistance R : or, E e -+- 
 I R. Now let part of the load be removed; a less 
 amount of energy is now required to carry the decreased 
 load. The amount taken by the motor is measured by 
 
MOTOR TESTING. 479 
 
 the product E I, and since E is constant, / must de- 
 crease. As /decreases, the drop through armature and 
 field decreases; now since I R has decreased, the C. E. 
 M. F. must increase, for E c -\- f R is a constant. 
 TheC. E. M. F. can be increased in two ways: by strength- 
 ening the field, by raising the speed: but in this case the 
 field is weakened as /grows less, and hence the speed in- 
 creases at a rapid rate to compensate for the weakened 
 field and the decreased drop. As the load is further 
 lessened, the speed soon passes the safety point and the 
 belt flies off. The armature unhindered speeds on, its 
 current becoming a minimum, and its C. E. M. F. 
 approaching the impressed E. M. F. as a limit. The 
 less the machine's internal resistance, the greater its 
 C. E. M. F. and the smaller the internal drop. 
 
 It may be at first puzzling to find a way to check the 
 series machine's tendency to race: In the equation E 
 c A- I R we see that if / R can be made large, Es value 
 may be correspondingly small, also if R be made large, the 
 drop / R will be large, and without any abnormal value 
 of /. To secure this the starting box is thrown into 
 series connection with the motor, and is gradually cut 
 out as the device in operation gains headway. It is thus 
 that the speed is held within proper limits. 
 
 In a shunt motor the field and the armature circuits 
 are in multiple, so that if the applied voltage is main- 
 tained constant the current which each gets depends 
 solely upon its effective resistance. The field circuit 
 containing only ohmic resistance its current remains 
 constant, excepting in so far as this resistance may be 
 increased by heat. The armature when at rest also 
 contains only ohmic resistance, and this being very low 
 
480 TESTING OF DYNAMOS AND MOTORS. 
 
 would admit an abnormal current were no steps taken to 
 prevent it. The starting box is therefore used to limit 
 the current value at starting. As the armature begins to 
 turn it generates a C. E. M. F., which may replace the 
 ohmic resistance of the box; as the speed rises the 
 C. E. M. F. does also, and the box is gradually cut out, 
 until the C. E. M. F. is finally adequate to control the 
 armature current. 
 
 When full load is on a shunt motor and the armature 
 current reaches its maximum value, the brushes have 
 their greatest backward lead, and it is here that the 
 armature reaction is greatest, and weakens the field 
 most. If at this point the load be suddenly removed, 
 the first effect is to strengthen the field by withdrawing 
 the armature's reaction, and secondly to lessen the drop 
 by decreasing /. The first effect tends to lower the 
 speed and the latter to raise it, with the result that the 
 speed rises about 20 $. 
 
 The difference between shunt and series motors in this 
 respect is, that on series motors the increase of 
 C. E. M. F. due to increase of speed which follows the 
 removal of load weakens the field, while on shunt 
 motors the reverse is true. Another difference is in the 
 effect of heating. On a series machine at constant 
 voltage and a given current the speed is higher at the 
 beginning than at the end of a heat test, because the 
 increased internal resistance raises the drop, thereby 
 lowering the effective impressed voltage, and as the speed 
 depends upon this, it decreases with it. On a shunt 
 machine, the initial speed is lower than the final, because 
 as the shunt field heats its increased resistance admits less 
 current, thereby weakening the field and raising the speed. 
 
MOTOR TESTING. 481 
 
 The fact that the field and armature of a series motor 
 are in series adapts it for use where jarring is excessive 
 or where the external circuit includes a sliding or rolling 
 contact. If the circuit is momentarily interrupted, field 
 and armature are alike deprived of current, and when the 
 circuit is closed the same current flows through both, so 
 that a prolonged short circuit is impossible unless the 
 rotation of the armature is prevented. If the circuit of 
 a shunt motor is broken for a moment, the motor loses 
 its field, and when the circuit is again Closed the arma- 
 ture short circuits the field and admits an abnormal cur- 
 rent, which not only gives a violent wrench to the belt or 
 gearing, but also endangers the life of the armature. 
 The result of these facts is that, as a rule, portable 
 motors such as are used on street railroads and the like 
 are series motors; while stationary motors are generally 
 shunt-wound. 
 
 T EST XVIII. Series Motor- Generator Test. O n ace o u n t 
 of complications which arise due to the fact that series 
 machines support a field only on closed circuit, and that 
 generators and motors run in opposite directions forgiven 
 connections, special precautions must be taken in running 
 a series motor-generator test. The loss is best supplied 
 by an engine or by a motor whose speed can be easily 
 controlled. The following scheme is adapted to testing 
 street railway motors where made in quantities, and 
 when it is necessary to use as little power as possible 
 from the engine. Although not in general use the plan 
 has commendable features. A and B, Fig. 158, are two 
 motors geared to axle, 6", which is belted to engine, E, 
 and turns always in the same direction. A is connected 
 to generate for this direction of rotation, and B is con- 
 
482 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 nected so that it will not generate. It will therefore run 
 as a motor, and in the same direction as the axle. To 
 run the test with both motors connected alike, it is nec- 
 essary either to have an intermediate gear on one motor, 
 or to have the motors run on opposite sides of the same 
 gear, so that the armatures will turn in opposite direc- 
 tions. Usually the best method is to reverse the con- 
 
 =CH 
 
 
 
 
 -m- -| 
 
 
 B 
 
 =D> 
 
 A 
 
 
 
 
 
 
 FIG. 158. 
 
 nections and leave the direction of rotation the same. 
 A and B have their fields and armatures in series as 
 shown in the diagram and include in circuit a switch K 
 and a variable resistance, R, capable of carrying the 
 machine's full current. R exceeds the critical resistance 
 of the machine for the given speed, so that upon closing 
 K the dynamo will not generate until part of R is cut 
 out. Before starting a test for the first time it is well to 
 preliminarily determine the correct connections for the 
 dynamo to generate. This is done by connecting the 
 machine as shown in Fig. 159 and bringing it up to 
 speed. In Fig. 159 R is the starting box, a l # 3 the 
 
MOTOR TESTING. 
 
 483 
 
 f 
 
 A 
 
 f. 
 
 |R 
 
 ,K 
 
 
 _1. \ 
 
 oJ 
 
 armature terminals, and /, / a the field terminals. If now 
 the machine refuses to generate upon closing K and 
 slowly working R out, the armature leads should be 
 interchanged. 
 
 There is also a question of speed to be decided 
 before starting. Running at its rated current and 
 speed, a machine as a generator 
 will not produce an E. M. F. as 
 great as the impressed would 
 have to be in order to run the 
 machine as a motor under simi- 
 lar conditions, and hence the 
 system is not subjected to as 
 severe a test as it ought to be. 
 To compensate for this, the 
 gearing or belting should be 
 arranged to run A above its 
 
 rated speed and thus raise the E. M. F. to its rated 
 value. The necessary increase in speed can be approxi- 
 mately calculated if A's internal resistance is known. 
 Let the machine's internal resistance be r; then i r is the 
 internal drop when current, /", flows; let the E. M. F. 
 desired to be impressed at the motor terminals be E. 
 Then will the C. E. M. F. be E i r = e. Let n be the 
 speed of A running as a motor with an impressed 
 E. M. F. of E volts and a current of /' amperes. Since 
 as a motor the machine generates a C. E. M. F. of c 
 volts, and as a dynamo it must generate an E. M. F. of 
 E volts, we may write, e : E \\ n : n', where ' is the 
 required speed. This gives us 
 
 n E n E 
 
 FIG. 159. 
 
 n = 
 
 E-ir' 
 
484 TESTING OF DYNAMOS AND MOTORS. 
 
 where all the terms are known and hence ;/' deter- 
 mined. 
 
 On account of the series machine's ability to rapidly 
 pick up as soon as it begins to generate, it is well to pro- 
 vide belt guards to avoid the annoyance of losing the 
 belt under sudden overloads. A further precaution is to 
 insert a light fuse at the start, and then cut it out when 
 the test is under way. If the motor shaft is arranged to 
 be thrown in by a clutch the start is much smoother. 
 In starting up, the machine is brought up to speed, K 
 closed, and R slowly worked out, at the same time weak- 
 ening B's field by means of the shunt r shown in Fig. 
 158. As soon as the ammeter shows A to be generating 
 R must be very carefully handled to avoid precipitating 
 a heavy overload and throwing the belt. A will refuse 
 to generate until a certain amount of R has been cut out, 
 and will than pick up very rapidly. It is absolutely 
 necessary that means be provided for weakening the 
 motor field, otherwise since the same current must pass 
 through both machines, and since they run at the same 
 speed, the C. E. M. F. of B will be the same as the 
 E. M. F. of A, and a load cannot be worked on. The 
 shunt affords the same regulation as obtains on a car, 
 but has a different relation, in that on a car its value is 
 constant and the speed variable, while in this test the 
 shunt is variable and the speed constant. 
 
 A's current passing through B runs it as a motor, and 
 helps to turn the system, thus lessening the demand on 
 the supplier, which then supplies only energy enough to 
 cover the losses, which may amount to from 25$ to 35$ 
 of the motor's output. After running A for a stated 
 time as a dynamo it is changed over and run as a motor. 
 
MOTOR TESTING. 
 
 485 
 
 This change is most rapidly effected by using a crossed 
 belt to reverse the direction of rotation; it is then only 
 necessary to move the shunt from B to A. 
 
 TEST XIX. Testing Series Machines with Water Rheostat. 
 In Fig. 160 are shown connections for a second 
 method of testing series machines. It is a modified form 
 of motor-generator 
 test, but differs from 
 it in that the dyna- 
 mo's energy is not 
 returned to the mo- 
 tor. The amount 
 of energy supplied 
 must be something 
 over the rated out- 
 put of one machine. 
 G is a generator of 
 the same voltage as the machines under test, D is the 
 dynamo armature, M, that of the motor; F and F' the 
 dynamo and motor fields respectively. It will be seen 
 that F, F\ and J/, are in series, and that F is, therefore, 
 excited from G. W is a water rheostat to which Z>'s 
 terminals are connected: A' is a switch across which is 
 the starting box R. M and D are geared or bolted to 
 the same shaft or to each other and hence start up to- 
 gether. When R is all cut out it is short circuited by 
 means of K-, if the speed is then too high and the load is 
 low, the plates of the water rheostat must be brought 
 nearer together, thereby decreasing its resistance, in- 
 creasing the load, and lowering the speed. If the load is 
 apparently all right and the speed still high, the am- 
 meter or voltmeter may be "off." Only frequent cali- 
 
 FIG 160. 
 
486 TESTING OF DYNAMOS AND MOTOKS. 
 
 bration can eliminate liability to this error. If the load 
 is certainly all right and the speed high, the indications 
 are of a weak field, which may be due to short circuited 
 turns in a field coil; to reversal of a field spool, on any 
 motor having more than two coils; to a loose joint in the 
 magnetic circuit; to an abnormally wide air gap; to an 
 inferior quality of iron in the armature core or frame; 
 or to displacement of brushes. Whether there are any 
 short circuited field turns or not can be decided by pass- 
 ing current through all and taking the drop on each. 
 If one coil is reversed on a motor having but two coils, 
 it will not start at all. If one of four is reversed it is 
 readily located by trying to start the motor using but two 
 coils at a time. When the pair in use are of proper 
 polarity the motor will start, though more current will 
 be required than ordinarily to do so. When of the 
 wrong polarity no safe current will start the motor. We 
 emphasize safe because it is a fact that sufficient current 
 through the armature would enable it to react upon the 
 pole pieces, and produce field enough to start. Once 
 started, the armature would race. The loose joint and 
 air gap difficulty can best be located by inspection, while 
 the question of inferiority of iron can be considered 
 after all other probable difficulties have been eliminated. 
 If there are no facilities for magnetic testing, we must 
 resort to the somewhat crude method of interchanging 
 the parts of the faulty motor with those of one that we 
 know to be all right. 
 
 A modification of the above test consists in running D 
 as a self-exciting series dynamo, with the advantage 
 that a greater E. M. F. is applied to M, because there 
 is no impressed voltage lost in the resistance of D's field, 
 
MOTOR TESTING. 487 
 
 and this may be considerable. In this test the connec- 
 tions depend upon the direction of rotation. If with 
 given connections D refuses to pick up, the direction of 
 rotation must be reversed. Of course with separate ex- 
 citation the direction of rotation need not be considered. 
 When the dynamo does "pick up" it does so with a 
 rush and the speed drops rapidly. This effect is more 
 marked in warm weather than in cold, for the water box 
 resistance is lower. On the other hand the machine may 
 refuse to generate for want of a sufficiently strong re- 
 sidual field. In this case the practice is to short circuit 
 the leads with a piece of fuse wire in multiple with the 
 water box. As soon as the machine generates it blows 
 the fuse, and then continues to send current through the 
 water box, in virtue of the strong field provided by the 
 short circuit. The above test method is much used in 
 testing street railway motors, and for this purpose a box 
 5 feet x i l /z feet -j- i^ feet is sufficient to carry the 
 load. 
 
 TEST XX. Efficiency Test. A modification of Test 
 XVIII. is used for determining the efficiency of series 
 motors. In Fig. 158 shaft 5", instead of being driven 
 by the engine , is driven by a motor which we will 
 call M. The test is made in the following steps: 
 i. The energy necessary to turn M free at the proper 
 speed is taken. 2. M is then belted to S and all pin- 
 ions on A drawn. The energy required by M to run 
 A's armature at proper speed is then measured. 3. A's 
 pinion is replaced and 's removed and readings taken 
 on M. 4. B is geared to A in the usual way, and the 
 energy again measured. These successive measure- 
 ments give the data for finding the frictional losses on 
 
4 88 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 all moving parts. A is then connected as a generator 
 and runs B as a motor. Readings are taken as in former 
 efficiency tests and from these readings the efficiency is 
 deduced. 
 
 TEST XXI. Efficiency Test with Prony Brake. An- 
 other good way of finding efficiency is by means of the 
 
 DETAILS OF E 
 
 FIG. 161. 
 
 Prony brake, a device much used in testing rooms. One 
 form used is shown in Fig. 161, where M is an iron 
 pulley keyed to the armature shaft. About the outside 
 of this pulley is placed a steel strap, one of whose ends 
 fastens to the lower arm of clamp E, the other end to 
 the upper arm of E and ending at Z>, where the weight 
 hanger W is attached. The grip of the strap can be 
 varied at will by adjustment of cord /. Falso is vari- 
 
MOTOR TESTING. 489 
 
 able. The inner rim of the pulley is flanged and will 
 hold a considerable amount of water, which, poured in 
 after the wheel is in motion, is kept in by the centrifugal 
 force, and serves to cool the rims. When the wheel 
 stops a large portion of this water is ejected, so that the 
 operator must be on the lookout. The action of the brake 
 is as follows: The motor armature turns left to right 
 and tends to carry the mass IV around with it as soon 
 as the steel strap is tightened. On the other hand, }V 
 resists by its weight the tendency to be carried around. 
 To measure the load on the motor, the weights on Wand 
 the tension on E are so adjusted that the weight is bal- 
 anced exactly at the horizontal line; the weight pulls 
 on the wheel with the leverage OT. The work done 
 when a balance is effected is measured in mechanical units 
 by taking the product of the circumference of the wheel, 
 the number of revolutions per second, and the weight W. 
 If d be the diameter of the wheel the circumference is 
 3.1415 d or 7t d. The power or work per second then is 
 
 Power = 7t d n W, 
 
 where n is the revolutions per second, and W the weight 
 on the hanger. The intake is measured electrically, 
 and is the product E x /, and is measured in watts. 
 
 If d\s given in centimetres and Win grams, then the 
 output is measured in ergs per second, and the watt is 
 equal to 10,000,000 ergs per second: therefore we have 
 
 Output in watts = - - = W , 
 
 10,000,000 
 
 also 
 
 Intake in watts = E x I W^\ 
 
490 TESTING OF DYNAMOS AND MOTORS. 
 
 we then have 
 
 n d nw 
 
 _- . 
 
 Efficiency = -=~ = 
 
 10,000,000 E X / 
 If d is given in feet and W in pounds we then have 
 
 Output in watts = 1.36 it d W, 
 
 and 
 
 1.36 n d W 
 
 Efficiency = 
 
 E X 
 
 The advantages of this method are that the apparatus is 
 compact and simple, and the readings easily taken. The 
 drawback is that the energy is entirely lost, being con- 
 sumed in friction on the pulley and strap. If the test is 
 long continued the rise of temperature of the pulley will 
 cause the water in the rim to rise to the boiling point. 
 
 In testing series machines rigidly connected it is a 
 wise precaution to have the scheme of wiring embodied 
 in a drawing, which can be followed in connecting up 
 each set of motors. This drawing should include, not 
 only the motors themselves, but the controller and instru- 
 ments as well. Then if a controller should have a 
 wrong connection, or if the polarity of armature or 
 field is reversed, through some error in winding, or in 
 bringing out the leads, this fact is likely to be discovered 
 by a reversal in the direction of rotation, or by a refusal 
 to run at all. 
 
 Having discovered that an error of this kind exists 
 somewhere in the set under test, the next step is to 
 locate it. Since there are two motors concerned it is 
 highly likely that one of them is correct, and a compari- 
 
MOTOR TESTING. 49! 
 
 son of the wiring, when the machines are made to run, 
 with the diagram, will identify the faulty machine. 
 
 If no knowledge exists as to the relation of direction 
 of rotation and connections, so that it is not known when 
 the machine is running correctly, the best procedure is 
 to connect up and run a motor which has been tested, 
 and is known to be correct. In this case the motor in 
 the rack which runs like it is all right, and the other one 
 probably at fault. To make perfectly sure, however, the 
 test motor must be run on both sets of wires, then if its 
 direction of rotation is the same on both sets of wires, it 
 locates the latter set, with its motor, as the faulty one. 
 If the rotation of the test motor is reversed in passing 
 from one set to the other, it shows that the motors are 
 right and the trouble is in the controller. 
 
 Having now identified the faulty machine we must de- 
 termine whether the trouble is in the armature or fields. 
 The best way to do this is to place an armature which is 
 known to be correct in the frame, and to run it; if the 
 two run in the same direction it shows that they are alike 
 and that the fields are at fault. The methods of testing 
 out faults in armature and field have already been given 
 in earlier chapters, and will not here be repeated. 
 
 In such tests it simplifies matters if, instead of a con- 
 troller, with its complicated wiring, an ordinary rheostat 
 is used as a starting box. The motors can then be con- 
 nected as shown in Fig. 162, where //' are the terminals 
 of the starting box. In connecting the two motors cor- 
 responding field and armature terminals go together, 
 leaving the same leads on the two motors to be connected 
 to the box. It then makes no difference which lead is 
 connected to / or /' of the box, for, since both field and 
 
492 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 jwwmi 
 
 armature are included between the leads, the direction of 
 rotation will be the same for either connection. 
 
 The compound-wound motor combines the general 
 characteristics of both the series and of the shunt motor. 
 A dynamo compounded 
 for constant potential, if 
 differentially connected, 
 will as a motor automati- 
 cally regulate for con- 
 stant speed. By con- 
 necting differentially is 
 meant that the series 
 field opposes the shunt 
 field in its magnetizing 
 effect. Supposing a com- 
 pound-wound motor to 
 be running at its proper 
 speed, and let us suppose 
 
 the load to be increasing, the tendency of the added load is 
 to cause the speed to drop, and it will momentarily do so. 
 With this fall of speed comes a decrease in the C. E. M. 
 F., and the current in the armature rises, and with it the 
 strength of the series field; but the series field opposes 
 the shunt field and hence the field of the motor is weak- 
 ened, and the speed increases. 
 
 In connecting a compound-wound motor differentially 
 the connections can be tested as follows: With the given 
 connections the armature current is read, and the speed 
 noted. The connections of the series field are then re- 
 versed, and the armature current adjusted to the same 
 value as before and the speed taken. The differential 
 connection will give the higher speed, as it gives the. 
 
MOTOR TESTING. 493 
 
 weaker field. The test can also be made by short cir- 
 cuiting the series field so that the motor is first running 
 on the shunt field alone, and reading the current and 
 speed as above. The series coil is then cut in, the cur- 
 rent adjusted, and the speed read. In this case the ter- 
 minal E. M. F. will have to be raised and the latter speed 
 will be found to be lower than the first, if the connections 
 are differential, but the fall of speed will not be so great 
 as when cumulatively connected. If instead of keeping 
 the current constant the E. M. F. is constant, the differ- 
 ential connection will give a higher speed than with the 
 shunt field alone, while the cumulative connection will 
 give a lower speed. The statement above that the differ- 
 ential connection gives a lower speed than the shunt field 
 alone may at first seem contradictory. If we remember, 
 however, that when the E. M. F. is raised to compensate 
 for the / R drop in the series coil, the current in the 
 shunt field is thereby increased, and that the shunt field 
 is relatively stronger than the series field, we see that in 
 this case the motor field is strengthened, and hence the 
 speed falls. With the cumulative connection the field is 
 strengthened to a greater extent, and the fall in speed is 
 greater. 
 
 In testing rooms it is common practice to run com- 
 pound-wound motors with the fields cumulatively con- 
 nected, and in motor-generator tests this connection is 
 necessary unless there is a heavy shunt on the motor 
 series field to weaken its effect. In the motor-generator 
 tests already cited it was found impracticable to get the 
 load on slowly when the fields were connected in opposi- 
 tion, and it would be only theoretically possible with the 
 shunt field rheostats so arranged as to gradually intro- 
 
494 TESTING OF DYNAMOS AND MOTORS. 
 
 duce resistance into the shunt field circuit. Under these 
 conditions the system would be very unstable, and the 
 least variation in speed precipitates an overload or even 
 a short circuit. 
 
 To render the control in putting on a load more 
 complete, it is sometimes necessary to use a shunt board 
 on the series field of the cumulatively connected motor: 
 this is the case where the movement of the rocker arm 
 is too limited to admit of working on the load by shifting 
 the brushes. The shunt board is also needed when the 
 shunt field has been so weakened that the machine is 
 practically running as a series motor. This condition is 
 generally indicated by a steady uncontrollable sparking 
 at the brushes, which is only relieved by strengthening 
 the shunt field. The reason for this lies in the fact that 
 if the motor is running as a series motor the neutral 
 point is not the same as under normal conditions, and 
 may have shifted out of the range of the movement of 
 the rocker arm. A trial is the best test of the necessity 
 of a shunt board. 
 
 We see then that there are two conditions under which 
 a shunt board can be profitably used. i. With fields 
 cumulatively connected, where the shunt board weakens 
 the assisting power of the series coils and does away with 
 sparking. 2. Where the fields are differentially con- 
 nected and the opposing power of the series field is to be 
 weakened. 
 
 The series motor property of the compound-wound 
 motor makes it safer to break its shunt field under full 
 load than under light load, provided the fields are 
 cumulatively connected. In this case at full load the 
 current in the series winding is large and the field is 
 
MOTOR TESTING. 495 
 
 mainly due to it, so that the withdrawal of the shunt 
 field will have no further effect than to increase the 
 sparking, and perhaps raise the speed somewhat. At 
 light load the field may be considered as entirely due to 
 the shunt coils, and the series field is very small, so that 
 to break the former would precipitate a short circuit. 
 If the armature does not burn out the motor will start to 
 racing and throw its belt. If while the armature is at a 
 high speed the shunt field be again closed, the motor will 
 generate a higher C. E. M. F. than the impressed E. M. F. 
 and will momentarily reverse the dynamo that is running 
 it. This action lasts only so long as the momentum of 
 the motor enables it to hold its speed, and the system 
 soon returns to its normal state. The above reactions 
 are especially pronounced where the machines are heavy 
 and strongly overcompounded. 
 
 If both machines are belted to the same countershaft 
 the above reversals cannot take place until the motor 
 belt is thrown, as neither can run faster than the engine 
 does. If the machines are simply belted together and 
 the loss supplied from a dynamo no restriction is present, 
 and the cycle of reactions is similar to several that have 
 already been described. 
 
 Compound-wound motors of large output are not in 
 general use because for most machine work a shunt motor 
 running on constant potential mains regulates sufficiently 
 well. 
 
 In belting a motor to a shafting attention must be paid 
 to two points, i. The higher the speed allowed on the 
 motor the higher will be the efficiency. 2. The tools 
 operated by the shafting must be run at a speed sufficient 
 to secure a high shop efficiency. The designed speed of 
 
496 TESTING OF DYNAMOS AND MOTORS. 
 
 most types of commercial motors is stamped on the name 
 plate, and is the speed at which the motor should run at 
 full load. 
 
 It sometimes happens that it is desired to add tools 
 to a line of shafting that already fully loads the 
 motor running it. It is possible to do this without over- 
 loading the motor, by putting a larger pulley on the 
 countershaft or a smaller one on the armature, thereby 
 decreasing the speed of the countershaft, the motor 
 armature keeping the same speed as before. Since now 
 the countershaft is running at a slower speed, each 
 machine turns slower and does less work in a given time, 
 the new machines making up the difference. It is thus 
 seen that the total work turned out is the same as before, 
 since the load on the motor is the same whether it carries 
 a few machines heavily loaded or more machines less 
 heavily loaded. 
 
 At times in testing rooms it is desirable to throw a 
 load on to a line of shafting greater than the engine 
 driving the shafting is able to carry. In this case a 
 motor may be belted to the shafting, and run from a 
 dynamo attached to another engine. In making the 
 motor connections one must be careful not to make the 
 motor oppose the engine as regards its direction of 
 rotation. On self-excited machines, one familiar with the 
 rules for connections can predict the direction of rota- 
 tion, and act accordingly; but if the motor is separately 
 excited no rule is available and recourse must be had to a 
 preliminary test. To insure that a motor attached to an 
 engine shall concur in effort with the engine, the motor 
 fields must be so excited that the E. M. F. of the arma- 
 ture is opposed to that of the dynamo that is to run it. 
 
MOTOR TESTING. 497 
 
 It then comes to this, that since the E. M. F. generated 
 by the motor armature is to become its C. E. M. F. it 
 must oppose the impressed E. M. F. Having secured 
 this, and adjusted the fields so that the voltage read 
 across the switch is low, the switch is closed and the field 
 weakened. The motor then takes load and relieves the 
 engine. 
 
 In belting up the motor for such work, care must be 
 taken to select pulleys of such a size that countershaft 
 and motor are both running at their proper speed. Sup- 
 posing the motor to make 1,000 revolutions and the shaft 
 300, if the diameter of the pulley is 4 feet, that of the 
 armature pulley can be found as follows: The circum- 
 ference of the shaft pulley is TT Z> or 3. 1416 x 4 which 
 equals 12.56 feet, and in one minute the belt travels 
 12.56 x 300 = 3,768 feet. Now the armature pulley must 
 travel at the same rate, and if d be its diameter we have 
 d X 3.1414 X 1,000 = 3,768, which gives us 
 
 d - - 1.2 feet. 
 
 The general formula is 
 
 D X TT X // = d X n X ', 
 or 
 
 d n 
 
 where </and D are the diameters and ;/ and n' the revo- 
 lutions per minute of the shaft and armature. In the 
 above case this gives us 
 
 d__ 300 
 4 1,000 
 
498 TESTING OF DYNAMOS AND MOTORS. 
 
 or 
 
 rf=I^ = I . z feet. 
 
 I,OOO 
 
 This method of relieving an overloaded engine is fre- 
 quently resorted to, and sometimes at the end of a test 
 the engineer is surprised to find that he is unable to stop 
 his engine by shutting off the steam. In shutting down, 
 the motor switch should be opened first. It may even 
 happen that the motor is running the engine; this is apt 
 to happen when two dynamos running from different 
 engines are placed in multiple, and by careless handling 
 of a rheostat one of them is changed over into a motor. 
 
 If a machine, as a dynamo, is designed to give a cer- 
 tain voltage, it will not as a motor give the same speed 
 when run at that voltage, but will give a speed from 10 $ 
 to 20 <f> lower. The reason for this is as follows: In the 
 case of the dynamo the losses due to friction, belt ten- 
 sion, armature, and field resistance, etc., are looked after 
 by the engine and the voltage is maintained over and 
 above these losses. In other words, the total E. M. F. 
 of the dynamo is greater than its terminal E. M. F. In 
 the case of the motor the losses must be supplied by the 
 energy given to the motor electrically, and consequently 
 the voltage available for producing rotation is not equal 
 to the impressed E. M. F., but is lower than this by an 
 amount equal to the drop through the motor. Hence 
 the speed of the motor is lower than that of the dynamo. 
 
 The case of the motor is not then the exact counter- 
 part of that of the dynamo. Thus, if the output of 
 the dynamo be 25 amperes at 100 volts, the output is 
 2,500 watts. If the losses cared for by the engine are 
 
MOTOR TESTING. 499 
 
 equal to 500 watts, then the energy involved in the 
 dynamo is 2,500 -(- 500 or 3,000 watts. On the other 
 hand, if the motor be running with an impressed E. M. F. 
 of 100 volts and is taking 25 amperes, and if the losses 
 in the motor are 500 watts, then the energy effective in 
 producing rotation is 2,500 500 or 2,000 watts, as 
 against 3,000 in the dynamo. In order then that the 
 motor shall run at the same speed as the dynamo, its total 
 energy of rotation must equal the output of the dynamo, 
 and we have 2,500 -f- 500 = 3,000 watts necessary to be 
 supplied to the motor terminals. To do this the im- 
 pressed voltage must be raised from 100 to 120 volts. 
 Since TOO is 17 % less than 120 the speed of the motor at 
 100 volts will be 17 $ below that of the dynamo, on the 
 supposition that the speed varies directly as the im- 
 pressed E. M. F. 
 
 The ratio of speeds of the same machine, when run as 
 a dynamo and motor, is given approximately by the 
 fraction 
 
 // 
 
 where Jf 7 d is the total intake of the dynamo, and W m is 
 the output of the motor, /. e., the energy available for 
 producing rotation after all losses are deducted. 
 
 Small machines which are successful as motors do not 
 give satisfaction as dynamos, for of their output a large 
 part is consumed in energizing the magnetic circuit. 
 For since the reluctance of the magnetic circuit de- 
 pends very largely upon the air gap, the magnetic circuit 
 of small machines is of such high reluctance that nearly 
 or quite all of the output, of the machine would 
 
500 TESTING OF DYNAMOS AND MOTORS. 
 
 be absorbed by the fields. On the motor, energy is 
 furnished for exciting the fields over and above that 
 constituting the rated output of the machine, and on 
 small motors the question of efficiency is not so much 
 considered as that of getting the required work done; 
 for, since the entire output of the motor is small the 
 expense is not greatly increased even if the efficiency 
 be but 50 <f> or less. 
 
 Where a considerable amount of work is to be done the 
 -question of efficiency and of the number of machines 
 involved is of importance. It is an advantage to run one 
 machine at full load rather than two at half load, for the 
 efficiency is highest at full or overload; also the fric- 
 tional losses are less the less the number of machines 
 in use. To eliminate this evil as much as possible, 
 power plants are divided into large and small units. 
 When the load is very light a single small machine is 
 placed in service and, as the load increases beyond its 
 capacity, larger machines take its place, so long as 
 a single machine can carry the load. The machines are 
 so handled that all machines in service are running from 
 three-fourths to full load. 
 
 Motors are run, in common practice, in series and in 
 multiple, and it is possible, though not usual, to run 
 them in series-multiple if proper precautions are taken. 
 In general, it may be said that shunt and compound- 
 wound motors are run in multiple, as in the case of 
 factories or of isolated plants on lighting mains; and 
 series motors in both multiple and series as in street-car 
 work. In testing rooms separate excitation is also 
 resorted to where special advantages are sought for. 
 
 When running motors in series the prime requisite 
 
MOTOR TESTING. 501 
 
 is that their armatures be rigidly connected, as, for 
 example, where all are belted to the same countershaft, 
 or when two motors are geared to the car axles. To see 
 the necessity of this let us consider the case of two 
 shunt motors connected in series and run from a con- 
 stant potential dynamo, but without any mechanical con- 
 nection with each other. If both motors have a switch, 
 one may be closed and a starting box placed across the 
 other. As the box is cut out only that motor will start 
 across whose switch the box is placed, for this motor 
 alone possesses a field. Should the motors be separately 
 excited or be series wound, both may start, and that one 
 will start first whose frictional resistance and inertia is 
 the least, both being started without load. If load is 
 now placed on either one it will immediately slow down 
 and stop, while the speed of the other one will rise. If, 
 instead of placing the box across one switch, both 
 switches are closed, and the box is in series with the 
 system, then upon cutting it out that machine will start 
 whose friction is least and whose field is the strongest. 
 Both fields, however, will be very weak, and only one of 
 the motors will take a load. The effect of placing the 
 box in this last manner is to reduce the E. M. F. applied 
 to the fields as well as to the armatures, thereby greatly 
 reducing the starting power of the motors. Lastly, 
 both motors can be started by using two starting 
 boxes, which are worked simultaneously or by con- 
 necting the box so as to span both switches. But 
 while both machines will start one will stop as soon as 
 the load is put on. 
 
 These somewhat curious actions depend upon the dis- 
 tribution of the impressed E. M. F. between the two 
 
502 TESTING OF DYNAMOS AND MOTORS. 
 
 machines, and can be readily accounted for as follows: 
 An E. M. F. impressed upon a circuit is distributed 
 along the circuit according to the ohmic resistance and 
 the C. E. M. Fs. present in the circuit. With the two 
 motors in series the ohmic resistance of the circuit is 
 small and may be considered as uniformly distributed; 
 the C. E. M. Fs. are therefore the controlling factors. 
 In the first case considered, one motor, which we will 
 call A, has its switch closed, and the resistance from 
 terminal to terminal being very low, the potential differ- 
 ence (measured by / 7?) is also small; also since the 
 armature short circuits the fields, they cannot acquire 
 strength. The other motor, B, having an open switch 
 across which the box is placed, its potential difference 
 from terminal to terminal is comparatively high; so that 
 a strong field is obtained and, as the box is cut out, the 
 motor easily starts. As the speed rises the C. E. M. F. 
 rises, the armature current falls, and the I R drop in the 
 first motor, A, standing stationary, is lowered, while B 
 absorbs a still greater proportion of the impressed 
 E. M. F. 
 
 The manifest remedy for this state of affairs is to 
 rigidly connect the armatures, so that when B starts up 
 A is also set in motion. A would then develop a C. E. 
 M. F., build up a field, and take a share of the load. 
 
 In the case of the series motors connected in series, 
 the case is modified in that both motors start up, and 
 attain a good speed; but if a load is put onto one, its 
 speed will go down, while that of the free motor will go 
 up, until the loaded one is stopped and the free one is 
 racing. The reason is that the system is an unstable 
 one, and when once the equilibrium is upset it does not 
 
MOTOR TESTING. 503 
 
 return to a balance, but the disparity becomes greater 
 and greater. Thus when both are running the E. M. F. 
 may be said to be equally divided between the two 
 machines, but as the load is put on the speed of the 
 motor falls, its C. E. M. F. falls, and the potential 
 difference across the machine is lessened, and that across 
 the other machine is increased so that its speed rises. 
 This rise of speed in the free motor raises its C. E. M. F., 
 thus robbing the loaded motor of further voltage, so that 
 its speed continues to fall until at last it stops. Here 
 again the remedy is to rigidly connect the armatures. 
 
 In experimenting with motors in series, it is wise lo- 
 use only that voltage which each can safely stand singly; 
 for if one stops practically the whole E. M. F. is applied 
 to the running motor, and it may suffer injury. If the 
 plan of belting both motors to the same countershaft is 
 adopted, care must be taken that the connections are such 
 as to drive the shaft in the same direction. It is equally 
 important that the pulleys or gear wheels be of such size 
 that each motor shall run at its proper speed. Failure 
 to comply with this requisite may result in an overload of 
 one motor, and finally in burning it out. In making 
 these adjustments it is well to first run the motors alone, 
 and to take the proper observations. 
 
 If a motor designed to run at i, 200 revolutions per min- 
 ute is rigidly geared to one designed for 800 revolutions per 
 minute in such a way that they must have the same speed ; 
 the resultant speed will be a compromise between the two. 
 As to which motor is overloaded, depends upon whether 
 they are connected in series or in multiple. If they are in 
 series the current is the same in both, and the amount of 
 work done by each depends upon the distribution of im- 
 
504 TESTING OF DYNAMOS AND MOTORS. 
 
 pressed E. M. F. ; and this in turn depends upon the 
 C. E. M. Fs., and in lesser degree upon the ohmic resist- 
 ance. Any properly designed motor will for any current 
 generate its proper C. E. M. F. when running most nearly 
 to its rated speed, hence that motor which is nearest its 
 proper speed will have most nearly its proper load, and 
 the other one must take the balance. Thus if A is 
 designed for 1,200 revolutions per minute and is running 
 at 900 revolutions per minute, and B is designed for 800 
 revolutions per minute, then B has about its proper load 
 and A may be overloaded. 
 
 If the machines are in multiple then that machine 
 which has the highest relative speed has the highest 
 C. E. M. F., and the lowest current in it. It thus 
 throws the load onto the machine which is running below 
 its proper speed. 
 
 From the preceding pages and chapters some con- 
 ception can be formed of the amount of manipulation 
 and contriving that must necessarily be resorted to in 
 large testing rooms, whsre machines of different output 
 and various types are tested ; where machines of the same 
 output range in voltage from 25 to 500 volts, and are of 
 such dimensions that all the armatures can be run in the 
 same field frame, if the proper precautions are taken; or 
 where the field excitation may be varied, and the field 
 strength on all made the same, and a difference made in 
 E. M. F. or speed. In bringing the chapter to a close 
 some stray facts may be gathered up, and some of the 
 less frequently met with difficulties mentioned. 
 
MOTOR TESTING. 505 
 
 On fields having the same excitation or ampere-turns, 
 the number of watts consumed will be the same but the 
 resistance of 500 volt fields will be twice that of 250 
 volt fields, or four times that of 125 volt fields. Taking 
 the bipolar machine as the simplest type, involving the 
 use of but two spools, we may say that a pair of 500 volt 
 fields placed in series on 500 volts give the same excita- 
 tion and consume the same amount of energy as when 
 placed in multiple on 250 volts. 
 
 In running a 125 volt, 250 volt, or 500 volt armature, 
 as a motor or generator, in any of the various fields, care 
 must be taken that they are properly connected, and 
 when necessary the fields separately excited. One unac- 
 customed to handling machines in this way is apt to 
 forget that it will not do to place a 125 volt field across a 
 500 volt armature, and he may be forcibly reminded of 
 his error by a burnt out field or rheostat. Where the 
 fields are separately excited care must be taken that there 
 is no metallic connection between field and armature. A 
 mistake in this regard may throw the machine and its 
 exciter together, either in series or multiple, with the 
 result that the field is unduly strengthened in the first 
 case, or weakened in the second. If this mistake is made 
 on a separately excited loss supplier in a motor-generator 
 test, it is likely to manifest itself in uncalled-for variations 
 of speed. 
 
 As a general thing it is only the shunt fields that are 
 ever connected in multiple, but sometimes occasions 
 arise when an armature is run in a compound-wound field 
 whose current capacity exceeds the capacity of the series 
 coils. In such a case the series coils can be placed in 
 multiple to avoid cutting them out of service, care being 
 
5 6 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 taken that they are so connected as to strengthen and 
 not weaken the fields. 
 
 It sometimes happens that it is desired to reduce the 
 E. M. F. of a dynamo below its critical value. Separate 
 excitation facilitates this without the danger of the ma- 
 chine dropping its field; and this is often resorted to 
 where current is to be sent through a very low resist- 
 ance, which is to 
 be measured by the 
 method of fall of 
 potential. This ex- 
 treme weakening of 
 the field requires 
 the introduction of 
 a great deal of re- 
 sistance in the field 
 circuit, and a suffi- 
 cient number of 
 rheostats is not 
 
 QD-WWWH 
 
 FIG. 163. 
 
 always available. A very neat way of avoiding this diffi- 
 culty is to connect half of the field spools in opposition to 
 the other half, and to place the two halves in multiple. A 
 resistance box is then placed in series with one half as 
 shown in Fig. 163. As long as the box is cut out and 
 the current in the two sides is the same, the fields oppose 
 each other and the armature voltage is zero. If now Jt 
 is gradually cut in, the opposing force of F is weakened 
 and ^ produces a field, and the armature a correspond- 
 ing E. M. F. This method of regulation is becoming 
 much used in putting a full load current on a machine at 
 a very low voltage, for the purpose of a heat test. In 
 this way a dynamo may be short circuited through an 
 
MOTOR TESTING. 507 
 
 ammeter and full current gotten with an E. M. F. of 
 three or four volts. In the case of the large Edison 
 multipolar direct driven machines, the armatures are 
 always tested in this way, without the machine being 
 assembled in the factory. 
 
 The same range that is secured for the E. M. F. of the 
 dynamo by separate excitation is secured for the speed of 
 the motor by the same means. In this case the fields are 
 connected in multiple, and by raising the voltage the 
 field is made very strong and the speed correspondingly 
 reduced. 
 
 This method of field excitation probably has its ex- 
 treme example in the following method, which has given 
 satisfaction in the hands of the writers, and was called 
 into use when the field rheostats were not of sufficient 
 current capacity to be placed in the field circuit whose 
 resistance was to be varied. The method consists in ex- 
 citing the fields from two generators connected in oppo- 
 sition. When the voltage of the generators is equal and 
 opposite no current flows and there is no zero field. By 
 lowering the voltage on one machine a current flows of 
 the value 
 
 E - E' 
 
 E and E being the E. M. F. of the two generators and 
 r the resistance of the field circuit. To get the best 
 results the two generators should be themselves sepa- 
 rately excited. 
 
 On separately excited machines the brushes seem to 
 spark less for the same load variation than do those that 
 are self-excited; also the variation of potential attheter- 
 
508 TESTING OF DYNAMOS AND MOTORS. 
 
 minals, for a given change in the current, is less on a 
 separately excited than on a self-excited shunt dynamo. 
 The reason for this is that any change in the armature 
 current produces a weakening of the fields, while on a 
 separately excited machine the fields are constant aside 
 from the slight effect of armature reaction. 
 
 If, in separately exciting, the fields are placed in multi- 
 ple, it must be remembered that the current carrying 
 capacity of the boxes must be doubled, or a burn out 
 may ensue. When a box shows signs of giving out, it 
 should be relieved by placing a second one in multiple 
 with it, and then the defective box cut out. To cut out 
 the box it is not always necessary to remove the connect- 
 ing wires, but it is sufficient to turn the handle to the 
 "all out" position. Unfortunately, makers of rheostats 
 have not yet agreed upon a uniform construction for 
 rheostats, so that some turn from right to left and some 
 from left to right to accomplish the same result. Thus 
 to introduce resistance, the handle of an Edison box is 
 turned from left to right, while with the Thomson-Hous- 
 ton box the opposite is true. These minor points must 
 be familiar to the tester, for a wrong movement at a 
 critical time may precipitate serious trouble. 
 
 A recent writer has drawn attention to the need of uni- 
 formity in the rating of and specifications for dynamos 
 and motors. Ample justification can be found in the 
 preceding chapters for this remark, and in closing this 
 survey of electrical testing, the writers would add their 
 voice to this timely plea. No tester who has been long 
 upon the floor but knows the added troubles resulting 
 from even slight differences of detail in the instructions 
 for the various tests. Perhaps no one thing would 
 
MOTOR TESTING. 509 
 
 more help to make electrical testing Simpler, and freer 
 from needless annoyances, than such a reform. Ex- 
 perience would soon crystallize into better-defined and 
 more uniform methods, and the special test no longer 
 arise, like a spectre, to trouble the dreams of the Weary 
 Tester. 
 
CHAPTER XV. 
 
 INSTALLATION CAR EQUIPMENT TESTS. 
 
 The subject of testing a railway car equipment will, 
 in this chapter, be considered from the standpoint of the 
 operator, rather than from that of the designing or in- 
 stalling engineer. It will be assumed that the equipment 
 motors, controllers, trucks, etc. has been purchased 
 from a responsible concern and is in every way up-to- 
 date in design and workmanship. It is quite safe under 
 present conditions of competition to assume that all ques- 
 tions of efficiency, temperature rise, speed, torque and 
 current consumption have been reduced to such a basis, 
 as to be almost as readily comprehended by the buyer as 
 by the maker. 
 
 The controller maker, for example, has so completely 
 mastered all the points involved in starting and stopping 
 a car under various conditions of load, grade, speed, 
 schedule and even handling, that he knows just about 
 what device to recommend to fulfil certain maximum 
 conditions of use and abuse in a given service. So also, 
 the truck builder is willing to vouch for his part of the 
 equipment, feeling that in the article offered, a liberal 
 factor of safety based upon experience makes all due al- 
 lowance for square curves, bad rails, and to some extent, 
 lack of attention to wearing parts. The claims of the 
 
 510 
 
INSTALLATION CAR EQUIPMENT TESTS. 51 1 
 
 various parties relative to the excellence of their wares 
 can, as a rule, be conceded, provided they show a willing- 
 ness to waive payment until a fair trial in service shall 
 confirm their claims. 
 
 After a high grade equipment has been secured, in- 
 stalled and is in satisfactory operation, eternal vigilance 
 is the price of continued good service and the object 
 here will be to point out the conditions of abuse 
 under which trouble is most apt to arise in the several 
 parts of the equipment, to show how to avoid these condi- 
 tions and to give the most common methods of locating 
 and remedying the faults to which even the best of 
 equipments must ever be liable. There is, perhaps, no 
 more logical order in which to discuss the devices com- 
 posing the equipment of a car than to take them up in 
 the order in which the current passes through them on 
 its way from the trolley wire to the rail. This order will 
 be as follows : I The trolley, II The Overhead Switches, 
 III Fuse Box, IV Lightning Arrester, V Controller (in- 
 cluding the starting coil and motors.) 
 
 I THE TROLLEY. This device comprises two prin- 
 cipal parts the stand and the/WI?. The stand comprises 
 an upper pivotal member called the base and a lower 
 stationary member called the foot. Fig. 164 shows a 
 base complete: C is the foot, all the rest of the device is 
 the base. C screws to the top of the car and acts as a 
 pivot for the base to turn on. A A' is a single casting, 
 the upper end of which is a socket adjustable to receive 
 the pole; the lower end, A ', is forked or yoke shaped 
 and engages the compression rods, Z>, that actuate spring 
 F, when the pole is pulled down. A A' turns on center 
 
 
 
512 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 f. Stem G is fixed to casting B B, and serves to sup- 
 port and guide compression cups, IT, and the springs 
 F and E, Spring E is the kick spring and H is a nut 
 used to vary the compression on F. The action of the 
 stand is as follows: in the figure it is in its neutral posi- 
 tion and the pole should stand nearly vertical. To put 
 the pole on the wire it must be rocked down clock- 
 wise; this swings A' clockwise around J, moves rods 
 
 FIG. 164. 
 
 D to the left, and with them compression cup, /', 
 thereby compressing F. The lower the pole is pulled, 
 the more is F compressed and the more upward pressure 
 does the pole exert against the wire. Spring E is pro- 
 vided to cushion the blow received should the pole get 
 away in service. Suppose the rope breaks; as soon as 
 A A' rocks past the neutral position, rods, D D, pull 
 on r and compress E\ this cushions the blow and lessens 
 the chances of breaking any of the parts. By tightening 
 nuts If, cup / is forced along stem G, compressing spring 
 
INSTALLATION CAR EQUIPMENT TESTS. 513 
 
 /', and thereby increasing the upward pressure of the 
 pole at all positions. To keep a stand in good W0 rking 
 order, pins J and K and the pivot around which base B H 
 turns on C C, should be oiled auout once a week. 
 Also alxMit once a month, stiff grease should be applied 
 to stem G so spring F may work freely. If G and J are 
 not kept lubricated, the action of the whole device be- 
 comes sluggish, the pole will not work up and down 
 freely and will tend to jump the wire even on a straight 
 track. If the main pivot is allowed to run dry, the pole 
 is sure to leave the wire on curves and cross-overs. The 
 first symptom that a trolley stand needs oil, is that the 
 whole thing works hard when the wheel is placed on 
 the wire from which it has jumped. With neglect all 
 moving parts wear loose. 
 
 THE POLE. This comprises the pole proper, the ferrule 
 the harp and the wheel. Fig. 165 shows a pole complete; 
 P, is the pole proper; H, the harp; F. the ferrule; S t the 
 contact spring and W, the wheel. Most poles are made 
 of hard drawn steel, which will bend only under a heavy 
 force; once bent such a pole should be straightened cold 
 as the character of the steel is such that it looses temper 
 if heated. Poles are from 12 to 16 ft. long, according to 
 the height of the car and the height of the wire. 1 1 is 
 i^ in. diameter on the big end, holds this diameter for 
 about 2 ft. and then tapers to i in. on the harp end. The 
 ferrule is a brass or iron ring with an eye hole to take 
 the trolley rope. It is free to turn on the harp stem as a 
 center. The harp, H, is a brass or malleable iron fork 
 designed to be riveted on to the small end of the pole to 
 hold the wheel, If. Iron has almost entirely supplanted 
 
TESTING OP DYNAMOS AND MOTORS. 
 
 brass in such uses, as it offers less temptation to thieves 
 and costs less. The tangs of the fork are drilled to take 
 the axle, A, around which W turns as a center. The 
 axle is secured on both ends by a cotter 
 pin, which lies in a groove and keeps the 
 axle from turning. A good harp saves 
 line work, the main requirement being 
 that it have 110 corners or swells to catch 
 in a line frog if the pole goes off under 
 headway. The harp should be light. 
 The wheel is made of brass or gun metal ; 
 if too soft its life is short, -if too hard it 
 wears the wire. The amount of wear 
 between the two also depends upon the 
 shape of the groove, the condition of the 
 line work and the care bestowed upon 
 the adjustment of the base and pole. 
 Some wheels are solid, others have re- 
 movable centers called bushings. A 
 trolley bushing is a spiral brass casting 
 whose outside is hollow milled to drive 
 into a hole in the wheel and whose inside 
 is bored to fit the axle. The air chan- 
 nels beeween the metal spirals are filled 
 with a graphite compound. As the bear- 
 FIG. 165. ing is the first part to wear out under 
 normal conditions, the use of a bush- 
 ing prolongs the life of a wheel. To replace a worn 
 bushing, it is driven out and a new one driven in. L,ives 
 of trolley wheels vary widely with local conditions, but 
 should average 5000 miles under good care. 
 
INSTALLATION CAR EQUIPMENT TESTS. 51 5 
 
 Every road should experimentally determine the trolley 
 characteristics best suited to its own conditions. By 
 trolley characteristics are meant : pole tension, weight 
 and material of the pole ; size, weight, shape and mate- 
 rial of the wheel ; width, depth and general contour of 
 the groove ; shape of the groove lips or flanges. Such 
 data can be gotton only by a series of careful tests. Such 
 a series of tests alw r ays discloses facts not looked for and 
 should be conducted with great care so that results may 
 be relied upon. The test consists in installing several 
 kinds of wheels in service and keeping track of their 
 records. Each kind should be put on at least six cars 
 and on more if possible. The mileage of each wheel 
 must be kept exactly, noting delays, accidents, etc., and 
 one wheel of each kind must be kept so that the shapes 
 of the grooves of the old wheels and new ones can be 
 compared. As the condition of the trolley stand influ- 
 ences the life of the wheel greatly, a series of preliminary 
 tests must be made to insure that all of the wheels start 
 out on an even basis. All poles must be of the same 
 size, weight and length and must be of the same tension. 
 The tester must be certain that all parts of the trolley 
 outfit are greased and kept greased, so that the motion 
 of the wheel on the axle, of the socket up and down and 
 cf the base on the foot pivot will always be free and easy. 
 When a trolley outfit is set up correctly, all parts are well 
 oiled and aligned, the harp is on the pole straight and 
 the pole is in the socket straight, so that the wheel when 
 on a straight wire has its flanges parallel to the wire. 
 To insure that all these conditions obtain, select inside of 
 the shed, a stretch of the wire about 50 ft. long, drop a 
 
516 TESTING OF DYNAMOS AND MOTORS. 
 
 plumb line down from the ends to the track below and draw 
 a chalk line through the two points touched by the bob. 
 This chalk line should be everywhere equidistant from 
 the two rails ; if it is not so, the trolley wire must be 
 pulled over until it is. The car can now be run under 
 the truly centered wire and the trolley device adjusted 
 to the proper tension (about 16 Ibs. as a rule) and all the 
 parts aligned. If the adjustment is good, the pole, when 
 on the wire, will lean to neither one side or the other, 
 the wire will rest in the bottom of the groove and will be 
 parallel to the flanges. The set should be the same from 
 both ends of the car, and in the case of double spring 
 bases whose poles rock over, as well as swing around, 
 the tension should be the same in both cases. Rock-over 
 bases have the advantage that if the pole gets bent con- 
 ditions can be improved by rocking it over so as to bring 
 the bend down. Sometimes, even when the parts are 
 all well oiled and adjusted, the base will work hard, be- 
 cause it is loose on the pivot, due to wear, and binds 
 when one tries to swing it. This fault, as a rule, occurs 
 only on old stands. 
 
 Assuming the bases and poles to be set right, the fol- 
 lowing conclusions can be drawn in regard to the wear of 
 the wheels. If both flanges of the wheels persist in 
 wearing to a razor edge before the bushing is much worn, 
 the groove is too deep or too narrow, or both, or the in- 
 side edge of the flange is not flared enough. If the wire 
 wears down into one side of the groove bottom and one 
 flange only gets sharp, the trolley wire is out of center 
 somewhere or the pole is out of set. The first conclusion 
 is verified if the same wheel does not give this trouble on 
 
INSTALLATION CAR EQUIPMENT TESTS. 517 
 
 other parts of the road. Where the wire cuts down into 
 the center of the groove very rapidly, it prooves that the 
 metal is too soft and the tension too strong for that par- 
 ticular make of wheel. A chattering noise when in 
 motion indicates either that the wheel is flat or the bush- 
 ing worn. A flat wheel may be due to lack of oil causing 
 it to stick, or to a bent harp or to soft spots in the metal. 
 A worn bushing may be due to want of oil or to soft- 
 ness. The makers may get it too soft in the effort to 
 have it self-lubricating. It is better to have it hard 
 enough and to keep it oiled. Also the axles themselves 
 wear small in course of time ; all wheels should be gauged 
 as well as the axles. Occasionally a road will get hold 
 of a lot of wheels that are bored out of center and that 
 will cause them to emit the flat wheel chatter. Where 
 the wheels show a rough pitted appearance on the lip of 
 the flanges, it shows poor handling in running through 
 overhead frogs where the flange must carry all the cur- 
 rent ; the same roughness is carried to a degree where the 
 poor design of the line causes the wheel to jump at every 
 ear. The rougher a wheel is, the more apt is it to leave 
 the wire at curves and crossings. Finally the best shape 
 for a wheel is indicated by the wear of those that have 
 been in service. 
 
 THE SPRING. The trolley wheel spring is made of 
 copper and serves two purposes ; it acts as a soft metal 
 washer between the hub of the wheel and the side .of the 
 harp and saves wear. It also conducts the current from 
 the wheel to the harp. Where no spring is used, the bear- 
 ing surface between the wheel and axle must carry most 
 of the current with the result that i* gets pitted and rough. 
 
TESTING OF DYNAMOS AND MOTORS. 
 
 II THE OVERHEAD SWITCH. This device is placed 
 under the hood just above and a little in front of the 
 motorman's head, with its handle so turned that the 
 most involuntary movement of the motorman in the time 
 of trouble will be to slap it " Off." The object of the 
 
 G 
 
 FIG. 1 66. 
 
 device is to provide a simple and certain means of break- 
 ing the current, should a ground or other fault render 
 the controller unable to do so. When this switch is open, 
 no current can get to the motors or controllers and so it is 
 used for ' ' killing ' ' the circuit w r hen it is desired to in- 
 spect or work on any part of it. Again, there may be 
 
INSTALLATION CAR EQUIPMENT TESTS. 519 
 
 trouble with the controller or some other device, such 
 that it may be convenient to put the controller on the ist 
 notch, and run the car to the house by means of this over- 
 head switch. The most complete type of the device in 
 use to-day, is the combined overhead switch and cir- 
 cuit breaker. Fig. 166 shows such a device very much 
 in use. In the Fig. A, is the positive terminal; /?, the 
 negative; C, a coil which operates the tripping device and 
 also operates the blow-out; 2), the push button by means 
 of which the tripping device may l)e operated by hand; /?, 
 a spring that holds the keeper, A', away from its coil, C\ 
 unless the current exceeds the value 
 at which the breaker is set to act ; JL. JL 
 
 F is a graduated scale ; / is the 
 sighting disc, by means of which 
 and thumbscrew, G, the tension on 
 E is regulated; H, is the handle C 1 
 for resetting the breaker. Fig. 
 167 is a diagramatic sketch of the 
 connections revealed by removing 
 the name plate. A, B, C and 
 // are the same as in Fig. 166. The 
 breaker is as follows : suppose G is adjusted till / is on a 
 level with 175 on the scale; the breaker is then supposed 
 to act at 175 amps. For any current less than this, E 
 holds K against the pull of C ; at 175 amps, the pull of 
 C overcomes E and draws A' down, thereby liberating the 
 trigger on K from a shoulder carried on handle, H. A 
 spring then carries the handle to the "off" position, 
 opening the circuit between E and E. The arc formed 
 moves up to the auxiliary breaking blocks E ', 17, and is 
 
520 TESTING OF DYNAMOS AND MOTORS. 
 
 there extinguished by the magnetic action also provided 
 by coil C. For use on cars the breaker is enclosed in a 
 wooden case, out of which sticks the handle, A, and the 
 push button, D. To reset the device after action, the 
 handle is shoved to, the same as on any ordinary switch. 
 To operate the switch by hand, simply press the button; 
 this shuts the keeper down the same as the current does 
 on overload and liberates the handle. 
 
 FIG. 168. 
 
 There are several important points to be watched 
 about a circuit breaker, ist. It must be tested at inter- 
 vals to see that it acts at the current value for which it 
 is set. To do this have a reliable ammeter and a water 
 rheostat connected in series, as indicated in Fig. 168, 
 where A is the meter; R, the water rheostat and K an 
 ordinary switch. Test line T terminates in a hook that 
 can be hooked over the negative post on the breaker, 
 and all disconnecting being thereby avoided. To test a 
 breaker on a car, throw the controller off and hook test line 
 
INSTALLATION CAR EQUIPMENT TESTS. 
 
 521 
 
 T on to the negative side of the breaker; after removing 
 the cover, note the value to which the breaker is set; if the 
 value is right, close K and work the current up to that 
 value ; the breaker should act within 5 amps, of it. 
 
 TROLLEY 
 
 FIG. 169. 
 
 If it fails to, readjust by means of G. 2nd. The blow, 
 out chamber should not be allowed to accumulate carbon 
 dust. At least once a year the insulation between the 
 break points should be measured as follows : as indicated 
 in Fig. 169, run a test line from the trolley wire to the 
 positive side of a 500 volt voltmeter; from the negative side 
 of the meter run a second test line terminating in a hard 
 point ; next tie down the trolley pole and see that the 
 breaker is opened, and put one controller on the first notch. 
 The test consists in touching the positive side of the open 
 
522 TESTING OF DYNAMOS AND MOTORS. 
 
 breaker with the free test line. If the arc chamber is car- 
 bonized, a current flows from the trolley wire, through the 
 voltmeter, across the carbonized path to the ground. If 
 the deflection is more than 450 volts, the name plate should 
 be removed and the arc chamber cleaned. 3rd. Under no 
 circumstances should the breaker be placed where the 
 driver cannot reach the handle without jumping ; if it is, he 
 will use a switch iron with the result that in a short while 
 the interference lug on the handle will be so impaired 
 that the breaker cannot be set ; if this occurs on the road 
 it will be necessary to cut the device out entirely or to tie 
 the handle over in order to run the car to the house. 
 4th. Where a breaker has been neglected and allowed to 
 stick, or where the capacity of the motors has been in- 
 creased without regard to the breaker, the insulation on 
 the wire of the blow coil may get so badly roasted as to 
 short circuit it. In such a case, the coil will not operate 
 at the value set and hence the device is of little pro- 
 tection to the motors ; also, when it does act, there is bad 
 arcing in the arc chamber. When there is reason to 
 suppose that the blow coil is baked, it can of course be 
 tested by one of the several ways given in Chaper VI, 
 but the best way is to remove a part of the insulation 
 from the coil and look at the wire ; if the cover is brown 
 and can be scraped off with the nail, it is baked, and the 
 coil must be renewed. When a coil is baked, there is 
 always a protracted characteristic sputtering when the 
 arc is extinguished. When the breaker is in good 
 order, there is no noise save a puif somewhat similar to 
 the exhaust of a loaded engine. 
 
INSTALLATION CAR EQUIPMENT TESTS. 523 
 
 The strongest point about a breaker-overhead-switch is 
 that it will not allow a motorman to "notch" his controller 
 above a certain rate, and in emergency reversals he does 
 not put the controller beyond the first notch. Unscrupu- 
 lous men can and do get around this feature sometimes by 
 holding the handle over with a switch iron, but such prac- 
 tice should be met by severe discipline. A car breaker is 
 much more of a protection than a car fuse, as is evidenced 
 by the fact that cars run into the house on account of 
 roasted fields, could not, after being equipped with a 
 breaker set to act at the fuse rating, be even started, be- 
 cause the breaker would act on the first notch. 
 
 On all up-to-date cars there is a hood-switch or breaker 
 on both ends, so that in time of trouble the cutting off of 
 the power is under the control of both members of the 
 crew. It is safe enough, and less expensive, to have a 
 plain switch on one end, and a breaker on the other. 
 
 Ill THE FrsE-Box. This is also a safety device. As a 
 rule a car has a fuse-l>ox,even if it has a breaker, and where 
 there is no breaker it is absolutely essential that there be a 
 fuse to protect the devices, in case of a short circuit. The 
 weakest part of a string will break ; so, also, the weakest 
 part of an electric circuit will burn out when the circuit 
 is overstrained. This weakest part might prove to be a 
 loose connection or a bad contact, but it is more than likely 
 that the weak spot will show up inside of a motor where it 
 costs more to repair the damage. The idea of the fuse-box 
 is to provide a weak spot, which, in case of a short circuit, 
 will give way before any other spot does. To insure this, 
 the fuse, if of copper, is made smaller than any other wire 
 to be found in the main motor circuit or any of the other 
 
524 TESTING OF DYNAMOS AND MOTORS. 
 
 devices that carry the main current. A 30 h-p street rail- 
 way motor armature is wound with a No. 8 or No. 9 B. & 
 S. wire, and the field with a No. 4 or No. 5 B. & S., ac- 
 cording to the work to be done. The size of the armature 
 wire is one-half that of the field, because the two halves of 
 the armature are in multiple, and each wire carries 
 but half the current, so the field and armature have 
 the same current capacity. The fuse wire must, however, 
 be a great deal smaller than the field wire, because, being 
 in an exposed place, its facilities for radiation are greater, 
 and a much larger current is required to melt it than is 
 required inside of a hot, closed motor. Again, the shorter 
 a fuse, the smaller must it be to melt at a given current 
 value, because the lugs to which it attaches conduct the 
 heat off. As a final result of all these influences, it is 
 found that two 30 h-p motors should be protected by a 
 fuse wire no larger than a No. 12 B. & S., and two 50 h-p 
 or 60 h-p motors, properly handled, can get along on a 
 No. 9 B. & S. This assumes a distance of 4 in. between 
 lugs. 
 
 Every road using copper fuses should determine, by 
 test, the size of wire needed to protect each type of equip- 
 ment. Such a test must be based upon a knowledge of 
 the full load capacity of the motors; their average daily 
 current, and the length of the fuse to be used. The cur- 
 rent capacity of a motor is gotten by multiplying the rated 
 horse-power by 746, the watts per horse-power, and divid- 
 ing by the line voltage. For example, a 3O-hp motor on a 
 line averaging 500 volts would have a current capacity of 
 
 30X746 = ^>3 = 44.7 Now, it is the current 
 
 500 500 * ' 
 
INSTALLATION CAR EQUIPMENT TESTS. 525 
 
 that burns out a wire, so it is not hard to sec that where 
 the voltage on a line is allowed to get low and the time 
 table is kept the same, the motorman, in his effort to keep 
 on time, must abuse his motors, increase their average 
 current consumption, and, therefore, the size of the fuse 
 required. The average current consumption can be got- 
 ten best by the use of a wattmeter. The wattmeter is 
 connected on the car (the field in series with the overhead 
 switch, the armature across the line), and its reading 
 
 FIG. 170. 
 
 taken just before the car goes out on its first run. When 
 the car comes into the house at night the meter is read 
 again ; the difference between the two readings gives the 
 number of watt-hours absorbed during the day of, say, 
 twelve hours. Dividing this by 1 2 and the voltage on the 
 line gives the average flow of current during that time. In 
 the absence of a wattmeter, an ammeter must be used, but 
 this is very tedious, for, in order to get correct results, the 
 readings must follow each other at very close intervals, say, 
 ten to the minute. At the end of the test the current read- 
 ings are all added together and divided by the number of 
 readings; this gives the average current. Knowing the 
 
526 TESTING "OF DYNAMOS AND MOTORS. 
 
 average current and the full load current, a wire is selected 
 which, at the given distance between lugs, will stand con- 
 tinuously that current which is half-way between the full 
 load and average values. Say the full load current of the 
 car is 90 amps, and the average current 40 amps. , then 
 the current for which the fuse must be selected is 65 amps. 
 
 Let us assume that the length of the fuse is to be 
 4 ins. Two metal blocks provided with connecting posts 
 must be rigged up as shown in Fig. 170, where T is the 
 trolley wire ; K, a switch ; BB, the two blocks ; F , the fuse 
 wire R,< a water resistance; A, an ammeter, and G, the 
 ground. // are two pieces of asbestos or fiber, through 
 which the fuse is run to prevent the blocks, BB, from be- 
 ing burned by the arc. With the fuse short-circuited, J? 
 is adjusted until A registers 65. The short circuit is then 
 removed and the fuse let into circuit and left there for 
 about five minutes, when it is just as hot as it will ever be. 
 A dark box is then set over the fuse, blocks and all, the 
 box having in the top a small hole through which the fuse 
 can be seen. The fuse seen in the dark should show the 
 faintest pink color ; if it does not, a smaller size wire must 
 be selected, and so on until the right size is obtained. A 
 fuse selected on this basis will, under normal service con- 
 ditions, blow once in about every two weeks, from 
 gradual deterioration. 
 
 EXTINGUISHING THE ARC. Were no means provided 
 for extinguishing the arc that tends to hold when a fuse 
 blows under heavy current on a 5oo-volt circuit, the fuse 
 blocks would soon be burned to a state of uselessness. 
 Several schemes have been devised for suppressing the 
 arc. Among them are : confinement in an air tight 
 
INSTALLATION CAR EQUIPMENT TESTS. 527 
 
 chamber ; the use of spring flaps to break the arc ; the 
 use of an extra long fuse ; the use of a magnetic blowout. 
 The magnetic blowout type made by the General Flectric 
 Company has successfully withstood the test of years, and 
 will be selected for description. 
 
 Fig. 171 shows the fuse-box complete ready for a fuse, 
 and Fig. 172 is a diagrammatic sketch of the inside con- 
 nections. In Fig. 171 are holes, through which the 
 two circuit wires go to connect in the box : through holes 
 bb a screw-driver can be put to tighten the connections, 
 
 abe, shown in Fig. . 7 
 
 172; cc in both V t* 
 
 fi g u r e s are the 
 thumb-screws to se- 
 cure the two ends 
 of the fuse wire. /, 
 Fig. 171, is a raw- 
 hide cover to keep p I(; j^T 
 out the water. No 
 
 substantial lid is needed, as there is no demonstration 
 when a fuse blows. Fig. 172 (b) shows the special fuse 
 used ; a special fuse is not absolutely necessary, as it is 
 very easy to secure a plain wire under the thumb-screws, 
 but it is a good idea, in that it % lessens the chances of the 
 wrong size of fuse being used. 
 
 POINTS ON FUSE- BOXES. On many roads the proper 
 ' ' fusing ' ' of the cars seems to be treated as a matter of 
 minor importance. Trucks, motors, etc. , are shifted from 
 one car to another without any apparent attention being 
 paid to the requirements of the fuse-box, which being 
 fastened to the car body, remains there, a victim of circum- 
 
528 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 stances. Most roads operate several capacities of motors ; 
 as a rule, a motorman carries a bunch of fuses, any one of 
 which, in the majority of cases, he regards as adapted to 
 use on any car he may happen to be running. If allowed 
 to choose his own fuse he is apt to choose the biggest. In 
 this way, large motors get small fuses and small motors 
 large fuses. The use of a fuse that is too small entails no 
 bad results beyond the annoyance of having it blow con- 
 tinually for no other reason than that it is too small. On 
 
 0000 
 
 (a) 
 
 the other hand, the use of a fuse that is too large deprives 
 the car of the very protection that the fuse is intended to 
 give. With a fuse of the right size a car would never 
 get away from the shop with the fields on the motor con- 
 nected wrong, and the motorman would have to take a 
 loaded car up a grade and around curves on the series 
 notch of the controller to avoid blowing the fuse. With 
 a fuse of the right size it would not be possible for a car 
 to stay in service for weeks at a time with its fields baked, 
 calling for a new controller or brush holder or armature 
 once a week. A fuse wire may be said to be of the right 
 size when, under normal conditions of service, it will blow 
 
INSTALLATION CAR EQUIPMENT TESTS 
 
 529 
 
 in about two weeks after being put in. The reason it is 
 more apt to melt after two weeks of service is that the 
 heat oxydizes the wire, thereby impairing its contacts. 
 
 It matters not what kind of a fuse-box may be on a car, 
 it is safest to throw off one of the overhead switches be- 
 fore renewing a fuse. It is then impossible to get burned, 
 shocked or have the car start unexpectedly. To get a 
 shock, it is only necessary to leave the pole on the wire, 
 
 N 
 
 
 
 r 1_ 
 
 j 
 
 t 
 
 
 r~ 
 
 ["LLSj" 
 
 
 , -, 
 
 
 JLAJ 
 
 
 
 1?9 
 
 
 
 
 o 
 
 j, 
 
 
 j 
 
 
 \ 
 
 
 I 
 
 I' i 
 
 - ; 
 
 /' i 
 
 
 E 
 
 FIG. 173. 
 
 both switches in and, standing on the ground or resting 
 the hand on any part of the truck, dash or gate irons, to 
 touch the positive side of the fuse-box. To get burned 
 under like conditions, it is only necessary that there be a 
 ground on the trunk wire between the fuse and the con- 
 troller trolley post. 
 
 Fig. 173 is a diagramatic sketch of the trunk wire 
 from the positive side of the fuse-box to the negative side 
 of the blow coil on one controller, and to the trolley post 
 on the other. Nc is the wire that runs from the No. 2 
 overhead -switch to the fuse-box, FB; LA is the arrester; 
 
530 TESTING OF DYNAMOS AND MOTORS. 
 
 //, the two controller trolley posts, and tot, the car 
 wire that connects them ; BC is the blow coil. As long as 
 the condition of the circuit is normal, the path of the cur- 
 rent is Nc-FB-LA to the tap, x; from here the path de- 
 pends upon which controller is being used. If A is used, 
 the path from x is x-o-t-BC-t'; on through the controller 
 and motors to the ground. Suppose that a car under head- 
 way suddenly develops a ground on the trunk wire at 
 any of the points indicated by the dotted lines, g^ , g 2 . . . . 
 "5> g*\ the car will halt in its speed; if on a grade, it will 
 do so suddenly, as inertia lends less influence and the 
 blowing of the fuse cuts off the ' ' power. ' ' The motor- 
 man feels the check in the speed, and throws the controller 
 to the "off" position, which act can make no difference, 
 as the fault is ahead of the moving part of the controller, 
 and therefore cuts it out. Such a fault is often due to a 
 grounded blow coil, or to some part of the trolley wire 
 being rubbed by a brake-rod or car wheel, both of which 
 are dead grounds. An old motorman's first move is to 
 try his car lamps, to see if the line is dead. If the lamps 
 light, he looks at the fuse to see if it is gone, and in this 
 particular case, finding that it is, proceeds to renew it. 
 Once in a while the fault will burn itself out at the same 
 time that the fuse does, but rarely. The fault that caused 
 the fuse to blow in the first place is ready to do so again. 
 The result, then, of trying to replace a fuse, with the pole 
 on and switches in, under any of the above conditions, 
 would be to have the fuse blow in the motorman's hand or 
 face and, perhaps, burn him badly. 
 
 Now, suppose that for no other reason than that the 
 fuse is old, it gives out while the car is under head- 
 
INSTALLATION CAR EQUIPMENT TESTS. 531 
 
 way, and that the motorman, for some reason or other, 
 fails to throw his controller to the "off" position before 
 getting off the car. As soon as both ends of the new 
 fuse touch the blocks the car starts. In nine cases out of 
 ten the motorman will get a shock or burn, which will 
 place him in very poor condition to catch the car should 
 the fuse catch or weld, thereby enabling the car to keep 
 going. 
 
 IV THE Lir.HTNiNr, ARRESTERS. The fields of motors 
 and dynamos are powerful electro-magnets of great self- 
 induction and on some of them it takes the line voltage a 
 quarter of a minute to work the current up to its full 
 value. The voltage of lightning is up in the mil- 
 lions and, when it strikes a circuit, something must break- 
 down to give it a path to earth. One side of a street rail- 
 way motor is always grounded, when the line has a 
 ground return. The path from the trolley, through the 
 motor, to the rail, tempts the lightning and it moves along 
 this path smoothly, until it gets to the motor itself, where 
 its path is blocked by the self-induction of the motor parts ; 
 but the discharge with such an enormous pressure behind 
 it can not be withstood ; it must get to earth some way, 
 so it jumps right through the field or armature insulation 
 to the motor frame. The path through the motor wind- 
 ing is an inductive path ; the path through the insulation 
 is a non-inductive path. Lightning always takes a non- 
 inductive path where it has a choice, but where there is no 
 choice, it has the power to create for itself a non-inductive 
 path by forcing a short cut through the inductive path. 
 
 In Fig. 174, T-a-G is a piece of wire bent into a loop; 
 one end sticks up in the air at T, the other into the ground 
 
532 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 \ 
 
 G 
 
 FIG. 174. 
 
 at G; o is an air gap; T-o-G is a non-inductive path; 
 
 T-a-G, an inductive path. If by means of an influence 
 machine a discharge be passed into the wire 
 at T, it will jump the gap at <?, rather than 
 force its way through the self-induction of 
 the single loop. 
 
 The function of a lightning arrester is to 
 give to the discharge a non-inductive, low 
 resistance path to the ground, so that the 
 motor insulation will be spared. The 
 arrester humors the lightning it dares not 
 combat, 
 In Fig. 175, T is connected to the trolley wire and G to 
 
 the rail ; A and F are the 
 
 armature and field of the 
 
 motor to be protected ; L is 
 
 a simple form of aresrter, 
 
 L consists of two brass 
 
 plates held so close together 
 
 that the air gap between 
 
 them is thinner than the 
 
 thinnest insulation to be 
 
 found on the motor ; one 
 
 plate connects to the trunk 
 
 wire from the fuse-box, the 
 
 other plate to the ground. 
 
 A discharge coming in at 
 
 T, jumps the gap at L and 
 
 goes to earth at G. The 
 
 arrester outfit shown in 
 
 Fig. 176 differs from that 
 
 1 II 
 
 
 v * 
 
 nr^ 
 
 L 
 
 
 
 
 
 A? 
 
 AC 
 
 2 
 
 
 i 
 
 
 G |G 
 FIG. 175. FIG. 176, 
 
INSTALLATION CAR EQUIPMENT TESTS. 
 
 533 
 
 in Fig. 175 in two details: 
 First, the two plates have 
 comb-shaped edges, be- 
 cause of the greater dis- 
 charging power of points; 
 second, it has coil .9, call- 
 ed a "kicking coil," or 
 "kicker." The kicker 
 is inserted to make the 
 motor path more uninvit- 
 ing to lightning on the 
 principal that lightning 
 discharges are alternating 
 in character, and that an 
 inductive resistance 
 properly placed renders it 
 
 FIG. 178. 
 
 FIG. 177. 
 
 practically im- 
 possible for the 
 lightning to take 
 that path. The 
 kicker is noth- 
 ing more than 
 about a dozen 
 turns of the 
 trunk w ire 
 wound on a 
 wooden core. All 
 arrester wires 
 should be 
 straight so that 
 their self-indue- 
 
534 
 
 TESTING OP DYNAMOS AND MOTORS. 
 
 tion will be a minimum. The air gap in an arrester 
 should not be over 1-64 in., to be reliable in railway 
 work. The 500- volt trolley pressure can not jump 
 across the gap in an arrester, but it is strong enough 
 to follow the lightning discharge over. Means must be 
 taken to stop the arc set up by the trolley voltage, or the 
 points of the arrester will be destroyed. To extinguish 
 
 FIG. 179. 
 
 FIG. 1 80. 
 
 the arc that follows the discharge several devices have 
 been used. The one that has perhaps been most effective 
 in this field is the magnetic -blowout incorporated in the 
 arrester of the Gen. Klec. Co. 
 
 Figs. 177, 178, 179, 1 80 and 181 show a type of mag- 
 netic blowout arrester that has taken a strong hold with 
 the railway public. Fig. 177 shows the general appear- 
 ance of the arrester, and Fig. 178 shows it in a box ready 
 
INSTALLATION CAR EQUIPMENT TESTS. 555 
 
 t'o be put on a car. Fig. 179 shows the bare arrester with 
 the lid off, and Fig. 180 shows the lid. In Fig. 179, c is 
 ihc blow coil; r, a non-inductive resistance; h, h, the 
 magnet poles, and k /, the knife blades that enter jaws 
 magnet poles, and A'A'. the knife blades that enter jaws 
 K K ^ Fig. 180, of the lid when it is on. The air gap. a, 
 Fig. 1 80, is only .025 in. wide. The air gap is cut into cir- 
 cuit when the lid is put on, and it falls directly between the 
 horns of the blow magnet, so that any arc is quickly sup- 
 pressed. The clip, /, Fig. 180, engages knob, /, Fig. 179, 
 and holds the lid in place. The main case and lid are 
 porcelain ; there are no moving parts. The gap will pass 
 a spark to earth at 2000 volts. The spark points are on 
 the lid and are accessible for inspection, renewal or ad- 
 justment. In series with the spark gap is a non-inductive 
 carbon resistance that limits the value of the line current 
 that can follow the spark and lessens the burning 
 of the points. In order that the blow coil may offer no 
 inductive resistance to the discharge, it is in multiple with 
 a part of r, Fig. 1 8 1 , showing the connections, z I is a car- 
 bon resistance divided into two parts, r' and r; r' is in 
 multiple with blow coil, d, while r is in series with both ; 
 one end of the blow coil connects to one side of the air 
 gap, and to one end of the carbon resistance at z; the 
 other end of the coil connects to the resistance at point p. 
 the junction of r and r 1 . The trolley comes in at the top 
 and connects to one side of the gap. For simplicity, the 
 air gap connections are shown as permanent, but, in fact, 
 the air gap is not in circuit till the lid is on. In Fig. 181, / 
 is the wire leading to the fuse-box ; t' t the wire from the 
 fuse-box ; k is the kicker, and / and a, the motor field 
 
536 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 and armature ; g is the motor ground wire, and g' that of 
 the arrester. Confusion of the trolley and ground wires 
 can do no harm, as the main current does not enter the 
 device except when a discharge occurs. Ordinarily, the 
 current path is /, t'-k-f-a-g ; if a discharge occurs the 
 path is t-t'-t"-a-n-z-p-r-l-g' . The trolley current that fol- 
 lows splits at n ; part takes the path n-z r'-p, and the 
 larger part takes the path n-x-d-y to />, through the blow 
 
 coil. The lightning discharge, being alternating in char- 
 acter, avoids the blow coil on account of its self-induc- 
 tion, and takes the carbon path ; the trolley current 
 being direct, avoids the carbon on account of its high 
 resistance, and takes the blow coil path. Fig. 181 
 shows the two leads to come out at the bottom and top ; 
 when a wooden case is used (Fig. 178) both leads come 
 out at the bottom of the case. As the arrester has no 
 moving parts, it can be set up in any position. The de- 
 vice is giving satisfaction, notwithstanding the many pre- 
 
INSTALLATION CAR KQUIPMENT TESTS. 537 
 
 dictions that the porcelain parts would never stand the 
 heat of the arc. 
 
 POINTS ON ARRESTERS. All arresters, whatever their 
 make, should be inspected after each storm; even if the 
 device itself is in good order, there may be some broken 
 or burnt-off connection leading to or from it. An open 
 ground or trolley wire renders the device inoperative. 
 The main point of care on an arrester is to keep the air 
 gap thinner than any of the insulation the device is to 
 protect. Only constant inspection can insure this adjust- 
 ment, as each discharge burns the gap a little wider and 
 the jolting of the car may not help matters any. The 
 space between the points must be kept clean, and all con- 
 nections tight. An arrester in good order will always 
 protect a car from the effects of mild discharges due to 
 the static induction of charged clouds. There are some 
 bolts of lightning against which there can be no pro- 
 tection. If from static causes the line potential rises to 
 several thousand volts, an arrester in good order will pass 
 a spark to earth and relieve the tension ; but if a heavy 
 bolt of lightning, such a one as jumps a quarter of a mile 
 through air, splintering a tree or shattering the side of a 
 house, strikes a car, the arrester is by no means a cer- 
 tain protection. 
 
 The insulation on street car motors is tested to only 
 2000 volts at the factory. The gap on an arrester, then, 
 must be adjusted to pass a spark before the line reaches 
 this potential, or the device will be of no protection. The 
 insulation of motors alternately idle and active, absorbs 
 more or less moisture and is, therefore, not as high when 
 old, as when new. 
 
538 TESTING OF DYNAMOS AND MOTORS. 
 
 It would be a simple matter to make an arrester ex- 
 tremely sensitive, if the lightning were the only factor 
 to contend with ; but the normal current that follows the 
 discharge on a high- voltage circuit must be cared for 
 or it will burn the points so as to render the device use- 
 less for the next discharge, and discharges often follow 
 each other at very close intervals. Again, the two 
 spark points may weld, so that the arrester ground or 
 trolley wire must be cut before the car can move itself at 
 all. On a road several miles long, the same discharge 
 has been known to injure cars at opposite ends of the 
 line ; this accounts for the fact that cars are sometimes 
 struck, when no thunder or lightning are in evidence. 
 Certain conditions raise the potential of the whole line, 
 and the most sensitive arrester or the weakest insulation 
 passes the first discharge. 
 
 It has been found impracticable to work an arrester on 
 an air gap thinner than .025 in., which is thin enough for 
 ample protection, as none of the insulation of the motor 
 or controller ever approaches this value. 
 
 V THE CONTROLLER, (i) RHEOSTAT, RESISTANCE, OR 
 STARTING COIL. The starting coil is a resistance used 
 to limit the value of the starting current. It permits the 
 car to start smoothly and saves the motors undue strain. 
 In connection with the controller it regulates the speed 
 after the car is started. When a current passes through 
 resistance, heat is developed and energy lost; if it were not 
 for these facts, the motors might be wound to give resist- 
 ance enough to limit the starting current ; but this resist- 
 ance would be in circuit all the time and cause a constant 
 loss of energy, whereas the starting coil is cut out on cer- 
 
INSTALLATION CAR EQUIPMENT TESTS 539 
 
 tain notches. Motormen have had occasion to observe that 
 some cars run slower after having made several trips. The 
 fact is most noticeable on heavy cars equipped with old style 
 motors, and is due to the heating which raises the resist- 
 ance of the motor winding. The practice of to-day is to 
 minimize the heating and loss of engery in the motor by 
 making the resistance of the winding as low as possible. 
 In order to keep the 5OO-volt line pressure from sending 
 an abnormal current through the low resistance motors 
 at starting, the starting coil is used. It is true, the start- 
 ing coil gets very hot, and this heat represents lost en- 
 ergy ; but the coil is used only on resistance notches, and 
 does not affect the car's maximum speed. 
 
 There are three good reasons why a car should not be 
 run for any length of time "on resistance." First. It in- 
 jures the coil, as it is not designed for continuous run- 
 ning. Second. It is very uneconomical, the heat repre- 
 senting so much lost energy. Third. If one resistance 
 notch is used to the exclusion of others, one part of the 
 coil heats excessively, its resistance trebles or quadruples, 
 and causes the car to jump badly when that section is cut 
 out by the controller. Ordinary starting coils, when used 
 with the proper sized motors and cars, will about double 
 their cold resistance and hold this value in continuous 
 normal service. But when the coil is abused, either by 
 the motorman or the management, it may get so hot as to 
 set the car floor on fire. A very hot coil is certain to 
 make the car jump on some notches, while other notches 
 will not be felt at .all. 
 
 Resistance coils as they leave the factory are designed 
 so that on a car of the right weight, on a level, equipped 
 
540 TESTING OF DYNAMOS AND MOTORS. 
 
 with the motors for which the coil is intended, the car 
 will start with a jerk when the coil is cold, but will set- 
 tle down to smooth starting as the coil warms up. 
 
 The motors and controlling devices on many roads are 
 very much abused in the following manner : A lot of 
 motors of a certain size are bought, for example, to put 
 under i6-ft. horse cars that have been strengthened for 
 the purpose. The cars are light, the road level, and the 
 runs easy, with, perhaps, a lay over at both ends. In 
 course of time the traffic grows, the road is extended, the 
 time-table is revised and two small cars are spliced to- 
 gether to make one big one. Almost before it is realized, 
 the motors that handled the smaller cars so well under 
 the proper conditions, are required to tug an "eight- 
 wheeler" over 5 to 10 per cent grades; then because 
 they revolt at such abuse, are condemned, and new ones 
 bought from a different company. 
 
 In many cases, the starting coil is the first device to 
 show the effects of such abuse. As soon as the coils be- 
 gin to open circuit, short circuit, roast and fall to pieces, 
 they are replaced by coils that will "stand more current" 
 coils better suited, perhaps, to start motors of twice the 
 size with the result that the car starts with a current 
 that ought to blow the fuse. The final effect of such a 
 change is to roast the motor fields and blow up the con- 
 trollers. 
 
 When starting coils begin to give general trouble, the 
 idea that more current capacity is needed is correct, but 
 it must be carried out with the condition in mind that the 
 resistance be kept the same. , To fulfil this condition, the 
 coil must be four times as heavy, when the current ca- 
 
INSTALLATION CAR EQUIPMENT TESTS. 541 
 
 pacity is doubled, because not only is the cross-section 
 of the resistance wire doubled, but the length is 
 doubled too. Fig. 182 (a, b, c) will help to make 
 this idea clear. In sketch (a) r is a coil of given resist- 
 ance, say two ohms. It is desired to combine the least 
 number of these coils that will double the capacity of fa) 
 without changing the resistance. The resistance from .r 
 to y, (a), is two ohms: that from .r to v (b) is one ohm, 
 because there are two two-ohm paths in multiple. 
 
 HWWWWWW 
 
 Sketch (b) represents the condition where the current 
 capacity has been doubled, but the resistance halved. To 
 restore the resistance to its former value, two ohms, take 
 two units like that in sketch (b) and connect them in 
 series as shown in sketch (c). Combination (c) con- 
 tains four such coils as r, and is, therefore, four times as 
 heavy. 
 
 Fig. 183 shows a substantial form of coil very much in 
 use. Fig. 1 84 shows how any number of these coils can 
 be hung under a car. The coil is built up of band iron 
 
542 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 and mica, and up to certain reasonable limits is not in- 
 jured by heat or water. A single coil is called a barrel, 
 and the proper resistance for starting any railway motor 
 
 FIG. 183. 
 
 can be made up by combining two or more of these bar- 
 rels in series or multiple, or both. 
 
 (2) THE SHUNT AND LOOP. These devices are used 
 to increase the speed of a car. They have some virtues, 
 but their failings have condemned them to be set aside. 
 As has been shown on pages 51 and 52, weakening the 
 
 FIG. 184. 
 
 fiield of a motor increases its speed. The field can be weak- 
 ened either by shunting some of the field current or by 
 cutting out some of the field turns; either procedure lessens 
 
INSTALLATION CAR EQUIPMENT TESTS. 
 
 thfc ampere-turns. The shunt does the former, the loop the 
 latter. Fig. 185 is the wiring- diagram for a car equipped 
 with one motor, one old style rheostat and a reverse 
 switch, A'. T is the trolley; FB, the fuse-box; LA, 
 the arrester; R the resistance which is cut out gradually 
 as the shoe B moves from C to D. Arm, AB, is turned 
 by a handle in the driver's control. .-/-}-, A is the arma- 
 ture and F-\-.F , the field. is the shunt of low resist- 
 
 ance, one end of which is tapped on the field wire at O. 
 The other end goes to plate E on the rheostat. To start 
 the car B is put on C and the current takes path T-FB- 
 A-B-C-R-D-F+-F O-A+-A to the ground at G. As 
 the shoe B revolves toward plate D, R is gradually cut 
 out, until, when B touches D, all of A is cut out and the 
 current path is T-FB-A-B-D-F+-F O-A+-A G. 
 The motor is directly across the line but has a full field, 
 it B is advanced until it touches both D and , the shunt 
 
544 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 becomes connected to the motor field at both ends and is in 
 multiple with it, so that the current leaving shoe B, 
 
 reaches by way of two paths : B-D-F-\--F O through 
 
 the field, and B-E-S-O through the shunt. Plates D and 
 E are insulated from each other so that the shunt is idle 
 until B touches both. Were D and E connected the shunt 
 would be active throughout from C to , greatly impair- 
 ing the starting power of the car. An open circuit in the 
 
 F- 
 
 shunt or its wire simply decreases the maximum speed 
 of the car. 
 
 Fig. 1 86 is the same as 185, except that a loop replaces 
 the shunt. When B touches D, the current path is T-FB- 
 
 A-B-D-F+-F A+-A G. When B touches both D 
 
 and E, the current reaches O by two paths : B-D-N-F-\--O 
 and B-E-L-O ; the latter is a short circuit so that the 
 lower part of the field is cut out. In shunt control, then, 
 a part of the current goes through all of the field ; in loop 
 control, all of the current goes through a part of the field. 
 
INSTALLATION CAR EQUIPMENT TESTS. 545 
 
 An open circuit in cither the shunt wire or L>op wire 
 does not affect the starting of the car. An open circuit 
 in the D-F -\- field wire, Fig. 185, will make it im- 
 possible to start the car at all, and will burn the shunt out 
 if B is put on E and allowed to stay there a few seconds. 
 In Fig 1 86, an open circuit between O and 7 ; kills the 
 car ; with an open circuit between C and D, the car is 
 dead until B touches /:, when the car starts with a jump. 
 The more the shunt is used, the cooler the fields run, for 
 the shunt relieves them of current. When the loop is used 
 continuously, one part of the field cools off and the other 
 parts gets hotter, because one part is cut out and the other 
 part gets more current than ever. Shunts and loops have 
 been set aside, because from a practical standpoint their 
 troubles outweigh their advantages. 
 
 (3) CONTROLLER PROPER. Figs. 185 and 186 give an 
 idea of the old rheostat formerly in general use; with this 
 device, the motors were permanently connected, either in 
 series or multiple. Both of these methods have given way to 
 series-parallel control. In series-parallel control, the 
 motors are started in series and after the car is up to half 
 speed the motors are thrown into multiple. This method 
 requires less current to start a car, because, the motors 
 being in series, each gets the total main current. When 
 the two motors are started in multiple, each motor gets 
 but half the total current, so that in order for each to get 
 a certain current, the total current must be twice as great 
 as when the motors are in series. This means that the 
 line losses are four times as great. If the station voltage 
 is 500 and the line drop 25 volts when starting a car at a dis- 
 tant point with the motors in series ; to start it with the same 
 
546 TESTING OF DYNAMOS AND MOTORS. 
 
 impulse, but with the motors in multiple, calls for twice 
 as great a line current, which causes twice as great a drop, 
 or 50 volts. Since the current and line drop have both 
 been doubled, the power lost in the line has been -quad- 
 rupled ; for, call E the volts lost ; C, the current ; W, the 
 watts lost. In the first case, W^E^C. In the second 
 case, lV z ~2Ey^2C=4EC^W^ When the two motors 
 are in series, their C.E.M.F's are in series, and hence the 
 spurious or non-heat generating resistance is much great- 
 er than when the motors are in multiple. This means, 
 practically speaking, greater economy at the medium 
 speeds, because the C. E. M. F's take the place of 
 the starting coil to a greater extent, and diminish the heat 
 losses in that coil. Of course, a small starting coil is re- 
 quired to start, because, until the car begins to move, the 
 C.E.M.F. is zero. With no coil at all, the car would 
 start with a jerk on good rails and the wheels might slip 
 so badly as to prevent its starting at all on bad rails. A 
 well designed and well handled series-parallel controller 
 effects an economy in starting and on the low and medium 
 speeds. To get the full benefit of the control, though, it 
 is necessary to let each controller notch have its full effect 
 before moving the handle to the next one ; also to allow 
 the motors to attain their full series speed before throw- 
 ing them into multiple. A series-parallel controller can 
 be just as much abused as any other kind of controller, 
 and when it is, is apt to give less economical service than 
 some of the older types. 
 
 Fig. 187 is a general view of the General Electric Com- 
 pany's type K, magnetic blowout, series-parallel con- 
 troller. Fig. 1 88 gives the internal controller wiring; 
 
INSTALLATION CAR EQUIPMENT TKSTS. 
 
 547 
 
 Fig. '189 is a diagrammatic sketch of the connections to 
 the motors. J> is the controller drum, made up of five 
 separate castings that are insulated from each other. The 
 
 castings have tips that engage the fingers successively as 
 the drum is revolved. Tips a lt a*, etc., belong to one 
 casting, and tips t^ t ^ 2 , etc., to another and so on. The 
 tips are so marked as to indicate the position in which 
 
54 8 TESTING OF DYNAMOS AND MOTORS. 
 
 CONN-BOARD 
 
 nan 
 
 i 
 
 i- c: cr 
 
 
 L 
 
 n 
 
 HID 
 
 1 
 
 
 1-1 
 
 
 i i- 1 1 
 
 T- (N 
 
 Ul O _. 
 
 ^ 
 p 
 
 
 F" 
 
 UJ 
 
 as 
 
 o> 
 
 1 
 
 1 
 
 UJ 
 
 :nig_aj 
 
 
 
 
 
 a 
 
 5- 
 
 LJ^nn 
 
 a 
 
INSTALLATION CAR EQUIPMENT TESTS. 
 
 549 
 
 they come into action; thus, tips </, and <-, make contact 
 on the first position; tips </, on the tenth or last position. 
 
 MOTOR NO. 1 MOTOR NO. 2 
 
 FIG. 189. 
 
 B is the reverse drum operated by handle Z, Fig. 187. A 
 car must be so connected that the inclination of L indi- 
 cates the direction of motion of the car. K is the niDtor 
 
550 TESTING OF DYNAMOS AND MOTORS. 
 
 cut-out switch. The three blocks to the left are for 
 motor No. i ; the two to the right for motor No. 2. When 
 both switches are down as far as they will go, as in Fig. 
 187 both motors are cut in. To cut a motor out, its 
 switch is thrown up. M is the blow-out coil ; the current 
 from the fuse-box enters this coil first, so that the whole 
 device is protected from arcs. In Fig. 189 FF' and A A' , 
 are the fields and armatures, respectively, of the Nos. i 
 and 2 motors ; / is the No. i, and /' the No. 2, loop parts 
 of the fields. R is the two-part starting coil. The top 
 row of numbers indicates the positions, the bottom row, 
 the notches of the drum. The blocks to the left of the 
 drums represent the fingers. 
 
 In Fig. 189, suppose that the reverse lever points ahead 
 and that tips, /, x, y and 2, therefore, engage the re- 
 verse fingers. When the power drum, D, is put on the 
 first notch, the current path is, T-M-T-a^-a^-R^-r^-r^- 
 7-3-19-19-19-2-^ -A-AA^ -y-F^-F-l-E^ -E^ -E^-E^-c^- 
 1-15-15- 15- i5-x-A t -A'-AA ir w-f l s-ir s -f t t -J7'-t'-G. 
 The blow coil, all of the starting coil and the two motors 
 are in series. On the second notch finger, R% touches tip 
 a% and short circuits the part of the starting coil that lies 
 between r 1 and r 2 . The current path is the same as be- 
 fore, excepting that R\, r are cut out. On the third 
 notch finger, Jt z touches tip a 3 , and cuts out the rest of 
 the resistance. The letters ^ 3 , r z can now be left out. 
 The blow coil and two motors are in series across the line. 
 On the fourth notch, the three c and e tips fall under 
 fingers L^ , Z 2 and G. The current path is now : T-M- T- 
 
INSTALLATION CAR EQUIPMENT TESTS. 551 
 
 L^-e^-e^-G-G-G. The loop wires tap on to the field coils 
 of the two motors at O and O', when the current gets to 
 these points it takes the shorter path through the loop 
 wires, cutting out the lower part of both fields. The 
 motors are in series with the blow coil, but each motor has 
 a part of its field cut out. The fifth position is not a 
 notch, and the current path is the same as on the second 
 notch. The blow coil, half the starting coil, and the two 
 motors are in series. Resistance is cut in again to les- 
 sen the current before passing to parallel. On the sixth 
 position, the two </ 6 tips touch the E^ and G fingers; the 
 current path is, T-M-T-a^-a^-R^-r^-r^-i^-i^-i^-z-A ,- 
 A- A A \-y-P \-F-l-E ^-E \-E \-d \-J 6 -G. In passing from 
 the fifth to the sixth position, motor Xo. 2 is short cir- 
 cuited by fingers E x touching tips r, and </ 6 at the same 
 time, the current grounding through finger G and lower 
 tip '/ 6 , so that on the sixth position motor No. 2 is dropped 
 out of the circuit entirely. The blow coil, half the starting 
 coil, and motor No. i are in series. The seventh position 
 is the same as the sixth. On the fifth notch (eighth po- 
 sition) tips 8 fall under fingers 19 and 15, and the cur- 
 rent path becomes : T-M-T-a^a^R^-r^r^ : ' ^" r9 " T 9" g * 
 
 ,-,-,-t-t-. 
 
 8 - 1 5- 1 5- 1 5- 1 5-*-^ 2 -A'-AA 2 - w-F % -F z -F* -F'-l'- G 
 r s the current splits; part takes the path r 3 - 19- 19- 19-3 
 through motor Xo. i, and part, path *V-# 3 -i9- 8 - 8 , etc., 
 through motor No. 2. The blow coil and half the resist- 
 ance are in series with the two motors which are in mul- 
 tiple with ^ach other. On notch six (position nine), -ff s 
 
55 2 TESTING OF DYNAMOS AND MOTORS. 
 
 touches tip a 9 and cuts out the starting coil; leaving out 
 It-g, r z , the path of the current is the same as above. On 
 the seventh and last notch (position ten), tips d l touch 
 fingers Z t and Z 2 , and cut the loop wires into action again, 
 cutting out the lower parts of both motor fields. With this 
 exception, the current path is the same as on the sixth 
 notch. All controllers that use a shunt or loop have two 
 running notches in series and two in multiple. In this case, 
 the running notches are three, four, six and seven. Three 
 is used for climbing grades; four, for moderate speed on a 
 level; six and seven, for high speeds. Four and seven 
 should never be used for climbing grades. When one motor 
 is cut out the controller is operated on the first four notches. 
 The throwing of either cut-out switch so disposes an in- 
 terference that the controller drum can not be moved past 
 the fourth notch, the controller and reverse drums 
 interlocking. The controller drum can not be operated 
 while the reverse drum is at the "off" position, and the 
 reverse drum can not be operated while the controller 
 drum is on. This feature makes it impossible for the 
 motorman to reverse the motors without first throwing 
 off the power, and it also gives him control of the car 
 when he may not be on it; because the reverse handle 
 can be lifted only at the "off" position, so that, with the 
 reverse handle in his possession, the power drum can not 
 be turned by any one who does not understand the nature 
 of the locking device. 
 
 (4) THE MOTORS. Figs. 190 and 191 show a recent 
 type of motor made by the Lorain Steel Company. Fig. 
 
 190 shows it complete, ready to go on the axle, and Fig. 
 
 191 shows the case open, exposing the armature. The 
 
INSTALLATION CAR EQUIPMENT TESTS. 
 
 553 
 
 case is cast in two halves linked together, as shown at ////, 
 Fig. 191. Of low carbon steel, it is stronger than cast 
 iron and magnetically nearly as good as wrought iron. 
 In order that the armature or a field coil may be taken 
 out from the pit without disturbing any other part, all up- 
 to-date motors are so made that the removal of a couple 
 of bolts on each end will permit the lower casing to swing 
 
 down, either with or without the armature. Yokes, YY, 
 Fig. 191, have a split key on one end and a lock nut on 
 the other ; one shank passes through a lug in the upper, 
 and the other shank through a lug on the lower half of the 
 case. To swing the armature with the lower case, the 
 split keys are removed ; to retain it in the upper case, the 
 keys are left in and nuts KK* ', Fig. 190, removed, DD, Fig. 
 191, are oil rings that turn with the armature and throw 
 all grease and oil down on the track through the aper- 
 tures, one side of one of which is seen at O. Motors on 
 
554 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 which this construction is used are said to have "out- 
 board" bearings. The grease cups are over the bearings 
 and, are cast with the top case. The bearings are malle- 
 able iron sleeves lined with babbitt and provided with 
 
 FIG. 191. 
 
 grooves and grease holes; they are centered on guide 
 pins that keep them from turning with the axle. 
 
 Fig. 192 shows the armature; it is built up of thin 
 sheets of soft iron, pressed together and held by steel head- 
 plates. The teeth are punched in the plates before as- 
 sembling, and immersion in japan insulates the plates 
 from each other. The slots are big enough to take three 
 coils, so that there are but one-third as many slots as coils. 
 
INSTALLATION CAR EQUIPMENT TESTS. 555 
 
 This lessens the cost of the armature and strengthens the 
 teeth. The eoils are machine wound ; three of them are 
 bound together and encased in an insulating armor, which 
 is pressed to shape in a press through which steam cir- 
 culates ; thirty-three of these armored coils complete an 
 armature. 
 
 In the connecting of the armature shown, no solder is 
 used ; the commutator slots are milled a trifle too small 
 for the leads; the slots and leads are tinned; the leads 
 
 are then driven into the slots and secured by a strong 
 band, shown at 5", Fig. 192. On the rear of the coils, is 
 a metal shield, F. The commutator and pinion are on a 
 tapered seat. The commutator is of unusually liberal di- 
 mensions. 
 
 Fig- J 93 shows the field coil; there are four of these 
 to slip over as many laminated pole pieces ; the pole pieces 
 are detachable, and, when drawn home, their flanges bear 
 against guide ribs, GG, and hold the coil firmly in place. 
 The field is wound on a double flanged shell, which ren- 
 ders the winding immune from direct pressure of any kind, 
 and makes it impossible for the pole piece to cut into the 
 
556 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 corner of the coil or for the shrinkage of the insulation 
 to loosen it. This manner of securing is positive. O and 
 / are the connecting lugs ; the ends of the coil winding 
 connect to these lugs, also the wires used to connect the 
 coils to each other and to the car hose. The ridge L is 
 a pocket cast in the shell, to allow the inside end of the 
 winding to come to the surface. This end must pass 
 
 every layer of wire in 
 the coil, and is extra 
 well insulated to insure 
 safety. The end is 
 usually brought out in 
 the form of a terminal 
 made of sheet copper; 
 this flat terminal is 
 soldered to the end of 
 FIG. 193. the wire and insulated 
 
 before the coil is begun. 
 
 The coil shown, as well as all recent types of coil, are 
 rounded off on one side to fit the inside lines of the motor, 
 so that there is little chance of getting the coil into the 
 motor upside down ; but on most all coils of this type, it 
 is easy enough to get a coil in end for end, and it is very 
 often done. The effect of such a mistake is to bring next 
 to each other in the motor two leads that should not con- 
 nect together. The first symptom of such a mistake is 
 that the brushes spark badly on the series notches, and 
 when the controller is thrown over to multiple, the main 
 motor fuse blows. If the confusion in connecting is not 
 bad enough to continually blow the fuse, the final effect 
 is to bake the insulation on the field wire. It is possible 
 
INSTALLATION CAR EQUIPMENT TESTS. 557 
 
 for the field coils to be so badly confused as to render the 
 motor unable to start at all alone, but if there are two 
 motors on the car, the fault may not be suspected until 
 the motors are thrown into multiple and the fuse blows. 
 It is a conservative estimate to say that one-fifth of the 
 baked fields, blown fuses, grounded brush holders and 
 knockcd-out controllers can trace their failure directly 
 or indirectly to wrongly connected field coils. Coils 
 should be made and the leads brought out so that it would 
 be impossible to get one in upside down or end for end. 
 If the motor makers would pay half the attention to this 
 detail that they do to less important ones, they would soon 
 begin to reap the product of their sowing. 
 
CHAPTER XVI. 
 
 CAR EQUIPMENT TESTS. 
 
 Having acquired some familiarity with the devices 
 which go on a car, the next step is to give them a prelim- 
 inary test before installing them. The testing of the trol- 
 ley stand, the hood switch or breaker, the fuse-box, and 
 the lightning arrester have been considered, so that it is 
 in order to take up the testing of the resistance coils, the 
 shunt, the controllers and the motors. 
 
 THE TEST CIRCUIT. The test circuit available around 
 a car house usually consists of a magneto bell or lamp 
 circuit. The lamp circuit will be selected in this case, 
 because its indications are more positive and it can 
 be operated by one man. The lamp circuit, as a rule, 
 consists of a single row of five lamps in series, but it is 
 much more satisfactory to use two five-lamp rows in mul- 
 tiple, as shown in Fig. 194. A better flash is gotten on 
 low voltage, and if one row gives out another remains 
 to test with. This lessens the liability of false indications, 
 due to the tester not knowing that the circuit is open. 
 The test circuit shown on Fig. 194 is made as follows: 
 take two boards ^ in. X 7 ins. X 13 ins. finished off, 
 groove one, as shown in view (c), to bury the connecting 
 wires shown by dotted lines in view (a) ; on the opposite 
 side from the grooves mount ten keyless lamp sockets, as 
 
 558 
 
CAR EQUIPMENT TESTS. 
 
 559 
 
 indicated by the dotted circles in (c); connect these all 
 in scries and bring the two opposite ends to posts T and 
 T' as shown in the figure. The second piece of board is 
 now screwed to the back of the grooved one to protect the 
 the connecting wires. The whole is given a coat of thick 
 
 - T 
 
 (a) 
 
 ?Q L -CK>O L -Q 
 
 (I) 
 
 Ma 
 
 FIG, 194. 
 
 shellac, such as comes out of the bottom of a shellac pot. 
 After this is dry, the device is ready for the test lines. 
 The test lines are best made of silk covered, rubber cored, 
 flexible lamp cord ; each line is a complete cord, the wires 
 being skinned on the end, twisted together and dipped 
 in solder. Both ends of both lines have a terminal. One 
 terminal, Fig. 194 ( d), goes to the post T or 7 V . It is 
 made of hard spring copper, so that when the milled head 
 
560 TESTING OF DYNAMOS AND MOTORS. 
 
 is screwed down, it will hold the terminal securely. The 
 terminal for the test ends of the lines are points well in- 
 sulated, except at the ends, so that the tester may not be 
 so liable to a shock. The point on one line is straight 
 and on the other is both straight and crooked, to hang 
 over the trolley wire, or to stick in a lamp plug switch. 
 Fig. 195 shows a good form of test point. T is a brass 
 rod, pointed at one and drilled at the other, to re- 
 ceive the test line which is soldered in place. ^ is a hard 
 rubber or red fiber tube into which rod T is forced snug- 
 ly. This sleeve extends well back over the test line in- 
 
 sulation, so that the working up and down will not break 
 the wire off. 
 
 TESTING THE STARTING COIL. ^. new starting coil 
 must be tested for insulation and for continuity. In 
 Fig. 183 the coil can be tested for insulation as follows: 
 Lay the coil down so that one of the end nuts, c, touches 
 the rail ; then, with the test lamps connected, as in Fig. 
 196, touch any of the lugs, L, L, L, with the test point, T; 
 if the lamps light, it shows that the resistance is grounded 
 to the rod on which it is built up. If the lamps do not 
 light, touch the test point to the rail, to insure that the 
 test circuit is all right. If the circuit is all right, the in- 
 sulation is O. K. To test for continuity, that is, to see 
 that the coil has no open circuit in it, lay the coil down so 
 
CAR EQUIPMENT TESTS. 5 61 
 
 that one of the end plates rests on the rail, and touch the 
 other end plate with the test line. If the lamps light the 
 coil is O. K. If they da not, an open circuit exists. To 
 find out in which end of the coil the open circuit is, lay 
 the coil down so that the middle lug L rests on the rail, 
 and touch both end plates with the test line. The open 
 circuit lies nearest the end plate that fails to complete the 
 
 TROLLEY LINE 
 
 TEST LAMPS 
 
 TC^O., T | 
 FIG. 196. 
 
 circuit, and that, therefore, fails to light the lamps. 
 Knowing in which end of the coil the break lies, a great 
 deal of labor is saved in unbuilding the coil to repair it. 
 
 TESTING THE CONTROLLER. The controller must be 
 tested for grounds, open circuits, and short circuits. 
 Besides this electrical test, it must be seen to that the in- 
 terlocking device and motor cut-out switches are in 
 working order, and that all fingers make good contact, 
 and are in alignment. To test the controller for grounds, 
 lay it on its back, so that the iron frame rests on the rail ; 
 see that both drums are at the off position, that both cut- 
 out switches are down as far as they will go, and that 
 
562 TESTING OF DYNAMOS AND MOTORS. 
 
 the ground wire that runs from G on the right-hand cut- 
 out, Fig. 1 89, to a connecting lug on the bottom bearing of 
 the controller drum, is disconnected. (This ground wire 
 is used to prevent the controller frame from ever getting 
 charged). The test line is then touched to all parts of 
 the controller, one at a time : The power fingers, reverse 
 fingers, cut-out switches, posts on the connecting board, 
 all the castings on both drums, and the blow coil. If 
 the lamps fail to light under all these tests, the insulation 
 to base is O. K. To test the parts for cross and open 
 circuits ; the test lines are disposed, as shown in Fig. 197. 
 Here there is an additional test line, T l run from the 
 
 TROLLEY 
 
 RAIL 
 
 TEST LINE J * T| -TEST LINE 
 FIG. 197. 
 
 rail. If these two test lines are touched together, the 
 lamps light. To use them for testing the controller, one 
 test line is held in each hand, and the test conducted as 
 follows: (See Fig. 188). Holding one test line on T 
 finger and the other on T post on the connecting board, 
 if the lamps light, it shows the connections from T to T, 
 through the blow coil, to be all right. Next place T on 
 finger R^ and 7\ on connecting post R^. The lamps 
 should light. Fingers R% and 19 should light up with 
 cut-out blocks 19, reverse finger 19, and connecting post 
 
CAR EQUIPMENT TESTS. 563 
 
 ^? 3 . Power fingers, -", , should light up with cut-out block, 
 El. The four armature connecting posts light up with 
 the reverse fingers of the same mark.... and so on. 
 All drum tips of the same letter (Fig. 189) should light 
 the lamps; hut where the test lines rest on tips of dif- 
 ferent letters, the lamps should not light. 
 
 The parts of the controller are tested for short circuit 
 by holding a test line on one part, and with the other 
 test line touching every other part ; the lamps should in 
 no case light up between two parts that the drawing does 
 not show to be connected. A man who is perfectly 
 familiar with the controller he is testing, can test it in a 
 very short time. 
 
 MOTOR TESTS. The first test to make on the mo- 
 tors is the insulation test ; this is done by setting the 
 motor on the rail ; this grounds the frame. L'sing the 
 connections of Fig. 196, touch the test line to all of the 
 motor leads, one after the other. If one of the field 
 leads, say, lights the test lamps, it indicates one of the field 
 coils to be grounded ; before drawing any final conclu- 
 sion, see that the opposite end of the field whose lead is 
 being touched is not resting against the motor frame ; if 
 the field circuit proves to be grounded, disconnect the 
 back field connection, thereby separating the two top 
 fields from the two bottom ones, and by testing them 
 alone, determine in which half of the motor the fault lies. 
 Having determined this, open the motor and separate the 
 two field coils in the faulty half; test these two coils one 
 at a time, and take out the grounded one. Very often, 
 upon opening the motor, the ground will be found to be 
 due to a lead being pinched between the two halves of 
 
564 TESTING OF DYNAMOS AND MOTORS. 
 
 the case. If this is the case, it will not, as a rule, be 
 necessary to remove a coil ; this pinched lead can be re- 
 leased, spliced if neccessary, taped up and shellaced. If 
 one of the armature leads shows a ground, draw the 
 brushes out of the holders and test both leads and the 
 armature. If the drawing of the brushes removes the 
 ground, the armature winding itself or the commutator 
 must be the faulty member. If the drawing of the 
 brushes does not remove the ground, the fault must be 
 in one of the brush holders. This is not certain, though, 
 for the brush holders may have dropped down far enough 
 to touch the commutators themselves, in which case the 
 drawing of the brushes does not separate them from the 
 commutator. If the fault proves to be in the armature 
 itself, it must be taken out, the leads lifted, and each coil 
 tested for a ground. The motors are always tested be- 
 fore they leave the factory, but in spite of every precau- 
 tion, a grounded one will get through occasionally, and it 
 is very unfortunate if it finds its way into a trial equip- 
 ment. In the test and on a car, one field or brush holder 
 on one motor is always permanently grounded; if the 
 permanently grounded member happens to be defective, 
 it is very apt to get out on the road in that condition. 
 
 If the motor fails to show up a ground, it is next tested 
 for an open circuit, as follows: First take a field lead, 
 lay it on the rail or motor frame, and lay a weight on top 
 of it to keep it grounded; then touch the other end of 
 the field with the test line. If the lamps light, the field 
 circuit is O. K. Next put the brushes back in the brush 
 holders and test the armature in the same way, seeing to 
 it that both brushes make contact with the commutator.. 
 
CAR EQUIPMENT TESTS. 
 
 565 
 
 Sometimes shellac on the commutator, or a piece of 
 foreign matter under the brush, will keep the lamps from 
 lighting, and thereby give the symptom of a more serious 
 break. 
 
 TESTING THE MOTOR POLARITY. Being assured that 
 the motors are neither grounded nor open circuited, 
 the next step is to get the polarity tested to insure that 
 
 PINION END 
 
 PINION END 
 
 the two motors shall urge the car in the same direction 
 after they are mounted on the truck. The motors when 
 swung on the axle are back to back, so that their arma- 
 tures must turn in opposite directions, in order to urge 
 the car the same way. To get the polarity right, proceed 
 as follows : Set the two motors on the floor back to back ; 
 
566 TESTING OF DYNAMOS AND MOTORS. 
 
 their commutators will then point in opposite directions, 
 as shown in Fig. 198. Find the field and armature leads 
 on the two motors. To be certain that the leads have not 
 been confused in bringing them out of the motor frame, 
 run the hand inside the motor and trace them from their 
 starting point. Mark the field leads with a piece of chalk 
 and leave the armature leads unmarked. Assume a given 
 direction for the car and, therefore, for the motors. In 
 Fig. 189, it will be seen that the field leads on the No. I 
 motor are marked JF^ and E^ , and on No. 2 motor, F% 
 and E z or G. The No. i armature leads are marked A ^ 
 and A A 1 , and the No. 2 armature leads, A 2 andAA 2 . 
 The E z lead on the No. 2 motor is grounded, so that on 
 both motors the fields are next to the ground. The fields 
 are supposed to be so connected that the current enters at 
 their F ends and leaves at their E ends. The current en- 
 ters the armatures at their single letter ends and leave at 
 the double letter ends. Take either of the motors as it 
 sits on the ground and connect one field and armature to- 
 gether ; ground the remaining field to the rail, and by 
 means of a wire run through a starting coil or other re- 
 sistance, touch the remaining armature lead and give the 
 armature a spin. If it turns in the right direction, mark 
 the armature lead where the current enters, Ai, the other 
 armature lead, A A i; the grounded field lead, El ; the 
 other field lead, Fi. Should the armature turn in the 
 wrong direction, let the grounded field and its mate ex- 
 change places, and mark them accordingly. The No. 2 
 motor is tested and marked in the same way. Now the 
 taps out of the car wiring hose have the same marks on 
 them, and like marks connect together. The final result 
 
CAR EQUIPMENT TESTS. 567 
 
 of the test is that the current enters the two motor ar- 
 mature leads at like ends of the motor ; but it enters the 
 front field lead on the No. I motor, and the rear field 
 lead on the No. 2 motor. This polarity test is often de- 
 ferred until the two motors have been hung on the car 
 axles. It is just as well one time as another. It is nec- 
 essary to mark the leads for only one equipment, by test, 
 when the leads on all the others can be marked similarly. 
 A man familiar with the motors, controllers and car wir- 
 ing of a given equipment, does not need any marks at all. 
 
 The motors being electrically all right, are ready to hang 
 on the axle. At the same time that the motors are being 
 mounted, the carpenter is setting up the controllers, in- 
 stalling the trolley stand, overhead switches, etc., so 
 .that all branches of the work are ready at about the same 
 time, and the car can be set down on the truck. The car 
 wiring hose is, of course, made up and installed before 
 the car body is let down. The truck boxes, and motor 
 axle, and armature boxes are oiled or packed before the 
 truck is moved. The most particular part of the truck 
 work is the setting of the gear ; the gear is to the truck 
 what the commutator is to the motor. 
 
 SETTING THE GEAR. To put on a gear properly 
 is more of a job than some people trunk it is. It will be 
 assumed that all the gear teeth are the same size, and that 
 the two halves of the gear are mates ; that is, that they 
 have been bored together, and that the hole is in the center. 
 The first thing to do is to inspect the gear, to see that it 
 is of the size and bore wanted ; that it has no irregular 
 teeth ; that the teeth have not been hammered on the end, 
 making a ridge to bend the armature shaft ; that it has no 
 
568 TESTING OF DYNAMOS AND MOTORS. 
 
 cracks visible to the eye, or audible when the gear is struck 
 with a hammer. If there are any ridges, they must be 
 chipped and dressed with a file. The next step is to lay 
 the gear down and take the bolts out ; never use a monkey 
 wrench on gear nuts ; use a socket wrench or an S wrench. 
 When the gear comes apart, sometimes, though not al- 
 ways, four little strips of sheet iron, called shims or liners, 
 fall out. The gears are cut with these shims in between 
 the two halves, with the result that both halves are a lit- 
 tle less then half a circle, so that when the gear is put on 
 an axle that is slightly too small, it will pull up tight. 
 The position of the keyway for the key that secures the 
 gear must be selected with due regard for the dimensions 
 of the motor, or the result may be to have one end of the 
 motor interfere with one of the car wheels. The key. 
 should be fit so snugly in the keyway that, when struck 
 with the hammer, the sound and feel should be the same as 
 if the axle were struck. One may then be certain that the 
 key is in solid and not apt to get loose. Before inserting 
 the key, all burrs should be smoothed off with a file, and 
 the key tried in both the axle and gear keyways. The key 
 snugly in place, the keyway half of the gear is put on. 
 If the gear fits the axle as it should, it should require 
 some pounding to force it into place. Never pound a 
 gear with a hammer without laying a piece of copper or 
 wood on top of it. If pounding fails to force the gear on 
 it may be bored too small, or the axle may be a trifle 
 large. In such a case, ascertain which is at fault and 
 remedy it, with the end in view of keeping everything to 
 standard size. If the gear fits the axle exactly, it is best 
 to put a piece of thin paper in between the two and force 
 
CAR EQUIPMENT TESTS. 569 
 
 the gear on over it ; the gear is then much less apt to get 
 loose. In some cases, where a new gear is put on an 
 old axle, it sometimes goes on so loosely that it is ne- 
 cessary to put in a liner of iron or emery cloth. Where 
 a gear is bored too small, it can be rebored, but where it 
 is too large, if it must be used, it must be shimmed. It 
 is even the practice to put 4 in. gears on 3f in. axles ; but 
 in such cases, the lining must be done with great care, and 
 the key must be made deeper to allow for the thickness of 
 the liner. When a liner or shim gets to be 3-16 in. thick, 
 it is called a bushing, and the best way to make such a 
 bushing is as follows : Take a wrought-iron ring of the 
 width of the gear, and of such thickness that the outside 
 can be turned to fit the bore of the gear, and the inside to 
 fit the axle. The ring is put in a lathe and finished inside, 
 so it can be driven on a mandrel and finished outside. A 
 piece is then sawed out the width of the key, and the ring 
 split to go on the axle. A gear bushed in this way will 
 give no trouble. 
 
 If the upper half is a good fit, the lower half is very 
 apt to be. The upper half is the more particular of the 
 two, because it has the key seat in it. As soon as the 
 upper half is on, the axle is turned, the lower half put on 
 and the bolts put in. On all modern gears there are lugs 
 cast to keep the bolt heads from turning when tightening 
 up the nuts. Lock washers are put under the nuts, so that 
 they can not work off and get between the gears and bend 
 the armature shaft or break the motor frame. In tighten- 
 ing up the bolts in a gear the four nearest the axle should 
 be tightened up first. When a gear is on right, the face 
 of the hub should square up with the axle. The better 
 the job done on a gear, the longer it will last. 
 
570 TESTING OF DYNAMOS AND MOTORS. 
 
 MOUNTING THE MOTORS. The gears on, it is in 
 order to hang the motors. Before doing so, pour a little 
 oil in the armature boxes and give the armature a spin, 
 to -see that the armature has end play, that none of the 
 bearings bind, and that the core does not rub the pole 
 pieces. The motor must be so placed that when one axle 
 bearing is snug against the gear and the other against 
 the axle collar, the suspension lug will fall in line with 
 its support, so that it will not be necessary to bar the 
 motor over and put a wearing strain on any part of it. 
 
 TESTING THE GEARS. To test the gears, jack the 
 truck up on one end until the wheels on that end are off 
 the rail ; also block up the motor on that end, to take its 
 weight off the axle, because the car journal brass is on top 
 of the axle, and as soon as the truck frame is raised, the 
 weight of the motor and wheels draws the axle down on 
 the bottom of the journal box, where there is no bearing, 
 and where the axle is apt to be badly cut. Next connect 
 up the motors as in the test for polarity, and, if possible, 
 insert a water rheostat, so that the speed can be regulated. 
 The water resistance is then adjusted so that the car 
 wheels make about 150 r. p. m. Any fault in the gear- 
 ing will manifest itself in any of several ways.- The gears 
 are first given a spin without the gear case on. If the 
 gear emits a grinding noise once per revolution of the 
 car wheel it indicates that the car axle is sprung ; or 
 that the gear is bored out of center; or that the gear is 
 put on lop-sided. If the axle is sprung, the whole motor 
 will be seen to give back and forth, and up and down, 
 every time the axle revolves, and the side frames of the 
 truck will also be seen to work up and down. If the gear 
 
CAR EQUIPMENT TESTS. 571 
 
 is bored out of center, the fact is most readily detected In- 
 putting a little stiff grease on the gears and noting how it 
 wears ; on the high side of the gear the grease will work 
 further down into the trough of the teeth than on the low 
 side ; also, one familiar with the work is able to recognize 
 a difference in the sound when the high side passes under 
 the pinion. If the gear is on lop-sided, it will be seen to 
 wobble when looked at from the rear. If the grinding 
 noise occurs several times per revolution of the car wheel, 
 it indicates an imperfection in the pinion, or a bent ar- 
 mature shaft. If the shaft is bent, the motor brushes 
 will chatter. 
 
 A clicking or knocking noise once per revolution of the 
 gear, indicates that the gear has an odd sized tooth in 
 it (due to the machine or to the man that cut it) ; or that 
 the two halves of the gear are open on one side ; or the 
 gear may be loose on the axle; or the key may be loose 
 or worn. If the gear has one or more big teeth in it, their 
 wear will show it, and the trouble can be relieved by 
 dressing the faulty teeth with a chisel and file. If the 
 gear has several small teeth, the best thing to do is to 
 change the gear. If one side of the gear is open on one 
 side, close it, if possible, by drawing up on the bolts ; if 
 not, take the gear off and set it again. If the gear or key 
 is loose, the chances are that both key and keyway are 
 badly worn, and the best thing to do is cut a keyway on 
 the other end of the axle, in which case the axle must be 
 turned end for end, to bring the gear on the right side. 
 A loose gear can be detected by pinching the gear back 
 and forth with a bar. 
 
 If the gears run smoothly with the cases off, put the 
 
572 TESTING OF DYNAMOS AND MOTORS. 
 
 cases on and try them again. Any grinding noise will be 
 due to the gear or pinion rubbing the gear case. Anyone 
 familiar with the noises can tell whether it is the gear or 
 pinion, by the pitch of the noise. To minimize the 
 chances of the gear case rubbing, all burrs should be 
 cleared off inside of both ends before the gear case is put 
 on. The gear case should be put on centrally, or the side 
 play in the armature will let the pinion over against it, 
 and, in course of a short while, the pinion will mill a hole 
 in it. 
 
 STARTING THK CAR. The controlling devices in- 
 stalled, the wiring hose in place, and the motors mounted, 
 the truck is run under the car and all connections made. 
 If everything is all right, the car is ready to start as soon 
 as the pole is on, the overhead switches closed, and the 
 controller put on the first notch. It is a wise precaution 
 when installing a car for the first time to insert a 
 resistance in the circuit somewhere, so that in case a 
 ground or other wrong connection exists at the time of 
 trying to start, there will be no violent demonstration. 
 As good a place as any to insert this resistance is on top 
 of the car, between the roof wire and the trolley stand. 
 Take a starting coil and lay it on a sheet of ^-in. asbestos, 
 so that in case it gets hot, the car roof will not be dam- 
 aged. Everything ready, one controller is put on the first 
 notch (the pole being on and the hood switches in). If 
 the car fails to start, it may be due to absence of power on 
 the line, to an open circuit, or to a ground or wrong con- 
 nection. If the controller shows no flash when thrown 
 off, the trouble is due to lack of power or to an open cir- 
 cuit. If the power is off the line, the car lamps will fail 
 
CAR EQUIPMENT TESTS. 
 
 573 
 
574 TESTING OF DYNAMOS AND MOTORS. 
 
 to burn, and so will the test lamps. The car lamps can 
 not be relied on, alone, for the lamp circuit may be de- 
 fective. If the lamps burn, the trouble is due to an open 
 circuit. Try the controller on the other end of the car; 
 if the car still fails to start there also, cut in one motor at a 
 time and try both motors; if the car still fails to start, the 
 open circuit must be in some part of the circuit common 
 to both motors, anywhere from L to S, Fig. 199, or in the 
 ground wire, and may be due to any of the following 
 causes : The trolley wheel may be resting on a line 
 breaker; an overhead switch may be open or defective; 
 there may be too thick a coat of paint on the base of the 
 pole, insulating it from the socket ; the roof trolley wire 
 may be disconnected from the stand ; the lightning ar- 
 rester may be connected in wrong; the ground connec- 
 tion may be bad ; the car may be standing on a dead rail. 
 The first step to take is to see that the pole is on the wire, 
 that both hood switches or circuit breakers are closed 
 they are supposed to have been tested before being in- 
 stalled and that the fuse-box has a fuse in it. If these 
 are all right, inspect the trolley stand wire and ground 
 wire. Other causes are harder to locate, and will be 
 deferred until the lamp test is taken up. 
 
 If the car will start on both motors, on one controller, 
 but will not start at all on the other, it indicates the 'open 
 circuit to be local to one controller circuit. In Fig. 199, 
 an open circuit from tap 5 to finger ^ on either end, 
 will disable the car on that end. For example, if one of 
 the blow coil connections gets loose, or if the trolley 
 finger fails to make contact on either controller, no cur- 
 rent can get to either motor, so the car can not be operated 
 from that end. 
 
CAR EQUIPMENT TESTS. 575 
 
 If the car fails to start on both ends with both motors 
 cut in, but will start on both ends with one motor cut out, 
 it goes to show that the break is local to the motor that is 
 cut out. One of its leads may be pulled out of its connect- 
 or, or a brush may be missing, or a brush-holder spring 
 may not be resting on the brush. In any case, the car will 
 not start with the motors in series, but will start on the 
 good motor, as soon as the controller is thrown to the 
 multiple position. 
 
 With an open circuit, anywhere from finger A', to the 
 A*, splice of the starting coil, or if the A', finger fails to 
 make contact with the drum, the controller affected will 
 not start on the first notch, but will start on the second. 
 If the break is in the A*, tap wire, or in the starting coil 
 itself, between A*, and R z the car will not start on the 
 first notch at either end. An open circuit on the A* 3 tap, 
 or in the coil itself, between A* 2 and A'., will prevent the 
 car from starting on either end until the controller is put 
 on the third notch. The most common cause of failure 
 to start is want of contact between a finger and the drum. 
 If the R finger fails to make contact on one controller the 
 car will start on the first notch of that controller, but will 
 loose the power on the second notch, and pick it up again 
 on the third notch. If the break is in the R tap, both 
 ends of the car will behave as in the last case. Any trouble 
 due to bad finger contacts is always accompanied by more 
 or less sizzling that can be heard and felt in the handle. 
 
 Some open circuits can be readily located by inspec- 
 tion or symptom, and others can not. Any kind of an 
 open circuit can be quickly located with a test lamp cir- 
 cuit, as follows : Hang one of the lamp test lines over 
 
576 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 the trolley, or stick it into the trolley side of the car lamp 
 circuit leaving the other test line free to test with. L,et us 
 assume that the positive test line is hung over the trolley 
 wire. In this case, the car pole is tied down, the over- 
 head switches put in, and one of the controllers put on the 
 
 FIG. 200. 
 
 first notch. Suppose that the open circuit is due to paint 
 on the bottom of the pole. The only way the lamp circuit 
 can get a ground, and light its lamps, is by way of the 
 main motor circuit, and if there is a break in this circuit 
 the lamps can not, of course, light, until the test line gets 
 on the ground side of the break. Accordingly, the lamps 
 
CAR EQUIPMENT TESTS. 577 
 
 will not light when the test line is touched to the trolley 
 harp or wheel, because they are on the positive side of the 
 break ; but as soon as the test line touches the trolley 
 stand, or the roof wire running to it, the lamps light. 
 
 The idea is more clearly seen by means of Fig. 200. L 
 is the test bank ; P is the pole ; A', the socket ; S, the 
 stand; K, K, the two hood switches; AA',i\\c two mo- 
 tors. The break, caused by the paint, is between P and 
 N. When the test line is touched to P, the lamps do not 
 light, because no current can get through the paint ; but 
 when the test line is touched to 5", the current can pass 
 through the path, T-L-S-K-K-A-A' , to the ground at G. 
 
 Suppose the break to be due to a wrong connection of 
 the lightning arrester; on that type of arrester through 
 whicn the motor current does not flow (the type indi- 
 cated in Fig. 199), there is little chance of a wrong con- 
 nection being made; such arresters have but two leads, 
 and these are of such light wire that no one with common 
 sense would connect the arrester in series the only con- 
 nection that can make an open circuit ; but on arresters 
 that have three connecting wires or posts, it is an easy 
 matter to create an open circuit. In Fig. 201, views A 
 and B show two three-post arresters properly connected; 
 a is the air gap in all the views. C and D show the same 
 arresters so connected that the trolley wire is on the 
 wrong side of the air gap. In this case, the current can 
 not get to the motor circuit, as there is a break at a, so 
 the test lamps can not light until the test line crosses the 
 
 gap- 
 
 In Fig. 199 suppose that one of the No. i motor 
 brushes is left out the A l brush, for example touch the 
 
578 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 test line successively to the T f Jf l , ^? 2 > -^s> ^i an d 
 connecting 1 posts ; the lamps will not light until the test 
 line touches the AA i post, showing that the fault lies 
 somewhere between A^ and A A j Its location can be 
 still further determined by touching, successively, the A^ 
 reverse finger and brush holder, and the AA t brush 
 holder. The lamps will light on AA lt but not on A ly 
 
 T 
 
 r 
 
 f 
 
 showing the fault to be between the two. In all cases, 
 when testing for an open circuit with the lamp circuit, 
 the fault always lies between the two points between 
 which the lamps light and fail to light.. 
 
 There is no condition that will give a novice more 
 trouble than a "dead rail," which gives the same symp- 
 tom as an open circuit. A dead rail is due to the break- 
 ing of the bond wires that connect the rails together. In 
 
CAR EQUIPMENT TESTS. 579 
 
 such a case the open circuit exists between the dead rail 
 and the rails adjacent to both ends of it. A dead rail is 
 called dead because no current can get from it to adjoin- 
 ing rails ; but it gets very much alive when a car runs 
 on to it, and a person standing on an adjacent rail and 
 touching the dead rail, or any part of the iron work of 
 the car, will get a severe shock. A dead rail can be de- 
 tected \vith the lamp circuit. Neither the car lamps nor 
 the test lamps will light on the dead rail, but will light on 
 the rails next to it. To have a car move itself off a dead 
 rail, a switch iron or other piece of metal must be shoved 
 down between the dead rail and the one next to it. 
 
 GROUNDS AND SHORT CIRCUITS. If a car fails to 
 start on the first notch and the controller flashes when 
 thrown off, it may be due to any of several causes : It is 
 most apt to be due to a ground, short circuit or wrong 
 connection, but sometimes it may be due to stiffness of 
 the bearings or sticking of the brake-shoes. With the 
 lamp circuit, the existence of a ground can be quickly 
 determined, and can be located in a very short while. 
 To determine if a ground exists, throw the reverse switch 
 to the "off" position, in order to disconnect the field that 
 is permanently grounded, and touch every part of the 
 controller with the test line. No part of the controllers 
 should cause the lamps to light, except the fingers which 
 are permanently grounded to the upper bar of the right- 
 hand motor cut-out. If the test shows the existence of a 
 ground, it can be located by a process of elimination. 
 The heater and lamp switches must be turned off, or re- 
 sults may be misleading. A ground anywhere from L 
 to the fuse-box will cause a demonstration as soon as the 
 
580 TESTING OF DYNAMOS AND MOTORS. 
 
 pole is put on the wire, and it must either burn itself out 
 or blow the station breaker, because the car fuse can not 
 act. If, however, breakers are used instead of hood 
 switches, they will act and relieve the situation promptly. 
 If a ground takes place bet ween the fuse-box and the trolley 
 finger on either controller, the car fuse will blow as soon 
 as the pole is put on. Such grounds generally announce 
 themselves very forcibly and do enough burning to make 
 the location easy to find. If the ground is between the 
 fuse-box and the controller trolley finger, and it can not 
 be found by inspection, proceed as follows : Disconnect 
 the trolley side of the lightning arrester and test with the 
 lamp circuit ; if the ground is due to contact between the 
 arrester points, the disconnection of the arrester will re- 
 move the ground. If it does not, disconnect the blow 
 coils in the controllers, one at a time, and test after each 
 disconnection. If the disconnection of a blow coil re- 
 moves the ground, the ground is in that coil. If it is in 
 neither blow coil nor the arrester, it must be in one of the 
 connecting wires or in the main car trolley wire. 
 
 A ground in the starting coil or any of its connecting 
 wires or fingers will cause the lamps to light if any of 
 the wires marked Jt or 19 are touched with the test line, 
 In order to locate the affected part exactly, the wires 
 must be disconnected, one at a time, and a test made after 
 each disconnection. As soon as the faulty wire is dis- 
 connected the remaining ones will not light the test lamps, 
 but the faulty one will. Grounds on the resistances are 
 often due to brake-rods, levers, chains, etc. A ground 
 on motor field No. i, or any of its connections, will cause 
 the test lamps to light up, when the test line is touched 
 
CAR EQUIPMENT TESTS. 581 
 
 to any connection marked /^ or E^ ; a ground on the No. 
 i motor armature will show up the lamps on any A , or 
 AA l connection, and on the armature itself. To tell 
 whether the fault is in the armature itself or in some wire 
 leading to it, draw the brushes and test both the brush 
 holders and the commutator. A ground on motor arma- 
 ture No. 2 will show up on any A 2 or A A* connection, and 
 must be located, as in the last case. Any connection marked 
 F 2 or E 2 will always show up a ground, because one end 
 of the No. 2 field is permanently connected to the ground 
 wire, as in Fig. 199, or to the motor frame, so that in or- 
 der to test it for a ground, the ground wire must be dis- 
 connected. The whole principle of locating a ground 
 lies in disconnecting the wires, one at a time, and elimi- 
 nating the faulty one. 
 
 It is possible for a ground to exist in such a place, 
 that the car will start up all right and run on the series 
 notches, but will blow a fuse as soon as the motors are 
 thrown to multiple. Take, for example, a ground on the 
 motor field No. i, or the No. 2 motor brush holder. In 
 Fig. 202 it can be seen that a ground at G, the dotted line, 
 cuts out motor No. 2 entirely ; as long as the motors 
 are in series, the current passes through motor No. i 
 and to the ground, through the fault. As soon as the 
 motors go over to multiple, the fuse will blow if the 
 ground is on the No. 2 armature brush holder, but there 
 will be no demonstration at all if the ground is on the 
 negative end of the No. i field. A car may run along 
 for days in this condition with no other symptom than 
 that the car starts slowly, and runs at a lower speed on 
 series points. 
 
582 TESTING OF DYNAMOS AND MOTORS. 
 
 If when the car fails to start the controller flashes on 
 being thrown off, it can not be told immediately whether 
 the fault is due to a ground, short circuit, or wrong con- 
 nection. To tell if it is a ground, disconnect the motor 
 ground wire and try the controller again ; if it shows no 
 flash, the removal of the ground wire has removed the 
 
 T 
 
 FIG. 202. 
 
 only ground, so that the fault must be due to a short cir- 
 cuit or a wrong connection. 
 
 Short circuits in the controller or in the car wiring 
 hose, as a rule, do so much burning that they are easily 
 located by the eye. A short circuit between controller 
 parts can be located with the lamp circuit, as has been 
 shown. It can also be used to test the hose, as follows : 
 The hose must be disconnected at both ends; the resist- 
 ance coil and both motors must also be disconnected, for, 
 of course, if the wires are left in the coil, the lamps will 
 light from one wire to the other through the coil; the same 
 
CAR EQUIPMENT TESTS. 583 
 
 is true in regard to the motor wires. The test lamp con- 
 nections of Fig. 196 are used; one test line is hooked on 
 to one of the car wires and the other test line used to 
 touch every other car wire. If the tester is familiar with 
 the controller under test, it is unnecessary to disconnect 
 the controllers, as all the blocks to which the car wires 
 connect are insulated from each other when both drums 
 are at the " off " position. 
 
 There are two very common sources of short circuit 
 that care must be taken to avoid. The trolley and ground 
 wires being alike, are easily confused; if they become in- 
 terchanged in either controller, the car fuse blows as soon 
 as the pole is put on the wire, or as seen in Fig. 199, the 
 current passes through the fuse-box directly to the ground. 
 The other common source is due to the workmen's practice 
 of using a wrench, instead of the handle, to turn the con- 
 troller drums ; the result is that a reverse drum is some- 
 times left at an "on" position on one of the controllers ; 
 let it be left at "go ahead," for example; if anyone goes 
 to the other controller and throws its reverse lever to "go 
 ahead" and tries to start the car, the car fuse blows ; be- 
 cause the two controllers arc set to run the car in opposite 
 directions, with the result that there is a short circuit, as 
 soon as enough of the starting coil is cut out. 
 
 It will be seen in Fig. 1 99 that the armature wires are 
 crossed in one controller ; that is, the A l armature wire 
 goes to the A l connecting board block on the front con- 
 troller, but it connects to the A A l block on the rear one; 
 the same is true of the Xo. 2 armature ; this is done so that 
 the car will move in the direction that the reverse lever 
 points, irrespective of which controller is in use. The 
 armature wires are said to run to the front controller 
 
584 TESTING OF DYNAMOS AND MOTORS. 
 
 straight, and to the rear controller crossed, and the de- 
 vice is necessary from the fact that the controllers face op- 
 positely, as regards the direction of motion of the car. 
 
 It has been shown how to test the polarity of the motors 
 so that the proper direction of rotation shall be insured. 
 Experts, as a rule, do not resort to this preliminary test, 
 preferring to connect the motors up "hit or miss," and 
 correcting the job where a trial proves it necessary. 
 This "hit or miss" method of connecting can result in 
 any of the following complications : The motors may 
 oppose or "buck" each other on one or both ends of the 
 car, so that the car can not move at all on both motors. 
 The car may move oppositely to the indication of the re- 
 verse lever, on one or both ends. One motor 
 may obey the indication of the reverse switch 
 on both ends of the car, and the other motor 
 disobey it on one or both ends. Any of the above com- 
 plications can be straightened out as follows: (i.) If 
 the motors buck each other, the gears will be heard to 
 give a click when the controller is put on the first notch, 
 and the controller will flash when thrown to the "off" po- 
 sition. To settle the matter conclusively and decide 
 which motor must be corrected, try them one at a time, 
 on both ends of the car, and note which motor obeys the 
 reverse lever. If motor No. i is right on the front 
 end and wrong on the rear end, while motor No. 2 is 
 wrong on the front end and right on the rear end, re- 
 verse the No. 2 armature wires in the front controller, and 
 the No. i armature wires in the rear one. If the No. I 
 motor is right on both ends, and the No. 2 wrong on 
 both ends, the armature leads of the No. 2 motor must 
 
CAR EQUIPMENT TESTS. 
 
 585 
 
 be reversed at the motor. If the Xo. I motor is wrong 
 on botli ends, reverse its armature leads at the motor. 
 (2.) If both motors are right on one end and buck on the 
 other, reverse the controller armature wires on the motor 
 that is wrong. (3.) If the car moves contrary to the 
 indication of the reverse lever on one end and is all right 
 on the other, reverse the controller armature wires of 
 T T 
 
 _D 
 
 FIG. 203. 
 
 both motors on the contrary end. If the movement of 
 the car is contrary on both ends, reverse both motor ar- 
 mature leads at the motors. 
 
 Assuming that the car starts properly on the first 
 notch, it should notch in evenly distributed impulses up 
 to the last series notch, where the loop or shunt is cut 
 into action. If any of the resistance wires get inter- 
 changed in a controller, it will cause the car to accelerate 
 in jumps on that controller; if the resistance wires get 
 confused at the coil itself, it will cause bad notching at 
 
586 TESTING OF DYNAMOS AND MOTORS. 
 
 both ends. For example, suppose the .^i and ^ 3 wires get 
 interchanged at the coil itself. In Fig. 203 view a shows 
 the resistance wires properly connected; view b shows 
 the J^ l and ^ 3 wires interchanged ; in view b, on the first 
 notch, the current takes the path, T-T-L-M-N-O-P-G, 
 to the ground; the starting coil is cut out and the car 
 starts with a jump. On the second notch, the current 
 path is T-T-L-B-D-P-G ; half the coil is cut in, so the 
 car runs faster on the first notch than on the second. On 
 the third notch, the current path is T-T-L-S-C-A-P-G ; 
 the whole coil is cut in and the car runs slower than ever. 
 Such a mistake in connections is most noticeable on 
 grades and curves and with loads, the car jerking and 
 halting in a very disagreeable manner. The controller 
 tips burn and blister, and the starting coil heats. Con- 
 fusion of the R z and ^ 3 , or the R^ and R l wires causes 
 the same action, but to a less degree. If the car starts 
 smoothly and notches evenly, it should be run a mile or 
 so with the motor armature and axle cap bolts loosened 
 up a little to give the bearings a chance to find a seat. 
 In starting a car for the first time, the whole action of the 
 car is more or less stiff, and especially is this true of the 
 bearings. It is a good idea to prime the bearings with a 
 good quality of cylinder oil to start with, and pack the 
 boxes with regular motor grease. The gear cases are 
 filled to a depth of about 6 in. with gear grease. If none 
 of the bearings show a tendency to heat, the caps can be 
 screwed down and the car put on a regular run, which, 
 if possible, should be so selected as to have the car pass 
 the shop, the first day, so that if any trouble arises, it will 
 be easy to get the car in. If any bearing shows a ten- 
 
CAR EQUIPMENT TESTS. 587 
 
 dency to heat, take it out and scrape it and try the car 
 again. Under no circumstances should a car with a 
 faulty bearing be turned over to the operating company, 
 as it is liable to tie up the road. 
 
 This completes the first series of tests on a car; all of 
 them can be made with the lamp circuit. To make sonic 
 of the tests that follow, an ammeter or voltmeter, or bo:h 
 will be required. 
 
 If either pair of car wheels shows a tendency to spin 
 at starting, any of several things can be the matter. The 
 brakes may be poorly adjusted, so that when released, Un- 
 shoes hug one pair of wheels ; in this case, it is the wheels 
 that are free that spin. Xo chronic fault to which cars 
 are heir can cause more trouble than sticking brake- 
 shoes. There is a constant and heavy extra load imposed 
 upon the motors. This causes heating of the motors, 
 sparking, blowing of fuses, breaking down of controllers, 
 and grounding of brush holders. When the shoes stick, 
 the shoes and wheels get very hot and give off an odor 
 that is easily recognized. If the wheels start to spin in 
 passing to the second or third notch, it shows that the 
 resistance coil is not divided up right ; too much of the 
 coil is cut out at a time. If the spinning takes place on 
 the first notch, it shows the entire resistance of the coil 
 to be too low. A two-section coil is generally divided, so 
 that about five-eighths of it is cut out on the second notch 
 and three-eighths on the third. On a three-section coil, 
 half of the coil is cut out on the second notch, quarter on 
 the third, and the remaining quarter on the fourth. 
 Sometimes a car will spin its wheels when it first goes 
 on duty, but will cease to as soon as the coil heats up. 
 
588 TESTING OF DYNAMOS AND MOTORS. 
 
 A very simple test with a voltmeter will deter- 
 mine whether the resistance coil is properly divided or 
 not. The car is allowed to run along on the first notch, 
 and when at a steady speed, the drop is taken on the 
 whole coil, and on each section. Suppose that the drop 
 over all this is 130 volts on a two-section coil; then the 
 drop from JK t to R z should be 81 volts, and the drop 
 from R% to RZ , 49 volts; 49 : 8 1 : : 3 : 5 ; or 49 is ^ of 1 30, 
 and 8 1 is f of 130, and since the drops are proportional to 
 the resistances across which they take place, the resist- 
 ances are divided in the same ratio, and are therefore 
 right. 
 
 Spinning of the car wheels is also caused by a short- 
 circuited or wrongly connected loop, or baked or wrong- 
 ly connected field coils. Any of these faults can be 
 tested for with a voltmeter. Suppose the loop part of 
 the field is short-circuited ; as this part of the field is not 
 cut out till the fourth notch is reached, its effects can be 
 felt on all notches except the loop notch. If the loop 
 wire is confused with one of the end field wires, one part 
 or the other (according to which end field wire is in the 
 confusion) of the field coil is cut out even on the first 
 notch, with practically the same result as if part of the 
 field were short-circuited. If the field coils are baked or 
 wrongly connected, or if the motor case is partly open, 
 or a pole piece or two loose, the conditions and symptoms 
 are the same as in the above cases. The net result of any 
 of these faults is to weaken the field of the motor, thereby 
 decreasing its starting power and increasing the speed. 
 The car is slow in starting and takes more time to get under 
 full headway, but its maximum speed is greater in most 
 
CAR EQUIPMENT TESTS. 589 
 
 cases than it would be were the field in perfect order. A 
 weak field on one motor causes the motor to shirk its 
 load on the series notches, and to take more than its share 
 of the load on the multiple notches. The c. e. m. f. 
 of a motor, being 1 dependent upon the field strength, is 
 low when the field is weak ; so that when the motors are 
 in multiple, and each has an independent path, tije one 
 with the weak field and low c. e. m. f. will take the most 
 current, and in some cases the faulty motor will take more 
 current than both ought to take under normal conditions ; 
 the result is, the car blows fuses and gives general trouble. 
 When the motors are in series, the c. e. m. fs. are in 
 series, and the conditions are entirely different. Between 
 the line and the ground, the line e. m. f. distributes 
 itself according to the resistance it meets ; where the re- 
 sistance is greatest, the greatest drop takes place. It is 
 upon this fact that is based a voltmeter test for determin- 
 ing whether or not the motor fields are unbalanced. The 
 test is conducted as follows : The controller is put on the 
 third notch, and the car allowed to get under full head- 
 way ; the voltmeter is then, by means of test lines, ap- 
 plied to first one armature and then the other. That is, 
 the drop is taken on the two armatures, so that they may 
 be compared. If the two motors were perfectly balanced, 
 the two readings would be the same; but there are sev- 
 eral factors that lend an influence toward making a 
 slight difference in the readings ; among them are the fol- 
 lowing: Slightly different shape in the pole pieces; dif- 
 ferent quality of steel in the two frames or cases ; dif- 
 ferent dimensions of cases due to the difference in shrink- 
 age on cooling; difference in the setting of the brushes; 
 
590 TESTING OF DYNAMOS AND MOTORS. 
 
 difference in the size of the car wheels to which the two 
 armatures are geared; condition of the commutators in 
 regard to size and cleanness ; condition of the arma- 
 ture bearings in regard to wear the mo^e the bearings 
 are worn, the wider is the air gap on top and the nar- 
 rower is it on the bottom; the internal resistance of the 
 two motors may not be the same. All of these minor in- 
 flnences should not conspire to produce a total differ- 
 ence of over 10 volts in the two readings. 
 
 If there is an irregularity in the fields of either motor, 
 it may cause a difference of anywhere from 30 to 150 
 volts in the two readings. Let us suppose that the line 
 voltage is 500 volts and that the drop across the two 
 armatures at uniform speed is, respectively, 285 and 195 
 volts, a difference of 90 volts, their sum being 480 volts ; 
 the 20 volts unaccounted for being dropped across the mo- 
 tor fields and car wires. Where there is a difference of 90 
 volts in the amount of line pressure absorbed by the two 
 motors when in series, there is something radically wrong 
 with the motor whose armature gives the lowest reading. 
 Its field is abnormally weak. Inspection will prove whether 
 or not the motor case is partly open or any of the pole 
 pieces loose ; if these parts are all right, the trouble must 
 be in the field coils themselves. Suppose that the loop 
 wire and one of the end field wires of one motor are con- 
 fused in one of the controllers; in this case, if the two 
 motors are otherwise all right, the voltmeter will show a 
 discrepancy in their c. e. m. f. only on one end of the 
 car; if the loop wire and field wire are confused at the 
 motor itself, the discrepancy in e. m. f, will show at both 
 ends. The voltmeter will tell if the loop wire is con- 
 
CAR EQUIPMENT TESTS. 59 1 
 
 fused with either field wire; apply it as follows: (Figs. 
 199 and 1 86. The dotted lines in Fig. 186 indicate the 
 loop wire and F+ wire to be interchanged ; in this case 
 only the upper part of the field coil is in use, and it only 
 takes the current. The result is that when the volt lines 
 are applied to the F l and E\ , and the F and field 
 connecting board blocks, the drop on the wrongly con- 
 nected field will be only about half what it is on the good 
 field. To prove conclusively if the loop and field wire 
 have been interchanged, disconnect the loop wire from 
 both controllers; on the controller where no confusion 
 exists the car will start on the first notch, as usual ; but 
 on the end where the trouble is, the car will not start un- 
 til the controller reaches the loop notch. In Fig. 186, 
 loop wire L and field wire F-\- have exchanged places, 
 as indicated by the dotted lines. With the normal con- 
 nection the current path on the first position is T-FB-A- 
 B-C-R-D-F+-O-F A+-A N-G; with the inter- 
 changed connection, the current path becomes T-FB-A- 
 B-C-R-D-O-F A+-A .V-G; the loop wire has be- 
 come a part of the main circuit, and disconnecting it, pre- 
 vents the car from starting on the first notch. As soon 
 as the shoe, Fig. 186 or the drum, Fig. 199 reaches the 
 loop position (with the loop wire disconnected), the cur- 
 rent passes through the whole field, and since all resist- 
 ance is cut out, the car starts with a jump. If then, the 
 voltmeter shows half the normal drop between the field 
 connecting board blocks, shows no drop at all between the 
 loop block and one of the field blocks (the field block to 
 which the loop wire has been run by mistake) ; and the 
 car fails to start until the loop notch is reached, the evi- 
 
592 TESTING OF DYNAMOS AND MOTORS. 
 
 dence is complete enough to convict the loop wire. If 
 the symptoms exist on one controller, the trouble is in 
 that controller' s wires ; but if on both controllers, the 
 confusion is between the motors and the car wiring hose. 
 To finally clinch the diagnosis, it is the practice to dis- 
 connect the wires involved in the trouble and test them 
 out with the test lamps. 
 
 If the connections of the field and loop prove to be 
 all right, the trouble must be due to baked or short-cir- 
 cuited or wrongly connected field coils, inside of the mo- 
 tor. To tell if the coils are baked or short-circuited, pro- 
 ceed as follows : Get a field coil just like the ones in the 
 motors and set it inside of the car ; lift the trap doors and 
 motor covers, so that the motor fields may acquire about 
 the same temperature as the test field (the motors should 
 be allowed to cool off over night). Take the insula- 
 tion off the field connectors, so that the ends of the field 
 may be accessible to the voltmeter test lines ; connect the 
 test field in series with one of the motors ; reverse the ar- 
 mature terminals on one motor so that the two motors 
 will buck each other, and will, therefore, be unable to start 
 on the first notch. Everything being ready, have a helper 
 put one of the controllers on the first notch ; quickly take 
 the drop on the two motor fields and test field and throw 
 the controller off to avoid heating the starting coil too 
 much. Repeat this operation several times, so that there 
 may be several sets of readings to compare. -Each read- 
 ing in the last set will be lower than the corresponding 
 reading in the first set, because the heating of the start- 
 ing coil raises its resistance rapidly and diminishes 
 the current; but the relation between the readings 
 
CAR EQUIPMENT TESTS. 593 
 
 in the several sets should remain about the same. Since 
 eacli motor has four coils in it, its field resistance 
 should be four times as great as that of the test coil 
 and the drop, therefore, four times as great. If the 
 drop on one of the motors proves to be only three 
 times that on the test coil, the insulation must be 
 skinned off the back field connection of the suspicious 
 motor, so that the drop can be taken on two coils at a 
 time. The current is turned on again and several more 
 sets of drops taken on the test coil and the motor. The 
 drop on each pair of motor coils should be twice that on 
 the test coil. If the drop on one pair of coils is twice as 
 much, but that on the other considerably less, it indicates 
 a fault local to the low drop pair of coils ; but if the drop 
 on both pairs of coils is low, it indicates the coils to be 
 baked. Acting upon this indication, the best thing to do 
 is to lift the insulation on the end of one of the bottom 
 coils and look at the wire. To do this, the motor must 
 be opened. The bottom coils are selected, because, being 
 on the bottom where the ventilation is poor, they are, as 
 a rule, the first to bake. If the coils are not baked, the 
 fault is local to one coil. To locate this coil, take the 
 drop on all four of the coils, one at a time. Single coils 
 are sometimes short-circuited by the inside end touching 
 one of the top layers in its path from the bottom layer 
 to the surface. 
 
 On equipments that permanently ground the field on 
 motor No. 2, as in Fig. 199, it is possible for a second 
 ground to develop in the field and cut out a coil or two. 
 In such a case, the voltmeter will show no drop when its 
 lines are held on the coils on the ground side of the fault, 
 
594 TESTING OF DYNAMOS AND MOTORS. 
 
 as no current passes through. With such a fault, it is 
 possible for a car to run for days, chewing up commuta- 
 tors, throwing solder in armatures and grounding brush 
 holders. 
 
 If the coils prove to be neither baked nor short- 
 circuited, the trouble must be due to a wrong in- 
 ternal field connection. One coil or two may be con- 
 nected in wrong, thereby reversing them. It is a conserv- 
 ative estimate to say that one-fourth of the fuse blowing 
 and field roasting troubles encountered on modern types of 
 motors is due directly or indirectly to a field coil 
 being put in wrong or connected wrong, and, in modern 
 motors, all the blame does not rest on the man that puts 
 in the coil. In the effort to economize on space inside 
 of the motor, and to facilitate opening the motor without 
 disconnecting any of the wires, the path given to them 
 inside of the motor is a very circuitous one. There are 
 points where the field connecting wires can be neither 
 seen nor felt, and when two or more wires disappear 
 under a field where they can not be followed with the eye 
 or hand, it is impossible to tell whether they are con- 
 nected right or not. On such a motor it is the practice 
 of the pitmen to connect a new field in just as the one 
 taken out was connected. If a motor ever happens to 
 leave the factory with a wrongly connected field, the 
 chances are that every field put into the motor for some 
 time after it is installed will be put in the same way and 
 cause trouble, until the fault is detected by test. If one 
 of the motors on a car has a reversed coil, the following 
 symptoms prevail : The car will start with the motors 
 in series, even though the faulty motor might be unable 
 
CAR EQUIPMENT TESTS. 595 
 
 to start the car alone, because the car starts on the good 
 motor, the faulty one acting simply as a means of com- 
 pleting the circuit. As soon as the motors are thrown 
 over to the multiple and each has an independent path to 
 earth, the faulty motor, having a very low c. e. in. f., 
 lets in a current that blows the breaker or main motor 
 fuse. 
 
 Where the car is too heavily fused, or the breaker is 
 out of order, a car will run along with a reversed field 
 until all the fields begin to roast, when the fuse begins 
 to give trouble. A reversed coil will not only roast itself, 
 but will roast all coils in series with it, so that when one 
 coil in a motor is found to be roasted, all of them had 
 better be taken out. 
 
 A reversed coil can be tested for by means of a compass 
 or wire nail. The test with the compass docs not require 
 that the motor case be opened, but it must be made by one 
 familiar with the tricks of compasses, or the results will 
 be misleading. On any street car motor the poles should 
 alternate, in polarity ; that is, if any given pole is north, 
 the poles adjacent must be south, and vice versa. To 
 test with a compass, a current must be sent through the 
 motor as it hangs on the truck. This is best done 
 by using an outside resistance a water box or a lot of 
 old starting coils because it injures the regular starting 
 coil to subject it to the full voltage for so long a time. 
 A current of about half the full load capacity of the motor 
 is sent through the field alone. Holding the compass per- 
 fectly level, so that the needle will not stick, it is passed 
 entirely around the motor in a circle whose plane is per- 
 pendicular to the axle and suspension bar, and a little to 
 
596 
 
 TESTING OF DYNAMOS AND MOTORS. 
 
 one side of the center. If the fields are connected right 
 the needle will reverse ends every time a pole piece is 
 passed. Where two adjacent poles are alike, and there- 
 fore of the wrong relative polarity, the needle will not 
 reverse ends. 
 
 The test with the nail is along the same lines as that with 
 the compass, but as it is made on the inside of the motor, it 
 requires that the motor be opened, the armature taken out, 
 and the case closed again. This is necessary from the fact 
 that the space inside of a modern motor is so limited that 
 
 FIG. 205. 
 
 FIG. 206. 
 
 there is hardly room to reach all parts with the hand. 
 Also, when a current passes through the field coils and 
 the armature is in place, the poles induced in the arma- 
 ture confuse the tester who is not accustomed to make 
 the test. The test is especially useful in shops where 
 whole sets of fields are put in ; also at depots, to test a 
 car whose record leads it to be suspected that there is 
 something wrong with the field connections. The test 
 is conducted as follows: The motor case is opened, the 
 armature taken out and the case closed again ; a current 
 is then sent through the field coils as in the compass test ; 
 
CAR EQUIPMENT TESTS. 597 
 
 if the coils are connected properly, the poles will alter- 
 nate in polarity and the path of the lines of force will 
 be as shown in Fig. 204 but if one of the coils is of the 
 wrong- polarity, there will be three like poles adjacent to 
 each other, and the general path of the lines of force will 
 be as shown by the dotted lines in Fig. 206. A mag- 
 netizable piece of metal, if free to move, will arrange 
 itself parallel to the general direction of the lines of force 
 of the field that it is in. If the fields are right and the 
 lines flow as in Fig. 204, a wire nail held loosely between 
 the thumb and forefinger will, when passed by hand 
 from one pole piece to the other, take an easy, natural 
 path, never turning, but persevering in the general di- 
 rection of the lines of force. Its pointed end, say, is the 
 last part to leave the one pole piece, and the head end is 
 the first part to touch the approaching one. If, however, 
 one or more coils are reversed, and the path of the lines 
 of force is that of Figs. 205 or 206, the nail, in its passage 
 from one pole piece to the other, will show a tendency to 
 take up a position at right angles to the general path of 
 the lines of force. In Fig. 204 a, a, a, a, show the posi- 
 tions taken by the nail in passing from one pole to the 
 other, when all the coils are connected properly. In Fig. 
 205, the right-hand bottom coil is wrong, making the 
 pole S instead of N; the result is that the nail, as indi- 
 cated at a', a', takes the perpendicular position on both 
 sides of the faulty coil ; whenever a single coil is re- 
 versed, then the action of the nail is irregular on both 
 sides of it ; also its pole piece is much weaker than any 
 of the others ; the pole piece opposite being the strongest 
 of the four. To right matters, it is only necessary to re- 
 
598 TESTING OF DYNAMOS AND MOTORS. 
 
 verse the connections of the faulty coil, when it becomes 
 an N pole, and the condition of alternate polarity is re- 
 stored. 
 
 In Fig. 206, two coils have been reversed, with the re- 
 sult that the top and bottom halves of the case work 
 against each other, and with the final result that the mo- 
 tor, as far as the field is concerned, becomes a bipolar 
 motor with compound pole pieces. It can not act as a 
 bipolar motor, however, because the armature is not 
 adapted to run in that kind of a field. The letters, a, a, 
 and a', a', show the positions taken up by the nail ; be- 
 tween the unlike poles, its position is regular, but be- 
 tween the like ones, irregular, so that when the nail takes 
 up the cross wire positions at opposite ends of a di- 
 ameter, the conclusion is in order that two of the field 
 coils are connected wrong, and the diameter is the di- 
 viding line between the two halves, each of which has a 
 reversed coil. In order to right matters, do either of 
 two things reverse the connections of both coils in either 
 half, or just have these two coils exchange places. It is, 
 of course, possible for the coils to be so reversed as to give 
 the polarity shown by the dotted letters in Fig. 206. In 
 such a case, the irregularity in the position of the nail 
 takes place on the dotted vertical diameter, dividing the 
 field into a right and left-hand half, instead of top and 
 bottom. The best way to familiarize oneself with the 
 test is to take a motor, connect the fields up right and 
 wrong, and try them. After the "feel" of the nail is once 
 acquired, there is nothing difficult in the test. 
 
 A two field- coil motor with one coil reversed, or a four- 
 coil motor with two coils reversed, will not start a car 
 
CAK EQUIPMENT TESTS. 
 
 599 
 
 alone with normal current ; upon this fact is based a very 
 good dynamic test for determining if the field coils on a 
 motor are reversed, and with very little trouble. The 
 test is conducted as follows: Disconnect the back field 
 connection, thereby dividing the motor field into two 
 
 halves of two coils each. Regarding each half as if it 
 were the complete field, connect the two halves in, one 
 at a time, and try to start the car. Suppose that the 
 coils in the top half are connected right and that those 
 in the bottom half oppose each other; then, when the top 
 half is connected in, the car will start on about the second 
 notch ; but when the bottom half is in, the car will not 
 start at all, because the two oppositely directed pole 
 
6oo TESTING OF DYNAMOS AND MOTORS. 
 
 pieces try to start the car in opposite directions, and it 
 can not, therefore, start at all. If the car will not start 
 or spin the wheels on the third notch, it indicates the field 
 coils are connected wrong in both halves. This test is 
 especially valuable in that it can be conducted by anyone 
 who knows how to connect the fields and start the car. 
 Pitmen can, therefore, use it for determining in which 
 half of the case the faulty field lies, and thereby, per- 
 haps, save labor in opening up the motor. 
 
 Fig. 207 (a), (b) and (c), show two field coils of the 
 general shape used on modern motors. There are four 
 coils to a motor, and the coils being alike, are inter- 
 changeable. This is a good feature, in that only one style 
 of coil need be kept in stock, but it increases the liability 
 of confusing connections. Each coil has an inside end, 
 marked 7, and an outside end, marked O; the inside ends 
 can be generally distinguished by the fact that, as a 
 rule, there is a bump on the coil where the inside end is 
 brought to the surface. In the motor, an inside end al- 
 ways connects to an inside end, and an outside end to an 
 outside end, leaving two inside ends or two outside ends 
 to be brought out of the motor to tap on to the car wiring 
 hose. Fig. 207 (b), shows the proper connection for two 
 fields ; for example, suppose that these two fields go into 
 the top half of the motor ; two others, similarly connected, 
 go into the bottom half; the second pair must be put in 
 so that similar ends of the two pairs shall be opposite; 
 two of the opposite ends connect together, and the two 
 remaining leads are brought out as motor leads. There 
 are two ways in which two coils may be connected to- 
 gether wrongly. Fig. 207 (a) shows one way ; here one 
 
CAR EQUIPMENT TESTS. 6oi 
 
 coil has been turned end for end. Fig. 207 (c) shows a 
 second way ; here one coil has been turned over on its 
 back. The result of both of these errors is to bring two 
 adjacent coils together in such a way that the inside end 
 of one connects to the outside end of the other. The re- 
 sult of this is to have the current enter and leave the two 
 coils in the same way, thereby making their polarities 
 alike when they ought to be opposite. If the current goes 
 into one coil at its inside end, it should enter the coils 
 next to it on either side, at their outside ends. There is 
 some excuse for getting a coil put in end for end, for 
 where the lugs or leads are not marked in any way, and 
 are disposed as in the figure, it is difficult for the un- 
 practiced eye to distinguish them. There is, however, 
 no excuse for putting a coil in top side down, as the one 
 side is always curved to fit the inside shape of the motor 
 case, and to get the flat side next to the case, the coil 
 has to be forced. The only absolutely certain way to in- 
 sure that the coils are connected right, where there is any 
 doubt, is to test with the compass or nail. 
 
 Where an armature develops an actual ground or short 
 circuit, or open circuit, no instrumental test is necessary 
 to locate it, as there is always more or less demonstra- 
 tion. An open-circuited or grounded armature can not 
 run the car alone, but if both motors are cut in and are 
 in series the good motor can run the car, unless the 
 ground is on the motor next to the trolley wire, in which 
 case the current passes to the ground without reaching 
 the good motor at all. When the motor armature next the 
 trolley has a ground the action of the armature, when the 
 power is applied, depends upon whether the field is next to 
 
602 TESTING OF DYNAMOS AND MOTORS. 
 
 the ground, as in Fig. 199, or whether the armature is, as 
 is the case on some wiring diagrams. If the field is next to 
 the ground, the current does not pass through it, as the 
 fault on the armature cuts it out ; so the main motor fuse 
 blows as soon as the controller reaches the second notch. 
 If the armature is next to the ground, it will, when the 
 power is applied, turn partly over, far enough for the 
 ground to come under the positive brush holder, and 
 stops. If the controller is advanced beyond the first notch 
 the main motor fuse blows. 
 
 A ground on the motor next to the ground will per- 
 mit the car to run on the good motor as long as the two 
 are in series, but the faulty motor will run with a jerky 
 motion. A short circuit in the armature winding or com- 
 mutator will cause this same jerky motion of the arma- 
 ture; but the two faults can be readily distinguished by 
 their action when the ground wire is disconnected. If 
 the. fault is a short-circuited armature, disconnecting 
 the car ground wire will open the circuit, so that the 
 car can not start on either motor ; if the fault is a ground, 
 the current still has the fault through which to pass to 
 earth so that the removal of the car ground wire does not 
 keep the car from starting on the good motor. A short- 
 circuited armature can be readily detected by means of a 
 pocket knife or a piece of iron. If either is held up 
 near the head of the armature it will, if the armature 
 has a short circuit in it, vibrate or pulsate ; on modern 
 types of armatures the slots are so wide that they will 
 cause some pulsation of the test piece, but that due to 
 short circuit is readily distinguished, because it is more 
 violent and is less frequent per revolution of the armature. 
 
CAR EQUIPMENT TESTS. 603 
 
 The first symptom of an open circuit in an arma- 
 ture is a chewing away of the mica from between the 
 commutator bars to which the open -circuited coil is con- 
 nected. This takes place as soon as the open circuit 
 starts while there is still a contact, but a poor contact. 
 As soon as the rupture is complete a ball of fire follows 
 the commutator around when the armature is in motion. 
 A motor will start with a single open circuit in the arma- 
 ture, because the single open circuit docs not open the ar- 
 mature circuit entirely, there being two paths through it. 
 After the armature is in motion, both halves are active in 
 turning it, because one half is intact, and the other half 
 gets current through the arc that holds across the break. 
 An armature sometimes gets partially open-circuited all 
 around the commutator, due to the fact that the ex- 
 cessive current, due to abuse of some kind, has melted 
 the solder out of the connections, impairing them. 
 
 This chapter would be incomplete were no allu- 
 sion made to the property that street car motors 
 have of acting as electric brakes in time of emergency. 
 This property is based upon one or two properties of 
 dynamos, and to understand the brake action, these must 
 be briefly outlined. Two dynamos are said to oppose each 
 other when they are in scries, as far as connections go, 
 but have their polarities opposed, i. e., the two dynamos 
 try to send current through each other in opposite di- 
 rections. In Fig. 208, ^,-f, A , and F l -f- , are the 
 armature and field leads of one dynamo; A t + t A* 
 and F those of the other. The connections are such 
 that machine No. i tries to send the current around 
 the circuit clockwise, while No. 2 tries to send it counter 
 
604 TESTING OF DYNAMOS AND MOTORS. 
 
 clockwise. If the conditions are such that the two ma- 
 chines have the same e. m. f., no current can flow, be- 
 cause the e. m. f. of one is counteracted by the equal and 
 opposite e. m. f. of the other. If, however, for any rea- 
 son, the field on one machine becomes stronger than that 
 on the other, the machine with the stronger field 
 will force a current back through the one with the 
 weaker field and run it as a motor. If the two ma- 
 
 'A1- A2-N 
 
 A1+ A / \ A A2+ 
 
 F2 - 
 
 FIG. 208. 
 
 chines have no field save that due to their residual 
 magnetism, the machine with the stronger residual 
 will run its mate as a motor ; before this can happen, 
 however, one other condition must be fulfilled. The 
 machines shown in the figure are series machines ; 
 street car motors are series machines. For given 
 connections, series machines run in opposite directions, 
 as dynamos and motors; therefore, a street car motor 
 mounted on a car under headway can not generate unless 
 either its field or its armature connections are reversed. 
 Fig. 209 is a diagrammatic sketch of the general con- 
 
CAR EQUIPMENT TESTS. 
 
 605 
 
 nections of two motors under a car, the motors being in 
 series; Fig. 2 10 is the same for the two motors in multiple. 
 In Fig. 210 as long as the car is taking power, the current 
 splits at ,r and divides between the two motors. In con- 
 nection with Fig. 210 suppose that the car is going at a 
 fair rate of speed, and that the controller is thrown to the 
 
 0T 
 
 RESISTANCE 
 
 RESISTANCE J 
 
 a 
 
 X d 
 
 1 ' 
 
 G 
 FIG. 209. FIG. 210. 
 
 off position, and the reverse lever thrown back ; the 
 throwing of the reverse switch reverses the two arma- 
 tures, and thereby connects the two motors up as dyna- 
 mos in a position to generate as soon as conditions per- 
 mit. The condition necessary is that the controller be 
 advanced to a multiple notch, the overhead switch being 
 thrown, to cut off the power ; the result of this is to throw 
 the two motors together in a local circuit, as shown in 
 
606 TESTING OF DYNAMOS AND MOTORS. 
 
 Fig. 210. The car being still in motion, in virtue of its 
 momentum, and both motors being connected up as 
 dynamos, each machine tries to act as a generator and 
 run its mate as a motor, and the machine that has the 
 most residual magnetism can generate the higher voltage, 
 and thereby back a current through the lower voltage 
 machine and run its armature as a motor. The throwing 
 of the reverse lever connects both motors up as dy- 
 namos, for the given direction . of rotation, as has been 
 stated ; so that as soon as one machine becomes a motor 
 with this connection, the wheels to which it is geared, 
 spin around in the opposite direction to what the direc- 
 tion of motion of the car calls for. If the car is under 
 good headway, when the motors are reversed, as soon as 
 one of them takes hold as a dynamo, the speed of the car 
 gets a sudden and violent check, for the following rea- 
 sons : In the first place, the spinning of one pair of 
 wheels in the wrong direction has a retarding effect. 
 Also, in the local circuit that includes the two dynamos, 
 there is no resistance save that of the motors themselves, 
 so that the machine that becomes a generator, acts 
 through a short circuit and generates a very large cur- 
 rent; this gives the momentum of the car a lot of work 
 to do, and consequently checks the speed. 
 
 It is a good thing to know how to use the motors as 
 brakes, for there is no reckoning on when a brake chain 
 or rod may give way at the same time that the line power 
 fails, leaving the car helpless, except for the braking 
 power of the motors. It is not well, however, to make a 
 practice of stopping a car in this way, for since the re- 
 sistance coil and fuse-box are outside of the circuit in 
 
CAR EQUIPMENT TESTS. 607 
 
 which the dynamo acts, the machines are not protected 
 from overload, save by the spinning of the wheels. 
 Again, the sudden reversal of one armature strains its 
 pinion and gear, just as a sudden reversal with the power 
 does. There have been several plans devised for using 
 the generative ability of the motors to stop the car. They 
 all depend upon the use of a resistance to keep the current 
 and braking power down to a safe value. Most of these 
 devices require that the hand brake be used to bring the 
 car to a stand, or to hold it on a grade, because the mo- 
 tors, of course, cease to generate as soon as the car comes 
 to a stop. 
 
 To stop a car. then, by means of the motors and an 
 ordinary controller, when the car is moving "ahead," 
 throw the overhead switch off, throw the reverse lever 
 back (to do this the controller must be thrown to the "off" 
 position), and advance the controller to the multiple po- 
 sition. When the car is backing up, with the reverse lever 
 at the "back-up" position, throw the overhead switch, 
 throw the reverse lever "ahead," and put the controller 
 on a multiple notch. If, however, the car is ascending 
 a grade, and the loss of the power and the failure of the 
 brake rigging starts it down the hill backward, do not 
 move the reverse lever; leave it where it is and simply ad- 
 vance the controller to the first multiple notch ; as far as 
 the motors' generating is concerned, it makes no differ- 
 ence what multiple notch the controller is on, because the 
 starting coil is not in the local generating circuit ; but if 
 the power should happen to come back on the line while 
 the car is descending the hill, it will check the speed too 
 suddenly, if all resistance is cut out, will blow the 
 
608 TESTING OF DYNAMOS AND MOTORS. 
 
 fuse, and perchance, strip a pinion. When the reverse 
 lever is set for the motors to generate, it may take sev- 
 eral seconds for them to take hold, but when they do, 
 they do so very suddenly. 
 
 One other strange feature about street car mo- 
 tors, is the peculiar action known as bucking, which is 
 closely related to their braking ability. Bucking takes 
 place most commonly on equipments that have the motor 
 armature next to the ground ; why this is, can be readily 
 seen in Fig. 211, where T is the trolley; a, the motor ar- 
 mature ; F, the field, and G the ground. If a ground oc- 
 curs on the field at a' for example, that part of the field 
 
 
 FIG. 211. 
 
 that lies between the fault and the trolley wire becomes 
 separately excited ; both ends of the armature being 
 grounded, one through the permanent ground and the 
 other through the fault, the faulty motor runs as a sep- 
 arately excited short-circuited dynamo; the current gen- 
 erated by the armature being very heavy, the drag be- 
 tween the armature and pole pieces is correspondingly 
 so, and so is the work thrown upon the car. There be- 
 ing nothing to do this work except the momentum of the 
 car, the speed receives a check which constitutes the evi- 
 dence of the "buck." To aggravate matters, the short 
 circuiting of the trolley current through the fault robs 
 
CAR EQUIPMENT TESTS. 609 
 
 the car of all line power. A ground at a, usually on the 
 positive brush holder of the motor, separately excites the 
 whole field and makes the action more violent. Very 
 often on account of "crowding" the motor beyond its 
 capacity, the current will jump over from the brush 
 holder to the motor frame, but the arc does not hold ; in 
 such a case the car gives a violent kick, and, perhaps, 
 blows a fuse ; if it does not it runs on as usual until more 
 abuse is heaped on it. A grounded armature on a ground 
 return circuit will cause the car to give a succession of 
 kicks, until it is brought to a stand. The motor will have 
 to be cut out before the car can be run to the house. A 
 short-circuited armature will cause a car to buck on any 
 kind of a circuit, the violence of the bucking depending 
 upon the gravity of the short circuit. 
 
APPENDIX. 
 

APPENDIX. 
 
 TABLE I. 
 PROPERTIES OF COPPER WIRE. 
 
 
 
 
 
 RESISTANCES PER 
 
 co ! 
 
 en . 
 
 3 
 
 z < M 
 
 WEIGHTS. 
 
 i ooo FEET IN 
 INTERNATIONAL 
 
 W aH 
 
 H C* 
 
 ij-/j~ i 
 
 OHMS. 
 
 1" 
 
 5 Feet. 
 
 Mile. 
 
 At 60 F. 
 
 At 75 F. 
 
 0000 
 000 
 
 460. 211 600. 
 
 410. I 168 too. 
 
 6 4 i. 
 509- 
 
 3382. 
 2687. 
 
 .048 ii 
 .060 56 
 
 .049 66 
 .062 51 
 
 00 
 
 365. 133 225. i 403. 
 
 2 129. 
 
 .076 42 
 
 .078 87 
 
 
 
 325- 
 
 105 625. 
 
 320. 
 
 i 688. 
 
 09639 
 
 .09948 
 
 I 
 
 289. 
 
 83521. 
 
 253- 
 
 ' 335- 
 
 .1219 
 
 .125 8 
 
 2 
 
 258. 
 
 66564. 
 
 202. 
 
 i 064. 
 
 152 9 
 
 '57 9 
 
 3 
 
 229. 
 
 52441. 
 
 '59- 
 
 8^8. 
 
 .194 i 
 
 .2004 
 
 4 
 
 204. 
 
 41 616. 
 
 126. 
 
 665. 
 
 .244 6 
 
 .252 5 
 
 5 
 
 182. 
 
 33 124. 
 
 100. 
 
 529. 
 
 37 4 
 
 3'7 2 
 
 6 
 
 162. 
 
 36 244. 
 
 79- 
 
 419. 
 
 3879 
 
 .4004 
 
 7 
 8 
 
 144. 
 128. 
 
 20 736. 
 
 16 384. 
 
 63. 
 So. 
 
 &: 
 
 .491 
 .621 4 
 
 .506 7 
 -64' 3 
 
 9 
 
 114. 
 
 12 996. 
 
 39- 
 
 208. 
 
 .7834 
 
 .8085 
 
 10 
 
 103. 
 
 10 404. 
 
 32. 
 
 1 66. 
 
 9785 
 
 1. 01 
 
 ii 
 
 91. 
 
 828l. 
 
 25- 
 
 132- 
 
 1.229 
 
 1.269 
 
 12 
 
 8t. 
 
 6 5 6l. 
 
 20. 
 
 105. 
 
 1.552 
 
 i. 60 1 
 
 3 
 
 72- 
 
 5184. 
 
 '5-7 
 
 83- 
 
 1.964 
 
 2.027 
 
 '4 
 
 64. 
 
 4 096. 
 
 12.4 
 
 65- 
 
 2-485 
 
 2.565 
 
 II 
 
 57- 
 5'- 
 
 3249- 
 2 601 . 
 
 9.8 
 7-9 
 
 52- 
 42- 
 
 3.133 
 3-9U 
 
 3-234 
 4.04 
 
 17 
 18 
 
 45- 
 40. 
 
 2 025. 
 I 600. 
 
 6.1 
 
 4.8 
 
 25^6 
 
 5.028 
 6.363 
 
 5.189 
 6.567 
 
 ig 
 
 36. 
 
 1296. 
 
 3-9 
 
 20.7 
 
 7-855 
 
 8.108 
 
 30 
 
 
 I 024. 
 
 
 16.4 
 
 9.942 
 
 10.26 
 
 21 
 
 28.5 
 
 812.3 
 
 , a<5 
 
 13- 
 
 12-53 
 
 12.94 
 
 22 
 
 25-3 
 
 640., 
 
 1.9 
 
 10.2 
 
 15.9 
 
 16.41 
 
 23 
 
 22.6 
 
 510.8 
 
 1-5 
 
 8.2 
 
 19-93 
 
 20.57 
 
 24 
 
 20.1 
 
 404. 
 
 1.2 
 
 6-5 
 
 25.2 
 
 26.01 
 
 25 
 26 
 
 17.9 
 
 15-9 
 
 320.4 
 252.8 
 
 97 
 77 
 
 4- 
 
 3>-77 
 40.27 
 
 32-79 
 41.56 
 
 27 
 28 
 
 I 4 .2 
 12.6 
 
 201.6 
 
 158.8 
 
 .61 
 48 
 
 3-2 
 
 2.5 
 
 50.49 
 64-13 
 
 52.11 
 66.18 
 
 29 
 
 ".3 
 
 127.7 
 
 39 
 
 2. 
 
 79-73 
 
 82.29 
 
 30 
 
 10. 
 
 100. 
 
 3 
 
 1.6 
 
 101.8 
 
 105.1 
 
 3' 
 
 8.9 
 
 79 2 
 
 24 
 
 1.27 
 
 128.5 
 
 132-7 
 
 32 
 
 8. 
 
 64. 
 
 .19 
 
 1.03 
 
 '59-1 
 
 164.2 
 
 33 
 
 7.1 
 
 50-4 
 
 IS 
 
 .81 
 
 202. 
 
 208.4 
 
 34 
 
 6-3 
 
 39-7 
 
 .12 
 
 .63 
 
 256.5 
 
 264.7 
 
 P 
 
 5-6 
 5- 
 
 3 r -4 
 25. 
 
 2J 
 
 5 
 4 
 
 407.2 
 
 335-1 
 420.3 
 
614 APPENDIX. 
 
 TABLE II. 
 TEMPERATURE COEFFICIENTS. 
 
 Table of Temperature Variations in the Resistance of Pure Soft Copper, 
 according to Matthiessen's Standard and Formulae. 
 
 w 
 
 
 
 
 p^ c/3 W 
 
 
 
 
 Hg< 
 
 TEMPERATURE 
 
 
 
 gss 
 
 COEFFICIENT 
 
 OF 
 
 LOGARITHM. 
 
 INTERNATIONAL 
 OHMS. 
 
 sg 
 
 RESISTANCE. 
 
 
 
 
 
 I 
 
 2. 
 
 i. 
 
 1.003 876 
 1.007 764 
 
 o. 
 
 0.001 680 I 
 
 0.003 35 8 8 
 
 0.141 73 
 0.142 28 
 0.142 83 
 
 3 
 
 i. 01 1 66 
 
 0.005 036 2 
 
 o. 143 38 
 
 4 
 
 1.015 58 
 
 0.006 712 i 
 
 0.143 94 
 
 5 
 
 1-019 5 
 
 0.008 386 4 
 
 0.144 49 
 
 6 
 
 1.023 43 
 
 o.oio 059 3 
 
 0.14505 
 
 7 
 
 1.027 38 
 
 o.on 730 7 
 
 0.145 61 
 
 8 
 
 1-031 34 
 
 0.013 4 3 
 
 0.146 17 
 
 9 
 
 1-035 3i 
 
 0.015 68 3 
 
 0.146 73 
 
 10 
 
 1.039 29 
 
 0.016 734 6 
 
 o.i47 3 
 
 ii 
 
 1.043 28 
 
 0.018 399 3 
 
 0.14786 
 
 12 
 
 1.047 2 8 
 
 O.O2O 062 I 
 
 0.148 43 
 
 13 
 
 1.051 29 
 
 0.021 723 
 
 0.149 
 
 14 
 
 1-055 32 
 
 0.023 382 I 
 
 0.149 57 
 
 15 
 
 J -o59 35 
 
 0.025 039 
 
 0.150 14 
 
 16 
 
 1.06339 
 
 0.026 694 
 
 0.150 71 
 
 I 7 
 
 1.067 45 
 
 O.O28 348 
 
 0.151 29 
 
 18 
 
 1.071 52 
 
 0.029 999 
 
 0.151 86 
 
 19 
 
 1 -75 59 
 
 0.031 644 
 
 0.152 44 
 
 20 
 
 1.079 68 
 
 0.033 294 
 
 0.153 02 
 
 21 
 
 1.083 78 
 
 0.034 939 
 
 0.1536 
 
 22 
 
 1.087 88 
 
 0.036 581 
 
 0.154 18 
 
 23 
 
 1.092 
 
 0.038 222 
 
 o.i54 77 
 
 24 
 
 1.096 12 
 
 0.039 859 
 
 
 25 
 
 1. 100 26 
 
 0.041 494 
 
 0.155 94 
 
 26 
 
 1.104 4 
 
 0.043 I2 7 
 
 0-15653 
 
 27 
 
 1.108 56 
 
 0.044 758 
 
 
 28 
 
 1. 112 72 
 
 0.046 385 
 
 157 7 
 
 29 
 
 1.116 89 
 
 0.048 on 
 
 0.158 3 
 
 3 
 
 1. 121 07 
 
 0.049 633 
 
 0.158 89 
 
 40 
 
 
 
 1.163 3 2 
 1. 206 25 
 I 249 65 
 
 0.065 699 
 0.081 436 
 0.096 787 
 
 0.164 88 
 0.17095 
 0.177 n 
 
 70 
 
 1.293 27 
 
 0.111 687 
 
 0.183 29 
 
 80 
 
 1.336 81 
 
 0.126 069 
 
 0.183 46 
 
 90 
 
 1-37995 
 
 0.139 863 
 
 0.195 58 
 
 TOO 
 
 1.422 31 
 
 0.152995 
 
 0.201 58 
 
APPENDIX. 
 
 6I.S 
 
 TABLE III. 
 GALVANIZED IRON WIRE. 
 
 (Taken from John A. Roebling's Son's Wire in Electrical Construction.) 
 
 
 
 
 WEIGHTS, 
 
 RESISTANCE PER MILE 
 
 
 
 
 POUNDS. 
 
 IN OHMS. 
 
 HO 
 
 $6 
 
 Cfl 
 
 is 
 
 
 
 
 w . 
 
 H 
 
 
 
 C/3 
 
 
 
 
 
 
 D . 
 
 < 55 
 
 
 
 
 
 Z ffl 
 
 fc=Q 
 
 5"" 
 
 
 
 
 
 
 
 
 IOOO 
 
 Feet. 
 
 One 
 Mile. 
 
 Iron. 
 
 Steel. 
 
 00 
 
 
 I 
 
 
 I 
 2 
 
 340 
 300 
 
 284 
 
 304 
 237 
 212 
 
 I 607 2.93 
 I 251 3.76 
 I 121 4.19 
 
 4-05 
 5-2 
 
 5-8 
 
 2 
 
 3 
 
 259 
 
 177 
 
 932 5-04 
 
 6.97 
 
 3 
 
 4 
 
 238 
 
 149 
 
 787 5-97 
 
 8.26 
 
 5 
 
 220 
 
 127 
 
 673 6.99 
 
 9.66 
 
 4 6 
 
 203 
 
 109 
 
 573 ; 8.21 
 
 11-35 
 
 5 
 
 7 
 
 180 i 
 
 85 
 
 450 10.44 
 
 M-43 
 
 6 
 
 8 
 
 165 
 
 72 
 
 378 12.42 
 
 17.18. 
 
 7 
 
 9 
 
 148 
 
 58 
 
 305 
 
 15 44 
 
 21-35 
 
 S 
 
 10 
 
 134 
 
 47 
 
 250 
 
 18.83 
 
 26.04 
 
 8 
 
 ii 
 
 120 ! 
 
 38 
 
 200 
 
 23.48 
 
 3247 
 
 9 
 
 12 
 
 109 | 
 
 31 
 
 I6 5 
 
 28.46 
 
 39.36 
 
 ii 
 
 13 
 
 95 
 
 24 
 
 125 
 
 37-47 
 
 51.82 
 
 12 
 
 14 
 
 83 
 
 18 
 
 96 
 
 49-08 
 
 67.88 
 
 13 
 
 15 
 
 72 
 
 13-7 
 
 72 
 
 65.23 
 
 90.21 
 
 
 16 
 
 65 
 
 n. I 
 
 59 
 
 80.03 
 
 110.7 
 
 15 
 
 17 
 
 58 
 
 8.9 
 
 47 
 
 100.5 
 
 139- 
 
 16 
 
 18 
 
 49 
 
 6.3 
 
 33 
 
 140.8 
 
 194.8 
 
6i6 
 
 APPENDIX. 
 
 TABLE IV. 
 CURRENT CAPACITY FOR IRON WIRE. 
 
 (From American Electrician, March, 1897.) 
 
 NUMBER 
 B. & S. G. 
 
 SAFE 
 CURRENT 
 
 IN 
 
 WOOD 
 FRAME. 
 
 SAFE 
 CURRENT 
 
 IN 
 
 IRON 
 FRAME. 
 
 SAFE 
 CURRENT 
 
 FOR 
 
 ONE 
 MINUTE. 
 
 NUMBER 
 OF FEET 
 PER OHM. 
 
 8 
 
 17.4 
 
 20.3 
 
 43-6 
 
 250 
 
 9 
 
 14.6 
 
 17.1 
 
 36.6 
 
 i?3 
 
 10 
 
 12.3 
 
 14-3 
 
 30.8 
 
 137 
 
 ii 
 
 10.3 
 
 12 
 
 25.8 
 
 108 
 
 12 
 
 8.7 
 
 10 
 
 21.7 
 
 86.4 
 
 13 
 
 7-3 
 
 8.5 
 
 18.3 
 
 68.5 
 
 14 
 
 6.1 
 
 7 
 
 15-3 
 
 54-3 
 
 15 
 
 5-i 
 
 6 
 
 12.9 
 
 43-i 
 
 16 
 
 4-3 
 
 5 
 
 10.8 
 
 34-1 
 
 17 
 
 3-6 
 
 4.2 
 
 9.1 
 
 27.1 
 
 18 
 
 3 
 
 3-5 
 
 7.6 
 
 24-3 
 
 19 
 
 2.52 
 
 2 -9 
 
 6.3 
 
 16.5 
 
 20 
 
 2.17 
 
 2.5 
 
 5-4 
 
 13-5 
 
 21 
 
 1.82 
 
 2.1 
 
 4-5 
 
 10.7 
 
 22 
 
 1-53 
 
 1.77 
 
 3-8 
 
 8.49 
 
 23 
 
 1.28 
 
 1.49 
 
 3-2 
 
 6.73 
 
 24 
 
 i. 08 
 
 1.20 
 
 2-3 
 
 5-34 
 
APPENDIX. 
 
 6i 7 
 
 TABLE V. 
 
 FUSING EFFECTS OF CURRENTS. 
 
 Table giving the diameters of wires of various materials which will be 
 
 fused by a current of given strength. 
 
 W. H. PREECE, F. R. S. 
 
 d= ( - 
 
 i d = diameter. 
 
 -. / = current in amperes. 
 
 ( a = constant. 
 
 
 DIAMETERS IN INCHES. 
 
 z. 
 
 
 
 
 
 
 
 
 
 
 iS 
 
 
 
 
 iT 
 
 o> 
 
 
 
 
 o' 
 
 
 gg 
 
 
 
 1* 
 
 c* " 
 
 iz 
 
 c *" 
 
 00 
 
 V? 
 
 o - 
 
 s 
 
 u 
 
 8.3 
 
 |tl 
 
 X "* 
 
 c || 
 
 l\ 
 
 1 II ; . II 
 
 II 
 
 III 
 
 6 
 
 
 * 
 
 ^ 
 
 es * 
 
 4> 
 
 s s " .E * 
 
 n "3 
 
 cs 3 
 
 
 u 
 
 * 
 
 fc 
 
 C 
 
 o. ^ i 
 
 H 
 
 f 
 
 ^ 
 
 i 
 
 0.002 I 
 
 0.002 6 
 
 0.0033 
 
 0.0033 
 
 0.003 5 
 
 0.0047 
 
 o 007 2 0.008 3 
 
 0.008 i 
 
 2 
 
 0.003 4 
 
 0.004 I 
 
 0.0053 
 
 0.005 } 
 
 0.005 & 
 
 0.007 4 
 
 o.oi i 3 0.013 2 
 
 0.012 8 
 
 3 
 
 0.0044 
 
 0.005 4 
 
 0.007 
 
 0.0069 
 
 0.007 4 
 
 0.009 7 
 
 0.014 9 0.017 3 
 
 0.016 8 
 
 4 
 
 0.0053 
 
 0.006 5 
 
 0.008 4 
 
 0.008 4 
 
 0.008 9 
 
 0.011 7 
 
 O.OlS I ' 0.021 
 
 0.020 3 
 
 5 
 
 0.006 2 
 
 0.007 6 
 
 0.009 8 
 
 0.009 7 
 
 o.oio 4 0.013 6 
 
 0.021 i 0.024 3 
 
 0.023 6 
 
 10 
 
 0.009 8 
 
 0.012 
 
 0.0155 
 
 0.015 4 
 
 0.016 4 i 0.021 6 
 
 0.033 4 
 
 0.038 6 
 
 0.037 5 
 
 15 
 
 0.012 9 
 
 0.015 8 
 
 0.020 3 
 
 O.O2O 2 
 
 0.021 5 
 
 0.028 3 
 
 0.043 7 
 
 0.0506 
 
 0.049 x 
 
 20 
 
 ' 0.015 6 
 
 0.019 i 
 
 0.024 6 
 
 0.024 5 
 
 0.026 i 
 
 0.034 3 
 
 0.052 9 
 
 0.061 3 
 
 0.059 5 
 
 25 
 
 0.018 i 
 
 O.O22 2 
 
 0.028 6 
 
 0.028 4 
 
 0.0303 
 
 0.039 8 
 
 0.061 4 
 
 0.071 i 
 
 0.069 
 
 30 
 
 \ 0.020 5 
 
 0.025 
 
 0.032 3 
 
 0.032 0.034 2 
 
 0.045 
 
 0.069 4 
 
 0.080 3 
 
 0.077 9 
 
 35 
 
 1 0.022 7 
 
 0.0277 
 
 0.035 8 
 
 0.015 6 
 
 0.037 9 
 
 0.049 8 
 
 0.076 9 
 
 0.089 
 
 0.0864 
 
 4 
 
 0.024 8 
 
 0.030 3 
 
 0.039 * 
 
 0.038 8 
 
 0.041 4 
 
 0.054 5 
 
 0.084 
 
 0.097 3 
 
 0.0944 
 
 45 
 
 0.026 8 
 
 0.032 8 
 
 0.042 3 
 
 0.042 
 
 0.044 8 
 
 0.058 9 
 
 0.0909 
 
 0.105 2 
 
 0.102 I 
 
 50 
 
 0.028 8 
 
 0.035 2 
 
 0-045 4 
 
 0.045 
 
 0.048 
 
 0.063 2 
 
 0.097 5 
 
 0.1129 
 
 0.109 5 
 
 60 
 
 0.032 5 
 
 0.039 7 
 
 0.051 3 
 
 0.050 9 
 
 0.054 2 
 
 0.071 4 
 
 O.IIO I 
 
 0.127 5 
 
 0.123 7 
 
 70 
 
 0.036 
 
 0.044 
 
 0.056 8 
 
 0.056 4 
 
 0.060 i 
 
 0.079 
 
 0.122 
 
 0.141 3 
 
 0.1371 
 
 80 
 
 0-039 4 
 
 0.048 i 
 
 0.062 i 
 
 0.061 6 
 
 0.065 7 
 
 0.0864 
 
 - I 334 
 
 0.154 4 
 
 0.1499 
 
 9 
 
 0.042 6 
 
 0.052 
 
 0.067 2 
 
 0.066 7 
 
 0.071 i 
 
 0.093 5 
 
 O.J443 
 
 0.167 I 
 
 0.162 i 
 
 100 
 
 120 
 
 0-045 7 
 0.051 6 
 
 0.055 8 
 0.063 
 
 0.072 
 
 0.081 4 
 
 0.071 5 
 0.0808 
 
 0.076 2 
 0.086 I 
 
 o.ioo 3 
 0.113 3 
 
 0.1548 
 0.1748 
 
 0.179 2 
 O.2O2 4 
 
 o.i739 
 0.1964 
 
 140 
 
 0.057 2 
 
 0.069 8 
 
 O.OOX) 2 
 
 0.0895 
 
 0.0954 
 
 0.125 5 
 
 0.193 7 
 
 0.224 3 
 
 0.217 6 
 
 160 
 180 
 
 0.062 5 
 0.067 6 
 
 0.0763 
 0.082 6 
 
 0.098 6 
 o. i 06 6 
 
 0.097 8 
 0.105 8 
 
 0.1043 
 0.112 8 
 
 0.1372 
 0.1484 
 
 0.2II 8 
 
 0.229 l 
 
 0.245 2 
 0.265 2 
 
 0.2379 
 0.257 3 
 
 200 
 
 0.072 5 
 
 0.0886 
 
 0.114 4 
 
 0.1135 
 
 o.iai 
 
 0.1592 
 
 0-245 7 
 
 0.284 5 
 
 0.276 
 
 225 
 
 0.078 4 
 
 0.095 8 
 
 '0.123 7 
 
 O. 122 8 
 
 0.1309 
 
 O.I72 2 
 
 0.265 8 
 
 0.307 7 
 
 0.2986 
 
 250 
 
 0.084 * 
 
 O. IO2 8 
 
 0.132 7 
 
 0.131 7 
 
 0.1404 
 
 0.184 8 
 
 0.285 * 
 
 0.3301 
 
 0.3203 
 
 275 
 
 0.0897 
 
 0.1095 
 
 0.141 4 
 
 0.1404 
 
 0.149 7 
 
 0.1969 
 
 0.3038 
 
 0.3518 
 
 0.341 7 
 
 300 
 
 0.095 
 
 0. I I'l I 
 
 0.149 8 
 
 0.148 7 
 
 0.1586 
 
 0.208 6 1 
 
 0.322 
 
 0.3728 
 
 0.361 7 
 
6i8 
 
 APPENDIX. 
 
 <D y 
 
 tf 
 
 ? "3 
 2^3- 
 
 B K- (B 
 
 IP. 
 
 g.0.0 
 
 - 
 
 < O ST. 
 
 ^ 3 
 
 3 en S 
 K- " o 
 P 5^ 
 
 ? 3 
 
 
 
 
 O OO^J Ut W 
 
 o ~j w 
 
 No. B. & S. G. 
 
 00000 
 00000 
 
 o S - So N 
 
 00 OMO 
 
 00000 
 
 85i a& 
 
 00 ONUt * H 
 
 Diameter 
 in inches. 
 d. 
 
 O 
 
 88888 
 S%j?ti& 
 
 M H M <JJ U) 
 
 00000 
 
 o b b b b 
 |o\oo\K 
 
 OsOOtn H o\ 
 
 ^b\-2^ 
 
 fc. 
 
 (fijee 
 
 M ^ O31JI U) 
 
 fe-b 1 ^ 
 
 Copper, 
 a = 10 244. 
 
 O O -^ OOUl 
 
 ^?&s 
 
 Aluminum, 
 a = 7 585. 
 
 O>vi v) 5 5 
 
 (0 OJ (Jl OO M 
 ^. tn 4>. ^ ^) 
 
 Platinum, 
 a 5 172. 
 
 Os<l vo W -5 
 
 (0 W Ul 00 M 
 
 German Silver, 
 a = 5 230. 
 
 ut o O M ON 
 
 (0 W Ln ^J O 
 (0 10 O <J 00 
 
 Platinoid, 
 a = 4 750. 
 
 *. A ON 00 O 
 
 M to OJ Ln ^J 
 Ol M OJ M IH 
 
 Iron, 
 a = 3 148. 
 
 K) to W * tn 
 
 OO H VI V) ^J 
 
 Tin, 
 a = i 642. 
 
 M 10 10 U) -^ 
 
 o\o i S'o 
 
 Tin-lead Alloy, 
 a = i 318. 
 
 ^ M . w ^ 
 
 O\O tn K) M 
 
 Lead, 
 a = i 379. 
 
 00 
 
 s 
 
 ft) 
 
 M 
 O 
 
 g n 
 
 O > 
 
 ^ td 
 o r 1 
 a w 
 
 i 
 
APPENDIX. 
 
 TABLE VII. 
 DYNAMO TESTING RECORD. 
 
 Dynamo Class Volts no load Volts full load... 
 
 Machine No Frame No Spool No 
 
 Armature No Rheostat Type No 
 
 Trimmings Plate No Location Date 
 
 619 
 
 
 ^0 
 
 ?c 
 
 ^c 
 
 Volts Field. 
 Ins. No.... 
 
 d* 
 II 
 
 1 
 
 1 
 
 n 
 
 Remarks. 
 
 
 COLD TEST--Time 
 
 
 
 
 
 
 
 7V MI f> 
 
 
 S'o load, field rheo. all out 
 No load field rheo all in 
 
 
 
 
 
 
 
 Room 
 
 |-...*C. 
 
 No load, normal voltage 
 
 
 
 
 
 
 
 
 
 Full load, no G. S. Shunt, field 
 rheo. left as in previous readings 
 
 Put miscellaneous readings on 
 back of this sheet, 
 
 COMPOUNDING TEST HOT. 
 Time 
 
 
 
 
 
 
 
 
 
 No load 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 i-ull 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 iJo load 
 
 
 
 
 
 
 
 
 
 HOT TEST Time 
 
 
 
 
 
 
 
 Temp. 
 
 ' C, 
 
 No load, normal voltage 
 
 
 
 
 
 
 
 Room 
 
 \ 
 
 No load, field rheo. all out .. 
 
 
 
 
 
 
 
 
 
 No load, field rheo. all in 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Temperature C. after hours run at Volts Amps. 
 
 Armature * Surface Air ducts Commutator 
 
 ( End connections Spider 
 
 Spools, by increase of resistance .By Therm Frame 
 
 Resistance by bridge ] Shunt Coil Cold Hot .'. 
 
 < Series Coil Cold Hot 
 
 Insulation R. of machine Size of Shunt 
 
 Insulation R. of Spools Tested by 
 
 Insulation R. of Armature. . . 
 
620 
 
 APPENDIX. 
 
 TABLE VIII. 
 DYNAMO TESTING RECORD. Continued. 
 
 REMARKS. 
 
 MISCELLANEOUS READINGS. 
 
 Size of G. S. Shunt. 
 
 Volts. 
 
 Amp. 
 
 Volts 
 Field. 
 
 Amp. 
 Field. 
 
 Speed. 
 
 Pts. 
 Rheo. 
 
 
 
 
 
 
 
 
 
 
 
INDEX. 
 
 Adding tools to a loaded motor, 
 496 
 
 Alternating current circuits, 
 grounds on, 456, 457, 459 
 
 Ammeter, calibration by volt- 
 ameter, 77; construction of 
 shunt for, 99; graduation of, 
 93; measuring resistance of 
 grounds by, 447; several in 
 multiple, 100; shunts used 
 with, 95 
 
 Ampere, basis of, 70; definition 
 of, 14; hour. 15; international, 
 17; measurement by copper 
 voltameter, 71 
 
 Ampere's principle, 20 
 
 Arc dynamos, Brush, 260; brush 
 regulation, 238; regulation, 
 249; in series and multiple, 
 247; Thomson-Houston, 250; 
 Westinghouse, 274 
 
 Arc lamps, 247 
 
 Armature, 25, 26; bar to bar 
 test, 151; cross connecting, 
 469; drop in, 281; high volt- 
 age in low voltage fields, 505; 
 insulation measurement, 194; 
 location of faults, 149, 151; lo- 
 cation of grounds, 152, 242, 
 243; "lost volts," 282; reac- 
 tion and compounding, 315; 
 reaction in series machine, 233; 
 resistance and efficiency, 284; 
 
 rewinding, 237; Siemens, 27, 
 28; Siemens defect of, 28; 
 Siemens E. M. F. of, 28; test 
 for open and short circuit, 241 ; 
 winding and temperature limit 
 295; artificial cooling, 296; 
 grounds in street-car type, 
 6or; open circuits in, 603; size 
 of wire on street-car type, 524; 
 car motor type of, 554 
 Astatic needle, 123. 
 
 B 
 
 Back induction, 34 
 
 Baked field coils, 556, 592 
 
 Barlow's wheel, 3, 4, 23 
 
 Battery, amalgamation, 120; 
 Daniells cell, 115; efficiency, 
 118; E. M. F., 119, 120; in- 
 ternal resistance, 119, 211; 
 Leclanche cell, 121; maximum 
 activity, 117; maximum cur- 
 rent, 116; maximum current 
 with given resistance, 118 
 
 Brake action of series motors in 
 parallel, 603 
 
 Brush test, 433; brush regula- 
 tion on arc dynamos, 238; on 
 shunt dynamos 291 
 
 Brush arc dynamo, 260; arma- 
 tures, 262; brushes, 271; com- 
 mutator, 264, 272; controller, 
 266; dial, 274; E. M. F. 263; 
 field spools, 260; regulator, 
 
 621 
 
622 
 
 INDEX. 
 
 264, 268; starting up, 269; 
 switches, 270; troubles of, 
 272; under light load, 271 
 Bucking of car motors, 608 
 Burning out a ground, 466 
 
 Calomel cell, 136 
 
 Car brake shoes, 587 
 
 Car control, modern controller, 
 545; old style rheostat, 543; 
 series parallel, 545; use of field 
 shunts and loops, 542 
 
 Car controller, abuse of, 546; 
 General Electric Go's., 547; 
 tests of, 561 
 
 Car circuit breaker, 518 
 
 Car, failure to start, 572 
 
 Car motors, 554; lubrication of, 
 586 
 
 Car, starting coils, 542; starting 
 up first time, 572 
 
 Car switches, 518 
 
 Car wheels, spinning of, 588 
 
 Car wiring diagram, 573 
 
 Car wiring, grounds and shorts, 
 572; locating faults in, 572, 
 579; tests of, 572, 579; wrong- 
 ly connected, 584 
 
 Cell (see battery) 
 
 Centimeter, 6 
 
 Choke coil, street-car type, 533 
 
 Circuit breaker, street-car type, 
 518; cleaning contacts of, 522; 
 testing for faults in, 520 
 
 Clark cell, 136 
 
 Commutator, 27; insulation of, 
 200 
 
 Compensating coil for galvano- 
 meter shunt, 98 
 
 Compound-wound machine, 303, 
 314; and armature reaction, 
 315; compounding factors in, 
 308; troubles in, 400; drop on 
 series coils, 328 ; German silver 
 
 shunt, 321, 323; hit and miss 
 method of compounding, 318, 
 324; limits to, 316; practice of, 
 318; series coils on large and 
 small dynamos, 329; and speed 
 323; compounding volts per 
 revolution, 324 (see also motor 
 generator test); direction of 
 rotation as dynamo and motor, 
 303; introducing into circuit, 
 335J over compounding, ' 320, 
 322; running in multiple, 337, 
 340; running in series, 342 
 
 Compound-wound motor, break- 
 ing shunt field on, 494; speed 
 regulation, 492; speed with 
 differential connections, 492; 
 and shunt board, 493, 494 
 
 Conductivity, specific, 61 
 
 Controller, 'for street cars, 538, 
 545 
 
 Copper voltameter, directions for 
 use, 72; formulae for, 74; plate 
 form, 71; spiral coil form, 75 
 
 Core-loss test, 419 
 
 Coulomb, 15 
 
 Counter E. M. F., 37 
 
 Critical speed on series machine, 
 
 303, 314 
 Cross connecting armatures and 
 
 commutators, 469 
 Cross induction, 33 
 Current, definition of, 14 
 
 Dead rails, 578 
 
 Differential galvanometer, 174 
 
 Differential connections, tests 
 for, 492 
 
 Direction of rotation dynamos 
 and motors, 300 
 
 Distribution test for E. M. F., 
 Mordey's method, 429; Swin- 
 burn's method, 431; Thomp- 
 son's method, 432 
 
INDEX. 
 
 6*3 
 
 Dynamo armature, 25, 26; as 
 motors, 54; efficiency test, 433; 
 field of, 24; losses in, 32; of 
 different type run together, 
 
 305 
 
 Dynamometer, Siemens, 104,106 
 Dyne, 6 
 
 Eddy current test, 423 
 
 Edison dynamo field connections, 
 236 
 
 Efficiency, commercial, 35, 42; 
 electrical, 34, 40; at maximum 
 activity, 40; maximum com- 
 mercial, 43 
 
 Efficiency test of dynamo, 433; 
 of motor, 487; by Prony brake, 
 488 
 
 Electromagnetism, 18 
 
 E. M. F., definition, 14, 29; and 
 potential difference, 112; and 
 cost of field winding, 283; dis- 
 tribution test, 429, 431, 432; 
 and motor speed, 471; low, by 
 opposed dynamos, 507; low, by 
 opposed fields, 505, 506; regu- 
 lation on shunt machine, 282; 
 sources of, 114 
 
 Energy, 4; laws of, 5 
 
 Equalizing bar, theory of, 331 ; 
 running without, 335: on com- 
 pound-wound dynamos, 337; 
 on series machines, 246 
 
 F 
 
 Faults in armatures, 149, 151 
 Faults, location of, in car wir- 
 ing, 572, 579 
 
 Field coils, car motor type, 555; 
 locating baked coils, 592; use 
 of shunt and loop, 542; wrong- 
 ly connected, 556, 594 
 
 Field connections on Edison dy- 
 namo, 236; excitation methods 
 29, 30, 31 ; excitation voltage 
 and walls consumed, 498; sat- 
 uration test, 426; test for open 
 and short circuit in, 239; re- 
 sistance and efficiency, 284; 
 shunt, watts expended in. 284 
 Friction losses in dynamo, 423 
 Fuses, copper for street car 
 work, 524; deterioration of. 
 529; proper size of, 524; mag- 
 netic blow-out for, 527 
 
 Galvanometer, astatic needle for, 
 123; differential, 174; direct- 
 ing magnet, 124. instrumental 
 constant, 85; insulation for 
 keys, 128; proportion box, 129, 
 131; proportion lines, 128; 
 range, 90; range varied by re- 
 sistance box, 86; shunt, 86; 
 requisites for, 122; resistance 
 boxes and moisture, 132; set- 
 ting up, 83, 88. 134; shunt and 
 compensating coil, 98; shield 
 of iron rings. 124; taking rapid 
 readings, 127; tangent, prin- 
 ciple of, 78; wiring of, 133 
 
 Gears, bushings for, 569; for 
 street car motors, 567; test- 
 ing. 570 
 
 Gram unit of mass, 6 
 
 Grounds in street car armatures, 
 601; in street car wiring, 579 
 
 Grounds, detector for alternating 
 circuits, 456; telephone, 457; 
 Stanley static, 459; lamps and 
 bell, 440: delusive, 441; volt- 
 meter, 442; methods of locat- 
 ing, 460, 466; measuring re- 
 sistance of, ammeter method, 
 447, 452; voltmeter method, 
 445, 453. 455 
 
624 
 
 INDEX. 
 
 II 
 
 High voltage and line loss, 207 
 Horse power, definition, 16 
 Hysteresis, 423 
 
 I 
 
 Induction, mutual and self, 23 
 Insulation, galvanometer and 
 grounds, 203, 205; high volt- 
 age tests, 207; of long lines, 
 443; measurement of, first 
 method, 183; second method, 
 185; third method, of insula- 
 tors, 188; fourth method, elec- 
 trometer, 193; fifth method, of 
 armatures by voltmeter, 194; 
 sixth method, by galvanometer, 
 203; bar to bar test, 179; of 
 commutator, 200; of marine 
 cables, 191; of underground 
 cables. 192; and temperature, 
 202; and voltage, 201; test in 
 motor generator test, 409; test 
 cases, portable, 210 
 
 J 
 
 Joule, definition of, 16 
 
 K 
 
 Kicking coils, for use with car 
 lightning arresters, 533 
 
 Lamp bank, care in handling, 
 
 354 
 
 Leclanch6 cell, 121 
 Lightning, nature of discharge, 
 
 531 
 
 Lightning arresters, 531; adjust- 
 ment of air gap, 534, 538; 
 choke coils for use with, 533; 
 General Electric Go's, type of, 
 
 534 
 Line loss and high voltage, 207 
 
 M 
 
 Machines of different type run 
 together, 305 
 
 Magnetic field, 8; meridian de- 
 termination of, 125 
 
 Magneto, detecting, and locating 
 grounds by, 461 
 
 Manganin, 219 
 
 Matter, 4 
 
 Motor-generator tests, brushes 
 for, 386; changing over with 
 compound-wound machines, 
 372, 405; compounding, 398; 
 data, hot, 404; engine as loss 
 supplier, 356; field connections 
 380, 381; field polarity, 383, 
 397; full test with details, 378; 
 free data, 396; German silver 
 shunt, 404; load precipitated, 
 reasons for, 400; load failure 
 to go on, reason for, 401; 
 machines of different current 
 capacity, 366, 373, 374; ma- 
 chines of different current 
 capacity and E. M. F., 374; 
 putting on load, 383, 385, 397; 
 reversal when starting up, 372; 
 rocker arm, loose effect of, 
 406; series field, test for, 388; 
 sudden removal of -load, 406, 
 417; voltage to start multipolar 
 motors, 387; with lamp bank, 
 3591 with street car-motors, 
 481; with three machines, 354, 
 411 
 
 Motors, belting up for shop work, 
 495; classification, 50,468; dif- 
 ferential, 53; helping over- 
 loaded engine, 496-498; prin- 
 ciples of, 36; proper diameter 
 for pulley, 497; separately ex- 
 cited, 50; series, 51; shunt, 50; 
 speed, 49, 52, 471 ; armature for 
 street car type, 554 ; connecting 
 up on car, 565; field coils for 
 street car type, 556; grounds 
 
INDEX. 
 
 62.5 
 
 on, 602; mounting on car, 570; 
 
 street car types, 552; tests of 
 
 street car type, 563; trial runs, 
 
 566, 570 
 Multi polar motors running on 
 
 two brushes, 470 
 Mutual induction, 23 
 
 O 
 
 Oersted, 17 
 
 Ohm, 13 
 
 Ohm's law, 15, 59; and C. E. M. 
 
 F., 64; self-induction, 65 
 Open circuit, locating, 114; in 
 
 armature, 149, 151, 241; field, 
 
 239 
 
 Thomson's slide bridge, 166; 
 high, by differential galvanom- 
 eter, 175, 179; of grounds by 
 ammeter, 447; of grounds by 
 voltmeter, 453; Wheatstone's 
 bridge, 156, 167; box bridge, 
 160; slide bridge, 159; inter- 
 polation 173; of multiple cir- 
 cuits, 101, 102 
 
 Resistance specific, 221, 224; 
 table, 223; and temperature 
 coefficient, 220; water rheo- 
 stat, 214 
 
 Retentivity, 20 
 
 Rheostats, caution in using, 508 
 
 Running two stations in multiple 
 343 
 
 Permeability, 10; table, 12 
 
 Pole strength, unit of, 7 
 
 Portable test cases, 210 
 
 Prody brake, 488 
 
 Proportion box and galvanom- 
 eter, 129; and direct reading 
 scale, 131 
 
 Proportion lines and galvanom- 
 eter, 128 
 
 R 
 
 Radiating surface and weight, 
 296 
 
 Rating of dynamos and motors, 
 508 
 
 Residual magnetism, 20 
 
 Resistance, 13, 69, 145; boxes 
 and moisture, 132; liquid, 213; 
 liquid measurement; by tele- 
 phone, 216; measurement of 
 burning lamp, 154; method of 
 constant deflection, 156; by 
 differential galvanometer, 174; 
 low, by comparison of poten- 
 tial, 147; low, differential gal- 
 vanometer, 175, 179; low, 
 Vienna method, 153; low, 
 
 Saturation test, 426 
 
 Self-exciting dynamos, trouble 
 on, 234 
 
 Self induction. 23 
 
 Separate excitation and sparking 
 507; and variation of E. M. F. 
 with current, 508 
 
 Separately excited machine di- 
 rection of rotation, 305 
 
 Series machines, 231; arc lamps 
 run in multiple and series, 247; 
 armature reaction, 233; critical 
 speed, 232, 244: direction of 
 rotation as dynamos and motor 
 302; flashing at commutator, 
 233; motor tendency to race 
 on no load, 476, 479; proper 
 speed and sparking, 245; re- 
 moval from service, 244; run 
 in multiple and series, 246, 348 
 
 Series motor efficiency test, 487; 
 locating faulty motor, 490, 491 ; 
 run in series, 500 
 Shop efficiency, 500 
 
 Short circuit locating in arma- 
 ture, 241; in field, 239 
 
 Shunt, theory of, 81; board on 
 
626 
 
 INDKX. 
 
 compound-wound motors, 493, 
 494; box, 83; and series motors 
 run in series, 500, 503; stand- 
 ard, 91, 92; used with am- 
 meters, 95 
 
 Shunt machine, 279; direction of 
 rotation as dynamo and motor, 
 300; danger of reversal when 
 multiple, 300; E. M. F. regula- 
 tion, 282; by brushes, 291; 
 effect of adding a dynamo to a 
 multiple circuit, 310; field rheo- 
 stat, 288, 297; and E. M. F., 
 290, 292; loss in, 294; test for 
 connections, 298, 299; for high 
 voltage, 304; motor and start- 
 ing box, 472, 479; on short cir- 
 cuit, 288; putting into multiple 
 circuit, 306, 308; putting in 
 multiple circuit to be watched, 
 309; run in multiple, 300; series 
 298; and series motors, speed 
 on removal of load, 480; signs 
 of reversal, 312; sparking upon 
 reversal, 313; sudden removal 
 of load, 289; taking from mul- 
 tiple circuit, 310, 311 
 
 Shunt winding substituting for 
 series, 475, 476; and tempera- 
 ture limit, 295 
 
 Shunts for field coils, 542 
 
 Single machine compounding on 
 water box, 375 
 
 Slide wire bridge, 159 
 
 Speed, 49 
 
 Standard cell, Clark's 136; 
 Daniell's 115 
 
 Stanley static ground detector 
 459 
 
 Starting box selection, 473 
 
 Starting coils, construction of, 
 542; continuity tests, 560; de- 
 terioration of, 539; mounting 
 on car, 542; tests of insulation 
 560 use and abuse of, 538 
 
 Storage batteries and station 
 
 efficiency, 347 
 Switches, car, 518 
 
 Temperature coefficient for cop- 
 per, 217; table appendix; by 
 rise of resistance, 408; and in- 
 sulation, 202; limit in arma- 
 ture and field, 295 
 
 Testing low voltage machines, 
 351; machines below their nor- 
 mal voltage, 363 
 
 Tests of street car wiring, 572, 
 
 579 
 
 Thomson-Houston arc dynamo, 
 250; air blast, 254; brushes, 
 254; causes of flashing, 253; 
 cut-out, 258; E. M. F., 252; 
 grounds on armature, 257; 
 grounds on field, 256; grounds 
 on regulator, 256; induction 
 due to, 252; lead of armature, 
 258; method of controlling 
 sparking, 253; regulator, 251; 
 /'.anting the field, 258 
 Torque, definition of, 44 
 Trolley, characteristics, 515; de- 
 scription of parts, 511; troubles 
 with, 516; lubrication of parts, 
 513; pole, adjustment of, 516; 
 pole, straigthening, 513; stand, 
 511; spring, 517; wheel, bear- 
 ings, 514; wheel life of, 514; 
 wheel, tests of different makes, 
 515; wheel, wearing of. 516 
 
 U 
 
 Underground circuits, grounds 
 
 on, 462, 463 
 Unipolar motors, 468 
 Units, absolute and practical, 6; 
 
 length, mass, time, force, 6; of 
 
 pole strength, 7 
 
INDEX. 
 
 627 
 
 Volt, definition of, 14 
 
 Voltage, regulation by opposed 
 dynamos, 393 
 
 Voltmeter, calibration of, 138; 
 calibration coil, 139; calibra- 
 tion box method, 142; Poggen 
 dorf's method, 140; measuring 
 current by, 103; measuring 
 resistance of grounds by, 445, 
 453; measuring insulation by, 
 194; use of multipliers with, 
 143 
 
 W 
 
 Watt, 15; hour, 16; meter, 109 
 meter recording, no 
 
 Westinghouse arc dynamo, 274; 
 armature, 275; commutator, 
 277; fields, 274, 275, 277; reg- 
 ulation, 275, 278 
 
 Wheatstone's bridge, general re- 
 marks, 167; interpolation, 173 
 theory, 156 
 
 Wiring, street car, 573 
 
 Work, Joule unit of, 16 
 
THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 AN INITIAL FINE OF 25 CENTS 
 
 WILL BE ASSESSED FOR FAILURE TO RETURN 
 THIS BOOK ON THE DATE DUE. THE PENALTY 
 WILL INCREASE TO SO CENTS ON THE FOURTH 
 DAY AND TO $1.OO ON THE SEVENTH DAY 
 OVERDUE. 
 
 
 ICLF 
 
 IN) 
 
 
 
 Wf 28, 99 ., 
 
 
 FEB 24 1938 
 
 OCT 
 
 MAR 8 1941 M 
 LIBRARY USE 
 
 SEP 15 ;ggg 
 
 NOV2 J975X 
 
 "038 
 
 LD 21-50m-8,-32 
 

 
 YC 19844