, INTERNAL COMBUSTION ENGINES INTERNAL COMBUSTION ENGINES / THEIR THEORY, CONSTRUCTION AND OPERATION BY ROLLA C. CARPENTER, M.M.E., LL.D. AND H. DIEDERICHS, M.E. PROFESSORS OF EXPERIMENTAL ENGINEERING, STBLEY COLLEGE, CORNELL UNIVERSITY WITH 379 ILLUSTRATIONS SECOND EDITION REVISED NEW YORK D. VAN NOSTRAND COMPANY 23 MURRAY AND i 9 o 9 27 WARREN STS. Copyright, 1908, by D. VAN NOSTRAND Co. The Plimpton Press Norwood Mass. U.SA. PREFACE THE intention of the authors in the preparation of this book has been to present in as simple terms as possible the fundamental and theoretical principles relating to the internal combustion engine, and to describe the various methods- of applying these principles to practical construction. The book does not in any way treat of the proportioning and the strength of the various machine parts. The general treatment of the subject is indicated by the various chapter headings. Thus the first five chapters relate to definitions and theoretical considerations, the subjects being as follows: CHAPTER I. DEFINITIONS AND CLASSIFICATION. CHAPTER II. THERMODYNAMIC PRINCIPLES. CHAPTER III. THEORETICAL DISCUSSION OF VARIOUS CYCLES. CHAPTER IV. THEORETICAL CYCLES MODIFIED BY PRACTICE. CHAPTER V. THE TEMPERATURE-ENTROPY DIAGRAM. In the discussion on theoretical cycles in Chapter III, very little reference has been made to cycles not in actual use. The cycles are considered principally with reference to their practical application and any danger of confusing the mind of the student by a multiplicity of theoretical cycles of no practical value is avoided. The main idea of Chapter IV is to show how the lines of the real cycles differ from those of the theoretical cycles laid down in the previous chapter, and to discuss briefly the reasons for such difference. The five chapters following, VI to X inclusive, take up the phenomena of combustion, the various gas-engine fuels, and the formation and properties of the fuel mixture. Thus, Chapter VI treats of combustion in general and discusses the most important properties of the gases usually found in gas-engine practice. Tables are given embodying the important constants for many of 179717 vi PREFACE these gases, and the chapter ends with some type computations on fuel mixture and exhaust, gas constants which may serve as a guide for similar work by the student. The question of gas-engine fuels is treated in the next three chapters by dividing all the fuels into three classes: the solid, the liquid and the gas fuels. Chapters VII and VIII take up the first two of these classes. Broadly speaking, neither class of fuel is directly available for gas-engine work, hence it was thought desirable to show in these chapters also the means by which these fuels are rendered available. Accordingly, Chapter VII discusses producer gas and describes the construction and operation of the most important types of producers, while Chapter VIII con- sists largely of a description of the various vaporizing devices used for crude oil, gasoline, kerosene, and alcohol. The latter fuel, recognizing its growing importance, has been treated in some detail. Chapter IX relates to industrial gases, pointing out briefly the method of their manufacture and giving the most important gas constants. In Chapter X, after discussing the fuel mixture and its most important properties, an attempt has been made to collaborate the most important experiments on the variation of the specific heat of the fuel gases with a view to ascertaining the present state of our knowledge on this point. Chapter XI gives a brief outline of the historical development of the gas engine, the various types being described only where they are of importance in connection with the development of modern forms. This is followed in Chapter XII by an extended description of the most important forms of internal combustion engines found in the market at the present time. The aim of this chapter has been not only to show how the various manu- facturers have solved what is fundamentally the same problem, but also to familiarize the student, by means of a large number of illustrations, with the main constructive features of the gas engines at the disposal of the man requiring power. Two of the important problems connected with the gas engine are the questions of ignition and of governing. The former is taken up in detail in Chapter XIII. This chapter, after mention- ing briefly other methods of ignition, concerns itself mainly with ignition by electric spark, as being the most important method PREFACE vii used to-day. The chapter also takes up briefly two other gas- engine auxiliaries: mufflers and starting apparatus. Chapter XIV treats the governing problem by discussing first the principles of the various systems of governing employed, and afterward shows the mechanical details of the governors used. It is beyond the scope of this book, however, to treat of the prin- ciples of governor design. Chapter XIV discusses various methods in use for determining the necessary cylinder dimensions of a gas engine to develop a certain given power, or conversely, to determine the probable power for a given engine. The most reliable of these methods appears to be based upon the necessary charge volume for the given power. This method, the authors believe, was originally due to Giildner, and was adapted from that author's hand-book " Entwerfen und Berechnen der Verbrennungsmotoren." For the determination of the power of automobile engines, two additional semi-empirical formulae are given, and the results of the computations by the various methods are compared by means of a type example. The remaining three chapters of the book treat of what might be called the economic side of the internal combustion engine. Chapter XVI takes up the methods used in/ the testing of gas engines. The rules followed in this country will be of course the code laid down by the American Society of Mechanical Engi- neers. The Code of the German Society of Engineers, however, is appended, because it gives additional information upon some of the points involved and because it treats in greater detail of gas producers. The results of tests on engines and producers are discussed in Chapter XVII. The various factors affecting economy are taken up in somewhat greater detail than has been done in any of the previous chapters. Tables are given showing the results of numerous tests and these should prove valuable in furnishing a guide as to what may be expected of other installations in the matter of fuel economy. Finally, while Chapter XVII treats only of the question of fuel economy, Chapter XVIII takes up the entire financial problem relating to the gas engine. It shows in brief, as far as the infor- mation is available, the probable capital cost of the installation, viii PREFACE the cost of erection, the operating expenses, etc. It is shown most strongly that the question of fuel cost is not always the im- portant item of the problem, a point which is often lost sight of in discussions regarding the comparative merits of various prime movers. It must be confessed that reliable information regard- ing some of the factors in this financial problem is still very scarce, owing, of course, directly to the comparative youth of the gas engine. It should also be pointed out that many so-called com- parisons between various prime movers, as between steam and gas, are often based upon hypothetical assumptions that _ fit only the particular case under discussion, and any generalization of the results obtained often leads to serious misrepresentation. The book is of course largely compiled from different sources and is in the main an outgrowth of a course of lectures on the internal combustion engine delivered to students of Sibley College for the past three years. Acknowledgments are due to numerous writers upon the subject for the facts and statements presented. The acknowledgments have generally been made in the body of the book, but the authors desire to extend herewith special acknowledgment to the following European authorities: Messrs. H. Giildner, Dugald Clerk, Bryan Donkin, and Aime Witz, and to Professor C. E. Lucke of Columbia University and Mr. F. E. Junge of New York. Thanks are also due to various manufacturers for kindly fur- nishing the text and illustrations of Chapter XII. ITHACA, N. Y., April 23, 1908. PREFACE TO THE SECOND EDITION IN presenting the second edition of this book 'the authors wish to express their obligation to all those who kindly aided in pointing out errors, and especially to Professor H. P. de Schweinitz of Lehigh University, for a thorough revision of the first two chapters. February 1, 1909. TABLE OF CONTENTS CHAPTER I INTRODUCTION, DEFINITIONS AND CLASSIFICATION, INDICATED AND BRAKE HORSE-POWER PAGE Mechanical Work 1 Heat and Temperature 3 Thermometers and Pyrometers 6 Specific Heat 12 Heat Unit and Mechanical Equivalent of Heat ....... 15 Entropy 15 Classification of Engines 17 The Steam Engine 18 Hot Air Engines 22 Classification of Internal Combustion Engines 27 The Engine Indicator . 32 Indicated and Brake Horse-power 38 Forms of Indicator Diagrams 40 CHAPTER II THERMODYNAMICS OF THE GAS ENGINE Characteristics of Perfect Gases 45 Relation of Heat Transmission to Changes of Volume and Pressure . 48 Transformation to Different States 51 Work of Isothermal and Adiabatic Expansion 53 Relation of Heat and Entropy 54 Second Law of Thermodynamics 54 Graphical Relations 55 Comparison of Theoretical and Actual Heat Engines 61 CHAPTER III THEORETICAL COMPARISON OF VARIOUS TYPES OF INTERNAL COM- BUSTION ENGINES. Theory of the Constant Volume, Beau de Rochas or Otto Cycle ... 65 Theory of the Constant -Pressure or Bray ton Cycle, the Diesel Cycle . 69 ix X TABLE OF CONTENTS PAGE Comparison of Various Cycles 73 Conditions affecting the Choice of Best Cycle 79 CHAPTER IV THE VARIOUS EVENTS OF THE CONSTANT-VOLUME AND CONSTANT- PRESSURE CYCLE AS MODIFIED BY PRACTICAL CONDITIONS The Four-Stroke or Otto Cycle 84 The Suction Stroke 84 The Compression Stroke . 87 The Combustion Line, Typical Indicator Diagrams 90 The Expansion Line 97 The Exhaust Stroke 100 The Two-Stroke Cycle . . . . ' 102 The Constant-Pressure Cycle . . 104 CHAPTER V THE TEMPERATURE ENTROPY DIAGRAM APPLIED TO THE GAS ENGINE General Relations Involved 107 Mathematical Construction of the Entropy Diagram 112 Interpretation of the Diagram 119 Graphical Method of Constructing the Entropy Diagram 120 CHAPTER VI * COMBUSTION Perfect Gases 126 Combining Weights and Volumes, Combustion, Heating Value, Air Required, etc 127 Calorimeters 129 Tables of Constants and Typical Example of Gas Computations . . 138 CHAPTER VII GAS-ENGINE FUELS, THE SOLID FUELS, GAS PRODUCERS The Production of Air Gas 147 The Production of Water Gas 147 The Production of Producer or Power Gas 149 Gas Producers in Practice 156 Classification of Producers 158 Description of Pressure Producers 159 Description of Suction Producers 165 Description of Combination Producers 173 Some Producer Details 175 TABLE OF CONTENTS xi CHAPTER VIII THE GAS-ENGINE FUELS, LIQUID FUELS: CARBURETERS AND VAPORIZERS PAGE Crude Oils and their Distillates, Gasoline, Kerosene, Alcohol . . . 178 Mixing Devices for Liquid Fuels 185 Description of Various Types of Vaporizers and Carburetars for Gaso- line, Kerosene, Crude Oil and Alcohol 186 Conditions required for Proper Gasification of Alcohol 205 CHAPTER IX GAS-ENGINE FUELS, THE GAS FUELS Illuminating Gas 206 Oil Gas 207 Coke Oven Gas 208 Blast Furnace Gas 209 Aeetylene 211 Water Gas 212 Natural Gas 212 Table of Constants for Above Gases .213 CHAPTER X THE FUEL MIXTURE, EXPLOSIBILITY, PRESSURE AND TEMPERATURE Explosibility and Explosive Ranges 215 Pressure and Temperature after Combustion, Experiments of Clerk, Langen, etc 220 Velocity of Flame Propagation and Time of Explosion 227 CHAPTER XI THE HISTORY OF THE GAS ENGINE Origin . 232 Period of Speculation and Invention 232 Period of Development 239 Period of Application 257 CHAPTER XII MODERN TYPES OF INTERNAL COMBUSTION ENGINES General Features of Design 263 Gas Engines : Small and Medium Sized Engines 265 Large .Gas Engines 304 xii TABLE OF CONTENTS Liquid Fuel Engines : PAGE Stationary Gasoline Engines 358 Marine Gasoline Engines 361 Automobile Gasoline Engines 372 Oil Engines 375 Alcohol Engines 390 CHAPTER XIII GAS ENGINE AUXILIARIES IGNITION, MUFFLERS AND STARTING APPARATUS Ignition: Ignition by Open Flame . 392 Ignition by Hot Tube 394 Ignition by Heat of Compression 396 Ignition by Electric Spark 397 Make-and -Break Ignition 397 Jump Spark Ignition 401 Timers 405 Spark Plugs 406 Auxiliary Spark Gap 409 Relative Advantages and Disadvantages of the two Systems of Electric Ignition ...'.. 410 Sources of Current . 410 Chemical Generators 410 Primary Cells 410 Storage Cells 411 Mechanical Generators 415 Dynamos and Magnetos 416 Methods of Connecting up Primary and Secondary Batteries . . . 420 High -Tension Distributors 423 Mufflers 426 Starting Apparatus 428 CHAPTER XIV REGULATION OF INTERNAL COMBUSTION ENGINES General Considerations 439 Systems of Governing 441 The Hit-and-Miss 442 Quality Governing 444 Quantity Governing 446 Combination Systems 447 Governing by Timing of Spark 449 Governing of 2-Cycle Engines 449 Mechanical Details of Governors . 450 Pendulum or Inertia Governors for Hit-and-Miss Regulation . . 450 TABLE Of 1 CONTENTS xiii PAGE Mechanical Centrifugal Governors for Hit-and-Miss Regulation . . . 454 Governors for Quality Regulation 457 Governors for Quantity Regulation 460 Governors for Combination Systems 462 Governing Details of 2-Cycle Engines 467 CHAPTER XV THE ESTIMATION OF POWER OF GAS ENGINES Limits of Piston Speeds and Rotative Speeds 471 Method of Determining Power by Assuming Mean Effective Pressure: Grover's Formula 472 Method of Determining Power by Calculating Mean Effective Pressure from Tables of S. A. Moss 472 Method of Determining Power, or Cylinder Dimensions, by Giildner's Method 477 The Power Rating of Automobile Engines 483 CHAPTER XVI METHODS OF TESTING INTERNAL COMBUSTION ENGINES Rules for Conducting Tests of Gas and Oil Engines 486 A. S. M. E. Code of 1901 487 Rules for Testing Gas Engines and Gas Producers Code of the German Society of Engineers . .511 CHAPTER XVII THE PERFORMANCE OF GAS ENGINES AND GAS PRODUCERS The Performance of Engines as affected by: Cooling Water Conditions and Piston Speed 530 Compression 532 Varying Fuel Mixture 533 Point of Ignition 534 Engine Economy Depending upon Load 537 The Heat Balance ; ... 539 Results of Tests on Engines and Producers 542 Table of Engine Tests 544 Table of Tests of Producers and Producer Plants 546 CHAPTER XVIII COST OF INSTALLATION AND OF OPERATION Cost of Producers and Engines 547 Cost of Erection 548 xiv TABLE OF CONTENTS PAGE Piping and Auxiliaries 548 Floor Space and Buildings 549 Cost of Operation 553 Interest 553 Depreciation 553 Insurance 553 Fuel Costs 553 Cost of Water for Cooling and Washing . . 561 Oil and Waste '. . 564 Attendance 565 Maintenance and Repair 567 Total Operating Costs, and Costs as Compared with other Prime Movers 568 CHAPTER I INTRODUCTION, DEFINITIONS, CLASSIFICATION, AND FORM OF INTERNAL COMBUSTION ENGINES, INDICATED AND BRAKE HORSE-POWER i. Mechanical Work. Work is done when resistance is overcome; it is measured by the product of the resisting force and the distance through which that force is moved. If one pound is lifted one foot high in opposition to the force of gravity, a quantity of work, measured by the product of one pound by one foot, is performed, which quantity is known as a foot-pound, and is the unit of measurement for mechanical work in countries where the pound is a unit of weight and the foot a unit of dis- tance. If 20 pounds are lifted 15 feet the work performed would be similarly 20 X 15 foot-pounds = 300 foot-pounds. In countries where the metric system is used mechanical work is measured by the product of the resisting force in kilo- grams (2.2046 pounds) multiplied by the distance in meters (3.2808 feet); the product is expressed in kilogr ammeters (7.233 foot-pounds). The foot-pound or kilogrammeter is a gravity measure which depends on the intensity of the force of gravity at the place, and varies with that force. The variation is, however, so slight for different positions on the earth's surface that for all practical engineering work no sensible error is produced by considering it a constant quantity. The unit of measurement usually employed by engineers for expressing the rate of work, or the quantity of work done in a given time as one second or one minute, is the horse-power, which has been arbitrarily defined as equivalent to 550 foot- pounds per second or 33,000 foot-pounds per minute. This quantity is considerably greater than the power a horse can exert, at least for any considerable length of time; it was first 1 \> INTERNAL COMBUSTION ENGINES used by James Watt in defining the power of the steam engine and has been established by long use as a definite measure of power. In France the term Force de Cheval is applied to a rate of work of 75 kilogrammeters per second (542 J ft. Ibs.) or 4500 kilogrammeters per minute (32549 ft. Ibs.). In general, if W be the work performed against the pressure or resisting force p while moving through the space or volume v, W = pv. (1) Work is done when force is applied so as to produce motion in the direction of action of the force, and also when force is employed in changing the velocity of a body already in motion. The latter condition is of considerable practical importance and can be considered as follows: Suppose a body whose mass is M be moving in a certain direction with a velocity u, and let a force exerting a momentum P be applied in the direction of motion, required to find the effect produced by this force acting through the small time t, during which the body moves through the distance v, and has at the end of the time the velocity u'. The momentum produced by the force in one unit of time is P, and in t units of time it is Pt. Since this is equal to the increase of momentum produced, we have Pt = M (u' - u). As the distance is equal to the mean velocity multiplied by the time, we have v = %(u' + u) t. By multiplying the above equations, Pvt = J (Mu' 2 - Mu 2 ) t; dividing by t, Pv = i (Mu' 2 - Mu 2 ) (2) . Pv is the mechanical work done in overcoming a resistance; the expression J Mu 2 is the kinetic energy. From this it is seen that the mechanical work done is measured by the increase in .the kinetic energy produced. The mechanical work done by a fluid during a change of volume from v to v f is equal to the mean resistance overcome, or INTRODUCTION, DEFINITIONS, ETC. 3 pressure exerted, p, multiplied by the change of volume. That is, in general, W = f v pdv (3) W = p(v' - v) In the operation of an engine, the working fluid expands and contracts as the piston moves forward and backward, and in one or more revolutions returns to its initial condition, so far as pressure, volume and temperature are concerned, and then passes through the same stages of expansion and contraction as before. The period through which these changes take place is termed a cycle. The work performed in a cycle would be equal to the mean pressure, p, exerted, multiplied by the total volume, v r , swept through; that is W - = pv'. (4) Mechanical work can be represented by a diagram in which the pressure exerted or resistance overcome, p, is represented by the ordinates, and the volume v by the abscissa. Such a diagram is called a pressure-volume diagram; its area is equal to fpdv, and is proportional to the work performed. Thus in Fig. 1-1, if the distances parallel to OY represent the pressure at any given point, and the distances parallel to OX the corresponding volume, then will the total work done in chang- ing from the highest to the lowest pressure and from least to greatest volume be represented by the area of the figure a b d e /. 2. Heat. Heat is a peculiar form of energy; it may be generated by the application of mechanical work, the amount so produced being exactly proportional to the mechanical energy which disappears. Conversely, mechanical work may be done by the action of heat, and for every foot-pound of work so done a definite amount of heat is put out of existence. Heat is also produced by a form of chemical action known as combustion, during which operation fuels are burned. 3. Temperature. The temperature of a body is defined by Maxwell * as " its thermal state with reference to its power of communicating heat to other bodies." * Theory of Heat. 4 INTERNAL COMBUSTION ENGINES A body transmitting heat to another is at a higher tempera- ture and is said to be hotter; conversely, one receiving heat is at a lower temperature and is said to be colder. Heat flows from a hotter to a colder body, but not conversely, 100 -Ibs. 01- 0.0 0.1 0.2 0.3 O.I cu. ft. FIG. 1-1.- Pressure-volume or Work Diagram. and the rate of flow increases with the difference of temperature, although probably not exactly in the same ratio. The difference of temperature thus causes a flow of heat in a manner somewhat smiliar to that caused by a difference of pressure in the case of water. , The terms hotter and colder are relative ones commonly applied to distinguish substances having relatively a higher or lower temperature. It should be noted that temperature is that property of heat which refers to its intensity or transmission power also, that heat energy may exist at different tempera- tures, and, furthermore, in one condition may be much colder than in another. INTRODUCTION, DEFINITIONS, ETC. The following scales of temperatures are in common use in which the temperatures, of freezing and boiling water under a barometric pressure of 29.92 inches are taken as points of reference. The Centigrade scale was introduced by Celsius, professor of astronomy in the University of Upsala about 1742; in it the freez- ing-point is marked degrees and called zero, and the boiling- point is marked 100 degrees. The simplicity of dividing the distance between the points of reference into 100 parts and call- ing each of them a degree has caused it to be generally adopted along with the Metric System for scientific use, especially on the Continent of Europe. The other scales are called by the names of those who introduced them. Fahrenheit of Dantzig, about 1714, introduced a thermometer scale in which the freezing-point was marked 32 degrees and the boiling-point 212 degrees, the space between the reference points being divided into 180 equal parts called degrees, and the graduation extended above and below the points of reference. A point 32 degrees below freezing was called zero. Despite the inconvenience of the scale of the Fahrenheit thermometer it is in general use by English-speaking people for commercial and business purposes, and for that reason will be used principally in this treatise. Reaumur introduced a thermometer scale about 1730 in which the freezing-point is marked degrees and the boiling- point 80 degrees, which is used to some extent on the Continent of Europe for medical purposes. The following table gives the comparative value of the three thermometric scales: THERMOMETRIC SCALES Fahrenhei t Centigrade Reaumur Degrees between freezing and boiling . . Assumed temperatures at freezing-point 180 32 100 80 Assumed temperatures at boiling-point Comparative length of a degree 212 1 100 1.80 2.25 Comparative length of a degree f 1 To transform into absolute tempera- ture add 460 273 218 6 INTERNAL COMBUSTION ENGINES 4. Absolute Temperature is an expression for the value of temperature measured from an ideal point called the absolute zero, which is assumed to be the lowest possible point on any scale of temperature. The position of the absolute zero can be calculated by the law of expansion of a perfect gas, which is expressed by the simple equation in which p = the pressure, v = the volume of a given mass of gas, T = the absolute temperature, and R = a constant which varies only with the different kinds of gas. This equation can be considered as the characteristic equation of a permanent gas, from which T can be computed if p, v, and R are known, which is the case with most of the gases. It is evident that the value of one degree of absolute tempera- ture can be taken at pleasure as equal either to that on the Cen- tigrade or Fahrenheit scale. The exact location of the position of absolute zero is some- what in doubt since it is determined by the relative expansion of air, nitrogen, or hydrogen under a constant pressure; these gases are not perfect gases and the expansion in volume per degree of increase in temperature may not be exactly the same as for a gas which could not be liquefied for any conditions of pressure or temperature. Preston in his work on the Theory of Heat states that the most trustworthy observations indicate that the absolute temperature of freezing water is 273.14 Centigrade, which would correspond to 491.65 Fahrenheit. It is sufficiently near for all practical purposes to consider the temperature of freezing water on the absolute scale as 273 degrees Centigrade or 492 degrees Fahrenheit, and these numbers will be used in this treatise in reducing to the absolute scale. From this it is seen that to reduce to the absolute scale it is necessary to add to . the temperature, if expressed in degrees, Fahrenheit 460, or if in degrees Centigrade 273. 5. Thermometers. Instruments for measuring temperature are called Thermometers. The expansion of a gaseous, liquid or solid body under con- INTRODUCTION, DEFINITIONS, ETC. stant pressure is almost, if not exactly, proportional to the increase of temperature, estimated from absolute zero. In the thermometer in common use the tempera- ture is measured by the expansion of mercury, con- fined in a glass tube, from which the air has been exhausted. Such a thermometer is quite satisfactory within the range of temperature through which mer- cury will remain liquid. In the better grades of mer- curial thermometers the graduations are cut with extreme care directly on the stem. The glass is care- fully selected and is permitted to season or age until molecular changes have stopped before graduation. The general appearance of such thermometers is shown in Fig. 1-2. When thermometers are likely to be used in temperatures which would send the mercury column above the limits of the graduations, it is desirable to have an extra bulb, called a safety-bulb, at the top, to prevent breaking from overheat- ing. Mercurial thermometers can be used from a temperature about 40 degrees below zero to 600 degrees above zero Fahrenheit. By filling the space above the mercury with some neutral gas as N or CO 2 under pressure the upper limit may be raised some hundred degrees; but as the melting-point of glass is low, FlG - 1 ~ 2 - the upper limit can scarcely ever exceed 800 degrees to 900 degrees Fahrenheit. METALLIC THERMOMETERS in which the expansion of a metal, or the difference in expan- sion of metals of two different kinds, is multiplied FIG. 1-3. Metallic Pyrometer. Mercurial Thermom- eter. by a system of levers so as to move a hand over a dial are fre- quently used for the measurement of temperature. Such ther- mometers are sometimes called pyrometers. An illustration of such a thermometer is shown in Fig. 1-3. The metallic thermometer can be used for temperatures not exceeding 1200 degrees to 1500 degrees Fahrenheit, but it is sel- 8 INTERNAL COMBUSTION ENGINES dom an instrument of accuracy and is extremely liable to acci- dent. The scale of these instruments should be frequently com- pared with the boiling-point, and adjusted if not found correct. It has already been shown that air or permanent gases like nitrogen and hydrogen when under a constant pressure will ex- pand in volume in proportion to the absolute temperature, or when confined so as to have a constant volume will increase in pressure in proportion to the absolute temperature. It follows from this that if air be maintained at a constant volume and heated, its absolute pressure will increase with the absolute temperature, or vice versa, if it be maintained at con- stant pressure, its volume will vary with absolute temperature. The air thermometer constructed in accordance with either principle is used as a standard way of measuring temperature, but because of the extreme difficulty of maintaining constant pressures or constant volumes it is an awkward instrument to use and is employed very little in the ordinary measurement of temperature. The Jolly form of constant volume air thermom- eter is shown in Fig. 1-4. The leg C F has a flexible connection at the bottom and may be raised to maintain a constant .volume of air from the bulb L to the line A B. The increase of pressure is measured by a scale attached to C G. ELECTRICAL THERMOMETERS. Temperature may be measured by elec- trical thermometers, of which there are two classes. In one class a conducting circuit is formed of two different metals, such a construction being frequently termed a thermo- element, a number of these connected together is known as a thermopile. In this construction an electro-motive force is pro- duced which is proportional to the difference of temperature of the junctions and may be measured by a sensitive galvanometer. If one of the junctions be maintained at a constant or known temperature, the temperature of the other may be computed from the reading of the galvanometer. For the measurement of FIG. 1-4. Jolly Air Ther- mometer. INTRODUCTION, DEFINITIONS, ETC. 9 high temperature, metals having a high melting-point, such as platinum and a platinum-iridium alloy, may be used for the ele- ments. The LeChatelier pyrometer is an instrument of this class; FIG. 1-5. The LeChatelier's Electrical Pyrometer. it consists of a galvanometer connected to a thermo- element B as shown in Fig. 1-5 and is extensively used; the thermo-element is constructed of plati- num and platinum-rhodium and is enclosed in a porcelain tube, as shown in Fig. 1-6. The Bristol pyrometer which may be had with a recording device belongs to the above class. Another class of electrical thermometers is based on the law of increase of electric resistance of metals due to the rise of tem- perature. With this class of thermometers the difference in temperature can be determined from the measurement of the drop in potential for a known current passing through a coil. This method is employed in the platinum thermometers of Sie- mens, Calendar, and various others. The electrical thermom- eters are superior to all others for many uses. OPTICAL PYROMETERS. The approximate temperature of in- candescent bodies may be determined by the color of the radiant rays. Pouillet, as the result of a large number of experiments, concluded that all incandescent bodies have a definite and fixed 10 INTERNAL COMBUSTION ENGINES color corresponding to each temperature, as shown in the follow- ing table: Color Temp. C. Temp. F. Faint red 525 927 Dark red 700 1292 Faint cherry 800 1482 Cherry 900 1652 ' Bright cherry 1000 1932 Dark orange 1160 2120 Bright orange White heat 1200 1300 2192 2372 Bright white 1400 2552 Dazzling white 1500 2732 The fixed relation between color and temperature is due to the fact that the color of an incandescent body varies with the wave length which is a function of the tempera- ture. A number of optical pyrometers have been devised which determine the temperature by the appearance of the heated body. The Mesure and Noel pyrometer changes the wave lengths by the rotation of the plane of polariza- tion of light passing through a quartz plate cut perpendicularly to its axis. In the use of the instrument the temperature is measured by noting the angle through which the analyzer is turned in order to produce a lemon yellow color. The Morse thermo-gage, which is extensively used in the steel industry, consists of an incan- descent lamp with a rheostat arranged so that the current flowing through it and its conse- quent brightness may be regulated. When the FIG. 1-6. The ^ m of the incandescent lamp becomes the same Le Chatelier color as that of the object the temperature is Thermo-Ele- computed from the reading of a milli- volt meter arranged to measure the current. The tempera- ture corresponding to a given electrical reading is determined by calibration. The optical pyrometers are convenient and of approximate INTRODUCTION, DEFINITIONS, ETC. 11 accuracy in determining high temperatures of incandescent bodies. They are of no value in determining temperatures of combustible objects. VAPOR THERMOMETERS. -- The pressure produced by a satu- rated vapor confined in a closed vessel increases with the tem- perature, in accordance with a known law or a law which may be determined. By providing a suitable pressure gage and attaching it to a closed vessel of the proper shape the tempera- ture may be obtained from the pressure readings, the value of which are known, or from the dial readings of the pressure gage, which may be graduated by trial into degrees of temperature. An instrument of this type is. made by Schaeffer and Budenberg and called by them a Thalpotasimeter. CALOMETRIC THERMOMETER. The temperature can also be measured by calometric methods, by heating a body of known weight and specific heat to the temperature which it is desired to measure, then transferring this body with as little cooling as possible to a vessel containing a known weight of water. The equation for this operation will be expressed as follows: W (t -O = ws (t x - t), in which W = equivalent weight of water and its containing vessel. w = weight of heated body, s = specific heat of hot body. f = original temperature of water. t = final temperature of water. t x = temperature of hot body. A ball of platinum, copper, porcelain, or burned fire clay answers nicely for the body to be heated. FUSION THERMOMETERS. The temperature can be approxi- mately determined from the known melting-points of metallic bodies, on the principle that the temperature will be higher than the melting-point of a body that melts, and lower than the melting temperature of one that does not melt. In place of metallic bodies a series of fusible clay cones called " Seger cones," whose melting-points are known, are often employed in the same manner. 12 INTERNAL COMBUSTION ENGINES 6. Specific Heat. Different materials of the same weights have different capacities for absorbing heat for a corresponding change of temperature; thus one pound of water will absorb about nine times as much heat as one pound of wrought iron for the same change of temperature. This peculiar heat capacity of bodies compared with water is termed specific heat, which is usually defined as follows: The specific heat of a body is the ratio of the quantity of heat required to raise that body one degree in temperature, to the quantity required to raise an equal weight of water at standard temperature one degree. The specific heat of water is not quite constant, being nearly three fourths of one per cent higher at the boiling and freezing temperatures than at fifteen degrees Centigrade; this is shown by the following table from the book of Steam Tables by Professor Peabody : SPECIFIC HEAT OF WATER Centigrade Fahrenheit Specific Heat 5 32 41 1.0072 5 10 41 50 1.0044 15 20 59 68 1.00 2025 68 77 0.9984 2530 77 86 0.9948 30 35 8695 0.9954 4045 104 113 1.00 45155 113311 1.008 155200 311 392 1.046 7. Specific Heat of a Permanent Gas. In general the conditions under which the change of temperature occurs should be distinctly specified, for the temperature of a body may be varied by the mechanical work done in the compression or ex- pansion which occurs. The change of volume due to increase of temperature is so small for solids and liquids that the external work in change of volume may be neglected, but such is not the case for gases. For this reason the conditions under which the heating of a gas takes place must be stated when referring to its specific heat, and it has become customary to speak of two spe- INTRODUCTION, DEFINITIONS, ETC. 13 cific heats in connection with any gas, namely, the specific heat at constant volume, and the specific heat under constant pressure. The former is the quantity of heat required to raise the tempera- ture of a unit mass of the gas one degree when its volume is kept constant, and the latter the quantity of heat required to raise the temperature of a unit mass one degree when the pressure is kept constant, compared with that of a unit mass of water. In the first case the pressure increases while the volume is kept constant and no external work is done; in the latter the volume increases under constant pressure, and an amount of external work is done which is measured by the product of the pressure by the change of volume. The specific heat of a gas at constant volume bears a known relation to the specific heat at constant pressure, so that if one be determined experimentally the other may be computed. It will be shown later that the specific heat at constant pressure equals the specific heat at constant volume plus a constant which depends on the nature of the gas. That is, C p = C v + R when expressed in heat units, and K p = K v + JR when expressed in foot-pounds. The following table gives the specific heats of constant pres- sure and volume of the principal gases : TABLE OF SPECIFIC HEATS Gas e# C v 7 Gas Cp C v 7 H N 3.4090 0.2175 0.2438 2.4177 0.1543 0.1729 1.41 1.41 1.41 CO CO 2 CH 4 0.2479 0.2169 0.5929 0.1738 0.157 1.41 1.29 Air HsO( vapor) 0.2375 0.4805 0.1684 0.3585 1.41 1.34 C 2 H 4 0.4040 0.320 1.26 In the above table C p = Specific heat of constant pressure. C v = Specific heat of constant volume. 7 = C-, + C., The specific heat of a perfect gas is considered constant by most English writers who discuss the subject. Quite a number 14 INTERNAL COMBUSTION ENGINES of experiments have shown, however, that it varies with both the pressure and the temperature. In the French work by A. Witz the following formula is given for the change of specific heat with pressure: C p = a -f b (p 1), in which a and b are constants for each gas and p the pressure in kilos per square centimeter. The specific heat of all vapors which are liquefied at moderate temperature undoubtedly increases with the increase of tempera- ture. This subject is fully discussed in Chapter X of this work. In this work the specific heat of gases will be considered con- stant unless otherwise mentioned. Any error caused by such consideration will not usually be serious, and by so doing the various formulas which express the heat capacities under actual working conditions are much simplified. The specific heat of a mixture of various gases constituting a known weight or mass is equal to the mean specific heat of the mixture, and is found by multiplying the weight of each component part by its specific heat and dividing the sum of the products by the total weight. 8. The Heat Unit. Heat is measured by its capacity to raise the temperature of a known weight of water. The unit of measurement is termed a calorie in the metric system, and a British thermal unit (B. T. U.) in the English system. A calorie is commonly defined as the heat required to raise one kilogram of water from the freezing-point to one degree Centigrade, and a British thermal unit (B. T. U) that required to raise one pound of water from 32 to 33 degrees Fahrenheit. Because of the varia- tion in the specific heat of water near the freezing-point Professor Peabody in his Tables of Saturated Steam defines the thermal unit as that required to raise the temperature from 62 to 63 degrees Fahrenheit, or from about 15 to 16 degrees Centigrade. The specific heat of water changes slightly for different tem- peratures as already noted, but it can be considered as constant without sensible error for all ordinary purposes in the measure- ment of heat when the temperature is maintained between the freezing and boiling points. It is noted from the above statement that heat is measured not by the temperature alone but by its ability to heat a mass of water from one temperature to another ; the total heat expressed INTRODUCTION, DEFINITIONS, ETC. 15 in thermal units being equal to the product of the weight of water by the change of temperature. Thus if 20 pounds of water be heated 25 degrees Fahrenheit, the heat required is the product of 20 times 25 = 500 B. T. U. 9. Mechanical Equivalent of Heat. The mechanical equivalent of heat is the amount of work expressed in mechanical units which may be performed by the transformation of one heat unit into mechanical work. The value of the mechanical equiva- lent of heat was determined experimentally by Joule who found that 772 foot-pounds was equivalent to one B. T. U. or 425.6 kilo- grammeters to one calorie. The determinations made later and with more accurate instruments by Rowland, reduced to the sea level and to 45 degrees of latitude, give the following values which are now generally adopted: Expressed in calories, J = 426.9 kilogrammeters. Expressed in B. T. U., J = 778 foot-pounds. In this work J is used as the symbol for the mechanical equivalent of heat, and A as the reciprocal of J. That is, J =- If A. The experiment by means of which the equivalent value of the heat unit was determined in units of mechanical work serve to prove the general mechanical principle of the conservation of energy, which had been previously stated by Clausius as follows: 11 In all cases where work is produced by heat, the quantity of heat consumed is proportional to the work done; and conversely, by the expenditure of the same amount of work the same quantity of heat may be produced.'' This principle is often called the first law of Thermodynamics. 10. Entropy. One of the qualities or properties of heat which cannot be measured by any simple physical apparatus is termed "Entropy." This same quality was named by Rankine "the Thermodynamic Function"; its value is such that its change during a given time multiplied by the absolute temperature equals the total heat which may be transformed into mechanical work. From this definition it is noted that the product of abso- lute temperature by change of entropy is the measure of the capacity of heat for performing mechanical work. As an illustration, if a gaseous body under pressure be allowed 16 INTERNAL COMBUSTION ENGINES to expand without receiving or giving off heat its entropy would remain constant and any mechanical work performed would be done at the expense of the heat existing in the body. A change which takes place without gain or loss of heat is termed adiabatic, which condition corresponds to that of constant entropy. Expansion or compression of a body taking place without change of temperature is called isothermal expansion or compres- sion, and lines drawn in a diagram indicating this property are termed " isothermal" lines. Change of entropy with respect to heat is in many respects analogous to change of volume in respect to mechanical work; it has already been shown that the mechanical work is equal to the change in volume multiplied by the resistance or pressure overcome. If we denote the total heat capable of being trans- formed into mechanical work by Q, the change of entropy by and the absolute temperature by T , we shall have the following equations : ^ , N W (v v) p. Q = (*- ') T. From the first of the above equations it is noted that no mechanical work can be done without a change in volume, and further, that the amount of work done is measured by the change of volume multiplied by the pressure. From the second it is seen that no change in the amount of heat which a body contains can take place without a change in entropy, for when = <', Q = 0. The amount of heat transferred is measured by the change of entropy multiplied by the absolute temperature. From the above , W v -v' = - p From which is seen that the change in volume is equal to the mechanical work performed divided by the mean pressure. From which it is seen that the change in entropy is equal to the total heat transformed into work divided by the absolute temperature. INTRODUCTION, DEFINITIONS, ETC. 17 11. Classification of Engines The action of an engine is, in general, to produce motion against a resistance or to per- form work. Engines are popularly classified in accordance with the nature of the working fluid, as hydraulic engines, steam engines, gas engines, oil engines, etc. They may be more scientifi- cally classified in accordance with the nature of the working process as pressure engines and heat engines. In the pressure engine work is produced by change of pressure without change of temperature, as illustrated in the piston water engine. In the heat engine work is produced by transforming heat into mechani- cal work, which process is accompanied with change of tempera- ture, and usually, also, a change of pressure. In mechanical structure engines are of two classes, recipro- cating and rotary. In the reciprocating engine a piston free to move in a cylinder is pushed backward and forward by alternate changes in pressure of the fluid against either face. The recipro- cating motion of the piston is converted into a continuous rotary motion by a mechanism usually consisting of crank and fly-wheel which will be described later. In the rotary engine a rotary motion is directly produced by the force due to pressure, impulse, or reaction acting upon revolving blades or pistons arranged in a suitable casing. The term is often confined to a structure with a revolving piston in which motion is produced by a difference of pressure, whereas the term turbine is applied to the structure when rotation is produced by impulse or reaction of the jet. In a general way the turbine is a species of rotary engine. This treatise will be principally confined to internal combustion engines having a piston with reciprocating motion. Engines are classified as single acting when the propelling force is 'applied to one side of the piston only, &nd as double acting when it is applied alternately to both sides. Engines are classified as simple, compound, triple, expansion, etc., depending on the number of cylinders through which the working fluid passes in succession as it expands from highest to lowest pressure. 12. Classification of Heat Engines. Heat available for use in a heat engine is usually produced by a species of chemical action termed combustion. Heat engines may be classified in accordance with the location of the place of combustion with 18 INTERNAL COMBUSTION ENGINES respect to the working cylinder as external combustion engines and internal combustion engines. The external combustion engines include steam and other vapor engines, hot air engines, and some forms of gas or oil engines; the internal combustion engines include all the usual forms of gas and oil engines in which the fuel is consumed in the working cylinder. In this work the term gas engine will be frequently used as including all forms of internal combustion engines adapted to burn gas or vapor irrespective of the nature of the fuel. Of these various types of engines the steam and the gas engines are the only ones of practical commercial importance at the present time. The hot air engine was built extensively about fifty years ago, and its theory was thoroughly investigated at that time. It failed as a commercial machine because of the high cost of repairs and operation; it is principally useful at the present time as illustrating the practical application of certain thermodynamical principles. As the steam engine and hot air engine have a practical bear- ing on the internal combustion engine, a short description is inserted. 13. The Steam Engine. The mechanism of the steam engine and its mode of operation should be familiar to all students of the internal combustion engine. The term steam engine is used here in its broad sense, including the boiler, the engine proper, and all the accessories necessary for its operation. In the mode of operation of the steam engine, steam is produced at any de- sired pressure by the combustion of fuel in a furnace beneath the boiler, which latter is a strong closed vessel containing a certain amount of water and into which water is introduced by a feed pump as desired. The steam engine proper is pro video! with a cylinder in which is fitted a piston which is propelled by the steam pressure acting on one or both sides. The admission and discharge of the steam are controlled by a valve or valves moved by the mechanism of the engine, the form of which varies greatly with different types. In the more common form, a slide valve is used which is propelled backward and forward by a valve rod moved either by an eccentric or by a short crank attached to the main shaft. The valve is operated so as to admit steam at nearly boiler pressure back of the piston for a portion of the stroke, and INTRODUCTION, DEFINITIONS, ETC. 19 then to cut off communication with the boiler, after which the piston is pushed forward by the expansive force of the steam. The valve is moved at the end of the stroke so as to open com- munication between the cylinder and the exhaust pipe, which in the case of a non-condensing engine discharges into the atmos- phere and in case of a condensing engine discharges into a more or less perfect vacuum. The motion of the piston is communicated, by means of a piston rod which slides through a stuffing-box at the end of the cylinder, to a block called a cross-head which moves in guides, from which motion is communicated by a connecting rod to the crank of the main shaft so as to produce rotary motion. S' FIG. 1-7. Double Acting Steam Engine. The train of. mechanism of the ordinary double-acting steam engine which is used to communicate motion from the piston to the ily-wheel is shown in Fig. 1-7, in which P represents the piston, p the piston rod, H the cross-head, C the connecting rod, M E the crank, E the crank pin, F the wrist pin, M the main shaft, G the fly-wheel, S the stuffing-box, which is used to prevent leakage of steam around the piston rod. The cross-head moves in guides K which direct its motion. One end of the connect- ing rod E has a circular motion, the other a rectilinear motion. In the view referred to the valves which admit steam alternately to the ends of the cylinder are not shown. A view of a two-cylinder single-acting engine is shown in 20 INTERNAL COMBUSTION ENGINES Fig. 1-8 in section. In this engine the steam is admitted only at one end of the cylinder at the proper intervals of time by the sliding motion of the piston valve V, which is operated by the FIG. 1-8. Two-cylinder Single-acting Steam Engine. mechanism of the engine by means of a small crank attached to the main shaft and the various link 'connections shown. Steam engines are frequently built compound, in which case the steam works in succession in a small and large cylinder, termed respectively the high-pressure and low-pressure cylinder. These cylinders may be arranged side by side as in the view of the single-acting engine, Fig. 1-8, or they may be arranged in tandem as shown in Fig. 1-9. In the tandem compound engine shown FIG. 1-9. Tandem Compound Steam Engine. INTRODUCTION, DEFINITIONS, ETC. 21 the supply of steam is admitted, to the high-pressure cylinder by the slide valve V operated from the small crank E f ', and to the low-pressure cylinder by the slide valve V operated from the eccentric E. Since the application of pressure to the piston as described above is periodic and not constant, a fly-wheel consisting of a heavy mass of metal must be applied as shown, to produce uniform motion. To regulate the speed a governor is used, which is constructed so that the variation in centrifugal force either tends to cut off the supply of steam or to close the steam valves if the speed become higher than desired, or the reverse. An indicator or pressure volume diagram of a steam engine is shown in Fig. 1-35. In this diagram vertical distances are pro- portional to pressures per square inch acting on the piston, and horizontal distances to the space passed through by the piston. The diagram shows the relation of pressure and volume at any point. The work done on the piston is proportional to the area of the diagram. The steam engine, as will be shown later, does not realize in the work performed as great an efficiency on the basis of the heat values of the fuels employed as the gas engine; but it has the advantage of being adapted for the burning of solid fuels under its boiler without converting the same into gas, and this in a measure sometimes compensates for the greater heat losses. The steam pressure is exerted on the piston for a much larger portion of the stroke than an explosion in a gas engine piston, and as a consequence the inertia of moving parts is depended upon less for uniform speed than in the gas engine, as will be afterwards shown. In the past the steam engine has had a great advantage over the gas engine, due to the fact that it has been more reliable in operation and could produce a more nearly uniform motion. The development of the gas engine has, however, removed in a large measure such defects, and at the present time there is no great difference in respect to reliability and regulation on the part of these two classes of engines. The use of a producer in which gas can readily be made, from solid fuels also equalizes any advantages which the steam engine has had from that source. 22 INTERNAL COMBUSTION ENGINES 14. Hot Air Engines. The hot air engine is principally of importance to-day for its scientific value. Its actual com- mercial use is confined to pumping small quantities of water under favorable conditions. It is of scientific interest because it is the only heat engine yet produced which represents almost perfectly the standard ideal cycle of Carnot, with which the operation of nearly all heat engines is compared. In the hot air engine a mass of air is successively heated and cooled; during the time that it is heated either its volume or pressure increases, and during the time it is cooled the reverse operation takes place. Mechanical work is performed by the change of volume or pressure which is utilized for moving a piston. The principal varieties of air engines may be classified by the following distinctive features: 1. Change of temperature at con- stant pressure. 2. Change of temperature at constant volumes. 3. Heat received and rejected at a pair of constant pressures. Ericsson's engine, best known as the caloric engine, may be taken as an example of the first class. In this en- gine, air is admitted from the atmosphere to the compressing pump at the lowest working tempera- ture, and compressed, the temperature being main- tained constant by the action of some refrigerat- ing apparatus. The air when compressed enters a receiver. It is then ad- mitted to the working cyl- || inder, being heated on its passage t6 the higher tem- perature, so that its vol- FIG. 1-10. Ericsson Hot Air Engine. ume is increased and the pressure remains constant under the movement of the piston, then expands with its temperature main- tained constant at the higher limit, and is finally expelled into the atmosphere, giving up its heat* to the regenerator, to be used in heating the volume of air next introduced. INTRODUCTION, DEFINITIONS, ETC. 23 This engine is represented in Fig. 1-10. B is a working cylin- der, placed over the furnace H. This cylinder consists of two parts; the upper part in which the piston works is accurately turned, and the lower part in which the air receives heat from the furnace is less accurately made. A is the piston of the cylin- der, consisting of an upper part which is accurately fitted and pro- vided with metallic packings so as to work air-tight in the upper part of the cylinder. The lower part is somewhat smaller than the cylinder, is hollow, and filled with brick dust, fragments of fireclay, or some slow conductor of heat. The cover of the cylin- der B has holes in it marked a to admit the external air to the space above the piston. D is the compressing pump with piston, C, which is connected to the piston A by three or four piston rods, of which two are shown at d d. The space between the piston C and A is open to the external air. In operation the air is drawn into the compress- ing pump through the valve c, and is forced out after being com- pressed through the valve e into a receiver marked F. It is admitted at proper intervals of time by the valve 6 into the working cylinder B, being heated in its passage by hot plates in the vessel G, termed the regenerator. It is further heated while in the working cylinder by the heat from the furnace ///the effect of which is to increase the tem- perature and volume underneath the piston. The increase of volume drives the piston to the end of its stroke. The exhaust valve / is opened by a mechanism connected to the engine and the working gases are forced outward through the regenerator, G, and discharge into the atmosphere through the pipe g. The regenerator is a vessel nearly filled with metallic plates which are heated by the escaping gases and give up heat to the engine gases, thus reducing in large measure the heat wastes from the engine. In some forms of the Ericsson engine the air entering the com- pressor through the valve c is taken from the exhaust opening g; with this arrangement the same mass of air is repeatedly warmed and cooled, the changes of temperature taking place at constant pressures. Ericsson's caloric engine was employed to drive a ship across the Alantic in 1853. The ship was 250 ft. long, had paddle 24 INTERNAL COMBUSTION ENGINES wheels 32 ft. in diameter. On its first trial trip the ship made twelve knots an hour with the wind, burning six tons of fuel per day. On the second trial the maximum speed was nine knots. After this unfavorable circumstances came to light and in 1855 the engine was taken out and a steam engine substituted. Extended accounts of the Ericsson and other hot-air engines will be found in Knight's Mechanical Dictionary, 1873, in Apple- ton's Cyclopedia of Mechanics, 1878, in Bourne's work on the Steam, Hot Air, and Gas Engine, and in Rankine's Steam Engine. Figure 1-11 represents the reciprocating parts of an air engine of the class in which temperature is changed at constant volume. Such an engine was designed and built by Dr. Robert Stirling about 1850, but improved by various other inventors, and is still built and sold for pumping small quantities of water as redesigned by Rider. In the figure D C A B is the air receiver or heating and cooling vessel, which is provided with a furnace underneath, not shown in the diagram; G is the working cylinder with the working piston H. The receiver and cylinder communicate freely through the passage F, which is open at all times when the engine is working. Within the receiver is an inner receiver or lining of a similar figure, which has its bottom pierced with many small holes shown with dotted lines in the cut. The annular space between the receiver and its lining extending along the side of the receiver contains the regenerator, which consists of a series of oblong strips of metal with narrow passages between FIG. 1-11. Stirling Hot them . The inner surface of the cylindri- cal part of the lining from A A to C C is turned and the plunger E is fitted nicely in this portion. The upper portion of the receiver D D is supplied with a horizon- tal coil of fine copper tube, through which a current of cold water is forced or it is jacketed with cold water, and is termed the refrigerator. It is thus noted that the bottom of the receiver is kept hot while the top is kept cold. The plunger E is con- INTRODUCTION, DEFINITIONS, ETC. 25 nected to the working mechanism of the engine so as to transfer a certain mass of air, which may be called the working air, from the hot to the cold end of the receiver, and in so doing making it pass up and down through the regenerator. The mechanism for moving the plunger E is so adjusted that the up-stroke of that plunger takes place when the piston, H , is at or near the beginning of its forward stroke, and the down stroke of the plunger when the piston H is at or near the beginning of the back stroke. An air engine of class 3, which received and rejected heat at constant pressures, was designed by Jewell, but probably was never put into practical use. HOT AIR ENGINES OPERATED BY PRODUCTS OF COMBUSTION. Another form of air engine, which Rankine in his Steam Engine terms a furnace gas en- gine, was first designed by Cayley of England and Barre of France, and was redesigned and improved and put on the market in this country about 1865 by Wilcox. The engines of this class operate similar to a steam engine, pressure being pro- duced by combustion in a closed furnace instead of in a steam boiler. In the operation of this FIG. 1-12. Wilcox Hot Air Engine. engine (see Fig. 1-12) a pump draws air from the atmosphere through valve F, compresses it, and forces it into a strong air- tight furnace C through pipe H , where its oxygen combines with the fuel; then the hot gas produced by the combustion mixed with air is admitted into the working cylinder through pipe B and valve 7, under pressure produced by the temperature of com- bustion, where it drives the piston P through part of its stroke at full pressure and through the remainder by expansion, until it falls to atmospheric pressure. It is discharged through an ex- haust valve not shown and a pipe X. The entering air is heated by a regenerator which is warmed by the exhaust gases. In the 26 INTERNAL COMBUSTION ENGINES form shown (U. S. patent May 19, 1865) the furnace is fed with coal through a double valve, which is so constructed that it can be introduced without permitting the escape of more than a very small quantity of the compressed air. In the Wilcox engine, patented Sept. 19, 1865, the furnace was fed with petroleum oil under pressure. Respecting Cayley's engine, Rankine states: "The cylinder, piston, and valves of this engine were found to be so rapidly destroyed by the intense heat and the dust from the fuel that no attempt was made to bring it into general practical use." Wil- cox's engine never met with commercial success. Most forms of the gas turbine belong to a class, in which pressure is produced by the heat of combustion of the gases be- fore entering the turbine, either in a closed combustion chamber or in the inlet pipe. As the supply to a turbine is continuous no inlet or exhaust valves are required, and hence the trouble experienced in the early forms of this engine is greatly reduced. 15. Structure and Mode of Operation of the Gas En- gine. The mechanism of the ordinary gas engine, using the term in its general sense as covering all internal combustion engines, is similar in most respects to that of the steam engine, which has already been briefly described. It consists of a cylinder containing a piston which is moved by the pressure produced by the explosion of a charge consisting of a mixture of gas or vapor and air in the cylinder. The motion of the piston is communi- cated to a main shaft by a connecting rod similar to that used in the steam engine. The valve mechanism of the gas engine serves to admit and discharge the charge at the proper interval of time and is operated by the mechanism of the engine. In the early development of the gas engine a slide valve was used to a con- siderable extent, but at the present time the poppet valve is commonly used as it has been found to withstand high tempera- ture better than the other form. In order to start the gas engine some external force must be provided to introduce the combustible charge into the working cylinder and give it the initial compression. This may be done in the first instance by revolving the engine by extraneous power, which puts the piston in motion and serves to draw in the neces- sary gas and air by suction and to compress the same. After the INTRODUCTION, DEFINITIONS, ETC. 27 engine is in operation the inertia of the moving parts keeps it in motion and serves to draw in and compress the charge. It is quite evident that the amount of work performed will be proportional to the mass or weight of the charge. For this reason it is desirable that the charge be under as much com- pression as practicable at the time of ignition, and all modern internal combustion engines provide means for compressing the charge previous to ignition either outside or inside of the work- ing cylinder. In nearly every case, in the modern engine, the compression is completed in the working cylinder. The principal events in the operation of an internal combus- tion engine are as follows: 1. Charging or suction, during which time the charge is drawn into the cylinder. 2. Compression, during which time the charge is compressed. 3. Ignition, explosion, and expansion, during which time heat is supplied which causes the combustion or explosion followed by the expansion due to increase of volume caused by motion of the piston. Ignition may take place under conditions of (1) constant volume, (2) constant pressure, or (3) constant temperature. 4. Exhaust, during which time the products of combustion leave the cylinder. 16. Classification of Internal Combustion Engines. Gas engines are scientifically classified in accordance with the mode of applying heat during ignition as follows: 1. Engines receiving heat with charge at constant volume; these will be called in this work Explosion Engines. 2. Engines receiving heat with charge at constant pressure; these will be called in this work Pressure Engines. 3. Engines receiving heat with charge at constant tempera- ture; these will be called in this work Constant Temperature Engines. In the first of the above classes of engines the charge may be ignited with or without previous compression; consequently this class may be subdivided into non-compression and compression engines, the non-compression engine has entirely gone out .of use because of its low efficiency and small capacity for a given size. Engines of any class may be either two or four stroke cycle engines as explained. 28 INTERNAL COMBUSTION ENGINES This classification is based on the characteristic equation ex- pressing the relation between pressure, volume, and temperature of a given weight of a perfect gas, which, as will be shown later, is P T"= R in which p = absolute pressure. v = the volume. T = absolute temperature. R = a constant for any given gas. It is evident that the heat may be received while any of the variables, pressure, volume, and temperature, remains constant, that is, at constant volume as in Class 1, constant pressure as in Class 2, or constant temperature as in Class 3. Internal combustion engines are often unscientifically classi- fied by the nature of the working fluid as gas engines, petrol engines, and oil engines. This classification gives no considera- tion to the fact that any of the above classes will operate with any of the fuels named. The term gas engine is frequently used in this work in its general sense, as applying to any form of in- ternal combustion engine. ENGINES IGNITING AT CONSTANT VOLUME, OR EXPLOSION ENGINES. In these engines, which are the ones commonly used, the various operations are performed in the order mentioned above in each working cycle of the engine. These engines may be divided into two classes accordingly as they perform these operations, in one end of the cylinder, (a) in four strokes or (6) in two strokes. In this class of engines the combustion is practically in- stantaneous and of the nature of an explosion, taking place under normal conditions while the volume of gas remains constant, thus producing an extremely rapid rise of pressure. (a) Four-Stroke Cycle Engine. The internal combustion engine most commonly used ignites the charge while its volume remains stationary, and requires four strokes for one cycle of operation. For this reason it is known as the jour-stroke cycle or four-cycle engine. This engine as ordinarily built is a single- acting engine with all the operations performed on one side of INTRODUCTION, DEFINITIONS, ETC. the working piston. A diagram showing its general construction and mode of operation for each stroke is shown in Fig. 1-13. In the operation of this en- gine the charge is drawn in during the first out stroke of the piston, is compressed during the return stroke, is ignited with the piston station- ary at the end of the stroke and with the vol- ume of the charge con- stant. It expands during the next out stroke and is exhausted and expelled -from the cylinder during the next in stroke. An engine of this kind was first described by Beau de Rochas in 1861, it was first built by Otto in 1876. The cycle on which it oper- ates is for this reason often called the Beau de Rochas or Otto cycle. (b) Two-stroke Cycle Engine. An internal combustion en- gine, igniting as before, which is used for many purposes, is de- signed to perform the four operations above referred to in two strokes and is known as a two-stroke cycle engine or a two-cycle engine. A common form of such engine is shown in Fig. 1-14, in which form the engine is in part a double-acting engine and both sides of the piston constitute closed chambers. The crank case is made tight by the use of stuffing-boxes on the main shaft, and the suction operation is performed by the inward stroke of the piston which draws the charge into the crank case through the inlet /, where it is partially compressed by the out or return stroke of the piston and transferred through port a, which is un- covered at the proper time by the piston, to the ignition side of FIG. 1-13. Diagram of Four-cycle Engine. 30 INTERNAL COMBUSTION ENGINES the piston. At the same time the charge of previously burned gases is escaping through the port E. The compression is completed in the working cylinder C, after the piston has closed the transfer port A. Ignition occurs at the beginning of the out stroke and when the piston and vol- ume in the working cylinder are sta- tionary as in the preceding case. In the action of this engine suction takes place below the piston at the same time that compression takes place above, and compression takes place below the piston at the time of expansion above, as shown by the diagram of the crank circle in the figure. In certain large engines which are now made to operate on the two- stroke cycle system the suction and preliminary compression of the charge, which is performed in the engine just described by the work- ing piston, is performed in a separate cylinder, the compression being completed in the working FIG. 1-14. Two-cycle Engine. cylinder. The method of igniting the charge in common use will be described at length later in the book. In the class of engines re- ferred to it consists of means for firing the charge instantaneously when under compression, and with its volume constant by an electric spark, a hot tube, or an open flame. ENGINES IGNITING THE CHARGE AT CONSTANT PRESSURE. The only engine of this class of practical importance is the Brayton, although under many conditions the Diesel engine ignites under constant pressure. The Brayton engine was at one time used extensively in America but is not now manufac- tured. In this engine the combustible and air for supporting combustion were supplied to the working cylinder under pressure which remained constant until the inlet valve closed; the com- bustion taking place during the admission of the air and com- INTRODUCTION, DEFINITIONS, ETC. 31 bustible. The work is performed by pressure acting during in- crease of volume in much the same manner as in the steam engine. In the Brayton engine, one form of 'which is shown in Fig. 1-15, the compression is performed in a compressor distinct from the power end of the working cylinder, and the heat is supplied at constant pressure. In the figure B is the working piston arranged to move in the cylinder A. The lower part of the cylinder A is the working cylinder, the upper part the air compressor which is arranged to deliver air into the reservoir C. FIG. 1-15. The Brayton Engine. Oil is injected by the pump G into a vaporizing device (shown on a larger scale in 3 of Fig. 1-15) when it comes in contact with compressed air from the reservoir C. The inlet valve b and ex- haust valve c are operated by suitable cams on the cam shaft E. ENGINES IGNITING THE CHARGE AT CONSTANT TEMPERA- TURE. Another class of engine, patented by Diesel, is ar- ranged to supply the fuel during a portion of the working stroke at such a rate as to maintain the temperature constant, the working cylinder having previously been filled with air during 32 INTERNAL COMBUSTION ENGINES a suction stroke and compressed during a return stroke. In the Diesel engine the compression is sufficient to raise the air to a temperature high enough to ignite the fuel as it enters the working cylinder. The Brayton engine as above described is a two-stroke cycle engine and the Diesel a four-stroke cycle engine, but both engines could be constructed to operate with either cycle. 17. The Engine Indicator. This is an instrument de- signed to draw a diagram with ordinates proportional to the pressure which acts inside the cylinder at each point during the working and return stroke of the piston. It is briefly described here in order to give the student an idea of the method of ob- taining the pressure volume diagrams which are frequently re- ferred to in the work. It consists essentially of (1) a part carrying a sheet of paper which is moved by proper mechanism in corresponding directions and proportional to the piston of the en- gine, and of (-2) a part which carries a pen- cil which is moved a distance proportional to the pressure per square inch acting upon the piston. The engine -indicator was first designed by James Watt, substantially as shown in Fig. 1-16. This indicator was constructed with (1) a flat plate, D B, on which paper could be mounted which w r as moved in proportion to the motion of the engine piston, and (2) a cylinder A A which could be put in communication with the working cylinder of the engine by a three-way cock H and pipe B. In the cylinder A A was a piston whose motion was resisted by a spiral spring arranged to carry on its piston rod a pencil so constructed as to draw a diagram on the moving plate with ordinates proportional to the pressure. The cock // can be turned so as to put the cylinder A A in communication with the air for the purpose of drawing a line showing the atmospheric pressure, which is called FIG. 1-16. The Watt Engine Indicator. INTRODUCTION, DEFINITIONS, ETC. 33 the "atmospheric line." This simple form of indicator, although containing the essential elements of the modern indicator, was crude in construction and gave results which were far from accurate. The modern indicator is an in- strument of precision, and differs principally from that designed by Watt by tjie substitution (1) of an oscillating drum, called the paper drum, for the flat reciprocating plate, for carrying the paper, (2) of movable indicator springs in place of a fixed one, making it possible to regulate the length of the ordinates, and (3) a multiply- ing pencil motion in place of the direct one, whereby the motion of the pencil on the indicator drum is made greater than that of the indi- cator piston. FIG. 1-17. The Thompson Indicator. made FIG. 1-18. The Thompson Indicator. A sectional and perspective view of the Thompson indicator, by the American Steam Gauge & Valve Company, is 34 INTERNAL COMBUSTION ENGINES FIG. 1-19. The Crosby Indicator. shown in Figs. 1-17 and 1-18. A perspective view of the Crosby indicator, which differs from the Thompson principally in the construction of the indi- cator spring and pencil motion, is shown in Fig. 1-19. For gas engine work the indicator spring is liable' to be injured by heat. To lessen these difficulties most of the makers supply indicators with ex- ternal springs, as shown in the attached view of the Tabor indi- cator, Fig. 1-20. The authors have found from an extensive experience in indicat- ing gas engines that the indicator spring when arranged as in Fig. 1-17 can be kept from injury by surrounding the working cylinder with a water jacket or cup filled with water. The indicator cylinder is connec- ted to the working cylinder by a pipe containing an indicator cock, which is arranged as in the Watt indicator to connect either with the air or the engine. The indi- cator drum is usually connected to some form of reducing motion by the indicator cord, which will move the surface of the drum proportional to the motion of the piston and not to exceed two or three inches, re- gardless of the stroke of the piston. Several forms of reducing motions with schemes for connecting will be shown in the chapter on the testing of gas engines, but will not be further referred to here. The class of engine indicator as described above is not adapted to take diagrams at extremely high speeds because of the inertia of the moving parts and because of the error due to stretching FIG. 1-20. The Tabor Indi- cator with External Spring. INTRODUCTION, DEFINITIONS, ETC. 35 of cords or flexible parts which connect the paper drum to the moving parts of the engine. For high speeds the optical indica- tor is preferable. The optical indicator has been designed so that the only mov- ing part is a small mirror which is arranged so as to project a ray of light on a ground glass screen or on a photographic plate. FIG. 1-21. Perspective View of Manograph. The mirror is moved in one direction an amount proportional to the pressure acting on the piston of the engine, and in a direction at right angles an amount in proportion to the motion of the piston, so that the joint movement is proportional to the pressure volume diagram, the area of which represents the mechanical FIG. 1-22. Horizontal Section of Manograph. work performed by the working fluid in the engine cylinder. Such indicators are affected only to a slight degree by the rotative speed of the engine. There are two instruments of this class on the market, one the Manograph, manufactured by J. Charpentier, Rue de Lambre, 20, Paris, the other the Optische Indicator, constructed by the 36 INTERNAL COMBUSTION ENGINES Elsaessische Electricitaets-Werke, Strassburg. A perspective view of the Manograph mounted on a tripod is shown in Fig. 1-21, a vertical section of same in Fig. 1-22; a detailed view of the mirror and engine connections are shown in Figs. 1-23 and 1-24. FIG. 1-23. Section of Manograph Mirror. The general construction is that of a photographic camera connected to the engine by tube T, and by flexible shaft R con- FIG. 1-24. Manograph Engine Connection. nected to the main shaft of the engine. An acetylene or electric lamp is located at G and its ray of light is projected through a perforated diaphragm to a prism, H, and thence to mirror N, at the back of the camera. The mirror N, supported on springs, INTRODUCTION, DEFINITIONS, ETC. 37 Fig. 1-23, is connected by a pin with the diaphragm M, which is in communication with the engine cylinder, so as to be tilted in one direction an amount proportional to the change of pressure in the engine cylinder. The mirror N is also tilted in a direction at right angles to the first motion by means of a crank with lever connection which is rotated from the shaft in proportion to the motion of the engine. The motion of the small crank can be set FIG. 1-25 Diagram with Manograph. in phase with that of the engine crank by the thumb screw F so that it will give a motion directly proportional to the piston of the engine. To make the errors as small as possible the angular motion of the mirror is made very small. The pressure scale of the manograph should be determined by carefully comparing the photographic diagram with a known pressure. The instrument FIG. 1-26. Optical Indicator. is arranged to give a diagram which would have exactly correct proportions when the connecting rod has a length 4.5 times that of the crank, which is nearly the average proportions in actual practice and would make the resulting error small for other conditions. Figure 1-25 represents a diagram taken with the Manograph from an engine making 1500" turns per minute and giving a maxi- mum pressure corresponding to 158 pounds per square inch. 38 INTERNAL COMBUSTION ENGINES The Optische Indicator differs from the Manograph princi- pally in details of construction. Its general appearance is shown in Fig. 1-26. A diagram taken from a motor operated with gasoline making 1000 turns per minute is shown in Fig. 1-27, the pressure scale of which is about 120 pounds per square inch. FIG. 1-27. Diagram with Optical Indicator. 18. Indicated and Brake or Delivered Horse-Power. The indicated horse-power which is generally denoted by the symbol I. H.P. is proportional to the area of the diagram obtained by use of the engine indicator, since this diagram has ordi- nates which are proportional to the pressures acting upon the engine piston at each point during the working and return strokes, and abscissa proportional to the corresponding space moved through by the engine piston. The indicated horse-power (I. H. P.) is computed by use of the formula plan I. H. P. = 33000 in which p = the mean effective pressure (m. e. p.) for the cycle of operation, acting on each square inch of the piston. I = length of stroke in feet. a = net area of piston in square inches. n = number of cycles per minute. 33,000 = number of foot-pounds per minute in one horse-power. The mean effective pressure (m. e. p.) is the mean ordinate INTRODUCTION, DEFINITIONS, ETC. 39 for all the strokes constituting the cycle multiplied by the proper pressure scale; it is best obtained by finding the net area by use of a planimeter (an instrument which will be described later), which is to be divided by the length of the diagram and multiplied by the scale of the indicator spring. The other quantities in the formulae depend upon the dimensions and speed of the engine. The brake or dynametric horse-power, for which the symbol in this work will be D. H. P., is that delivered from the main shaft of the engine and is consequently less than the indicated V horse-power by an amount equal to the engine friction and internal losses. The brake horse-power is usually measured by use of a spe- cial form of absorption dynamometer known as the Prony brake. Various forms of this brake have been employed, of which FIG. 1-28. A Prony Brake. two only are considered in this place. One form is shown in the diagram Fig. 1-28, which consists of a series of blocks connected by a leather strap or strip of iron and arranged so as to rub on the surface of a wheel attached to the main shaft. The brake is provided with two arms, the free end of which rests on a pair of scales. The amount of friction may be varied by use of a hand wheel or similar device as shown at S. The horizontal distance from the center of the wheel to the end of the arms is known as the arm of the brake and is denoted in the formulae which fol- low by a. In the use of the brake the load is applied by turning the screw and is measured by the reading on the weighing scale. In a brake with the arm on one side only, as shown in the figure, the amount required to balance the overhanging brake arm must be deducted from the reading of the scales to give the net load. 40 INTERNAL COMBUSTION ENGINES The horse-power is calculated from the formula D TT p 2 TT a n W 33000 in which n = number of revolutions per minute. a = the brake arm in feet. W = the net load on scales corrected for unbalanced effect of brake. Another form of brake is shown in diagram, Fig. 1-29, which is convenient for testing small engines. It consists of a rope or strap which makes one or more turns around the wheel, the tension or pull on both ends of which must be known or measured. In the form shown a weight of known amount, w, is applied at one end, and a spring balance is em- ployed to measure the resistance at the other end. The formula for this brake is as fol- lows: FIG. 1-29. The Rope /TJ7 x D - H -^ 2 " in which r = radius of the wheel in feet to center of strap. W = the principal scale reading. w = the lesser scale reading or weight carried. In a modification of this brake the principal tension is received by a framework resting on a pair of scales, and the smaller resist- ance is absorbed by an upward pull on the platform of the same scales. With this arrangement the scale reading gives directly the difference of weights W w. In the use of the Prony brake heat is generated equivalent to the mechanical work absorbed. If the load is heavy it will be necessary to circulate water or some other heat-removing fluid inside of the rim of the revolving wheel or in some equivalent place. 19. Forms of Indicator Diagrams. A few of the typical forms of indicator diagrams are considered in this place for the purpose of making the student familiar with the subject. They will be discussed at length later in the work. INTRODUCTION, DEFINITIONS, ETC. 41 Figure 1-30 is a hypothetical diagram of a four-cycle explosion engine with the events which take place on the various strokes marked. Thus the suction stroke is represented by d e, the com- pression by e /, the explosion by / a, the expansion by a b, the exhaust by 6 c d. The atmospheric line not clearly shown in the diagram would occupy a position intermediate between c d HYPOTHETICAL DIAGRAM, showing Cycle. FIG. 1-30. Four-cycle Engine. and e d. The lines c d and e d on the indicator usually coincide with the atmospheric line for the reason that the spring used is too stiff to show such small variations in pressure as exist between the atmospheric line and the exhaust and suction lines. Figure 1-31 shows diagrams of a two-cycle explosion engine in which the upper diagram, abcf, is taken in the working cyl- FIG. 1-31. Two-cycle Engine Diagram. inder, and the lower diagram, e' /', in the compressor. In this diagram the net work is the difference between that shown on the first diagram and that on the second. Figures 1-32 and 1-33 are diagrams from a Brayton or constant pressure engine, Fig. 1-32 being taken from the working cylinder and Fig. 1-33 from the compression cylinder. It will be noted 42 INTERNAL COMBUSTION ENGINES from Fig. 1-32 that the pressure remains constant from a to a', at which time communication is cut off from the compressor, after which the fluid expands from a' to c in much the same FIG. 1-32. Diagram from Brayton Working Cylinder. manner as in the steam engine. The net work is proportional to the difference of the areas of the two diagrams. FIG. 1-33. Diagram from Brayton Compressor. Figure 1-34 is a diagram from a Diesel engine in which the temperature is supposed to be constant during that portion of the stroke represented by a b, during which time fuel is being sup- plied to the working cylinder. FIG. 1-34. Diesel Engine Diagram. Figure 1-35 is a diagram from a steam engine, the expansion line of which as referred to an hyperbola, c e f g h, which is asymp- totic to the line of no pressure, C D, and of no volume, C B. Points in the hyperbola are obtained if the initial point, c, is INTRODUCTION, DEFINITIONS, ETC. 43 known, by drawing a vertical from c; then from C draw diagonals crossing c b and A B. The intersection of a horizontal line from the intersection of the diagonal and c b, with a vertical line from B rr FIG. 1-35. Method of Drawing Hyperbola. the intersection of the same diagonal and the line A B, give points in the hyperbola. Another method of drawing an hyperbola is shown in Fig. 1-36, d' FIG. 1-36. Method of Drawing an Hyperbola. which represents an indicator diagram referred to lines of no volume and no pressure. This method is founded on the prin- ciple that the intercepts made by a straight line intersecting an 44 INTERNAL COMBUSTION ENGINES hyperbola and its asymptotes are equal. Beginning at any point as a, draw the straight line, a' b f , and lay off from the line CD, b' b equal to a' a, then will b be a point in the hyper- bola. Draw a similar line through b as c' d' and find another point as c. Repeat the method until all the points required for drawing the curve are found. The hyperbola is a useful line of reference in connection with indicator diagrams. As will be shown later, it represents the condition of isothermal expansion in the gas engine. CHAPTER II THERMODYNAMICS OF THE GAS ENGINE i. Notation. In order to comprehend the limitations of the actual engine it is necessary to understand how the working fluid behaves when subject to definite changes in an engine un- affected by friction or mechanical limitations, and we will for that reason give attention to the theoretical considerations relat- ting to the " internal combustion engine" which forms part of the science of Thermodynamics. In the consideration of the theoretical action of a perfect engine the following symbols will be used: A = the reciprocal of the mechanical equivalent of heat. a = the absolute temperature of the freezing-point = 273 C. or 492 F. C p = the specific heat of constant pressure in heat units. C v = the specific heat of constant volume in heat units. J = mechanical equivalent of heat = 778 in foot-pounds = 426 K. G. M. Kp = specific heat of constant pressure in mechanical units. K v = the mechanical heat at constant volume in mechanical units. p = absolute pressure for condition denoted by postscript. Po = absolute pressure at freezing-point. Q = the total heat of a given mass. R = a constant for a given gas. T = absolute temperature. t = temperature Fahrenheit or Centigrade as marked. o = volume of a given mass of gas, its condition being denoted by postscipt. V Q = the volume of a given mass of gas at freezing-point. W = the mechanical work performed in mechanical units. W = the mechanical work in heat units. a = the coefficient of expansion of a perfect gas = the reciprocal of a. 7 = specific heat of constant pressure divided by the specific heat of constant volume. 2. Characteristics of Perfect Gases. In an internal combustion engine the work is produced by the change of volume and pressure of a gaseous mixture composed of atmospheric air 45 46 INTERNAL COMBUSTION ENGINES and the various products of combustion. This gas mixture, within the working limits of temperature, is obedient to the laws of the perfect gases, and for that reason in any theory of the internal combustion engine we are principally concerned with such laws and with the changes of volume and pressure in a per- fect gas, in relation to its change of temperature. By combining Boyle's law with Gay-Lussac's law it is learned that the product of pressure and volume of a given mass of a perfect gas varies directly as its absolute temperature, that is: pv = p Q v Q (1 + a t) (1) in which p equals the absolute pressure and v the volume of a given mass of gas at a temperature, t, above the freezing-point, p and v represent the pressure and volume of the same mass at the freezing-point, and a represents the coefficient of expan- sion of the gas per degree of absolute temperature. As has already been shown a will equal when expressed in the Centi- grade system the reciprocal of 273 and when expressed in the Fahrenheit system the reciprocal of 492. If we denote the number of degrees between the freezing-point and absolute zero by a, then from the preceding explanation a will equal the re- ciprocal of a. If we denote the absolute temperature by T we shall have T = a + t. (2) substituting for a in equation (1) we have 13 V in the above equation - is a constant for each gas. a then we have Let R = (4) a p, = RT . (5) The above equation may be considered the characteristic equation of a perfect gas, since it shows the relations between the pressure, volume and absolute temperature. R is a constant which depends on the nature of the gas and THERMODYNAMICS OF THE GAS ENGINE 47 can be computed if the specific pressure, p , and specific volume, v at standard pressures and temperatures are known. It follows from the above that, p^ = RT 1} from which by comparing with (5) po'.p&t ::7\ :T (5a) The specific pressure p is the weight of the atmosphere at the freezing-point under normal conditions; it is equivalent to that of a column of mercury 760 mm. high (29.921 inches). This re- duced to pressure per unit of area is p = 10333 kilograms per square meter; or in English units, p = 31 fr6.83 pounds per square foot. , = 14.696 pounds per square inch. = 29.921 inches of mercury. The specific volume is determined from the density of the gas. The following table gives the specific volume in Metric and English measures for some of the more common gases: VALUE OF v AT LAT. 45 Cubic meters per Kilogram Cubic feet per Pound Air 7735327 12 3909 Nitrogen (N) 0.7963291 12.7561 Oxygen (O) 0.6996231 11 2070 Hydrogen (H) 1 116705 178 881 Carbonic Acid (CO?) 5058741 8 10324 By substituting the values of p v , and a in the equation R = the value of R can be found. Thus for air, in French units R = 10333 X 0.77353 - 273 = 29.20 ; in English units R = 2116.3 X 12.391 - 492 = 53.22. 48 INTERNAL COMBUSTION ENGINES The following table gives values of R for a few gases : VALUES OF R v English | Metric Hydrogen (H) 770.3 48.74 35.41 53.22 422.68 26.475 19.43 29.20 Oxvffen fO^ Carbon-dioxide (CO 2 ) Air The following table showing the specific heat of the ordinary gases is inserted here for convenience. TABLE OF SPECIFIC HEATS SPECIFK : HEAT * NAME OF GAS SYMBOL Constant Pressure C P Constant Volume c v C V y Air 2375 1684 1 406 Oxvffcn o 2175 1552 1 403 Nitrogen Hydrogen N H 0.2438 3 4090 0.1727 2 4110 1.416 1 414 Nitric oxide NO 0.2317 0.1652 1 402 Carbonic oxide Carbon dioxide CO CO 2 0.2450 2000 0.1736 1550 1.413 1 261 Steam H 2 O (saturated) 4805 3700 1 298 Methane Acetylene Disulphide carbon Olefiant gas CH 4 C 2 H 2 CS 2 C H 0.5930 0.3460 0.1569 4040 0.4680 0.2700 0.1310 3330 1.198 1 125 Ammonia. . .-. . 24 NHj 05084 3910 1 300 Alcohol CH O fl 04534 4100 1 150 A 2 6 As before noted, the specific heat of gases increases with the temperature and possibly also with the pressure, which law will be referred to in discussing the application to special cases. The expanding products of combustion are composed of N, CO 2 H 2 O, O, and possibly a trace of NO. For approximate computations the value of C p /C v for the burned gases may be taken at 1.37. 3. General Relations of Heat Transmission to Changes of Volume and Pressure. When a quantity of heat dQ is sup- plied to a mass of gas it produces a complex result; the heat THERMODYNAMICS OF THE GAS ENGINE 49 warms the gas and raises its temperature, at the same time per- forming internal work by overcoming the molecular forces; it then develops external work, which is made apparent by the expansion of the gas against an external resistance. Denoting by dU the quantity of heat employed in warming the gas and in molecular work we will have dQ = dU + Apdv, in which pdv is the external work expressed in mechanical units and A pdv its equivalent in heat units. In the operation of a gas engine the mass of gas constituting the working substance expands during each cycle from one volume to another, passing through a series of successive changes of volume and pressure, and finally returns to its initial state. It is evident that when the-mass returns to its initial state that the quantity represented byfdU is equal to zero. For this condi- tion we shall have dQ = A pdv. Calling U the internal heat of the gas, which has already been shown to be a function of the volume pressure and temperature, we have TT , . U = f(v. p. t.). Since t can be determined from the values of p and v as indicated in equation 5, we can consider that in practice U is a function only of v and p and may write U = } (v. p.). The increase of internal heat for a variation of volume dv and of pressure d p may be expressed as a total differential of the function, /, and we have By substituting the above value in the expression d Q the thermal state of the gas may be represented as follows: or 50 INTERNAL COMBUSTION ENGINES It may be noted that the above equation cannot be integrated in its present form unless Q can be expressed as a function of the initial and final volumes and pressures of the gas. This demon- strates that the expenditure of heat required to make a gas pass from one state to another cannot be deduced from a knowledge of the extreme states, if one does not know the order and relation of the intermediate states. The heat transferred at constant volume is equal to the specific heat C v multiplied by the change of temperature. That is, Q = C v (T l T), from which dQ = C v dt. For a similar reason that transferred at constant pressure 'Q = C p (T,-T} dQ = C p dt. For the condition of transfer of heat at constant volume, dv of equation (7) will equal 0, and equation (7) will become Since the heat interchange for this case takes place at constant volume, it has been shown that dQ = C^dt. By placing these two values of dQ equal we have c - - - (m C > Sp *TT from which the value of g for heat interchanges at constant volume can be found. For the transfer of heat at constant pressure the temperature changes take place without change of pressure, in which case d p of equation (7) = 0, and we have by substitution Since the heat interchange for this case takes place at constant pressure, dQ = Cpdt, and by substitution dt SU THERMODYNAMICS OF THE GAS ENGINE 51 &TT from which the value of y- for heat transformation at constant pressure can be obtained. Substituting the values of eq. (8) and (9) in eq. (7) we have dQ = C^dp + C p f v dv (10) which gives the value of the heat interchange for successive change of pressure and volume. This can be reduced as follows: From eq. (5), pv = RT, hence vdp = Rdt when v is constant, and pdv = Rdt when p is constant. From this it follows that Bt v , Bt p - = D an d T = D Bp R W R Substituting the above values in eq. (10) we have dQ = | (C v vdp + Cppdv) (11) Now, from eq. (5) , we obtain by complete differentiation pdv + vdp = Rdt, from which vdp Rdt pdv Substituting this in eq. (1 1) we may write dQ = jJC v (Rdt - pdv) + Cppdv] But from eq. (5) p _ T R~ v hence finally ~ x T , dQ = C^dt + (C p - C v ) - dv (lla) 4. Transformation to Different States. The modes of trans- formation from one state to another depend upon the relation of p and v at different points, T being always determined, by the relation of p to v, from eq. (5). For simplicity of treatment and for producing a standard for comparison it is assumed that the changes in the relations of volume, pressure, and temperature take place with one of the variables constant in the general equation (5a) , p. 47. Thus, if the volume remains constant during the change, we have v = v lf and - (12) 52 INTERNAL COMBUSTION ENGINES which is the equation for constant- volume conditions. If the pressure remain constant, p = p l and V -2- (13) which is the equation for constant-pressure conditions. If the temperature remain constant, which latter is termed an isothermal condition, T = 7\. i-LJt (14) Pi v which is the equation of an isothermal line for a perfect gas in a pressure- volume diagram. Another standard of comparison is the transformation in pressure, volume or temperature which takes place without gain or loss of heat. This latter condition is called adiabatic and cor- responds to that of constant entropy. It represents the conditions of the equation (11) when dQ = 0, in which case C v vdp + C p pdv = C 1 vdp + -~rpdv = Cy C 1 Substitute y for -^, then CT, vdp -f- ypdv = 0. which integrated between the limits p l v l and pv gives -7 log. ) from which p^ pv y = constant (15) which is the equation of an adiabatic line for a perfect gas in a pressure- volume diagram. The equation for adiabatic transformation in terms of v and T can be obtained by substituting for p and p^ the values as given in (5a) and reducing, which will give TV-' - T> -' ^ In a similar manner the adiabatic equation in terms of p and T can be obtained, which is as follows: THERMODYNAMICS OF THE GAS ENGINE 53 5. Work Performed in Isothermal Expansion. The work, W, performed when the gas expands isothermally from an initial volume, v to a volume v l can be calculated as follows : The general formula for mechanical work is W -i fpdv but as pv = p^ for isothermal expansion /Virlii -ji -p,,iog. (is) Since pv = p l v l = RT this may be written The heat applied during isothermal expansion can be obtained by making dt = and T a constant in (11 a) and integrating. We will have Q=(C P - C\)T P d = (C P -C V )T \og e ^ = ARTlog e -' (19) This value being the same as that of the external work indicates that the heat applied during isothermal expansion is equivalent to the external work performed. It will be noted from (18) that an infinite increase in isother- mal expansion will lead to an infinite amount of work. Thus in the equation W = pv loge if v 1 be made equal to infinity the value of W also becomes infinite. 6. Work Performed in Adiabatic Expansion. The work performed when the gas expands adiabatically from an initial volume, v l to a volume, v 2 , can be found by substituting the value f P = ^ f rom formula (15) as follows: dv W pi F-'i = I pdv = p.v I Jv l ^ l Jv 54 INTERNAL COMBUSTION ENGINES 1 - ~ '* Therefore W = ^ J 1 - () ^ (20) For infinite adiabatic. expansion the work, W, does not become infinite as in isothermal expansion, since for this case = 0, and W 7-1 7. Relations of Heat to Entropy. The heat transfor- mations can be expressed as a function of the absolute tempera- ture and entropy, which expression possesses some advantages for tracing heat interchanges over the pressure- volume equations. For this case the variables in the equation become tempera- ture, T, and entropy, <, instead of v and p as in the preceding cases; in the diagram representing such conditions, horizontal lines would represent equal temperatures or isothermal conditions, while vertical lines would represent equal entropy or adiabatic conditions. The ordinates in such a diagram would then repre- sent temperature, T, and the abscissa entropy, . Since the heat interchange d Q is equal to the product of the absolute temperature into the corresponding change of entropy, we have dQ = *- For gases, if heat is supplied at constant pressure dQ = Cpdt r (21) 8. Carnot or Reversible Engine, Second Law of Ther- modynamics. A reversible engine is one that may be run in one direction so as to transform heat into work, or in the opposite direction so as to transform work into heat. No actual heat engine is built in this manner, since such an hypothesis requires that all the gases exhausted shall pass through all the states in a reversed direction during compression and return to the initial state, which, because of the chemical THERMODYNAMICS OF THE GAS ENGINE 55 changes during combustion, is impossible in the internal combus- tion engine. The internal combustion engine can be considered as approximating the theoretical reversible engine, which thus becomes useful as a standard of comparison. For the cycle of a reversible engine This is the highest attainable result with any heat engine, since it indicates that the heat transferred into work from motion in one direction would be returned to its source by an equal amount of work applied to drive the engine in an opposite direction. The above statement is Carnot's principle, which is often called the Second Law of Thermodynamics. It follows from this:* (1) All reversible engines working between the same source of heat and refrigerator have equal efficiencies. (2) The efficiency of a reversible engine is independent of the working substance. (3) A self-acting machine cannot transfer heat from one body to another at a higher temperature. It further follows from this that for any irreversible engine cycle the work for a given expenditure of heat is less than for a reversible engine; that is, in which N represents the mechanical results of the work performed. 9. Graphical Relations. The relations of the heat inter- changes to the transformations of pressure, volume, and tempera- ture will be more clearly understood by reference to a diagram. The pressure- volume diagram, which has for its ordinates lines corresponding to pressure and for its abscissae distance corre- sponding to volumes, shows the conditions of uniform pressure by a horizontal line and of constant volume by a vertical line. On this diagram an isothermal line is represented by equation (14), pv = p^v^ which is the equation of an equilateral hyperbola of which the axes are the lines of zero volume and the line of zero pressure. Two methods of drawing the hyperbola have been *Peabody's Thermodynamics, page 30 56 INTERNAL COMBUSTION ENGINES given in Art. 19, Chapter I. In the case of a steam engine it will be remembered that an isothermal condition is represented by an equal pressure line. The equation of an adiabatic line on a pressure-volume dia- gram, as given in (15) is log = y log -. The values of the coor- Pi v dinates for drawing this curve can be found by assuming values 11 of -and finding the corresponding values, by use of a table of T) Naperian logarithms, of . A table giving the values of y for Pi different gases has been given. It is usually assumed as 1.37 for gas engines and is subject to some correction for changes due to rise of temperature. The general relations of isothermal and adiabatic lines to pressure, volume and temperature is shown on the diagram Fig. 2-1. 10 20 30 40 50 60 70 80 90 100 10 30 40 50 60 70 80. 90 200 % 1 CU.D'IC foot. 2 cubic feet. FIG. 2-1. Relations of Isothermal and Adiabatic Curves. Figure 2-1 shows isothermal and adiabatic expansion and com- pression lines drawn from the same points. This diagram shows that for a given number of expansions the adiabatic line falls THERMODYNAMICS OF THE GAS ENGINE 57 below the isothermal line, and also that the area between the adiabatic line and the base line is less than that between the isothermal and the base line. As this area represents the exter- nal work performed, it indicates that for a given number of ex- pansions the work is greater in isothermal than in adiabatic expansion, which also follows from the demonstrations which have been given. 150\ FIG. 2-2. Isothermal and Adiabatic Changes Compared. The various fundamental changes which may take place in a perfect heat engine are represented by the diagram, Fig. 2-2, in which V is the base line from which pressures are measured and P the zero volume line from which volumes are measured. In the diagram G A is a vertical line and represents the condition of receiving heat at constant volume; A B, a horizontal line, repre- sents the condition of receiving heat at constant pressure ; B C, a plain hyperbola, represents isothermal expansion, and B D, a logarithmic curve, represents adiabatic expansion; D E, a vertical line, represents the discharge of heat at constant volume; E F, a horizontal line, represents the discharge of heat at constant pressure; FG, a logarithmic curve, represents adiabatic compression, and F H, a plain hyperbola, represents isothermal compression. The area of the diagram, ABDE FG, represents the external work done with adiabatic expansion and compression, and ABC E F H A represents the work done with isothermal expansion and compression. During the period of receiving 'heat at constant volume, represented by AG, the absolute temperature may be computed 58 INTERNAL COMBUSTION ENGINES at A, from (12), provided it is known at G, since it is proportional to the pressures at those points. During the period of receiving heat at constant pressure, represented by A B, the absolute temperature increases in pro- portion to the volume, as shown in equation (13), and if known at one point may be computed at any other. As the volume is proportional to the distance from the line P, the temperature on the line A B will be proportional to that distance. If the expansion is isothermal the temperature would remain constant from B to C. If the expansion is adiabatic, as from B to D, the temperature could be calculated from either equation (16) or (17) for various points of the curve. The actual pressure- volume diagram as taken with the indicator shows lines which only approximate those for constant pressure, constant temperature, or constant volume as shown in Fig. 2-2. The corners of actual diagrams are likely to be rounded to a considerable extent and many of the transformations indicated may not appear. 100 ribs. 0.0 0.1 0.2 0.3 0.4 cu. t. FIG. 2-3. Four-cycle Engine Diagram. THERMODYNAMICS OF THE GAS ENGINE 59 Figure 2-3 represents a diagram of a 4-cycle engine in which the ordinates have been enlarged relatively with respect to the abscissae and on which there have been drawn a line of no volume, often called the clearance line, and a line of zero pressure. The scale of ordinates is attached to the diagram. If we assume that the line a f is a vertical line and that the temperature at the lower end of that line is 567 degrees absolute, then we find by computation, as explained above, that for the point a it is 1995 degrees absolute. As absolute temperature F is 460 degrees higher than that shown on a thermometer, the temperature F at these two points would correspond to 107 degrees and 1535 degrees. The heat transformations may also be represented by the entropy temperature diagram, in which case the ordinates be- come temperature and entropy instead of pressure and volume, p o v FIG. 2-4. Pressure Volume Diagram of Isothermal and Adiabatic Changes. A simple illustration is shown in Figs. 2-4 and 2-5. In Fig. 2-4 is shown a pressure-volume diagram of a working material which expands isothermally from A to B, then adiabatically from B to C. It is then compressed isothermally from C to D and adiabati- cally from D to A, when it reaches the initial condition. The me- chanical work performed during these operations is proportional to the area of the diagram A B C D. This diagram may be 60 INTERNAL COMBUSTION ENGINES transformed into a temperature entropy diagram very easily, since for that case the isothermal lines A B and B C, would be horizontal and parallel to the line of zero temperature, and the adiabatic lines A D and D C would be vertical and parallel to the line from which entropy is reckoned. Fig. 2-5 shows the tem- perature-entropy diagram constructed as described. The area of this diagram shows the heat trans- ferred into work. For the case considered this is a rectangle and its area represents the maximum amount of heat available for work within the tem- perature limits A B and DC. The case considered is that of the perfect rever- sible engine which operates ~~^ in a Carnot cycle, as al- FIG. 2-5. Entropy-Temperature Diagram. rea dy described. It appears from the diagram, Fig. 2-5, as well as from the demonstration, that an engine working on this principle can transform the maximum amount of heat into work. An engine working on any other principle, as for instance that shown in the Fig. 2-2 A B D E F G A, will trans- form a less amount of heat into work. For this case there is both adiabatic expansion B D and adiabatic com- pression F G. The entropy-tempera- ture diagram for this case is repre- sented, but not to scale, in Fig. 2-6 by the diagram A B D E F G, which by in- spection is smaller and shows less heat available for work than the diagram FIG. 2-6. m B n F, which is drawn between the same temperature limits. For the figure A B C E F H, shown in Fig. 2-2, as a pressure- THERMODYNAMICS OF THE GAS ENGINE 61 volume diagram we have isothermal expansion B C and isother- mal compression F H. The entropy-temperature diagram for this case is represented by the T diagram, not drawn to scale, in Fig. 2-7, A B C E FH, which by in- spection is smaller and shows less heat available for work than the diagram m C n H, which is drawn between the same temperature limit. The transformation of the pressure-volume diagram as taken the indicator into entropy- on FIG. 2-7. temperature diagram is given at length later in the book. 10. Comparison of Theoretical and Actual Heat Engines. - When a gas after a series of transformations of pressure- volume and temperature passes through a series of intermediate states and of physical and chemical changes and returns to the same condition, in all . respects, which it possessed at the beginning of the transformations, it is said to operate in a closed cycle. It is evident that if a change of composition occurs during the course of the cycle the body may return to its initial condition so far as pressure and volume are concerned without returning to its initial condition in other respects; for example, a mixture com- posed of hydrogen, carbon monoxide, methane, carbon dioxide and nitrogen, with which the cylinder is charged in the initial condition before combustion, may be changed during combustion to the vapor of water, carbon dioxide, and nitrogen, which change would not be shown on the indicator diagram. In the actual operation of the internal combustion engine the gas or vapor is subjected to periodical changes, as outlined above which do not constitute rigorously closed cycles. The operation is, however, approximately a closed cycle since the state of the working fluid after each series of changes returns to its initial state, from which all the operations as to chemical and physical changes are reproduced as in the preceding phase. The machine in which the changes take place will be a perfect neat engine if all of the heat disappearing has been transformed into w r ork; it has been demonstrated practically, however, that it 62 INTERNAL COMBUSTION ENGINES is not possible to utilize all of the heat, consequently it is necessary to expend in heat a larger quantity than its equivalent in units of work. It has already been shown in effect that it is not only necessary to have a place of combustion which produces a high temperature at the beginning of the period of movement, but it is also necessary to have a point of lower temperature which will act as a refrigerator and permit the flow of heat from a higher to a lower temperature level. The quantity of heat Q supplied by the combustion less the amount q taken up by the refrigerator leaves a difference Q q which may be utilized. From this statement it would appear that the amount of work possible would be increased either by increasing Q or by diminishing q. The temperature of the discharge heat q must evidently be con- siderably above that of absolute zero because of the difficulty of obtaining a refrigerant of low temperature and of disposing of the heat which would be discharged under such a condition. Generally in practice the temperature of the discharge heat is considerably above that of the surrounding air, which is much in excess of absolute zero. The ratio of Q q to Q measures the perfection of the heat engine. This we will call the cyclic efficiency ~ = cyclic efficiency v It is of great practical interest to know how to determine the value of this coefficient for any case, but we should at first estab- lish the maximum value that can be obtained. In this connection the cycle of Carnot which is formed of two isothermal lines and two adiabatic lines, Figs. 24 and 25, should receive consideration, since it is the one which gives the maximum work returned for the heat expended. The Carnot cycle is represented in the pressure-volume dia- gram Fig. 2-4, in which a mass of fluid in its initial state p v , with temperature T as shown at A, expands in volume isothermally to B, at which point it has pressure and volume p^. From B it expands to C without gain or loss of heat, following the adiabatic B C. From C to D there is a discharge of heat into a colder body or refrigerator at constant temperature, during which time the volume is reduced from v 2 to ^3. From D to A compression takes THERMODYNAMICS OF THE GAS ENGINE 63 place without gain or loss of heat, which raises the temperature to that of the initial state at A. In order to carry out the cycle of operation described, sufficient heat must be supplied during the isothermal expansion from A to B to keep the temperature constant; this amount by equation 7) (19) will be Q = A R T log- 1 -. In the adiabatic expansion from B to C, during which there is neither gain nor loss of heat, the rela- tions of the volumes to the temperatures are expressed by equa- tion (16) T AM 1 =( _ 2 T 2 W During the third period from C to D heat is discharged at con- stant temperature and can be expressed as before q = A R T log - 2 . During the fourth period the quantity of heat remains con- stant and we have >T A,. \ v l -to fV\y i The cycle as above described is a closed one, and as the heat re- ceived and discharged is of constant temperature it follows that The work produced is equal to the area A B C D. The efficiency of the cycle is Q q T T ~Q~ ^TT This is equal to the ratio of the fall of temperature to the absolute temperature of the initial condition. The Carnot cycle may also be represented, as already shown, by the temperature-entropy diagram, in which case the diagram will be a rectangle, Fig. 2-5. It is doubtless true that no better method exists for utilizing the heat furnished by combustion than that of supplying it at con- stant temperature, permitting the body to expand without gain or loss of heat, discharging it to the refrigerator at a constant tem- perature, which should be as low as possible, and compressing without gain or loss of heat to the original temperature. Theoretically it is possible to equal but not to surpass the return from the Carnot cycle. The maximum effect that can be 64 INTERNAL COMBUSTION ENGINES obtained from a heat engine working between the temperatures T and T f , and in which Q is the heat expended, is expressed by the equation , In order to judge the theoretical value of a cycle on which any heat engine operates it is desirable to calculate at first the coefficient of economy of the proposed cycle, then compare this coefficient with the Carnot coefficient between the same temperature limits. The knowledge which is given by comparing the cycle of the engine with the Carnot cycle is not sufficient to determine the practical value of the engine, since the effect of friction, shocks due to inertia, and the passive resistance of the various mechanical parts consume a portion of the work supplied by the transfor- mation of heat and do not appear in the useful work delivered by the machine. It is quite possible that machines which have a high degree of perfection for the transformation of heat will still give small return as practical, useful machines. The hot-air engines for example, which have a perfect cycle of operation, have proved in practice of little value because of the small amount of heat that would pass through a metallic wall in a given time, and as a consequence the return in useful work is small in proportion to the expense of construction. There are few engines of any kind which operate in a cycle approximat- ing that of Carnot, among these should be mentioned the Sterling air engine, which theoretically operates on the Carnot cycle. The cycle of operation of the steam engine is incomplete in many respects; it however resembles that of Carnot in that heat is received into the engine cylinder, until the valve closes connec- tion with the boiler, at practically constant temperature. After cut-off the steam is expanded approximately without gain or loss of heat. It is then discharged through the condenser at constant temperature. The adiabatic compression is sometimes considered as being performed by the feed pump which supplies water to the boiler. The steam engine cycle is thus seen, when the boiler furnace and boiler feed pump are included as a part, to approxi- mate that of the Carnot cycle. The Diesel motor is the only gas engine which approximates in the theory of its operation to the Carnot cycle. CHAPTER III THEORETICAL COMPARISON OF VARIOUS TYPES OF INTERNAL COMBUSTION ENGINES 1. Throughout the following discussion of the theoretical cycles it will be assumed that the specific heats at constant volume C v , and at constant pressure C p , do not vary either with pressure or temperature. It has been shown that they vary, but the question is unsettled. If the variation is such as determined by the experiments of Mallard and Le Chatelier, which are extensively quoted, our present-day gas engine does not admit thermally of any further improvement. In view of this unsettled con- dition it is best to assume the specific heats constant. It is further assumed, in this theoretical discussion, that the value C 1 of y = -^ is the same for the burned gases as for the fresh fuel C 7 ; mixture. It is further understood that, wherever heat supplied to a cycle is mentioned, it refers to the lower heating value of the fuel con- cerned, either per pound or per standard cubic foot, as stated. 2. The cycle receiving heat at constant volume, Beau de Rochas or Otto cycle. The principles upon which the present-day constant volume combustion engine is based, and which helped it to its commercial success, were first clearly enunciated by Beau de Rochas in a written pamphlet in 1862. It remained for Otto, however, to construct the first practically successful machine operating with this cycle, hence the cycle receiving heat at constant volume is more often known as the Otto cycle. In what follows let Q = quantity of heat received by the theoretical cycle, q = quantity of heat rejected by the theoretical cycle, then E c = the cycle efficiency = -^--. 65 66 INTERNAL COMBUSTION ENGINES It shows the highest efficiency which an engine could possibly show if it could follow exactly the lines of the theoretical cycle. Practically this can never be realized, but the conditions which determine why the actual thermal efficiency of an engine is always less than the cyclic efficiency will be treated in detail later on. In Fig. 3-1, at the end of the charging stroke, whether that be in the two- or in the four-cycle, the charge is under a pressure, temperature, and volume determined by the point 1. Adiabatic compression then takes place to 2. A quantity of heat, Q, is next received at constant volume to 3. From 3 to 4 adiabatic expansion takes place, and finally the quantity of heat, q, is re- jected along line 4-1 at constant volume. 1 3, FIG 3-1. Let the total charge weight be G Ibs. This consists of G a Ibs. air, G g Ibs. gaseous fuel, and G r Ibs. burned gases from the previous cycle. a) (2) Q = GC, (T, - T,) q == GC V (T, - T,) Hence the heat utilized by the cycle is ' Q - q = GC V (T, - T 2 - T t + T and the cyclic efficiency E _Q-V= GC v( T 3- T 2- T 4 +T J (3) (4) (5) T t -T, THEORETICAL CYCLES 67 Again, in Fig. 3-1, let v c = clearance volume, and v s = stroke volume, so that v t = v c + v s = total volume. We may write, since compression and expansion are assumed adiabatic, (6) and P^l = P*$- (7) Dividing (6) by (7), ^ = ^. (8) Ps P* Now 2 =? T 3 p 3 4 p, hence finally ^ = J or ^ = ^ (9) ^3 ^4 ^3. ^2 With the aid of (9), equation (5) may be written S) | (10) There are two other ways of stating the cyclic efficiency which will be developed next. Again we can write PiV=P 2 ^ Y (11) also , f\ Dividing (11) by (12) v 1 v-l A or Y i aV'V "V- 1 (13) in which r = the compression ratio . Hence B ' =1 ~fi~ 1 ~7?= I (14) 68 INTERNAL COMBUSTION ENGINES Finally, raise equation (12) to the y power y y y y Pi Vt = P-2 Vc (15) and divide (11) by (15). We shall have y-l from which 1 E=l_j!=l_(P) y (17) 1 2 F2 From an examination .of equations (14) and (17) it will be seen C that E c depends upon r or p 2 , and y = ~. Lv To make clear the influence of the value of y, assume that in two given cases the value of r = 5, but that in the first case a rich gas mixture with y= 1.35, in the second case a lean gas mixture with y = 1.39 be employed. Then for the two cases we shall have ^1- T ~ 5 -431 and E c =l-~~=A6Q (&) The advantage in favor of the lean gas mixture is therefore .466-. 431 _ -431" Besides this the use of lean gas mixtures in practice usually shows a smaller jacket water loss, owing to the lower mean tem- perature of the entire cycle. The value of E c also depends upon the ratio &, according to ? 7 2 eq. (17). The smaller this ratio the greater the efficiency. But the value of p lt the suction pressure, is almost entirely out of our control, so that the problem narrows down to making p 2 , the com- pression pressure, as high as possible. This brings out clearly the value of high compression. The practical limits to this state- ment will be pointed out later on. To show the combined influence upon E c of r or p 2 and y the following table, from Giildner, is given: THEORETICAL CYCLES 69 CYCLIC EFFICIENCIES, E c ., FOR THE OTTO CYCLE r = 2.0 2.5 3.0 3.5 4.0 4.5 5.0 6.0 7.0 8.0 9.0 10.0 y = 1.20 .129 .167 .197 .221 .242 .260 .275 .301 .322 .340 .356 .361 = 1.25 .159 .205 .240 .269 .290 .313 .331 .361 .385 .405 .423 .438 1.30 .188 .241 .281 .313 .343 .363 .383 .416 .442 .464 .483 .499 1.35 .216 .274 .319 .355 .384 .409 .431 .466 .494 .517 .537 .553 1.40 .248 .313 .363 .402 .434 .460 .483 .520 .550 .574 .594 .611 3. The cycle receiving heat at a constant pressure, the Bray ton cycle, and the approximate Diesel cycle of to-day. The Brayton Cycle. In the older machines of the Brayton type, suction and com- pression of the fuel mixture or of air took place in one cylinder, a i FIG. 3-2. while the combustion, expansion, and exhaust took place in another. In Fig. 3-2, area b 1 2 a b represents the pump diagram, area a 3 4 b a the diagram from the power cylinder. For the purpose of theoretical discussion of this cycle, the two diagrams may be combined as shown, giving in area 1234 the useful work developed. The compression line, 1-2, and the expansion line, 3-4, are again assumed adiabatic. Expansion is carried to exhaust pressure. The quantity of heat Q is received along 2-3, the amount q is rejected along 4-1, at constant pressure in both cases. 70 INTERNAL COMBUSTION ENGINES Let the charge weight be G pounds made up as in the pre- vious ease. Then Q = GC p (T 3 - rjB.T.U. and q = GC p (T, - 7\) B. T. U. Her E _Q~q_ ~ GC P (T 3 -T 2 ) _ l T 4 -T, (18) But from the adiabatic law we may write (19) and (20) Divide (19) by (20), /t>A Y /uA Y or (21) Also, J^L _ PjP or since p* = p 4 , = ^ (22) rtl FJ~1 ' 1 A i / rry fT\ ^ / * 1 * 4 ^1^4 and similarly ^ = ^3 (23) From (22) ~ = Y ( 24 ) From (23) 5 = ft (25) Equations (25) and (24) in combination with (21) finally give rr\ rp ^**- rp rn Substituting (26) in (18), we have T T * J7i 1 J l_1 J 4 /Q7\ &c l -nrl fjr (21) THEORETICAL CYCLES 71 But it has already been shown, equation (13), that for adiabatic compression so that finally the cyclic efficiency for combustion at constant pressure is .. Ec=l ~ r -^r V < 28 > which is the same as for combustion at constant volume. The Diesel Cycle of to-day. The Diesel cycle of to-day approximates the constant-pressure form outlined in Fig. 3-3. Compression line 1-2 and expansion j. FIG. 3-3. line 3-4 are assumed adiabatic. Heat is received at constant pressure along line 2-3, and rejected at constant volume along line 4-1. In the Otto cycle, with the machine at full load, the ratio of compression is equal to the ratio of expansion. In the Diesel, the ratio of compression , (see Fig. 3-3), is always greater V c than the ration of expansion, . Ve Let the charge weight again be G pounds. Diesel machines, as constructed, are oil engines, so that the increase of charge weight along line 2-3 is small and 'may be neglected without serious error. That is, we may assume G constant for the cycle. If the investigation, however, is carried through for a gas, especially a lean gas, this assumption is not permissible. 72 INTERNAL COMBUSTION ENGINES To develop the efficiency formula, we may write Q - GC p (T a - T 2 ). (29) q ^GC^Tt-TJ. (30) Now from Fig. 33 & T"* '7^ ^ '7~' ^ ^Q 1 ^ /TI /T7 3 2 .. 2 \ / j. 2 3 ^ where, 8 = ratio of cut-off volume to clearance volume. Also Pf = |l from which T 4 =T^ (32) But from the adiabatic law p 4 v 4 v = p 3 v s > from which p 4 = and ^9,7?^ = p 2 v 2 "* from which p t = Hence (32) may be written Substituting (31) and (33) in (29) and (30) respectively, Q = GC P T 2 (8-1). q =GC V T V (8 V -1). The cyclic efficiency for the Diesel cycle consequently is _ Q-g _ q _ GC y g\(y-l) - 7 1 1 But it has shown, equation (13), that -^ = = y . Y _ 1 .. fuNlVERSlTY ^V ^ THEORETICAL CYCLES 73 hence finally l) (34) Equation (34) shows that the expression for the theoretical efficiency of the Diesel cycle is the same as that for the Otto and 8 7 -l the Brayton, with the exception of the factor^ -- . The effi- ciency thus depends not only upon r and y, but also upon 8, that is, finally, on the volume at cut-off. In actual practice the cut-off volume v e is about 10 per cent of the stroke at full load. With a compression ratio of r = 13, this makes 8 about = 2.5. To show the influence of the factor 8 upon E c , assume y = 1.35. The efficiencies for full load, 8 = 2.5, an overload, 8 = 3.0, and some partial load, 8 = 1.5, will be as follows: Cyclic Efficiency E c for the Diesel cycle. For 8 = Ve . = 1.5 2.5 3.0 v c and for r = 13, and y 1.35, E c = .560 .509 .487. It appears from this that, other conditions remaining the same, the smaller the value of 8, the greater E c . This result is actually borne out in practice, within limits, where a large num- ber of tests of Diessl engines have often shown a greater thermal efficiency at three-quarters than at full load. That this condition does not hold for still lower loads is due to other circumstances. 4. Comparison of Various Cycles. The question of the best gas-engine cycle has often been dis- cussed. In general, there is no best gas-engine cycle, but that cycle should be chosen which will give the best return for the practical conditions existing. To give some insight into the problem of choosing the cycle best adapted to given conditions, we will first obtain some theo- retical basis of comparison, and show afterwards how this is affected in practice. Many methods of comparison have been employed by various writers, but the following, due originally to E. Meyer * seems to the writer to be the clearest and most comprehensive. *Zeitschrift des Vereins deutscher Ingcnieure, 1897, p-. 1108, 74 INTERNAL COMBUSTION ENGINES Using the Carnot cycle as a basis of comparison, it is clear that the amount of heat transferred into mechanical work de- pends only upon the total amount of heat and the temperature difference in the cycle. But to become available as a basis for the comparison of gas-engine cycles, only that Carnot cycle can be used for which the heat element 8^, supplied at a constant temperature 7\, and the heat element 8g 2 rejected at a constant temperature T 2 , are of infinitesimal amount, so that the two adiabatics forming the rest of the cycle are infinitely close to- gether. Then every closed cycle may, by a number of adiabatics, be divided into an infinite number of elementary cycles, Fig. 3-4, FIG. 3-4. for each of which we may with very small error assume that the heat element 8^ is supplied at a constant temperature T lf and the heat element 8 V' > V dq 2 ' = dq 2 " = dq 2 f " Ag/< Ag/< Ag/" From all this it follows that, starting with the same pressure of compression, the constant-volume combustion is more efficient than that at constant pressure; and this in turn is more efficient than isothermal combustion. Isothermal combustion, being so obviously inferior to the other two in theory, is also difficult to carry into operation prac- tically, and for these reasons is practically obsolete. THEORETICAL CYCLES 79 5. Practical conditions affecting the choice of best cycle for any given case. It is evident from the preceding article that the most impor- tant item in the efficiency question is the pressure at the end of compression. The aim should be to use as high a pressure as possible in order to introduce the first heat element at the highest possible efficiency. But the compression pressure governs the maximum pressure and temperature occurring in the cycle, and that brings us to a consideration of the pressure and temperature limits. Of these two the temperature limit plays but a secondary part, because it is possible to operate on a cycle whose maximum temperature may be 3000 degrees Fahrenheit for the reason that FIG. 3-9. these maximum temperatures exist but for a very short time. The pressure, limit is more important, because no matter how short a time the maximum pressure lasts, the driving mechanism of the machine must be built for this pressure. In modern prac- tice 550-600 pounds seems to be about the maximum pressure limit that can be economically handled. Herein we find a condi- tion which may modify the conclusion arrived at in the previous article as regards comparative efficiency of combustion at con- stant volume and at constant pressure. Assuming the more practical condition, that the maximum pressures in the two cycles, instead of the compression pressures, shall be the same, the diagrams would be placed as shown in Fig. 3-9, in which the broken line represents the Otto cycle. It is evident in such a case that only the last heat element of the 80 INTERNAL COMBUSTION ENGINES Otto cycle is introduced at the same efficiency as the first, and consequently all, of the heat elements of the cycle at constant volume. Hence, upon these premises, the constant-pressure cycle is more efficient than the constant volume cycle. Under some conditions, however, favorable to the Otto cycle, this advan- tage of the constant-pressure cycle is not very great. Thus in a Diesel engine cutting off at full load at 10 per cent, 8 = 2.5, assume r = 13, and y = 1.41, then E c will be .564. An Otto engine cycle of the same maximum pressure limit, = about 460 pounds, would show a compression pressure of from 190-225 pounds, r would be about equal to 7, and with y = 1.41, E c would be .550. This shows a gain for the Diesel engine of only 1.5 per cent, but it should be pointed out that a compression pressure of 200 pounds in an Otto cycle can only be reached with extremely lean fuel gases, or with separate fuel and air compression. However that may be, the advantage of the constant-pressure cycle over the constant-volume cycle, presupposing equal maximum pres- sures in each, is comparatively small, and hence the undoubted gain that the Diesel engine shows in practice over the average constant-volume engine is by some writers attributed not so much to the cycle as to the greater perfection of combustion. That this is approximately true has been proven by Giildner who constructed engines operating on the Otto cycle upon the most advanced ideas, and obtained efficiencies fully as good as those obtained by Diesel. A second limit set to the compression pressure that can be carried is due to pre-ignition. All fuel mixtures will ignite spon- taneously if the temperature becomes high enough, but the critical temperature varies greatly for the different fuels. This fact directly governs the compression pressure. While it is hardly advisable to use more than say 80-90 pounds in the case of a gasoline mixture, a blast furnace gas mixture will easily stand 150 pounds without pre-ignition. Since the efficiency of the cycle depends directly upon the compression pressure, as above shown, we should expect a better efficiency for blast furnace gas than for gasoline, and this is actually so in practice. The difference, however, is due to the nature of the fuel. A remedy for this state of affairs would be to compress the air separately and introduce the gasoline or other fuel oil only at the moment THEORETICAL CYCLES 81 combustion is desired. This leads to the constant-pressure cycle, and thus the fuel to be employed should be considered in the choice of cycle. A final point to be considered is this: By theory, the greater the compression pressure, the higher the efficiency. This holds for either type of cycle. But how far can this compression be carried, outside of the questions of upper pressure limit and pre- ignition above considered, before the added gain due to higher compression is in practice balanced or overbalanced by attendant losses. The following discussion of this question by Giildner, as applied to combustion at constant volume, is instructive. The efficiency of any engine should not be judged upon cylin- der performance, but upon the performance at the shaft. Usually this is called the " thermal efficiency per brake horse-power," but a shorter and more expressive term which the authors prefer and will use in this treatise, is " Economic Efficiency E e ." Now E. = E,E m = XE c E m . where E t = thermal efficiency per indicated horse-power. u ' i- m r Brake H. P. E m = mechanical efficiency of engine = - E c = cyclic efficiency as derived in the preceding articles = = 1 --- - for combustion at constant volume. 7-1 r and X = a factor such that E t = XE C , so that X is always less than 1. The mechanical efficiency may be expressed by Pi where Pf = mean indicated pressure per square inch of piston, and PI = pressure lost in friction per square inch of piston. 82 INTERNAL COMBUSTION ENGINES Hence Pi - Pf Pi The value of E e therefore depends upon the relation between the factors X, p i} and p f , as modified by a variation in r, that is, in the pressure of compression. The change in the value of X due to a variation in r is quite unknown, and is in any case so small that it may be neglected. Regarding the relation between p { , p { , and r, numerous tests have shown that, as r increases, the mechanical efficiency ^ y , is at first quite con- stant, but that after a certain point it commences to decrease quite rapidly. This is due to the fact that for the first part of the range any increase in p f , due to an increase in r, is counter- balanced by a corresponding gain in p^ Beyond a certain point, however, there is a loss in p., owing to the fact that the fuel mixture has to be made more and more lean to prevent pre-ignition, and the mechanical efficiency consequently ^de- creases. Hence we have the net result that as long as an increase in p } , due to an increase in r, is met by a proportionate gain in pi, the Economic Efficiency E e will increase. Just as soon, how- ever, as, with increase in r, the gain in pi is no longer sufficient to overcome the increase in p/, E e will commence to decrease, and the economic compression limit will have been passed. Tests and computations have shown that up to r = 6, which means a compression pressure of about 160 pounds, the mechani- cal efficiency does not change materially, but that beyond this it commences to decrease. This is probably due to the necessary increase in the size of machine parts due to the greater pressures and to the fact that leaner mixtures than would ordinarily be used must be employed to prevent pre-ignition, with a correspond- ing loss in the value of p { . Beyond r = 6 the gain in E e is small and at r = 10 it practically ceases to increase. Hence we con- clude that for combustion at constant volume the use of values of r greater than about 8, for which the compression pressure equals about 225 pounds, is no longer accompanied by any useful gain in the economic efficiency of the machine. This statement does not apply to combustion at constant pressure because its E c follows THEORETICAL CYCLES 83 a different law, and E m decreases more slowly on account of the generally greater value of p.. Finally, in laying out an Otto cycle, to obtain high engine capacity per unit volume of cylinder means making the maxi- mum pressure as high as possible to obtain a high value of p.. With fuel mixtures which cannot stand a high amount of com- pression this would mean the use of rich gas mixtures. Where j ^ ^ e max. pressure the compression can be made higher, the ratio of - , comp. pressure instead of being from 4-5, may be profitably made from 2.5 to 3 by the use of leaner mixtures, so that the maximum pressure shall be in the neighborhood of say 450 pounds. The chances are, as the gas engine develops, that higher maximum pressures will be employed, but according to Giildner the use of working pressures exceeding 600 pounds is neither economical nor safe. CHAPTER IV THE VARIOUS EVENTS OF THE CONSTANT VOLUME AND THE CONSTANT-PRESSURE CYCLES AS MODIFIED BY PRACTICAL CONDITIONS 1. IN the previous chapter the various cycles were discussed and compared on theoretical grounds. For this purpose several things were assumed which in practice are only approximately true; thus compression and expansion lines were assumed adi- abatic and the surrounding walls impermeable to heat. It is, how- ever, true that the heat interchange between the charge and the walls may be such as to give a line on a diagram which strictly follows the adiabatic law. Such a line is by some writers called a false or pseudo-adiabatic. Again it has been assumed in the theo- retical discussion that ignition is perfect, that the composition of the charge is uniform, and that combustion is complete and per- fect ; none of these things quite obtain in practice and all the va- riations have their influence upon engine performance. Thus it happens that the cyclic efficiency above computed is in any given case never realized, but that the actual thermal efficiency is al- ways less than E c . The following paragraphs will point dut these modifications more in detail. 2. The four-stroke Otto cycle. (a) THE SUCTION STROKE. At the end of the exhaust stroke, the clearance volume V c , Fig. 4-1, is filled \\ith burned gases under a pressure p e and a temperature T ' e . The weight of these gases can only be approximately computed, since nothing definite is known of the temperature T e . It is in most cases probably between 12-1400 degrees Fahrenheit, while the pressure p e in well-designed machines may be from 16-18 pounds absolute, but circumstances may alter these figures considerably. At the commencement of the suction stroke the pressure falls from p e to the suction pressure p s along a curve deter- mined by the re-expansion of the burned gases in the clearance spaces. Only after this re-expansion will the fresh charge be drawn into the cylinder. It is thus seen that the 84 THEORETICAL CYCLES MODIFIED BY PRACTICE 85 volumetric efficiency, E v , of the cylinder, that is, the ratio volume of fresh mixture , , . '. , 7 r j; \ -,-> depends directly upon the weight volume of piston displacement of burned gases remaining, thus affecting cylinder capacity. If through bad form of combustion chamber, too small an exhaust opening, or a restricted exhaust pipe, the exhaust pressure should be kept too high, or too much burned gas remain behind, this effect of re-expansion will be more marked than above stated. A too early closure of the exhaust valve may have the effect shown in Fig. 4-1 in dotted line. It is evident, however, that in the ordinary four-cycle engine without scavenging this loss of volumetric effi- ciency will always be present, depend ing upon the clearance volume. FIG. 4-1. A second factor affecting the volumetric efficiency of the cylinder is the suction pressure p s . At the end of the suction stroke the cylinder contains a volume of gas, v t , made up partly of burned gas and partly of fresh mixture, under a pressure p s . The compression stroke raises this amount of gas to the pressure p c with a volume v c . The compression curve crosses the atmos- pheric line when the stroke volume is only v s , and not the full volume v r . Hence the volume v represents a loss in volumetric efficiency, and this loss is the greater the smaller p s . It follows that the inlet pipes and valves should be so designed as to cause a minimum suction Bunder-pressure," that is, to keep p s as close to atmospheric pressure as possible. This also makes clear why vaporizers and carbureters always decrease engine capacity somewhat. 86 INTERNAL COMBUSTION ENGINES Thus there is a loss of volumetric efficiency, and hence of engine capacity, at each end of the suction line. The real volu- metric efficiency, E v , is found in any given case by dividing the volume represented by the line a-b, measured along the atmos- pheric line, by the volume v s of the piston displacement. Attention should at this point be called to the fact that cer- tain systems of speed regulations depend upon the variation in the suction pressure. By throttling the mixture from the begin- ning of the stroke, or by cutting off the supply completely at a given point in the stroke, p s is increased or decreased depending on the load on the engine, thus directly controlling the charge volume, and hence the engine capacity. This is explained more in detail in the chapter on governing. Since it is quite evident that neither the loss by re-expansion or that due to the suction pressure at the end of the suction stroke can be entirely avoided, it becomes interesting, at least from the point of design, to know approximately \vhat values of E v to expect in different types of engines. Naturally E v decreases as engine speed increases, because since high-speed engines are usually small engines, the difficulty of placing valve openings large enough to prevent serious throttling of the charge becomes greater as speeds increase. Computations are of little avail in this matter since little is definitely known of the temperature of the charge at the end of the suction stroke. For that reason more reliance is to be placed in figures based upon practical experience. The following table is due to Giildner: E v p s Ibs. persq. inch absolute 1. blow-speed engines with mechanically oper- ated inlet valve 88 - .93 12.9 - 13.7 2. Slow-speed engines with automatic inlet valve .80 - .87 12.5 - 13.2 3. High-speed engines with mechanically oper- ated inlet valve 78 - .85 11.7 - 12.5 4. High-speed engines with automatic inlet valve .65 .75 11.4 12.2 5. Very high-speed engines with automatic inlet valves and air cooling 50 - .65 8.8-11.0 Suction gas generators and vaporizers may in unfavorable cases decrease the above figures for E v as much as .05. (6) THE COMPRESSION STROKE. The compression line may THEORETICAL CYCLES MODIFIED BY PRACTICE 87 be taken to follow the general law pv n = constant. During the first part of the stroke there is probably a flow of heat from the walls to the comparatively cool charge, but this is soon over- balanced by the heat of compression, so that during the last and greatest part of the stroke the flow is into the walls. The com- y pression curve is therefore rarely an adiabatic pv = constant, C 1 where y = ~ , as computed for the charge. In most cases the L> v line is intermediate between an adiabatic and an isothermal, and strictly also the value of the exponent, n, is not constant along the entire line. The value of n in actual cases lies between 1.30 and 1.38, with an average of about 1.35. In cases of very ineffective C 1 cooling n may exceed y = -~ . It should also be noted that C w leaky pistons and valves cause a flattening of the compression curve, which apparently decreases the true value of n. From the equation pv n -= constant, we may derive the follow- ing equation for the absolute pressure p c at the end of compres- sion, see Fig. 1. The absolute temperature at the pressure p c will be y = T I l n m n-. n This equation requires an assumption for the value of T s , the temperature at the end of the suction stroke. As already stated, not a great deal is known about this. S. A. Moss states that experiments have shown it to be between 200-300 degrees Fahrenheit, that is, 660-760 degrees absolute, but gives no details. Schottler in his examples on type-cycles assumes in most cases 350 degrees Centigrade absolute, = 632 degrees Fahrenheit absolute. The clearance volume required to produce a pressure p c and a temperature T c will be The following table shows values for the absolute compression pressure, p c and the end temperature T c for various values of n, of r, the ratio of compression, and of T s . The value of p s has been assumed at 12.5 pounds absolute: 88 INTERNAL COMBUSTION ENGINES _. oo o p o oo co o 10 S ' ' 00 CO - P 00 CM *-> rH O O rH CM CM ^ '- 1 ^ ^ 01 oo r>. CD iO O rH Ca CO Tt< CO (M O OO 00 S P O P to OO O to O t^l CO CM CM rH rH t^ rH CM CO "tf IO ^ ^_ ^H rH rH rH "5 CO 00 O ro b* Tf 1 CS i^ o? o * 00 00 00 T* O5 OO t^ CD to CO O r- CM CO TJI C\) rH rH rH rH rH i 10 co o o I-H O5 l^ iO ^ 00 l> o to oo tO rH Oi 00 CD Tfl in o cr> rH CM co illll o OS "H 2 2 3 fe g s OS 05 2 2 CO CO Tfi CO Jo tt '- | ^ ^ rH ,-H CM 188111 ^*: CM 00 * Q> Oi tO CM t>. i>- oo GJ o o o co 5 b- CO O O t^ oo os O5 CO O g- O rH IO t^ rH t^ CO 00 00 05 CM t- CD CO CM 00 - !> 00 <^ O CO * 1C ^i * O CO 00 00 Oi (N CO t- CM CO $2 00 Tt< rH t^ ^ t- 00 05 05 ^ t^ CO CO I-H O t^ O O CD t^ 00 00 O5 P X 00 00 _ T-H t^* CO CO t"- t^ 00 ^ ^ CO l>- I> 00 O5 O5 & *O O iO CO I s " !> 000 iO O O co t t> s 1 o o >o co i>- i^ tO C >O CD t^^ t^* THEORETICAL CYCLES MODIFIED BY PRACTICE 89 It has already been shown in the previous chapter how the cyclic efficiency and consequently also the thermal efficiency of an engine depends upon the compression pressure. It has also been shown that there are commercial limits to the compression pressure due to pre-ignition of charge. As regards pre-ignition of charge, which is to be distinguished from back firing or explo- sions in the exhaust pipe, the greater danger of this exists with fuel mixtures high in hydrogen. Hydrogen, next to acetylene, possesses the lowest ignition point of any of the gaseous fuels commonly employed. Hence illuminating gas with about 45 per cent by volume of hydrogen is much more liable to pre-ex- plosion under the temperature of compression than a producer gas with 15 per cent. Hence we may use a higher compression pressure in the case of the latter gas, and may expect, and actually obtain, a higher thermal efficiency in practice. With gases very rich in CO and low in hydrogen there is little danger of pre-igni- tion even up to the commercial limit of high pressures. Lucke estimates that for every 5 per cent of hydrogen that the gas con- tains 15 pounds should be subtracted from the otherwise allow- able compression pressure. In general it should be borne in mind, and this applies to all fuel mixtures, the better and more effective the cooling of the cylinder, the higher can be the com- pression without danger of pre-ignition. Anything which draws down the temperature during compression, as water injection, is also favorable to the same thing. A case in point is Banki's method of water injection with gasoline as fuel, b}' which means compression pressures could be employed which gave thermal efficiency results equal to the best obtained on lean gases. Premature explosions are sometimes directly due to faulty design of the combustion chamber. Any projecting point or edge in the chamber which cannot be effectively water-cooled may become red hot under compression, and thus locally raise the temperature high enough to cause pre-ignition. This action is so certain that it has been proposed to use it as a method of ignition by placing a projection on the piston face. It was found that while this scheme would work, it was not susceptible of con- trol, and was abandoned. The following table shows the safe compression pressures in use with the fuels commonly employed, as given by Lucke, The 90 INTERNAL COMBUSTION ENGINES second column gives the percentage of clearance required to produce the pressure p c in terms of the piston displacement. This column has been computed assuming p s = 13 Ibs. absolute and n = 1.35. Fuel Compression Pressure Lbs. by gage % Clearance in terms of Piston Displacement Gasoline: Auto-engines with carbureters, cooling not very effective high speeds. . 45 - 95 Ave 65 35 Gasoline : Stationary, slow-speed engines, more effec- tive cooling but usually less simple com- bustion chamber 60 - 85 Ave 70 32 Kerosene : Hot bulb injection and ignition . . . 30 75 35 40 Kerosene: Previously vaporized in devices not requir- ing a vacuum . . 45 - 85 Ave 65 35 Natural Gas Natural Gas : Average for large and medium engines. . . Illuminating gas . 75 - 130 Ave., 115 60 100 22 Producer Gas . Ave., 80 100 - 160 26 Producer Gas: In large engines water-cooled on pistons and valves Ave., 135 20 Blast Furnace Gas 120 - 190 Ave., 155 17 (c) THE COMBUSTION LINE. The shape of the combustion line depends primarily upon the interrelation of three things: composition of charge, point of ignition, and piston speed. For every fuel there is a certain fuel-air mixture which gives the greatest rate of flame propagation, i.e., the most rapid com- bustion. Any further admixture of neutral gases, whether these be air or burned gases, results in a slower combustion, until there comes a time when ignition fails. Suppose, therefore, that in any given engine with full throttle, constant speed, and proper igni- THEORETICAL CYCLES MODIFIED BY PRACTICE 91 tion, we obtain diagram, Fig. 4-2, a. Now if the throttle is partly closed, making the dilution of the charge by the burned gases greater than before, card, Fig. 4-2, 6, results. Further closure of the throttle makes the combustion still slower, Fig. 4-2, c. Similar effects would have been obtained with full throttle if the proportion of air in the fresh charge be seriously increased. For proper combustion the time of ignition should be so chosen that the combustion line is vertical or nearly so. This means that every different fuel mixture and every different piston speed will have its own proper point of ignition. For that reason the ignition apparatus in every engine should be made adjustable, because the only way to determine the proper time is by trial. For small gasoline engines this adjusting may be done by ear, since the proper point of ignition corresponds nearly with the highest speed, for a given throttle position. For more accurate work the indicator, preferably with a constant-speed drum motion, should be used. It will in general be found when ignition is right that it occurs some time before the piston has reached the dead center, the amount of this lead de- pending upon the mixture and the piston speed as above stated. What im- properly timed ignition results in is shown in a diagram, Fig. 4-3, given by Clerk. With proper ignition the normal diagram is indicated by a. As the time of ignition is made later and later, cards 6, c and d result, in the last case flame propagation starting so FIG. 4-2. 92 INTERNAL COMBUSTION ENGINES late that it barely overtakes the piston before the end of the stroke. If instead of changing the time of ignition the piston speed had been increased, effects very similar to those of Fig. 4-3 would have been obtained. Thus the dependence of the shape of the combustion line upon the three factors mentioned at the outset is clear. The maximum pressure attained during combustion depends upon the heating value of the charge. With all conditions favor- able, the maximum pressure should be reached at or before one- tenth stroke. The rich fuel mixtures, those for illuminating gas, natural gas, gasoline, etc., show rapid combustion. They FIG. 4-3. cannot in general stand high compression, and the ratio of maximum pressure , ... , simply called the pressure ratio, is usually compression pressure high, between 3 and 5. The naturally leaner fuels like producer gas and blast furnace gas ignite better when highly compressed, but on account of the generally low heating value of their fuel mixtures the pressure ratio for these gases is usually less, between 1.5 and 3. The maximum explosion pressure p x , Fig. 4-1, even assuming complete combustion at constant volume, is in no actual case as high as that computed on theoretical grounds for the heat received by the cycle. There may be several reasons for this. One is un- doubtedly the loss of heat to the jacket water during explosion. This amount of heat is a dead loss since it does not even enter the cycle. It is something, however, which cannot be avoided, since cooling is a necessity for other reasons. Another is due to the undoubted fact that the specific heat of the gases increases THEORETICAL CYCLES MODIFIED BY PRACTICE 93 with the temperature. We can not, however, as yet mathemati- cally gage this effect, because nothing definite is known of the law of increase.* Another theory to account for the failure to realize theoreti- cally maximum pressure assumes that combustion is not com- plete, that is, that not all of the heat of the charge is liberated when the piston starts forward. This results in after-burning, which will be considered later. It is quite likely that in most cases these three things combine to keep the observed pressure below that calculated. Assume that the combustion takes place at constant volume, and let p x and V c , Fig. 4-1, be the pressure and volume at the end of the explosion. mi x , m Then p x = ^ and T x = PC If the combustion line is other than vertical, as indicated by dotted line in Fig. 4-1, let p x , T x ' and V x ' be the data for the end of the combustion. The above equations then become If in the above equations the values of p x or p x , are taken from actual diagrams, the equations for T x or T x l will give real temperatures. But if p x should be computed from one of the theoretical diagrams of the previous chapter, then the value of T x should be multiplied by a factor which expresses how much the real value of p x falls below the theoretical value of p x due to imperfections of combustion. This factor is approximately equal to the ratio Indicated thermal efficiency cyclic efficiency An idea of the maximum pressures p x or p/ existing in the cycle may be gained by considering that in most cases the ratio can be made equal to 3. Turning to the table, page 88, we PC * See Chapter X. 94 INTERNAL COMBUSTION ENGINES see that for n = 1.35, and r = 6, p c is equal to about 140 Ib. absolute. Thus p x would be about 3 X 140 = 420 Ib. The temperature T ' x , assuming T s = 700 degrees, which for the same values of n and r, makes T =1114, would then be about T x = 42 ** 114 = 3342 F. absolute. J.4U The following figures give a few of the characteristic diagrams for various fuels. FIG. 4-4. FIG. 4-4. From Struthers-Wells hit-and-miss engine. ll|"x 18", 30 H.P., 200 r.p.m., natural gas. Card good throughout, compression 80 Ib., max. pressure, 320 Ib. Pressure ratio W = 3.5. FIG. 4-5. FIG. 4-5. From Struthers-Wells automatic engine, 16$" x 22", 150 H.P., 200 r.p.m., natural gas, full load. Card good, compression 96 Ib., max., pressure 330 Ib. Pressure ratio ftf = 3.1. THEORETICAL CYCLES MODIFIED BY PRACTICE 95 FIG. 4-6. From Stock- port engine. Given by Clerk, the Gas and Oil Engine, p. 321. 9f" x 17", 9 h.p., 182 r.p.m. Illuminat- ing gas. Compression 90 lb., max. pressure 270 lb. Pressure ratio = 2.7. FIG. 4-6. FIG. 4-7. FromHornsby- Akroyd kerosene en- gine. 6 H.P., 225 r.p.m., Card at about J load. Hot bulb vaporization, hence pressures low. Com- pression 45 Ibs., maxi- mum pressure 116 lb. Pressure ratio = VV = 2.2. FIG. 4-7. FIG. 4-8. Card from a gasoline engine given by Lucke, Gas En- gine Design, p. 71. Vapor prepared outside cylin- der. Compression 80 lb., maximum pressure 372 lb. Pressure ratio V/ = 4.07. FIG. 4-8. 96 INTERNAL COMBUSTION ENGINES FIG. 4-9. FIG. 4-9. Taken from Koerting engine. 700 H.P. Blast furnace gas. Max. pressure 242 lb., compression pressure 127 lb., Pressure ratio = m = 1.8. FIG. 4-10. FIG. 4-10. From American Crossley producer gas engine, given by Langton. 18J x 24", 65 K.W.. 200 r.p.m. Hit-and-miss governor. Gas rather high in H and low in CO, hence compression low. Compression 83 lb., maximum pressure 248 lb. Pressure ratio W = 2.7. The following two diagrams show abnormal conditions: FIG. 4-11. FIG. 4-11. Pronounced case of pre-ignition in 6 H.P. Hornsby-Akroyd kerosene engine due to too high compression. This was cured by the addition of a little water to the charge. THEORETICAL CYCLES MODIFIED BY PRACTICE 97 FIG. 4-12. FIG. 4-12. Case of back-firing as distinguished from pre-ignition. Power, Oct. 15, 1900. Explosion in the suction pipe at a during the suction stroke. Observed in an engine which attempted scavenging by means of os- cillations of burned gases in the exhaust pipe. Back-firing more often occurs in the exhaust pipe due to accumulation of unburned gas. (d) THE EXPANSION LINE. The expansion line, like the compression line, may be taken to follow a general law p v n = C constant, in which n is rarely equal to y= *, for the burned CMP gases. Assuming that combustion is complete when the piston starts forward, the loss of heat to the jacket during expansion should cause the expansion line to lie below the adiabatic, that is, n should be greater than y. (It should be remembered in this connection that leaky pistons and valves cause an increase in the true value of n.) Now it is very often found that the expansion line falls off much more slowly than this, coinciding now and then with the adiabatic, but very often lying between this and the isothermal, that is, n is less than y. The explanation of this phenomenon has been held to be an evolution of heat along the. expansion line equal to or exceeding the loss of heat to the jacket during this period. Several theories have been advanced to explain this. The older machines of Lenoir and Hugon showed values of the exponent n for the expansion line between 1.4 and 1.6, while the earlier Otto machines showed values approximating 1.3. To explain this, Otto, and after him, Slaby, at least for a time, advanced the theory of stratification. It was supposed that the charge of a 4-cycle machine could be so arranged as to have practically nothing but burned gases against the piston, next practically air, then a layer of poor mixture, and finally near the igniter the fuel mixture in its full strength. It was further 98 INTERNAL COMBUSTION ENGINES assumed that this arrangement was disturbed but little during compression. Hence ignition was sure; but as the combustion progressed and reached the leaner layers of mixture, it became more and more slow, was not completed when the piston started forward, and was continued along the expansion line, showing a so-called after-burning. Thus Otto attempted to explain the somewhat slower rise of the combustion line and the absence of the serious shock at the moment of explosion in his gas engines. In the fight against the Otto patent, tests were made by the Deutz Company, by Dewar and by Teichman, all of which seem to support Otto's claim of stratification. Slaby defended it vigorously. It has, however, been pretty clearly shown to-day that, while stratification is not at all impossible, in fact it is not easy to get a uniform mixture, it cannot have any marked effect upon economy or performance; that is, the diagrams would not be very far different. The opinion of the best writers of to-day leans toward the requirements of most uniform mixture and rapid combustion. Clerk has attempted to show that some heat is " suppressed" during the combustion period in every gas engine and released along the expansion line, although to attempt to express these quantities in thermal units would seem superfluous in view of the fact that we know nothing definite of the variation of specific heat at high temperatures. It would seem, therefore, that after-burning, if it exists, is not a peculiarity common to Otto engines only. It is merely evidenced more strongly in the Otto diagrams by the fact of higher piston speeds and porportion- ately smaller enveloping surface, giving less time and opportunity for heat losses along the expansion line, as compared with gas engines before Otto's time. Hence, as Clerk puts it, the slow dropping of the expansion line is not the cause of the greater economy of the Otto engine, but rather the effect and evidence of it. The second explanation of the supposed after-burning rests on the so-called dissociation theory. It has been shown by Bunsen that a composite gas breaks up into its elements when the tem- perature exceeds certain limits. Conversely, chemical combina- tions, such as combustion, can no longer take place when this temperature limit is reached. This theory applied to the gas engine would mean that at the inner piston position, if the tem- perature rises to the limit, combustion no longer takes place, THEORETICAL CYCLES MODIFIED BY PRACTICE 99 but just as soon as the piston starts forward, resulting in a drop of both temperature and pressure, combustion again ensues and is, by the same method, continued along the expansion line until no combustible remains. Clerk strongly leans to this view of the matter. But it has been pretty definitely shown that the tem- peratures in gas engines rarely exceed 28-3000 degrees Fahrenheit, and that these temperatures are below the dissociation limit. It has also been pointed out by Schottler that if this theory holds good the expansion line should be an isothermal as long as there is com- bustion. It may therefore be concluded that dissociation plays but a small part in the combustion phenomena found in gas engines. Witz, in some tests made with an engine whose piston speeds could be varied, found that in each case after-burning occurred, but that the combustion was the more rapid and the maximum pressure the higher (that is, after-burning the less noticeable), the greater the piston speed and the warmer the jacket walls. Thus he concludes that after-burning largely depends upon the influence of the walls. Slaby, corroborated by E. Meyer, on the basis of other tests, has tried to show that these conclusions of Witz are not generally applicable, but on weighing the evidence, and in the light of later achievements, it must be concluded that they and the principle based upon them, i.e., rapid combustion of the leanest possible mixture at the greatest possible piston speed, at least represent a step in the right direction for gas engine economy. From the above it is quite evident that none of the theories advanced explain satisfactorily all of the phases of the question of after-burning, the occurrence of which must be held as proven, especially in the light of later tests. Schottler, indeed, has offered another explanation for the so-called abnormal position of the expansion line by showing that a natural solution of the question may be found in the variation of the specific heat of the expanding gases with temperature. Some figures quoted by him show that this may be the case, but before the idea of after-burn- ing can be dispensed with, a great deal more experimental work is needed and desirable. The requirements for best efficiency of combustion and ex- pansion have already been briefly explained. In a little greater detail they are as follows: 100 INTERNAL COMBUSTION ENGINES 1. Highest possible compression pressure before ignition. The effect of this is, (a) less admixture of burned gases to the fresh charge, (b) less loss to jacket because smaller volume is involved, (c) greater mean effective pressure, (d) greater ease of ignition of charge. 2. Pure and uniform mixture and rapid combustion to avoid after-burning. The bad effect of after-burning is due to the great jacket loss along the expansion line, and its effect has been com- pared by Koerting to that of a leaky valve in a steam engine. 3. Avoid external cooling. This of course cannot be entirely eliminated. But since the amount of heat lost by cooling is a function of both time and superficial surface, this requirement calls for high piston speeds and a form of cylinder in which the superficial surface . ratio - - is the smallest possible, volume To return to the expansion line, at the moment the exhaust valve opens, the gases have expanded to a volume V y , with a pressure p y , and a temperature T yJ see Fig. 4-1. We may write The ratio is the real ratio of expansion.' The expressions for ^c the value of p y and T y for the constant-pressure cycle are anal- ogous; care should be taken, however, to use the proper ratio of expansion, which in this case is quite different from the ratio of compression. In the case of the Otto cycle the ratio of ex- pansion is in most instances nearly as great as the ratio of compression r, arid we may therefore write, from the above ex- pression for p py - Px r n It appears from this equation that, with p x remaining the same, the terminal pressure decreases as the compression in- creases. In practice, however, the terminal pressure in most cases shows an increase as compression is increased, due, no doubt, to the fact that p x also increases with compression, unless the mixture is made correspondingly leaner. (e) THE EXHAUST STROKE. The velocity of efflux of gas at the instant the exhaust valve opens is very high, approximat- ing 25-3500 ft. per second. The valve should start to open at about THEORETICAL CYCLES MODIFIED T 9 = entropy. It may be remembered from the statements in Chap. II that When heat is supplied to a mass of gas the volume, pressure, and temperature of the gas may vary simultaneously. Hence we may write the energy changes occurring under the conditions in the following general terms, following Grover: Additional internal f External effect of 1 Addition of heat may produce energy of the gas in- volving rise of tern- \ -|- perature. work done by the gas expanding be- tween its contain- ing walls. With the notation above given we may write this + ySF (2) * Grover, Modern Gas and Oil Engines. 110 INTERNAL COMBUSTION ENGINES but 2? = R = J(C, - Cy (3) /TT hence p = J(C P C v }-- (4) v Substituting (4) for p in equation (2) & (cv - c> r-^ (5 and r S* = f =C.-^ + lum . Cubic Fe it .04 .08 .12 .10 .20 .24, .28 .32 ^36 JO & 3440 31804 323.J KiC5- r l,T85 1058- 1005 te- Figure indicH \ diagra Tem'p. *F .50 .00 .6^ 4 08 32 .70 FIG. 5-1. 2. Heating value of the fuel is found by computation to be 617.5 B. T. U. per cubic foot. Heating value of the charge as computed for the weight of gas in a charge is 29.06 B. T. U. THE TEMPERATURE-ENTROPY DIAGRAM 115 3. The absolute pressure at the end of the suction stroke, see diagram, Fig. 5-1, was 12.3 Ib. Brooks and Stewart calcu- lated the absolute temperature for the same point at 739 degrees Fahrenheit. By using the equation T T l temperatures were computed for various points around the cycle as indicated on the card. In Fig. 5-1 the full line shows the actual card, enlarged from the original of Brooks and Stewart, while the broken line indicates the ideal cycle which receives the same amount of heat. 4. Stroke volume = .460 cu. ft. Clearance volume = .280 cu. ft. Total volume = .740 cu. ft. 5. Index for expansion and compression lines. The index of the ideal card is 1.39 for the compression and 1.37 for the expansion line as computed under (1, e). For the real diagram it is often found that the index is not a constant for the entire line. For that reason each line should be divided into a number of parts and the index determined for each. FIG. 5-2. The method of doing this is as follows: To determine the index between any two points a and b on the expansion or com- pression line, Fig. 5-2, find the volumes and absolute pressures 116 INTERNAL COMBUSTION ENGINES for each point from the diagram. Then the exponent or index in the equation pv n = constant is log p a - log p h n + -0976 12.3 739 - .196 log e T l|| - - .0978 - .0002 Having thus worked around the cycle, the error seems to be very slight in view of the fact that the exponent, n, for the pro- longed expansion line has been assumed = 1 .433. Plotting these values of entropy and temperature finally. re- sults in the diagram shown in full line in Fig. 5-3. INTERPRETATION OF THE ENTROPY DIAGRAM, FIG. 5-3. In order to evaluate the diagram it is necessary to know the number of heat units per square inch. This is most easily ob- tained by multiplying one inch of temperature scale by one inch of entropy scale, and then multiplying the result by the charge weight per cycle, since the entropy diagram is drawn for 1 pound of charge weight. In this case we have: Value of 1 square inch in B. T. U. = 800 X .04 X. 03477= 1.113. The individual areas of the original diagram were next gone over with a planimeter, and after multiplying each area by the square inch equivalent, the results were as follows: %of B. T. U. Total Heat 1. Heat received during explosion, area a A Bba 18.086 62.24 2. Heat received during expansion, area b B C D d 1.002 3.44 3. Total heat received, as shown by diagram 19.088 65.68 4. Total heat supplied as calculated (see footnote) 29.316 5. Difference in heat loss to Jacket and Radiation = area dDC BE F fd 10.228 34.89 6. Heat loss to exhaust, area gG Ddg 13.422 45.78 7. Heat loss during compression, area a A G g a 233 .79 8. Total heat lost per cycle 23.883 81.46 9. Indicated work, area A BC DG A 5.431 18.54 29.314 10000 120 INTERNAL COMBUSTION ENGINES The above results do not agree with those of Brooks and Stewart for the same test. Their results are given in the follow- ing table: Heat in indicated work ........... 17.0 per cent. Heat loss in hot gases ............. 15.5 per cent. Heat loss in water jacket .......... 52.0 per cent. Heat loss in radiation ............. 15.5 per cent. Total .............. .......... 100.0 per cent. The agreement as regards indicated work is fair. The rest of the figures of Brooks and Stewart are abnormal, in that the radiation loss is as large as the exhaust loss, and in that the jacket water loss is much too large. Brooks and Stewart them- selves admit that the jacket water loss was not accurately deter- mined. For that reason, too, the radiation loss cannot be found separately in the entropy analysis above given. If the jacket water loss had been accurately found, item 5 in the above analysis could have been separated into jacket water and radiation loss. NOTE. This quantity is greater than the latent heat energy in the gas = 29.06 B. T. U. per cycle, by the heat equivalent ofc the area A X g a = .256 B. T. U. II. GRAPHICAL CONSTRUCTION OF THE ENTROPY DIAGRAM The graphical method to be described is due to Prof. H. T. Eddy, and is by him explained in the Transactions of the Ameri- can Society of Mechanical Engineers, Vol. 21, p. 275. It is based upon the following considerations: Equation (6), p. 110, after integration may be written: Entropy difference = < t - 4> 2 = C v log e - 1 + (C P - C v ) logZ 1 (24) 1 2 * 2 Dividing equation (24) by (C p C v ) we have The problem then resolves itself into rinding graphical represen- tation for the quantities -^ log. and log. J Uf-L't) ^2 '2 The construction divides itself into two main parts : 1. The change of the pressure- volume diagram into a tempera ture-volume diagram, and THE TEMPERATURE-ENTROPY DIAGRAM 121 2. The change of the temperature-volume diagram so ob- tained into a temperature-entropy diagram. Actual p-V Diagram Ideal " Actual T-V I'deal " Actual Entropy Ideal Hi |.144 i .I'.rj ; | 1.2-10 vi.VolinieTcub"idFeeti I .04 ,08 .12 .10 .20 .24 .28 .32 .30 .40 .44 .48 .52 .50 .QO .04 122 INTERNAL COMBUSTION ENGINES 1. In Fig. 5-5 the actual and ideal p-v diagrams of Fig. 5-1 have been reproduced. Choose any convenient volume ordinate, in this case V = .38 cu. ft. has been taken, and from the points of intersection of the various pressure levels with this ordinate draw straight lines through the origin. These straight lines are constant-pressure lines in a temperature- volume field. Thus the .line A B in this case is the 60 pounds constant-pressure line. To see the reason for this consider the general equation This may be written v _R T _p T~p > T v~R rn But - is the tangent of any given angle, BAG, and hence ^ is v H constant for any point along A B. The same holds for any straight line drawn through any other pressure level. Next, to construct the temperature-volume diagram, con- sider any pressure level as E F = 100 Ib. This cuts the real diagram in the points a and 6. At a and 6 erect perpendiculars until they intersect the constant-pressure line = 100 Ib. in the points a' and b'. These will be two points in the temperature- volume diagram. In the same manner the entire diagram may be outlined. The temperature scale, which up to this point has been arbitrary, may next be determined. We know that G, the lowest point on the temperature diagram, must represent the temperature in the cycle at the end of the suction stroke = 739 degrees Fahrenheit. This at once determines the tempera- ture scale. 2. To obtain the graphical representation of the expression y \og e r 1 choose any convenient horizontal line as the zero. In *j this case the line A C has been taken. With A as a center and any radius AX, draw the arc XY to cut any convenient line, as A B, passing through the origin. From Y draw the perpendicu- lar YX'. Again with A as a center and AX' as radius, draw the arc X'Y' to intersect with A B, and draw the perpendicular Y' X". Where. these perpendicular lines YX l and Y V X" cut any successive pair of equidistant horizontal rulings will be found two THE TEMPERATURE-ENTROPY DIAGRAM 123 points on the required curve. The horizontal rulings chosen in this case are marked serially on the ordinate V = .22 cu. ft. Thus m and n are two points on the curve MN sought. It is evident that the intersections of perpendiculars YX f and Y'X" with any other two successive horizontal rulings might have been chosen as two points on the curve. This would have resulted merely in moving the curve bodily up or down on the field, the shape would have been exactly the same. For the same reason it is immaterial whether the line A C or any other horizontal line is used as the base of construction. The resulting curve is in any case asymptotic to the vertical line V = 0. Choosing any other set of equidistant horizontal rulings, nearer together or further apart, has the effect of making the 'curve rise slower or faster as the case may be. As will be seen later by inspection, this merely changes the position of the entropy diagram in the coordinate field but does not affect the final result. To prove that the ordinates of the curve M N represent the T7- values of log ^, Professor Eddy proceeds as follows : Let the volume AX = V , AX' = V v AX" = V 2 , AX n = V n . It can be shown by plane geometry that, with the construction used, _!= 2 = ___ = Z or ZV = V v ZVt = V 2 l . . . " . ! . -ZV n _=V n Hence 7217 . _ v 73 v . _ v 7"V - V ^ v o ~ K 2> z/K o~ K 3 ^ v o ~ y n Now let Z = e y in which e = Naperian base = 2.7 +, and y = distance between the equidistant horizontal rulings above as- sumed. Then e y J_i e zy _L? e ny J_ v ' ~ v ~ v V V K and taking the logarithm of both sides of the last term, we have n y = the ordinate at any given volume V n = \og e ^. It is ^0 evident that V , the unit of comparison, may be arbitrarily chosen. The next step is to obtain the graphical representation of the C T T expression - ^ log e ~. Considering the part log e =i by itself, Lp L> v 1 2 1 2 124 INTERNAL COMBUSTION ENGINES it is evident that the curve representing this may be constructed in the same manner as curve M N. But the coordinates are in this case the line of zero temperatures, A C, and any arbitrarily chosen vertical line, in this case, V = .74. The curve must be asymptotic to A C, but the choice of the other line of reference is unrestricted as it merely moves the curve bodily to the left or right. In this case equidistant vertical rulings having the same common distance as those used for curve M N have been em- ployed, hence curve P is a duplicate of M N. The ordinates of this curve P must next be multiplied by C the factor -^-'^-^ . Professor Eddy, in order to make the con- CpC v struction graphical throughout, assumes that the value of this factor is 2.45 in all cases. The limits of accuracy regarding this assumption have already been pointed out. It is probably not sufficiently accurate for most cases, and hence a separate compu- tation of this factor is necessary for every given case. This de- stroys a great deal of the value of the entire method, but enough is left to make the method much less laborious than the mathe- matical construction. In the case under discussion, for the burned gases, C v = .196, and C p = .268; hence: ~~J =2.72 The ordinates of the curve P are therefore multiplied by 2.72, giving the curve marked R S. By the aid of the curves M N and R S, the entropy diagram may now be constructed in the following manner: Take any temperature level as T U. This cuts the tempera- ture-volume diagram of the real cycle in the points c and d, and the curve R S in e. With a pair of dividers determine the ordi- / y \ nate g h of the curve M N, I log c * j, corresponding to the volume of point c, and since this ordinate is positive, add it to the ordi- / C T \ nate / e ( v log e .. l ) of the curve R S. This gives the point \C/> LV 1 2/ e' as one point of the entropy-temperature diagram. In the same manner, for the second point of intersection, d, of the tem- perature level T U, determine the ordinate g' h', corresponding to its volume, and add it to the ordinate fe of the curve R S, This THE TEMPERATURE-ENTROPY DIAGRAM 125 gives e" as a second point on the entropy diagram. By taking a sufficient number of temperature levels, the entire diagram may be closely outlined, as shown by the full line. The same thing has been done for the ideal temperature- volume diagram, giving the ideal entropy diagrain indicated in broken line. The last step in the construction is the determination of the entropy scale, if this is desired. Determine by planimeter the area under the combustion line of the ideal diagram down to the line T = 0. In the original diagram this was found to be 28.32 square inches. Since the heat applied the cycle was 29.06 oo C\R B. T. U., the thermal value of each square inch of area = ^^ = 28. 32 1.026 B. T. IT. Each inch of ordinate was equal to 614 degrees, hence the entropy scale per inch is -r~ = .00167 for the charge weight of .03477 pounds. This is equivalent to an entropy scale of .048 per inch for one pound of charge weight. The following table shows how closely the entropy diagrams obtained by the two methods outlined agree : Math. Method Graph. Method Max. Entropy of Ideal Cycle . ... .3166 .3158 Max. Entropy of Real Cycle ..... .2557 .2572 The agreement may be pronounced quite satisfactory. It is quite likely, however, that the lower part of the graphical entropy diagram will show discrepancies. These are in great part due to the fact that toward the lower end of the curve R S, the inter- sections with the horizontal temperature levels become less defi- nite, impairing the accuracy. Another source of error may lie in the fact that the compression line has been constructed from the curve R S, which was itself constructed from burned gas data, while evidently the data of the fresh charge should have been used. CHAPTER VI COMBUSTION i. The Perfect Gases. The perfect gases are those which follow the general law: Y= R C 1 ) where p = pressure expressed in pounds per square foot. v = volume in cubic feet. T = absolute temperature. R is the amount of work done by 1 pound of gas when heated 1 degree Fahrenheit, the pressure remaining constant at p pounds per square foot. R is thus a constant for any one gas, but differs for different gases. The data for the gases of most use in gas-engine practice are those of atomic and molecular weight, density and weight per cubic foot. The following table gives these figures for some of the more important gases: Gas Atomic Weight Molecular Formula Molecular Weight Density Air=l Weight per cu. ft. at 29.92" Hg and 32F Hydrogen 1 Ho 2 0692 . 00559 Oxygen. 16 Oo 32 1 106 08921 Nitrogen 14 N 2 28 .971 .07831 Carbon Monoxide . 14 CO 28 .967 .07807 Carbon Dioxide . . . 14.6 00, 44 1.529 .11267 Dry air ___ __ 29* 1.00 .08072 Water vapor 6 H 2 O 18 .623 .05020 v. Acetylene 6.5 C 2 H 2 26 .915 .07251 Methane ......... 3.2 CH 4 16 .554 .04464 Ethylene 4 7 CofL 28 .974 . 07809 " Benzol 5 75 CH 78 2 695 .21758 Alcohol 6.50 C,H fi O 46 1.601 . 12958 *Only apparent value. The weight of a cubic foot of any of the above gases under standard conditions may be found with sufficient accuracy by 126 COMBUSTION 127 dividing the molecular weight of the gas by the constant 359.* The weight of a cubic foot of CO 2 under standard conditions, for 44 instance, is - = 0.1225 pounds, as computed by slide rule. This is sufficiently accurate for all practical purposes. To find the weight of a cubic foot of gas under other than standard conditions, the formula m rp \6) 1 * 1 may be employed for a change of either pressure and tempera- ture separately or a simultaneous change of both. The weight of a cubic foot is inversely proportional to the volume. 2. Combining Weights and Volumes, Combustion, Heating Value, Air Required. COMBINING WEIGHTS AND VOLUMES. All elements, when they do combine, unite only in certain fixed proportions, although there may be several proportions for any given pair of elements. Thus carbon forms two combinations with oxygen, CO and CO 2 . The gases also combine in definite volume proportions. The resulting volume is -either the sum of the original volumes or is in a definite ratio less than this sum. It is necessary to remember merely that anything that can be said of molecules according to Avogadro's Law, applies with equal force to combining volumes. Thus take the combination of C and O to CO. Always re- membering to use the combining weights of the various elements we can write C 2 + O 2 = 2CO 1 vol. C + 1 vol. = 2 vols. CO. Similarly, 2 H 2 + O 2 = 2 H 2 O. 2 vols. H + 1 vol. O = 2 vols. H 2 O andagain 1 vol. C + 4 vols. H = 2 vols. CH 4 . The elements mostly concerned in combustion phenomena are carbon and hydrogen, together with the compounds carbon monoxide and the various hydro-carbons. The following table * Derived, in connection with Avogadro's Law, from the fact that the kilogramme- volume of perfect gases is equal to 22.33 cubic meters under standard conditions. 128 INTERNAL COMBUSTION ENGINES shows the combustion formula, also oxygen and air required, for the first three of these combustibles: # oo oo ^ 3 1 CO t *". ^T 1 | . S3 3 :'1 O^ ^O "^ ^^ ^t 1 T 1 .S ! * * *}j 2 I 1.1 oo oo b t^ 1> " ^ rh w o PH CO H S OOO _: 1 to 'S 1 I 6 : d M - ^ 8 c ^8 ^o C 'S 15 S W > o ^ c . (N . CNJ >l 8 ! O 1 || 1 || 1 H 'o || .2 CO? S" SiOCOCO COOi>OOOrH l-I>-OO 3l>iOOi lOOiCOiOtO * fa rH t- ;llgiSslss w (N a ft CO * 1? ^ CO Tt^ CO ^ ^^ t^w IN- CO 3 * 1 1 3 Oi O "^ CO Oi CO Oi } !> rt* CO rt< rH rH rH rH o fi ^ 8^ 00 - Oi co i^ co 7) Oi "* O * CO CO CO CO :d o 1 2- ~ ^11 to c to o 81 ^ CO CO CO TH CO CO CO CO rH E rH * | 1 s - 3 l> C ^0 ^ _] t-1 %t CM 5 tO CO CO TH a"E ^^0 8 ( n 3 < > ( 3 : ^ 12 5 OX 5 p^ 3 Hydrogen n_i TIT iiliii!1 liilillill 140 INTERNAL COMBUSTION ENGINES 2SS s i!. TH T-H 00 rH O5 *-l CO O 00 8 8 % CO CO CO OO 't' O 00 TP Oi Oi CO "^ Oi OO CO O O CD t^ O 00 r^ co ' t-- iO Oi CO T^ T^ t^ Oi O ^ ^^ 10 Tt . CO CO !> O ^ r-t Tt< CO i-H i-l (N S3 t^ ^^ CO 00 Ol CO ^D CO ^^ CO C^ GO -^ 00 00 O f> -^H |^ j^ oo o o o o o CO OO O CO O O co i> I-H I-H oo GO *O C^ ^^ 00 I-H (N O O Tt< CD CO CO GO CTj ^ ^p 1 > per pound C v ) C P C v " COMBUSTION 143 The values of C p and C v in the foregoing table are found as follows. Consider the mixture with ratio = 1.5: Burned Gas By Volume By Weight CO 2 H 2 O o .3257 .1997 0798 .0399 .0100 0071 N 1 6748 1311 2.2800cu. ft. . 1881 pounds For CO 2 C p = .0399 X .20 H 2 O N C p = .0100 X .48 = .0080 - .0048 C p = .0071 X .217 - -0015 C p = .1311 X .244 = .0320 For .1881 Ibs. 2C p = 1)463 .0463 C P= Since C p - C v = R T1881 = .2462 C v = .2462 - = .1787 and ^ = 1.379 778 Attention should be called to the fact that although consider- able contraction of volume occurs, in the case of this gas, during combustion, still the values of R, C p and C v are not greatly dif- ferent from the corresponding values before combustion. In some other gases, as illuminating gas for instance, the change is even less. So that in most ordinary cases it is sufficiently accurate to assume that these gas constants are the 'same before and after combustion. Only in cases where extreme accuracy is desired is this assumption not permissible. 3. Computation of the amount of air used in excess of theoretical requirements from the exhaust gas analysis. In actual practice the exhaust gases are analyzed for C0 2 , O, and N.' By the ordinary method of collecting these gases, the water vapor originally present is -thrown down and does not appear in the analysis. As mentioned before, if the fuel gas itself carries 144 INTERNAL COMBUSTION ENGINES no nitrogen, the excess coefficient for the air actually used may be computed from the formula given on page 137. To show an example of the method of computation when the fuel gas carries N, we will take the case of the Dowson gas above given, and assume that the exhaust gas analysis gives the following results: CO 2 - 14.36%, O - 5.31%, and N - 80.33% by volume. We proceed as follows: Products of Combustion for theoretical ratio per cubic foot of gas, are: CO 2 .3257 cu. ft. H 2 O .1997 " O .0000 " N .4898 " due to gas itself 1 XT , = 1-3746 cu. ft. of N, N .8848 " due to air used j . [35.63% is due to gas. of which { _ __ . , [ 64.37% is due to air. Total 1.9000 cu. ft. On the basis of the above exhaust gas analysis, we now have: Total N 80.33 Of this amount, N due to excess air will be 3.76 X 5.31 . . = 19.96 Leaves N due to the gas itself and to air actually burned =60.37 Of this remainder, as above shown, 35.63% is due to the fuel gas =21.49 Leaves N due to the air actually burned =38.88 Hence the excess coefficient 38.88 + 19.96 = 58.84 = ~38^8~ ~ 3^88 " and the real ratio of air to gas for the original fuel mixture was 1.5 X 1.12 = 1.68. 4. Calorific Intensity. --By calorific intensity is meant the temperature that can be realized theoretically when a unit weight of any fuel is completely burned under stated conditions of oxy- gen or air supply. If H represents the heating value of the fuel in B. T. TL, A, B, C, etc., the weights of the various resulting products of combustion, and C pA , C pB , C pc , etc., the specific heat at constant pressure of these products, the general state- ment for calorific intensity, supposing the pressure to remain constant, is COMBUSTION 145 TT Theoretical Temperature Rise = -j-^ - ^ - -=-= - F. -f CCc + Thus the calorific intensity of hydrogen with theoretical air would be, the products of combustion being water vapor and nitrogen, 52230 [9 X .48] + 126.8 X .244] That of C to CO 2 with theoretical air would similarly be 14500 [3.66 X. 20] + [8.91 X .244] Such high temperatures, however, are practically never realized, due probably to two causes. On the one hand it is claimed that the specific heat of gases is not constant at all temperatures, but that it rises with the temperatures; on the other, dissociation is supposed to set in before such temperatures are reached. These matters will be taken up somewhat more in detail in a later chapter. CHAPTER VII GAS-ENGINE FUELS; THE SOLID FUELS; GAS PRODUCERS THE general requirement for a gas-engine fuel is that it must mix readily with air to form a combustible gas or vapor. Further, it should burn with little or no residue. This latter requirement is not met by the solid fuels, as coal dust for instance, and while isolated attempts at using powdered coal directly have been made, they have so far not been successful, owing to the fact that the resulting ash soon seriously interferes with operation. The gas-engine fuels may be classed under three heads: 1. The solid fuels. 2. The liquid fuels. 3. The gas fuels. It is the rule that the working medium in all internal com- bustion engines is either a combustible gas or a combustible vapor, no matter what the fuel may have been from which it was derived. This implies gasification of the solid, and vaporization of the liquid, fuels. As already pointed out above, the solid fuels cannot be em- ployed in their natural state. From coal we derive by distillation illuminating gas, and from coal and sometimes other materials, as wood, refuse, etc., by gasification, the various classes of pro- ducer or power gas. Illuminating gas will be further considered under the head of gas fuels. i. The Conversion of the Solid Fuels to Gas: Producer Gases. - Gasification of solid fuel differs from distillation in the fact that the process is carried one step further, i.e., not only are the gases, if any, driven off from the fuel, but the carbon itself is gasified, leaving behind nothing but ash. The fundamental principle of all producer-gas processes is 146 GAS-ENGINE FUELS 147 therefore, first, dry distillation of the fuel, and, second, the con- version of the solid carbon into a combustible gas, which can only be carbon monoxide. If the producer gas is found to contain other gases than those mentioned, it can only be due to changes in the process, unavoidable or otherwise. Producer practice may be carried on in the following ways: 1. No steam or water introduced with the air, resulting in air gas. 2. Producer blown up with air for one period, then blown with steam alone. Product during first stage is air gas, during the second, water gas. 3. Producer furnished with air carrying a certain quantity of water vapor. Product is ordinary producer gas, Dowson gas, etc. The first of these is seldom employed on account of limita- tions pointed out below. The second and third introduce modi- fications into the simple air-gas process owing to the presence of water or steam. AIR GAS. Considering the case of the gasification of carbon alone, resulting in the production of the so-called air gas, assume the combustion of C to CO complete; we then have 1 Ib. of C + 1.33 Ib. O = 2.33 Ib. CO . If C had been burned completely to C0 2 , the calorific power would have been 14647 B. T. U.; burning only to CO, however, we obtain only 4429 B. T. U., so that the remainder or 10218 B. T. U. is carried out of the producer by the gas. made, and thus repre- sents its heat energy. 4429 B. T. U. appears as sensible heat in the gas, and if the gas is cooled before entering the engine cylinder, as it usually is for good reasons, the greatest possible efficiency which can be realized from the gasification of 1 pound of carbon 10218 in this way is " = 69.6 per cent. It will be shown below that this is by no means the maximum possible producer efficiency. WATER GAS. When water vapor is led through or over incandescent carbon the following reactions take place: I. C 2 + 4 H 2 O = 2 CO 2 + 4 H 2 . II. C 2 + 2 H 2 O = 2 CO + 2 H 2 . 148 INTERNAL COMBUSTION ENGINES I occurs at temperatures less than 1250 degrees Fahrenheit, while II alone occurs at temperatures exceeding 1800 degrees Fahrenheit; both may occur between these temperature limits, but the higher the temperature the greater the formation of CO. The maximum amount of CO is of course the end in view, and assuming that no CO 2 is formed, i.e., temperature at or above 1800 degrees Fahrenheit, we have the following quantitative statement : 1 Ib. C + 1.5 lb. H 2 O = 2.33 Ib. CO + .17 Ib. H 2 from which 1 pound of water gas must contain 2.33 2.33 f .17 and .17 = .932 lb. Carbon monoxide = .068 lb. Hydrogen 2.33 + .17 The gasification of 1 pound of carbon, therefore, in the pres- ence of water vapor results in products which, on complete com- bustion, develop the following amount of heat: 2.33 Ibs. CO X 4380 = 10205 B. T. U. .17 Ibs. H 2 X 62100 = 10557 B. T. U. Total, 20762 B. T. U. Heating value of C to CO 2 , 14647 B. T. U. Excess, 6115B.T. U. The excess of 6115 B. T. U. can only be due to heat rendered latent during the process. Water vapor on coming in contact with incandescent carbon dissociates into H 2 and O. The latter unites with C to form CO 2 , but as the temperature of the producer is at or over 1800 degrees Fahrenheit, CO 2 is dissociated to CO. The heat thus rendered latent accounts for the excess above shown. Now it is evident that this heat can only come from the stock of heat present in the producer when the blowing with steam first starts. Hence there must be a continual cooling of the producer contents during the period of water-gas making. This results finally in a serious production of CO 2 according to reaction I, when the steam must be shut off and the contents of the producer brought back to incandescence by blowing with air. GAS-ENGINE FUELS 149 The most unfavorable operation of the producer occurs when the reactions are according to equation I. Under this condition 1 Ib. C + 3 Ib. H 2 O = 3.66 CO 2 + .34 H 2 so that 1 pound of water gas then contains 366 3.66 + .34 and .34 3.66 + .34 = .915 Ib. CO, = .085 Ib. Hydrogen The production of water gas reckoned on the basis of coal or carbon is not at all efficient, since the heat in the poor gas made during the blowing-up period is very often wasted, and only in rare instances of utility. Other losses are, of course, those through incomplete combustion, radiation, etc.; but these are inherent in all producers to a greater or lesser extent. PRODUCER GAS. Midway between air gas and water gas we find the great class of power gases for the production of which the producer is blown continuously with a mixture of air and water vapor. To get a fair insight into the working of a power gas pro- ducer and of the efficiencies that may be realized, we will assume the following definitions and quantities.* 1. The heat supplied to a producer consists of the heat fur- nished to it in the fuel plus the heat contained in steam and air above a certain fixed temperature, say 32 degrees Fahrenheit. 2. The heat leaving the producer in the gas is made up of the latent heat of the gas plus the sensible heat. What the quantity of heat considered as the useful effect in efficiency should be depends upon circumstances. In furnace work, where it may be of advantage to employ the hot gas, the useful effect would be the sum of the latent and sensible heats of the gas. In gas-en- gine practice, on the contrary, the opposite is the case, and the useful effect would be the latent heat only. In the first case we speak of the hot-gas efficiency, in the second of the cold-gas efficiency. * Adapted from the discussion of E. Meyer, Zeitschrift des Yereins deutscher Ingenieure, 1895, p. 1523. 150 INTERNAL COMBUSTION ENGINES 3. Inside of the producer the following reactions take place, some endothermic, others exothermic. Of every pound of carbon the larger part burns to CO, the remainder to CO 2 . The heat generated from these two combus- tions is utilized in the following ways: Part of it dissociates the steam present, forming H, and CO or CO 2 , or both. How this action varies with the temperature of the producer has al- ready been pointed out. A second part of the heat serves to bring the fresh fuel up to the temperature of the producer, a third is lost by radiation from the exterior producer walls, and the remainder appears as sensible heat in the gas made. The formulae to be derived will be based upon one pound of C rather than upon one pound of coal, for the reason that coals vary greatly in composition, and it is in every case quite easy to change from this basis to that of coal if the qualities of the coal be known. Let 14500 B. T. U. = heat of combustion of 1 Ib. of C to 3.66 Ib. CO 2 . 4400 B. T. U. = heat of combustion of 1 Ib. C to 2.33 Ibs. CO. 10100 B. T. U. = heat of combustion of 2.33 Ib. CO to 3.66 Ib. CO 2 *() 1 C\C\ 6900 B. T. U. = - - = heat required to dissociate 1 Ib. of 9 water vapor under producer conditions. x = part of 1 Ib. of C burning to CO 2 . (1 x) = part of 1 Ib. of C burning to CO. y = pounds of steam introduced per Ib. of C. A = heat furnished in steam. B = heat furnished in air. C = heat required to bring fresh fuel to temperature of producer. R = heat lost by radiation. S = sensible heat of the gas. All of the above heat quantities are per pound of carbon gasi- fied, and above a temperature of say 32 degrees Fahrenheit. With this notation the general heat equation for the producer may be stated as follows, based on 1 Ib. of carbon: 4400 (1 - x) + 14500 x + A + B = 6900 y + S + R + C The heat that will be generated by the combustion of the volume of gas formed comes from CO and H. GAS-ENGINE FUELS 151 Heat generated in gas per pound of C gasified = [10100 (1 - x) + 6900 y.] B. T. U. 6900 y for the heat generated by H is obtained by considering that we must receive as much heat from the combustion of the H in the gas as was rendered latent during dissociation of the amount of H 2 O required to furnish it. Heat supplied per pound of C gasified = 14500 + A + B. Hence 10100 (1 - x) + 6900 y Cold-gas efficiency 145 QO + A + B 10100 (1 - x) + 6900 y + S Hot-gas efficacy 14500 + A + B In the latter case, no part of the sensible heat of the gas is abstracted before the gas is used. In practice there is always some loss of temperature between the producer and the place where the gas is used, hence S is never fully obtained. If in the above general heat equation the values of x, A, B, C, R and S are known, the value of y, i.e., the pounds of steam to be used per pound of carbon, may be computed. The composition of the resulting producer gas may be computed as follows, using the above notation: COMPOSITION BY WEIGHT PER POUND OF CARBON GASIFIED. CO = (1 - x) ~= 2.33 (1 - x) pounds C0 2 =~x = 3.66 x pounds H = y = -- pounds The amount of N is found as follows: i (\ Oxygen required for CO = (1 - x) -^ = 1.33 (1 re) pounds 32 Oxygen required for C0 2 = y~ x = 2 - 66 x pounds Oxygen produced by dissociation of H 2 O. 16 8 = 18 2/= 9 2/P UndS . . O required from air blast 152 INTERNAL COMBUSTION ENGINES = 1.33 (1 - x) + 2.66 x - | y pounds = 1.33 (1 + x) - y pounds y [8 ~1 100 1.33 (1 + x ) - y ^- pounds 9 J 26.5 Hence also N brought in by air blast is o ~| 76 'i 1.33(1 + x) -^y ^| pounds The total weight of gas produced by one pound of carbon therefore is = 2.33 (1 - x) + 3.66 x + | + 1~ 1.33 (1 + x) - ^y 1^| pounds L J = (6.67 + 5.67 x 2.77 y) pounds of producer gas. COMPOSITION BY VOLUME PER POUND OF CARBON GASIFIED. Volumes at 32 degrees and 14.7 Ib. pressure. Weight per cubic foot of the various gas under standard con- ditions. CO = .07807 Ib., CO 2 = .12267 Ib., H = .00559, N = .07831. Hence, from the above weight computations, we may write directly. o 0.0. r Volume of CO = '*** (1 - x) = 29.84 (1 - x) cu, ft. .07oU7 3.66 .12267 Volume of CO 2 = ^^x = 29.02 x cu. ft. Volume of H = _. L~^= 19.87 y cu. ft. 76.5 [ J3.5X. 07831 = (55.51 + 41.74 x - 37.10 y) cubic feet, From this Total volume of gas per pound of C gasified = [29.84(1 -x) + 29.02 x + 19.87 y + 55.51 + 41.74 x - 37.10 cubic feet. = (85.35 + 40.92 x - 17.23 y) cubic feet. GAS-ENGINE FUELS 153 2. Theoretical Yield of Producer. If it is supposed that no CO 2 is formed, that all the sensible heat produced in the generator is recovered in making steam and preheating air and fresh fuel, and that no radiation has taken place, we shall have x and R = 0, and S = A + B. Under these theoretical conditions, the general heat equation (p. 151) then becomes 4400 = 6900 y + C. The value of C is approximately .2 X 1800 = 360 B. T. U. Hence 4140 - 6900 y, 4140 from which y = 7^.^. = -600 pounds of steam per pound of C gasified. ^ The theoretical yield of gas per pound of carbon under these conditions will be 29.84 cubic feet of CO. 19.87 X .600 = 11.92 cubic feet of H and 55.51 - 37.10 X .600 = 33.25 cubic feet of N. Total volume of yield = 75.01 cubic feet per pound of C. Composition of gas, per cent by volume, 39.8 per cent CO, 15.9 per cent H, 44.3 per cent N. Taking the higher heating value of R at 346 B. T. U. per cubic foot, and the heating value of CO at 342 B. T. U. per cubic foot, the gas yield from 1 pound of C will develop (29.84 X 342) + (11.92 X 346) = 14249 B. T. U. The higher heating value per cubic foot of this theoretical gas is therefore 14940 4^=* = 187.2 B. T. U. per cu. ft. 7 o.Ul In actual practice, however, x and R cannot = 0, neither is S = A + B; i.e., not all of the heat appearing as sensible heat in the gas leaving the producer is ever recovered. To 154 INTERNAL COMBUSTION ENGINES make clear what happens under these circumstances, the follow- ing table is constructed. In this table it is assumed that the heat furnished in steam and air per pound of carbon equals 1000 B. T. U. = (A + B) in all cases, and that the sum of the heat losses due to radiation, R, and sensible heat of the gas, S, = 1640 B. T. U. per pound of carbon. The heat furnished the producer per pound of C will then in all cases be 14500 + 1000 = 15500 B. T. U., while the heat accounted for will be 15500 - (1640 + 360) = 13500 B. T. U., 1 S500 so that the generator efficiency on hot gas in all cases = looUU = 87.2 per cent. It is further assumed for illustration that only the amount of carbon burned to CO 2 per pound of carbon gasi- fied varies. GAS-ENGINE FUELS 155 o> .2 o> 3 -^ "2 6 & p *> t>. p CO *O O lO i-H rH i-H i-H i-H Composition, % by volume IO O O O i-l 00 CO CO r-1 oi w q I-H T^ t^. oo 1-1 I-H 1-1 co M *> o co i i i i Cubic feet of gas per pound of C 1 3 o i> co O5 rt< CO Tt* CO t^. Qi s a o O O5 O O CO O 00 C3 i-l CO l> 1> 00 00 00 - (N Tt< TfH 1-1 00 (N Oi 00 O ^j l>- id rti CO (N CO CO CO CO CO w o o >o o o OO- t^- lO lO TP O5 (N ^O 00 i-l i-H i-H i 1 C^ o o TJH 10 O O O 00 00 Oi Oi O5 Oi CO CO O t>> CO oi d 06 i-l Jl* CO Oi CO O? Oi Oi CO 00 CO t"^ 1*1 2 S o o o o o o 1-1 ca co TP 156 INTERNAL COMBUSTION ENGINES In commenting upon the above table, Meyer points out that as the percentage of CO 2 in the gas increases, the heating value of the gas decreases, but since at the same time the percentage of H and the volume of gas per pound of C increase, it is not always right to conclude from analysis alone that the efficiency of the generator is less with a fairly high than with a low percent- age of CO 2 in the gas. In general terms it can be stated, how- ever, that the lower the temperature of the generator, the greater the formation of CO 2 , and the greater also the loss of heat in sensible heat of the resulting gas. Referring to the gas engine itself, a high percentage of CO 2 in the gas means a low engine capacity, since this high percentage is usually also accompanied by an increased amount of the other indifferent gases. 3. Gas Producers in Practice. Turning now to actual gen- erator practice, we find the following main points of difference: The fuel is not pure carbon, but some impure form of it, as coal, coke, lignite, peat, or wood. This in itself merely results in a lower yield of gas per pound of fuel fired than that above com- puted, and this decrease is further emphasized by the fact that some of the unburned carbon in the fuel is always lost in the ash. A complex fuel being used containing gases which are distilled off during the first part of the process, the resulting producer gas will have a somewhat different composition than that above computed. The main difference is due to the addition of hydrocarbons, and this difference is therefore greater with bituminous coals than with any of the other fuels. The use of any of the above-mentioned fuels results also in other complications more or less difficult, depending upon the fuel used. Such are the formation of tar, dust carried by the gas, etc., all of which make a cleaning of the gas imperative before it can be used. The primary consideration in the operation of the gas pro- ducers is perhaps the kind of fuel used. The points to be con- sidered in this connection are: percentage, of water carried by fuel, amount and kind of ash, tar-forming ingredients of the fuel, size of fuel, and whether it cokes or not. A high percentage of water has a direct effect in lowering the temperature of the producer, besides lowering the heating value GAS-ENGINE FUELS 157 of the gas per cubic foot as made. A large amount of ash makes more frequent cleaning out necessary, or it is likely to result in a partial stoppage of the air supply. If the ash should be easily fusible the case is much more complicated, as this results in bad clinkering. The size of the fuel should be a happy medium. Small fuel, i.e., screenings, etc., clog up easily and in any case require a higher blast pressure. Large fuel, on the other hand, offers too little surface for gasification and is apt to let much water and CO 2 escape unreduced. A coking coal nearly always gives trouble from this cause, and it necessitates constant breaking up of the charge. The formation of tar, which results especially when bitumi- nous coals are gasified, makes a cleaning of the gas for engine pur- poses indispensable. Tar results when some of the hydrocarbon gases are condensed through cooling in the gas mains and pipes. If these gases reach the cyclinder their combustion is likely to result in a strong deposit of soot. In either case the operation of the engine will soon be seriously interfered with. Tar can be almost entirely removed from the gas by washing it, but this process requires constructions fully as costly as the producer itself and hence other methods have been employed. The tar-forming gases are always those which are formed from the dry distillation of the coal, hence most trouble is en- countered with bituminous coal, less with lignite and still less with anthracite. For this reason anthracite and coke producers have been most successful, although producers using brown coals and lignites are in operation, as are also those using bitu- minous coal, but with less success. This does not apply to steel works where bituminous coal is used extensively for gasification. But there the gas is used mostly hot and less trouble from tar is experienced. The tar gasee can be "fixed," i.e., changed to permanent gases when the producer gas containing them is led through an incandescent bed of fuel before entering the gas mains. In this case the tarry hydrocarbons are changed either to H 2 O and CO 2 or split up into CO and H. In some producers only the gases resulting from the dry distillation are handled in this way. In either case the tarry hydrocarbons are fixed, and no elaborate cleaning apparatus for the gas is required. The necessity for 158 INTERNAL COMBUSTION ENGINES treating bituminous producer gas in this way has resulted in various constructions of producer, a few of which are given below. Gas producer installations may be divided into three classes : a. Pressure Producers. - In these air and steam are furnished to the fuel bed by a blower or fan. The ash pit of the producer must be enclosed, making the removal of ash com- plicated or the action of the producer intermittent, unless the water-bottom type is used. Steam for blowing is usually fur- nished by a separate boiler. Since the rate of production of gas is usually not regulated according to the demand for gas directly, a gas holder is usually necessary for this type. b. Suction Producers. In this class the air and steam are drawn through the producer by the suction of the engine cylinder. The production of gas is thus directly regulated by the demand. The ash pit remains open, and steam enough can usually be generated by the sensible heat of the gas. Suction gas producers have nearly replaced pressure producers for gas-engine purposes. Some of the obvious advantages, as open ash pit, absence of separate boiler and of gas holder, have been pointed out above. The dangers at first supposed to be inherent in this system have failed to materialize. Leaks in a pressure system may lead to a poisoning of the atmospheric air by the color- and odorless CO, positively dangerous to attendants. Leaks in a suction system only result in an in-leakage of air. That this can never happen to such an extent as to form an ex- plosive mixture, except through a combination of extraordinary circumstances, is at once evident when we consider that the ratio of air to producer gas for such a mixture would have to be at least 1 to 1. In spite of such advantages the suction system is by no means perfect. The regulation of the water supply to control the amount of H in the gas, the vaporizer for the water, and the cleaning apparatus are still points which admit of improve- ment even in the most recent form. c. Combination Producers. The air and steam mixture is drawn through the producer by a fan and the resulting gas forced by the same fan to the engine. The producer in this system is thus of the suction type. GAS-ENGINE FUELS 159 PRESSURE PRODUCERS. Taylor. Fig. 7-1 shows the Taylor producer made by R. D. Wood & Co. of Philadelphia. In their publications on the producer this company lays down the following requirements for a successful pressure producer. Most of these, however, apply to producers in general. 1. A continuous and automatic feed; the former for regularity and uniformity of gas production with improved quality, the latter for eliminating negligence of attendants. 2. A deep fuel bed carried on a deep bed of ashes; the first to make good gas, and the second to prevent waste of fuel. 3. Blast carried by. conduit through the ashes to the incandescent fuel. 4. Visibility of the ashes, and accessibility FIG. 7-1. Taylor of the apertures for their removal, arranged Producer. so that operator can see what he is doing. 5. Level, grateless support for the burden, insuring uniform depth of fuel at all points, and consequent uniformity in the pro- duction of gas. These points are well covered in the design of the producer. The fuel is admitted through a distributing hopper which keeps the layer of fuel level over the cross-section. The bed of ashes is kept at about 6 inches over the top of the air pipe, thus protect- ing it from direct heat. The entire charge in the producer is supported by a plate whose diameter is somewhat greater than that at the bosh. As necessity requires, this plate can be revolved and the ashes are scraped, or they fall off, into the closed ash pit, which is under blast pressure. The grinding action ensuing when the plate is revolved settles the contents of the producer and thus closes up any free air channels that may have been formed. Once a day the pit must be opened for the removal of the ash. Blast is supplied generally by a steam jet. Morgan. Somewhat similar in design is the Morgan pro- ducer, Fig. 7-2. The main point of difference is in the removal of the ash. The fuel here is also admitted through a -continuous automatic feeding device. The blast is controlled by a steam injector so designed as to maintain a proper proportion between 160 INTERNAL COMBUSTION ENGINES air and steam. The make of gas can be completely controlled by the adjustment of a j-inch steam valve. This producer is of the water- bottom type, i.e., the ashes fall into a water seal at the bottom, and may there be removed without stopping the operation of the producer. This is not easily done when a grate or similar device is used in a pressure producer. For certain fuels, especially those highly bituminous or those where ash is apt to clinker, the water-bottom producers possess some advantage over the others. About three feet above the water level in the ash pan, and some inches above the top of the blast dis- tributing pipe, a number of sight holes are arranged around the circumference of the producer. Through these the zone of combustion may be watched. The top of the producer is covered with a shallow water pan. Poke holes through the top with water-sealed covers are also provided. FIG. 7-2. Morgan Producer. H i FIG. 7-3. Wile Producer Installation. Wile. A complete Wile pressure gas plant is shown in Fig. 7-3.* Steam under about 40 pounds pressure is generated in the boiler, A, and enters the generator through the injector, 7, mixed with air. The gas made passes through the seal box, D, and the * J. I. Wile in Power. GAS-ENGINE FUELS 161 scrubber to the gas holder. This is the ordinary arrangement, but it is open to objection when the load is extremely variable. In such a case it is usual to arrange the gas holder so that it shuts off steam at / when the holder is full. The contents of the pro- ducer then cool, and a further cooling results when the steam is next turned on. The temperature ranges in the producer are therefore apt to be high under such conditions. To meet this difficulty the design can be changed by placing the steam injector at B, above the seal box D. The gas holder is connected with the seal box by a return pipe E. When the gas holder is up, catch H in the gas holder opens the return valve^and the injector B merely draws on the gas holder, placing the generator temporarily out of commission. When the gas holder falls, the return valve closes, the injector draws on the generator, and gas is again made. In this design the arrangement at / is then merely a saturator, and the plant is really a combination plant, the generator being under suction. FIG. 7-4. Koerting Pressure Producer. Koerting, Hannover* Fig. 7-4 shows a Koerting pressure plant. The gases made during the firing-up period escape through the pipe a. Should the natural draft not be strong enough, it may be increased by means of a small steam blower in pipe a. After the flame at the try cock shows dark red, and not blue, the valve b is closed, and the gas made sent through the pre-heater, scrub- * Giildner, Entwerfen und Berechnen der Verbrennungsmotoren, p. 384. 162 INTERNAL COMBUSTION ENGINES ber, and sawdust purifier to the gas holder. The make of gas is controlled by the gas holder through the chain C, which acts upon the blast through a throttle valve at d. According to pub- lished figures, the average analysis of the gas made is, by volume, H, 18 per cent; CO, 26 per cent; C M H 2w , 2 per cent. CO 2 , 7 per cent; N, 47 per cent; Efficiency, 80-82 per cent. '///////////////////////^ FIG. 7-5. Deutz Pressure Producer. Deutz* The steam used for the blower a, in the Deutz plant, Fig. 7-5, is superheated in a coiled pipe above the fuel in the boiler. FIG. 7-6. Poetter Producer. * Giildner, Entwerfen und Berechnen der Verbrennungsmotoren, p. 384. GAS-ENGINE FUELS 163 The air is pre-heated by the heat of gases in the heater C. The opening b is used to try the gas during firing up. Arrangements of scrubber and purifier are similar to those already described. Poetter. The Poetter producer,* Fig. 7-6, is especially de- signed for bituminous coal. The fuel charged is distilled while still in the hopper, the gases formed are drawn off by special steam blower and are led through the pipe Cd under the grate. Air for blast is provided through e /. The gas made escapes at a, and is first used to raise the steam required in a boiler. It is then led through cooler, scrubber, coke- and sawdust purifier to the gas holder. Schottler, in describing a plant of Poetter pro- ducers at Johannesburg, surmises that their operation might cause trouble, although the kind of coal used at Johannesburg is not stated. FIG. 7-7. Mond Producer Plant. Mond. When a producer is blown with a large excess of steam, a great deal of it will go through undecomposed. At the same time, however, the quality of the gas made undergoes some radical changes. The quantity of H in the gas will be high, sometimes up to 25 per cent, while on account of the low pro- ducer temperature, a great deal more of the C is burned to CO 2 than is ordinarily the case. A further, and important, change is that a great deal of the N is changed to ammonia, a reaction which does not take place in producers run under the ordinarily higher temperatures. To recover this ammonia is an important consideration. This is the process of Mond. A Mond gas plant consists essentially of two parts, the pro- ducer and the condensing and recovery plant, Fig. 7-7. * Rev. Mec., 1904, p. 484. 164 INTERNAL COMBUSTION ENGINES The details of the producer A do not differ much from those already described. It is of the water-bottom type. The gas made passes through the regenerators B. These consist of double wrought-iron tubes united alternately top and bottom. The hot gases pass through the inner tubes, heating the mixture of steam and air, used for blowing the producer, which flows through the outer tubes in the opposite direction. The gas next passes through the washer C. This is a large chamber partly filled with water. By means of dashers, driven by D, the chamber is kept filled with water spray. The gas is cooled considerably in pass- ing through C, the water vapor in the gas being condensed to a great extent. Up to this point all of the apparatus is necessary even if no ammonia recovery is attempted. For a gasification capacity of less than 30 tons of coal in twenty-four hours it is not usual to install a recovery plant on account of the high cost of installation. Beyond this capacity the installation is justified. A recovery plant consists of the acid tower E, the gas cooling tower F, and the air heating and saturating tower G (see Fig. 7-7). The gas after leaving the condenser C enters the tower E at the bottom and flows upward through firebrick checker work. In doing so it meets a descending rain of sulfuric acid liquor containing about 4 per cent of free acid, a^d by this the free ammonia in the ascending gas is fixed, being changed to sulfate. The acid liquor is circulated by a special pump, and kept at the proper strength by drawing off the sulfate liquor and adding a corresponding amount of fresh acid solution from time to time. The gas then enters the cooling tower F, at the bottom, and in its ascent is cooled by descending cold water. The gas gives up its burden of steam, which in turn heats the descending cold water. The gas is then conveyed to the gas mains. The hot water leaving the gas-cooling tower is pumped to the top of the air-saturating tower. Here it flows downward through checker work, and in its descent saturates and heats the air, which is driven upward through the tower by blowers. Leaving this tower, the air-water vapor mixture then goes to the regenerator B, where it is further pre-heated before entering the producer. GAS-ENGINE FUELS 165 Sexton* states that the amount of steam used is about 2i tons per ton of fuel, and estimates that about two tons of this go through undecomposed. The method has the advantage that slack coal may be gasified with success. The recovery of ammonia amounts to about 1 ton for 23 tons of coal gasified, or adding the fuel required for making steam, about 1 ton for 28.5 tons of fuel. The average analysis of Mond gas is, by volume, 11 per cent CO, 17.1 per cent C0 2 , 1.8 per cent CH 4 , .4 per cent C M H 2w , 27 per cent H, and 42.5 per cent N. This gas is free from tar and excess of moisture, and burns with a non-luminous flame. FIG. 7-8. Deutz Suction Producer. SUCTION PRODUCERS. Deutz.^ In the Deutz suction pro- ducer, Fig. 7-8, the vaporizer is arranged around the top of the generator. During the suction stroke of the engine, fresh air enters the vaporizer, is here mixed with water vapor, and then, flowing through the connecting pipe at the left side of the genera- tor, reaches the under side of the grate. The gas made leaves the generator through the pipe at the right and enters the wet scrub- ber at the bottom. From here it passes to the engine through a * Sexton, Producer Gas, p. 90. f Giildner, p. 386. 166 INTERNAL COMBUSTION ENGINES regulator. Just before entering the engine the gas passes a vessel containing a wire brush for a final cleaning of tar. The larger Deutz suction installations show a somewhat different construction. FIG. 7-9. Koerting Suction Producer. Koertmg* Fig. 7-9 shows a Koerting suction plant designed for from 30-35 horse-power. In this case the vaporizer is inde- pendent of the generator. The latter is heated by the hot gases on their way to scrubber and purifier. During the heating-up period, the air is furnished by the blower shown. The smoke escapes through a pipe which branches off beyond the vaporizer, so that this will be heated during the blowing-up period. FIG. 7-10. Koerting Installation. A complete view of a Koerting plant is shown in Fig. 7-10. * Giildner, p. 386, GAS-ENGINE FUELS 167 American Crossley. The American Crossley suction gas plants in this country are built by the Power & Mining Machinery Com- pany. Fig. 7-11 shows a complete plant. The producer is of con- ventional design. The fuel bed rests on a shaking grate. The fuel is admitted as shown. The waste heat boiler or saturator surrounds the feed tube. The water level in this saturator is automatically regulated. During the firing-up period the pro- ducer is blown by means of the fan shown, the gases escaping by the purge pipe. A new idea seems to be the saturating of only part of the air supply. The makers point out the difficulty of FIG. 7-11. American Crossley Producer. making uniform gas at all, especially at low, loads, owing to the difficulty of controlling the amount of steam admitted. To remedy this, part of the air only passes through the saturator, while the rest is directly admitted to the ash pit. Valves on these two air inlets can be set so as to give satisfactory gas over wide ranges of load. Another arrangement in which this producer plant differs from most others is that of the cleaning apparatus. This consists of a wet scrubber, an hydraulic box, and a combination wet and dry scrubber. The gas enters the first and passes downward, passing upward in the combination scrubber. The operation is shown plainly in the cut. 168 INTERNAL COMBUSTION ENGINES Fairbanks-Morse. Fig. 7-12 shows a Fairbanks-Morse installa- tion for 21 horse-power, the producer itself being shown in Fig. 7-13, in greater detail. The general design of the producer is very FIG. 7-12. Fairbanks-Morse Installation. similar to that of the American-Crossley, except that all of the air passes through the vaporizer. Apparently only a wet scrub- ber is provided, it being evidently intended to use anthracite only. A gas. tank of considerable volume is interposed between scrubber and engine. The suction producers so far described have been for anthra- cite, coke, or charcoal. The desire to utilize soft coal, lignites, etc., has led to special designs, of which the following are a few examples. In some of these all of the gas made from the fuel is passed through other highly incandescent but gas-free fuel to fix the tar-forming gases, in others only the gases of distillation are so treated. GAS-ENGINE FUELS 169 FIG. 7-13. Fairbanks' Producer. Riche. * Riche's producer, Fig. 7-14, is especially designed for wood, but it may also be used for hard coal and coke. The FIG. 7-14. Riche's Producer. * Genie Civil, 1901-2, p. 398. 170 INTERNAL COMBUSTION ENGINES column at the right serves as a magazine for the fuel bed at b. The gases formed are sucked up through C, which is kept filled with incandescent charcoal. The fixed gas passes through pipes d and e, being washed by means of sprayed water in its passage. Lencauchez. * Instead of using two separate stacks, the two fuels may be put together in one furnace, as is done by Lencauchez, Fig. 7-15. Long-flaming bituminous coal is charged through a, while through 6 coke from coke ovens or gas works is charged. Water is evaporated in the ash pit. The openings c and d show only black smoke, while e shows a colorless gas. The gases from the soft coal are fixed in passing through the incandescent coke layer. Schottler mentions that this gas is not used as power FIG. 7-15. Lencauchez Producer. gas but for heating purposes. The proportions of the charge are about 20-25 per cent coke to 80-75 per cent soft coal. Lencauchez. f Another way of reaching the same end, that is, fixing the tarry gases, is by the use of so-called double genera- tors. In these generators the charge burns both top and bottom, the gases being drawn off at about the middle of the producer. Fig. 7-16 shows a generator of this type designed by Lencauchez. It consists essentially of three conical parts. The upper one serves as a magazine. In the middle one the fire burns on top, in the bottom one at the bottom. Air enters at D and B. Water is introduced at A and evaporated in the ash pit. The air and gas currents are shown by the arrows. The gases of distillation formed in the magazine are drawn downward through the * Zeitschrift des Vereins deutscher Ingenieure, 1905, p. 1903. f Zeitschrift des Vereins deutscher Ingenieure, 1905, p. 1905 GAS-ENGINE FUELS 171 FIG. 7-16. Lencauchez Double Zone Generator. incandescent fuel in the middle section and are so fixed. Any unburned coked fuel then passes on toward the grate and is gasified on reaching the incandescent zone in the lower section. FIG. 7-17. Koerting Producer for Peat. 172 INTERNAL COMBUSTION ENGINES Koerting. * Koerting's producer for peat, Fig. 7-17, is built along lines similar to the above. The producer consists of two parts, of which the upper and larger one is fitted with step grates, a a, the lower one with an ordinary grate b. When in operation, the charge of peat in the upper parts burns only near the grates a a, while the inner core of peat is only coked. The gases so formed pass into the perforated pipe c and are led through d under the grate b. The coked peat formed in the upper part passes downward arid is gasified over the grate b. The gases from d pass through the incandescent layer above 6, are fixed, and together with the gases formed over b f ass out at e. One would naturally assume at first sight that the gases of distillation would pass directly downward and out at e, instead of entering c and passing through d. The direction of the movement is regulated by creating a slight vacuum under the grate b, and by contract- ing the producer section at /. This contraction produces an extra resistance to the passage of gases downward, and they conse- quently take the more convenient way through c and d. Schottler gives the analysis of a peat gas so made as 14 per cent C0 2 , 4 per cent CH 4 , 15 per cent CO, 10 per cent H, and 57 per cent N, producer efficiency being about 75 per cent. FIG. 7-18. Crossley Producer. * Zeitschrift des Vereins deutscher Ingenieure, 1905, p. 1909. GAS-ENGINE FUELS 173 COMBINATION SYSTEMS. Crossley. The Crossley producer, Fig. 7-18, is suitable for bituminous coal, or other fuel high in volatile matter. The fuel is first charged into the retorts A, and is there distilled by the sensible heat of the gases formed below. When charging, the valve B is drawn tightly against its seat, and the fuel is projected downward by turning the spiral D by means of the capstan. When the distillation is complete the casting B is lowered and the coked fuel is thrown into the main part of the producer by turning D. At the same time it is broken up by the spirals. The gases of distillation are drawn out of the retorts through the pipe F, and pass under the grate T 7 , at M. The suction is produced by a fan blower, which at the same time draws air and steam to the main bed of fuel, and discharges the resulting producer gas into a holder at suitable pressure. FIG. 7-19. Deutz Double Zone Producer. Deutz. The Deutz plant for bituminous coal is shown in Fig. 7-19. The producer is of the double-generator type. The air enters the vaporizer at a, and reaches the ash pit through pipe b. The fire burns top and bottom. Green coal is charged at the top. It is distilled, the gases formed pass downward through incan- descent fuel and are so fixed, passing out at c. The greater part of the fuel coked near the top passes downward and is finally gasified in the lower part of the producer, the gases also passing out at c. Suction is maintained by a fan connected beyond the coke scrubber. Owing to the fixing of the tarry gases it is found 174 INTERNAL COMBUSTION ENGINES that a coke scrubber is all that is required. The fan, which is almost a necessary part of such a plant, owing to the increased friction entailed by the design of the producer, delivers the gas to a holder. The holder regulates the make of gas by controlling the fan. During the firing-up period, a purge pipe is lowered over the top of the generator, which has no special charging bell. The same fan is then used as a blower to furnish air to the pro- ducer, by-passing the scrubber. FIG. 7-20. Loomis-Pettibone Producer. Loomis-Pettibone. * The Loomis-Pettibone gas apparatus, Fig. 7-20, may be used for the making of producer gas alone, or for alternate manufacture of water gas and lean power gas. Each plant consists of two producers, a boiler, a combination wet and dry scrubber, and an exhauster which delivers the gas made to the proper holder or the purge pipe, depending upon the positions of the valves x, y and z. The producers are charged through the top. Through the same doors the air also enters. Thus the gases pass downward through incandescent fuel, so that any kind of fuel may be employed. Leaving the producers at the bottom, the gases pass through the boiler, where they are cooled, and then through the scrubber and purifier. The make of gas is con- trolled by the speed of the exhauster, the suction pressure being ordinarily from 12 to 24 inches, the delivery pressure about 6 inches of water. Suppose it is desired to make water gas. The feed doors E * The Power Plant of the Montezuma Copper Co., John Langton. GAS-ENGINE FUELS 175 and F, and the valve B, are closed. Steam is then admitted under the grate of producer No. 2, the right-hand producer in Fig. 7-20. The water gas formed in No. 2 is passed through the top connection to No. 1, passes down No. 1, out at A, and thus up the boiler and through the scrubber to gas holder. The next time water gas is run, the directions are reversed, or this may be done when one half of a water-gas run is completed. To make water gas, the aim is to blow the producers up with air as hot as possible, and then pass steam only through them until a great deal of it goes through undecomposed. Then air alone is again used to get the fires hot. This procedure means a wide range of temperature in the producer, and while this does not affect the quality of the water gas seriously, it does affect the producer gas made during the blowing-up periods, as a great deal of CO 2 is formed when the producers are comparatively cool. In a large plant where several producers are at one time delivering to the same gas holder, this irregularity is not much felt at the mixing valve of the engine where the two gases are mixed in proper proportion. For a single set plant this feature of varying pro- ducer gas make is more serious, and for this case the producers may be operated in a different manner. To this end the steam connections, M, are used so that steam may be introduced along with the air, thus making water gas and lean gas concurrently, maintaining the producer tempera- ture fairly constant for a long time; in other words, true producer gas is made. To get around the too fine adjustment of air and steam for constant temperature, somewhat less top steam than could be used is employed, thus allowing the temperature to slowly rise. A short water-gas run brings it down again to the proper point. The water gas made is slowly fed from its holder to the main gas holder through a controlled throttle valve so that its rate of feeding approximately equals its rate of production. 4. Some Producer Details. - - Mathot, in The Engineering Magazine, May, 1905, gives the following data for suction por- ducers : To get the greatest production of H and the most effective reduction of CO 2 , make the cross-section of the base of the pro- ducer from 0.6 to 0.9 of the piston area of the engine. This is for a single cylinder 4-cycle engine at from 600-800 feet piston 176 INTERNAL COMBUSTION ENGINES speed. The depth of fuel bed should be from 3 to 5 times the diam- eter at the base, for -inch to f-inch lump coal. Amount of water dissociated is from 0.8 to 1.2 times the weight of coal con- sumed. More than twenty analyses give the following average figures: CO, 24 per cent; CO 2 , 5 per cent; H, 17 per cent; N, 54 per cent. Calorific power from 135 to 150 B. T. U. per cubic foot. Average coal used shows 89 per cent C, 2 per cent H, 4 per cent O + N, and 5 per cent ash. Best size coal is from J-inch to 1^-inch lump with 8 to 10 per cent of ash, and not more than 8 to 10 per cent volatile matter. Water in the ash pan is satisfactory for steam production if the air is pre-heated. But preference is now given to internally heated vaporizers. Tubular vaporizers produce sufficient steam after 10 to 13 minutes. With well-designed apparatus, any suc- tion plant should be in operation 25 minutes after lighting up. The volume of the scrubber should be from 6 to 8 times that of the producer. Its height should be from 3 to 4 times its diameter. The filling is coke in pieces 3 to 4 inches in diameter, the coarser pieces near the bottom, the smaller near the top. Amount of water used for washing is from 3 to 5 gallons per horse-power per hour for anthracite gas. According to the DeLavergne Company, the impurities to be removed from one ton of fuel are, for anthracite, from 1 to 2 pounds of ammonia, traces of sulfur, and from 5 to 10 pounds of tar; for bituminous coal, from 4 to 5 pounds of ammonia, sulfur from traces to 5 per cent, and from 10 to 12 gallons of tar. The Morgan Construction Company point out the effect of sulfur on the formation of clinker. A coal with a large per cent of ash may work satisfactorily in a producer provided the sulfur does not exceed 1 per cent. Above 3 per cent the effect of sulfur in forming clinker is badly felt unless special facilities are at hand to break up such formations. The same company makes the following statements regard- ing pressure-producer capacities. The best rate of combustion seems to be about 10 pounds of bituminous coal per square foot cross-section per hour, the coal carrying 10 per cent of ash and 1 per cent of sulfur. With coals of lower ash content the rate I GAS-ENGINE FUELS 177 may be increased to 12 pounds, and in some special cases to 15 pounds. For poorer coals, however, the rate may sink to 6 or 8 pounds. Regarding the size of gas flues and pipes, the statement is made that each square foot of gas flue will take care of a gasifica- tion of 200 pounds of coal per hour, and will serve a gas-making area of 16 square feet in the producer. t CHAPTER VIII THE GAS-ENGINE FUELS'. LIQUID FUELS, CARBURETER AND VAPORIZERS i. The Crude Oils and their Distillates. The crude petroleum oils are practically mechanical mixtures of various hydrocarbon families. It is therefore to be expected that crude oils from various fields show compositions considerably different. From these crude oils we obtain by distillation all of the mineral oils now so widely used for a great variety of purposes. Of these oils the various gasolines, kerosenes, and the so-called "distillates, " besides crude oil itself, are used for gas-engine fuels. CRUDE OIL. The following table* gives a few analyses of crude oils from different parts of the world. These oils are usually of a dark green color, the specific gravity is between 0.80 to 0.90 at 32 degrees, and the flash point between 76 and 93 degrees Fahrenheit. Sp. Gravity at 32 C H Oand Impuri- ties Lower Heating Value per pound Heavy crude W. Virginia .873 .841 .886 .826 .886 83.5 84.3 84.9 82.0 84.9 13.3 14.1 13.7 14.8 13.7 3.2 1.6 1.4 3.2 1.4 18324 18400 19210 17930 19210 19463 20410 21600 18781 18010 18416 19440 19496 18330 Light crude W. Virginia Heavy crude Pa. Light crude Pa Heavy crude Ohio Salt Creek, Wyoming Bothwell, Canada .857 .861 .872 .885 .884 .956 84.3 80.2 86.2 82.2 85.3 86.3 86.6 85.0 13.4 17.1 13.3 12.1 12.6 13.6 12.3 11.2 2.3 2.7 .5 5.7 2.1 .1 .1.1 2.8 Lima, Ohio Schwabwjller, Lower Rhine . . . Gallicia East Gallicia, East Light Crude Baku Heavy Crude Baku Keudong Java * Poole, the Calorific Power of Fuels. 178 GAS-ENGINE FUELS 179 GASOLINE. After the distillation of the very light products from crude oil, we obtain a series of gasolines of varying flash points and* gravity. The first of these is 86 degree gasoline, as measured on the Beaume scale. This forms a mixture with air so readily even at ordinary temperature that it is somewhat dangerous and not often used. The next gasoline, the original liquid gas-engine fuel, is 74-degree gasoline. Owing to the much greater consumption of gasoline, due to the introduction of the automobile mainly, the gravity of gasoline and also the flash point have slowly gone up, until to-day 69-degree gasoline is quite common. , Gasoline vaporizes readily at ordinary room temperatures, and it is therefore necessary to keep it covered, not only to pre- vent loss but also accidents due to explosions. Insurance com- "panies usually specify that any quantity of it must be kept in an underground tank outside of the building, and that 13 undoubtedly the best way. Data on the heating value of gasoline is not at all plentiful. A sample the writer had the opportunity of testing recently gave the following figures: f Specific gravity, Beaume 69.5 at 60 degrees Fahrenheit. Composition, 84.76 per cent C, 15.24 per cent H. Lower heating value as computed from analysis, 20411 B. T. U. Heating value as determined by Junker's calorimeter, higher value 19606, lower value 18482 B. T. U. tf f It should be noted that this is one of the instances where the heating value as computed is much higher than the actual value, nearly 2000 B. T. U. = 10 f)er cent in this case. It shows the necessity of calorimetric determination for accurate work. J^EROSENE. Kerosene, the next heavier distillate beyond gasolines, is not as extensively used as gasoline in this country for gas engines. It will not form an explosive mixture with air at ordinary temperatures, and therefore requires more elaborate apparatus for the formation of such a mixture^ The following data is given by Mr. S. A. Moss for ordinary American kerosene: Sp. Gr. at 60 Flash point C H Lower Heating Value open by weight per Ib. B. T. U. .80 100 .85 .15 18520 180 INTERNAL COMBUSTION ENGINES Specific gravity at 24C Average composition, C H O Further reliable data on kerosene is given in a lecture by Diesel in 1897, in connection with Schottler's test of the Diesel engine. (75.2 F.) .789. 85.13 per cent. 14.21 per cent. .66 per cent. Lower heating value computed from analysis, 19874 B. T. U. Lower heating value by Junker's calorimeter, average of five tests, 18242 B. T. U. DISTILLATES. A distillation product resembling kerosene in its general properties is sometimes used as fuel. These so- called distillates are not as well refined as kerosene, but are handled the same when used for engine work. The following table, due to Hofer and transposed from Giild- ner, shows a good method of presenting the various distillation products from crude oil: (a) Volatile Oils. SP. GR., .65-. 75, FLASH POINT BELOW 70 F., BOILING-POINT BELOW 300 Sp. Gr. Boiling-Point Petroleum Ether . . . 65-. 66 95- 122 F. Benzine 66- 68 112- 158 Various Gasolenes .67-. 74 149-300 (b) Illuminating Oils. SP. GR,, .78-.S6, FLASH POINT 70-158, BOILING-POINT ABOVE 300 F. Sp. Gr. American Kerosene . . . 78 -.81 Russian Kerosene . . .82-825 Standard White 808- 812 Prime White 80 -.806 Astralin . .85 -.86 GAS-ENGINE FUELS 181 (c) Heavy Oils. SP. GR., .8G-.96, FLASH POINT 374-482 F. Sp. Gr. Solar Oil 86-.8S Lubricating and Cylinder Oils. .88-.90, etc., can be used as lubricating oils only. 2. Alcohol. The use of alcohol as a gas-engine fuel in this country is as yet of no great importance, although the recent action of Congress in removing internal revenue under certain restrictions will do a great deal toward helping alcohol to the place it deserves as a fuel. In some European countries, Germany for instance, the price of alcohol is not much greater than that of gasoline, and it may therefore compete with gasoline with some success, especially when some of its advantages are considered. It is much safer than gasoline as regards fire risks, and since it always contains some water, a higher degree of compression may be employed in the engine, guaranteeing better thermal efficiency. These advantages compensate largely the greater spe- cific heat cost of alcohol. Ethyl alcohol, * whose chemical formula is C 2 H 6 O, may be made in various ways, but the commercial alcohol of to-day is the result of fermentation, generally of grape sugar, in the final stage. The raw materials are various. Thus, according to Sand f they may be divided into three classes: 1. Those containing starch potatoes, with 15 to 24 per cent starch, rye, with 50 to 56 per cent starch, corn, with 60 per cent starch, 2. Those containing sugar sugar beet, with 8 to 18 per cent sugar, sugar cane, with 12 to 16 per cent sugar. 3. Those containing alcohol wine with 9 to 16 per cent alcohol. * What follows is a reprint from an article by the writer on the use of alcohol as a fuel for gas engines in Marine Engineering, June, 1906. f Sand, Zeitschrift des Vereines deutscher Ingenieure, 1894, p. 933. 182 INTERNAL COMBUSTION ENGINES The method of manufacture, of course, varies with the raw material, but need not be described in detail here. Theo- retically, 100 pounds of grape sugar should yield 51 pounds of pure alcohol; in reality the yield is from to $ less than this amount. A second method of producing alcohol, notably mentioned by Witz in his "Moteurs a Gaz et a Petrole," is to start with calcium carbide as a raw material. This, by a somewhat com- plicated process, can be changed from CaC 2 through the stages of C 2 H 2 and C 2 H 4 to alcohol, C 2 H 6 O. Barium carbide or strontium carbide can be used in the same way. Witz states that from 1 kilogram (2.2 pounds) of calcium carbide 0.8 liter (1.69 pints) of alcohol can be obtained; this is equivalent to 0.096 United States gallons of alcohol from 1 pound of carbide. Estimating the price of carbide at 3 cents per pound, which is even now some- what below the market price, 1 gallon of alcohol would therefore cost, in raw material alone, 31.2 cents, to say nothing of the cost of the chemical operations. The by-products in the case of the calcium carbide do not amount to much. There is consequently little likelihood that the so-called synthetic or mineral alcohol will ever seriously compete with gasoline or kerosene for power. The heating value of alcohol cannot be accurately computed from its chemical composition, because nothing definite is known of the arrangement of the atoms entering the composition. We therefore have to depend upon the calorimeter. The figures determined for absolute alcohol by various experimenters are as follows: Higher Heating Value per pound Lower Heating Value per pound Thomson 13310 B T U. 12036 B. T. U. Favre & Silberman 12913 B. T. U. 11664 B. T. U. The value 11664 B. T. U. is the one most generally used. Absolute alcohol has a specific gravity of 0.7946 at 15 degrees Centigrade (59 degrees Fahrenheit), so that one gallon of pure alcohol weighs 6.625 pounds, and has a lower heating value of 77274 B. T. U. GAS-ENGINE FUELS 183 One pound of C 2 H 6 O contains 0.522 pound carbon, 0.130 pound hydrogen, and 0.348 pound oxygen. According to this there will be required for the combustion of one pound of absolute or 100 per cent alcohol, (0.522 X 2.66)+ (0.130 X 8) - 0.348 0.23 = 9 pounds of air. This is the equivalent of 111.5 cubic feet of air at 62 degrees Fahrenheit, per pound of C 2 H 6 O. Commercial alcohol, however, is never pure, but nearly always contains a certain quantity of water, the admixture being measured according to volume per cent. Thus, 90 per cent alcohol means that the mixture carries 10 per cent by volume of water. The heating value of such alcohol is of course correspondingly reduced from that of 100 per cent alcohol according to the following table, due to Schottler: Absolute alcohol volume per cent Specific gravity Absolute alcohol weight per cent Lower heating value per pound B. T. U. 95 0.805 93.8 10880 90 0.815 87.7 10080 85 0.826 81.8 9360 80 0.836 76.1 8630 75 0.846 70.5 7920 70 0.856 65.0 7200 70 0.856 65.0 7200 It is required by law, in countries where alcohol is now used in the industries, to so fix the fuel that it is rendered undrinkable. This process is called " denaturizing " the alcohol. The bill passed by Congress at the last session provides for the same thing. The materials used for this purpose differ in the various European countries. Some of them try to keep the process a secret, hence some of the information given in the following table * is based upon analyses. * Zeitschrift des Vereines deutscher Ingenieure, June, 1905. 184 INTERNAL COMBUSTION ENGINES MATERIALS USED TQ DENATURIZE ETHYL ALCOHOL Country Sp. Gr. of Denat- urized Alcohol at 15 C Methylene (wood alcohol) and its impurities per cent Pyridine or Pyridine Bases per cent Ace- tone per cent Benzol per cent Benzine per cent France 0.832 7.5 2.5 0.5 Germany Denat. Alcohol. . 0.819 1.5 0.5 0.5 Motor Alcohol . . 0.825 0.75 0.25 0.25 2.0 Austria Denat. Alcohol. . 0.835 3.75 0.5 1.25 Motor Alcohol . . 0.826 0.5 trace trace 2.5 Russia 0.836 10.0 0..5 5.0 Italy Motor Alcohol . . 0.835 6.5 0.65 2.0 1.0 Switzerland 0.837 5.0 0.32 2.2 It will be noted from the table that the material most used for denaturizing ethyl alcohol is wood alcohol. The heating value of the fuel is by the addition of the denaturizing liquid changed but little in most cases. Benzol, C 6 H 6 , besides being used for denaturizing, is some- times used in larger quantities than indicated in the above table, for the purpose of increasing the heating value of the fuel mixture per pound. Benzol has a specific gravity of 0.866 and a heating value of 17190 B. T. U. per pound. A mixture of x per cent by weight of absolute alcohol with y per cent of benzol, will there- fore have a heating value of [11664 x + 17190 y] B. T. U. per Ib. If the alcohol is not absolute, its proper heating value should be substituted from the table above given. In this way from 10 to 40 per cent of benzol is sometimes employed, thus raising the heating value of the fuel, and at the same time decreasing the specific heat cost, i.e., the cost per heat unit. There is a second reason why benzol is employed. Under cer- tain circumstances there will be formed acetic acid in the products of combustion of alcohol. This causes rusting of the engine parts. On examination it will be found that this is due to a combustion GAS-ENGINE FUELS 185 with insufficient air supply, and the surest way to prevent rust- ing, therefore, is to use a good excess of air, and to have a perfect mixture. Under such conditions there will be no danger of cor- rosion. It is also found that a good addition of benzol acts as an additional safeguard. It should not be forgotten in this connec- tion, however, that the great advantage possessed by alcohol in its odorless exhaust is sacrificed to some extent by the use of ben- zol. 3. Mixing Devices for Liquid Fuels. As has been already pointed out, the fuel mixture of a gas engine is always a gas or vapor mixed with air. Hence in the case of liquid fuels special devices are required to convert these fuels into gases or vapor. Such devices are indiscriminately known as carbureters, vapor- izers, mixers, or mixing valves. In the strict sense, a carbureter is a device in which the mixture is formed by passing air over or through the liquid fuel. When no special heating of the fuel is done, these devices are applicable to the more volatile fuels only, as gasoline. The name carbureter, however, is also used when in this device gasoline is mechanically atomized or sprayed into the current of incoming air. The term vaporizer is usually employed when considerable heat for gasification or vaporization is re- quired, as is the case for kerosene, crude oil, and alcohol. MIXING DEVICES FOR GASOLINE: CARBURETERS. Carbu- reters in which the air is passed over or through gasoline have the drawback that when the carbureter chamber is filled with fresh gasoline the lighter constituents of the liquid distil off first. The parts harder to vaporize remain behind, so that the fuel mixture is anything but constant, growing leaner as the vaporization progresses. Atomizing or spraying carbureters are not open to this objec- tion, a definite amount of gasoline being injected each time. Hence they are much used*, especially in automobile work. Figure 8-1 shows a carbureter which is of what has been very aptly termed the bubbling type. The suction stroke of the engine establishes a partial vacuum in a through the check valve e, and the chamber d, interposed for reasons of safety. This causes air to enter through the screen 6, rising through c in a spray. Bubbling up through the gasoline it carries with it some of the vaporized liquid and goes through d and e to the engine, 186 INTERNAL COMBUSTION ENGINES FIG. 8-1, where a normal combustion mixture is formed by the addition of more air. The jacket water from the engine may be circulated through /, assisting in the vaporization. In special cases a part of the exhaust gases may also ^ be circulated through the bottom, g. Figures 8-2 and 8-3 are properly termed surface carbureters. In the former, the Reithman carbureter, the gasoline is fed from the reservoir a in drops into the chamber b, where it is absorbed by broad suspended wicks. From the surfaces of these the gaso- line vaporizes, adding itself to the air which strikes through chamber b on every suction stroke. Chamber / is filled with clean gravel or wire FIG. 8-2 Reithman Carbureter. screens, to act as a safety device against possible back firing. Chamber b is surrounded by a water jacket, the water in which is heated by the exhaust gases passing through the tubes d d. GAS-ENGINE FUELS 187 The Petreano carbureter, Fig. 8-3, will serve for either gasoline or alcohol. The exhaust gases pass through a central pipe, r, which is surrounded by a jacket, V. The pipe r has a covering, d, of spongy asbestos film which is constantly kept moist with the liquid to be vaporized. The liquid fuel enters by one open- ing in the top, the air by an- other. The chamber V has four cones, as shown, two of which are also partially covered with asbestos. By means of these cones the fuel vapor and air are thoroughly mixed before enter- ing the chamber M and finally the suction pipe to the engine. The small openings, o, are for the purpose of drawing off the heavier parts of the liquid, those that vaporize less easily, if there be any, in order not to interfere with the regularity of carbureting. FlG - *- 3 - Petreano Carbureter. Atomizing or spraying carbureters are, however, much more frequently employed, and the following to be described are .all of this type. 1 e Semi -Sectional SECTION ON A-B. 'Views. SECTION OMA-t. Angle Pattern. Vertical Pattern. FIG. 8-4. Lunkenheimer Mixing Valves. 188 INTERNAL COMBUSTION ENGINES The simplest design of this type of carbureter is shown in Fig. 8-4, which shows two of the Lunkenheimer carbureters. In either design, gasoline is furnished through the opening 0, and its admission to the valve seat of the main suction valve E is regulated by the needle valve A. On the suction stroke of the engine the valve E automatically lifts, and the air rushing through this opening carries with it a certain amount of gasoline flowing from the small hole K. It is quite evident that this type of car- bureter gives best service on hit-and-miss engines. For any other type a varying load would probably cause trouble, as there is no automatic regulation of the gasoline supply. Still of very simple design, but capable of regulation of the gasoline supply, is the Sintz carbureter, Fig. 8-5. The lift of the gasoline needle valve b is regulated by the lift of the mechanically regulated suction valve a. By means of the cock c it is possible to admit part of the air uncarbureted to the cylinder. Figures 8-6 and 8-7 are examples of the so-called float-feed type of carbureters. The Daimler, Fig. 8-6, is well FIG. 8-5. Sintz Carbureter, known. A float, B, in the cham- ber, A, operates a pair of counterweight levers, E, and through them the valve spindle, D, which controls the admission of gasoline at C, keeping it at a constant level in the nozzle. Gaso- line enters at N from a tank under slight pressure or at a higher level. is a cloth filter and the plug P serves to catch any grit that may be brought in. The float chamber is vented through a small hole in the cap over the valve spindle to relieve any pres- sure which may be formed due to varying positions of the float B. The action of the carbureter is as follows: At each charging stroke of the engine air is drawn into the annular chamber H and passes with great velocity through the drop tube F, surround- ing the gasoline nozzle. The gasoline is drawn in a jet from the nozzle and, with the air, striking the deflector K, the two are very thoroughly mixed, passing to the engine through M. An aux- GAS-ENGINE FUELS 189 iliary air supply can be admitted through the cap at the top, the openings through which can be regulated. FIG. 8-6. Daimler Carbureter. FIG. 8-7. Abeille Carbureter. The Abeille carbureter, Fig. 8-7, embodies the same idea as the Daimler. The float B, by operating the lever D, opens and closes a needle valve at the lower end of the -weighted spindle C, to maintain a constant level just below the opening in the nozzle 190 INTERNAL COMBUSTION ENGINES E. On the suction stroke, air enters at H, and rushing up through the double cone draws the gasoline from the nozzle, atomizing it by striking the perforated cone G. A secondary air supply is admitted through the cap L and the holes in G, being regulated by the position of the cap L. FIG. 8-8. De Dion Carbureter. Of more complicated design, but in principle the same as the other float-feed carbureters, is the De Dion, Fig. 8-8. Air is admitted through the tube t. By the position of the valve V (see horizontal section), which position can be regulated by the lever/ (see vertical section), part of the air goes directly into the dis- charge pipe t' ', while the rest is deflected downward to be car- bureted. The course of this latter body of air is indicated in the right vertical section by the arrow. The level of the gasoline in the nozzle D is again kept constant by the annular float C in the chamber H, operating the lever G and through it the valve F. It will be noted that none of the atomizing or spraying carbu- reters for gasoline so far described are heated in any manner. Occasionally, however, we find one in which the atomizing is assisted by heat. The W. Hay vaporizer, Fig. 8-9, is of this type. Gasoline enters the annular chamber, a a, through the pipe d. From this chamber a number of small openings lead through the seat of the suction valve E. Some of these openings are provided with adjusting screws as shown at the left. On GAS-ENGINE FUELS 191 the opening of E, the inrush of air atomizes the gasoline flow- ing from these small openings, and the current of air and gaso- line striking the wings e of the fan h, supported on the spindle /, sets the fan in motion, thus promoting a thorough mixture of air and vapor before the fuel finally passes through x to the admission valve A. The ex- haust gases from B are passed through the chamber F, finally escaping through the slotted openings ' g. In their passage they heat both the gasoline storage chamber a a and the fan chamber. Occasionally we also find carbureters which combine the several principles of those de- FIG. 8-9. W. Hay Vaporizer, scribed. Thus in the Gautier carbureter, Fig. 8-10, the gasoline is admitted through A, the supply being regulated by the valve FIG. &-10. Gautier Carbureter. K, which opens at the proper moment owing to suction through pipe E. Part of the supply falls on to the saucer F, and from there into the reservoir H. Into the liquid at H dips a pipe G r 192 INTERNAL COMBUSTION ENGINES supported as shown. The air supply through / circulates through the chamber L, bubbles through the liquid at H into the chamber surrounding the saucer F, licks up some of the gasoline in the saucer, and the mixture finally escapes through E. VAPORIZING DEVICES FOR CRUDE OIL AND KEROSENE. While it is fairly easy to satisfactorily atomize or vaporize gasoline and to maintain the mixture, the thing is a little more difficult with kerosene, and much more so with crude oil. In either case the agency of heat is required, and this is applied either in a separate vaporizer or retort, or the kerosene or crude oil is injected directly into the cylinder, the vaporization taking place in the combustion chamber. One trouble with kerosene is the readiness with which some of the vapor is condensed on striking comparatively cool sur- faces. This may happen on the mixture striking parts of the water-cooled walls. In such a case part of the fuel may go through the cylinder unburned, and this is a point that should be care- fully guarded against when a special vaporizer for kerosene is employed. In the case of crude oil, the heating in the vaporizer results in the distillation of the lighter products first. The amount of vapor formed will naturally be less and less as the distillation proceeds, resulting in a constant impoverishing of the fuel mix- ture. The remedy would therefdre seem to be a constant supply of fresh oil and a removal of the old before it has commenced to seriously decrease its yield. This is what is actually done in some devices. The method, however, naturally results in the fact that the plant can only use part of the crude oil. It is claimed by some that only 10 per cent of the oil is so withdrawn from the vaporizer, but this seems extremely doubtful. A crude oil-air mixture is open to the same objection as a kerosene mixture as regards condensation of some of the heavier hydrocarbons and consequent loss. Both of these mixtures are also subject to cracking; that is, a breaking up of the heavier hydrocarbons into the lighter with a consequent deposit of carbon. For large power plants the best solution in the case of crude oil would seem to be the use of some type of oil-gas producer. Some of these are in actual use along the Pacific coast, and will be described later. GAS-ENGINE FUELS 193 Figure 8-11 shows the kerosene atomizer used on the Hornsby engine. With it is combined the regulating valve. Kerosene is pumped up through the lower right-hand pipe in the sectional cut. The governor regulates through c the position of the over- flow valve bj the surplus kerosene flowing back through d to the reservoir. The kerosene at the moment demanded by the engine FIG. 8-11. Hornsby- Akroyd Atomizer. is forced by the pump through the plug a, issuing from its end in a fine spray. This plug is kept cool by a water jacket as shown. Another atomizing kerosene carbureter is the Gibbon, shown in Fig. 8-12. In this a valveless pump x w, actuated by a cam, injects kerosene from the tank up through the atomizing opening into the chamber U. This is furnished with wings, U', to pre- sent a larger surface for heating. Chamber U is directly connected with the combustion chamber, and is surrounded by a light case to prevent radiation. At the start it is heated by a lamp, but after a short period of operation the charge ignites by compres- sion, as is the case also in the Hornsby engines. In the Crossley vaporizer, Fig. 8-13, a certain fixed amount of oil is measured and drawn into the vaporizer by the air on the . .?. . ... FIG. 8-12. Gibbon opening of the valve. In this case ignition Kerosene Vaporizer. 194 INTERNAL COMBUSTION ENGINES is produced by means of a hot tube, and the engine is of course governed on the hit-and-miss system. The lamp heating the hot tube at the same time heats the vaporizer. FIG. 8-13. Crossley Vaporizer. The Priestman engine is designed for either crude oil or kero- sene. Its vaporizer is shown in Figs. 8-14 and 8-15. Oil is ad- mitted to the spraying nozzle K from a reservoir under pressure. EXHAUST FIG. 8-14. Priestman Vaporizer. Petrofeum FIG. 8-15. Priestman Vaporizer. GAS-ENGINE FUELS 195 The pump furnishing the air for this purpose also furnishes air to the annular space J , surrounding the nozzle K, for the purpose of atomizing the oiK The finely divided oil enters the vaporizer chamber and, mixing with the larger body of air admitted on the suction stroke to the vaporizer chamber through the valve L, passes on to the cylinder. The vaporizer chamber, Fig. 8-15, is heated by the exhaust gases passing through the jacket space C, thus vaporizing the oil spray in its passage. On starting, the vaporizer chamber is heated by a lamp whose hot gases pass through the jacket spaces d. The governor, by regulating the position of the plug valve H , regulates at the same time both the oil and the air supply to the vaporizer. A crude oil vaporizer quite extensively used on the Pacific coast is the " Economist" retort, Fig. 8-16. This consists of FIG. 8-16. " Economist" Crude Oil Vaporizer. two concentric shells as shown. In the jacket space between them the engine exhaust gases are circulated to furnish the heat for vaporization. The inner drum is slowly rotated upon its axis, as shown by the gearing at the right, and it is furnished on the interior with a continuous spiral rib. The crude oil is admitted through the central pipe at the left. By the rotation of the drum it is carried up the sides, turned over and over, thus exposing a thin sheet all the time to the action of the heat, which is most favorable for complete vaporization. At the same time, owing to the action of the spiral rib, the oil is slowly carried forward through the drum. The residue is finally discharged through an outlet at the right, in a manner not clearly shown. The makers claim that during the progress of the oil through the drum the temperature is maintained just high enough to get all the gas from the oil. The method of regulation by this device is not clearly stated. If it is done by regulating the oil 196 INTERNAL COMBUSTION ENGINES supply to the retort according to the load, there would be con- siderable lag in the regulation; if done by regulating the air supply through the retort there would be a large variation in the composition of the mixture. An auxiliary air supply between generator and engine would help, but whether this is used or not is not stated. Retorts as above described are made up to from 125 to 150 horse-power. Beyond this they become unsatisfactory, and other devices have to be employed. The one that seems to promise best at present is an oil-gas generator of the type of the Lowe. The principle of the Lowe system is to heat up to a very high temperature a mass of firebrick checker work contained in a fire- brick-lined steel shell, by means of a crude oil-air blast. When the desired temperature is reached the chimney connection to the generator is closed off, and crude oil and superheated steam in an intimate mixture are sent down through the highly heated checker work. The result is the formation of an impure oil water gas with a good deal of lampblack as a by-product. This by-product may be utilized as fuel to furnish the power necessary for blowing engines, etc., around the plant. The gas itself is washed in the ordinary manner, and is a high-grade gas of good illuminating power. The production of this gas is therefore carried on in the two stages, a heating-up period and a period of make. From published figures the efficiency of these oil-gas pro- ducers is as yet not very high. In the Journal of Electricity, Power and Gas, September, 1904, a consumption of 11.22 gallons of oil per 1000 cubic feet of gas made is given as the average for one plant for July. Other plants have shown 9 and 10 gal- lons per 1000 cubic feet of gas of about the same heating value. Assuming the crude oil to weigh 7 pounds per gallon, its heating value at 19000 B. T. U. per pound, and the heating value of the gas at 650 B. T. U. per cubic foot, the efficiency of generation on cold gas would be for the most favorable case above given 1000X650 . = 9X7X19000 VAPORIZING DEVICES FOR ALCOHOL. The alcohol engine has perhaps received its greatest development in Germany. GAS-ENGINE FUELS 197 It is for that reason that we shall have to turn to the literature of that country for the best and latest information on the details of alcohol engines. The following material is a reprint from an article published by the writer on the alcohol question in Marine Engineering, June, 1906. The information is mainly due to the work of E. Meyer and of A. Schottler, also to the discussions by Giildner, Diesel, and others. * Regarding the formation of the fuel mixture with alcohol, it is found that it is less volatile than gasoline, but easier to handle than kerosene. In nearly all of the vaporizing devices for alcohol now on the market, the agency of heat, usually the exhaust heat of the waste gases, is used to aid in the formation of the mixture. This scheme has the drawback that no heat is available at the start when the engine is cold. To avoid an open flame for the purpose of heating the vaporizer at the start, which is both dangerous and cumbersome, the engines in most cases are started with gaso- line, and when, after a few strokes, enough heat is available, the change is usually made by throwing over a single lever. In tests of ten different engines made by Meyer, it was shown that this change to alcohol could be made in the slowest case in 6 minutes and 40 seconds, the time of the fastest being 55 seconds. Based on the manner of heating the vaporizer, we can dis- tinguish the following classes: 1. Those in which no heat is employed. 2. Those in which the air is pre-heated. 3. Those in which the mixture is heated and superheated. Of the first type is the Deutz, Figs. 8-17 and 8-18. When the engine is regulated by the throttling method, and not by the hit-and-miss system, it has been found that no pre-heating of air or fuel mixture is required. The reason for this is un- doubtedly that in a hit-and-miss engine, under less than normal load, a succession of misses cools the cylinder down so far as to throw down some of the alcohol vapor on the next explosion, unless it is superheated. The Deutz engine is governed by throttling. The inlet valve / is actuated, through the levers * E. Meyer, Zeitschrift d. V. d. I., 1903, pages, 513, 600, 632, 669; R. Schottler, Zeitschrift d. V. d. I., 1902, pages, 1157, 1223; H. Giildner, Zeit- schrift d. V. d. I., 1902, page 623; Diesel, Zeitschrift d. V. d. I., 1903, page 1366. 198 INTERNAL COMBUSTION ENGINES shown, by the cam a, which is of taper form and under the control of the governor Fig. 8-17. Upon the position of a depends the length of time the valve / is open. Through the bell crank c d e the cam also acts upon the plunger of the fuel pump h, operat- ing in such a way as to cause suction during the first part of the cam movement, and pumping of the liquid during the second. Thus the fuel is injected during the second FlG> 8 _] 7. _ Deutz Alcohol half of the suction stroke only, insur- Vaporizer. ing a rich mixture around the igniter. The alcohol is forced through the sprayer or atomizer i, Fig. 8-18, into the current of air which enters through the valve k. Thus no pre-heating what- ever is done, but the atomizing is thorough; and the ports into the cylinder are as direct and short as possible, hence no vapor is thrown down. The Altman vaporizer, Fig. 8-19, is of the second class. The air pipe a b is surrounded at its lower end by the exhaust pipe; the air is thus pre-heated by the exhaust gases. A regulating valve for the air is placed at e. This, when drawn up- ward, decreases the amount of air passing, but always makes the air current strike through the upper part of the pipe, in this manner directing it always against the fuel nozzle d. The inlet valve c is operated by the lever /, actuated by the cam /, through the pendulum hit-and-miss governor o m p. This valve lever / at the same time opens the fuel valve d, through the reach rod shown and the finger h i, Fig. 8-21. How this is done is shown in Fig. 8-20. The lever /, on being depressed, forces down the point of the screw k, thereby turning the reach rod about its axis, which depresses FIG. 8-18. Deutz Alcohol Vaporizer. GAS-ENGINE FUELS the point i, Fig. 8-21 , opening the valve d. The amount of opening depends upon the position of the screw k, and this can be very finely adjusted by the worm and wheel arrangement shown. In this vaporizer the fuel supply is atomized partly by the current of ir ; and is afterwards vaporized by the heat of the pre-heated air. FIG. 8-19. Altman Alcohol Vaporizer. FIG. 8-20. FIG. 8-21. The following three vaporizers are of the third class. Fig. 8-22 shows the Swiderski-Longuemarre. Here also the exhaust gases are used for heating. They pass through the annular chamber a, and their action is aided by the radiating webs b b. The float d maintains a constant level in the supply chamber. From this chamber the flow of alcohol is regulated by the needle valve /. 200 INTERNAL COMBUSTION ENGINES The liquid flows into the space g, and overflows through a num- ber of small openings h h. Air entering through i is made to pass partly outside, partly inside, the concentric spaces created FIG. 8-22. Swiderski-Longuemarre Alcohol Carbureter. FIG. 8-23. Dresden Alcohol Vaporizer. FIG. 8-24. Dresden Alcohol Vaporizer. GAS-ENGINE FUELS 201 by the sleeve k. The amount of air passing outside is regulated by the openings n n which are controlled by the lever I. The air currents passing upward carry along with them some of the liquid, the mixture being heated by the exhaust gases in a. The per- ,. forated plate o tends to aid in forming a uniform mixture. The vaporizer of the Dresdener Gasmotorenfabrik is shown in Figs. 8-23 and 8-24. In this case the warm cooling water of the engine is used for heating. It enters the water space at x, Fig. 8-24. On very cold days the vaporization may be assisted at the start by pouring some hot water into the funnel a. Air enters at y, 8-23. The inlet valve b is automatic. It may be pushed down at will at the start by pressing down on the projecting stem c. The downward movement of the inlet wave opens the fuel valve d, to which alcohol is furnished through the needle valve e, Fig. 8-24. Through a number of fine openings the fuel flows into the current of air and is carried along with it, the thorough mixing being assisted by the current striking the cone g. As will be seen from the drawing, the heating of the charge cannot be very high. In the first place only the comparatively cool jacket water of the engine is used, and secondly the mixture itself is not in the heated chamber for any length, of time. In contradistinction to the Dresden vaporizer, the Diirr, Fig. 8-25, pro- duces a highly heated mixture. Air enters at x and its amount is regu- lated by the throttle valve a. The inlet valve b is automatic. Alcohol is supplied through the needle valve c, as shown, so that when b is closed no flow of alcohol takes place. The current of air charged with alcohol particles passes down through d, up the annular space e, and out at y to the cylinder. The exhaust gases enter at z, and by means of baffle plates FIG. 8-25. Diirr Alcohol Vaporizer. 202 INTERNAL COMBUSTION ENGINES are made to take the course shown by the arrows, through the space /. Further, the space e is filled with a large number of metal spirals, which connect the outside wall of e with its inside wall, thus furnishing a large heated surface to the passage of the charge, and facilitating the transfer of heat from the space / to the space d. Every possible way is therefore made use of to apply the heat of the exhaust gases, and this vaporizer consequently fur- nishes a mixture more highly superheated than that of the others. FIG. 8-26. Gasoline-Alcohol Vaporizer. Finally, Fig. 8-26 shows what may be called a double float carbureter, which is the form that alcohol vaporizers are likely to take. This is used on the Marienfelde machines. Assume that the chamber a is used for gasoline, b for alcohol. The needle supply valves can be held closed by the springs c and d, as shown. On starting with gasoline, the chamber a is used. Spring c is pushed aside so that fuel can enter, being kept at constant level by the float. The valve g is so set that the path is open for the air from h past the gasoline nozzle e, through g into the cylinder. At every suction stroke the in-rushing air is then charged with gasoline issuing in a small jet from e. If it is de- sired to change to alcohol, spring c is pushed into place, spring d GAS-ENGINE FUELS 203 is pushed aside, and valve g is thrown over into the position shown in Fig. 10, all the work of a moment. The air supply to this vaporizer is pre-heated. It is quite evident from an examination of the vaporizers above described that the final temperature of the mixture is very differ- ent in the different devices. Upon this temperature, however, de- pends in a great measure the only other point of difference between gasoline and alcohol engines, i.e., the amount of compression. All other things being the same, that fuel mixture entering the cylinder at the highest temperature will soonest give rise to pre- ignition, or at least to pounding, under an increase in compression. High temperature of charge also effects engine capacity unfavor- ably. It therefore becomes important to determine approxi- mately the lowest practical temperature of vaporization, and the heat necessary. Of course the amount of heat required depends upon the amount of alcohol (and its purity) per pound or cubic foot of air. Assuming that 90 volume-per cent alcohol is used, the theoretical amount of air required for perfect combustion is 7.8 pounds. Assuming that an excess of 50 per cent of air is used, which is a desirable allowance, 1 pound of 90 per cent alcohol would require in round numbers 11.7 pounds of air. With the air temperature at 60 degrees Fahrenheit, and the atmospheric pressure 14.7 pounds per square inch, this amounts to 0.0065 pound of 90 per cent alcohol per cubic foot of dry air. 90 (volume) per cent alcohol is equivalent to 87.7 (weight) per cent, so that 1 pound of air will carry, according to the above assumed ratio of mixture, 0.877 X j-j-s = 0.075 pound of absolute alcohol, and 0.123 X ^ = 0.010 pound of water. To compute the air temperature required so that it may take up the above quantities of alcohol and water vapor, we must know the relation between the temperature and the degree of saturation. Meyer in his computations used the data contained in the Physikalisch-Chemische Tabellen of Landoldt & Bornstein. 204 INTERNAL COMBUSTION ENGINES Temperature Degrees F Vapor Tension Inches Mercury 1 pound of air contains in saturated condition, in pounds Alcohol Vapor Water Vapor At 28.95 inches Hg. Press. At 26.05 inches Hg. Press. Alcohol Vapor Water Vapor Alcohol Vapor Water Vapor 50 0.950 0.359 0.055 0.008 0.061 0.009 59 1.283 0.500 0.075 0.011 0.084 0.013 68 1.733 0.687 0.104 0.016 0.117 0.018 77 2.325 0.925 0.144 0.022 0.162 0.025 86 3.090 1.240 0.200 0.031 0.227 0.036 104 5.270 2.162 0.390 0.063 0.450 0.072 122 8.660 3.620 0.827 0.135 1.002 0.164 For our purpose the figures of the table have been transposed into English units. In the ordinary case the air drawn into the vaporizer is not dry, but contains a certain quantity of water. Assume that the air is at a temperature of 59 degrees and just saturated. At a pressure of 26.05 inches of mercury this would correspond to 0.013 pound of water per pound of, air in its initial condition. Now in the case of the average mixture above computed, the tem- perature of vaporization must be high enough to vaporize an additional 0.010 pound of water, making the total 0.023 pound that the air must contain per pound. It is seen from the table that a temperature of 77 degrees is quite sufficient to do this. It is also seen that at this temperature the air may take up 0.162 pound of absolute alcohol, while the quantity in the above mixture is only 0.075 pound per pound of mixture. At a tempera- ture of 77 degrees the mixture ready for the cylinder may there- fore contain the alcohol vapor in a state of some superheat. If therefore the temperature of the walls with which the mixture comes in contact is not less than 77 degrees, no fear of condensa- tion of alcohol vapor need be entertained. In this connection a statement in the Engineering Record is of interest. It is there claimed that the consumption of alcohol with the jacket water leaving at 60 degrees Fahrenheit is 100 per cent higher than with jacket water leaving near 212 degrees; i.e., with cooling by GAS-ENGINE FUELS 205 vaporization. In the light of the above facts, some such increase in the consumption is quite possible. In order to convert the liquid alcohol into vapor, a certain quantity of heat is required. According to Regnault, this amount is, for the various temperatures given, and computed above 32 degrees Fahrenheit, as follows: At 32 F 425.7 B. T. U. per pound 68 F 453.6 B. T. U. per pound 122 F 475.2 B. T. U. per pound 212 F 481.1 B. T. U. per pound The specific heat of liquid alcohol is close to 0.6, so that in order to convert the quantity of 90 volume-per cent alcohol con- tained in the assumed mixture to alcohol vapor at 77 degrees Fahrenheit would require approximately, assuming the liquid alcohol at 60 degrees Fahrenheit, 0.075 X [458 - (28 X 0.6)] + [0.010 X 1100] = 44.1 B. T. U., where 1100 B. T. U. is assumed as the heat of vaporization of water under the existing conditions. Now the heating value of 90 volume-per cent alcohol has been shown to be 10080 B. T. U. per pound, so that the heating value on one pound of our assumed mixture will be 0.075 X 10080 = 44 1 756 B. T. U. The heat of vaporization required is therefore ^~ 756 = 5.8 per cent of the heating value of the fuel. It can be shown that the amount of heat is easily obtainable from the exhaust gases. It can also be shown that the problem may be solved by pre-heating the air only, for, assuming that the specific heat of air at constant pressure is 0.238, we would have to pre-heat the 44.1 air for the assumed mixture to ~ + 77 = 262 degrees Fahren- .Zoo rieit, which is easily possible. If, on the other hand, not the air but the mixture is heated, then the walls need to have a temperature only sufficiently higher than 77 degrees to transfer the required amount of heat for vapori- zation to the mixture in the time available. To furnish more heat than this is harmful, if anything, for it affects unfavorably both the possible degree of compression and the capacity of the machine. The cooler the mixture after formation and vaporiza- tion, the better. CHAPTER IX GAS-ENGINE FUELS! GAS FUELS OUTSIDE of producer gas, which has been treated in a previous chapter, the gases used for gas engine fuel are: 1. Illuminating gas. 2. Oil gas. 3. Coke oven gas. 4. Blast 'furnace gas. 5. Acetylene. 6. Water gas. 7. Natural gas. i. Illuminating Gas. -- Illuminating gas is made by dis- tilling bituminous coal in retorts. From 100 pounds of average coal are obtained about 400-450 cu. ft. of cooled gas, 50-70 Ib. of coke, 4.25-4.75 Ib. of tar, and 8-10 Ib. of ammonia liquor. Each 100 pounds of coal also require about 20 pounds of coke for the heating of the retort. The composition of the gas varies constantly somewhat even in the same plant. The average composition is about 45-48 per cent by volume of hydrogen, 35-38 per cent CH 4 , 5-8 per cent CO, and the rest heavy hydrocarbons, oxygen, nitrogen, and carbon dioxide. The gas owes its illuminating power to the heavy hydrocarbons it contains. Its heating value, however, is not proportional to its candle power. To determine the heating value the best way is to use a calorimeter. It may,, however, with sufficient accuracy be computed from the analysis of the gas. Varying somewhat in this same locality, the average lower heating value is probably not far from 600 B. T. U. per cubic foot. Its density averages about .4, air being 1.0, its average weight per cubic foot, therefore, being not far from .032 Ib. The following table shows a few typical analyses of illuminat- ing gas.* * Mostly from Poole, the Calorific Power of Fuels. 206 GAS-ENGINE FUELS 207 H CH 4 Hydro- carbons C0 2 CO O N B.T.U. cu. ft. Lower Value Newton, Mass Cleveland 50.59 34.80 34.80 28.80 5.23 11.20 1.16 .20 6.16 10 40 40 2.06 14 20 599 657 Boston 47.49 38.67 5.21 1 04 6 74 85 651 Cincinnati 45.85 39.26 5.17 .82 4.78 41 3.71 645 Birmingham 40.23 39.00 4.76 1.50 4.05 36 10 10 671 Glasgow 39.18 40.26 1000 29 7 14 06 3 830 1 Liverpool 36.44 44.28 7.90 1 70 3.39 1P 6.10 792 1 Hanover Paris 46.27 50.10 37.55 33.10 3.17 5.80 .81 1 50 11.19 630 50 1.01 2 70 661 667 Average f 43.44 3730 6 49 1 00 6 68 32 5 77 686 2. Oil Gas. Oil gas is made by vaporizing and superheat- ing crude oils. It may be made by vaporizing these oils in retorts which are externally heated, as in the case of the Pintsch method, or the manufacture may be carried on as in the Lowe process, in which, as previously described, the generator is first internally heated by burning crude oil, the oil to be gasified being then sent into the heated chamber together with steam under exclusion of air. Pintsch gas, much used for railway car illumination, contains, according to Giildner,* 17.4 per cent C 2 H 4 , 58.3 per cent CH 4 , and 24.3 per cent H by volume. Another gas obtained from a by-product paraffin oil showed the following composition by volume: 28.9 per cent C 2 H 4 , 54.9 per cent CH 4 , 5.6 per cent H, 8.9 per cent CO, and 9 per cent CO 2 . Oil gas as made by the Lowe process is a water gas ; the com- position will therefore show much more H than is indicated in the above analysis of Pintsch gas. Wyer f in his Gas Producer gives the following figures : 32 per cent H, 48 per cent CH 4 , 16i per cent C 2 H 4 , 3 per cent N, .5 per cent O. 1 Values evidently too high. * Giildner, Entwerfen und Berechnen der Verbrennungsmotoren. f Wyer, Producer Gas and Gas Producers. 208 INTERNAL COMBUSTION ENGINES Giildner estimates that the yield of gas from 1 pound of oil in the Pintsch process is from 7-10 cu. ft. of cooled gas, about .75 Ib. of coke being used in the same time for heating. This amounts to about 100 Ib. of oil or about 14 gallons of oil per 1000 cu. ft. of gas, in the most favorable case, as against 9-10 gallons per 1000 cu. ft. in the Lowe process. The heating value per cubic foot of the Pintsch gas, however, is higher than that of the Lowe in the ratio 900 of about - , so that on the basis of thermal efficiency the two 650 methods of making oil gas are probably not very far apart, with the chances in favor of the Lowe system on account of the coke used for heating in the other system, which must, of course, be considered in making thermal calculations. 3. Coke Oven Gas. Coke oven gas when made in the old type Bee-hive oven is fundamentally the same as illuminating gas. Compare the following analysis given by Wyer of a sample of this gas with the average analysis given for illuminating gas on page 208. H, 50.0 per cent; CH 4 , 36.0 per cent; C 2 H 4 , .4 per cent; N, 2 per cent; CO, 6 per cent; O, 5 per cent; CO 2 , 1.5 per cent. When made in modern by-product ovens, however, the gas yield is sometimes divided and that, part of the gas used for fuel has a somewhat different composition. From a diagram published by the United Coke and Gas Company of New York,* it appears that the gas evolved during the coking of the charge in a retort is divided into two parts. The entire coking period covers nearly 25 hours. The gas evolved during approximately the first ten hours, called the surplus or rich gas, is separated from that made during the rest of the period, called fuel gas. The surplus gas is high in illuminating power and in heating value, approximately 720 B. T. U. per cubic foot. The fuel gas has an average heating value of about 560 B. T. U. per cubic foot. The figures quoted are for a medium volatile coal. The rich gas from a ton of this coal in an actual case amounted to 4300 cubic feet, which was 46 per cent of the total yield per ton, this part of the gas carrying 52 per cent of the total heat value of the gas. A ton of this coal therefore yields about 9400 cu. ft. of gas. The same treatise above quoted gives the following gas * The United Otto System of By-product Coke Ovens. GAS-ENGINE FUELS 209 analysis for a coal carrying from 30-32 per cent of volatile matter: Illuminating or Rich Gas Fuel Gas Illuminants 58 28 CH 4 40.8 29.6 H 37.6 41.6 CO 5.6 6.3 CO 2 3.7 3.2 o .4 .4 N 6 1 16.1 B T U per cu ft higher value 100.0 7303 100.0 551.3 Where no illuminating gases are desired the entire gas yield is recovered together. The gas is excellent for power purposes except for the somewhat high percentages of H which render the fuel mixtures liable to pre-ignition under high compression in the cylinder. 4. Blast Furnace Gas. The blast furnace is really a large gas generator, with the difference that to the charge of fuel is added the burden of ore and of flux, and that the blast is air alone without admixture of steam. Owing to the calcination of the flux, which is limestone ordinarily, and to the fact that no steam is used, the gas is high in CO 2 and contains little H, the main combustible being CO. This gas had been used a long time in hot blast stoves for blast heating and under boilers to produce steam for power purposes around the works. It was Thwaite in England and Liirman in Germany who about twenty years ago called attention to the fact that this gas, although low in heating value, could be readily burned in gas engines when suitably com- pressed. The credit of having carried out this idea first on a large scale belongs to the Societe Cockerill of Seraing, Belgium, who about 1898 put a 150 horse-power engine using this gas into operation. It is estimated roughly that for every ton of pig iron produced one ton of coke is required, the combustion resulting in about 5 tons of gas. Taking a furnace, therefore, with an average daily capacity of 200 tons of pig iron, the gas available per hour amounts 210 INTERNAL COMBUSTION ENGINES to about 41.6 tons or 1,000,000 cu. ft. Estimating that 500,000 cubic feet are necessary for the operation of the hot blast stoves, this leaves 500,000 cu. ft. available, which if directly used in gas engines would develop about 5000 I. H. P. Of this amount 1000- 1200 horse-power are probably required for power purposes around the furnace, leaving from 3800-4000 horse-power avail- able for other work. The same amount of gas, if used under boilers, would have resulted in only about 1200 boiler horse-power, or perhaps about 2400 horse-power total in steam engines, leaving 1200 horse-power available for other purposes. The composition and heating value of blast furnace gas natu- rally varies somewhat in different furnaces, and even in the same furnace under varying accidental conditions of operation. A large number of determinations led M. Witz * to the conclusion that the average heating value of a standard cubic foot was 110 B. T. U. and that it very rarely fell be'low 95 B. T. U., or rose above 118 B. T. U. The average composition of the gas, according to Ledebur, appears to be % by Volume by Weight CO 24.0 24.0 CO 2 12.0 17.0 H . . 2.0 .2 CH 4 2.0 .8 N . 60.0 58.0 The above analysis shows no water vapor, some of which is present in the gas as it leaves the stack, and it therefore probably refers to cleaned gas. It is apparent that it is an excellent gas for internal combustion engines. Its low content of H makes it suitable for high compressions, thereby overcoming any objec- tion that may be made regarding its low heating value and difficulty of ignition. The most serious trouble encountered in the use of blast fur- nace gas is the fact that it carries more or less dust, which renders cleaning of the gas imperative before use in engines. It is also apt to carry metallic vapors, which do not, however, become * Moteurs a Gaz et a Pet role, p. 267. GAS-ENGINE FUELS 211 harmful until after combustion in the cylinder. The amount of impurities carried depends altogether upon the kind of ore re- duced in the furnace. In some cases it is so slight that the ordi- nary dust settlers combined with a scrubber of some sort are quite sufficient to reduce the amount to below the allowable limit. This is, however, the exception, and the fact that the gas must be cleaned, and thoroughly cleaned, cannot be emphasized too strongly. , It is comparatively easy to take out the coarser dust carried by the gas by appliances which have long been in use for this purpose to prepare the gas for stoves and boilers. The fine dust, however, causes more trouble, and special cleaning apparatus is necessary to reduce the amount carried. The ordinary method of procedure is to give the gas a prelim- inary cleaning by allowing the coarse dust to settle. The fine dust, together with the water vapor and the metallic vapors, are then taken out by passing the gas through washers, of which there are various forms, as spray towers, centrifugal fans, etc. Coke scrubbers are not satisfactory on account of the clog- ging up by dust which soon takes place. A dry scrubber, filled with sawdust or shavings, is sometimes used to complete the outfit of cleaning apparatus. The amount of water used in the washers varies with different types. It may be from 5-50 gallons per 1000 cu. ft. of gas cleaned, depending upon the efficiency of the apparatus. The amount of dust finally carried by the gas should not be higher than about .2 grains per cubic foot. 5. Acetylene. It is only in recent years that the means for making acetylene gas in any quantity were found. To-day calcium carbide is made in quantity in the electric furnace. The gas is produced by decomposing this carbide by means of water, as per following reaction: CaC 2 + 2 H 2 O = C 2 H 2 + CaOH 2 O The generators employed usually regulate the amount of water supplied to the carbide receptacle. The gas is led to a holder, which by the position of its bell regulates the water supply. The amount of gas produced per pound of carbide should be theoreti- cally 5.45 cu. ft. of dry acetylene gas. Owing to impurities, how- 212 INTERNAL COMBUSTION ENGINES ever, this is usually reduced to about 4.8 cu. ft. Combustion of this gas takes place according to the formula C 2 H 2 + 5O - H 2 + 2CO 2 Its heating value is 20673 B. T. U. per pound, or 1499 B. T. U. per cubic foot lower value. The gas is distinguished by low temperature -of ignition, lower than that of H, approximately 865 degrees Fahrenheit, high velocity of flame propagation at the best ratio of air to gas, about 12 to 1, and high maximum temperature of explosion owing to the high heating value. The first of these, low-ignition tempera- ture, leads to pre-ignition and requires the use of comparatively low compression pressures. 6. Water Gas. The theory of the production of water gas has been already outlined in a previous chapter. The average composition of the gas may be taken to approximate by volume 42 per cent CO, 44.5 per cent H, 3.5 per cent CO 2 , the rest being O and N. In medium sized well-handled generators each pound of coke will yield about 32 cu. ft. of gas, each pound of good anthracite coal from 24-30 cu. ft. The average lower heating value of the gas may be taken at 290 B. T. U. per cu. ft. 7. Natural Gas. Natural gas is found in many parts of the world. It has, however, perhaps received the most extended use as a fuel for power in the United States. It is there found in New York, Pennsylvania, Ohio, Indiana, West Virginia, Kentucky, Tennessee, Colorado and California. This gas is not of constant chemical composition in the different wells, and not constant even in the same well. Marsh gas, CH 4 , however, is nearly always the main constituent. According to Poole, the Ohio and Indiana fields yield a gas of the most constant composition. The follow- ing is the composition at Findlay, Ohio, and is typical of the field. H CH 4 C 2 H 4 O CO CO 2 N H 2 S 1.64 93.35 .35 .39 .41 .25 3.41 .20 per cent by vol. The above composition, however, is sometimes radically changed. Thus a gas well near Pittsburg changed the composi- GAS-ENGINE FUELS 213 tion of the gas in three months from 9.64 per cent H to 35.92 per cent H, mostly at the expense of Marsh gas. The heating value of this gas is high, the above Findlay gas showing a lower heating value of 962 B. T. U. per cubic foot as com- puted. It is a good fuel for gas engines, as it is cheap and not very liable to pre-ignition when the hydrogen is low. It is, how- ever, of decreasing importance on account of the gradual failure of the supply. The following table, and Fig. 9-1, give a recapitulation of the most important data for the fuel gases most often found in gas engine practice. It is to be remembered that the figures given represent approximate average results only, but for rapid calcula- tion they are sufficiently accurate. AVERAGE APPROXIMATE DATA FOR FUEL GASES Wt. per Lower Least air No. cu. ft. Standard Density Heating Value required for Combustion Kind of Gas in Air = l per cu. ft. cu. ft. pounds B. T. U. per cu. ft. 1 Illuminating gas .032 .40 565 5 25 Natural gas .045 .55 950 9.10 3 Blue water gas . .057 .71 290 2.45 4 Oil gas Pintsch .056 .70 1000* 9 50 5 Oil gas, Lowe .040 .49 650 7.75 6 Coke oven gas . . .029 .36 545 5.00 7 Producer gas from coke 075 .93 135 1 00 8 Producer gas from anthracite .... .065 .80 145 1.15 9 Producer gas from soft coal .073 .90 145 1.25 10 Blast furnace gas 079 98 100 .70 214 INTERNAL COMBUSTION ENGINES grains <| e P 1L| si II i CO P^ OQ O o "I a a 3 o \ ce \ s 1 *.y XJV - > *^. 2 ^ Ratio of Explosion to Atn X c o e *Cl x Tests by Bun ! sen -o- 1" " Berf helot und^ Heille Mui iird t nd 1C Cliatc ier 1 heoveticnl Kesult ests by Laiigen 1 1 1 3 4 Ratio of Inert Gases to Fuel Gas; (CO.) by Volume = /t FIG. 10-2. \ re to Atm. o -i Bunscn Berthelot Mallard and le Chute | 3 I' 1. Theoretical Result! 2. Tests by JUangen 234 00 Ratio of Inert Gases to Fuel Gas.^J ) by Volume = 'in FIG. 10-3. THE FUEL MIXTURE 223 Langen, in analyzing the results shown in the above diagrams, makes the following observations: 1. The computation of explosion pressures on the assumption of constant specific heat and complete combustion furnishes values which considerably exceed those actually observed. 2. For equal ratios of inert diatomic gases to fuel gases, the kind of inert gas used seems to have no influence upon the explo- sion-pressure, as far as the same observer is concerned. This would lead to the conclusion that the molecular heats of the so-called simple or diatomic gases are equal to each other, at least up to 4500 degrees Fahrenheit. Regarding the results of his own experiments, Langen is of the opinion that in fuel mixtures containing CO, dissociation of CO 2 sets in when the temperature exceeds 1900 degrees Centigrade (3450 degrees Fahrenheit). He bases his opinion on the abnor- mal position of the cooling curve as observed on the diagrams taken for such mixtures. And since the amount of this dissocia- tion is indeterminate, no definite equation expressing the relation of maximum pressure to initial pressure can be established, at least for temperatures exceeding 3400. For hydrogen mixtures, on the other hand, the cooling curves on the diagrams are always normal. Hence the dissociation limit for steam does not seem to have been reached even with the strongest mixtures. From that part of his experiments for which complete com- bustion can be assumed, Langen derives equations for mean molecular heat of diatomic gases, and for carbon dioxide and steam. The temperature limits for the field so covered are not very wide, 1500-1700 degrees Centigrade (2730-3100 degrees Fahrenheit), and further it was assumed that the molecular heat is a linear function of the temperature. Transposed to read mean specific instead of mean molecular heats, these equations are as follows : For C0 2 , C v = .152 + .0000591 t. For H 2 O, C v = .328 + .000119 t. For N, C v = .171 + .0000215 *. For O, C v = .150 + .0000188 t. where t is in degrees Centigrade. The formula of Mallard and Le Chatelier agree with the 224 INTERNAL COMBUSTION ENGINES above as regards the diatomic gases, O and N. For CO 2 and H 2 O these observers obtained results which gave the following relations : For CO 2 , C v = .143 + .0000834 t. For H 2 O, C v = .312 + .000182 t. These formulae show a somewhat more rapid increase of C v with temperature than do those of Langen. Langen observes in explanation of this discrepancy that Mallard and Le Chatelier's formula for CO 2 is obtained from results for which the tempera- tures were from 1700 to 2000 degrees Centigrade, and that the formula for H 2 O is based on experiments for which the tempera- tures were very much higher than for his experiments. In the former case dissociation was shown to be more than likely, in the latter the formula gives results which do not seem to apply very closely for the important temperature range between 2250 and 4000 degrees Fahrenheit. It is plain, therefore, that Langen's formulae promise greater accuracy. The second important investigation in this field was made by Clerk, and by him reported to the Royal Society. His method of operation is so decidedly different from that of Langen and the earlier experimenters that it becomes both interesting and im- portant to see how far his results agree with those already men- tioned. The method of experiment is best described in Clerk's own words : "It consists in subjecting the whole of the highly heated products of the combustion of a gaseous charge to alternate com- pression and expansion within the entire cylinder while cooling proceeds, and observing by the indicator the successive pressure and temperature-falls from revolution to revolution, together with the temperature and pressure rise and fall due to alternate compression and expansion. The engine is set to run at any given speed, and at the desired moment after the charge of gas and air has been drawn in, compressed, and ignited, the exhaust valve and charge inlet valves are prevented from opening, so that when the piston reaches the termination of its power stroke, the ex- haust gases are retained within the cylinder, and the piston com- presses them to the minimum volume, expands them again to THE FUEL MIXTURE 225 the maximum volume, and so compresses and expands during the desired number of strokes." To attempt to explain the method of evaluating the expansion and compression lines so obtained would lead too far for the scope of this book. The reader is referred to the original article.* The engine operated with coal gas. The average composition of the working fluid, as calculated from the analysis of the gas, was H 2 O (assumed gaseous), 11.9 per cent by volume; CO 2 , 5.2 per cent; O, 7.9 per cent, and N, 75 per cent. The mixture as actually used varied somewhat from this composition, but since the percentage of N is nearly constant, this variation can have but small effect upon any specific heat calculations. For this mixture, Mr. Clerk, on the basis of his experiments, found the following mean specific heats, expressed in foot pounds per cubic foot of working fluid at 760 mm. and C. Range of Temperature C F Mean specific heat in ft. Ibs. per cu. ft. at 760 mm and C. 0- 100 32- 212 20.3 0- 200 32- 392 20.9 0- 400 32- 752 21.9 0- 600 32- 1132 22.8 0- 800 32- 1472 23-6 0-1000 32- 1832 24-6 0-1200 32- 2192 24.6 0-1400 0-1500 32- 32- 2552 2732 25.0 25.2 Now to compare these results with those of Mallard and Le Chatelier and of Langen, the easiest way would be to reduce them to the ordinary specific heat basis, and then to compute a series of specific heats for the same temperature ranges and for the same mixture as used by Clerk by the aid of Mallard's and of Langen's formulae. This involves the assumption that these formulas hold for the lower temperature ranges. The following table shows the figures so obtained, and Fig. 10-4 gives a graphi- cal comparison. * See foot-note, page 221. 226 INTERNAL COMBUSTION ENGINES MEAN SPECIFIC HEAT FOR MIXTURE CONTAINING BY VOLUME 11.9 % H 2 O, 5.2 % CO 2 , 7.9 % O, and 75 % N. Temperature Range C F Mallard & Le Chatelier Langen Clerk 0- 100 32- 212 .1805 .1826 .1854 0- 200 32- 392 .1843 .1858 .1910 0- 400 32- 752 .1930 .1922 ,2000 0- 600 32-1132 .2006 .1985 .2083 0- 800 32-1472 .2083 .2047 .2157 0-1000 32-1832 .2161 .2112 .2202 0-1200 32-2192 .2238 .2176 .2248 0-1400 32-2552 .2315 .2239 .2284 0-1500 32-2732 .2355 .2271 .2303 0-150 0-140 0,130 )C / / / '/ / \ ariat f-a-A OI1S ( ixtu I'M i re-ro ll.!>5 5.2? 7JJJ 7.-I.O an S rtain tff,C "Till mg-b c He yVo. t / // 7 A ^ 0-110 o 2 g f // / / / / /, u /J rg /e 0-701 0-60C 0-50(1 0-40C 0-300 0-200 z & / $ / ^ F / / / / / / ^ > / y/ / // / . 7 .1 s V 9 lines .2 o,C. .: \ ! .1 FIG. 10-4. It is .plain from Fig. 10-4 that the question of the variation of specific heat with temperature cannot be considered entirely solved. It is true that Stevens * made experiments on air which check the results of both Mallard and Le Chatelier and of Langen * Ann. d. Phys., 1902. THE FUEL MIXTURE 227 very closely. It is therefore fair to assume that the formulae derived for diatomic gases are correct. The discrepancies ob- served between the results plotted in Fig. 10-4 are therefore with strong probability due to error in the available data for H 2 O and CO 2 . Further work is therefore required, and a check of Clerk's results is especially desirable. 3. Velocity of Flame Propagation and Time of Explo- sion. Experiments on the velocity of flame propagation in a given mixture, like the experiments on specific heat, have not led to any definite result. And although definite information on this point is desirable, on account of the connection between velocity of flame propagation and possible maximum engine speed, in any given case there are so many factors affecting the problem in actual practice, that the application of the results of laboratory experiments to actual conditions is of doubtful value. Thus the velocity with which the flame spreads through a mixture depends upon the kind of fuel, the composition and purity of the mixture, its temperature and pressure, and upon the location of the igniter and the shape of the combustion chamber. Further than this, the degree of mechanical agitation in the mixture at the moment of explosion has a marked influence upon the velocity. If we ignite a mixture in a tube, closed at one end, from the closed end, the pressure generated seems to project the flame ahead of the pressure wave toward the open end of the tube with a much higher flame velocity than would have been observed if the ignition had taken place from the open end. The same effect should be observed in a tube closed at both ends. Ignition from the open end gives the true velocity, as the flame then spreads by contact only. In an actual engine cylinder, with the volume in- creasing with the movement of the piston, we may expect the velocity of propagation to be somewhere between those found for ignition from the open end of a tube and those for ignition in a closed tube. Mallard and Le Chatelier used the open-tube method, measur- ing the time interval between the passage of the pressure wave between two points on the tube. Their results, as given by Clerk, for hydrogen were as follows: 228 INTERNAL COMBUSTION ENGINES Mixture Velocity of Pressure Propagation ft. per second 1 vol H -f- 4 vols air 6 56 1 vol H + 3 vols air 9 20 1 vol H + 2J vols air 11 10 1 vol H -f- 1 f vols air 12 40 1 vol H + l| vols air 14 30 1 vol H -(- 1 vol air . 12 30 1 vol H + ^ vol air 7 55 The true explosive mixture for hydrogen and air is 1 vol. of H to 2 vols. of air. It might be supposed that this would be the mixture showing highest velocity of propagation. For some unexplained reason a certain excess of hydrogen shows the highest 51)01 -300- osed e 2 3 4 5 6 7 b 9 10 11 12 13 Ratio Air to Gas by Vol. FIG. 10-5. result. A similar phenomenon was observed with coal gas and air mixtures by these observers. Meyer * used an apparatus very similar to that of Mallard and Le Chatelier, except that very sensitive platinum thermometers were employed to measure the passage of the flame, instead of diaphragms to measure the passage of the pressure wave. The fuel employed was coal gas. All tests were made with the mix- * E. J. Meyer, Sihley College thesis, 1905. THE FUEL MIXTURE 229 ture at atmospheric pressure at the time of ignition. Figs. 10-5 and 10-6 show the results graphically, the former being obtained for ignition from the closed end, the latter for ignition from the open end. The much higher velocity of the first curve bears out the statement previously made. The fact that the maximum velocity occurs with the same gas-air ratio shows that the true rate of inflammability has not been changed, but that mechanical actions alone are responsible for the difference in the observed results. Closely connected with the velocity of flame propagation, and subject to the influence of accidental accompanying conditions to the same degree, is the time of explosion. The most extensive work in this field was done by Clerk. The apparatus employed by him was very similar to that used by Langen in his specific / Vi ,-er, Ignition from jO-fL^ / \ opcr t end x o / / \ 1 Z / Clerk, X Expl. l_v.ess< rtoG 1 g ^-^^ 1 y Ei itio A 1S byj ^v, .0,. ^: i ===^ n i i 2 1 FIG. 10-6. heat experiments. The fuel mixture was ignited at constant volume and a pressure diagram obtained on the rotating drum of an indicator. Unfortunately the experiments were confined to coal gas, a few figures only being obtained for hydrogen and. none for the power gases so important to-day. The figures found for hydrogen were as follows, the time of explosion being the time interval from the moment of ignition to the attainment of maxi- mum pressure.* Mixture by Volume Time of Explosion in seconds Air Hydrogen 6 1 .150 4 1 .026 2.5 1 .010 * Clerk, The Gas and Oil Engine, p. 101. 230 INTERNAL COMBUSTION ENGINES Clerk's results for mixtures of air and Oldham coal gas are represented in Fig. 10-7. The volume ration showing fastest time of explosion, i.e., 6 to 1, agrees with the mixture for which Meyer found the greatest velocity of flame propagation. This mixture also showed about the highest pressure development in Clerk's experiments, 90 pounds per square inch. ^ ^~- ^ ^"^ / / > inn / o / Cle k,xi v ne of Expl. ssel, (JHdliai n closed iCoa/Gas. 9 , which on its previous suction stroke has drawn a mixture of air and gas into B, has completed about half of its in stroke and 256 INTERNAL COMBUSTION ENGINES displaced the mixture into G and A through the connecting pipe, W. About the time D has completed its in stroke, the power piston, C, has covered the exhaust ports on its return stroke, arid compression ensues in the main cylinder. The cylinder volumes are so proportioned that in theory no mixture can be lost through the exhaust ports. At the inner dead center of the power piston, C, or just before, the igniting cavity, 0, Fig. 11-23, comes oppo- site the port, N, and the charge is fired. The piston is impelled forward and the next charge, drawn into B in the meantime, com- mences to enter the space G as soon as the pressure in A has fallen enough, after the beginning of exhaust, to cause the valve in the pipe between B and G to lift. The intermediate valve FIG. 11-23. Horizontal Section, Clerk Engine. arrangement is shown in Fig. 11-22. P is an air chamber. The air drawn in by the suction of the displacer piston, 7J, passes through the valve, H, and in so doing is mixed with gas which enters through a number of fine holes ih the seat of the valve from the annular space, K. On the in stroke of the displacer piston the mixture under some pressure comes in under the valve, F, which it lifts as soon as the pressure of the exhaust gases above it has fallen low enough. The difficulty under which the engine labored was loss of mixture through the exhaust ports and consequent low economy. Although the respective volumes of the two cylinders may be in the proper ratio, the charge expands by heat on the transfer, and mechanical agitation favors the loss. In spite of this defect shop tests on 2, 4, 6, 8, and 12 horse-power engines in 1885 gave results fully equal to those obtained on the four-cycle machine HISTORY OF INTERNAL COMBUSTION ENGINE 257 of that day. The following table shows the results of some of these tests. Horse-power 2 4 6 8 12 Dia. Motor Cyl., inches . . . Stroke Motor Cyl., inches. Dia., Displacer, inches. . . . Stroke Displacer. inches . . R. P. M. . . . 5 8 6 9 212 6 10 7 11 190 7 12 7| 12 14fi 8 16 10 13 14.9 9 20 10 20 1 Q9 I. H. P D. H. P Mech. Eff.,% GasperB.H. P.-hr., cu. ft. 3.62 2.70 74.7 40.0 8.68 5.63 65.0 37.3 9.05 7.23 80.0 30.42 17.38 13.69 78.8 26.58 27.46 23.21 84.5 24.12 Figures 11-24 and 11-25 show a power diagram and a pump diagram from a 6 horse-power engine. 4. Period of Application. It is intended in what follows to give merely a brief resume of the further development of the in- ternal combustion engine say up to 1897 and to reserve a more detailed description of the most important forms in the market to-day for the next chapter. On account of the Otto '' 1 ^ ~ Power Card, Clerk Engine. patent, which practically amounted to a monopoly, other manu- facturers were forced to turn their attention to the development of the two-cycle en- gine. How Clerk solved the problem with fair success has been already shown. He was fol- lowed by Wittig and Hess (1880), Benz (1884), Sohnlein, Giildner (1893-1898), Oechel- hauser (1896) and Koerting (1898) in Germany, Robson, Southal, and Samson in England, Benier (1894) in France, and Mietz and Weiss in the United States. Many of these engines are in the market to-day, and some are described in the next chapter. With the fall of Otto's claims in Germany about 1885 the field FIG. 11-25. Pump Card, Clerk Engine. 258 INTERNAL COMBUSTION ENGINES became free, and many manufacturers of the two-cycle engine abandoned it for the more simple four-cycle. The fall of this patent was, in a sense, a misfortune as far as the two-cycle engine was concerned, as it held back the development of that type of machine at least ten years. It is only within the last six or eight years that the very obvious advantages of the two-cycle principle again received the attention they deserve. On the other hand, the development of the four-cycle engine after 1885 was extremely rapid. Thus while the limit of power was about 4 B. H. P. in 1878 and units of from 15-20 horse-power could be had in 1880, the limit soon rose to 100 horse-power in 1889 and 200 in 1893. Blast furnace gas called for units of 600 horse-power in 1898, while to-day engines developing up to 4000 horse-power are being built. With the increase in size up to the neighborhood of 200 horse-power, there comes an increase in thermal efficiency. Thus a Crossley engine of 12 horse-power soon showed a gas consump- tion of 24.3 cu. ft. of illuminating gas per B. H. P. hour. To-day efficiencies exceeding 25 per cent on the brake, with lean power gases, are not rare, and Giildner, by intelligently applying sound principles of construction, has succeeded in obtaining economic efficiencies exceeding 30 per cent. It would lead too far for the scope of this book to describe even a fair percentage of the various four-cycle engines brought out between 1885 and 1898. The most prominent names con- nected with the development of these engines are perhaps Loutzki in Germany (1888), Delamare-Debouteville & Malandin, who in 1900 brought out the first large blast furnace gas engine, in Belgium, Charon in France, Crossley Brothers in England (1892 and 1898), and Westinghouse in the United States (1896). Four- cycle const ant -pressure oil engines were developed by Capitaine (1889-91), Brunnler (1893-4), and Diesel (1893-97) in Germany. The Diesel engine is to-day one of our most important internal combustion engines, and is considered in greater detail below. Among the delevopers of the four-cycle constant volume oil engine, Daimler's work in connection with the high-speed engine deserves special mention. Other engines of this type brought out in this period are Spiel (1884), Capitaine (1885-90), Priest- man (1889), Hornsby-Akroyd (1892), Banki (1894), and Hasel- wander (1898). HISTORY OF INTERNAL COMBUSTION ENGINE 259 One of the greatest achievements of this period is the develop- ment of the Diesel engine. DIESEL. The history of the Diesel engine is interesting. It began in 1893 when Rudolf Diesel, in a pamphlet entitled "Theory and Construction of a Rational Heat Motor to replace the Steam Engine and other existing Heat Engines/' laid down the following " fundamental requirements for a perfect com- bustion." 1. Attainment of the highest temperature in the cycle, not by means of combustion and during the same, but before and inde- pendent of it by compression of air alone. 2. Gradual injection of atomized fuel into this highly com- pressed and heated air so that during combustion no rise of temperature takes place, i.e., the combustion shall be isothermal. For this purpose the process of combustion cannot after ignition be left to itself, but must be governed from the outside to main- tain proper relation between pressure, volume, and temperature. 3. Correct choice of weight of air with reference to the heat- ing value of the fuel and the desired compression temperature, so that the practical operation of the machine, lubrication, etc., shall be possible without water-cooling. It is interesting to follow out these points and to. see in how far their object has been attained. The intended fuel was coal dust, the cycle the Carnot. At the very outset, however, a modification was made in the cycle in cutting out the isothermal compression and substituting for it one stage adiabatic compression. But a jacket was not thought necessar-y, and in fact a non-conducting lining for the cylinder was demanded. As a consequence of the above pamphlet, two firms, Krupp in Essen and the Maschinen-fabrik Augsburg, undertook the con- struction of experimental machines. As was to be expected, further changes from the original idea were necessary, the two most important of which were the substitution of oil for coal dust, and the use of a water jacket. In 1898 the experimental stage had been so far passed that Schroter could report test figures which more than doubled the thermal efficiency of the then existing Otto engines. The final form of the engine as now constructed is shown in Figs. 11-26 to 260 INTERNAL COMBUSTION ENGINE^ 11-28, which represents a Diesel engine built by Krupp in 1898.* The internal-cylinder construction offers nothing new; a is the suction valve for air, b the fuel valve, and d the exhaust valve. c is the starting valve not in commission during ordinary opera- tions. All are actuated through levers by cams on the shaft, /, which is operated from the crank shaft through the intermediate shaft, e', h is the air pump to furnish compressed air for fuel injection and for starting. Both of these are taken from steel FIGS. 11-26 to 11-28. Diesel Engine, 1898. flasks into which h delivers; i is the oil pump under control of the governor which regulates the amount of oil per stroke to the load. On the first down stroke, the engine takes air through a and compresses it on the return stroke to a pressure of about 460 pounds, with a temperature of about 1100 degrees Fahrenheit. Just before the end of the up stroke the fuel valve, b, is opened to a width of only a few hundredths of an inch, and the injection air, previously compressed to about 650 pounds by pump h, flows * Guldner, p. 101. HISTORY OF INTERNAL COMBUSTION ENGINE 261 into the compression chamber, carrying with it and finely atomizing the oil furnished by the pump, i. The oil ignites on entering, due to the high compression temperature. In spite of the fact that isothermal combustion is intended, the lack of outside control causes a rise not only in the pressure of from 80-100 pounds, usually not very noticeable on the indicator card, but also a rise of temperature approximating 1800 degrees Fahrenheit. Thus while the intended mean temperature of the cycle of Diesel's pamphlet was about 350 degrees, that realized in actual operation is about 950 degrees Fahrenheit. The time during which the fuel injection valve, b, remains open, is constant -for all loads, but the effective stroke of the fuel pump, i, ends the sooner the lower the load, thus effecting regulation. After the closing of b, expansion commences and is followed by exhaust through d on the return stroke. The engine is started by the compressed air furnished by h during a previous operation and stored in a tank. To start, the cams on the shaft, /, are pulled to the right by the lever, g. This puts the valves a, b and d out of commission and starting valve c in commission. On placing the crank just beyond the upper center, and opening the tank valve, the engine takes compressed air for a few turns just like a steam engine takes steam. When the required momentum has been obtained, the cams are re- leased and are snapped back into place by a spring at the proper time. The history of the development of the Diesel engine is interest- ing in that the final construction departs so far from the patented ideal that the engine of to-day does not seem to be protected by the claims of that patent. The fact that no igniter was neces- sary was only incidentally considered by Diesel, and is not ex- pressly covered by the patent. The first test figures on a Diese! engine were published by Schroter in 1897. The dimensions of the engine were: Cylinder diameter, 9.85", stroke 15.7", rated B. H. P., 18-20. The fuel used was American kerosene having a heating value of 18400 B.T.U. The following table shows the results of two full load trials and compares them with some of the theoretical results figured by Diesel in his pamphlet of 1893.* *Guldner, p. 107. 262 INTERNAL COMBUSTION ENGINES Diesel Engine of 1897 Ideal Engine of 1893 Load ' Full Full R p M 171.8 154.2 300 B H P 19.87 17.82 ? Total I H P 27 85 24.77 100 Pump HP ...... 1.29 1.17 Net I H P . . 26.56 23.69 100 Oil per B. H. P.-hr., Ibs Cooling water pr B . H . P.-hr. , Ibs. Compression Press., Atm Max. Comb. Pressure, Atm Temp. End of Compression, C . . Temp. End of Combustion Indicated Thermal Eff ., % Mech Eff % .543 .523 154 203 32.5 32.5 36 36 550 550 1600 1600 33.7 34.7 71.0 72 .247 Ibs. of coal per I. H. P. Zero 250 250 800 800 73 Thermal Eff. on Brake, % 25.2 26.2 ? It is clear from the above table that, although the engine did not realize the early expectation of its designer, the results shown are a remarkable step in advance as regards the economy of the then ex- isting internal combustion engines. To-day even bet- ter figures are frequently obtained. E. Meyer, for instance, has lately tested FIG. 11-29. -Regulation Diagram, Diesel a DieSel en S ine showing Engine. an indicated thermal effi- ciency exceeding 42 per cent. Fig. 11-29* shows the general form of the indicator card obtained from Diesel engine. This is a regulating diagram, the size of the card decreasing with the load by shortening the cut-off. * Giildner, p. 109. CHAPTER XII MODERN TYPES OF INTERNAL COMBUSTION ENGINES THE previous chapter brought the development of the internal combustion engine up to 1897. Since then expansion has been very rapid, until to-day we find that the design of gas engines has been standardized in the most important particulars just as was the case with the steam engine some decades ago. It is in- tended in the present chapter to give a brief description of the most important engines found in the market to-day. An ex- amination of the market will show some fairly definite divisions among manufacturers as far as size of engine made is concerned. Thus there are but a half dozen firms in this country, and a few more than this in Europe, who make engines up to the very largest sizes. It is comparatively easy, therefore, to describe nearly all of these various engines, and this is very desirable on account of their importance. Next we find a somewhat larger number of medium sized engines of various types, and lastly a very large number of engines up to say 50 horse-power serving the general commercial field for small powers. It is of course impossible to cover the last two classes of engines to any great extent. The majority of these engines' do not differ except in minor details, and for these reasons the following list has been confined to what seem to be the most representative machines of each class. Regarding the general features of design, small engines are either horizontal or vertical. The cylinders are almost invariably single-acting, multiplication of power being obtained by increas- ing the number of cylinders. The vertical offers some advan- tages over the horizontal form, in that the foundation need not be as large or as heavy. Further it is claimed that it is easier to lubricate the cylinders uniformly, and that the wear on the cylinder is less. A favorite form of frame for this type of engine in this country is the box frame with enclosed crank case, using splash lubrication. Some European designers object to this 263 264 INTERNAL COMBUSTION ENGINES form, claiming that all supervision of crank pins and intermediate bearings is by this form of frame rendered impossible. The small two-cycle machine almost invariably uses the enclosed crank case for the pre-compression of the mixture. What has been said of the small machine applies in general also to medium sized engines. Vertical machines here possess the added advantage that it is easier to dismount them by means of overhead crane than is tho case with horizontal machines. The limit to a vertical engine comes in the head room required. For this reason all of the very large machines, as well as medium sized double-acting machines which require a cross-head, are horizontal. Another reason that may be cited is that it is easier to operate a medium sized or large horizontal engine than it is a vertical because all climbing or mounting platforms is avoided, and the whole installation is more completely under the operator's eye. Finally, the use of some of the industrial power gases favors the use of the horizontal machine, because any dust carried can be much more easily swept out of a horizontal than a vertical cylinder during regular operation. The double-acting engine is perhaps not used as widely as it deserves to be, increase in power being generally sought by multi- plying the single-acting cylinders. There is, however, to-day no reason why double-acting cylinders are not as reliable as the single-acting. For large machines, double-acting cylinders are almost an economic necessity, and the clumsy four-cylinder double-opposed large engine has become thoroughly obsolete. The largest engines of to-day are double or twin two-cylinder tandem double-acting engines. The very obvious disadvantages of the trunk piston can be quite successfully overcome for small and medium sized machines, and hence it is almost universally employed for these sizes. But for large 'machines the use of the trunk piston is indefensible, being opposed alike by considerations of manufacture and of reliable operation. The fitting of large pistons of this type offers grave difficulties in the shop, and in operation proper lubrication is difficult; further, the main office of the piston is to confine the gases without leakage; to make it also act as the machine member to take up the lateral thrust of the connecting rod may, in large machines, seriously interfere with its main purpose. MODERN TYPES OF COMBUSTION ENGINES 265 The crank shaft of small and medium sized engines is nearly always of the center-crank type. This type is very rigid and, above all, transmits the stresses equally to both sides of the frame. In double or twin machines, however, such a shaft would call for four main bearings, the proper alignment of which might cause some trouble in large engines of this type. For this reason, some American makers prefer the side-crank shaft, which is a much less costly shaft to make, and reduces the number of bearings for a twin engine to two. Of course, the side-crank frame takes up the explosion stresses eccentrically, and therefore has to be designed heavier than the center-crunk frame. On account of this fact and the generally higher stresses in gas engine frames as compared with steam-engine frames, Fluropean designers will not use the so-called side crank or Tangye form of frame. The period of experimentation in design, and of freak design, has largely passed in gas-engine practice, and, as mentioned at the outset of this chapter, standardization has made welcome progress during the last few years. The alcohol engine, now in its development as far as the United States is concerned, does not call for any radical changes in the existing designs of 'liquid fuel engines. As far as further progress is concerned, it is not unlikely that the next step will be the development of a highly efficient constant-pressure gas engine after the manner of the Diesel liquid fuel engines, taking up the thread again where Brayton left it in the seventies. A. Gas Engines 1. SMALL AND MEDIUM SIZED ENGINP:S. Among the best known makers of small and medium sized gas engines in this country may be mentioned the Otto Gas Engine Works of Phila- delphia, makers of the Otto engines; the Fairbanks-Morse Com- pany; the Jacobson Machine Manufacturing Company, of Warren, Pa.; the Jacobson Engine Company, of Chester, Pa., the Struthers- Wells Company of Warren, Pa., makers of the Warren engine; the Bruce-Meriam Abbott Company, of Cleveland; the Westing- house Machine Company; the Olds Gas Power Company, of Lansing, Mich.; the De LaVergne Machine Company of New York, makers of the four-cycle Koerting engines; the Weber Gas Engine Company of Kansas City, Mo.; the A, H, Alberger Company of 266 INTERNAL COMBUSTION ENGINES Buffalo, makers of the Buffalo tandem engines, the Foos Gas Engine Company of Springfield, Ohio, etc. In most cases the small gas engine operates on illuminating gas, natural gas, or gasoline, but in many instances attachments are furnished so that the engine can be run on either gas or liquid fuel as desired. Producer gas is not usually employed for very FIG. 12-1. Westinghouse Engine. small units. A 10 to 15 horse-power suction gas producer for hard coal is the present lower limit and perhaps the exception, while soft coal producers in their present state of development do not run less than 50 to 60 horse-power. The Westinghouse Gas Engine. The cross-sectional cut, Fig. 12-1, shows the essential parts of the Westinghouse vertical engine. In this type the crank mechanism is completely enclosed, and splash lubrication is depended upon for the proper oiling of MODERN TYPES OF COMBUSTION ENGINES 267 the intermediate bearings, the crank pins and the pistons. Both the inlet and exhaust valves are mechanically operated by cams and shafts driven from the main shaft; the inlet valve, J, by means of cam B and lever (7, the exhaust valve E by cam A and the roller lever shown. The igniter, operated from the inlet cam shaft, is located at F. Gas and air, after passing the chamber M in the proper proportion, enter the passage N on their way to the inlet valves. The engine is governed by means of a governor of the fly-ball type which regulates the amount of mixture enter- ing the passage N (see Chapter XIV). The smaller sizes of this FIG. 12-2. Westinghouse Engine. machine have two cylinders, and can generally be started by hand; the larger sizes, above about 85 horse-power, have three cylinders and are generally started by compressed air, which is admitted to one cylinder, starting the engine, while the other cylinders operate normally. Figs. 12-2 and 12-3 show general views of three-cylinder machines. Engines manufactured under the Jacobson Patents. There are two firms manufacturing engines under these patents, the Jacobson Engine Company of Chester, Pa., and the Jacobson 268 INTERNAL COMBUSTION ENGINES Machine Manufacturing Company of Warren, Pa. The engines made are of three types, hit-and-miss, automatic cut-off, and throttling. The Warren Company make hit-and-miss engines from 2^ to 25 horse-power and automatic engines from 8 to 27 horse-power. The Chester Company make hit-and-miss engines from 30 horse-power up, and automatic cut-off and throttling en- gines from 33 horse-power up. All of these engines so far men- tioned are single-acting. The automatic cut-off and throttling FIG. 12-3. Westinghquse Engine. types are built as single-cylinder or as tandem or twin-tandem units. The Chester Company now also undertake the building of double acting automatic cut-off or throttling engine as tandem or twin-tandem units up to any power desired. The structural features of the engines built by the two com- panies mentioned are of course very nearly the same, so that one description will do for both. The general features of the hit-and-miss machine are shown in Fig. 12-4. The most interesting detail of the design is perhaps the removable cylinder bushing, which is an unusual but highly commendable construction for small engines. This not only allows of choosing the proper grade of metal for the cylinder barrel, but thermal stresses are also avoided by admitting of MODERN TYPES OF COMBUSTION ENGINES 269 free expansion, in this case against the. packing at the head of the bushing. On account of the manner of holding it, this packing can never be blown out. FIG. 12-4. Jacobson Hit-and-Miss Engine. FIG. 12-5. Governing Mechanism of Jacobson Hit-and-Miss Engine. The inlet valve is automatic, and is held in a separate housing. The manner of operating the exhaust valve and the governor control are shown in Fig. 12-5. The side shaft is operated by means of screw gears, which is the most satisfactory way of 270 INTERNAL COMBUSTION ENGINES transmitting the motion. The governor is of the fly-ball type and is operated by the lay shaft. When the speed becomes too high, the blade, C, Fig. 12-5, put in position by the governor, engages the block A on the exhaust valve lever, prevents this lever from returning, thus holding the exhaust valve open. " VENT TO ATMOSPHERE FIG. 12-6. Gas Pressure Regulator for Jacobson Engines. Figure 12-4 also shows the make-and-break igniter. One block contains both electrodes and the location of the spark is central as regards the volume of the charge. This engine can- be run on either gas or gasoline, depending upon whether a gas regulator or a carbureter is employed. The type of regulator used is illustrated in Fig. 12-6. Its construction is very simple, there is nothing apparently to get out of order. It is claimed that this regulator furnishes the gas to the engine MODERN TYPES OF COMBUSTION ENGINES 271 FIG. 12-7. Jacobson Auto- matic Cut-off Engine. under atmospheric pressure at all times. The firm makes an attachment which allows of the change from one fuel to another at a moment's notice. The general design of the automatic cut-off is the same as for the hit-and-miss engine, the differ- ence being in the governor and the valve gear. Fig. 12-7 shows the method of operating the inlet and exhaust valves and the igniter. The valve gear is shown in greater de- tail in Fig. 12-8. The seats and stem bushings are all easily accessi- ble and replaceable. The exhaust valve is operated in the ordinary way by a cam from the lay shaft. For the inlet valve lever an eccen- tric is used. As the eccentric rod R rises, it "pushes upward the valve lever Q at the right end and opens the inlet valve. The inlet valve spindle carries two discs, the main inlet valve and the gas valve. Just before the exhaust valve closes, the gear commences to open the inlet valve, but only air enters since the gas valve V remains closed. The air thus enter- ing serves to scavenge out the combustion chamber. A moment later the valve spindle has descended far enough for the set collar on the spindle to depress the gas valve ; after which the mixture begins to enter the cylinder. Both valves close simul- taneously when the nose of the eccentric rod R is forced off the end of the valve lever by the action of the inclined plane P. The position of this plane is determined by the governor, which is of the fly-ball type and directly driven from the main shaft. Thus the valve always opens at the same point, but it closes sooner or later, depending upon the load. Figure 12-9 shows the general appearance of the automatic cut-off machine. They are built in sizes up to 27 horse-power by the Warren Company, and can be adapted to run on gasoline and natural, illuminating, or producer gas. The twin tandem type is shown in Fig. 12-10 and is built in sizes above 100 B. H. P. for producer gas and above 120 B. H. P. for natural gas by the 272 INTERNAL COMBUSTION ENGINES FIG. 12-8. Valve Gear, Jacobsori Automatic Cut-off Engine. Chester Company. One-half of this unit, the tandem engine, is built in sizes from 60 B. H. P. upward for natural gas and from 50 B. H. P. upward for pro- ducer gas. The Chester Company's throttling engine, a general view of which is shown in Fig. 12-11, differs from the automatic cut-off machine only in some details of valve gear. A cross^-section FIG. 12-9. -Jacobin Automatic Cut-off f this en S ine is shown in Engine. Fig. 12-12, while Fig. 12-13 MODERN TYPES OF COMBUSTION ENGINES 273 FIG. 12-10. Jacobson Twin-tandem Automatic Cut-off Engine. FIG. 12-11. Jacobson Throttling Engine. 274 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 275 FIG. 12-13. Valve Gear, Jacobson Throttling Engine. illustrates the valve gear. Both valves are operated by cams. The charge passes the mixing and throttling valve, which is con- trolled by the governor through the reach rod shown. The Bruce-Meriam-A bbott Engine. The Bruce-Meriam- Abbott Company, located in Cleveland, Ohio, manufactures engines for natural, illuminat- ing, and producer gas and for gasoline. The design used for natural and illuminating gas and for gasoline is shown in Fig. 12-14. The sizes range from 12 to 250 horse-power, two-cylinder up to 125 horse- power, and four-cylinder above that size. The design of cylin- der and frame is along conven- tional lines. The cam shaft , u , FIG. 12-14. Bruce-Meriam-Abbott across the top of and between Engine 276 INTERNAL COMBUSTION ENGINES FIG. 12-15. Detail of Bruce-Mer- iam-Abbott Engine. the cylinders is operated through spur and bevel gears as shown. The valves are of the poppet type and are located in the head. Above 55 horse-power they are held in separate cages which are easily removable. The cam shaft operates these valves by rocker arms on either side. This construction is more clearly shown in Fig. 12-15. An excellent feature of the machine is the purely cy- lindrical form of the combustion chamber. Governing is effected by a gov- ernor of the fly-ball type operated by the lay shaft. The governor sleeve operates one end of a lever passing between the cylinders, the other end supports the mix- ing valve, the'details of which are shown in Fig. 12-16. Gas and air enter the annular space shown. On the suction stroke of the engine the gas flows into the space surrounding the piston valve and mixes with the air by flowing out through the port about half way up. The two combined then enter the interior of the valve through the six ports shown; from this chamber the mixture goes to the engine. The position of the valve controls the- amount of throttling and thus regulates the weight of the charge going to the engine. It is stated that from full to no load this valve has to move only T^ inch. Jump spark ignition is used. The method of supporting the spark coil and the system of wiring is well shown in Fig. 12-15. The timer is very simple. There are two copper pins each about J-inch di- ameter and f inch long. One of them projects from the bottom of a small cup while the other is fastened to the end of a flat spring, and dips down in the same cup. The spring may be seen at A, Fig. 12-15. The cup is filled with oil and the points are normally held apart a small distance. Just before a spark is desired the spring is FIG. 12-16. Mixing Valve, Bruce-Mer- iam- Abbott Engine. MODERN TYPES OF COMBUSTION ENGINES 277 depressed by means of a cam, the circuit is made by bringing the points in contact and the spark is produced at the moment the cam releases the spring. What amounts to an outside spark gap (see next chapter) is provided so that the spark may be watched. To convert any gas engine of this design into a gasoline engine it is necessary merely to furnish a fuel pump and to replace the iron piston valve of the mixer by brass or bronze to prevent rust- ing. To retain the high compression used with gas, however, this company has adopted the Banki principle of injecting water into the cylinder. The device is said to give entire satisfaction. For producer and natural gas this firm makes engines ranging from 25 to 200 horse-power, two-cylinder units up to 90 horse- power, and four-cylinder above that. The design of these engines is apparently somewhat different from those above described, Fig. 12-17, the principal change being that the lay shaft evidently runs alongside the cylinders instead of between them. The cylin- der and frame construction is otherwise the same, but the interest- ing feature about the design is the fact that the center line of the cylinder is offset about one-half the length of the crank from the center line of the main bearing, as near as can be scaled from the drawing. This is on the principle of the Ramsey crank mechanism, the idea of which is to equalize the wear due to the side thrust of the piston against the cylinder on both sides of the cylinder and to improve the turning moment. Both of these aims are attained with a moderate offset such as here used. The Fairbanks, Morse & Company Engines. The Fairbanks, Morse & Company manufacture a number of different types of engines, both vertical and horizontal, for gas and liquid fuel. The general features of the horizontal design are shown in Fig. 12-18. It appears from this that the exhaust valve, at the side of the cylinder, is mechanically operated, while the inlet valve placed in the head is automatic. The governor is placed in the fly-wheel and may be either of the throttling or hit-and-miss type. In the latter case it operates to hold the exhaust valve open. The ignition gear is of the make-and-break type, arranged so that at starting the spark may be retarded. Engines above 9 or 10 horse-power are fitted with a self-start- ing device which consists of two parts, a match detonator and a hand pump. Both are shown in plan in Fig. 12-18 at the s;de 278 INTERNAL COMBUSTION ENGINES of the cylinder. The hand pump serves to pump a combustible mixture into the cylinder, with the crank just beyond the center, which mixture is fired by igniting a parlor match inserted into the detonator. The pressure so generated is sufficient to start the engine under some load. FIG. 12-17. Bruce-Meriam-Abbott Engine for Producer Gas. Engines are built for gasoline, naphtha and distillate, for kerosene, for alcohol, and for gas. In each case, of course, the fuel feeding and mixing arrangements differ somewhat. In the liquid fuel engines the feeding device acts positively, a pump being used to ^ inject the proper quantity of fuel into the charge of air. In MODERN TYPES OF COMBUSTION ENGINES 279 the gas engine there is instead a mechanically operated gas valve. If desired the engine may be arranged to run on either gas or FIG. 12-18. Fairbanks-Morse Engine. liquid fuel. Fig. 12-19 illustrates an engine having these features. The same company also builds two distinct types of producer gas engines. The first of these has the general features of the standard horizontal engine above described, the second is quite different, as shown in Fig. 12-20. In this design both valves are of the vertical poppet type, mechanically operated from the side shaft, the inlet valve on top, the exhaust at the bottom. As in the other Fairbanks en- gines, the ignition is by make and break. The governor is of the fly-ball type, operated from the lay shaft. Regula- tion is by throttling a mixture of constant proportion. The general type of Fair- banks vertical engine is shown in Fig. 12-21. These engines FlQ 12 _ 19 .- Fairbanks _ M orse Engine are built for any kind of fuel. f or Liquid or Gas Fuel. 280 INTERNAL COMBUSTION ENGINES Governing is effected by throttling the mixture; make-and-break ignition is used. Although the crank case is enclosed, lubrication is positive instead of by the splash method usually em- ployed in this design. The Koerting Four-cycle Gas Engine. This German machine is manufactured in this country by the De La- Vergne Machine Company of New York. The general fea- tures of the design are clearly shown in elevation, Fig. 12-22, and the two cross-sec- tions, Figs. 12-23 and 12-24. FIG. 12-20. Fairbanks-Morse Producer The entire design gives the impression of being very sub- stantial and thorough. The frame is a very rigid construction and the cylinder is supported by the frame throughout the length. The cylinder head is a somewhat complicated casting. The inlet and out- let valves are placed verti- cally over each other and are operated by cams from a lay shaft. The combustion chamber is divided by a water-cooled tongue project- ing from the cylinder head. The purpose of this projec- tion is to effectually cool the interior of the combustion chamber and to thus draw down the compression tem- perature, admitting of higher compression. The mixing FIG. 12-21. Fairbanks-Morse Vertical valve, shown at the left of Engine, the "transverse section in Fig. 12-24, is automatic. Regulation is effected by means of a governor of the Hartung type which MODERN TYPES OF COMBUSTION ENGINES 281 it -S 282 INTERNAL COMBUSTION ENGINES operates a butterfly throttle valve in the admission passage, as shown. The speed is thus controlled by throttling. In engines exceeding 100 horse-power the pistons are water-cooled. All engines above 12 horse-power have electric igniters while those below this power use the hot tube, at least as constructed by the parent firm. The Buffalo Tandem Engine. This engine is made by the A. H. Alberger Company of Buffalo. A full description will be MODERN TYPES OF COMBUSTION ENGINES 283 found in an article in Power for February, 1907, from which the following illustrations are taken. Fig. 12-25 shows a general .view, Fig. 12-26 a vertical cross-section, and Fig. 12-27 two cross-sections of the valve chest at right angles to each other. FIG. 12-24. Valve Gear, Koerting Four-cycle Engine. The engine differs materially from those previously described in having two cylinders in tandem, each acting on the four-cycle principle. The back piston is made like an ordinary steam engine 284 INTERNAL COMBUSTION ENGINES piston since there is no side thrust in this cylinder. Its piston rod passes through a water-cooled stuffing-box as shown. The back cylinder head is a simple flat plate not water-cooled; all h. FIG. 12-25. Buffalo Tandem Engine. parts requiring it, however, are thoroughly cooled. The valves are of the double-guide poppet type and are placed side by side in a valve chest at the side of the cylinder, Fig. 12-27. Cams FIG. 12-26. Section Through Cylinders and Pistons of Buffalo Gas Engine- on a lay shaft operate these valves, as shown in the transverse section, Fig. 12-27. The make-and-break igniter, which is ad- justable during operation, is placed over the inlet valve at the MODERN TYPES OF COMBUSTION ENGINES 285 right side of the valve chest. The exhaust pipe is fastened at the left side as shown in the longitudinal section of the valve FIG. 12-27. -*- Valve Details of Buffalo Engine. chest. The jacket water, after passing through the jacket, is made to enter the exhaust pipe, cooling the gases and thus acting as a muffler. The mixing and governing arrangements of this The Valve Cage FIG. 12-28. Mixing and Governing Arrangements Buffalo Engine. engine are shown in Fig. 12-28. The two inlet valve chambers are connected by a header, as shown in Fig. 12-25. At the center 286 INTERNAL COMBUSTION ENGINES this header carries the mixing valve. This valve, Fig. 12-28, is a hollow cylinder divided into two parts by a transverse partition. Each half has a number of slotted ports which in certain positions of the valve register with other similar ports in the cage. The valve has two offices. Its position up or down in the valve cage controls the ratio of air to gas. If moved up the effective gas port area is reduced, while that of the air ports is increased by the same amount, and vice versa. Thus no matter what the gas used, the total effective area is in all cases the same. Rotary motion of the valve controls the cut-off of the mixture along the suction FIG. 12-29. Buckeye Two-cycle Engine. stroke. This action is controlled by a Rites inertia governor which operates through linkage as shown in Fig. 12-25. The Buckeye Two-cycle Gas Engine. This engine, made by the Buckeye Engine Company of Salem, Ohio, illustrates a type of medium sized two-cycle engine. The following description is taken from Power, September, 1906. It is a single-acting scav- enging twin engine and can be made to operate also on gasoline or distillate. The present engine has two motor cylinders, with cranks set at 180 degrees, two fuel pumps and two air pumps. Fig. 12-29 shows the side of the engine on which the secondary shaft is located and gives a good general view of the valve gear. Fig. 12-30 is a sectional elevation of one element of the twin engine. The piston 2 performs a double duty; in addition to delivering MODERN TYPES OF COMBUSTION ENGINES 287 the power of the explo- sions to the crank, it compresses the charge in the chamber 18 for delivery to the com- bustion chamber. The cross-head 6 is in the form of a plunger and acts as an air pump and compressor piston in the chamber 7. The ex- haust ports are opened by the piston 2, as usual in engines working on the two-stroke cycle. These ports are shown at 14, 14, Fig. 12-30. Each piston compresses its own explosive mix- ture, but each cross- head plunger delivers compressed air for sca- venging to the combus- tion chamber of the other half of the engine unit. The cycle of operation is as follows: When the piston uncovers the exhaust ports, the scavenging valve 11 is opened by the valve gear and com- pressed air at about 8 pounds per square inch is admitted to the com- bustion chamber from the air pump of the other cylinder. This air blast sweeps out the 288 INTERNAL COMBUSTION ENGINES burnt gases, and the admission valve 10 is then opened, admit- ting a charge of gas and air from the compression chamber 18; this charge is also at a pressure of about 8 pounds per square inch. The piston further compresses the charge on its back- stroke, as usual, and it is fired by an electrical igniter. As the piston travels back, compressing the charge in the cylinder, it draws a fresh charge into the front end, 3, of the cylinder and the cross-head plunger of the other half of the unit similarly draws in a charge of air. The delivery of air from the chamber and pass- ages 7, 20, and 19 is controlled entirely by the valve 11, but the intake of air by the plunger 6 is controlled by a piston valve 63, Fig. 12-31, which takes air from the chamber 68, connecting with the atmosphere through the base of the engine, and delivers it to the chamber 69, which is connected to the cross-head cylinder of the other half of the engine. The piston valve 62 takes in a mix- ture of gas and air through the chamber 67, which is connected with the supply source, and delivers it to the fuel pump 3, Fig. 12-30, through a balanced throttle valve 60, Fig. 12-31. The fuel pump then forces the mixture back through the throttle valve 60 to the' admission valve cage. The throttle valve thus regulates both the quantity of mixture drawn in by the pump and the quantity delivered by the pump to the combustion chamber. The mixture and air-intake valves are operated by connecting rods from a rock-shaft, as indicated in Fig. 12-31; the rod 38 actuates the corresponding valves for the other half of the engine. This rock-shaft is oscillated by an eccentric on the main shaft and an eccentric rod. The governor is of the fly-ball spring-opposed type and serves merely to control the position of the balanced throttle valves in the fuel passages. Since the cylinder is filled with scavenging air every stroke, the variation of the amount of mixture admitted does not vary the compression pressure but merely varies the richness of the cylinder contents. The engine is equipped with both make-and-break and jump- spark igniters, but the former are ordinarily used. The Fairbanks Engine. The Fairbanks improved horizontal enigne is of the four-cycle type and made to operate on gas or on gasoline, distillate or alcohol. The general appearance of the engine is well shown in Fig. MODERN TYPES OP COMBUSTION ENGINES 289 290 INTERNAL COMBUSTION ENGINES 12-32 which shows the gas engine FIG. 12-32. Fairbanks Engine. head of the cylinder. Each valve and, when closed, the valve face is practically flush with the wall of the combustion chamber. This results in a combustion chamber of very simple form. The lay shaft at the side of the engine is driven by two-to-one gearing from the crank shaft. It car- ries, Fig. 12-32, first the bevel gear for operating the hit- and-miss fly-ball governor; second a clutch arrangement for making and breaking the connection between the inlet and igniter cam sleeve and the lay shaft; third, the exhaust valve cam, and lastly the sleeve carrying the inlet valve and igniter cams. The gov- ernor acts by interposing a pick blade, prevents the ex- from the governor side, while Fig. 12-33 gives a front view of the same ma- chine. The cylinder is partly supported by the frame, as shown. Cylinder jacket wall and cylinder head are cast in one piece, doing away with all joints. Both valves of the pop- pet type work upward and are mechanically operated. Fig. 12-34 shows the position of these valves at the is placed in a separate cage, FIG. 12-33. Fairbanks Engine. MODERN TYPES OF COMBUSTION ENGINES 291 haust valve lever from returning, and thus blocks the exhaust valve open. In this position a projection on the exhaust valve lever unlocks the clutch above mentioned, breaks the connection between lay shaft and inlet cam sleeve, causing the latter to remain stationary. The inlet valve then fails to open as long as the governor does not withdraw the blade. It is possible to change the speed of the engine through a certain range by adjust- ing the governor during operation. The ignition system is of the make-and- break type, as is clearly indicated in Fig. 12-33. The electrodes are contained in one block, which is easily removable for inspection. FIG. 12-34. Cylinder Construction, Fairbanks Engine. The method of operating the gas valve by means of the main inlet valve lever is also shown in Fig. 12-33. The Fairbanks gasoline engine is in all respects similar to the gas engine above described, except that the mixing valve shown at the left of Fig. 12-33 is replaced by a simple type of overflow carbureter which is supplied by a gasoline pump operated by a cam on the lay shaft. The Philadelphia Otto Engine. The general features of the design of this engine are shown in Figs. 12-35 and 12-36. These particular drawings refer to a 30 horse-power illuminating gas engine, but the same design is carried out in all horizontal engines from 5 up to and including 40 horse-power. Excellent features of the construction are the separate cylinder liner and the remov- 292 INTERNAL COMBUSTION ENGINES able valve cages for gas, inlet, and exhaust valves. Make-and- break electric ignition operated by a crank from the end of the lay shaft is used. FIG. 12-35. Philadelphia Otto Engine. The producer gas engines made by this firm, the Otto Gas Engine Works of Philadelphia, show the same general make-up. Fig. 12-37 gives a general view of a type made in 60, 75, 95, and 120 horse-power sizes. This design differs from that of Fig. 12-35 mainly in that the cylinder is supported also near the end by an extension of the frame. FIG. 12-36. Philadelphia Otto Engine. The suction gas engines are governed by controlling the fuel valve, regulation thus being effected by changing the quality of the mixture. The Olds Gas Engine. The Olds Gas Power Company of Lansing, Mich., build two types of gas engines: Type G from MODERN TYPES OF COMBUSTION ENGINES 293 8 to 100 horse-power and Type K from 25 to 300 horse-power, both being horizontal single-acting four-cycle engines. FIG. 12-37. Philadelphia Otto Producer-gas Engine. Type G is illustrated in Fig. 12-38. This engine may be used either for gas or gasoline. There appear to be no very unusual features in its general design. Both inlet and outlet valves are FIG. 12-38. Olds Type G Engine. of the poppet type, but an auxiliary exhaust opening is used to relieve the main exhaust valve. The exhaust valve is mechani- 294 INTERNAL COMBUSTION ENGINES cally operated by means of a straight push rod and cam. Igni- tion is by make and break. The governor is of the hit-and-miss type and operates to hold the exhaust valve open. The mixing arrangements are not specially described in the available informa- tion on this engine. Type K engines are of somewhat different design and embody in their make-up the best and most advanced ideas. A general view of the machine is given in Fig. 12-39. Among the excellent features of this design are the following: The jacket wall is in- tegral with the frame. The cylinder liner is made of a grade of FIG. 12-39. Olds Type K Engine. metal especially adapted to the service and consists of a straight cylinder with a flange at the outer end. This flange is received into the frame and is held in place by the cylinder head. This construction allows of even and unrestricted expansion. The cylinder head contains the openings for the inlet and outlet valve cages and is designed with the greatest possible regard to expan- sion and cooling stresses. The following description of the valve mechanism is taken from the catalogue published by the company. The inlet and exhaust valves are of the vertical poppet type, MODERN TYPES OF COMBUSTION ENGINES 295 mechanically operated and working in long guides in the same vertical axis with the inlet valve at the top and the exhaust valve at the bottom. The inlet valve and gas valve have a com- mon stem and the cage is so arranged that a thorough mixing occurs just at the entrance to the cylinder. The gas valve opens slightly later than the air valve and by special construction used a perfect seating of both valves is assured at all times. On the smaller sizes, by removing the inlet valve cage the exhaust valve is perfectly accessible, while on the larger sizes the exhaust valve may be removed together with its hollow water-cooled seat with- out disturbing the inlet valve. . This is easily done as its weight is counterbalanced. Both valves are operated by a single cam which is designed so as to have a quick, full valve opening, with- out noise or clatter. Valves and valve ports are of liberal sizes so undue throttling is avoided. The valve springs are of the highest grade spring steel, and rest on plates which, being sup- ported on ball and socket joints, do away with side thrust. All parts and particularly all bearings are made of ample dimensions with provision for oiling every wearing surface. The lay shaft which is driven from the crank shaft through spiral gears operates the valve mechanism, governor and ignition mechanism. The speed is controlled by a governor, apparently of the Hartung type, which is driven from the lay shaft and serves to throttle the mixture admitted to the cylinder. Figure 12-40 gives a good view of the ignition arrangements. The current is supplied by a low-tension make-and-break Bosch magneto, the operation of which is explained in Chapter XIII. The point of ignition can be easily changed during operation by adjusting the lever along the row of holes A-R. The Warren Engines. The Struthers- Wells Company of Warren, Pa., manufacture various types of engines, both hit- and-miss and throttling, all operating on the four-cycle prin- ciple. For ordinary power purposes the firm builds the single-cylinder hit-and-miss engine illustrated in Fig. 12-41. This engine is made in sizes from 10 to 90 horse-power. Above 30 H.P. the instal- lations are equipped with special starting devices. All of the other types are apparently governed by throttling. The next higher range of power, 35 to 200 horse-power, is covered 296 INTERNAL COMBUSTION ENGINES by the two-cylinder throttling engine shown in Fig. 12-42. The manner of governing is clearly indicated. A special type of the two-cylinder engine, for which maximum economy and closest regulation is claimed, is shown in elevation in Fig. 12-43. This engine is made in two sizes only, 110 and 125 horse-power. From the vertical section, Fig. 12-44, showing the cylinder construction, it is seen that the cylinder barrel and jacket are cast in one piece. The valves are of the vertical poppet FIG. 12-40. Igniter Details, Olds Type K Engine. type, opening upward, and are placed in a valve cage integral with the cylinder head. The manner of operating these valves by cams and levers from a valve shaft running across the engine under the cylinders is shown in both Figs. 12-43 and 12-44. Governing is effected by throttling the mixture in the supply pipe just before it divides. The centrifugal governor and its linkage are indicated in Fig. 12-45. In some cases the governor is located between the cylinders as shown in Fig. 12-43. A cross-section through the valve chest, Fig. 12-46, shows the method of operat- MODERN TYPES OF COMBUSTION ENGINES 297 FIG. 12-41. FIG. 12-42. Warren Throttling Engine. 298 INTERNAL COMBUSTION ENGINES FIG. 12-43. Warren Throttling Engine. FIG. 12-44. Vertical Section Warren Throttling Engine. Of THE UNIVERSITY OF MODERN TYPES OF COMBUSTION ENGINES 299 FIG. 12-45. Plan of Warren Two-cylinder Engine FIG. 12-46. Cross-section through Valve Chest, Warren Two-cylinder En- gine. 300 INTERNAL COMBUSTION ENGINES ing the make-and-break igniter, and the location of the starting valve in the valve chest cover just above the exhaust valve. Warren engines, covering the range from 200 to 325 horse- power, are of the tandem single-acting type, a general view of which is shown in Fig. 12-47. The cross-section, Fig. 12-48, shows the cylinder construction, which is unique in some of its features. The back part of the main frame forms the jacket wall for the front cylinder. The cylinder barrel itself is cast in one piece with the cylinder head. The construction of the back cylinder is similar. The distance piece between the two cylinders rests on a separate base, and forms, .for a portion of its length, the jacket wall for the rear cylinder. The front piston is of the ordinary trunk type. The back piston is longer than is usual in pistons not subject to side thrust. The method of water-cooling piston and rod is clearly shown. The valves are of the vertical poppet type held in separate cages which are easily removable. The exhaust valves, at the bot- tom of the cylinder, are water-cooled. The valve gear is shown in detail in Fig. 12-49. Both valves are operated from the same cam on the lay shaft. The operation of the exhaust valve is clear. The motion of the inlet valve varies, depending upon the load. The valve opens and closes always at the same time, because the lift of the actuating cam is not changed throughout the entire range of load. The governor, however, through the linkage shown, controls the position of the sliding block above the valve lever, moving it in or out, depending upon whether the load falls or rises. This block acts as the fulcrum about which the valve lever turns, and hence the lift of the valve is made proportional to the load. The inlet valve stem carries the gas valve. The latter opens somewhat later than the main inlet valve. The manner of mix- ing gas and air is clearly indicated in the figure. Butterfly valves in the air and gas passages serve to help control the proportions of the mixture. The ignition system, operated from the lay shaft as shown in Fig. 12-49, is of the make-and-break type. The Struthers- Wells Company also builds vertical engines up to 600 horse-power. The older multi-cylinder types of these machines are practically nothing but two or three separate engines direct connected. Thus a 450 horse-power unit now in operation consists of three engines with two fly-wheels between the cylinders. MODERN TYPES OF COMBUSTION ENGINES 301 302 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 303 FIG. 12-49. Valve Gear of Warren Single-acting Tandem Engine. 304 INTERNAL COMBUSTION ENGINES In the later designs this has been modified. A 600 horse-power vertical engine lately completed is of the four-cylinder A-frame type mounted on a solid bedplate with the fly-wheels at the ends. 2. LARGE GAS ENGINES. Large gas engines of American design are manufactured by the Westinghouse Machine Company, by the William Tod Company, Youngstown, Ohio, by the Snow Steam Pump Company, Buffalo, by the Riverside Engine Com- pany, Oil City, Pa., and by the Wisconsin Engine Co., of Corliss, Wis., makers of the Sargent engine. There are, however, a few other firms making large engines of foreign design. Thus the De La Vergne Machine Company of New York make the Koerting two-cycle engine, and the Power and Mining Machinery Company the Crossley engine. The Allis-Chalmers Company, who built the Niirnberg engine, have apparently an engine on the market that is not strictly of Niirnberg design. Besides these well- known machines made in this country, the Premier engine made in England, the Cockerill engine made in Belgium, and the Ger- man Oechelhauser and Deutz engines, should be mentioned. The Oechelhauser and Koerting engines are made under license by a number of firms in Germany, and in the case of the Koerting engine, also abroad. While it has in general been easy to get sufficient descriptive material on the above-mentioned engines, this does not apply to certain machines of American design, and the information given is hence somewhat meager. The Westinghouse Horizontal Engine. There seems to-be available practically no definite information on the constructive details of the Westinghouse horizontal double-acting tandem en- gine.* The older type apparently used the center crank and the combustion, or at least the valve, chambers were placed at the sides of the cylinders. The later types of horizontal engines built by this company approach much nearer to established European practice. Thus the engines installed in the power plant of the Warren and Jamestown railway system have the valves in the central lines of the cylinders, as shown in Fig. 12-50. The latest design is shown in Fig. 12-51, which represents a 3000 horse-power unit for the power plant of the Carnegie Steel Company at Besse- mer, Pa. In this type the side crank is used, but outside of this * A full description of the Westinghouse Engine has just appeared in Power, April, 1908. MODERN TYPES OF COMBUSTION ENGINES 305 306 INTERNAL COMBUSTION ENGINES nothing is definitely known to the writer regarding the mechanical details. An interesting development of very recent date is the FIG. 12-51. Westinghouse Double-acting Tandem Engine. Westinghouse vertical single-acting tandem engine shown in Fig. 12-52. This engine was built by the British Westinghouse Elec- tric and Manufacturing Company.* The Tod Engine. The following description of this engine is taken from the Iron Age of July 18, 1907: "The general ap- pearance of the engine is shown in Figs. 12-53 and 12-54. The cyl- inders are arranged in pairs, two cylinders being connected in tandem to each crank pin. The valve gear, igniter, switchboard and operator's hand wheels are all situated between the cylinders and are easily accessi- ble from an operating platform placed just FIG. 12-52. - Westinghouse Vertical Single- bel W the level f the acting Tandem Gas Engine. center lines of the cyl- * W. H. Booth, Cassier's Magazine, November, 1907. MODERN TYPES OF COMBUSTION ENGINES 307 308 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 309 inders and shaft. The foundation level under the cylinders being several feet lower than under the frames, leaves ample room be- low the cylinders to make the exhaust valves, etc., easily acces- sible, a very desirable but somewhat unusual feature. " The cylinders and water jackets are integral and cast in halves, secured together with flanged joints at the center. The cylinders are not attached directly to the main bedplates nor to each other, but are supported by the tie pieces in such a way that the barrels themselves are entirely free. All strains are transmitted through four heavy forged steel tie bolts extending the entire length of the cylinders, and attaching directly to heavy lugs on the bedplate. This obviates the transmission of the strains through the cylinder walls, and contributes to accessibility and the easy removing of parts. The pistons are of steel, with cast-iron junk rings, and the pistons and rods are water-jacketed in the usual manner. Adjustable tail-rod supports take the weight of the pistons and rods from the cylinders. " The valve gear is driven by eccentrics, eliminating entirely the cam drive, which has been an objectionable feature of many of the gas engines heretofore built. There is one eccentric for each end of each cylinder, which drives both the inlet and exhaust valves an arrangement that reduces the number of parts. The eccentrics are mounted on two lay shafts running parallel to the axes of the cylinders. The latter are driven by a cross shaft, which, in turn, derives its motion from two eccentrics mounted on the main shaft. The inlet valves are on top of the cylinders, and the exhaust valves on the bottom. The inlet valve proper is of the mushroom type, sealing the ports from the pres- sure in the cylinders. The main valve is operated by a rolling lever and returned to its seat by a spring. The mixing valves and governor valves are of radial gridiron type, and are located in the upper section of the valve bonnet. The mixing valves may be operated individually or collectively by a suitable hand mechanism. The governor valve which controls the admission of gas has constant travel, the time of opening being controlled by the governor by either increasing or decreasing the angle of advance of the crank which operates them, since the engine is of the constant-compression type. "The governor is of the fly-ball pattern and is in duplicate, 310 INTERNAL COMBUSTION ENGINES one controlling the operating valves on each side of the engine. The two are driven by Morse chains through a flexible coupling, and are connected by a cross rod which may be removed in case it is desired to operate either side of the engine alone. "Each end of each cylinder is equipped with two igniters, one operated mechanically and the other by a solenoid. Either may be used independently or the two together. The igniters are under control of the governor, and when the engine is at rest are automatically thrown back to the dead center. The ignition is on the make-and-break system, using direct current at 90 volts, supplied by a motor generator set, which is so designed that either end may be used as a motor or a generator, or both may be used as generators by driving directly from the engine shaft. Connected in series with each igniter is a tell-tale lamp on the switchboard, giving a positive indication as to whether or not the igniters are sparking, short-circuited or burnt out." The Sargent Engine. The Sargent complete expansion engine, made by the Wisconsin Engine Company of Corliss, Wis., is shown in general elevation in Fig. 12-55 and in transverse sec- tion in Fig. 12-56. As far as the constructive details of this en- gine are concerned, it has the merit of great simplicity. The engine is built as a double-acting tandem. There is but one valve to control admission and exhaust for each end of each cylinder, and but a single cam to perform these various offices, while a second cam operates the igniter. The lay shaft is driven from the main shaft by a pair of worm gears. The governor is of the inertia type, the Rites, and operates to advance or retard the lay-shaft, thus controlling the time of cutting off the admission of the incoming charge to the cylinder. This construction is decidedly different from that used by other designers. The combination valve is shown in cross-section in Fig. 12-56. Its operation is described as follows : "Gas is piped to the chamber A in the sub-base and air to the chamber B, which pass through the cylinder supports to the chambers A' and B', ready to pass into the mixing chamber when the cam depression M N passes the roller and the ports F in the piston valve register with the ports E and D in the bush- ing. When the piston valve goes down to this position, the con- fined air in the piston valve dash-pot forces open the poppet valve, MODERN TYPES OF COMBUSTION ENGINES 311 I 312 INTERNAL COMBUSTION ENGINES thus giving free admission to the charge. When the point N of the cam reaches the roller, it is forced down, while the other end of the lever goes up, carrying the piston valve which cuts off the admission. The poppet valve seats and both valves remain FIG. 12-56. Transverse Cross-section Sargent Engine. in normal position during compression, ignition, and expansion, or until the point L on the cam pushes the roller down and the piston up, which opens the poppet and the exhaust gases pass out through the ports K and the elbow W to the exhaust pipe under the floor. "The poppet valve seals the opening in the combustion chamber during the compression and inflammation, and the piston valve, MODERN TYPES OF COMBUSTION ENGINES 313 holding against no pressure, works loosely in its bushing, cutting off the admission and guiding the exhaust. " As the poppet valve controls both the inlet and outlet gases, both the valve and seat keep cool and need to be ground not over once or twice a year. The mechanism is simple and, as the roller is always bearing on the cam, the valve motion is practically noiseless." By revolving the piston valve by the index wheel, the blind port S varies the mixture to suit the gas whether it has 100 or 1000 B. T. U. per cubic foot. The fundamental idea of the Sargent engine is to get complete expansion of the charge in the cylinder. In the ordinary engine, which at full load draws in the charge to the full stroke of the piston, or nearly so, the terminal pressure is anywhere from 25 to 50 pounds, and the final temperature over 1000 degrees Fah- renheit. Sargent claims that his engine, cutting off the stroke at full load at three-quarters of the suction stroke, will show a ter- minal pressure slightly above atmosphere and a temperature of about 400 degrees Fahrenheit. The system of speed regulation by cutting off at various points along the suction stroke is one used by several makers, but in nearly all cases the engines cut off at nearly full stroke for full load, that is, at full load the ratio of compression is nearly equal to the ratio of expansion. In the Sargent engine the ratio of expansion always exceeds the ratio of compression, and a lowering of the terminal pressure and temperature is of course the result. There is no doubt that this lowering of the temperature has a great deal to do with any success that the use of a combination valve such as described above may have. In the Sargent engine the rated power is developed when the governor cuts the admission off at three- quarter stroke; the rest of the stroke may be considered potential over-capacity. In spite of the fact that ordinary experience shows a gas engine to be most efficient when it is developing its maximum, not the rated, load as long as normal speed is main- tained, the claim is made for this engine that its best efficiency is obtained when cutting off at three-quarter stroke. The Koerting Two-Cycle Engine. The Koerting engine is of German design, made in this country by the De La Vergne Machine Company of New York. This is one of the two successful large 314 INTERNAL COMBUSTION ENGINES two-cycle gas engines .in Europe, the other being the Oechel- hauser, not made in America. Koerting two-cycle engines are now made by several German firms, whose constructions differ somewhat among themselves and from the parent design, mainly in frame and governor details. The latter seem to have ah im- portant bearing upon the amount of work done by the pumps. The type made by the De La Vergne Machine Company is best illustrated from the catalogue of the firm. Fig. 12-57 shows a cross-sectional elevation and a plan of this machine, used in this case as a blowing engine. It is seen that the engine has but a single cylinder which is double-acting. There are thus exactly the same number of power impulses per turn as in a single-cylinder steam engine. The piston is very long, about seven-eighths of the stroke, and is of course water-cooled. This makes it necessary to support its weight at the cross-head and at a tail-bearing to save both cylinder and stuffing-boxes. At the side of the power cylinder there are a gas and an air pump which furnish gas and air by separate passages to the mixture inlet valves at the top of the cylinder. The manner of driving "these pumps from a side crank and of operating the piston valves, which control them, by means of rocker arms driven by an eccentric from the main shaft, is clearly shown in Fig. 12-58. Fig. 12-59 illustrates the valve shaft side of a partially constructed machine and shows the manner of operating the inlet valves and igniter gear. The exhaust gases are taken care of by a ring of ports in the middle of the cylinder. These ports are uncovered by the piston alter- nately on each side (see Fig. 12-57). The operation of the engine is as follows: It will be seen, from Fig. 12-58 that the pump crank is in the neighborhood of 100 degrees ahead of the main crank, i.e., when the latter is at either dead center the pumps have completed about one-half their stroke. From the beginning of its discharge stroke the air pump B, Fig. 12-60, has forced its charge of air into the passage leading to the main inlet valve and has forced some of this air also partly up to the gas passage, which is the inner concentric passage surrounding the inlet valve stem. In the meantime the piston of the gas pump A, the motion of which has been exactly the same as that of the air pump piston, has been allowed to shove back part of its charge into the gas suction pipe. MODERN TYPES OF COMBUSTION ENGINES 315 316 INTERNAL COMBUSTION ENGINES r 'So W MODERN TYPES OF COMBUSTION ENGINES 317 At the moment, therefore, when the exhaust gases have nearly completely escaped from the main cylinder through the ring of ports, and the main inlet valve is opened, there is at first a rush of air from both the air and gas passages. This serves to drive the remainder of the exhaust gases out of the cylinder. A mo- ment later the gas pump starts to deliver gas and the mixture enters the cylinder. By the time the power piston has covered the exhaust ports on its return stroke, the air and gas pistons FIG. 12-59. Valve Gear of Koerting Two-cycle Engine. have reached the end of their stroke, the inlet valve closes and the mixture is compressed in the power cylinder. Ignition is by electric spark. The speed of Koerting two-cycle engines is controlled by pro- portioning the amount of gas to the load. In the American type this is accomplished by putting butterfly throttle valves /-/, Fig. 12-60, in the discharge passages of the gas pump. The position of these valves is controlled by the governor, which, as some of the other illustrations show, is of the centrifugal type. This method of pump control- is open to the objection that if the 318 INTERNAL COMBUSTION ENGINES gas used is at all dirty, the frictional resistances soon become so great as to render the governor inoperative. Some of the im- provements of the later Koerting engines have been devoted to this very point. A modified design of gas pump is that of Klein Bros., Dahlbruch, Fig. 12-61.* In this construction the piston forces the charge back into the suction space through the ports shown at the middle of the cylinder, for one-half the stroke. The amount of gas delivered to the discharge passage during the last FIG. 12-60. half depends upon the position of the small overflow piston valves near the bottom. This is controlled by the governor through the linkage shown. Another point that has given some trouble in the first designs of Koerting engines is the fact that the time available for the opening and closing of the inlet valve was but a very small part of the time of one revolution. The inertia actions were very severe and hence the engine speed was somewhat re- stricted. This difficulty has been overcome in an elegant way in the design of the Siegener Maschinensban A-G as shown in Fig. 12-62. Here the lay shaft moves only at ane-half the speed of * Hoffman, Zeitschrift d. V. d. I., September 15, 1906. MODERN TYPES OF COMBUSTION ENGINES 319 FIG. 12-61. Gas Pump, Koerting Two-cycle Engine, Klein Bros. FIG. 12-62. Modification of Inlet Valve Gear, Koerting Two-cycle Engine. 320 INTERNAL COMBUSTION ENGINES the main shaft, the same as in four-cycle engines, but the valve is opened twice for every turn of the shaft. Reinhardt states that the Koerting engine is well adapted to blowing service because it starts easily under load and is certain in its operation through a wide range of speeds. The Riverside Engine. This engine is made by the Riverside Engine Company of Oil City, Pa., and embodies in its details several features decidedly different from those of other large engines. Fig. 12-63 shows the heavy duty double-acting tandem type and illustrates the peculiarity of the valve construction very clearly. Besides this type, this firm also builds single-acting, FIG. 12-63. Riverside Heavy Duty Double-acting Tandem Gas Engine. single-cylinder and twin engines, and single-acting tandem engines. One of the single-acting, single-cylinder engines is illustrated in Fig. 12-64. The following is a description of the main features of the double-acting machine. It applies with little modification also to the single-acting type. Fig. 12-65 shows in cross-section the cylinder, valve and piston details of the double-acting engine and will serve as an aid in following the description of these parts. The main frame or bed is of the heavy duty tangye rolling-mill type with bored guides and main bearings cast in a single piece, and is of great weight and extreme rigidity. The cylinder end of main frame is squared similar to cross-section of cylinder, and has machined holes for receiving the tie bars which attach the cylinder. The design of this frame, with its bored guideway, is such that a large portion of the metal is above the center line, making it exceptionally stiff. MODERN TYPES OF COMBUSTION ENGINES 321 The shaft is of side crank, built-up type and is machined all over. The cylinders are made in two halves, with heads and valve chest cast integral without joints or packing, and are held rigidly to the main frame by four heavy steel tie bars which take all tension strains. The cylinders are mounted on a heavy cast-iron sole plate, having a machined top surface which keeps cylinders in alignment, permits perfect freedom for expansion ami contrac- tion, and by removing distance piece, cylinders can be slid end- wise on sole plate, giving easy access to interior of cylinder, pistons Fi<;. 12 ()4. Riverside Single-art ing .Single Cylinder (las Engine. and piston rings. Rings can be cleaned or changed without dis- turbing piston or rod. The sole plate makes a drip pan under all the cylinders, keeping oil drip from foundation. The exhaust and inlet piping is attached to the sole plate, hence no piping except the water-jacket piping has to be disturbed when cylinders are moved. There is no overhead piping or wiring to interfere with traveling crane. Both inlet and exhaust valves are of the semi-balanced water- cooled poppet type, operated in a vertical position. All water passing to the cylinder jacket passes through valves first, giving perfect and positive cooling without any attention whatever. The balancing pistons run in renewable liners and are lubri- cated positively. These pistons form a large and perfect guide, 322 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 323 assuring positive alignment for these valves indefinitely. The valve seats are renewable and are located slightly below bottom of cylinder bore, so that all foreign substances are swept from the cylinder at each exhaust stroke. All valves are readily removed from the top of the cylinder without disturbing cam shaft. Piston and piston rods are water-cooled, the water entering through a telescopic joint connected to side of cross-head. Cir- culation through each piston is positive and the overflow is so arranged that pistons are kept full of water. Water passages through piston and rod are large and easy so that not over ten pounds' pressure is required for circulation. The overflow is visible so that water cannot come to a boiling-point without attracting engineer's attention. A heavy tail-rod support and adjustable shoe is provided for carrying weight of pistons and rod. The construction of the piston rod is such that a piston in either cylinder can be removed without disturbing the other cylinder or the connecting rod, cross- head or any part of the valve gear. The valve gear of the Riverside engine is very simple and con- sists of a single shaft mounted on top of the engine, running in self-oiling bearings. This shaft runs at one-half the speed of the main shaft and carries the inlet and exhaust cams, cams for operating the oil pumps, and the timers for ignition system: Power is transmitted to the inlet and exhaust valves by an inlet and exhaust lever hung on a single pin. All cams are keyed rigidly to the cam shaft. This construction makes a minimum number of joints subject to wear. Ample adjustments are pro- vided for taking up wear. Ignition is by an improved method throughout, consisting of two magnetically operated spark plugs in each cylinder. The sparking points being in series with the magnet coils give a posi- tive external indication by the movement of the armature as to whether the spark takes place within the cylinder. The timing is electrical; no gearing or mechanical trips being used, simply a No. 14 stranded wire running through iron armored conduits leads to each plug. The timer is mounted on the secondary shaft and is built in a heavy, substantial manner. The contacts are of the wipe type and are made of tool steel; and are adjustable for wear. A visible 324 INTERNAL COMBUSTION ENGINES indicating spindle shows the amount of contact. The entire wearing parts of timer are run in oil, which prevents burning of the contacts and reduces wear to minimum. Wiring from the timer to each spark plug is protected by a five ampere enclosed fuse; thus, should any spark plug become damaged or short-cir- cuited, it would have no effect on the rest of plugs. Any spark plug can be removed and replaced while the engine is in opera- tion. The time of ignition in all cylinders can be adjusted simul- taneously, and by an individual adjustment the time of ignition in any cylinder can be adjusted independently. All of these adjustments can be made when the engine is in operation. The method of starting the double-acting engine is as follows: A gas-engine driven air compressor is provided for supplying compressed air for starting this engine, air to be stored in a suitable steel receiver. The pneumatic starting gear on the engine consists of two shifter pistons mounted within the valve lever carry- ing pin. These pistons are shifted by throwing a small three- way cock which applies compressed air to one side of the piston, which in turn moves the valve levers f inch, bringing the cam rollers in line with an auxiliary cam which puts all the valves in both ends of one cylinder into two-cycle action, or, in other words, makes a poppet valve air or steam engine out of this cylinder, which permits the engine being started on any stroke and on either quarter. The operation being entirely automatic, the engine will run as long as compressed air is applied. The other double-acting cylinder continues to operate as a four-cycle gas engine and takes up its explosions after the first revolution. Compressed air is then shut off from the starting cylinder and the three-way cock reversed, which permits the cam ' levers being returned to their normal position and that cylinder immediately goes into four-cycle action and takes up its explosions on the next revolution. The starting gear adds practically no complication to the engine as the regular inlet and exhaust valves are used for distributing the air. There are no extra shaft or auxiliary valves and no tappings into the cylinders. There is only one compressed air pipe leading to the sole plate, air being delivered into the fuel duct and entering the cylinders via the inlet valves. The Crossley Engine. This English engine is made in this MODERN TYPES OF COMBUSTION ENGINES 325 country by the Power and Mining Machinery Company of New York, and known as the American Crossley. These engines are built as single-cylinder, two cylinder-opposed and double-opposed units, sizes ranging from 50 to 1300 B. H. P. This design differs radically from that of other large engines in that the single-acting cylinder is retained up to the largest sizes, i.e., the design is in a way nothing but an enlargement of the small four-cycle machine. It is fundamently the same design employed at first by the Deutz Company for their large engines, but has been given up by them some years ago in favor of double-acting engines. Figures 12-66 and 12-67 show a general elevation and a cross- sectional elevation respectively of the single-opposed type. The cylinder castings are securely bolted to each end of a central frame casting. Both connecting rods work on one crank pin. The pistons are water-cooled. The cylinder proper is a separate liner, resting at the crank end in the jacket casting and held at the head end by the cylinder head. The head carries the gas and inlet valves, while the exhaust valve is placed in a separate casting. The position of inlet and exhaust valve is shown in Fig. 12-68. Both are operated in a horizontal position by means of bell cranks by cams on a lay shaft. In this respect the engine differs from most others in which the valves are nearly always vertical. The objection to the horizontal form is at least partly overcome by giving the exhaust valve a guide at each end. This valve is water-cooled, as is necessary in large machines. The lay shaft s, Fig. 12-66, is operated by spiral gears, and actuates the main inlet, the cut-off valve, the gas valve, the exhaust valve, and the igniter gear. Gas and air enter the mixing chamber M, Figs. 12-66 and 12-69 through a proportioning valve, the air direct and the gas through a special gas valve. From here the mixture passes a multi-port cut-off valve, which surrounds the valve stem of the inlet valve and is shown in greater detail in Fig. 12-70. This cut- off valve is operated from the lay shaft through the rocker arm C, Fig. 12-71. The governor, shown in plan in Fig. 12-69, serves to shift the fulcrum / of the rocker arm c, Fig. 12-71, thus cutting off the mixture supply to the main inlet valve earlier or later as the load demands. Should the load drop so far that the com- 326 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 327 328 INTERNAL COMBUSTION ENGINES FIG. 12-68. American Crossley Engine, Inlet Valve at A, Exhaust Valve at E. r-S FIG. 12-69. Details of Valve Gear, American Crossley Engine. MODERN TYPES OF COMBUSTION ENGINES 329 bustion becomes sluggish under very low loads, the engine auto- matically shifts over to the hit-and-miss governing by the gover- nor withdrawing the block b, Fig. 12-69, thus keeping the gas valve closed altogether. In multi-cylinder engines the point at which this occurs is so adjusted is so that one cylinder after another changes over to the latter system of regu- FIG. 12-70. Inlet Valve, American Crossley Engine. lation and not all of them at the same time. The American company for some time adhered to hot tube ignition, but some of the later engines were fitted with make-and- break electric ignition. The Snow Engine. The Snow Steam Pump Company build FIG. 12-71. End View, American Crossley Engine. two types of horizontal double-acting four-cycle machines, called respectively Type "A" and Type "B." Both of these designs differ radically from the construction of other large gas engines 330 INTERNAL COMBUSTION ENGINES in that the valves are placed in chambers at the side of the cylin- ders. There are reasons for and against this construction, the main disadvantage being perhaps the cut-up form of the com- bustion chamber. Some of the many advantages are outlined below. Type B engines are built in single-cylinder units from 20 to 125 horse-power, and in tandem units from 80 to 500 horse- power. The general features of the design are shown in eleva- ion in Fig. 12-72 and in cross-section in Fig. 12-73. FIG. 12-72. Snow Type B Gas Engine. The most notable installations of Type A engines is found in the Martin station of the San Mateo Power Company in Cali- fornia. This station at present contains three of the twin- tandem units, a fourth is in course of erection, and the foundation is laid for a fifth. Fig. 12-74 shows a general view of the station, and also serves to give a general idea of the appearance of the engines. The engines of the Martin station were described in great detail by C. P. Poole in an article in Power, January 14 and MODERN TYPES OF COMBUSTION ENGINES 331 21, 1908, to which the writer is indebted for the following infor- mation: Figure 12-75 shows in cross-section the main features of the design, while Fig. 12-76 is a transverse cross-section through one of the valve chambers to show the valve construction. Each cylin- der of the twin tandem unit is 42 inches in diameter by 60-inch stroke. The engines are direct connected to Crocker- Wheeler generators rated at 4000 K.W. each. They are capable, however, of carrying momentarily an overload of 33 per cent, and have demonstrated their ability to carry 15 per cent overload con- tinuously. These figures make these units the largest gas power /HJl2X L , ^fir=rraaiS FIG. 12-73. Cross-section of Snow Type B Gas Engine. engines in the world. The fuel used is an oil-water gas made by the Lowe process. The main frame is of the side crank type. The cylinders are cast in two parts, fastened together by flanges at the center, as shown in detail in Fig. 12-77. This construction makes the outer wall independent of the cylinder wall proper, and avoids tempera- ture stresses. The valve chambers are cast integral with the cylinder casting. The cylinder heads are plain jacketed covers containing nothing but the stuffing-boxes. For the purpose of safety against leakage, however, each head has a double seat. The front and rear cylinders are connected by a distance piece of box-section, which has large openings in the side through which heads and pistons may be removed if required. The pistons are 332 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 333 made in two parts bolted together. These pistons, together with their rods, are of course water-cooled, the water entering at the front cross-head and leaving at the tail cross-head. The distinguishing feature of these engines, however, is the valve construction, shown in detail in Fig. 12-76. Mr. Poole, in the article mentioned, describes the gear as follows: "The cam shafts which drive the valve gear are driven by steel gears, bevel gears being used on the main shaft, driving back through spur wheels to secure the proper reduction in speed of two to one. The igniters and lubricators are also driven from the cam shafts, as are also the starting valves, which admit com- pressed air on the cylinders for starting. The inlet and exhaust FIG. 12-75. valves are driven by cams made of chilled cast iron, located on the cam shafts, which run in bearings bolted to the side of the valve chambers. "The inlet and exhaust valves are of the unbalanced mush- room type, working in cages secured to the top and bottom of the valve chamber. Both valves and their cages are water-jacketed, in order to prevent back-firing or pre-ignitions on account of the treacherous nature of the gas used. Each inlet valve is com- bined with a combination mixing and throttling valve, of piston form, so designed that when the inlet valve opens gas and air ports in proper proportion are open for the passage of gas and air in the ratio desired, the amount of the opening of both being fixed by the governor. From this it will be evident that the engines operate with variable compression and constant mixture, the supply of air and gas to each end of each cylinder being throttled directly at each inlet valve for its end of its cylinder. 334 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 335 "The exhaust valves are of cast iron with nickel-steel stems and are thoroughly water-jacketed, the water being fed to and carried from the valve by positive circulation. The connection of the exhaust valve cages with the exhaust pipe is so made that the cages can be readily removed without disconnecting any other part of the engine." In discussing the reasons for locating the valves at the side, Mr. Poole quotes the builders as follows: "It permits of the use of an absolutely continuous bedplate under all cylinders, which is considered essential on engines as long as these for maintaining absolute alignment and permit- ting unrestricted movement to compensate for variations in temperature in cylinder walls and connecting parts. "A solid, unbroken founda- tion from one end of the engine to the other is thus secured, ,and the builders consider that a solid, unbroken foundation is more essential for a gas engine than for any other prime mover on account of the enormous weight of the reciprocating parts and the inability to fully coun- terbalance. "It enables all working parts of the engine, without exception, to be located above the engine-room floor and therefore to be in full view of the attendants at all times. "Inlet- and exhaust-valve driving motions from the cam shaft are short and direct. "Exhaust valves are much more accessible for removal when located in a valve chamber on the side, since the crane can be used throughout the entire operation and all work in connection with the removal and displacement of the exhaust valves is done from the engine-room floor. "With valves located in a side valve chamber, broken inlet and exhaust valves cannot get into the interior of the cylinder and lodge between the piston and the cylinder-head, causing wrecks. FIG. 12-77. Cylinder Construction, Snow Type A Gas Engine. 336 INTERNAL COMBUSTION ENGINES "It has been very generally the opinion of gas-engine designers that the location of the valves in a side valve chamber of the kind used on the California engines entailed the certain disadvan- tage that foreign material entering the cylinder with the air and gas would deposit on the bottom of the cylinder counterbore, while if the exhaust valve be located on the bottom, such deposits will be carried out through that valve; furthermore, that lubricating oil collects in the bottom of the counterbore and is carbonized, causing back-firing and pre-ignitions. Experience with these engines thus far indicates that if lubrication is properly effected no carbonized oil is found in the cylinder, and that when oil is supplied excessively the resulting deposit of carbon is located not on the surface of the clearance space, but on the top of the piston barrel directly under the oil-inlet holes, on the piston rod where it wipes in from metallic packing, and on the face of the cylinder head directly under and close to the piston-rod hole in the head." The Cockerill Engine. This machine is built by the Soc. Anon. John Cockerill in Seraing, Belgium. This firm builds several types of engines, differing among themselves mainly in their valve constructions. Fig. 12-78* shows a 1200. horse- power tandem gas engine, in which the inlet valve is on top, the exhaust valve on the bottom of the cylinder. Both are operated apparently from the same cam on the lay shaft. This type of machine is built either for quality or quantity governing, in which case the construction of the inlet valve differs. Fig. 12-79 shows another form, a single-cylinder double-acting engine direct connected to a blowing cylinder. The details of the valve gear of this machine are quite clearly shown in Fig. 12-SO.f Both inlet and outlet valves are side by side at the bottom of the cyl- inder. Figure 12-81 shows the latest form of the double-acting tan- dem Cockerill engine. J The distinctive features of this machine are that all valves are placed below the cylinder and that eccen- trics only are used for operating these valves. The valve gear proper represents one of the most complicated constructions used *H. Dubbel, Z. d. V. d. I., Sept., 1905. fGiildner, p. 479. JGuldner, p. 481. MODERN TYPES Of COMBUSTION ENGINES 337 FIG. 12-78. FIG, 12-79. 338 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 339 to-day and is shown in section in Fig. 12-82. In this figure at the left, the eccentric, through its rod a and the lever b, operates the cam c which positively opens and closes the exhaust valve through the triple lever d. The inlet valve eccentric, Fig. 12-82, at the right, in a similar manner operates the gas valve cam c and the inlet cam c'. The lever b actuating the gas valve cam is disengaged from the rod a when the trip arrangement /-/' is brought into play by the cam disc e. The position of this disc on its shaft is regulated by the governor, through the linkage shown, thus effecting speed regulation. The dash pot h allows the gas valve to seat noiselessly after its linkage is released. FIG. 12-81. Cockerill Double-acting Tandem Engine The Cockerill engine has found considerable application in Europe, there being about 80,000 B. H. P. in operation in 121 engines, ranging from 200 to 3000 B. H. P. The fuel used is mainly blast furnace gas. The Niirnberg Engine. Although the Allis-Chalmers Company of Milwaukee have been the American licensees of the Maschinen- Gesellschaft Niirnberg, makers of the Niirnberg engine, the large gas engine now turned out by the American firm, differs in some of the important details from the original German design. The Niirnberg engine is to-day perhaps the most important four-cycle gas engine built in Germany, and for that reason a description is first given of the German machine, to be followed by a few details of the Allis-Chalmers engine as far as they could be obtained. 340 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 341 All of the medium sized and large Niirnberg engines are double- acting, either single-cylinder or tandem. They range in size from 250 to 6000 horse-power, the larger powers being the twin- tandem type. Probably the best description ever given of this machine is contained in Riedler's " Gross Gasmaschinen," from which the cross-section, Fig. 12-83, together with the two follow- ing figures, is taken. This cut shows frame, cylinder, and valve construction very clearly. The only part rigidly fastened to the foundation is the main frame, which is of the center crank type, as preferred by all European builders. The cylinders rest on supports such that they can freely expand and contract under temperature changes. The gas and inlet valves are placed on top and the exhaust valve at the bottom of the cylinder. The inlet and exhaust ports are cast in one part with the cylinder and jacket walls. This reduces the cylinder heads to simple water- jacketed covers, a point of great advantage as compared with the ^complicated cylinder heads of former days. All valve cages with their valves are easily removable for inspection. The exhaust valves are water-cooled. The valves are operated from a lay shaft running alongside, Fig. 12-84, by means of eccentrics and not by cams. How this motion is transmitted to the valve stems by roller levers to avoid shock and noise is very clearly shown in Fig. 12-85. The pistons are as simple as those of a steam engine. The method of fastening them on the shaft together with the water-cooling arrangements for rods and pistons are indi- cated in Fig. 18-83. The weight of piston and rod is carried by outside bearings, which relieves both cylinder and stuffing-boxes. The greatest attention to accessibility has been paid in the design of this engine. As shown in Fig. 12-86, by disconnecting the piston rod between the cylinders, disconnecting the connect- ing rod and sliding the rod and cross-head forward, and by talking off the front and rear cylinder covers, the entire engine interior is at once open to inspection. The Niirnberg engine regulates by pure quality regulation, that is, the amount of gas only is cut down as the load drops. To accomplish this there is a gas valve placed ahead of each main inlet valve on the top of the cylinder. Fig. 12-83 shows the con- struction of these valves in section and the method of operating them by eccentrics from the lay shaft may be seen in Fig. 12-84. 342 INTERNAL COMBUSTION ENGINES MODERN TYPES OF COMBUSTION ENGINES 343 FIG. 12-84. Niirnberg Double-acting Tandem Engine. FIG. 12-85. Details of 'Valve Gear, Niirnberg Engine. MODERN TYPES OF COMBUSTION ENGINES 345 The latter illustration also shows the centrifugal governor and the governor shaft running alongside of the cylinders. From this governor shaft reach rods run to the gas valves and control the time of the opening of these valves in the manner explained in Chapter XIV. In some of the later engines the details of the governor mechanism are changed somewhat * but the principle of operation is the same in all, i.e., to open the gas valve later in the stroke as the load drops and to keep the time of closure the same. The air is not throttled at any time, hence the compres- sion remains about the same. It is interesting to trace out the sequence of events in a four-cycle tandem engine, for which pur- pose Fig. 12-87 published by the Allis-Chalmers Company is given. SEQUENCE OF OPERATIONS IN THE FOUR-CYCLE. DOUBLE-ACTING SYSTEMl TANDEM OR TWIN TYPE. POWER Exhaust POWER Suction Exhaus Compression Suction POWER 4 POWER STROKES IN 2 REVOLUTIONS. FIG. 12-87. Most Niirnberg engines operate on blast furnace gas. The table on page 346, also from the catalogue of the Allis-Chalmers Company, will serve to give some idea of the cylinder sizes, speeds, floor-space, etc. It has already been mentioned that the engine now built by the Allis-Chalmers Company is not quite like the Niirnberg ma- chine. The following is a description of one of these engines, as given in Power, January 21, 1908. The engine, Fig. 12-88 is of the twin-tandem, double-acting type and direct connected to a 1040 K.W. Crocker- Wheeler generator: "It is of the side-crank type so generally used in steam engine * See Zeitschrift des Vereins deutscher Ingenieure, Aug. 11, 1906, and Sept. 22, 1906. 346 INTERNAL COMBUSTION ENGINES construction. The pistons are supported entirely by the piston rods, which are turned with sufficient camber to make them dead straight when the weight of the pistons is put on them. The rods are equipped with intermediate and tail shoes, as usual with large tandem construction, running on guides between the cylin- ders and behind the rear cylinder, in addition to the main cross- head. NURNBERG GAS ENGINE TABLE OF STANDARD SIZES TANDEM TWIN | TWIN-TANDEM Size of Floor Size of Floor Size of (3 Floor Cylinders 1 Space Cylinders 1 Space Cylinders Space 1 RH a P^ & PH PC H "& 1 K | H fi w "c 1 .2 pq W^ to pa (LI O PQ W fc 1 1 1 | s| 1 1 J 1 i 1 1 I 1 3 1 | ed *"ctf 1 I " H (S *0 M > o 6 This firm also builds vertical marine engines, the operation of which is the same as that of the horizontal machine except that the water injection scheme is not used. The Diesel Engine. The Diesel engine is to-day built by a number of firms in Europe and by the American Diesel Engine 386 INTERNAL COMBUSTION ENGINES Company in this country. In all cases, the various constructions are of the vertical four-cycle type, but European builders favor the open ^1-frame, while the American makers have adopted the enclosed box frame. Figure 12-126 gives a general idea of the appearance of the American Diesel engine, while Fig. 12-127 shows the construction. FIG. 12-126. American Diesel Engine. The massive self-contained build of this machine, necessitated of course by the high pressures occurring in its cycle of operation, is very noticeable. The method of operation of this machine has already been explained in Chapter XI, to which the reader is referred. The valve construction of this engine is very simple, as shown in Fig. 12-128. The exhaust valve opens upward, the inlet valve MODERN TYPES OF COMBUSTION ENGINES ' 387 downward; both are located in a small chamber at the side of the cylinder. The fuel injection valve / is opened by the bell-crank B, which is operated by a cam on the lay shaft, and closed by a helical spring S. Surrounding the spindle of the injection valve FIG. 12-127. Cross-section American Diesel Engine. are placed the atomizing arrangements by which the oil is very finely divided through the agency of highly compressed air, as soon as B opens the valve. The stroke of the lever B is uniform, hence the injection valve always opens to the same amount and for the same length of time, whatever the load on the engine. To govern the speed, the" governor controls the stroke of the 388 INTERNAL COMBUSTION ENGINES pump which furnishes the oil to the valve. The lower the load on the engine, the later in the stroke of the oil pump does the deliv- ery of oil to the injection valve commence. One method of doing this is explained in Chapter XIV, under Details of Governors. FIG. 12-128. Valve Construction, American Diesel Engine. The American Diesel engine is built in sizes from 75 to 450 B. H. P., mostly in three-cylinder units. The Priestman Oil Engine. This English machine, Fig. 12-129,* is mentioned here because the means used for forming the combustible mixture are different from those so far described. The engine is of the horizontal single-acting type, The exhaust valve is mechanically operated by an eccentric on the shaft. The same eccentric rod also operates a small air pump e which keeps up an air pressure of from 30 to 40 pounds on top of the oil in the *Guldner, p. 118. MODERN TYPES OF COMBUSTION ENGINES 389 supply tank b. The inlet valve c is automatic. On the suction stroke the piston draws a charge of finely divided oil, mixed with air furnished by the spray maker, into the vaporizer tt, together with a large quantity of auxiliary air to form the proper combus- tible mixture. The vaporizer, heated at the start by external means, is during operation kept by the exhaust gases at sufficient heat to completely vaporize the oil before it passes out on its way to the inlet valve. Thus the mixture reaches the cylinder com- pletely prepared. The spray maker and vaporizer are explained in greater detail in Chapter VIII. FIG. 12-129. Priestman Oil Engine. The cylinder operates on the ordinary four-cycle principle; ignition is by electric spark. In the later Priestman engines a small amount of water from the jacket is admitted to the cylinder each cycle, after the principle of Banki. The Fairbanks- Morse Crude Oil Vaporizer. In the fourth type of vaporizer the oil is not preliminarily sprayed or atomized, as is done in the Priestman vaporizer, but simply vaporized by the heat of the exhaust gases. The piston then generally draws a part of the necessary air through the vaporizer and saturates this with the oil vapor. The rest is added to form the proper mixture just before the inlet valve. Of this type is the " Econo- mist" retort, already described in Chapter VIII. Another type is that made by the Fairbanks-Morse Company, and illustrated 390 INTERNAL COMBUSTION ENGINES in Fig. 12-130. G is the main vaporizer chamber. The oil pump furnishes oil through the pipe F to the reservoir R on top of the chamber. The regulating valve T returns the excess pumped to the main supply through the pipe 0. From R the oil is allowed to trickle slowly downward over surfaces heated by the exhaust gases. These come in through the pipe N, but the volume enter- ing G is controlled by the position of the valve E, which sends a part into G, the rest directly out through X, depending upon the demands of the vaporizer. Air enters the chamber through C. FIG. 12-130. Fairbanks-Morse Engine with Crude Oil Vaporizer. On its way upward it saturates itself with oil vapor and finally flows out through B to the engine. The auxiliary air supply is furnished through A. Any vaporizer of this type labors under the disadvantage, already mentioned in the case of the "Economist" retort, that not all of the oil can be utilized. As the vaporization proceeds, the oil gets heavier, the amount of vapor evolved grows less at the temperature maintained, and the useless residue must finally be drawn off. In the apparatus above described provision for this has been made through the drain cock D. W is a heating lamp to heat the vaporizer on starting. 5. THE ALCOHOL ENGINE. The points of difference between the alcohol engine and any other gas engine have already been MODERN TYPES OF COMBUSTION ENGINES 391 mentioned in Chapter VIII. In constructive details these engines do not differ from the rest except that the compression is carried higher than in other liquid fuel engines. The main feature dis- tinguishing them is in the arrangements for forming the com- bustible mixture, and the various expedients adopted have been thoroughly discussed under the head of vaporizers in the chapter mentioned. Recent experiments have disproven the old state- ment that an alcohol engine cannot be started from the cold. Much depends upon the position of the carbureter with reference to the inlet valve ports, and after paying due attention to this point, a gasoline automobile engine has been successfully operated on alcohol without change. This bears out the experience of the Deutz Company in whose alcohol engines only a spray nozzle placed very close to the inlet valve is used. An ordinary gasoline carbureter can be made to act in somewhat the same way. Ameri- can practice regarding alcohol engines is for obvious reasons somewhat behind that of Europe, but the indications are at present that this condition will not long exist. CHAPTER XIII GAS ENGINE AUXILIARIES: IGNITION, MUFFLERS, AND STARTING APPARATUS Ignition. There are four methods of igniting the com- bustible charge in a gas engine. Some of these are still in use. others belong to the period of gas-engine development. These methods are the following : 1. Ignition by an open flame. 2. Ignition by hot tube. 3. Ignition by heat of compression, and 4. Ignition by electric spark. Ignition by open flame is practically obsolete, and ignition by hot tube is also fast falling into disuse. As a matter of fact, except for the Diesel engine, the Hornsby-Akroyd and a few others, which ignite the charge or fuel by heat of com- Explosion . - , . * , Port pression, the method of igniting the charge by the electric spark is to-day the means most generally employed. i. Ignition by Open Flame. This method has been superseded probably because of its occasional failure and the obvious danger con- nected with its use under certain conditions. The simplest arrangement of this type is Barnett's ignition cock, Fig. 13-1.* This consists of a hollow plug which works in a shell having two ports, 1 and 2. The former opens to the atmosphere and communicates with an outside flame, A, the latter opens into the cylinder. The port 3 in the plug is of such a size that it may communicate either with port 1 or 2. but never * Clerk, The Gas and Oil Engine, p. 207. 392 FIG. 13-1. Barnett's Ignition Cock. GAS ENGINE AUXILIARIES 393 with both at the same time. Inside the plug is placed a gas jet as shown. Gas should be admitted to this at such a rate that the flame burns inside of the plug and not out through the ports 3 and 1. The proper size of the ports has much to do with the proper admission of air to the inside of the plug to keep the flame alive, as shown by the arrows. If now the plug is quickly turned through 90 degrees, making the port 3 register with port 2, enough air is contained in the plug to keep the flame burning during the interval, and the combustible mixture, entering the plug much as the air enters it in the first position, is instantly ignited. Here again the proper size of the ports is of importance, for if the mixture cannot circulate through the plug, it must reach the flame by diffusion. The chances are that in that time the flame has died out and ignition fails. The flame is blown out by the force of the explosion, or dies out for lack of air, but is immediately relighted by the out- side flame, A, when the plug returns to its for- mer position. Barnett's scheme was open to the objec- tion that since the flame chamber in the plug was always under atmospheric pressure, the combustible mixture, if compressed to any extent, might extinguish the flame by sudden inrush when communication was established. Several methods to obviate this were invented, notably by Otto and by Clerk. The latter differs from the former in that it will operate at higher engine speeds; Otto's scheme failing at comparatively low speeds. A very simple solution of the problem is shown in Koerting's igniter, Fig. 13-2.* The plug, b, see left-hand section, is practically a divergent nozzle. The pressure of the combustible mixture during compression raises the plug, and some of the mixture, escaping through the fine opening, c, and expanding through b, is ignited by the open flame, d. At the instant the piston reverses, the plug is forced down by the plunger, * Schottler, Die Gasmaschine. FIG. 13-2. Koerting Igniter. 394 INTERNAL COMBUSTION ENGINES a, closing both the top and the bottom openings. But the mix- ture contained in the nozzle, 6, keeps on burning until, at the instant the side openings, e, are freed, the flame strikes into the cylinder and ignites the charge. The right-hand section, Fig. 13-2, shows the ignition position. 2. Ignition by Hot Tube. The simplest form of the hot tube ignition apparatus has already been shown in Fig. 11-5 Chapter XL It consists merely of a small tube 3 or 4 inches long, of steel, porcelain, or platinum. The open end of this tube is in communication with the combustion chamber, the other end is closed. The tube is kept red hot for a certain part of its length, generally by means of a Bunsen burner. A chimney surrounds both burner and tube to prevent loss of heat by radiation as far as possible. The action of the hot tube may be explained as follows : At the end of the exhaust stroke the tube is filled with burned gases, and these are not replaced by the fresh mixture even at the end of the next suction stroke, because the time avail- able is too short for diffusion. During the compression stroke the burned gases are being compressed into the closed end of the tube and are followed up by the fresh mixture. But no explosion follows even if this mixture reaches the red-hot part of the tube as long as the velocity of flame propagation out of the tube is less than the velocity of the fresh charge into the tube. When the former becomes greater than the latter, which happens at or just, before the piston reaches the dead center, the flame shoots out and ignites the charge. It is plain, however, that in the device shown in Fig. 1 1-5 the position of the hot zone along the tube must be about right or pre-ignition may result.' Want of adjustment in this arrangement has led to the improved hot tube shown in Fig. 13-3. In this case the position of the hot zone along the tube may be varied as shown, thus giving some control over the time of ignition. GAS ENGINE AUXILIARIES 395 In order to completely control the time of ignition the or- dinary hot tube has been perfected by the addition of a so-called timing valve. In Fig. 13-4, G is the tube kept hot in the ordinary way. Timing valve E is normally kept closed by the coil spring C. At the proper time in the cycle the ignition cam (not shown), through the link, B, and the bell-crank, A-D, compresses the spring, C, and opens the valve. Ignition then ensues. The valve is kept open during the expansion and exhaust strokes. The time of ignition may be changed by changing the posi- tion of the cam. Timing valves are open to the ob- jection that they are very difficult to keep in shape under the high tempera- tures occurring. To avoid the use of the small valve in the cylinder, Koert- ing has used the scheme shown in Fig. Tube with Timing Valve. 13-5; a is an open hot tube made of porcelain. In it there is placed the small platinum tube, c. Dur- ing compression some of the mixture es- capes through c and the valve k. When ignition is desired, k shuts the exit and the flame in a strikes back into the cylinder. Hot tubes may be from two to four inches long, and from one-quarter to one- half inches internal diameter. They may be made of steel, platinum, or Porcelain. Porcelain is best because cheap and nearly indestructible by heat. The hot tube finds application in small and medium sized stationary machines only. It is fully as cheap to operate as electric ignition and just as certain. In large machines this method of ignition is not as satisfac- tory, because the ignition itself is hardly sharp enough for the large volume of gas, and 'because in many cases the distance FIG. 13-5. KoertingHot Tube Igniter. 396 INTERNAL COMBUSTION ENGINES the flame has to strike is too great. Care should be taken to place the tube at the proper point, i.e., where a good mixture at the opening of the tube is insured, and where the opening cannot be clogged by oil or water. 3. Igniting the Charge by Heat of Compression. This method of igniting the charge is practically limited to liquid fuels and is carried out in several ways. (a) Only air is compressed to a very high degree so that its temperature is high enough to ignite the fuel as it is injected at the beginning of the working stroke. This is Diesel's method. (b) The charge may be ignited by means of a hot bulb or chamber connected to the combustion chamber proper by a narrow neck or opening. There are two modifications of this method. In the one the fuel is injected into the combustion chamber on the suction stroke. During the next stroke the mixture is compressed into the hot bulb and ignites. The com- bustion, however, is confined to the bulb until, near the end of the compression stroke, the velocity of flame propagation exceeds the velocity of gases entering the narrow neck of the bulb, when the flame strikes out and general ignition ensues. The action of the bulb is therefore very similar to that of the open hot tube. The main difference is that after the bulb has been externally heated at the start, the heat of compression soon keeps the walls of the bulb at a sufficiently high temperature so that the external flame can be extinguished. In the second modification, the hot bulb or chamber is used at the same time as a vaporizer. Thus in the Hornsby-Akroyd engine, the fuel is injected into the chamber by a pump at the beginning of the suction stroke. The piston draws nothing but air on this stroke, which air is partly forced into the bulb on the return stroke. Here it mixes with the oil vapor, which formed, due to contact with the hot walls, and while combustion may ensue it cannot be general because hardly enough oxygen is present in the bulb or vaporizer. Near the end of the com- pression stroke, however, the flame strikes out, and the combus- tion becomes explosive. As in the former case, the vaporizer is heated by a lamp at the start, but after a few minutes of opera- tion the walls of the vaporizer, if well protected, remain at a dull red heat, due to the heat of compression and explosion. GAS ENGINE AUXILIARIES 397 Capitaine * employs a method which differs from that used in the Hornsby engine in that, at the moment the fuel is injected into the vaporizer, a little auxiliary air is also admitted which sweeps the oil vapor formed into the combustion chamber proper, where it meets the main body of air and is compressed with it. Ignition ensues from the hot walls of the vaporizer as in the other cases. 4. Ignition by Means of the Electric Spark. Electric Igni- tion is to-day used more than any other. In fact in some branches of the industry, automobile work for instance, it is used exclu- sively. The reasons for this are not far to seek. As compared with the hot tube, there is no flame, and no fuel required to feed it. The system is perfectly flex- ible and susceptible of perfect timing. There are a number of electric-ignition sys- tems in use, differing in their methods of wiring, their sources of current, etc., but con- sidering for the moment nothing but basic principles, all the systems may be grouped under two heads. These are: 1. Make-and- Break Ignition, and 2. Jump-spark Ignition. In what follows, only the elementary principles of electric ignition will be discussed. For a comprehensive exposition of the subject, consult "Electric Ignition for Motor Vehicles" by W. Hibbert.f 1. MAKE-AND-BREAK IGNITION. The simplest kind of make- and- break circuit is shown in diagram in Fig. 13-6. J In this figure, B is a source of current and c a so-called spark coil. In this case such a coil consists merely of a number of turns of comparatively heavy wire wound about a bundle of wrought- * Zeitschrift d. V. d. I., 1907, p. 919. f Whittaker & Co., 64-66 Fifth Ave., New York City. j Roberts, The Gas Engine Handbook. FIG. 13-6. Make-and-Break Circuit. 398 INTERNAL COMBUSTION ENGINES iron wires. This coil is in series with the circuit, and acts as an inductive resistance. When the circuit is broken, it serves to intensify the pressure, causing a hot spark at the point of break. For this reason the writer prefers to call this kind of coil an intensifying coil rather than a spark coil, which, as used in jump-spark ignition, is a very different thing. The make-and-break mechanism consists in this case of a stationary electrode, e, insulated from the rest of the engine, and a movable electrode, p. The latter is connected to a flat spring, S, which in turn is in contact with the cam, C. The current Igniter flows from B, through the inten- sifying coil, c, to the electrode, p, and from here through the electrode, e, back to 5, thus completing the circuit. The operation is as follows: Cam C, rotating in the direction of the arrow, first presses electrode p against electrode e, making the circuit. At the proper moment, spring S slips off the cam, sud- denly forming a gap between p and e, across which the spark jumps. Thig Qf make d break . . mechanism is known as the FIG. IJM.-Make-and-Break Ignition Apparatus. hammer break. An example from practice is shown in Fig. 13-7.* The source of current in this case is a Bosch magneto. The current flows from d to e, the stationary electrode, and re- turns through /, the movable electrode, and the forked rod, h. Actuated by a latching arrangement on the half-time shaft, the armature lever is pulled to the left about 20 degrees, as shown in the lower figure. This puts the two powerful helical springs shown in tension, so that when the latch releases, the armature sleeve instantaneously returns to its normal position, generating the required current by cutting the lines of force with great rapidity. At the same instant the fork, h, strikes the bell crank, g, * Giildner, Verbrennungsmotoren, p. 365. GAS ENGINE AUXILIARIES 399 thus separating / from e, and causing the spark. This method is susceptible of adjustment by regulating the time of release of the latch. It should be noted that in this particular instance the two electrodes are in contact until the spark is desired. This is ad- missible because with the source of current used, electric energy is generated only for an instant, just before the break. If a con- tinuous source of current, such as a battery for instance, is used, it becomes necessary to modify the mechanism so 'that the elec- trodes are in contact only for a short time before the break; otherwise the system would be very wasteful of current. The hammer-break mechanism is open to two objections, rapid wearing away of the points and fouling. The former is aggravated if too strong a current is used. It is usual, therefore, to make the points of contact of some metal that will not easily corrode or wear away under heat. Platinum, or plati- num-iridium is extensively used for this purpose. There are, how- ever, some special alloys on the market, such- as Baker & Co.'s "Special," which are somewhat less costly, but do the work fully as well or better. Platinum is practically indestructible by heat, but it is hardly hard enough to stand the wear. The only remedy for fouling is periodic cleaning, although the claim is made for some of the special alloys that they remain bright indefinitely. To overcome the objection of fouling, a modification of the make-and-break system known as the wipe spark is sometimes employed. In this, one electrode is made to revolve and a pro- jection on it "wipes" across the other electrode at the proper time, causing a spark on the break. The spark produced in this way is perhaps hotter than that formed by the hammer break, and fouling of the sparking surfaces is effectually prevented. On the other hand, the wear is much greater. Make-and-break ignition has the advantage that only a low voltage is required to operate it. The pressure ordinarily used is from six to eight volts, while in many cases from two to four volts is quite sufficient. There is thus much less danger from leakage of current and short-circuiting in this system than there is in the jump spark method. The disadvantages of the system regarding w r ear and fouling have been already pointed out. Another, as compared with the 400 INTERNAL COMBUSTION ENGINES jump spark, consists in the fact that some mechanically operated gearing is required to "trip" the igniter. Both this fact and the rapid wear of the contact points have led designers to adopt the jump-spark system for high speed work. Lately, however, there high-speed trip gear. FIG. 13-8. Plan, Fay & Bowen Engine. seems to be a tendency to adapt the -make-and-break system also to high speeds, caused no doubt by the very obvious disadvantages of the jump spark. There are ways of efficiently operating a One of these, used by the Fay & Bowen Engine Co. for medium high speeds, is shown in Figs. 13-8 and 13-9. Fig. 13-8 shows a plan view of the vertical engine. The igniter shaft, A, passes ver- tically downward FIG. 13-9.-Igniter Block, Fay & Bowen fa h h _ Engine. . jacket space, and is driven from the crank shaft by a pair of bevel gears. It runs in bronze bearings, and where it passes through the jacket it is encased in a water-tight tube. At the top this shaft carries a small gear which meshes with a somewhat larger gear, B, under the igniter cam, C. All the gears are cut gears and run practically without noise. The driving is positive and GAS ENGINE AUXILIARIES 401 there can be no slip. As the cam, C, revolves, it engages the plunger D and forces it back against the spring G. When the plunger slips off the cam, the hammer, E, strikes the movable electrode, F, separating it from the stationary electrode, J, caus- ing the spark. This action is carried out with the same rapidity no matter how fast the fly-wheel is revolved at the start. For the greater part of its revolution, cam C is not in contact with the plunger, D, and hence the electrodes are separated, thus main- taining an open circuit for the greater part of the time and pre- venting the waste of current. Just as soon as C commences to push back the plunger Z>, the spring H pulls back the movable electrode F, and contact is made for a sufficient length of time to insure a good flow of current. The igniter plug, shown in greater detail in Fig. 13-9, is en- tirely independent of the driving gear and is held in place by four bolts, which can be removed at a moment's notice. The seat of the plug is a ground joint. The spark points can there- fore be examined and the plug replaced in a very short time, or a new plug may be substituted for the old one. The chances for wear in the whole arrangement, however, are very small and there seems to be no reason why this igniter gear should not be used for speeds much higher than those for which the designers now use it. Adjustment of the spark in this gear is made in a very simple way by pivoting the gear B about the center of the shaft A, thus changing the position of the cam, C, with relation to the plunger. The adjustment is controlled by the hand lever shown. Should the lever by any accident be left in the advanced spark position, so that the engine may get an explosion turning it the wrong way the next time it is started, a small clutch located under the igniter cam, C, immediately frees the cam so that no second back explosion can take place. 2. JUMP SPARK IGNITION. Figure 13-10* shows diagram- matically the simplest type of jump-spark system. There are in all cases a primary or low-tension and a secondary or high- tension circuit. The primary circuit is shown in heavy line and contains the source of current, B. The current flows from B through an arrangement, T, called the interrupter, commutator, or timer, which serves to make and break the primary current at * T. H. White, Petrol Motors and Motor Cars. 402 INTERNAL COMBUSTION ENGINES the proper time. It then passes through the primary winding, P, of the spark coil and returns to the source, completing the primary circuit. The secondary circuit, shown by a light line, consists of the secondary winding, S, of the spark coil and a spark plug in the cylinder of the engine, indicated in the figure by Z. It should be noted, in connection with the secondary circuit, that this cir- cuit is never actually closed, since a spark gap always exists in the spark plug. Hence current cannot be said to flow in this circuit until the tension or voltage becomes high enough to bridge this gap by a spark. To understand the operation of the jump-spark system it is necessary first to study the action of the spark coil. There are two kinds of these coils, the non-trembler and the trem- bler coil. The former is the type indicated in Fig. 13-10. Its actual construction is about a.s follows: / is the core of the coil consisting of a bun- dle of fine iron wires. This is covered with a layer of some insulating material, io_in an d around this is wound . 13-10. Simple Jump-spark System. the primary -winding. This consists generally of several layers of insulated copper wire, about No. 20 or 22. A light layer of insulation next sepa- rates this from the secondary winding, which consists generally of some 10 to 15000 turns of very fine insulated wire. Each layer- of this wire is separated from the next by a layer of insula- tion to prevent short-circuiting under the very high pressures occurring. To understand easily what follows, it is necessary merely to remember that if any conductor of electricity is moved across a magnetic field, or if a magnetic field is moved across a conductor, an electric current will immediately be set up in this conductor. Further, that if a current be passed through a conductor, a mag- netic field will immediately be set up around the conductor. Now, in the spark coil described, a current is sent through the GAS ENGINE AUXILIARIES 403 primary winding as soon as the contact is closed at T. This converts the iron core of the coil into an electro-magnet, setting up a strong magnetic field. The magnetic lines move outward across the windings of the secondary circuit and induce a high tension in this circuit. But, owing to self-induction, the build- ing up of the magnetic field is much slower than the collapse of the field when the primary current is suddenly interrupted at T. The magnetic lines then move inward across the secondary wind- ing with much greater rapidity, hence the pressure induced is much higher than that existing during the building up of the field, and, if the spark plug is right for the coil, a spark will bridge across the gap in the cylinder, igniting the charge. The fact that FIG. 1311. Jump-spark System with Trembler Coil. the voltage induced on the making of the current is not high enough to bridge the gap prevents the occurrence of a double spark in the cylinder, which might lead to pre-ignition of the charge. Since with the non-trembler coil the current in the primary is established only once when ignition is desired, only a single spark will occur in the cylinder. It is possible that this single spark may fail to fire and a series of sparks at the time of ignition is hence an advantage. This has led to the adoption of the trembler coil, Fig. 13-11*. The circuit shown in this figure is the same as that of Fig. 13-10, except that a trembler or " buzzer " T and a timer or commutator W have been substituted for the * T. H. White, Petrol Motor and Motor Cars. 404 INTERNAL COMBUSTION ENGINES simple make-and-break mechanism, T, of Fig. 13-10. The action of the trembler is simple. As soon as W makes contact, the primary current converts the core, /, into an electro-magnet which attracts the armature, A, of the trembler blade. This action, however, breaks the primary current by pulling the spring blade, T, away from the constant screw at E. A spark then jumps over in the cylinder as before explained. The break- ing of the primary current, however, releases the armature, A,' which returns to its normal position, again establishing the pri- mary circuit at E. The operation is then repeated. This action establishes a pulsating pressure in the secondary winding, causing a series of sparks as long as contact is maintained at W. The advantage of the trembler coil has already been pointed out. The disadvantages exist in the fact that a second moving part is introduced into the primary circuit which must be kept carefully adjusted if the system is to work satisfactorily. In both Figs. 13-10 and 13-11 it will be noticed that there is an arrangement, C, called a condenser, connected across the make-and-break mechanism, in the non-trembler coil across the interrupter, in the trembler coil across the vibrator or buzzer. The condenser consists of a large number of sheets of tinfoil, the number depending upon the capacity desired. Each sheet is separated from the next by a layer of insulation, and the alternate sheets are connected together. This manner of construction is clearly shown in the diagram. The object of the condenser is to prevent serious sparking at the make-and-break contacts in the primary circuit. The reasons why such a spark occurs at all in such a circuit is that the collapse of the magnetic field not only induces a high pressure in the secondary winding, but also causes a momentary increase in the pressure in the primary, thus bridg- ing any small gap by a spark, and causing rapid wear of the con- tact points of the trembler at E, Fig. 13-11. When the primary circuit is now broken at E, the current induced, instead of jump- ing across, is expended in charging the condenser. The action is very similar to that of an air chamber on a hydraulic pipe line, absorbing shock by compressing air. The next time the primary circuit is closed at E, the condenser discharges and helps to send a current through the primary. As actually constructed, spark coils are very compact. The GAS ENGINE AUXILIARIES 405 condenser is generally placed under the coils, and the whole is enclosed in a tight wooden box. Externally nothing shows but the terminals and the trembler, if the coil is of that type. In FIG. 13-12. Three-Terminal Spark Coil. some- coils one end of the secondary coil is connected to one end of the primary winding, so that only three terminals show, as in Fig. 13-12. Fig. 13-13 shows a four-terminal coil, the secondary terminals being on top. The tremblers are of various con- structions, nearly each maker having his own design. They must give a quick break. They should be easy of fine adjustment, but the adjustment, once made, should stay. In many cases, as for automobile and FlG 13 _ 13 _ Four . Terminal Coil . marine purposes, the entire spark coil is enclosed in a second box with tight cover, so as to prevent fouling by mud or water. An example of this is shown in Fig. 13-14. Timers. A very important part of a jump-spark system is the device making and breaking the primary circuit, for every- thing depends upon the non-failing regularity of its performance. 406 INTERNAL COMBUSTION ENGINES 13-14. Dash-board Coil. There is a large number of such timers or commutators on the market, all more or less good. Figs. 13-15 to 13-18 show a few of the designs. The fundamental idea in all of these is the same. The half-time shaft actuates a cam or wiper inside of a case, which cam, at the proper time, makes and breaks contact with insulated ter- minals held by the surrounding case. The number of such terminals de- pends upon the number of cylin- ders. A great deal of ingenuity is shown in the prevention of friction between the cam and the terminal. The action of the Sintz timer, Fig. 13-15, is obvious. Here we have roller contact, the ends of the ter- minals, are hardened steel, and the case is dust proof. Of somewhat similar, design is the Lacoste timer, Fig. 13-16. The cross-section shows clearly the manner of con- struction. Somewhat more complicated, but of excellent design, is the Pittsfield timer, Fig. 13-17. In the Grouse-Hinds double ball timer, Fig. 13-18, the cam on the half-time shaft passes between two steel balls, held as shown. This makes the con- tact positive, keeps the surfaces clean, and the wear is very small. With any of the above devices, the time of sparking may be varied by shifting the terminals with ref- erence to the cam or wiper on the half-time shaft. Some timers incor- porate governors to automatically time the spark. Spark Plugs. A spark plug consists of two electrodes or sparking points which are held a certain distance apart in the cylinder. The central electrode is insulated, while the metallic FIG. 13-15. Sintz Timer. GAS ENGINE AUXILIARIES 407 jacket enclosing the insulation generally carries the other spark point. This point, therefore, can be put in the circuit by fastening a wire anywhere to the engine. The essential requirements of I FIG. 13-16. Lacoste Timer. FIG. 13-17. Pittsfield Timer. the construction are that the insulation of the central electrode be sufficient and not liable to breaking down, and that the elec- trode points be so constructed that the plug is not easily subject to fouling. 408 INTERNAL COMBUSTION ENGINES The most important part of the entire plug is perhaps the insulation of the central electrode. Among the materials used for this purpose, porcelain and mica take the lead. Por- celain, while excel- lent, is very liable to break under any uneven expansion by heat, and the insu- lation must therefore be carefully designed with this point in view. Mica is not open to that objec- tion and its electri- cal resistance is very high, but owing to its laminated struc- ture, oil or soot may after a time be forced between the laminations under Fi. -13-18. Grouse-Hinds Double Ball Timer. the high pressures existing, thus short-circuiting the plug. How various manufacturers have tried to take into account the requirements mentioned, is shown in Fig. 13-19.* The first six plugs there shown have porcelain, and the last two mica insulation. It should be remembered in connection with spark plugs, that since the electrical resistance across the spark gap is greater when in actual operation in the engine than when in ordinary T? m tr ai -y FIG. 13-19. Various Designs of Spark air, a plug may give a fair spark pi ugs< * From Romans, Automobiles, p. 284. GAS ENGINE AUXILIARIES 409 when tested in air, and may still fail in operation. It should also be remembered that heat will lower the electrical resistance of porcelain, so that when the plug is very hot, short-circuiting through the insulation may result. It has been shown by ex- periment that while the resistance of porcelain cold was about 100 megohms, this fell to 2 megohms when the plug was at a dull red heat and under this condition sparking ceased. The spark, however, was immediately restored by an external spark gap, and continued even when the resistance had fallen to 800,000 ohms. Without the external gap, if the plug was allowed to cool down, sparking recommenced when the resistance of the porcelain had again risen to 5 megohms. Auxiliary Spark Gap. As the name implies, this is a second spark gap placed in the secondary circuit outside of the FIG. 13-20. Auxiliary Spark-gap. cylinder. This gap acts like an electrical condenser, above explained. The pressure builds up on one of the terminals of this gap, until it is high enough to break through the intervening air, causing an impulse of very high pressure through the circuit, thus giving a good spark across the main gap in the cylinder. Fig. 13-20.* shows one form of auxiliary spark gap. The advantages claimed for the device are: (a) Greater certainty of sparking in the cylinder, since the higher pressure generated will cause a spark even across a par- tially fouled plug. (b) Greater life of battery, since current cannot leap across a fouled plug as long as the auxiliary gap is not bridged. (c) The sparking can be watched, since a spark across the gap always means a spark in the cylinder. * Homans, Automobiles, p. 290. 410 INTERNAL COMBUSTION ENGINES In spite of these facts the auxiliary spark gap has not found extended application. RELATIVE ADVANTAGES AND DISADVANTAGES OF MAKE-AND-BREAK AND JUMP-SPARK SYSTEMS Make and Break Ignition. Low tension throughout the cir- cuit, requiring less thorough insulation, and causing less trouble from short-circuiting. The system is electrically more simple, while mechanically it is somewhat more complex than the jump- spark system. This latter fact makes it somewhat difficult to apply to high-speed engines. Jump-spark Ignition. Electrically more complex than the other, but has no moving parts inside of the cylinder. Can be operated under very high-speeds with entire success, and has the greatest flexibility with regard to spark adjustment. Sources of Current. All sources of electrical current used for electric ignition may be classed under two heads: 1. Chemical Generators, under which come (a) Primary sources, as wet and dry cells, and (6) Secondary sources, as the storage battery or accumulator. 2. Mechanical Generators, variously called dynamos and magnetos. 1. Chemical Sources of Current. (a) WET AND DRY CELLS. All chemical cells consist of three essential parts, a positive and a negative electrode and an exciting liquid, called the electrolyte. As the name implies, in the wet cell this electrolyte is used in its liquid form, while in the dry cell it is mixed with some absorbing material, and the paste is used to fill the space between the electrodes. Take the dry cell as an example. The negative element is usually a carbon rod placed at the center of the circular case which forms the envelope of the cell. This rod is surrounded generally first by a layer of manganese dioxide, the purpose of which will appear later, and the rest of the space between this and the positive element, usually zinc in the shape of a cylinder, is then filled with the electrolyte paste, the original liquid being usually sal- ammoniac and water. The top of the cell is then covered with pitch or other substance that prevents the evaporation of the liquid in the paste, except that a small vent hole is left to allow GAS ENGINE AUXILIARIES 411 of the escape of any gas that may form within the cell due to the chemical action going on. In such a cell the current generated by the action of the electrolyte passes from the zinc to the carbon electrode, so that as far as the terminals of the cell are concerned, the carbon is the positive terminal. The chemical action destroys the zinc after a time and produces hydrogen gas on the carbon element. The greater the amount of this gas deposited on this element, the slower the generation of current, so that it may finally cease altogether. The cell is then said to be polarized. In dry cells the gas is taken care of in two ways; the vent hole in the top allows some of it to escape, while the layer of man- ganese dioxide above mentioned absorbs another part. But it is a fact that by these means not all of the gas is rendered harm- less and hence the cells will polarize with more or less rapidity. This merely means that if current is drawn from them continu- ously for any considerable length of time, their strength will fail, making the cell appear dead. The same reasoning applies to wet cells where the hydrogen is allowed to escape through the liquid. Now, assuming that the zinc is not yet destroyed, if a cell so polarized is allowed to recuperate, it will again attain nearly its normal strength and may be used as before. The cells are said to be adapted to "open-circuit work.' 7 From all of this it is quite evident that in places where the requirement for current is not very great and, above all, not continuous, the primary cell will give satisfactory service. But where the draft of current is nearly continuous, as in high-speed four-cylinder machines for instance, the cell will rapidly polarize and soon fail to give suffi- cient voltage to operate the spark coil. The average size of a dry cell is about 2%" x 7" '. It will give when fresh from 1.3 to 1.5 .volts and from 12 to 15 amperes. It should be understood that there are other combinations of electrodes and electrolytes which may be used to generate cur- rent. Thus the so-called soda-cell is made up of a zinc plate and a copper-oxide plate with a caustic soda solution as the electrolyte. (b) STORAGE BATTERIES OR ACCUMULATORS. A storage cell, like a primary cell, consists of two electrodes dipped in an elec- trolyte, but contrary to the primary cell, it cannot give off elec- trical energy in its original state when the circuit is closed. It is necessary to charge a storage cell before it can return electrical 412 INTERNAL COMBUSTION ENGINES energy on the discharge. The charging action causes chemical changes in the material of the electrodes and in the electrolyte. The energy so rendered latent is nearly all restored, when, after the charging current is disconnected, the outside circuit is closed. Chemical changes, producing a current in the reverse direction, then take place in the cell, which return both the electrodes and the electrolyte to their original condition. Some exhausted primary cells may be partially restored by passing a current through them in the reverse way, but in most cases the trouble is not worth while. The possibility of a nearly complete re- generation of a storage cell is the chief difference between it and a primary cell. There are a number of materials which can be used as elec- trodes and electrolytes, but the usual type of storage cell to-day is that using some lead compound for the former and sulfuric acid and water for the latter. Hence only this lead storage cell will be here considered. In its modern form, both the positive and negative plates of a cell consist of cast grids of lead, to which antimony is sometimes added to stiffen them. The perforations in the positive plate are first filled with some compound of lead, as Pb 3 O 4 , which is after- ward converted to peroxide of lead, PbO 2 . Similarly the negative plate is filled with PbO which is afterwards converted into spongy metallic lead. A number of plates so prepared are then placed side by side in a glass jar, or if the battery is to be used for auto- mobile work, in a vessel of hard rubber or of wood lined with rubber or lead. Positive and negative plates alternate, and all the plates of like kind are connected together. There should always be one more negative than positive plates so that each side of, each positive plate shall face a negative plate. The arrangement of plates presents a large plate surface in a compact space. Suitable insulation separates the plates from each other and keeps them from touching the bottom, in order to prevent any short-circuiting by contact or by dipping into any sediment that may form. In automobile batteries, the top is enclosed to prevent the spilling of the electrolyte, and nothing shows but the two terminals and an opening for filling. This is usually kept closed by a rubber cork with a small vent hole to allow of the escape of any gases that may form. GAS ENGINE AUXILIARIES 413 The chemical reactions that occur during charging and dis- charging are not yet fully understood, but it is agreed that the main action is the formation of lead peroxide on the positive and metallic lead on the negative plate during charging, and the for- mation of lead sulphate on both plates during discharge. The action is best explained by the following diagram.* Discharging >- Charged Condition Discharged Condition + Plate, Electrolyte, - Plate, + Plate - Plate PbO 2 +2H2SO 4 +Pb = PbSO 4 + 2H 2 O + PbSO 4 + electric energy < Charging Such a lead cell when fully charged should show a voltage of from 2.2 to 2.25 on open circuit, and from 2.10 to 2.15 when the engine is running. The voltage soon drops to 2.0 and then slowly to 1.8. Three-quarters of the total discharge takes place between the latter figures. It is usual to discontinue discharging a cell when the voltage has reached 1.75. Beyond this point the formation of the insoluble lead sulfate becomes troublesome and, discharging much below this figure, the cell may be destroyed or at least seriously impaired. Rating of Storage Batteries. The amperage of storage cells depends on the weight of material in the cell converted by the chemical reactions, while the rate at which electrical energy can be taken off depends upon the surface of the active materials exposed to chemical action. Cells are rated by their ampere- hour capacity and nearly every maker states the normal rate of discharge recommended. For ordinary constructions the nor- mal discharge rate is about .04 ampere per square inch of total positive plate surface, and the discharge capacity about 4 ampere- hours per pound of plate, including negative and positive. In order to be able to compare different cells, the capacity rating is based upon a current that will cause the voltage of the cell to fall to 1.75 volts in eight hours. Thus, if to produce this result a current of say 25 amperes must be drawn, the capacity of the cell is said to be 8 x 25 = 200 ampere-hours. If the rate of dis- charge is faster than this, it is obtained at the expense of capacity. Thus if a current of 40 amperes were drawn, the capacity might- * International Library of Technology. 414 INTERNAL COMBUSTION ENGINES be only 160 ampere-hours. Conversely, if the rate of discharge is slower than the standard, the limiting voltage of 1.75 may not be reached for say twelve hours instead of eight. These varia- tions depend largely upon the make of cell. Charging a Cell. In charging a cell it is absolutely neces- sary to determine the polarity of the terminals of the source of current. The positive terminal must be connected to the posi- tive terminal of the cell. The charging rate of lead cells should be about the same as the normal eight-hour discharge rate. It is, however, possible to use smaller currents for a longer time. The voltage of the charging current must be somewhat greater, from 5 to 10 per cent, than the discharge voltage, on account of the internal resistance that must be overcome. In one charging test, the charging voltage rose from 2.05 to 2.15 at the end of two hours, to 2.20 at the end of six hours, and to 2.50 volts in eight hours and forty-five minutes. The rate of charging was thus about normal. If charging is continued beyond this point, the electrolyte will have the appearance of boiling, owing to the gas that is being evolved. Slight overcharging will not injure a cell, but a large amount of it leads to sulfating and permanent injury. Testing of Storage Batteries. Two tests may be made, one for voltage, the other for sparking. For the former a low- reading voltmeter, to 3 volts, is connected across the terminals of the battery, while the engine is in operation. The reading should be above 1.75 volts. Any cell may give 1.9 to 2 volts on open circuit, even if completely run down a short time before. The sparking test is made to determine in a way the state of the charge by noting the kind of spark. This test should be care- fully done and not repeated too often. It is a dead short-circuit method and therefore not good for the cell. The use of an ammeter is for that reason not recommended, as it would take too long to get a reading. The sparking test is made by placing one skinned end of a piece of insulated copper wire in contact with one end binding post of the battery, and then drawing the other end rapidly across the other post. The spark should be loud and snappy. Any storage battery should last from three to four years if properly treated. It is well to adopt a regular charging period, GAS ENGINE AUXILIARIES 415 say once in three weeks for the ordinary automobile battery, whether the battery is run down or not. 2. Mechanical Forms of Generators: Dynamos and Magnetos. - Mechanical forms of current producers have the advantage over primary and secondary batteries in that the energy required by them is derived directly from the engine they operate. Hence current will be produced as long as and only when desired. The other forms of generators depend upon sources of energy entirely extraneous to the engine plant, and the supply of current is there- fore not in any sense automatic, which would be the ideal condi- tion. The terms dynamo and magneto have been variously used. Some writers designate by " dynamo " any generator hav- ing electro-magnets serving to establish the magnetic field, and by "magneto" any machine employing permanent magnets for this service. Others define the difference as existing in the kind of current produced, a dynamo furnishing direct, i.e., continuous current, while a magneto produces alternating, i.e., pulsating, current. Whatever definition is adhered to, it should be remem- bered that in either machine the current is produced by an electrical conductor cutting the magnetic field. The current is pro- duced in exactly the same way, and for exactly the same reason, as that established in the secondary winding of a spark coil, as explained above. In this case the conductor of electricity is wound upon a piece of metal, called an armature, which is rapidly rotated in a magnetic field. It makes no difference whether this field is produced by permanent magnets or by electro-magnets. If there are a number of such conductors upon the armature, and the current induced in each is properly collected by a so-called commutator upon the armature shaft so as to be practically con- tinuous in its flow through the external circuit, we have what is generally called a dynamo. On the other hand, if the current in the external circuit rises to a maximum value and then dies out to give a maximum value next in the opposite direction, the machine is generally known as a magneto. While in all dynamos and most magnetos the armature constantly rotates in one direc- tion, it should be stated that in all magnetos this is not at all necessary. Thus in the Simms Bosch magneto, the armature is stationary, and only a sleeve surrounding the armature is rapidly oscillated in the magnetic ' field, It would be beyond the scope 416 INTERNAL COMBUSTION ENGINES of this book, however, to discuss all the possible modifications, and the reader is hence referred to the works upon this subject.* In general, the small dynamo used for ignition purposes is driven by means of a friction wheel from the fly-wheel of the engine. There is then no current available from the dynamo when the engine is started, and it becomes necessary to use a battery of some kind for the first minute or two, switching in the dynamo when it is up to speed. This scheme has the disadvantage that the battery is sometimes left in the circuit and the dynamos have been known to burn out under excessive engine speeds. A device called the Auto Sparker, Fig. 13-21, overcomes these difficulties. This little dynamo is fitted with a centrifugal governor which controls the position of the friction wheel on the fly-wheel rim, so that even at starting the ar- mature rotates rapidly enough to furnish start- ing current. This does away with an auxiliary battery. As the engine speeds up, the governor FIG. 13-21. Auto Sparker. of the dynamo acts to keep the armature speed constant, independent of the diameter of the fly-wheel or the engine speed. By adjusting the governor tension spring, it is possible to control the speed of the dynamo to get any current between 1 and 3 amperes and any voltage between 3 and 10 volts. Regarding magnetos, the following description of the action of a magneto, together with the explanation of the method of connecting it up, is taken from a catalogue of the Holley Bros. Company of Detroit. For clearness and simplicity this descrip- tion can hardly be improved upon. "A magneto, so far as its essential parts are concerned, is a very simple thing. It consists of a U-shaped piece of special steel, which is permanently magnetized; in other words, a com- mon horseshoe magnet and a rotating armature. The armature consists of a soft iron core of approximate H cross-section as * W. Hibbert, Electric Ignition for Motor Vehicles. GAS ENGINE AUXILIARIES 417 viewed along the shaft upon which it is supported and on which it is designed to rotate. The magnet, to the free ends of which are affixed soft iron arc-shaped pole pieces, and the armature core with the sides of the H correspondingly arc shaped, is shown in vertical section in Fig. 13-22. In the slot formed in the armature N N_ FIG. 13-22. core by the sides of the H, wire is wound in turns lengthwise of the armature shaft. So much for the construction of the ele- mentary magneto. In order to understand how it generates in its armature, when turned, an electric current, it is necessary to remember one law of physics, namely: Whenever a wire is wound about a magnetized soft iron core and the magnetism of the core suddenly dies out, there will be a tendency for a current to be 418 INTERNAL COMBUSTION ENGINES produced in the wire. A familiar example of the working of this law is found in the operation of the common jump-spark coil. Here we have a core made of soft iron wires and around it is wound a great many turns of fine wire, the ends of which are con- nected to a spark plug. The core is also wound with a coil of wire which is supplied with current from a battery, and when this current is flowing the core is magnetized. When the current from the battery is interrupted, the magnetism in the core sud- denly dies out, and, in accordance with the law above stated, a tendency is created for a current to flow in the fine wire coil which is connected to the spark plug and this ' induced' current jumps at the plug. "In order to explain how the iron core of the magneto arma- ture with its winding is magnetized and how the magnetism of the core is caused suddenly to die out, it is necessary to refer to four diagrams of Fig. 13-22, showing the armature in different positions of rotation with respect to the pole pieces. In diagram (I) the armature is represented with the two heads of its core in close proximity to the faces of the pole pieces. The space be- tween the pole pieces is thus almost -completely filled or bridged with iron, and magnetism passes from one pole piece to the other through the armature core, thoroughly magnetizing it. Next consider diagram (II). Here the armature is shown rotated into such a position that one edge of each pole of the armature core is just leaving the vicinity of one of the pole pieces. As soon as this position is passed, the space from pole piece to pole piece is no longer filled with iron, but with air which is not a conductor of electricity. Thus very little magnetism passes from one pole piece to the other and the core is no longer traversed by the magnetic influence and suddenly ceases to be mag- netic. This is exactly the condition prescribed by the above quoted law for the production of a current, and, in fact, when the armature in its rotation leaves position (II), there is a sudden impulse of current produced in 'the wire of the armature which dies away after the armature rotates a little beyond this position. In position (III), the conditions of armature magnetization exist- ing in position (I) are reproduced, except that the armature has changed ends in respect to the pole pieces and the magnetic in- fluence passes through it in the opposite sense, charging it oppo- GAS ENGINE AUXILIARIES 419 sitely, so that when the magnetism is discharged in position (IV) the current will be in the opposite direction through the wire of the armature winding. As the armature is turned upon its shaft, there are thus produced, in each complete rotation, two rather short impulses of current of opposite direction nearly correspond- ing with the instants at which the armature heads, so to speak, 'part company' with the pole pieces and are half a revolution apart. During the remainder of the rotation there is no current flowing. It may be readily seen that by connecting one end of the armature wire to the armature core, and by connecting the other to an insulated metallic contact segment, carried by the armature shaft, upon which bears a stationary insu- lated brush, the current impulses may be taken from the magneto for use. "Now as to the practical use of such a magneto for ignition purposes. Since it is only during a small part of the armature rotation that current is being generated, it is necessary to rotate the armature shaft at such a speed that these electrical impulses shall be so timed as to correspond with the periods when ignition is required by some one cylinder of the engine. If this were not attended to, the ignition periods of the engine might occur during the parts of the armature revolution, when no current was being produced. In order to bring about this result, the magneto and the engine must, at all times, run at a properly proportioned ratio of speeds and the positions of the engine, crank shaft, and armature must be adjusted right in the first place. If the magneto shaft is geared to the engine at the right ratio and the teeth of the two gears are correctly meshed, the desired re- sult will be brought about. For instance, if the engine be of the four-cylinder, four-cycle type, four sparks will be required for each two crank-shaft rotations. Four sparks will be produced for each two revolutions of the magneto, as well, and thus, if the magneto and the engine run at the same speed, the sparks will be numerically correct. If geared to the crank shaft, the crank- shaft gear and the magneto gear would have the same number of teeth, and if driven from a two to one shaft, the number of teeth in the two to one shaft gear would be twice as great as the teeth of the magneto gear. By changing the particu- lar teeth of one gear which are in mesh with certain teeth 420 INTERNAL COMBUSTION ENGINES i + of the other, the current impulses may be made to occur at the moments when the pistons are exactly in the firing positions." + In variable-speed engines, as automobile ma- chines, for instance, the service required of the ignition outfit becomes more exacting as the speed increases, owing to greater compression and less available time. This in the case of mechanical current generators is met by a natural increase in voltage with increase in speed, which constitutes another advantage of this type of generator as compared with primary and secondary cells. Thus less hand manipulation of the spark is required, but all magneto systems should be provided with means of altering the armature position relative to the crank -shaft position in order to alter the time of spark. METHODS OF CONNECTING UP PRIMARY AND SECONDARY BATTERIES, AND SYSTEMS OF WIRING USED. Primary and secondary cells may be connected in series, in parallel (or multiple) and in multiple-series. The meaning of these terms is explained in Fig. 13-23. For series connec- tions, Fig.- 13-23a, each positive element of one cell is connected to the negative element of the next, leaving free the negative element say of the first cell and the positive element of the last for connection to the outside circuit. In the second or multiple method of connection, Fig. 13-236, all the like elements of the cell are connected together. Fig. 13-23c finally shows six V J / cells in multiple series, i.e., three each are con- I nected in series, and these two sets in multiple I or parallel. To compute the voltage and am- perage that each one of these combinations will furnish to the outside circuit, let N = number of cells in the combination V = voltage of one cell, and A = amperage of one cell. GAS ENGINE AUXILIARIES 421 Then the following formulae will give the desired information: Kind of Combination Series Multiple Multiple-series Voltage of set NV V NV Amperage of set A NA 2A The ordinary dry cell, as stated, furnishes about 1.5 volts and 12 to 15 amperes. A make-and-break circuit should operate properly on about 8 to 10 volts, hence from 5 to 7 dry cells in series are required for this service. As far as jump-spark systems are concerned the following table gives pressures and currents required to operate some of the well-known spark coils, together with other interesting information.* From this table it is clear that from 4 to 6 dry cells in series are sufficient to operate most jump spark coils: Volts Amps. Vibration per sec. of j Trembler Prim. Res. Ohms. Sec. Res. Ohms. Kingston 89 Apple 5.20 2.2 94 .171 3715 Guenet 3.7 1.31 111 .300 2337 Guenet 5 8 123 2337 Hardv 3 78 1 05 122 274 2779 Fisher 5 8 82 149 613 2590 Dow 3.84 .57 149 .210 5394 Lacoste 3.72 1.46 177 .232 2006 Lacoste 5.62 1.94 197 .232 2006 Heinze 3.66 1.31 210 .320 1302 Pittsfield 228 Induction Coil Co 362 1.55 360 .312 6180 Milwaukee 390 .312 6180 Turning next to the systems of wiring used, all make-and-break systems are low-tension systems, i.e., the voltage does not gen- erally exceed 8 to 10 volts. Fig. 13-24 f- shows such a system in diagram with a magneto as the source of current. The circuit is easy to trace. One side of the electrical conductor on the armature is grounded, that is, connected to the engine frame *H. G. Chatain in the Automobile, July 18, 1907. f The following three figures ar.e from an article by C. B. Hayward in the Automobile, April 4, 1907. 422 INTERNAL COMBUSTION ENGINES through the armature shaft and the frame of the magneto itself, as shown at G^. The other end sends its current to one electrode FIG. 13-24. Simplicity of the Wiring of Low-tension Systems. of the make-and-break mechanism at A. When the commutator or timer, J5, makes contact, current flows, the circuit being com- pleted by grounding B, as shown at G 2 . Jump-spark systems are called high-tension systems, but a distinction should be made depending upon whether high or low FIG. 1325. Wiring Diagram, " High-tension with Coil System." tension magnetos are used. With a low-tension magneto it be- comes necessary to use the ordinary spark coil, and hence this method is sometimes called the high-tension with coil system. Fig. 13-25 shows the wiring for such a system, A being the con- GAS ENGINE AUXILIARIES 423 denser and B the primary and secondary windings of the spark coil. It is comparatively easy to trace out the complete primary and secondary circuits, if one takes into account the proper ground returns. The true high-tension jump-spark system differs from the above in the fact that the high-tension magneto embodies the secondary winding and the condenser of the spark coil. Hence the use of a separate spark coil is avoided. In Fig. 13-26 the two windings are indicated on the armature, but the condenser is shown at one side for the sake of clearness. This diagram shows the wiring for four plugs. FIG. 13-26. Wiring Diagram of True High-tension System. HIGH-TENSION DISTRIBUTOR. With the ordinary system of jump-spark ignition, as many coils as there are cylinders are required. It is possible, however, by placing a distributor in the high-tension side to serve a number of cylinders from one spark coil. The advantage of such a system is obvious, although it is bought at the cost of placing a make-and-break mechanism under very high voltage. The difficulties inherent in this, however, have been fairly successfully overcome. The difference in the wiring is made clear by Fig. 13-27 and Fig. 13-28,* both applying to four-cylinder engines. The former shows the four-part timer connected to the four spark coils serving the plugs S P^ to S P 4 . In Fig. 13-28, a high-tension distributor D connects the high- tension side of the single coil first with one plug, then with another as may be required. In practice the high-tension distrib- * Both from Hibbert, Electric Ignition for Motor Vehicles. 424 INTERNAL COMBUSTION ENGINES utor D and the primary commutator or timer, C M, are com- bined in one device. Fig. 13-29 shows the Grouse-Hinds Double Ball Contact distributor and Fig. 13-30 the Leavitt distributor. SP, Sfc, SP 3 FIG. 13-27. From "battery FIG. 13-28. The essential thing in high-tension distributors is that serious sparking in the high-tension side must be avoided. For that reason, in most of the devices the primary commutator does not establish the current until the high-tension distributor is in con- GAS ENGINE AUXILIARIES 425 tact with its proper plug segment and the primary current is broken before the contact in the high-tension side is over. To quote from the description of the Grouse-Hinds device: " The principle is exactly the same as that of the commutator already described, Fig. 13-18. The distributor has two cams and two sets of ball contacts, one set for the timer and the other for the distributor, the only difference being that in the latter the balls in each contact are about three-eighths of an inch apart and the cam insulated from the shaft. The connection is FIG. 13-29. Grouse-Hinds Distributor. made and the circuit is closed for each cylinder as the cam passes between the balls." The following is a description of the Leavitt distributor as given in a catalogue of the Uncas Specialty Company: "This device, two views of which are shown above, consists of a cylindrical casing, A, of hard rubber, into which is let a metal plate, B, at one end, and which is covered by a hard rubber cap, C, at the other end. Upon a ball bearing in the end plate, B, is mounted a driving sleeve, D, designed to be secured upon an extension of the cam shaft, and carrying fast upon it contact blocks, E, E, E, E, which make contact, successively, with the primary ball contact terminal, F. Thus far the device is identical with the ordinary timer. The commutator portion is located 426 INTERNAL COMBUSTION ENGINES at the opposite end of the cylindrical casing. The latter is en- larged at that end, and into the radial wall between the two cylindrical portions are clamped four flat-head studs, G, G, which serve as binding posts for the spark plug connections. Into the end of the metallic sleeve, D, is fastened a hard rubber stud, H, which at its outer end carries a radial arm, /, which is of metal with a relatively wide contact shoe at its end, which when the sleeve, Z), revolves, passes over the four contact studs, G, thus conducting the current succes- sively to the four plugs. The current is conducted to the ro- tating arm by a central ball con- tact, J, secured into the cap, C. When the hard rubber casing is moved around its axis by means of the arm, K, to vary the time of spark, both primary and secondary contacts are equally displaced." Mufflers. A muffler is an essential part of a gas-engine installation if quiet operation is desired. The sudden release of a body of gas at a pressure nor- mally of 40 pounds per square inch above the atmosphere causes a sharp noise very annoy- ing in the long run. A muffler is merely an enlargement in the exhaust pipe to allow of gradual expansion of the escaping gases. Many different schemes are used. Thus in some cases the muffler is merely a cast-iron pot or vessel of suitable vol- ume, in other cases the muffler is of more elaborate construc- tion consisting of a vessel filled with baffles or partitions in various ways and intended to expand the gas gradually and to break up the sound waves. Besides efficiency as a dampener of noise there are two other points that should be kept in mind FIG. 13-30. Leavitt Distributor. GAS ENGINE AUXILIARIES 427 with regard to mufflers, absence of any serious back pressure and durability. The increase in back pressure caused by a muffler depends upon the volume of the muffler and upon the amount of choking caused by the baffles. The minimum volume of the muffler should be at least five times the cylinder volume, but for com- plete silencing twice this volume is none too much. Outside of the plain cast-iron muffler pot, it is probably safe to say that nearly all baffled mufflers increase the back pressure somewhat. This fact is conceded by most manufacturers in that they furnish a cut-out which is called into service when the engine is to be called upon for a hard pull. At least one manufacturer, however, claims the production of a slight vacuum between muffler and engine due to the ejector action of the muffler. There is little doubt that cast iron is the best material to use for mufflers, as it is least attacked both by heat and the action of gases. This is especially true if a spray is used in the exhaust pipe for the purpose of cooling and condensing the exhaust gas. Many mufflers, however, for the sake of lightness and ease of manufacture, are made of galvanized sheet steel and give quite satisfactory service. The noise of the air rushing into the inlet pipes of an engine is also sometimes very annoying and in some cases may cause undesirable vibrations of doors and windows, and even walls, of the building. In such a case it is usual to muffle also the inlet pipes, and one of the best ways to do this in small and medium sized engines is to take the air from the hollow sub-base. In some yery large engines the proper silencing of the intake and the exhaust becomes quite a serious problem, as the ordi- nary muffler would become very large. The expedient sometimes used is to draw the air through an underground masonry duct of ample size leading in from the outside of the building, and to discharge through a similar duct into a chamber from which the gases finally escape. A spray of water into the exhaust pipe of such engines, close to the exhaust valve, helps materially. Figures 13-31 and 13-32 * show two types of muffler some- times used. In the first the stream of gas is merely divided, in the second each division is furnished with an enlargement de- * Mathot, Engineering Magazine, July, 1907. 428 INTERNAL COMBUSTION ENGINES FIG. 13-31. signed to decrease the gas velocity still further. Fig. 13-33 shows the Powell muffler. The partial section shows how the gas is broken up into many fine streams by passing it through perforated plates. In the so-called ejector muffler, Figs. 13-34 and 13-35, made by the Motor & Mfg. Works of Geneva, N. Y., a part of the gas passes straight out through a central pipe. The rest is made to pass a series of perforated cones as shown. It is claimed that the central pipe acts as an ejector serving to draw the gas through the cones, thus eliminating back pressure and even creating a vacuum ahead of the muffler. Starting Apparatus. The best way of starting small engines, up to say 10 to 12 horse-power, is to turn the fly-wheel over by hand either in the direction of normal rotation until the engine picks up, or, after a charge has been drawn in, by turn- ing it in the opposite direction against the compression and then snapping the spark by hand. In starting an engine in this way, it is essential to make sure first that the time of sparking is rather late, otherwise the engine may "buck," which may pos- sibly lead to an accident to the person starting it. As the size of the engine increases, however, the manual labor involved in the above scheme soon becomes too great and other means had to be devel- oped. These may be grouped under several heads. It should be noted that none of these are quite able to start an engine under load. (a) Starting crank. FIG. 13-32. ENGINE AUXILIARIES 429 (6) (d) Starting by smaller engine or other source of power. Starting by mixture. Starting by compressed air. (e) Electrical starters. Of the methods above mentioned, (a) is nearly universally used for engines up to 15 to 20 horse-power; beyond this, starting by compressed air, method (d), is generally employed. FIG. 13-33. Powell Muffler. FIG. 13-34. Ejector Muffler. FIG. 13-35. Ejector Muffler. (a) Most starting cranks are so arranged that, when turned in the direction of rotation of the engine, they grip the shaft. As soon as the first explosions accelerate the shaft so that it turns faster than the crank is being turned, the latter is released. This scheme does not prevent the crank from "kicking" back into the starter's hand if the spark should happen to be early, and many accidents have resulted therefrom. A crank which avoids this drawback is shown in Fig. 13-36. The following description of this device is from the Horseless Age, March 7, 1906: 430 INTERNAL COMBUSTION ENGINES "A bushing A, having a thread of exceedingly high pitch cut on its interior surface, contains a threaded sleeve B which serves as bearing for the shaft of the starting crank C. Rigidly secured to the shaft of the crank C are a ratchet wheel D and a ratchet cam E, the latter adapted to engage with the ratchet E' on the motor shaft. In an extension of the sleeve B are a set of spring press pawls F which are adapted to engage with the ratchet wheel D. " The device is mounted at such a distance from the end of the crank shaft that the ratchet cams E and E' cannot engage un- less the sleeve B is screwed to the limit of its motion into the bushing A by means of the hand wheel G. In starting the motor, the device being in the position shown in the as- sembly view, the action is the same as that of an ordinary starting crank. When the motor runs up to speed, owing to the pressure between the cam surfaces of ratchet cams E and E', the starting spindle is forced back into the sleeve B, and the ratchet cams E and E' are thereby dis- engaged. However, if the motor should kick back, the pawls F would engage into the notches in the ratchet wheel D, and the sleeve B would be rotated and draw the ratchet cams E and E' out of mesh. The pitch of the thread on the sleeve B is so steep that a very slight rotation of the sleeve in the bushing will carry it back far enough to pull the starting spindle out of engagement with the crank shaft. The pawl and ratchet mechanism is completely enclosed by a lateral extension on rHE HORSELESS AGE FIG. 13-36. Starting Crank. GAS ENGINE AUXILIARIES 431 the bushing A and an end plate bolted to the extension of the sleeve B." (b) If the plant has other engines in operation or if there is a shaft in operation it is quite easy to transmit the motion to the engine to be started. Some large engine installations were equipped with smaller engines the shaft of which carried a pinion which meshed with a ring of teeth on the fly-wheel of the large machine. When the large engine picked up its cycle, the small engine was automatically put out of mesh with the wheel. The cost of fitting a large engine for starting in this matter, however, is considerable. Hence this method has been largely replaced by the use of compressed air. (c) Starting by means of the fuel mixture is done in various ways. The scheme appears to be reliable in the case of engines using illuminating gas; for other gases it never was in any ex- tended use and in fact is to-day nearly obsolete. Besides the fact that if the first charge should fail there was generally not enough left for a second trial, the storing of an explosive mixture is obviously attended with some danger. In nearly all cases of starting by the fuel mixture, the engine crank is put a few degrees above center, i.e., well in the beginning of the expansion stroke: One method is to charge the com- pression space with air at atmosphere pressure, and then to open the gas cock. The mixture formed is allowed to escape through a special check valve and is ignited by a Bunsen burner just at the orifice. When the issuing flame burns reddish, showing that the mixture is rich, the gas valve leading into the cylinder is suddenly closed off. The flame at the orifice of the check valve strikes back into the combustion chamber and explodes the mixture remaining. The pressure generated closes the check valve and drives the piston forward. This is the original method of Clerk and of Green. The pressure generated behind the piston in the above manner is not very high, and the velocity attained therefore correspond- ingly low. The next step in the development was then to com- press the charge before ignition. The Clerk pressure starter and the Clerk-Lanchester high-pressure starter are of this type. Fig. 13-37 * shows the latter gear. The method of operation is as * Clerk, The Gas and Oil Engine. 432 INTERNAL COMBUSTION ENGINES follows: As the engine is slowing down for a stop with the main gas valve closed off, valve W is opened. This draws fresh air into the chamber D and the pipe D' through the valve Y. When next the engine is to be started, the gas cock F is opened. Gas flows into D and D' and forms a combustible mixture. Through a cock on the cylinder, or a slightly open exhaust valve, a part of the mixture is allowed to fill the combustion chamber through the valve W, while another part escapes through a check valve Y and is ignited by the Bunsen burner X. As soon as the mixture is right, F is closed off, the flow stops, and the flame strikes in through Y into D. The pressure generated closes Y. The flame FIG. 13-37. Clerk-Lanchester Starter. and pressure wave travels around into D f , compressing the still unburned charge ahead of it and finally ignites the charge so compressed in the combustion chamber A. The starting pressure so obtained is about twice that obtained by the method first described. Besides these flame starters, of which there are a number of modifications, other methods are in use. Thus in some engines a mixture is pumped under more or less pressure by hand into the combustion chamber, and ignited either by snapping the spark, or, as it is done by one manufacturer, by suddenly lighting an ordinary parlor match, the head of which is allowed to slightly project into the mixture. The device for doing this is quite simple but the method is not much used. Another scheme is to charge a pressure tank with the combus- GAS ENGINE AUXILIARIES 433 tible mixture. This is generally done by means of a special valve which, when the engine is in operation, opens somewhere along the compression stroke of the engine, and allows it to compress some of its charge into the tank. The starting is then done either by merely filling the combustion chamber under low pressure from the tank and igniting the mixture by suitable means, or by opening the tank valve wide, utilizing the pressure of the com- pressed mixture to start the engine, igniting the charge on the next compression stroke in the regular way. This scheme has the advantage that if the first trial fails there is generally enough in the tank for several more attempts, but the storing of any con- siderable quantity of mixture under pressure is not to be recom- mended on account of danger of explosions in the tank. Finally, at least one large German engine has been started by means of the mixture by first placing the crank in the proper position, and then blocking the fly-wheel by a plug of suitable material. A small gas engine is then started by hand and allowed to fill the combustion chamber of the larger engine under pressure in the same way as the tank was charged in the previous method. When the desired pressure is reached, the mixture is fired by hand manipulation of the spark, the impulse of the explosion breaks the plug holding the wheel and the engine picks up the cycle in the regular way. (d) To-day the starting of internal combustion engines of any size is generally done by compressed air. The latter may be obtained in various ways. In some of the smaller engines, when the engine is to be shut down, the fuel valve is closed and the engine as it slows down is allowed to compress the air drawn in into the tank. Some other installations have a hand-operated air compressor for charging the tank, or the compressor may be belt-driven from the engine for a few minutes. It is hardly neces- sary to say that all tank and pipe connections must be absolutely air tight, because the failure of the air pressure in a plant of some size would cause serious delay. For engines of the largest ca- pacity, it is usual to employ air compressors completely indepen- dent of the engine, thus obviating any failure due to tank leakage. The method of starting by compressed air is to set the engine beyond the center and then to give an impulse through the start- ing valve. In some engines- this valve is hand operated, in 434 INTERNAL COMBUSTION ENGINES others it is operated by the valve gear. Again in some, especially single-cylinder, engines the valve-actuating gear is not changed at all, except perhaps to relieve the compression somewhat, while in some multi-cylinder engines the gear on one cylinder is changed to make this cylinder act on the two-cycle principle, i.e., so that it can take an air impulse every turn, until the other cylinders have started regular operation. In any case it is best to retard the spark somewhat for starting. The starting pressures employed vary from 100 to 150 pounds, FIG. 13-38. Air-starting Apparatus. depending upon how the air is compressed. In general one or, at the most, two impulses are sufficient to start a machine. The following description of Fig. 13-38, given by F. E. Junge in Power, April, 1906, shows the method of starting a large engine by compressed air: " In this diagram A is the valve controlling the flow of air from a separately driven air compressor to the tank and B a similar valve in the pipe connecting the tank to the engine. Both valves are mounted on one pillar, which also has screwed on top of it a gage indicating continuously the pressure in the tank. Regulation of the supply of compressed air to the engine cylinder is effected by means of an automatic spring-loaded inlet poppet GAS ENGINE AUXILIARIES 435 valve, the stem and disc of which may be released or held fast by screwing down or unscrewing the hand wheel C at the engine end of the air pipe. A plug valve D is inserted in the air pipe immediately ahead of the inlet valve. Before starting the engine, the fly-wheel is turned into such a position that the crank is about 30 degrees above the inner dead center. The starting gear is adjusted so as to open both the inlet and the exhaust valves at the proper moments, the action being such as to allow part of the compressed air to escape during a fraction of the return travel of the piston and thereby reduce the compression to about 28 pounds per square inch for rich gases, and about 50 pounds per square inch for poor gases. In the meantime, the electric ignition device has been automatically adjusted so as to retard ignition for the first few strokes. .The main fuel or gas valves must, of course, also be set so as to produce the most favorable mixture for starting conditions. " To start the engine, the air stop valve B is opened, the auto- matic inlet valve released by screwing down the hand wheel C to the full extent, and compressed air is then admitted by turning the handle D 90 degrees. The piston will then begin to travel slowly on its outward stroke, and just before it reaches the outer dead center the handle D must be returned to its original position, shutting off the air supply. The first impulse given to the fly- wheel by compressed air will usually be sufficient to produce several revolutions at a speed of about one-fifth of the normal, when no load is on. During the following (suction) stroke a mixture of gas and air in the correct proportions is taken in, and on the next stroke compressed and ignited. If the right mixture does not happen to be obtained and ignition fails to occur, another compressed air impulse is given, which will always produce the desired result. After the first power stroke has been obtained the air supply valve B is closed and the automatic inlet valve held fast by unscrewing the hand wheel C until its hub bears against a collar on the valve stem, whereby the valve disc is firmly pressed on its seat. Then the starting gear is pushed back into the run- ning position, so as to allow the mechanism to open the valves at the regular intervals only. The point of ignition is thereby automatically advanced and may now be adjusted by hand or by the governor of the engine so as to suit the changed conditions. 436 INTERNAL COMBUSTION ENGINES When the main gas admission valve is set in the correct running position, all operations for starting have been duly executed. It may be added that it takes less time to perform the complete cycle of operations than it takes to describe it." (e) In central stations, or in other locations where electric current is available, starting by electricity is the simplest and most satisfactory met hod. There are again several ways in which this may be done. If the engine drives a dynamo, it is possible to drive this dynamo as a motor from another source of current, until the fly- wheel has attained suffi- cient velocity. The layout for doing this is not at all simple, however, and must be carefully handled. A storage battery charged during the previous oper- ation may be used as a source of current if no in- dependent source of cur- rent is available. FIG. ia-39. Motor Starter. A simpler way is to gear an electric motor to a rack on the fly-wheel as shown in Fig. 13-39.* When the engine picks up the motor is automatically thrown out of gear. This scheme has the merit of great simplicity, absolutely no change in the valve operation of the engine being required for starting. One method of automatically disengaging the motor is de- scribed in the Zeitschrift d. V. d. I., January 5, 1900. A general view of the Felten and Guilleaume electric starter for large en- gines is shown in Fig. 13-40, while Fig. 13-41 shows the diagram of connections and the method of operation for smaller sizes. The motor e by means of a chain drive operates the pulley s 2 . Rigidly fastened to this pulley is the gear 2 2 . This gear in turn * F. E. Junge, Power, April, 1906. GAS ENGINE AUXILIARIES 437 FIG. 13-40. Felten & Guilleaume Electric Starter. drives a gear z 3 which is pivoted on a swinging lever h t as shown. The arc z represents the pitch circle of the annular gear inside of the fly-wheel. Arrows indicate the direction of operation for each gear. One end of the lever h is connected by means of a strong helical spring / to the toothed segment m. Meshed with this segment is the gear z l which is turned in one direction or the other by the lever h 2 , which is the operating handle of the motor starting box a. In the diagram, lever h 2 has been pulled to the right as far as it will go. This action has started the motor e, and, by shoving the segment m to the left, has put the spring under considerable tension. This pulls the lower end of h^ to the right and causes z 3 to mesh with the fly-wheel z. The gear 2 3 , being positively driven from 2 2 , starts the fly-wheel revolving. The entire FIG. 13-41. mechanism is held in the position 438 INTERNAL COMBUSTION ENGINES shown by a latch k at the end of the segment m. If now the engine picks up its cycle, and the wheel z attains a greater velocity than it can receive from z 3 , the lever h^ by the action of the fly-wheel will be pushed toward the left, putting the spring / under higher tension. This motion proceeds until the upper end of h^ comes in contact with and unlatches k, when, under the influence of the spring, the segment m is pulled to the right very suddenly and the motor is shut off. At the same time the action of the fly-wheel upon z 3 throws this out of mesh with z. The entire labor connected with this starter therefore consists of the attendants turning the handle h 2 of the starting box to the right to start the motor. Beyond this no further attention is required, the action being entirely automatic. CHAPTER XIV REGULATION OF GAS ENGINES IN gas as in steam engines there are two kinds of speed vari- ation. The first of these is directly due to the impulse of the explosion, and it manifests itself in a variation of the velocity of the crank pin during one engine cycle. To confine this speed variation to within the allowable limits set by the particular service to which the engine is to be put is the function of the fly-wheel. Fly-wheel computations are beyond the scope of this work, but it will be of interest to point out briefly the relation between the various types and combinations of engines as regards the fly-wheel weight required for any given service. This weight depends not only upon the closeness of regulation desired, but largely also upon the variation of the crank effort during one cycle or revolution, and this in turn depends upon the cycle employed and the cylinder combination. The most complete exposition* of this subject relating to gas engines is given by Giildner,* and from the data given by him the following table (see page 440) is adapted. In deriving these figures it is assumed that the same coeffi- cient of fly-wheel regulation, designated by 8 W , is maintained in all cases. This coefficient o _ y max V min w~ y where V is the mean velocity of the crank pin. The second kind of speed variation above mentioned is the change in the number of revolutions of an engine in any given time due to a change of load. This variation is taken care of by the governor. Just as a limit is set to the speed variation during one cycle, so an allowable limit is set to the vari- * Entwerfen und Berechnen der Verbrennungs-motoren, 2d ed. 439 440 INTERNAL COMBUSTION ENGINES TABLE OF RELATIVE FLY-WHEEL WEIGHTS FOR THE SAME COEFFICIENT OF REGULATION; FOR VARIOUS TYPES OF ENGINES AND CYLINDER COMBI- NATIONS. WEIGHT OF THE FLY-WHEEL OF A SINGLE-ACTING, SINGLE-CYLINDER, 4-CYCLE ENGINE is TAKEN = 1.00. Type of Engine and Cyl- inder Combination Crank Travel Between Ex- plosions, Degrees Relative Fly-Wheel Weight for Equal Cyl. Di- ameter, Stroke, and Revolution Equal Maximum Horse-power 4-Cycle 2-Cycle 4-Cycle 2-Cycle 4-Cycle 2-Cycle I. Single-Acting, Sin- gle Cylinder II. Double- Acting, Sin- gle Cylinder III. 2-Cylinder Tandem Single -Acting, 1 Crank 720 540 & 180 360 540 & 180 360 540 & 180 240 180 360 180 360 180 360 180 120 1.000 1.230 .796 1.290 .792 1.290 .678 .335 .802 .424 1.595 .335 1 .602 .335 .237 1 .000 .615 .398 .645 .396 .645 .226 .084 .401 .106 .399 .084 .401 .0.84 .0395 IV . 2-Cy Under opposed , Single- Act ing, 1 Crank V.2 Cylinders in Par- allel, Single-Acting 2 Cranks together VI. 2 Cylinders in Par- allel, Single-Act- ing, 2 Cranks, 180 apart VII. 3 Cylinders in Par- allel, Single-Act- ing^ Cranks, 120 apart VIII. 4 Cylinders in Par- allel, Single-Act- ing, 4 Cranks, Nos. 1 & 3 together, Nos. 2 & 4 togeth- er, and 180 from Nos. 1 & 3 ation that may occur in the number of revolutions from full load to no load, and the governor must be designed and set accord- ingly The allowable variation, the so-called coefficient of gov- ernor regulation, may be expressed by 8 r = n. is the mean number of revolutions where n What REGULATION OF GAS ENGINES 441 the values of 8 W or 8 r may be depends altogether upon the kind of service, and the value of 8 r at least is, or should be, clearly stated in all engine specifications. The value of S w , that is, the variation in the angular velocity of the engine, can be made any- thing, depending upon the weight of the wheel put in. It usu- ally varies from 8 W = V, for ordinary commerical service, to 8 W = ?%* say, for the most exacting service required for the operation of alternating-current generators in parallel. The value of the co- efficient of governor regulation 8 r usually varies between $ and 2^7 i-6-, the speed variation is from 2 to 4 per cent. The theory of centrifugal governor design as applied to gas engines governing by throttling or cut-off cannot be taken up here, except merely to state that it is best, especially in the case of the power gases, to have the governor act upon the mechanism that positively operates the admission gear rather than upon the admission gear direct. In this way the governor does not in gen- eral have to be so powerful and above all it is unaffected by any clogging of the admission gear or valves due to tar and other in- crustations. It should further be noted that any governor of this type that is too sensitive may react with the speed variation within one revolution and will thus be hunting constantly. Dash pots only forcibly overcome these conditions, and it is better to design these governors with plenty of adjustment regarding their sensitiveness. The design of hit-and-miss governors (see below) is in general very simple. In contradistinction to the type of governors above mentioned, they can hardly be made too sensitive or astatic for the quicker they act, the better. In the following we take up, first, Systems of Governing, and, second, Mechanical Details of Governors. i. Systems of Governing. At the outset an essential difference between steam and gas engines in the matter of gov- erning should be noted. The working fluid in the steam engine is a comparatively stable medium, and, as long as the pressure remains constant, one position of the governor mechanism always corresponds to the same load, cycle after cycle recurring with the same development of power. This is absolutely essen- tial for close governing, and in this respect the steam engine has some advantage over the gas engine. The conditions in 442 INTERNAL COMBUSTION ENGINES the latter are very different. The working fluid is prepared by the engine itself, air and fuel being mixed at the engine to produce the medium. Various expedients to this end, more or less successful , are in use, but outside of this, due to accidents of design or other reasons, stratification of the charge more or less complete, and variation in ignition, may result in unequal veloc- ity of pressure propagation through the mass of the charge giving a bundle of different diagrams for the same heat value of the charge. Thus it may result that the same position of the governing mech- anism may not, and often does not, indicate the same power developed, and speed fluctuations are the inevitable result. For- tunately the mixing and ignition apparaus of our modern en- gines can be made perfect enough in their action to confine these fluctuations to within allowable limits. All gas engine governors come under the following systems as far as their effect upon the diagram is concerned. Mechan- ically they may be of various designs, as inertia, fly-ball, etc., as shown later: I. The Hit-and-Miss System. II. Variation of the Ratio of. Fuel to Air with Change in Load; Quality Governing. III. Variation of the Quantity of the Charge to suit the Load. Ratio Fuel to Air remaining constant; Quantity Governing. IV. Combination Systems. V. Governing by Varying Time of Ignition. I. The Hit-and-Miss System This system effects speed regulation by cutting out explosions altogether, depending on the load. Thus, for instance, if the engine is running at full load, the explosions or cycles will follow each other in regular order until the speed has increased enough above the mean to cause the governor to act, preventing the draw- ing in of the next charge, thus causing a "miss." This in turn causes the speed to fall sufficiently below^ the mean to make the governor act the opposite way, causing the explosions to recur. At any other load less than the full load the governor action is the same, except that as we go down in the scale the proportion of "misses" to "hits" constantly increases. This system may be operated in any of the following ways: REGULATION OF GAS ENGINES 443 (a) By keeping the fuel valve closed, so that the engine draws only air for the miss cycle, (b) By keeping the inlet valve closed, thus preventing the admission of both fuel and air. (c) By keeping the exhaust valve open. In this case the admission valve is usually automatic, and its opening is prevented by the fact that on the next stroke no vacuum is formed, the exhaust gases being sucked back into the cylinder. Theoretically this system of regulation is the simplest, and, from the standpoint of fuel consumption, the most economical; practically, however, it is beset with certain difficulties. In theory the cycles are all gone through under exactly the same conditions, and hence ratio of fuel to air, pressure of compres- sion and point of ignition can all be adjusted once for all to suit the requirements of best thermal efficiency. The thermal effi- ciency of the cylinder should therefore be the same at all loads. In practice there is some deviation from this ideal condition, even assuming perfect governor action, but the variation depends somewhat upon the manner of governing. Thus in engines in which only the fuel valve is kept closed to produce the miss cycles, it will generally be found that the card directly following a miss period is larger than those following it, at least for loads approach- ing full load. This is due to the fact that during the miss period the cylinder has been thoroughly scavenged by air, causing the next charge to be purer and somewhat larger in quantity than the average. Under very low loads the effect is apt to be the opposite, that is, owing to a prolonged period of miss strokes the cylinder has cooled so far as to make the first cycles following somewhat slow burning until the cylinder heats up again. It is evident that these variations must have their effect upon cylinder efficiency, but the effect is perhaps greater with liquid fuel engines than with gas engines proper, because a cool cylinder is likely to condense some of the fuel vapor, thus causing a direct loss. In engines that govern by keeping the exhaust valve open, drawing the exhaust gases back into the cylinder, the effects above outlined may be less marked, but the method cannot on that account be recommended as better than the other, because the inevitable mixing of the exhaust gases with the incoming charge has its own harmful effects-. 444 INTERNAL COMBUSTION ENGINES In spite of these facts, however, the hit-and-miss system of governing, no matter how carried out, usually shows a somewhat greater economy of fuel in practice than the other systems. We next turn to the efficiency of this system as a speed regu- lator. It is evident that the closeness of regulation in case cen- trifugal governors are employed depends altogether upon the sensitiveness of the governor, that is, upon the facility with which it changes from one position to the other; although it is possible here also to have a governor too sensitive, resulting in needless hunting. But whatever the type of hit-and-miss governor, the regulation will be closest if at the higher loads a constant series of explosions is followed by a single miss cycle, or if at the lower loads a single explosion is followed by a constant series of misses. Thus | load should be represented by the series Ill-Ill, etc., and load by 1 - - 1 - -, etc. Any disturbance of the gover- nor, accidental or otherwise, as through want of care, increased friction, wear, etc., will alter this ideal condition so that a f load, for instance, may be represented by the series 111 - 11 - 111 - 11, etc. But such variation at once unfavorably affects the regula- tion.* These accidental conditions are not under the control of the designer, and not always under the control of the operator, and the net result is that hit-and-miss regulation, though econom- ical, is somewhat unreliable and certainly not as close as that ob- tained by some of the other methods, unless a very heavy fly-wheel is employed. Hit-and-miss governing is therefore little employed where close regulation is essential, as for electric current generation. For ordinary commercial power operation, where the regulation need not be closer than say 3 to 5 per cent., the system is quite satisfactory, although it is being slowly replaced even in this field. It should be remembered in this connection that, if the engine is belt-connected to the power consumer, the flexible connection will tend to equalize the speed variations to a certain extent. II. Governing by Varying the Ratio of Fuel to Air: Quality Governing In this system the governor is usually made to act upon the fuel admission valve, so that as the load on the engine decreases * See Mollier, Zeitschrift d. V. d. Ingeriieure, 1903, p. 1704. REGULATION OF GAS ENGINES 445 the engine receives less and less fuel in the same total charge volume. This of course decreases the area of the indicator card developed to suit the load. Instead of acting upon the fuel valve, this method of governing has also been carried out by sucking back a certain amount of the exhaust gases, thus also decreasing the heat content of the charge. Another way is to regulate the air admission valve, making the fuel valve automatic. All things considered, however ? the first mentioned method is the best. Considered from a thermal standpoint this system has the advantage that, since the total charge volume remains practically the same for all loads, the compression pressure remains constant throughout. The cards obtained will therefore be somewhat as FIG. 14-1. shown in Fig. 14-1, in which the full line represents the full load diagram. It therefore should follow on theoretical grounds that the thermal efficiency of the cylinder should be about the same for all loads. In practice, however, it has been clearly shown that this system is inferior at low loads to the next one to be de- scribed. In fact, the fuel consumption per horse-power usually increases very rapidly as the load drops. The reason is that, as the fuel-ratio is decreased, the mixture rapidly becomes difficult to ignite and, above all, slow burning. This necessarily increases the heat loss to the jackets and the ignition difficulty may go as far as to prevent ignition altogether, causing a direct loss of fuel. In most cases after-burning is clearly recognizable by the slow dropping of the expansion line. Designers have tried to over- 446 INTERNAL COMBUSTION ENGINES come this difficulty by placing the time of ignition also under control, making it earlier as the load decreases. The scheme, however, does not appear to have been very successful. As a method of governing, this system is capable of giving close regulation with the proper weight of fly-wheel. The very fact, however, that the compression pressure does not drop in proportion to the maximum pressure introduces a disturbing factor into the crank effort diagram which would tend to make the regulation under this system less close at low loads than under System III. III. Governing by Varying the Quantity of Charge of Constant Composition to suit the Load: Quantity Governing Governing by changing the quantity of charge to suit the load may be carried out in three ways: (a) The engine draws a charge full stroke each time, but a part of the charge, depending upon the load, is forced back into the suction passages, the inlet valve being under governor con- trol. (6) The incoming charge is completely cut off by the gover- nor at the proper time, the charge expanding behind the piston for the rest of the stroke. This is known as the cut-off method. (c) The charge is throttled down throughout the entire suc- tion stroke, the governor determining the position of the inlet valves. This is called the throttling method. FIG. 14-2. Figs. 14-2 and 14-3 show the differences in the diagrams ob- tained under conditions (6) and (c) outlined above. REGULATION OF GAS ENGINES 447 Quantity governing in general is, on thermal grounds, open to the objection that the compression pressure decreases with the load, and hence the cylinder efficiency constantly decreases, On the other hand, the mixtures remain readily ignitable down to the friction load, with the result that quantity governing is on the whole more economical than quality governing. The fact, too, that the compression pressure decreases with the maximum pres- sure has a favorable influence upon the crank-effort diagram, ad- mitting of close regulation. As between methods (6) and (c), the former is slightly better because less work is lost in the lower loop. Everything consid- ered, where close regulation is essential, the cut-off method of quantity governing is the best, 14-3. Regarding the economy of the cut-off method as compared with the hit-and-miss method of governing, E. Meyer* finds that down to about \ load the two systems are about on a par. Below this load the efficiency of the cut-off as compared with the hit- and-miss method rapidly falls off. IV. Combination Systems It has been attempted to perfect quantity regulation by chang- ing the compression space so as to keep the compression pressure the same at all loads. Thermally this is a step in the right direc- tion, but no successful machine operating upon this system has yet appeared. Another combination system is that of Letombe, which is in successful use. Letombe regulates by lengthening the time of * Zeitschrift d. V.'d. Ingenieure, April 25, 1903. 448 INTERNAL COMBUSTION ENGINES opening of the inlet valve but decreasing the lift of the gas valve as the load decreases, As far as the fuel is concerned, this is quantity regulation, but the longer time of opening of the inlet valve increases the total charge volume, which means that the leaner mixtures will be more highly compressed than those for higher loads. This is thermaUy correct. Another point is that the richer mixtures at the higher loads, although less highly com- pressed, are less in total volume than the leaner mixtures. Hence as the load increases the ratio of expansion increases as compared 2Q\at FIG 14-4. to the ratio of compression, which tends to draw down the ter- minal pressure at the end of expansion and decreases the exhaust loss. The resulting diagrams are shown in Fig. 14-4, given by Giildner. The compression line a-b belongs to the full load card, line cd to the minimum load card. The latter card shows suc- tion full stroke and a compression pressure of about 190 pounds, the former shows a suction volume equivalent to about 55 per cent stroke and a compression pressure of about 115 pounds. In spite of the thermally excellent features, w r hich gives good REGULATION OF GAS ENGINES 449 regulation, the economy regarding fuel is no greater than that obtained by a purely cut-off system. Other combination systems that have been employed are quantity regulation at high loads combined with quality regula- tion at low loads or vice versa. This is done in some German engines. To compensate for the slow burning of the leaner mix- tures the spark is advanced. Lastly, engines that govern either by quantity or quality regulation at the higher loads have been governed by the hit-and-miss method at very low loads. This is done in the American-Crossley engines and also by Letombe in the system above described. V. Governing by Varying the Time of Ignition Strictly speaking, the time of ignition should be adjusted to suit the kind of charge. That means, for instance, that in qual- ity regulation the spark should be advanced with a decrease of load. It has already been mentioned that this has been tried both by governor control or by hand regulation. The former is rather difficult because proper ignition is subject to so many accidental variations, but hand control is quite practicable and all stationary engines should therefore be furnished with adjus- table spark gear. Automobile engines are generally governed by hand regula- tion of the throttle in combination with the spark. Governing of Two-Cycle Engines Small two-cycle engines are usually governed by throttling either the fuel or the charge. In the first case this results in what is practically quality regulation, in the second in quantity regu- lation, Liquid fuel engines of this type nearly always govern by adjusting the stroke of the pump to suit the load, resulting in quality regulation. In the larger machines, which are nearly al- ways served by separate pumps, it is absolutely essential that the cylinder be thoroughly scavenged. Hence it is usual to first admit air alone directly from the pump or an intermediate receiver, and a little later the fuel or the mixture as the case may be, the point of admission of the latter being under gov- ernor control. The governor may act either on the inlet valve or directly on the pump. It should be noted that if reservoirs 450 INTERNAL COMBUSTION ENGINES are used between pump and engine, there is likely to be a lag of several strokes between the action of the governor and its effect on the engine. For more detailed information on the governing of large two-cycle engines, see the descriptions at the end of this chapter. 2. Mechanical Details of Governors. For hit-and-miss governing, inertia governors of the so-called pendulum type are extensively used. Besides this, centrifugal governors of the fly- ball type also find application. For the "precision" systems of regulation, i.e., quantity and quality regulation, centrifugal governors of the fly-ball type are mostly used; shaft governors either of the centrifugal or inertia type are, however, lately finding application. A (a) PENDULUM OR INERTIA GOVERNORS FOR HlT-AND-MlSS REGULATION. - - The simplest form of this governor is shown in Fig. 14-5. The rod a-b carries the ball weight c, and receives an up-and-down motion usually from the lay shaft as shown by the arrows. As the rod moves down, the weight strikes the projection d, and the ' pendulum is thrown to the right as shown FIG. 14-5. by the dotted position. If the speed is right, the pendulum will have gone back to its normal position by the time the point b reaches the end of the valve stem e, and the gas or inlet valve will be opened. If the speed is above the normal, a-b will not be back in its normal position, and b will miss the valve stem e, causing a miss- stroke. A modification of this, the bell-crank pendulum, is shown in Fig. 14-6. The action of this governor is clear from the previous description. The disadvantage of these two forms may be said to be that they are thrown a considerable distance out of their normal posi- tion. This is overcome in what may be called spring-loaded pendulum governors, the fundamental type of which is shown in Fig. 14-7. As the bell crank a b c moves downward, the inertia REGULATION OF GAS ENGINES 451 of the weight, c, throws the governor into the dotted position. If by the time the stem of the valve, e, is reached, the spring has failed to pull the governor back into its normal position, the valve will not be opened, causing a miss-stroke. M\\\\\\M FIG. 14-6. FIG. 14-8. Fig. 14-8 shows the simplest kind of inertia governor for a horizontal valve stem. This may be used with or without the spring. Its action is sufficiently plain. There are many modifications of these fundamental types.* * See Giildner, Entwerfen & Berechnen der Verbrennungsmotoren, p. 354. 452 INTERNAL COMBUSTION ENGINES These will readily suggest themselves to suit any particular case. The following are some examples from actual practice. Fig. 14-9 shows a modification of the vertical pick blade gov- ernor that has been used on Crossley engines.* The lever, a, is Gas actuated by the cam, and as long as the speed is normal the spiral spring on one arm of the bell crank is strong enough to make the FIG. 14-10. weight, 6, attain its normal position in the time available. The blade, c, then opens the gas valve. If the speed rises above nor- mal, the inertia of the weight, 6, throws it down so far that it can- not reach its normal position in time and c misses the valve stem, * Giildner, Entwerfen & Berechnen der Verbrennungsmotoren. REGULATION OF GAS ENGINES 453 The governor on the Springfield engine, Fig. 14-10, is quite similar to the above, except that the blade acts horizontally.* In this case the bell crank is carried by the slide, A, which is ac- tuated from S. For normal speed the blade, P, is horizontal, and in this position hits the stem of the gas valve. For excessive FIG. 14-11. speed, the spiral spring fails to bring the weight W back into normal position in time and P misses the gas valve. An English design of hit-and-miss governor not quite so simple is shown in Fig. 14-1 l.f Here the inlet valve stem, 1, carries the FIG. 14-12. bracket, 2. To this is pivoted a blade, 4, at 3. For normal speed of the engine, the spring, 5, gently presses the inclined surface, 6, against the stationary roller, 7. In this position the blade, 4, will hit the lever, 8, thus opening the gas valve, .9. Should the * Power Quarterly, Oct. 15, 1900. f Clerk, Gas and Oil Engines. 454 INTERNAL COMBUSTION ENGINES speed, however, rise above normal, the upward throw given to the blade due to the inclined surface, 6, sliding against the roller, 7, becomes excessive, and blade 4 fails to assume its normal posi- tion by the time the lever 8 is reached. The gas valve then fails to open. An ingenious modification of the above types of hit-and-miss governors used by Delamare* is shown in Fig. 1412. The stem of the inlet slide valve carries a small cylinder, a, into which is fitted a stationary piston, 6, air- tight. The needle valve, c, serves to adjust the velocity of air escape from behind the piston, b, With the cylinder, a, in the extreme right-hand position, air is admitted behind b through the groove, d. As a travels to the left at the proper speed, the air is com- pressed, but it escapes through the needle valve, c, just fast enough to prevent its pushing out- ward the plunger, /, in the branch cylinder, e, against the spring. In FIG. 14-13. t n j s position the blade, g, hits the gas valve stem, h. For an excess in engine speed, cyl- inder a travels to the left so fast that the air cannot escape at the proper rate, the pressure generated forces the plunger, /, out- ward, and g misses the valve stem as shown in the dotted position. (b) CENTRIFUGAL GOVERNORS FOR HIT-AND-MISS REGULA- TION. A hit-and-miss governor of this type is used in the French engine shown in Fig. 14-13. f The centrifugal governor * Schottler, Die Gasmaschine. f Power Quarterly, Oct. 15, 1900, REGULATION OF GAS ENGINES 455 in the fly-wheel, when the speed becomes too high, actuates the latching arrangement i-k and throws the oil pump serving to charge the vaporizer temporarily out of action. FIG. 14-14. FIG. 14-15. The Robey governor,* Fig. 14-14, is of the fly-ball type; normally the roller, a, under the control of the governor travels on the cam, b, thus opening the gas valve at the proper time. Under excessive * Schottler, Die Gasmaschine. 456 INTERNAL COMBUSTION ENGINES speed, however, the roller is shoved to the left on its spindle by the governor linkage, and misses the cam. The Campbell oil engine has a fly-ball governor which operates to keep the exhaust valve open when the speed exceeds the nor- mal, Figs. 14-15 and 14-16.* The rising of the governor weights, Fig. 14-16, depresses the end, 0, of the lever, N, thus interposing the plate, P, Fig. 14-15, between the end of the exhaust lever, M, and the stationary bracket, Q. This prevents the exhaust valve from closing until the dropping of the speed causes the governor FIG. 14-16. to withdraw the plate, P, when the exhaust valve is again regu- larly opened by the sliding piece, K. A shaft governor acting to hold the exhaust valve open is used on the Perkins engine, Fig. 14-17.f The exhaust valve is operated by the lever R, which in turn is actuated by the block A, striking the end block, B y of the rod, C. As the speed ex- ceeds the normal, the governor weights move toward the circum- ference of the wheel, and, through suitable linkage, throw the latch, P, into a recess in the side of the block, B. This prevents the rod, C, from returning, and keeps the exhaust valve open. * Clerk, the Oil and Gas Engine, f Power Quarterly, Oct. 15, 1900. REGULATION OF GAS ENGINES 457 (c) QUALITY REGULATION. The quality method of regulation, as carried out for gas fuel, is best exemplified in the Niirnberg en- gine. This engine has separate gas and inlet valves. At all FIG. 14-17. loads the inlet valve opens and closes at the same time, thus ad- mitting a constant charge of volume. But the gas valve, under governor control, opens later as the load drops, and closes about the same time as the main inlet valve. Thus the composition or FIG. 14-18. quality of the mixture changes with every load. A certain dis- advantage of this method of proportioning the charge is that the mixture is not only affected by the relative time of opening of 458 INTERNAL COMBUSTION ENGINES the gas and inlet valves, but also by the relative velocities of the gas. and air currents. Thus at low loads, when the gas valve opens, the gas column has to start from rest, while the air column already has its maximum velocity. The time needed for the accel- eration of the gas column is practically constant, hence this factor affects the mixture differently for every load. Another point is that unless the gas valve closes quickly there is apt to be so much throttling of gas as to make the mixture drawn in at the end for low loads incombustible. This is especially bad since this mixture is apt to be located around the igniter. The governor used on the Niirnberg engine is of the ordinary fly-ball type. The trip gear for the gas valve is shown in Fig. 14-18.* The governor determines the position of the bell-crank FIG. 14-19. lever e-g, and through this the position of the wiper cam lever d. The eccentric rod, 6, through the latch b lf lifts the gas valve by depressing the valve lever, a. The position of d determines the time of opening. The valve is closed practically instantaneously by the coil spring shown in the top of the housing as soon as the small roller, c, has released the latch 6 X . The dash pot, /, serves to dampen the drop of the valve, making it close without shock. Many oil engines are governed by what is practically quality regulation. Thus the Hornsby-Akroyd oil engine, Fig. 14-19, f takes a full charge of air every stroke, but the quantity of oil is * Power, Feb., 1906. f Guldner, Verbrennungsmotoren, p. 119. REGULATION OF GAS ENGINES 459 changed to suit the load. The quality of the mixture therefore changes for every different load. The proportioning of the oil supply to the load is accomplished as follows. The oil pump supplies a constant quantity of oil every stroke to the vaporizer valve, c. This valve, however, is a two-way valve, one exit open- ing through an atomizing nozzle into the vaporizing chamber, while the other allows some of the oil to flow back into the oil tank. The fly-ball governor, through the linkage shown, controls the size of this overflow opening, c', and thus determines the amount of oil which enters the vaporizer chamber. FIG. 14-20. The governing mechanism of a small Diesel oil engine is shown in Fig. 14-20.* The pump plunger, a, is actuated by the lay shaft, h. The suction valve is shown at c and the automatic discharge valve at b. The oil under pressure flows through b to the atomiz- ing nozzle. The amount of oil required for the load on the engine at any given time is measured by controlling the time of closing of the suction valve, c, forcing more or less of the oil back into the * Giildner, Verhrennungsmotoren, p. 379, 460 INTERNAL COMBUSTION ENGINES oil supply tank. The motion of valve c is controlled through linkage not shown, by the governor lever e. At g this lever re- ceives an up-and-down motion from the lay shaft, but the manner in which this motion is transmitted to the suction valve depends upon the position of the fulcrum, /, which is determined by the governor, thus varying the time of closing of the valve. (d) QUANTITY REGULATION. Koerting Bros, govern their four- cycle engine as follows, Fig. 14-21*: Air enters through the pipe, D, and gas through B. The two are mixed in the proportion suited to the particular gas by the mixing valve A. This valve FIG. 14-21. when once adjusted therefore furnishes a mixture of constant qual- ity. To suit the quantity of this mixture to the load, the gover- nor controls the position of the throttle valve e in the passage leading from the mixing valve to the inlet valve. This valve thus throttles the charge during the entire suction stroke. A very similar arrangement is used in the Westinghouse vertical engine, Fig. 14-22. f A is the throttle valve controlling the entrance of the mixture to the cylinder. Its position depends upon the * Schottler, Die Gasmaschine. f L. S. Marks, Instruction Paper on Gas and Oil Engines. REGULATION OF GAS ENGINES 461 action of the fly-ball governor, B. Gas enters the interior of this valve through ports G and air through ports D. The .relative proportions of these ports are fixed by slides H-H, so that the proper relation of air to gas for the fuel used can be established and maintained. FIG. 14-22. Fig. 14-23 shows the arrangements for operating and control- ling the mixing valve in the Ehrhardt and Sehmer engine, which is a modification of the Deutz engine.* It will be readily seen, by a study of the figure, that the governor, through its linkage, con- trols the position of the fulcrum, about which the valve lever turns. The farther this fulcrum moves to the right, the greater will be the opening of the inlet-valve and the greater, consequently, also, the amount of mixture, proportioned in passing the valve, admitted to the cylinder. It will be noted that all of the above examples of quantity governing throttle the mixture throughout the suction stroke. It was pointed out, in the discussion on various systems of govern- ing, that the system employing quantity regulation with cut-off was superior to thai which throttles the charge throughout the stroke, and it seems surprising, therefore, that this system is not in more extended use. There is to the writer's knowledge only * K. Reinhardt, Bi-monthly Bulletin of the Am. Inst. of Mining Engineers, November, 1906, p. 1092 462 INTERNAL COMBUSTION ENGINES one engine in this country employing the cut-off system, the Sar- gent complete expansion engine. This engine employs a Rites inertia governor which acts upon the lay shaft driving gear which is loose on the crank shaft. As the load increases or decreases, the governor retards or advances this gear relative to the crank shaft, thus affecting all of the valve events depending upon the lay shaft at the same time. Since near the end of the stroke, however, the piston velocity is comparatively small, a relatively FIG. 14-23. large advance or retardation of the gear will not affect the exhaust events or the beginning of the suction stroke very much. But the cut-off occurs along the suction line some- where near the time when the piston has its maximum velocity, and hence a relatively small advance or retardation of the gear is sufficient to change the cut-off materially and hence to control the speed. (e) COMBINATION QUANTITY AND QUALITY REGULATION. - The purpose of these combination systems has already been REGULATION OF GAS ENGINES 463 pointed out. Fig. 14-24 shows the method as carried out by Reichenbach.* The governor, through the rod, c, controls the position of the bell crank lever, d. This lever has two slotted arms, a and 6. The former, by rod and lever, reg- ulates the position of the mixing valve, while the latter, by a similar arrangement, controls the position of the throttle valve. Suppose now that the sliding blocks a and b are in the position shown in the figure. Under these circum- stances, the movement of the throttle valve is insignificant FIG. 14-24. because b is so close to the center. Hence in this position the engine is practically governed by quality regulation. If the posi- tions of a and b were reversed, that is, b at the outer end of its slot and a at the inner end, there would be practically quantity regu- lation. In actual operation the positions of a and b in their re- spective slots are determined by trial and depend upon the gas used. The final adjustment is such that for the upper ranges of load the quality of the mixture is changed, while for the lower loads the throttle valve comes into action while the mixture is * Power,' July, 1906. 464 INTERNAL COMBUSTION ENGINES held at practically constant composition. In addition to this, Reichenbach also puts the point of ignition under governor con- trol, advancing it as the load drops. Reinhardt's method of governing does not come strictly under any of the heads so far discussed. The inlet valve is shown in Fig. 14-25.* The valve opens and closes with the beginning and end of the suction stroke. Above the valve there are arranged a series of ports, I for gas and II and III for air. The flow of gas FIG. 14-25. or air from these ports into the valve chamber proper is controlled by a slide guided by the valve spindle. At the beginning of the suction stroke the slide only frees the ports from chamber III, and hence only air is admitted. 'At some point in the stroke, how- ever, as determined by the governor, the slide is suddenly released and forced down by springs. This closes ports III and opens gas ports I and air ports II. Mixture of the proper proportion then * Reinhardt, Bi-monthly Bulletin, Am. Inst. Mining Engs., Nov., 1906. REGULATION OF GAS ENGINES 465 enters the cylinder until the end of the suction stroke. As far as the admission of a varying quantity of constant mixture is con- cerned, this system is pure quantity regulation; but since the amount of air drawn from III increases as the load decreases, the mixture in the cylinder as a whole grows leaner as the load drops. It is claimed, however, that combustion is good even under fric- tion load, because although the amount of constant mixture drawn in is small, more or less stratification of the charge ensures the presence of a fairly rich mixture around the igniter. The system has the advantage over pure quantity regulation in that the com- pression pressure remains practically unchanged. In general principle the above method of Reinhardt seems to be a modification of the method of Mees patented in 1901. FIG. 14-26. Letombe, recognizing the serious effect on thermal efficiency of the decrease in compression which accompanies pure quantity regulation, goes one step further and so arranges the valves that the lean mixtures under low loads are more strongly compressed than the rich mixtures at the higher loads. This is thermally an important step in the right direction. In Fig. 14-26,* a is the main inlet valve operated at a constant lift by the cam, d, through the lever, e. Ahead of this inlet valve are the air valve, 6, and the gas valve, c. Valve b is operated through the lever, g, by the cam, /. Gas valve c opens only when pushed up by 6. Cam / is shown in greater detail at the left. It has a number of elevat tions, each of which consists of two steps of different lengths, as shown at 1, 2, and 3. At full load the roller on the lever, g, mounts * Schottler, Die Gasmaschine. 466 INTERNAL COMBUSTION ENGINES the higher step of elevation, 1, opening both b and c wide, and the mixture, of constant proportion, enters the cylinder through a, which has been opened at the same time. After awhile the roller drops to the lower step of elevation, 1, air valve, 6, closes partly, closing gas valve, c, completely. Only air is then drawn into the cylinder until the roller, g, drops off the cam altogether. Now suppose that the load drops. The governor then pulls over the roller, g, until it comes in line with elevation 2 or 3, as the case WATCH IGNITION MECHANISM '"STARTING HANDLE FIG. 14-27. may be. The time of opening of the gas valve, c, is then short- ened, because the higher step of the elevation is shorter; but the time of admission of air alone to the cylinder is lengthened, be- cause the lower step of the cam is longer than it was before. Hence, although the mixture is leaner, the total charge volume is greater than at full load, and the leaner mixtures therefore receive a higher compression. Theoretically, this should give increased thermal efficiencies at low loads as compared with other systems REGULATION OF GAS ENGINES 467 of regulation. In practice, however, for reasons not explained, there does not seem to be much difference. (/) GOVERNING OF TWO-CYCLE ENGINES. Fig. 14-27 shows a type of two-cycle engine, the Lozier,* much used for small motor boat work. As in all machines of this kind, the mixture is formed and compressed in the crank-case, and flows from here through a communicating passage into the cylinder, as shown by the arrows. Speed is controlled simply by the throttle valve in the passage, thus controlling the amount of mixture entering the cylinder. This is a very simple case, because the load on the engine is prac- tically constant. In the new Buckeye two-cycle engine, which, as far as size is concerned, is intermediate between the small two-cycle gasoline engine and the very large two-cycle blast furnace gas engines of the Koerting or Oechelhauser type, the forward end of the cylinder is used as a mixture pump. Regulation is effected by means of the balanced throttle valve 60, Fig. 12-31, p. 289, which, under the control of the fly-ball governor, regulates both the suction to the pump and the delivery from the pump to the combustion chamber. Thus this engine governs by varying the quantity of the mixture; but since the cylinder is always thoroughly scav- enged by fresh air, the mixture will naturally grow leaner as the load drops, while the compression pressure remains practically the same. The governing details of the Oechelhauser two-cycle machine vary somewhat, depending upon the firm building them. Fig. 14-28 f shows the scheme adopted by Borsig. It may be remem- bered that in this machine air and mixture enter the cylinder through separate rings of ports in the side of the cylinder, as soon as the piston uncovers them. The air ports are uncovered first to admit the scavenging air, the mixture ports a little later. Each ring of ports is surrounded by a receiver into which the air and gas pumps deliver their charges under some compression. It is evident that, since the time taken by the piston to uncover and cover the gas and air ports is practically constant, the amount of mixture that can enter the cylinder during a given time must de- pend upon the pressure in these receivers, and upon the port area. * Power Quarterly, Oct. 15, 1900. t Hiedler, Gross-Gasmaschinen, p. 67. 468 INTERNAL COMBUSTION ENGINES The governor, therefore, is made to act upon the gas and air pumps to control the pressure in the receivers. Further, as the load de- creases and approaches the friction load, the amount of mixture entering becomes so small as compared with the amount of air in the cylinder that ignition is likely to become difficult. For this reason the mixture ports are surrounded by a slide under governor control, as shown at A, Fig. 14-28, which is operated in such a way that as the load decreases the ports opposite the igniter are FIG. 14-28. gradually closed first, thus insuring a comparatively rich mixture around the igniter at all times. The method of governing the Koerting two-cycle engine also differs somewhat, depending upon the manufacturing firm. In all designs there are a gas and an air pump, of which the air pump delivers its charge into the cylinder from the commencement of its discharging stroke in order to scavenge out the cylinder. The gas pump runs idle for a part of its discharge stroke, generally forcing the gas back into the suction main, until, at a point de- termined by the governor, the overflow valve closes, and the gas pump delivers the rest of its charge into the cylinder through a valve on top. Fig. 14-29 * shows the pump construction used by * F. E. Junge in Power, November, 1906. 470 INTERNAL COMBUSTION ENGINES the Gutehoffungshiitte. Both the gas and air pumps are con- trolled by piston valves operated by rocker arms and eccentrics. Situated below the piston valve of the gas pump there are two FIG. 14-30. auxiliary piston valves, h and i, which are under the control of the governor. Depending on the load on the engine, the gover- nor, through linkage shown in Fig. 14-30, and of which lever k, Fig. 14-29, is a part, rotates these auxiliary valves about their axis, thus controlling the time when the gas pump commences to deliver its charge to the power cylinder. CHAPTER XV THE ESTIMATION OF POWER OF GAS ENGINES IT is intended in this chapter to point out the various methods in use by which the power that a given engine may be expected to develop can be computed, or, what amounts to the same thing, to determine the cylinder dimensions for any given power. In steam engine practice this is a comparatively simple matter. It is necessary merely to lay down an ideal indicator card along well-defined lines, and then, by the use of card factors closely fixed by long practical experience, to determine the probable mean effective pressure in the cylinder or cylinders. Although this method is by some writers advocated also for the gas engine, the determination of the card factor for gas engines is based upon so many component factors, which in turn depend altogether upon the judgment of the designer, that the result is anything but certain. The difficulties attaching to this method are pointed out in greater detail below. Before describing the various methods of computing cylinder dimensions it is well to examine briefly into the allowable piston speed. The theoretical limit of piston speed of course depends directly upon the time of explosion, and this in turn depends upon the kind of mixture and to a certain extent upon the method of ignition. As far as modern practice goes the upper limit to-day seems to be at about 800 feet par minute. A certain formula, empirical as far as the writer is aware, makes the piston speed depend upon the horse-power of the engine, stating that piston speed = (660 + 6 \/B. H. P. ) ft- per min. (1) " H in which B. H. P. n is the normal brake horse-power. A similar empirical formula for the revolutions per minute 19QQ 13 r - p - m - = 65 + (2) 471 472 INTERNAL COMBUSTION ENGINES For stationary and especially low-compression engines the values of equations (1) and (2) should not be exceeded. For automobile and other high-speed work the revolutions may be increased up to 1.35 times the value computed from equation (2). I. The First Method of Horse-power Computation is more or less empirical in that it depends either upon the outright assump- tion of the mean effective pressure, or upon the computation of this factor from empirical or semi-empirical formulae. Thus Grover bases the determination of the M. E. P. upon the com- pression pressure, making M. E. P. = 2 C - .01 C 2 (3) where C = the pressure of compression in pounds above atmos- phere. This formula is derived from an examination of a large number of indicator diagrams, but leaves out of consideration the kind of fuel or the quality of the mixture. Further, the formula gives the maximum result when the compression pressure is 100 pounds by gage. Beyond this the M. E. P. drops. The results are therefore at best approximate and in some cases absurd. S. A. Moss, in an article entitled "Rational Methods of Gas Engine Powering/' in Power, July, 1906, goes further than Grover and, in fixing upon the probable M. E. P., takes into ac- count the size of engine as well as the kind of fuel. The cooling loss is relatively less in large than in small engines, hence the M. E. P. may be expected to increase somewhat with the size of the engine everything else remaining the same. The kind of fuel has an influence upon the power developed in any given cylinder due to the fact that the best air-fuel ratios, i.e., mix- tures, of the different fuels have a different heat content per cubic foot. Some of the mixtures are poorer, some richer than the average; as computed in Chapter X. Presupposing equal completeness of combustion, the richer mixtures yield the higher mean pressures, and hence the greater power. The following table shows the probable M. E. P., as computed by Moss. Only four-cycle engines are considered and the fuel is assumed to be average natural gas or illuminating ~gas (average heat content 87 B. T. U. per cu. ft. of mixture). POWER OF GAS ENGINES 473 TABLE I. VALUES OF MEAN EFFECTIVE PRESSURE FOR GAS ENGINES X<5T 81 MEAN EFFECTIVE PRESSURES, LBS. PER SQ. INCH. &** la Approximate Brake Horse-Power of Engine III hi- 5 or less 10 25 50 100 200 500 50 40 60 65 70 75 60 35 65 70 75 80 70 30 70 75 80 85 85 90 95 80 28 70 75 85 90 90 95 100 90 26 90 95 95 100 105 100 24 95 95 100 100 110 110 22 95 95 100 100 110 120 20 100 100 110 If some other kind of fuel than the two above specified is used, the M. E. P. of Table I should be multiplied by the proper factor from Table II following: TABLE II. RATIOS OF HEAT OF COMBUSTION PER CUBIC FOOT OF PERFECT MIXTURE, TO VALUE FOR AVERAGE NATURAL GAS OR MANUFACTURED ILLUMINATING GAS (87). Oil Gas 1.00 Water Gas (uncarbureted) 1.00 Coke Oven Gas 93 Air Gas (Siemens Producer Gas) 79 Carbureted Water Gas 1.05 Anthracite Producer Gas . 80 Bituminous Producer Gas 87 Blast Furnace Gas 67 Acetylene Gas 1.28 Gasoline (liquid, vapor, or vaporized and mixed with air) 1.12 Kerosene (vaporized alone or mixed with air), and entering cylinder .90 at about 150 F 90 That the maximum compression pressure allowable differs for every fuel has already been pointed out in Chapter IV, where a table giving the compression pressures ordinarily employed will be found. Having thus determined the probable M. E. P., the indicated horse-power developed at normal load by any given engine may then be computed from the formula (M.E.P.) Ian 33000 I.H.P.,- (4) 474 INTERNAL COMBUSTION ENGINES where I = stroke in ft. a = piston area in square inch. and n = no. ot explosions per minute = aj)out four-cycle machine at full load. It will generally be found that at full load the number of , . , ., i . ., r.p.m. explosions in a hit-and-miss engine is not quite equal to - 2i or that in a machine governed by any of the precision methods the card given at full load is not quite the maximum card obtain- able. This is done in both cases to give the engine some over- load capacity, say from 10 to 20 per cent. Assuming an average overload capacity of say 15 per cent, we then have the T TT P T TT P i.Ji.i -m ,-^ I.H.1.,- J15 where I. H. P. n = normal or full load I. H. P., and I. H. 1* = maximum I. H. P. obtainable. ' m Now at full load the mechanical efficiency of an average engine may be assumed to be 80 per cent, so that the normal brake horse-power will be B. H. P., -. .8 I. H. P. M = -4; LH.P. W = .695 I. H. P. m (6) i.io or I. H. P. m = 1.44 B. H. P. n (7) In the case of an engine to be constructed, B. H. P. n is usually specified, hence I. H. P. n can be found from equation (6). Sub- stituting this in equation (4) leaves as the unknown factors in this equation the factors I, a, and n. The latter depends upon the revolutions, which may also be specified or determined from equation (2). This leaves only I and a to be determined. It is next necessary to find some relation between cylinder diameter and stroke. An examination of existing engines shows that practice in this regard is pretty, well settled. High-speed S~f~ T*O Kf* engines show a ratio of - = 1.1 to 1.3. In medium speed diameter engines the ratio varies from about 1.2 to 1.6, and large machines vary between 1.5 and 2.0, POWER OF GAS ENGINES 475 To determine the cylinder dimensions, let d be the diameter of the cylinder in feet and let / = x d (8) Then equation (4) may be rewritten for maximum power, (M.E.P.) X xd X .785 d* X 144 X n 33000 300 From this the cylinder diameter is 3 ^V@iS^ ft - (10) EXAMPLES. 1. To show the method of using these equa- tions, suppose it is desired to determine the required cylinder dimensions for a 100 B. H. P., four-cycle throttling engine, using anthracite producer gas. From equation (2), the number of revolutions should be about 1200 From the table of Chapter IV, a compression pressure of 120 pounds by gage is about normal for the fuel used. From Tables I and II we should expect a mean effective pressure for this case of M. E. P. = 100 X .80 - 80 pounds. From equation (7), I. H. P. m = 1.44 B. H. P. n = 1.44 X 100 = 144 The engine being of the throttling type, r.p.m. 185 n = -^ = 92.5 and a;' may be assumed equal to 1.6. Substituting these values in equation (10), we finally have . = Appr. 184" Stroke = 1.6 X 18.5" = app. 29f" and r.p.m. = 185 476 INTERNAL COMBUSTION ENGINES 2. A four-cycle gasoline engine has a cylinder diameter of 6", a stroke of 8i", and the number of revolutions is 250 per minute. It is governed by hit and miss. What is its probable brake horse-power at normal load? The M. E. P. at normal I. H. P. from Tables I and II is about 1.12 X 75 = 84 pounds assuming that the compression pressure is about 70 pounds by gage. For the maximum indicated horse-power, the number of r i>m 250 explosions n may be assumed to be equal to ' v ' ~ = ^- =125. Substituting in equation (4) 84 X ^X 785X6 2 X125 I.H.P. JW =1.15LH.P. W = 144B.H.P.,= 33()QQ = 6.35 From which B.H.P. W = |~ = appr. 4.4 II. The Second Method of determining mean effective pres- sure follows steam engine practice, and is developed in detail by Lucke in his Gas Engine Design. The method consists of draw- ing first a standard air reference diagram. From this the mean effective pressure is determined by multiplying the mean effective pressure of the standard diagram by a card or diagram factor depending upon the kind of fuel used and the compression carried. Thus fundamentally this method does not differ greatly from that of Moss outlined above. The standard air reference diagram is obtained by working a pound of air, which is assumed to receive as much heat as a pound of the proper fuel-air mixture to be used in the real engine would contain, through the cycle. The following table shows the card factors as determined by Lucke from a large number of constructed engines: POWER OF GAS ENGINES 477 Kind of Fuel and Method of Use Range of Compression by Gage Card Factor % Kerosene, when previously vaporized Kerosene, injected on hot bulb, may be as low as Gasoline, used in carbureter requiring a vacuum, de- pending upon the extent of the vacuum 45-75 30-40 20 25-40 Gasoline with but little initial vacuum 80-130 50-30 100-160 56-40 Coal gas Av. 80 Av 45 Blast furnace gas .... 130-180 48-30 Natural gas .... 90-140 52-40 NOTE. Card factors for two-cycle engines may be taken as .8 that for four-cycle machines. The mean effective pressure of the standard air reference diagram, of course increases continuously with the compression. The above table, however, shows that the card factor decreases as the compression increases for all fuels except kerosene. This accounts for the fact that in Moss's Table I the M. E. P. does not increase beyond certain compressions. The thermal and mechan- ical reasons underlying this fact have been explained in Chap. IV. The table of card factors shows a large range of variation in the different factors and emphasizes the point made by Lucke that in determining cylinder sizes a great deal must be left to the personal experience and judgment of the designer. III. Third Method of Determining Cylinder Dimensions. It must be obvious from what has been said that it is not at all easy to predict with fair accuracy the mean effective pressure an engine may be expected to realize. For that reason a method of determining the cylinder dimensions of a proposed engine, or the power of an existing engine, without first determining the M. E. P., should be welcome. The method, developed by Giildner,* is based upon two well- known facts: 1. The power developed by any gas engine depends directly upon the volume of the mixture it can handle in unit time. 2. The power also depends upon the thermal efficiency with which the engine can handle this volume of mixture. Regarding the first point, the charge volume in unit time * Giildnor, Entwerfon & Berechnen der Verbrennungsmotoren. 478 INTERNAL COMBUSTION ENGINES involves cylinder diameter, stroke, and number of revolutions. Two of these factors can usually be fixed upon, which determines the third. The computation of the charge volume for a given time involves the assumption of a value for the volumetric effi- ciency of the suction stroke. There is now so much experimental data regarding this point that no great error can be made. A table of volumetric efficiencies E v for various types of engines is given on page 86. The same is true of the thermal efficiency and fuel characteris- tics. A large number of tests on various types of machines and different fuels enables us to-day to forecast with a fair degree of accuracy what thermal efficiency may be expected from a given engine when its approximate power and the kind of fuel used are known. The following derivation of equations for values of diameter, stroke and revolutions per minute is due to Giildner,* as is also the appended table. All metric measurements and values have been transformed to English units: Let B. H. P. n or N n = nominal brake horse-power. n = R. P. M. d = piston diameter in feet. / = stroke in feet. V h = .785 dH = piston displacement per stroke in cu. ft. V = E v V h = actual volume of mixture in cu. ft. per suction stroke, barometer 28.95", temperature 59 degrees Fahrenheit. y E v = = volumetric efficiency of suction stroke. L = the volume of air in cu. ft. required for 1 cu. ft. of gas fuel, or 1 pound of liquid fuel, under most favorable practical con- ditions. This is not the theoretical quan- tity, but some excess quantity as seems best for the particular fuel. * The article may be found either in the Zeitschrift des Vereines deut- scher Ingenieure, April 26, 1902, or in his book, Entwerfen und Berechnen der Verbrennungsmotoren, pp. 213-215. POWER OF GAS ENGINES 479 L h = the resulting actual quantity of air in cu. ft. for one explosion, for nominal horse-power. C s = the quantity of fuel used per hour, for gases in cu. ft., for liquids in pounds, at nominal horse-power. C = the same per horse-power hour. C h = the same per explosion. H = the lower heating value of the fuel, for gases per cu. ft., for liquids, per pound, in B. T. U. _ N n X 33000 X 60 _ 2545 N n _ ( thermal efficiency at the 778 C S H C S H ( brake = economic efficiency. From the above we can derive for four-cycle engines N n X 33000 X 60 2545 N n 778 He ~TU~ _2C J _2X2545JV 84.8AT, 60 n 60 nHe nHe C S L _ 2545 N n L _ 84.8 N n L h 30 n 30 nHe nHe For two-cycle engines equations (12) and (13) should be divided by 2, because there is a charging stroke for every revolu- tion. ENGINES FOR GAS FUEL The actual charge taken in by the engine during one suction stroke must be v - c* + ^ This volume requires a piston displacement of V - C h + L h _ 84.8 N n + 84.8 N n L V ~ - (14) E v nHeE v 84.8 #(! + L) nHeE v Solving (14) for d, I and n in' turn, we have 480 INTERNAL COMBUSTION ENGINES ENGINES FOR LIQUID FUEL In this case the fuel is introduced either in liquid form or in the shape of vapor. In either case the volume ratios of fuel to air are very much smaller than they are in the case of gaseous fuels. Take for instance the case of alcohol, a fuel of low heating value. Here we find that the vapor does not form theoretically more than 4 per cent of the volume of the charge. In reality it is even less than this on account of the excess air used. In view of this fact we may put for these engines .785 d*l = = ' n cubic feet (18) Solving this again for d, I and n we obtain for engines using liquid fuel: - > In these equations the quantities to which certain values must be assigned are L, e and E l v . As stated before, there is now so much practical data at hand that the proper determination of these should cause no difficulty. For values of E v see Table, p. 86. The table on the opposite page gives values of L, C and e, as found for various fuels. Since the nominal brake horse-power N n is 18 to 20 per cent below the maximum capacity of engines in ordinary cases, it is well to assume an excess of air of about 30 per cent at the outset, in the case of machines using gas. In the case of liquid fuel POWER OF GAS ENGINES 481 a H Q B < rf p o E || PS Sz SI to to i i t>- 00 CO to CO 00 T* 00 TjH -^ CO .tO . ; co ; ; 8 :S : : : :| : : . c^ . co oo t>- \ '. ; r ; CO ; I-H 00 Tf ; ; ; ; T-H c^ i- i Tj5 (N CO * / ' ; i> ; oo o I-H co ; ; ; w . p p co ; ; ; r i I i-H T I i ( ! ! rH C^ > ^ C^l 10 '. co '. '. '. to to co o i I ' i I i I r* co ' rH ' T ; : '. . co co co co~ to C^ (N O5 O 00 ^ (N CO CO CO CO m to o oo to tO to CO CO i i OS CO Tt< -(NO >-, . ow . v -S M . ^ O j, -S W '5 22 -S ' -2 's e 1 1. 1 1 cu ^ g fl o o P 73 O c O 3 S Woo^ O ^3 I-H >.' J> I I tH -H -H >C:K 482 INTERNAL COMBUSTION ENGINES engines a greater excess, 50 to 60 per cent, should be used, because these fuels are usually high in heating value, and if any economi- cal degree of compression is to be employed, the fuel mixture should be rather lean. Besides it is not easy to produce a uniform fuel mixture in the case of liquid fuels, and an excess of air would help to overcome this difficulty. These excess air allowances have been made in the table. The values of C and e are for the various fuels and for various sizes of engines, as determined from existing data. The fuel consumed by heating apparatus and ignition flames is not included in C. In the case of engines using generator gas it is supposed that suction generators are employed requiring no fuel for a separate boiler. To see how the results obtained by Giildner's method agree with those obtained by Moss, the same examples will be taken. EXAMPLE 1. Engine to be 100 B. H. P., throttling governor, fuel to be anthracite producer gas. Determine the cylinder dimensions, assuming as before that the number of revolutions is 185 per minute. For this case any one of the formulae 15 to 17 may be used. Taking the first, we have in which N n = 100 B. H. P. L = air used per cu. ft. of the fuel gas = say 1.3 for anthracite gas from the table. e = thermal efficiency that may be expected = .24 from the table. n = 185. H = heating value of fuel = 141 B. T. II. from the table. I = Stroke = x d = assumed as 1.6 d in previous case. E v == volumetric efficiency = say .90 from table p. 86. The, ff- 108 X 10 X 2 ' 3 -24X.185X 141 X 1.6 dX .90 rf3 = 108X100X2.3 .24 X 185 X 141 X 1.6 X .90 POWER OF GAS ENGINES 483 from which d = 1.42 ft. = appr. 17", and I = 1.6 d = 1.6 X 17" = appr.27i". This method, therefore, apparently calls for a smaller engine. EXAMPLE 2. Gasoline engine, cyl. diam. 6", stroke 8J", r.p.rrr. 250. Hit-and-miss governor. Determine probable B. H. P. For this, equation 21 may be rewritten _ nelHd?E v 108 L in which n = 250 r.p.m. e = thermal efficiency = .19 from table. I - stroke - 8J" = .71 ft. H == heating value of fuel = 19800, from table. d = 6" - .5 ft., d 2 = .25. j v = .75, from p. 86, since this engine is likely to have an automatic inlet valve. L = 250, for a good rich mixture. Then 250 X .19 X .71 X 19800 X .25 X .75 ^ p ^ = 108 X 250 The previous example gave 4.4 B. H. P. so that the agreement in this case is satisfactory. IV. For determining the power rating of automobile engines, several more or less empirical formula are in use, two of which will be cited. The Association of Automobile Manufacturers has fixed upon the following expression for four-cycle engines: where d = diameter of cylinder in inches, and N = number of cylinders. This is based on a speed of 1000 r.p.m. In this formula, however, the factor 2.5 results from the contraction of several other factors more or less arbitrarily assumed, and hence as far as design is concerned the formula is of little or no use. 484 INTERNAL COMBUSTION ENGINES where d = cyl. diam. in inches. I = stroke in inches. TV = no. of cylinders. n = r.p.m., and Cl = clearance expressed in terms of piston displacement. The following example shows the application of the methods discussed to the determination of the power of an automobile machine. EXAMPLE. Cylinder diameter 4", stroke 4", r.p.m. 1000, no. of cylinders 4. Determine the probable B. H. P. I. By determination of M. E. P. Assume compression carried is 80 pounds, which requires about 28 per cent clearance. From Moss's Tables M. E. P. = 75 X 1.12 = 84 pounds. From equation (4) therefore 84 X ~ X .785 X 4 2 X 500 LH ' P - ' 33000 ' Since the mechanical efficiency is in the neighborhood of .80, B. H. P. = .8 X 6 = 4.8 and for four cylinders Total B. H. P. = 4 X 4. 8 = 19,2 II. By Giildner's Method: nelHd 2 E v 108 L N n = - 1000 X .20 X ^ X 19800 X (AYx -75 = ylJ/ = 46 108 X 250 and total B. H. P. = 4 X 4.6 - 18.4 III. By Association Formula: POWER OF GAS ENGINES 485 IV. By Rice's Formula: 4' X 4.5 X 4 X IQOO/ _J \ 14000 {* h 10X.28j- Of the above results, those obtained by methods I, II, and IV agree very well. Method III gives a result considerably higher, but in view of the fact that the choice of the factor 2.5 is some- what arbitrary, this discrepancy does not mean much. If the factor had been taken equal to three, for instance, as is sometimes done, the result would be. brought down to 21.3 horse-power which is not so far different from the rest. CHAPTER XVI METHODS OF TESTING GAS ENGINES THE actual testing of gas engines, that is, the determination of indicated and brake horse-power, speed, fuel consumption, etc., does not differ materially from the methods long become standard in steam engine practice. But the calculations involved in the proper working up of the data obtained are enough different to have caused engineering societies at home and abroad to set up rules and regu- lations, so-called codes, for testing internal combustion engines. Thus in 1898 the American Society of Mechanical Engineers appointed a committee for this purpose which rendered a final report in an Appendix to the Steam Engine Code in 1901. The British Institution of Civil Engineers followed suit, the pre- liminary and final reports being rendered in March and December, 1905, respectively. Finally the Verein Deutscher Ingenieure adopted a code for the testing of gas engines and gas producers, which was published in the Zeitschrift for November 24, 1906. The latter code was translated nearly entirely by Mr. F. E. Junge and will be found in Power for February, 1907. Of these codes the American and German give specific rules for testing, the latter paying particular attention to acceptance tests of both -engines and gas producers. The British report concerns itself mainly with establishing efficiency standards. In describing the methods followed in the testing of gas engines, it was thought best to give the American code in its entirety, supplementing it by most of the references to the steam engine code contained in its original printed form. This is fol- lowed by a translation of nearly the entire German code, modi- fied where necessary to suit the American conditions. This code is also given because it not only supplements the American code in some important details, but on account of the rules given for the testing of gas producers. 480 METHODS OF TESTING GAS ENGINES 4s? RULES FOR CONDUCTING TESTS OF GAS AND OIL ENGINES. CODE, OF 1901 I. Objects of the Tests. At the outset the specific object of the test should be ascertained, whether it be to determine the fulfilment of a contract guarantee, to ascertain the highest econ- omy obtainable, to find the working economy and the defects as they exist, to ascertain the performance under special conditions, or to determine the effect of changes in the conditions; and the test should be arranged accordingly. Much depends upon the local conditions as to what prepara- tions should be made for a test, and this must' be determined largely by the good sense, tact, judgment, and ingenuity of the expert undertaking it, keeping in mind the main issue, which is to obtain accurate and reliable data. In deciding questions of contract, a clear understanding in regard to the methods of test should be agreed upon beforehand with all parties, unless these are distinctly provided for in the contract. II. General Condition of the Engine. Examine the engine, and make notes of its general condition, and any points of design, construction, or operation which bear on the objects in view. Make a special examination of all the valves by inspecting the seats and bearing surfaces, and note their condition, and see if the piston rings are gas-tight. If the trial is made to determine the highest efficiency, and the examination shows evidence of leakage, the valves and piston rings, etc., should be made tight, and all parts of the engine put in the best possible working condition before starting on the test. III. Dimensions, etc. Take the dimensions of the cylinder, or cylinders, whether already known or not; this should be done when they are hot, and in working order. If they are slightly worn the average' diameter should be determined. Measure, also, the compression space or clearance volume, which should be done, if practicable, by filling the spaces with water previously measured, the proper correction being made for the temperature. IV. Fuel. Decide upon the gas or oil to be used, and if the trial is to be made for maximum efficiency, the fuel should be the best of its class that can readily be obtained, or one that shows the highest calorific power. 488 INTERNAL COMBUSTION ENGINES V. Calibration of Instruments used in the Tests. All instruments and apparatus should be calibrated and their re- liability and accuracy verified by comparison with recognized standards. Apparatus liable to change or to become broken during the tests, such as gages, indicator springs, and ther- mometers, should be calibrated both before and after the experi- ments. The accuracy of all scales should be verified by standard weights. In the case of gas or water meters, special attention should be given to their calibration, both before and after the trial, and at the same rate of flow and pressure as exists during the trial. (a) GAGES. For pressures above the atmosphere, one of the most convenient, and at the same time reliable, standards is the dead-weight testing apparatus which is manufactured by many of the prominent gage makers. It consists of a vertical plunger nicely fitted into a cylinder containing oil or glycerine, through the medium of which the pressure is transmitted to the gage. The plunger is surmounted by a circular stand on which weights may be placed, and by means of which any desired pres- sure can be secured. The total weight, in pounds, on the plunger at any time, divided by the area of the plunger and of the bush- ing which receives it, in square inches, gives the pressure in pounds per square inch. Another standard of comparison for pressures is the mercury column. If this instrument is used, assurance must be had that it is properly graduated with reference to the ever varying zero point; that the mercury is pure, and that the proper correction is made for any difference of temperature that exists, compared with the temperature at which the instrument was graduated. For pressure below the atmosphere, an air 'pump or some other means of producing a vacuum is required, and reference must be made to a mercury gage. Such a gage may be a U-tube having a length of 30 inches or so, with both arms properly filled with pure mercury. (b) THERMOMETERS. Standard thermometers are those which indicate 212 degrees Fahrenheit in steam escaping from boiling water at the normal barometrical pressure of 29.92 inches, the whole stem up to the 212-degree point being surrounded by the steam; and which indicate 32 degrees Fahrenheit in melting METHODS OF TESTING GAS ENGINES 489 ice, the stem being likewise completely immersed to the 32-de- gree point; and which are calibrated for points between and beyond these two references points. We recommend, for tem- peratures between 212 degrees and 400 degrees Fahrenheit, that the comparison of the thermometer be made with the tempera- ture given in Regnault's Steam Tables, the method required being to place it in a mercury well surrounded by saturated steam under sufficient pressure to give the right temperature. The pressure should be accurately determined as pointed out in the above section (a), and the thermometer should be immersed to the same extent as it is under its working condition. Thermometers in practice are seldom used with the stems fully immersed; consequently, when they are compared with the standard, the comparison should be made under like conditions, whatever those happen to be. If pyrometers of any kind are used, they should be compared with a mercury thermometer within its range, and if extreme accuracy is required with an air thermometer, or a standard based thereon, at higher points, care being taken that the medium surrounding the pyrometer, be it air or liquid, is of the same uniform temperature as that surrounding the standard. (c) INDICATOR SPRINGS. The indicator springs should be calibrated with the indicator in as nearly as possible the same condition as to temperature as exists during the trial. This temperature can usually be estimated in any particular case. A simple way of heating the indicator is to subject it to a steam pressure just before calibration. Compressed air, or compressed carbonic acid gas, are suitable for the actual work of calibration. These gases should .be used in preference to steam, so as to bring the conditions as near as possible to those which obtain when the indicators are in actual use. When compressed carbonic acid gas is used, and trouble arises from the clogging of the escape valves with ice, the pipe between the valve and the gas tank should be heated. With both air and carbonic acid gas, the pipes leading to the indicator should also be heated if it is found that they are below the required temperature. The springs may be calibrated for this class of engines under a constant pressure, if desired, and the most satisfactory method is to cover the whole range of pressure through which the indicator acts; first, by grad- 490 INTERNAL COMBUSTION ENGINES ually increasing it from the lowest to the highest point, and then gradually reducing it from the highest to the lowest point, in the manner which has heretofore been widely followed by indicator makers; a mean of the results should be taken. The calibration should be made for at least five. points, two of these being for the pressures corresponding to the maximum and minimum pressures, and three for intermediate points equally distant. The standard of comparison recommended is the dead weight testing apparatus, a mercury column, or a steam gage, which has been proved correct by reference to either of these standards. When the scale of the spring determined by calibration is found to vary from the nominal scale with substantial uniformity, it is usually sufficiently accurate to take the arithmetical mean of the scales found at the different pressures tried. When, how- ever, the scale varies considerably at the different points, and absolute accuracy is desired, the method to be pursued is as follows: Select a sample diagram and divide it into a number of parts by means of lines parallel to the atmospheric line, the number of lines being equal to and corresponding with the num- ber of points at which the calibration of the spring is made. Take the mean scale of the spring for each division and multiply it by the area of the diagram inclosed between two contiguous lines. Add all the products together arid divide by the area of the whole diagram; the result will be the average scale of the spring to be used. If the sample diagram selected is a fair repre- sentative of the" entire set of diagrams taken during the test, this average scale can be applied to the whole. If not, a sufficient number of samples of diagrams representing the various conditions can be selected, and the average scale determined by a similar method for each, and thereby the average for the whole run. (d) GAS METERS. A meter used for measuring gas for a gas engine shouW be calibrated by referring its readings to the displacement of a gasometer of known volume, by comparing it with a standard gas meter of known error, or by passing air through the meter from a tank in which air under pressure is stored. If the latter method is adopted, it is necessary to observe the pressure of the air in the tank and its temperature, both at the tank and at the meter, and this should be done at uniform intervals during the progress of the calibration. The amount of METHODS OF TESTING GAS ENGINES 491 air passing through the meter is computed from the volume of the tank and the observed temperatures and pressures. The volume of the gas thus ascertained should be reduced to the equivalent at a given temperature and atmospheric pressure, corrected for the effect of moisture in the gas, which is ordinarily at the saturation point or nearly so. We recommend that a standard be adopted for gas-engine work, the same as that used in photometry, namely, the equivalent volume of the gas when saturated with moisture at the normal atmospheric pressure at a temperature of 60 degrees Fahrenheit. In order to reduce the reading of the volume containing moist gas at any other tempera- ture to this standard, multiply by the factor 459.4 + 60 b - (29.92 - s) 459.4 + t X 29.4 CALIBRATION OF A WATER METER SKETCH SHOWING METER CONNECTIONS ETC % in which b is the height of the barometer in inches at 32 degrees Fahrenheit, t the temperature of the gas at the meter in degrees Fahrenheit, and s the vacuum in inches of mercury corresponding to the temperature of t obtained from steam tables. (e) WATER METERS. A good method of calibrating a water meter is the following, reference being made to Fig. 16-1. Two tees A and B are placed in the feed pipe, and between them two valves C and D.* The meter is connected between the outlets of the tees A and B. The valves E and F are placed one on each side of the meter. When the meter is running, the valves E and F are opened, and the valves C and D are closed. Should an accident happen to the meter during the test, the valves E and F may be closed, and the valves C and D opened, so as to allow the feed water to flow directly to the point of use. A small bleeder G is FIG. 16-1. 492 INTERNAL COMBUSTION ENGINES opened when the valves C and D are closed, in order to make sure that there is no leakage. A gage is attached at H. When the meter is tested, the valves C, D, and F are closed, and the valves E and / are opened. The water flows from the valve / to a tank placed on weighing scales. In testing the meter the rate of flow should be the same as that on test, and the water leaving the meter is throttled at the valve / until the pressure shown by the gage H is the same as that indicated when the meter is running under the normal conditions. The piping lead- ing from the valve / to the tank is arranged with a swinging joint, consisting merely of a loosely fitting elbow, so that it can be swung readily into the tank or away from it. After the desired pressure and rate of flow have been secured, the end of the pipe is swung into the tank the instant that the pointer of the meter is opposite some graduation mark on the dial, and the water continues to empty into the tank. The tests should be made by starting and stopping at the same graduation mark on the meter dial, and continued until at least 10 or 20 cubic feet are discharged for one test. The water collected in the tank is then weighed. The water passing the meter should always be under pressure in order that any air in the meter may be discharged through the vents provided for this purpose. Care should be taken that there is no air contained in the water. The meter should be tested both before and after the engine trial, and several tests be made of the meter in each case in order to obtain confirmative results. It is well to make preliminary tests to determine whether the meter works satisfactorily before connecting it up for an engine trial. The results should agree with each other for two widely different rates of flow. VI. Duration of Test. The duration of a test should depend upon its character and the objects in view, and in any case the test should be continued until the consecutive readings of the rates at which oil or gas is consumed, taken at say half- hourly intervals, become uniform and thus verify each other. If the object is to determine the working economy, and the period of time during which the engine is usually in motion is some part of twenty-four hours, the duration of the test should be fixed for this number of hours. If the engine is one using coal for generat- ing gas, the test should cover a long enough period to determine METHODS OF TESTING GAS ENGINES 493 with accuracy the coal used in the gas producer; such a test should be of at least twenty-four hours' duration, and in most cases it should extend over several days. VII. Starting and Stopping a Test. In a test for deter- mining the maximum economy of an engine, it should first be run a sufficient time to bring all the conditions to a normal and constant state. Then the regular observations of the test should begin, and continue for the allotted time. If a test is made to determine the performance under working conditions, the test should begin as soon as the regular prepara- tions have been made for starting the engine in practical work, and the measurements should then commence and be continued until the close of the period covered by the day's work. VIII. Measurement of Fuel. If the fuel used is coal fur- nished to a gas producer, the same methods apply for determining the consumption as are used in steam boiler tests. If the fuel used be gas, the only practical method of measure- ment is the use of a meter through which the gas is passed. Gas bags should be placed between the meter and the engine to dimin- ish the variation of pressure, and these should be of a size propor- tionate to the quantity used. When a meter is employed to measure the air used by an engine, a receiver with a flexible diaphragm should be placed between the engine and the meter. The temperature and pressure of the gas should be measured, as also the barometric pressure and temperature of the atmosphere, and the quantity of gas should be determined by reference to the calibration of the meter, taking into account the temperature and pressure of the gas. (See Section V (d)). If the fuel is oil, this can be drawn from a tank which is filled to the original level at the end of the test, the amount of oil re- quired for so doing being weighed; or, for a small engine, the oil may be drawn from a calibrated vessel such as a vertical pipe. In an engine using an igniting flame the gas or oil required for it should be included in that of the main supply, but the amount so used should be stated separately, if possible. IX. Measurement of Heat-Units Consumed by the Engine. The number of heat units used is found by multiplying the num- ber of pounds of coal or oil or the cubic feet of gas consumed, by the total heat of combustion of the fuel as determined by a cal- 494 INTERNAL COMBUSTION ENGINES orimeter test. In determining the total heat of combustion no deduction is made for the latent heat of the water vapor in the products of combustion. There is a difference of opinion on the propriety of using this higher heating value, and for purposes of comparison care must be taken to note whether this or the lower value has been used. The calorimeter recommended for determining the heat of combustion is the Mahler, for solid fuels or oil, or the Junker for gases, or some form of calorimeter known to be equally reliable. (See Chapter VI, or Poole on " The Calorific Power of Fuels.") It is sometimes desirable, also, to have a complete chemical analysis of the oil or gas. The total heat of combustion may be computed, if desired, from the results of the analysis, and should agree well with the calorimeter values. In using the gas calorimeter, which involves the determina- tion of the volume instead of the weight of the gas, it is important that this should be reduced to the same temperature as that corresponding to the conditions of the engine trial. The formula to be used for making the reduction is that already given in Section V (d). For the purpose of making the calorimeter test, if the fuel used is coal for generating gas in a producer, or oil, samples should be taken at the time of the engine trial, and carefully preserved for subsequent determination. If gas is used, it is better to have a gas calorimeter on the spot, samples taken, and the calorimeter test made while the trial is going on. X. Measurement of Jacket Water to Cylinder or Cylinders. The jacket water may be measured by passing it through a water meter or allowing it to flow from a measuring tank before enter- ing the jacket, or by collecting it in tanks on its discharge. XI. Indicated Horse-Power. The directions given for deter- mining the indicated horse-power for steam engines apply in all respects to internal combustion engines. The indicated horse-power should be determined from the average mean effective pressure of diagrams taken at intervals of twenty minutes, and at more frequent intervals if the nature of the test makes this necessary. With variable loads, such as those of engines driving generators for electric railroad work, and of rubber-grinding and rolling mill engines, the diagrams cannot be METHODS OF TESTING GAS ENGINES 495 taken too often. In cases like the latter, one method of obtain- ing suitable averages is to take a series of diagrams on the same blank card without unhooking the driving cord, and apply the pencil at successive intervals of ten seconds until two minutes' time or more has elapsed, thereby obtaining a dozen or more indications in the time covered. This tends to insure the deter- mination of a fair average for that period. In taking diagrams for variable loads, as indeed for any load, the pencil should be applied long enough to cover several successive revolutions, so that the variations produced by the action of the governor may be properly recorded. To determine whether the governor is subject to what is called "racing" or "hunting," a "variation diagram" should be obtained; that is, one in which the pencil is applied a sufficient time to cover a complete cycle of variations. When the governor is found to be working in this manner, the defect should be remedied before proceeding with the test. AUTHOR'S NOTE. When the engine is governed by the hit-and-miss principle the diagrams taken on one card should in any case cover the series of consecutive explosions, and the mean diagram should be used as the basis of calculations. The most satisfactory driving rig for indicating seems to be some form of well-made pantagraph, with driving cord or fine annealed wire leading to the indicator. The reducing motion, whatever it may be, and the connections to the indicator, should be so perfect as to produce diagrams of equal lengths, and pro- duce a proportionate reduction of the motion of the piston at every point of the. stroke, as proved by test. To test the accuracy of the reducing motion without making special preparations for a thorough examination, it is sufficient to make a comparison between the actual proportion of the stroke covered and the apparent proportion measured on the indicator, and see how they agree. This may be done on a large engine by making the comparison \vherever it happens to stop, and repeat- ing the comparison when it has stopped with the piston at some other point of the stroke. With an engine which can be turned over by hand, or where auxiliary power is provided for moving it, the comparison may be made at a number of equidistant points in the stroke. To make the test properly, a diagram should be taken just before, stopping, and this will serve as a 496 INTERNAL COMBUSTION ENGINES reference for the measurements taken after stopping. The actual proportion of stroke covered is determined by measuring the distance which the piston has moved and comparing it with the whole length of the stroke, making sure that the slack has all been taken up. To obtain the apparent indication from the dia- gram, the indicator pencil is moved up and down with the finger so as to make a vertical mark on the diagram, and the distance of this mark from the beginning of the diagram compared to the whole length of the diagram is the proportion desired. It is necessary, of course, to go through these operations with- out changing in any way the adjustment of the driving cord of the indicator, or any part of the mechanism that would alter the movements of the indicator. In the manipulation of the indicator it is important to keep the instrument in clean condition and preserve it in mechanically good order. Ordinary cylinder oil is the best material to use for lubricating the indicator piston for pressures above the atmos- phere. It is better to have the piston fit the cylinder rather loosely so as to get absolute freedom of motion than to have a mechanically accurate fit. In the latter case, extreme care and frequent cleanings are required to obtain good diagrams. No diagrams should be accepted in which there is any appearance of want of freedom in the movement of the mechanism. A ragged or serrated line in the region of the expansion or compression lines is a sure indication that the piston or some part of the mechanism sticks; and when this state of things is revealed the indicator should not be trusted, but the cause should be ascer- tained and a suitable remedy applied. An indicator which is free when subjected to a steady pressure, as it is under a test of the springs for calibration, should be able to produce the same horizontal line, or substantially the same, after pushing the pencil down with the finger, as that traced after pushing the pencil up and subsequently tapping it lightly. When the pencil is moved by the finger, first up and then down, the piston being subjected to the pressure, the movement should appear smooth to the sense of feeling. The pipe connections for indicating gas and oil engines should be removed as far as possible from the ports and ignition devices, and made preferably in the cylinder head. The pipes should be METHODS OF TESTING GAS ENGINES 497 as short and direct as possible. Avoid the use of long pipes, otherwise explosions of the gas in these connections may occur. Ordinary indicators suitable for indicating steam engines are much too lightly constructed for gas and oil engines. The pencil mechanism, especially the pencil arm, needs to be very strong to prevent injury by the sudden impact at the instant of the ex- plosion; a special gas-engine indicator is required for satisfactory work, with a small piston and a small spring. See Chapter I for the description of various indicators. XII. Brake Horse-Power. The determination of the brake horse-power, which is very desirable, is the same for internal combustion as for steam engines. This term applies to the power delivered from the fly-wheel shaft to the engine. It is the power absorbed by a friction brake applied to the rim of the wheel, or to the shaft. A form of brake is preferred that is self-adjusting to a certain extent, so that it will, of itself, tend to maintain a constant resistance at the rim of the wheel. One of the simplest brakes for comparatively small engines, which may be made to em- body this principle, consists of a cotton or hemp rope, or a number of ropes, en- circling the wheel, arranged with weigh- ing scales or other means for showing the strain. Anordi- FlG - 16 ~ 2 ' nary band brake may also be constructed so as to embody the principle. The wheel should be provided with interior flanges for holding water used for keeping the rim cool. A self-adjusting rope brake is illustrated in Fig. 16-2, where it will be seen that, if the friction at the rim of the wheel increases, it will lift the weight A, which action will diminsh the tension in the end B of the rope, and thus prevent a further increase in the friction. The same device can be used for a band brake of the 498 INTERNAL COMBUSTION ENGINES ordinary construction. Where space below the wheel is limited, a cross bar, C, supported by a chain tackle exactly at its center point, may be used as shown in Fig. 16-2, thereby causing the action of the weight on the brake to be upward. A safety stop should be used with either form, to prevent the weights being accidentally raised more than a certain amount. The water-friction brake is specially adapted for high speeds and has the advantage of being self-cooling. The Alden brake is also self-cooling and is capable of fine adjustment. A water-friction brake is shown in Fig. 16-3. It consists of two circular discs, A and B, attached to the shaft C, and revolv- ing in a case, E, be- tween fixed planes. The space between the discs and the planes is supplied with running water, which enters at D and escapes at the cocks F, G, and H. The friction of the water against the surfaces constitutes a resistance which absorbs the desired power, and the heat generated within is carried away by the water itself. The water is thrown outward by centrifugal action and fills the outer portion of the case. The greater the depth of the ring of water, the greater amount of power absorbed. By suitably adjusting the amount of water entering and leaving any desired power can be obtained. Water- friction brakes have been used successfully at speeds of over 20,000 revolutions per minute. For methods of computing brake horse-power see Chapter I. XIII. Speed. There are several reliable methods of ascer- taining the speed, or the number of- revolutions of the engine crank shaft per minute. The simplest is the familiar method of counting a number of turns for a period of one minute with the eye fixed on the second hand of a time piece. Another is the use of a counter held for a minute or a number of minutes against FIG. 16-3. METHODS OF TESTING GAS ENGINES 499 the end of the main shaft. Another is the use of a reliable tach- ometer held likewise against the end of the shaft. The most reliable method, and the one we recommend, is the use of a continuous recording engine register or counter, taking the total reading each time that the general test data are recorded, and computing the revolutions per minute corresponding to the difference in the readings of the instrument. When the speed is above 250 revolutions per minute, it is almost impossible to make a satisfactory counting of the revolutions without the use of some kind of mechanical counter. The determination of variation of speed during a single revo- lution, or the effect of the fluctuation due to sudden changes of the load, is also desirable, es- pecially in engines driving electric generators used for lighting pur- poses. There is at present no recognized standard method of making such determinations, and if such are desired, the method employed may be devised by the person making the test and de- scribed in detail in the report. One method suggested for determining the instantaneous variation of speed which accom- panies a change of load is as follows: A screen containing a narrow slot is placed on the end of a bar and vibrated by means of electricity. A corresponding slot in a stationary screen is placed parallel and nearly touch- FIG. 16-4. ing the vibrating screen, and the two screens are placed a short distance from the fly-wheel of the engine in such a position that the observer can look through the two slots in the direction of the spokes of the wheel. The vibrations are adjusted so as to con- form to the frequency with which the spokes of the wheel pass the slots. 500 INTERNAL COMBUSTION ENGINES When this is done the observer viewing the wheel through the slots sees what appears to be a stationary fly-wheel. When a change in the velocity of the fly-wheel occurs, the wheel appears to revolve either backward or forward according to the direction of the change. By careful observations of the amount of this motion, the change of angular velocity during any given time is revealed. Experiments that have been made with a device of this kind show that the instantaneous gain of velocity, upon suddenly removing all the load from an engine, amounted to from one- sixth to one-quarter of a revolution of the wheel. In an engine which is governed by varying the number of ex- plosions or working cycles, a record should be kept of the number of explosions per minute; or if the engine is running at nearly maximum load, by counting the number of times the governor causes a miss in the explosions. One way of mechanically recording the explosions is to attach to the exhaust pipe a cylinder and piston arranged so that the pressure caused by the exhaust gases operate against a light spring and moves a register, which, is provided for automatically counting the number. AUTHOR'S NOTE. An instrument for this purpose has been devised by R. Mathot. The following description is from his book on " Modern Gas Engines and Producer Gas Plants:'' The instrument, Fig. 16-4, is somewhat similar in form to the ordinary indicator. Its record, however, is made on a paper tape which is continuously .unwound. The cylinder c is provided with a piston p, about the stem of which a spring .s is coiled. A clock train contained in the chamber b unwinds the strip of paper from the roll p' and draws it over the drum p", where the. pencil t leaves the mark. The tape is then rewound on the spindle p'" ' . A small stylus or pencil / traces the atmospheric line on the paper as it passes over the drum p" '. In order to obviate the binding of the piston p when subjected to the high temperature of the explosions, the cylinder c is provided with a casing e in which water is circulated by means of a small rubber tube which fits over the nipple e' '. This recorder analyzes with absolute precision the work of all engines, whatever may be their speed. It gives a continuous graphic record from which the number of explosions, together with the initial pressure of each, can be determined, and the order of their succession. Con- sequently the regularity or irregularity of the variations can be observed and traced to the secondary influences producing them, such as the action of the inlet and outlet valves and the sensiti\>eness of the governor. It renders it possible to estimate the resistance to suction and the back pressure due to expelling the burnt gases, the chief causes of loss in efficiency in high-speed engines. Furthermore, the influence of compression is markedly shown from the diagram obtained. The recorder is mounted on the engine; its piston is driven back by each of the explosions to a height corresponding with their force; and the stylus METHODS OF TESTING GAS ENGINES 501 or pencil controlled by the lever t records them side by side on the moving strip of paper. The speed with which this strip is unwound conforms with the number of revolutions of the engine to be tested, so that the records of the explosions are placed side by side clearly and legibly. Their succession indicates not only the number of explosions and of revolutions which occur in a given time, but also their regularity, the num- ber of misfires. The pressure of the explosions is measured by a scale con- nected with the recorder-spring. By employing a very weak spring which flexes at the bottom simply by the effect of the compression in the engine cylinder, it is possible to ascertain the amount of the resistance to suction and to the exhaust. It is simply sufficient to compare the explosion record with the atmospheric line, traced by the stylus /. By means of this appa- ratus, and of the records which it furnishes, it is possible analytically to regulate the work of an engine, to ascertain the proportion of air, gas, or hydro- carbon which produces the most powerful explosion, to regulate the com- pression, the speed, the time of ignition, the temperature, and the like. XIV. Recording the Data. The time of taking weights and every observation should be recorded, and note made of every event, however unimportant it may seem to be. The pressures, temperatures, meter readings, speeds, and other measurements should be observed every 20 or 30 minutes when the conditions are practically uniform, and at more frequent intervals if they are variable. Observations of the gas or oil measurements should be taken with special care at the expira- tion of each hour, so as to divide the test into hourly periods, and reveal the uniformity, or otherwise, of the conditions and results as the test goes forward. All data and observations should be kept on suitable prepared blank sheets or in notebooks. XV. Uniformity of Conditions. When the object of the test is to determine the maximum economy, all the conditions relating to the operation of the engine should be maintained as constant as possible during the trial. XVI. Indicator Diagrams and their Analysis. SAMPLE DIAGRAMS: Sample diagrams nearest to the mean should be selected from those taken during the trial and appended to the tables of the results. If there are separate compression or feed cylinders, the indicator diagrams from these should be taken and the power deducted from that of the main cylinder. XVII. Standards of Economy and Efficiency. The hourly consumption of heat, determined as pointed out in Article IX, divided by the indicated or the brake horse-power, is the standard expression of engine economy recommended. In making comparisons between the standard for internal 502 INTERNAL COMBUSTION ENGINES combustion engines and that for steam, it must be borne in mind that the former relates to energy concerned in the generation of the force employed, whereas in the steam engine it does not relate to the entire energy expended during the process of com- bustion in the steam boiler. The steam engine standard does not cover the losses due to combustion, while the internal com- bustion engine standard, in cases where a crude fuel such as oil is burned in the cylinder, does cover these losses. To make a direct comparison between the two classes of engines considered as complete plants for the production of power, the losses in generating the working agent must be taken into account in both cases and the comparison must be on the basis of the fuel used: and not only this, but on the basis of the same or equivalent fuel used in each case. In such a comparison, where producer gas is used, and the producer is included in the plant, the fuel con- sumption, which will be the weight of coal in both cases, may be directly compared. The thermal efficiency ratio per indicated horse-power or per brake horse-power for internal combustion engines is obtained in the same manner as for steam engines, and is expressed by the fraction 2545 B.T.U. per H.P. per hour XVIII. Heat Balance. For purposes of scientific research, a heat balance should be drawn which shows the manner in which the total heat of combustion is expended in the various processes concerned in the working of the engine. It may be divided into three parts: first, the heat which is converted into the indicated or brake work; second, the heat rejected in the cooling water of the jackets; and third, the heat rejected in the exhaust gases, together with that lost through incomplete combustion and radiation. To determine the first item, the number of foot-pounds of work performed by, say, one pound or one cubic foot of the fuel is determined; and this quantity divided by 778, which is the me- chanical equivalent of one British thermal unit, gives the number of heat units desired. The second item is determined by meas- uring the amount of cooling water passed through the jackets, equivalent to one pound or one cubic foot of fuel consumed, and METHODS Of TESTING GAS ENGINES 503 calculating the amount of heat rejected, by multiplying this quantity by the difference in the sensible heat of the water leav- ing the jacket and that entering. The third item is obtained by the method of differences; that is, by subtracting the sum of the first two items from the total heat supplied. The third item can be subdivided by computing the heat rejected in the exhaust gases as a separate quantity. The data for this computation are found by analyzing the fuel and the exhaust gases, or by measuring the quantity of air admitted to the cylinder in addi- tion to that of the gas or oil. For methods of making fuel and exhaust gas computations, see Chapter VI. XIX. Report of Test. The data and results of a test should be reported in the manner outlined in one of the following tables, the first of which gives a complete summary when all the data are determined, and the second is a shorter form of report in which some of the minor items are omitted. XX. Temperatures Computed at Various Points of the In- dicator Diagram. The computation of temperatures correspond- ing to various points in the indicator is, at best, approximate. It is possible only where the temperature of one point is known or assumed, cr where the amount of air entering the cylinder along with the charges of gas or oil, and the temperature of the exhaust gases, is determined. If the amount of air is determined for a gas engine, together with the necessary temperatures, so that the volume and the temperature of the air entering the cylinder per stroke, and that of the gas are known, we may, by combining this with the other data, compute the temperature for a point in the compression curve. In this computation we must allow for the volume of the exhaust gases remaining in the cylinder at the end of the stroke. The temperature at the point in the compression curve where it meets or crosses the atmospheric line will be given by the formula: where V is the total volume corresponding to the point where the compression curve meets or crosses the atmospheric line; V 504 INTERNAL COMBUSTION ENGINES the volume of the air at atmospheric pressure entering the cylin- der during each working cycle, reduced to the equivalent volume at 32 degrees Fahrenheit; V" the volunie of the gas consumed per cycle reduced to the equivalent at atmospheric pressure and 32 degrees Fahrenheit; and V"" the volume of the exhaust gases retained in the cylinder reduced to the same basis. To reduce the actual volumes to those at 32 degrees Fahrenheit, multiply by the ratios of 491.4- (T' + 459.4), where T' is the observed temperature of the air and of the gas used as fuel. For the exhaust gases retained in the cylinder at the end of the stroke T' may be taken as the temperature of the exhaust gases leaving the engine, provided the engine is not of the "scavenging" type. Having determined the temperature of a point in the com- pression curve, the temperature of any point in the diagram may be found by the equation 7\ = (T + 459.4) ^^i - 459.4. . . .(B). Here 7\ is the desired temperature of any point in the diagram where the absolute pressure is P l and the total volume V lf and P and V are the corresponding quantities for the point in the compression line having the temperature T computed from the formula (A). Formula (B) holds only where the weight of the gases contained in the cylinder is constant. It is also assumed in this formula that the density of the gas compared to air at the same tempera- ture and pressure is the same before and after the explosion. A second method may be employed, provided the air which enters the cylinder is measured. This will allow for any dif- ference in the density of the gas before and after explosion, and more exact values for temperatures on the expansion curve may be obtained than by the first method. In this method the density of the exhaust gases compared to air at the same temperature and pressure is computed, assuming perfect combustion, and including the effect of the water vapor present; and from this density the volume of the gases exhausted per cycle is determined. . If this volume exhausted per cycle, added to the volume of the gas retained in the clearance space at the end of the stroke, be called V in equation B, and T be the METHODS OF TESTING GAS ENGINES 505 observed temperature of the exhaust gases, this equation may be used for determining the temperature of any point in the dia- gram in the way already described. This method is more com- plicated than the first, as it involves the determination of the theoretical density after explosion, but it possesses the advantage that it may be applied to an oil as well as to a gas engine. A third method of computing the temperature of the various points in the diagram may be employed where analyses of the exhaust gases as well as of the fuel have to be made. This method is more complicated than the first, but, in common with the second, it possesses the advantage that it may be applied to an oil as well as to a gas engine. In applying the third method the volume of the exhaust gases discharged per working cycle would be given by the formula: 2 where D is the density of the exhaust gases at their observed temperature, computed from the analysis, assuming the vapor of water produced through burning the hydrogen in the fuel to be in a gaseous state; R the weight of the air which enters the cylinder per pound of fuel consumed per working cycle; the value of R, providing there are no unconsumed hydrocarbons, may be computed by employing the formula: 7? - NC .33(C0 2 + CO)" where N, CO 2 and CO represent the proportions, by volume, of the several constituents of the exhaust gases, and C the weight of carbon consumed and converted to CO 2 or CO per pound of fuel burned, computed from the analysis of the fuel and of the exhaust gases. Having determined the volume V 2 of the exhaust gases, formula (B) may be used in computing the temperature, in which case T will represent the temperature of the exhaust gases as in the second method, P the pressure of the exhaust, and V the volume of the exhaust gases V 2 discharged per stroke, added to the volume of the gases retained in the cylinder at the end of the stroke. The value of R given in equation (D) is approximate, on 506 INTERNAL COMBUSTION ENGINES account of the fact that the percentage of N should be that due to the air alone, and not that due to the air in addition to that contained in the fuel gas. Where extreme accuracy is desired, the value found for R may be used to determine the percentage of N which in the analysis of the exhaust gases is due to the N in the fuel gas, and this value may be subtracted from the total N shown by the analysis of the fuel gases, in order to obtain the correct value of N to be used in equation (C). TABLE NO. 1 DATA AND RESULTS OF TEST OF GAS OR OIL ENGINE Arranged according; to the Complete Form advised by the Engine Test Com- mittee, American Society of Mechanical Engineers. Code of 1902 1 . Made by of on engine located at to determine 2. Date of trial 3. Type of engine, whether oil or gas 4. Class of engine (mill, marine, motor for vehicle, pumping, or other) ..... 5. Number of revolutions for one cycle, and class of cycle ........ ........ 6. Method of ignition ............................................ ... 7. Name of builders ................................................. 8. Gas or oil used ................................................... (a) Specific gravity ............................... deg. Fahr. (6) Burning point ................................ (c) Flashing point . . ............................... " 9. Dimensions of engine : 1st Cyl. 2d Cyl. (a) Class of cylinder (working or for compressing the charge) .............................. (6) Vertical or horizontal ..... ............... t (c) Single or double acting ................... . (d) Cylinder dimensions ....................... Bore ................................ in. Stroke ............................. . ft. Diameter piston rod ................. . . in. Diameter tail rod ...................... in. (e) Compression space or clearance in per cent of volume displaced by piston per stroke. . Head end .............................. Crank end ............................. Average ............ . ................. (/) Surface in square feet (average) ....... ..... Barrel of cylinders ..................... Cylinder heads Clearance and ports Ends of piston Piston rod ..... METHODS OF TESTING GAS ENGINES 507 (g) Jacket surfaces or internal surfaces of cylinder heated by jackets, in square feet Barrel of cylinder Cylinder heads : Clearance and ports (h) Horse-power constant for one Ib. M. E. P., and one revolution per minute 10. Give description of main features of engine and plant, and illustrate with drawings of same given on an appended sheet. Describe method of governing. State whether the conditions were constant throughout the test. Total Quantities 11. Duration of test . . hours. 12. Gas or oil consumed cu. ft. or Ibs. 13. Air supplied in cubic feet cubic feet. 14. Cooling water supplied to jackets 15. Calorific value of gas or oil by calorimeter test, determined by calorimeter B. T. U. Hourly Quantities 16. Gas or oil consumed per hour cu. ft. or Ibs. 17. Cooling water supplied per hour Ibs. Pressures and Temperatures 18. Pressure at meter (for gas engine) in inches of water ins. 19. Barometric pressure of atmosphere: (a) Reading of height of barometer (6) Reading of temperature of barometer deg. Fahr. (c) Reading of barometer corrected to 32 Fahr ins. 20. Temperature of cooling water: (a) Inlet deg. Fahr. (6) Outlet 21. Temperature of gas at meter (for gas engine) 22. Temperature of atmosphere: (a) Dry-bulb thermometer (6) Wet-bulb thermometer (c) Degree of humidity per cent. 23. Temperature of exhaust gases deg. Fahr. How determined Data Relating to Heat Measurement 24. Heat units consumed per hour (Ibs. of oil or cu. ft. of gas per hour multiplied by the total heat of combustion) B. T. U. 25. Heat rejected in cooling water: (a) Total per hour (6) In per cent of heat of combustion of the gas or oil consumed per cent. 26. Sensible heat rejected in exhaust gases above temperature of inlet air: (a) Total per hour B. T. U. (6) In per cent of heat of combustion of the gas or oil con- sumed per cent. 27. Heat lost through incomplete combustion and radiation per hour: (a) Total per hour B. T. U. (6) In per cent of heat of combustion of the gas or oil con- sumed ; , . . . . per cent. 508 INTERNAL COMBUSTION ENGINES Speed, Etc. 28. Revolutions per minute rev. 29. Average number of explosions per minute How determined 30. Variation of speed between no load and full load rev. 31. Fluctuation of speed on changing from no load to full load measured by the increase in the revolutions due to the change. Indicator Diagrams 1st Cyl. 2d Cyl. 32. Pressure in Ibs. per sq. in. above atmosphere: (a) Maximum pressure (6) Pressure just before ignition (c) Pressure at end of expansion (d) Exhaust pressure 33. Temperatures in deg. Fahr. computed from diagrams: (a) Maximum temperature (not necessarily at maximum pressure) (6) Just before ignition (c) At end of expansion (d) During exhaust 34. Mean effective pressure in Ibs. per sq. in Power 35. Power as rated by builders: (a) Indicated horse-power H. P. (6) Brake 36. Indicated horse-power actually developed: First cylinder Second cylinder " Total 37. Brake H. P., electric H. P., or pump H. P., according to the class of engine 38. Friction indicated H. P. from diagram, with no load on engine and computed for average speed 39. Percentage of indicated H. P. lost in friction per cent. Standard Efficiency Results 40. Heat units consumed by the engine per hour: (a) Per indicated horse-power B. T. U. (6) Per brake horse-power 41. Heat units consumed by the engine per minute: (a) Per indicated horse-power " (6) Per brake horse-power " 42. Thermal efficiency ratio: (a) Per indicated horse-power per cent. (6) Per brake horse-power " Miscellaneous Efficiency Results 43. Cubic feet of gas or Ibs. of oil consumed per H. P. per hour: (a) Per indicated horse-power (6) Per brake horse-power Heat Balance 44. Quantities given in per cents of the total heat of combustion of the fuel: METHODS OF TESTING GAS ENGINES 509 (a) Heat equivalent of indicated horse-power per cent. (6) Heat rejected in cooling water : (c) Heat rejected in exhaust gases and lost through radia- tion and incomplete combustion " Sum = 100 " Subdivisions o*f Item (c) : (cl) Heat rejected in exhaust gases (c2) Lost through incomplete combustion (c3) Lost through radiation, and unaccounted for ..... Sum = Item (c) Additional Data Add any additional data bearing on the particular objects of the test or relating to the special class of service for which the engine is to be used. Also give copies of indicator diagrams nearest the mean and the corresponding scales. Where analyses are made of the gas or oil used as fuel, or of the exhaust gases, the results may be given in a separate table. TABLE NO. 2 DATA AND RESULTS OF STANDARD HEAT TEST OF GAS OR OIL ENGINE Arranged according to the Short Form advised by the Engine Test Committee, American Society of Mechanical Engineers. Code of 1902. 1 . Made by of on engine located at to determine . 2. Date of trial 3. Type and class of engine 4. Kind of fuel used (a) Specific gravity deg. Fahr. (6) Burning point (c) Flashing point 5. Dimensions of engine : IstCyl. 2dCyl. (a) Class of cylinder (working or for compressing the charge) (6) Single or double acting (c) Cylinder dimensions: Bore in. Stroke ft. Diameter piston rod in. (d) Average compression space, or clearance in per cent (e) Horse-power constant for one Ib. M. E. P. and one revolution per minute Total Quantities 6. Duration of test hours. 7. Gas or oil consumed cu. ft. or Ibs. 8. Cooling water supplied to jackets 9. Calorific value of fuel by calorimeter test, determined by . . calorimeter B. T. U. 510 INTERNAL COMBUSTION ENGINES Pressures and Temperatures 10. Pressure at meter (for gas engine) in inches of water ins. 1 1 . Barometric pressure of atmosphere : (a) Reading of barometer (6) Reading corrected to 32 degs. Fahr ! 12. Temperature of cooling water: (a) Inlet deg. Fahr. (6) Outlet (c) Degree of humidity 13. Temperature of gas at meter (for gas engine) 14. Temperature of atmosphere: (a) Dry bulb thermometer (6) Wet bulb thermometer 15. Temperature of exhaust gases Data Relating to Heat Measurement 16. Heat units consumed per hour (pounds of oil or cubic feet of gas per hour multiplied by the total heat of combus- tion) . B.T.U. 17. Heat rejected in cooling water per hour Speed, Etc. 18. Revolutions per minute rev. 19. Average number of explosions per minute Indicator Diagrams 20. Pressure in Ibs. per sq. in. above atmosphere: 1st Cyl. 2d Cyl. (a) Maximum pressure (6) Pressure just before ignition (c) Pressure at end of expansion (d) Exhaust pressure (e) Mean effective pressure Power 21. Indicated horse-power: First cylinder H. P. Second cylinder " Total 22. Brake horse-power " 23. Friction horse-power by friction diagrams " 24. Percentage of indicated horse-power lost in friction per cent. Standard Efficiency, and Other Results 25. Heat units consumed by the engine per hour: (a) Per indicated horse-power B. T. U. (5) Per brake horse-power " 26. Pounds of oil or cubic feet of gas consumed per hour: (a) Per indicated horse-power Ibs. or cu. ft. (6) Per brake horse-power . . . " Additional Data Add any additional data bearing on the particular objects of the test or relating to the special class of service for which the engine is to be used. Also give copies of indicator diagrams nearest the mean, and the correspond- ing scales. METHODS OF TESTING GAS ENGINES 511 RULES FOR TESTING GAS PRODUCERS AND GAS ENGINES. CODE OF THE GERMAN SOCIETY OF ENGINEERS * All metric units have been transposed to English units The preparation of the following rules for making gas engine and producer tests was undertaken by a committee appointed from the Verein deutscher Ingenieure, in collaboration with the German Society of Engine Builders, with the view of establishing definite general regulations governing such tests. It is desirable, by specifying the important proportions of the examined plants and the conditions under which the results were attained, to in- sure that these results are not only applicable to a single case, but that they have general value. To attain this end it is neces- sary that all data should be given uniformly according to a code of regulations such as that here presented. The execution of such tests should be intrusted only to per- sons possessing the required expert knowledge and practical experience. These persons must make a trial plan, or schedule, appropriate to the individual case in hand, which, in many in- stances, will not require that all of the investigations stipulated in the general code are actually carried out. They must further examine the instruments for measuring or recording purposes as to their fitness and must compile the results. The following rules, the adoption or selection of which must be left to the soundness of judgment of the investigator, are intended to serve as a basis on which to proceed. GENERAL REGULATIONS Object of Investigation 1. The object of a test made on a producer-gas plant may be to determine : (a) The quantity, composition, and calorific value of the fuel consumed. (&) The quantity, composition, and heat value of the gas produced. (c) The degree of efficiency of the producer-gas plant. (d) The separate heat Igsses in the plant. * Mainly from F. E. Junge's translation in Power, Feb., 1907. 512 INTERNAL COMBUSTION ENGINES (e) The quantity of impurities contained in one cubic meter or one cubic foot of gas (dust, tar, sulphur, etc.). (/) The moisture contents of the gas. (g) The water consumption of the producer-gas plant, either total or in the separate parts. (h) The mechanical work required for operating the plant, including apparatus. (i) The duration of time required for starting. (k) The stand-by losses during intervals of shutting down day or night. 2. The object of a test made on an internal combustion (gas) engine may be to determine: (a) The indicated capacity and the effective output. (b) The mechanical efficiency. (c) The fuel consumption and the heat consumption per horse-power hour. (d) The consumption of lubricants, separately for cylinder and engine. (e) The consumption of water and the heat conducted to the cooling water. (/) The fluctuations in number of revolutions. (g) The composition of exhaust gases. NUMBER AND 'DURATION OF TESTS Admissible Fluctuations 3. The number and duration of trials are determined by the purpose of the test as well as by a consideration of the conditions of installation and operation, and must be settled and previously arranged according to paragraphs four to eight. For trials of special importance the results of which are decisive for acceptance tests, for penalties or for premiums, this item deserves special consideration. 4. Acceptance tests should be made if possible immediately after a plant has been put into actual operation; the manufacturers, however, must be granted a reasonable time for making preliminary trials of their own and for carrying out alterations or improvements then necessary. The length of this time and other conditions are best agreed upon when drawing up the delivery contract. METHODS OF TESTING GAS ENGINES 513 5. In order to be able to get acquainted with the operation of the plant that is to be tested, to find time for examining the testing devices employed, and to break in observers and assistants, it is desirable that preliminary trials be allowed. 6. If the fuel consumption in gas producers is to be deter- mined, the trial run must be extended over at least eight hours under constant conditions and without interruptions. 7. For determining the consumption of liquid or gaseous fuel and provided the conditions are constant, it is sufficient for the higher loads to extend measurements over an hour, while for finding the consumption at the lower loads, measurements of even shorter duration are sufficient. To ascertain the constancy of the conditions the temperature of the outflowing cooling water must be read from time to time. These rules as to the duration of the tests are made with the provision that no interruption or disturbance of the trial takes place, and that intermediate read- ings show only slightly varying values for the consumption. 8. If only the mechanical efficiency of an engine is to be determined, trials of short duration under constant conditions are sufficient; but at least ten sets of indicator cards should be taken. 9. For investigations of special importance at least two tests should be made, one after the other. They should be accepted only if no interruptions occurred and if the results show no greater deviations than those due to unavoidable errors of observation. The mean of the two results is to be taken as the final result. 10. The extent to which the capacity and consumption of gas may differ from the guarantee or contract figures, without justifying a claim of breach of contract, is to be clearly stated before the tests (either in the original contract or in the schedule of tests). When no other agreement has been previously arrived at, the capacity guarantee is regarded as fulfilled if the figure obtained in the test is not more than 5 per cent below the value on which the guarantee was based. This margin, how- ever, is allowable only for the maximum output which was promised beyond the guaranteed continuous output. The latter must be rendered by the engine under all circumstances. The consumption of fuel and water as determined on test should not exceed the guaranteed figures by more than 5 per cent 514 INTERNAL COMBUSTION ENGINES even if, during the trial, the engine load fluctuated somewhat from the load upon which the guarantee was based, provided that fluctuation do not exceed an average of 5 per cent of such load, or a maximum of 15 per cent. Since it is often impossible when making .tests to have the internal combustion engine work at exactly the effective (horse- power) capacity on which the guarantee agreed upon in the con- tract is based, it is recommended that the agreement shall specify the expected fuel consumption for higher and lower outputs. The same provision is perferably made also with gas producers. UNITS OF MEASUREMENT AND DESIGNATIONS 11. When giving pressure data it must be stated whether absolute pressures or gage pressures above or below the at- mospheric are meant. Absolute pressure equals atmospheric pressure plus gage pressure. 12. All temperature and heat measurements refer to the Fahrenheit scale. 13. The mechanical equivalent of heat is taken at 778 foot- pounds. 14. The calorific value of a fuel is to be taken as its lower heating value; that is, the heat which is liberated by the com- plete combustion of the fuel when the burnt products are cooled down to the original (room) temperature at constant pressure, it being assumed, however, that the water of combustion and the moisture contained in the fuel remain vaporized. The calorific value must be based on the unit quantity or weight of original fuel, without deducting ash, moisture, etc., and is to be ex- pressed in heat units. For both solid and liquid fuels the unit of weight is the pound. The heat value of gaseous fuels is based on one cubic foot at 32 degrees Fahrenheit, and 760 millimeters barometer pressure, or must be expressed in thermal units as "effective" heat value, that is,' reduced to one cubic foot of actual gas used. If not specially stated, it is always understood that the heat value recorded is that of gas at 32 degrees Fahrenheit and 760 milli- meters barometer pressure. In this country the general standard so far recommended seems to indicate for "standard gas" a temperature of 60 degrees METHODS OF TESTING GAS ENGINES 515 Fahrenheit, and a pressure of 14.7 pounds per square inch, cor- responding to the usual atmospheric pressure. 15. The efficiency of a gas-producer plant is the ratio of the latent heat contained in the gas as produced to the heat of com- bustion of the total weight of fuel consumed in the plant, both items being computed from the lower heating value. In pro- ducer-gas plants having a separately fired steam boiler, it is advisable also to determine the ratio of the heat which is chemi- cally bound in the producer gas to the heat equivalent of that portion of the fuel which is consumed in the producer proper for making such gas. 16. The unit of measurement used for the power or work output of an internal combustion engine is the horse- power equal to 33000 foot-pounds per minute. It must be clearly stated whether the indicated power, or the useful or available power, is meant. If not otherwise designated it is understood that the figures refer to the useful or available output. 17. The indicated power of the engine, or the indicated work, is x the difference between the total power developed or work done, ana the indicated power, or work, which is consumed within the engine; in short, the difference between the positive and the negative indicated power or work. AUTHOR'S NOTE. This is the provision which caused considerable dis- cussion among gas-engine experts some time ago. It means as it stands, that in a 4-cycle machine, the indicated horse-power is that determined from the work diagram minus the work shown by the lower loop diagram ; and, in a 2-cycle engine, the total indicated horse-power as determined from the dia- gram of the power cylinder minus the pump work is considered as the indi- cated horse-power. This view is undoubtedly correct when the mechanical efficiency of the engine itself as a machine is to be determined. The power required at "no load" is the power indicated when no useful work is rendered by the engine. 18. Mechanical efficiency is the ratio of the useful power to the indicated power of the engine. 19. All consumption figures should be reduced to the hour basis, and if they are to be compared with the output of the engine they must be based on one horse-power hour. If not otherwise agreed upon, these. data refer to the useful or available output at full load. 516 INTERNAL COMBUSTION ENGINES EXECUTION OF TESTS 20. If the quantity of gas made in a producer or the weight of fuel consumed in an engine is to be measured, then all pipes or ducts which are not used in the test must be cut off from the pip- ing which leads to the producer and engine that are to be tested. This is best done by means of blind flanges. The active ducts, pipes, gas holders, etc., must be examined with regard to leakage and made tight if necessary. Unavoidable losses due to leakage must be determined. This holds especially for masonry gas mains. FUEL CONSUMPTION OF A GAS-PRODUCER PLANT 21. The kind, number, and duration of tests must be agreed upon according to the general rules laid down in paragraphs 1 to 10. 22. The constructive features and the operative conditions of gas-producer plants must be described and illustrated in the report by drawings, so far as this is necessary, to arrive at a clear understanding of the manner of working and of the results ob- tained. 23. Before making the test the plant should be examined as to whether or not it is in good working order. 24. The quantity of fuel consumed in the gas producer is determined by taking the weight of the fuel which is charged into the producer during the trials in order that the producer may contain at the end of the test exactly the same amount of heat - either liberated, or chemically bound in the fuel that it con- tained when starting the test. To meet this requirement it is not sufficient that the depth of the fuel bed be the same at the end as it was at the beginning; it must also be taken into con- sideration what influence the ash and the slag left in the pro- ducer, the location of the incandescent zone, the formation of fissures and cavities, the closeness or density of the producer charge, and the chemical composition of the burning fuel par- ticles exercise on the heat contents of the producer. In order to comply with this requirement the following rules should be followed: 25. When starting the test the plant should be in the condition of stability or normal working condition, if possible. METHODS OF TESTING GAS ENGINES 517 This means that after a period of shut-down for cleaning or repairs it should be in active operation for one or more days, running on fuel of the same characteristics and size, with the same depth of fuel bed, the same skill of attendance as regards the charging or feeding of fresh fuel and the removing of slag, and under the same load conditions that will obtain during the test. 26. During the trial the producer should be charged and poked as nearly in accordance with the requirements for attendance as possible. The level of fuel charged must be the same at the be- ginning and at the end of the tests and should be kept constant during the trial. About half an hour before starting and before stopping a test, the slag and ashes should be removed. If it is impossible to rake out the ashes during the operation of the producer, the plant must be shut down immediately after stopping the test, the ashes must be taken out at once and the producer refilled up to the same level that existed when starting the test. The weight of fuel used for this purpose 'must be added to the consumption. 27. The fuel consumed during the trial must be weighed, also the fuel which has not been burnt and remains useful; that is, that portion which drops down from above the grate while raking out the ashes, and that which is culled out from the ashes as unburnt. The weight of the former may be deducted from the consumption, but not the amount which is taken out from the ashes, nor the coal dust which accumulates in the scrubbers and in the flues between the producer and the engine. 28. To be able to determine the quantity of ash and slag produced during the trial, the ash box must be emptied before the test. If this is not possible, as when an inclined grate is used, the refuse in the ash box must be equalized before and after the run. 29. The stand-by losses during intervals of shutting down at day and night must be determined. 30. In order to get a representative sample of the solid fuel, the following course may be pursued: Of every carload, basket, or other measure of fuel, put a shovelful into a covered receptacle. Immediately after the test is over, the contents of the receptacle should be broken, mixed, spread and quartered by drawing the 518 INTERNAL COMBUSTION ENGINES two diagonals of a square. The two opposite quarters are re- jected, the two others broken up finer, mixed and quartered, and the two opposite quarters rejected. This is continued until a sample of some 10 to 20 pounds remains, which is preserved, in well-closed receptacles, for analysis. In addition to this a number of other samples must be put away in air-tight receptacles for use in determining the contents of moisture in the fuel. 31. The composition of the fuel shall be determined by chemical analysis. Its contents in carbon (C), hydrogen (H), oxygen (O), sulphur (S), ash (A), and water (W) must be given in percentage of weight referred to the original fuel. The con- tents, in the fuel, of nitrogen (N) can be disregarded. The be- havior of the fuel when being heated should be determined by a coking test. 32. The calorific value of the fuel must be determined by calorimetric analysis. An approximate determination of the heating value can be made on the basis of the chemical analysis by employing De Long's formula: Heating value = 145 C + 522.3 H - + 40 S - 9.66 W in which C, 0, H, S, and W are expressed in weight per cent. TESTING AN INTERNAL-COMBUSTION ENGINE 33. Kind, number, and duration of trials to be agreed upon according to the general regulations Nos. 1 to 8. 34. The constructive features and operative conditions of the engine must be so illustrated in the report as to enable one to form a correct idea of the manner of working and of the re- sults of operation. Especially important are the type and capacity of engine, diameter of cylinder and piston-rod, piston stroke, contents of clearance space, and other essential dimen- sions; the normal rate of revolution and the admissible fluctua- tions; kind and heat value of fuel for which the engine is intended. The diameter of the cylinder and the stroke should be actually measured if this is possible. The contents of the compression space are preferabty deter- mined by filling with water. If it is impossible to state the cubical contents of the compression space, then the compression METHODS OF TESTING GAS ENGINES 519 pressure at full load should at least be given. This is done by taking an indicator card while the ignition is interrupted. 35. Before making the test the engine must be examined internally and externally as to whether or not it is in good work- ing order. 36. The number of revolutions of the engine should be deter- mined by a continuous speed counter, the records of which must be noted at certain intervals, and must be checked or corrected from time to time by direct readings. If the speed conditions of the engine are to be investigated it is essential to determine the following items: (a) The number of revolutions under constant conditions at maximum load and at no load. (6) The fluctuations in speed at constant load: (c) The temporary change in the number of turns when the load is suddenly decreased or increased from a given constant load by a prescribed amount. These determinations can be executed with apparatus of the character of the Horn tacho- graph. The fluctuations of speed during the performance of one engine cycle above and below the mean value, expressed in parts of the latter, should be determined by calculation unless other- wise provided. The coefficient of fly-wheel regulation is AT" \ * y mm. \ A'mm./ g _ ^max. - JV mint = 2 /ALax. ~ *V max. H~. -/Vmin. \*V max. ~r 2 where N = number of revolutions. 37. The useful output can be determined either by brake test or by electrical measurement. The dimensions and weight of the brake should be determined before the trial. The electrical measurements can be made on a generator directly coupled to the gas engine. The useful work is com- puted from the output of the dynamo. The efficiency of the gen- erator should be determined by one of the methods as laid down in the "Rules for Judging and Testing Electrical Machinery and Transformers/' published by 'the association of German electrical engineers. If the efficiency is found approximately by measuring 520 INTERNAL COMBUSTION ENGINES the determinable losses, then an adequate amount (say 2 per cent of the full load output) must be allowed for losses not accounted for. The apparatus with which the electrical measurements are executed must be calibrated before and if possible also after the test. Whether anything besides the 2 per cent above allowed should be credited to the gas engine for increased bearing friction and windage of the generator must be settled in each individual case. Whether, in case the useful output can neither be determined by brake test or by electrical measurements, the code provision for testing steam engines can be admitted as correct for gas engines, namely, to designate the useful output as the difference between the indicated work at any load and the indicated work at no load, cannot be settled at the present state of development, since results of accurate investigations are not yet available. 38. Indicators must be connected immediately to the com- bustion chamber without employing long piping with sharp bends, and one indicator must be provided for every combustion chamber. For this purpose each compression chamber must have an opening for three-quarter or one inch Whitworth thread. The same holds true for pump cylinders. The indicators and their springs must be calibrated before and after the test according to the accepted standards. 39. During the test, cards should be taken quite frequently from every combustion chamber and from the pump cylinders. The cards should be designated by numbers, and the time when each card was taken, the scale of springs used and the number of single cards obtained must be recorded on the cards. At least five diagrams should be taken on one card successively. From time to time diagrams indicated with a weak spring should be taken from the combustion chambers. The indicated work at no load should be determined imme- diately after stopping the main test and while the engine is still warmed up ready for operation. Care must be taken that the no-load cards are not taken during an acceleration or during a retardation period of the fly-wheel. METHODS OF TESTING GAS ENGINES 521 ANALYSIS OF THE GAS GENERATED IN A PRODUCER-GAS PLANT OR CONSUMED IN AN INTERNAL COMBUSTION ENGINE, OR OF THE LIQUID FUEL USED 40. The samples for the chemical analysis of the gas must be taken during the trial at regular intervals and as frequently as possible. They must be either analyzed on the spot or preserved in glass tubes closed by melting the ends. The analysis is to deter- mine, in per cent of volume, the contents of the gas in carbon monoxide (CO), carbon dioxide . (CO 2 ) , hydrogen (H 2 ), marsh gas (CH 2 ), heavy hydrocarbons and oxygen (O 2 ). In addition it is recommended to determine the contents of sulfur. The gas samples should be taken from the gas main between the cleaning apparatus and the engine. 41. The heat value of the gas should be determined as often as possible by calorimetric analysis, and the burner of the calorimeter should be fed from the gas main without inter- ruption. In suction producer plants this can be done by means of a gas pump drawing from the main. If conditions should make it necessary that a sample be taken 'from the pipe while the calorimeter is shut off, such sample to be later transferred to and burned in the calorimeter, then the quantity of gas so taken should not be less than 300 liters (10.59 cubic feet), in order that the calorimeter may at first be brought into the condition of stability as regards the water of combustion, and in order that at least 100 liters (3.53 cubic feet) remain available for two successive analyses. The suction pump, the gas holder and the piping must be made tight with special care when making a calorimeter analysis of suction gas. 42. The gas meter of the calorimeter in which the heat value of the gas is determined must be calibrated. For determining the temperatures of the calorimeter water, only thermometers with calibration certificates or others compared with such should be used. The scales must be divided at least into tenths of a degree. On the basis of the chemical analysis the heating value per standard cubic foot of gases which do not contain heavy hydro- 522 INTERNAL COMBUSTION ENGINES carbons can be computed from the following formula, if a calorimetric analysis cannot be made Heating Value = 3.42 CO + 2.97 H 2 + 9.52 CH 4 where CO, H 2 and CH 4 are expressed in volume per cent. 43. The quantity of gas produced or consumed should be measured by means of a gas holder or a gas meter. The cross- sectional area of the holder should be determined by measure- ment of its circumference at several places. Consumption tests with the gas holder shall not be made while the latter is exposed to the" sun. 44. The gas meter must be calibrated and set level ; it must be so filled that the water level corresponds to the normal filling existing during calibration. Between the gas meter and the engine a pressure regulator must be installed or a large suction space provided so that the water level shows only small pulsations during the pressure fluctuations. 45. At intervals corresponding to the duration of test the following readings should be taken: position of the bell of the gas holder at three places or the records shown by the gas meter; the pressure in the bell or in the gas meter; the temperature of the gas when entering and when leaving the gas holder or the gas meter and before reaching the engine; the barometric pressure. 46. If the temperature of the gas is different when measuring the consumption than when measuring the heat value, the com- putation must also take into account the increase of volume which is due to the moisture contents of the gas at higher tem- peratures. 47. The consumption of liquid fuel must be determined either by weight or by measuring its volume. For determining heat value, composition, and specific weight of the fuel one representa- tive average sample is sufficient. 48. When measuring the fuel consumption of internal com- bustion engines, the consumption of lubricating oil for the cylinder should be determined at the same time. 49. If the consumption at low loads of a double-acting tan- dem or twin engine is to be determined, it is not allowable to shut off the gas from one or more ends of the cylinders, provided that no other arrangements have been previously agreed upon and METHODS OF TESTING GAS ENGINES 523 are mentioned in the report, or that the governor acts automati- cally in the way described. EXPLANATIONS TO VARIOUS ARTICLES OF THE CODE The main code is followed by a number of explanations from which the following extracts are taken. The figures refer to the paragraphs of the above code. 1 &2 In most cases only one or two of the objects of test mentioned are taken into account in any given trial. If in any exceptional case the object of the test should not be any of those mentioned, it should be a simple matter to adapt the rules given. Under 2 (c) the term horse-power hour is used. It is essential that in any given trial this term be more closely denned, as horse- power may mean indicated brake, or even horse-power developed by pumps. 4 It is extremely desirable that the contract state the time allowed the manufacturer for adjustment and trial runs, because his own interests may make him call sometimes for a long, sometimes for a short period. In the case of a small engine, more or less a commercial stock machine, he may wish to have the period as short as possible, and this the buyer may agree to without danger of loss to himself. If, however, the machine is of a special type, or one provided with special attachments, it is but a matter of justice to allow the manufacturer a reasonable time in which to break in the engine and to give him an oppor- tunity to correct any imperfections that may appear. It is to the interest of the buyer to grant such a period in order to become familiar with the machine before taking over the entire responsibility of operating it. It is also true that many faults appear only after some weeks of operation. On the other hand, too long a period of adjustment is in many cases not acceptable to the buyer, because any extended work of improvement usually seriously hampers operation; and because in many cases he desires an operative machine, which no longer requires the care of the manufacturer, as soon as possible. It frequently happens that no acceptance test is agreed upon. 524 INTERNAL COMBUSTION ENGINES In such cases it sometimes happens that the buyer comes back upon the manufacturer for faults which did not develop until the machine had been in operation some time. If the manufacturer then agrees to an investigation or a test, a sufficient period should be given him to make any investigation he sees fit or to correct any imperfections that may have appeared before the decisive trial or investigation is made. This sometimes leads to a simple settlement of the matter in that the manufac- turer discovers that ignorance or carelessness on the part of the operator have caused the imperfections complained of. The granting of such a period also guards the buyer against any later claim of the manufacturer that during the trial the machine was not in the condition in which he delivered it. Preliminary tests are always desirable, but not absolutely necessary. The cost of any kind of investigation is usually quite high and of course the cost increases directly with the time. The expert called will therefore make such tests when they seem to him essential. But the manufacturer should have the right to call for the time necessary for such trials if he is to present the machine in its best condition. 6 It cannot be denied that eight hours is a rather short time, because it is extremely difficult to determine whether the pro- ducer is in the same condition at the end as at the beginning of the test, and because this uncertainty may lead to large errors. On the other hand it is unquestionable that in many cases a longer time would call forth so many difficulties in operation that eight hours would seem the necessary limit. The rule is mainly framed to prevent trials of so short dura- tion that serious errors can hardly be avoided, but it leaves it to the judgment of the experimenter whether to make the tests longer than eight hours where it seems desirable and is possible to do so. 7 Intermediate readings are recommended without qualification, since they form the best criterion of the constancy of conditions. METHODS of TESTING GAS ENGINES 525 With liquid or gas fuels of constant composition, individual readings every five minutes apart sometimes show no variation for hours at a time. In such a case it is useless to extend the time of the trial. 8 In determining the mechanical efficiency of an engine it should not be forgotten that, although the average load may be con- stant, there may be speed . variations due to the inevitable in- equality of the indicator diagrams, so that during some cycles work is done in accelerating the fly-wheel, while during others the fly- wheel by retardation gives up some of its kinetic energy. To minimize any error ^ that this may introduce into the determination of the mechanical efficiency, at least ten diagrams should be taken. If the conditions are otherwise constant, however, it is not necessary to spread these diagrams over any considerable period of time. It is self-evident that during the time of taking the diagrams the supply of lubricating oil must not be increased. Changes in the mechanical efficiency of the engine, as for instance those due to fouling, cannot be detected with certainty even by a long test period; they become noticeable usually only after a period of operation extending over two weeks. The de- termination of the mechanical efficiency of an engine, after con- stant conditions of operation are attained, therefore only applies to the engine in its then existing state or condition. The number of diagrams to be taken on one card cannot be definitely stated. On account of variation in the diagrams, which is less at high than at low loads, care should be had not to take too few. On the other hand it is useless to take more than can be clearly distinguished. The running together of a larger number of diagrams only makes their evaluation more uncertain. 10 In consideration of unavoidable errors of observation, possible errors of the instruments used, etc., it is meet and usual to allow a certain margin between the figures found on trial and those guaranteed. In the steam engine code 5 per cent is allowed for this, and it seems reasonable to assume the same figure in this fctA H OF THE UNIVERSITY] OF 526 INTERNAL COMBUSTION ENGINES case. Only in one point, in the guaranteed normal capacity , does the gas engine call for an exception. A given steam engine gives its most economical results at a certain cut-off, but a higher capacity can always be obtained at the expense of a little economy, that is, a buyer is certain that even a machine slightly too small will give him sufficient ca- pacity. A gas engine, on the contrary, works with the greatest economy at its maximum load. It is to the interest of the buyer, therefore, to get an engine exactly suited to his needs and not to choose it too large. It is possible for the same reason that any engine, if lacking slightly in guaranteed capacity, may become absolutely useless to the buyer. For these reasons it was thought advisable not to grant the manufacturer any leeway whatever as regards guaranteed capacity. It is clear, therefore, that the manufacturer must take upon himself any possible inaccuracies in the measurements, unless he can show them up and demand a new trial. For that reason it is well for him to make his guarantee a little on the safe side of what he knows his engine is capable of developing. On the other hand, there is no harm done to the interest of the buyer if the manu- facturer underrates the normal capacity of his machine, because the former will always call for an engine of a certain normal capacity to suit his needs. If he fails to do this, but places his dependence in the guaranteed maximum -capacity, he is open to the charge of carelessness. Since during acceptance tests it is often not possible to keep the load quite constant, it became necessary, following the steam engine code, to allow a certain amount of variation, within which no just cause could be found for objection to the trial. There are cases where the variations occurring are much greater, as when a gas engine is used for driving a roll train. But no one set of "Rules" can possibly take into account all such extreme cases, and in such instances the contract should contain the necessary agreements to make any test clear and free from sub- sequent objections. The wish has been expressed from several quarters, that the " Rules" should contain a definition of the term " Normal Ca- pacity." On account of the peculiarity of the gas engine above discussed this is not quite feasible. But the term "Maximum METHODS OF TESTING GAS ENGINES 527 Continuous Capacity" perhaps defines most nearly what is in- tended in most cases. 14 It is sometimes the case that the heating value of the stand- ard cubic foot, that is, reduced to 32 degrees Fahrenheit and 760 mm. barometer, is so greatly different from the actual value of the gas as used, that any contract which contains only the heating value of the gas stated on that basis does not convey much meaning to the non-technical buyer. If for instance a given gas has a heating value of 135 B. T. U. per standard cubic foot, its effective heating value at a high altitude and in a warm climate, say at 68 degrees and 620 mm. barometer, will only be about 100 B. T. U. per cubic foot. To obviate any misunder- standing, it should be clearly stated that, when the "effective" heating value of the gas is not definitely specified, the heating value at 32 degrees Fahrenheit and 760 mm. barometer is meant. 19 By "full" load is meant the normal capacity, as per para- graph 10. 23 For acceptance tests, and all other tests which are intended to decide any disagreements between manufacturer and buyer, such examination should be carried out in the presence and with the aid of the former, as already mentioned under paragraph 4. 24-26 In all gas producer tests it is hardly possible with certainty to have all conditions exactly the same at the end as at the begin- ning. But since any difference in the beginning and end con- ditions may lead to considerable error, which can only be equalized by excessive length of test, the "Rules" are intended to operate to the end that such errors are not in any way magni- fied by the method of test. Hence the detailed statement in the regulations. 27 Since in actual operation, the fuel in the ash or the coal dust in the gas mains are hardly ever utilized, no correction should 528 INTERNAL COMBUSTION ENGINES be made for these on any trial. In order, however, to prevent the results from being influenced by insufficient cleaning of the producer, any fuel which falls out from above the grate dur- ing the cleaning period may be subtracted from the amount charged. 35 See explanation under 23. 37 A brake test of a large engine is in some instances not possible, and in any case a matter of considerable cost. In many cases, however, the larger gas engines are either direct connected to a generator or to some other power consumer, as a blowing cylinder. In the former case electrical measurements, from which the effective horse-power may be determined, are easily made. In the latter case the capacity guarantee will in most instances be based upon the performance of the power consumer, as for example the air compressed by the blowing cylinder. Outside of engines of this type, however; thero still remain many cases in which it would be of the utmost value to have some means of determining the effective capacity, and it should not be forgotten that, even in the case of medium-sized machines, a braking of the engine at the place of erection is often, on account of local restrictions, very difficult. The problem has been solved for steam engines by assuming, that the difference between the indicated horse-power at any load and the indicated horse-power at no load is the effec- tive or useful horse-power. It is quite possible that in many cases this is not quite correct, but the method is very generally accepted and followed. On account of the great overload capacity of the steam engine, a small error in this respect does not mean a great deal. But the case of the gas engine is quite different. The data on hand does not warrant the application of the same method to the gas engine, and the consequences of an erroneous conclusion are much more serious on account of the lack of overload capacity. For these reasons one is compelled in some cases to omit the determination of the effective capacity altogether and to be con- tent with the determination of the indicated power only. It is recommended in such cases that the mechanical efficiency be not METHODS OP TESTING GAS ENGINES 529 assumed too high and that any guarantees regarding fuel, etc., also be based upon the indicated horse-power. It is sometimes possible to brake an engine on the test floor of the factory. The mechanical efficiency may thus be previously determined when it is known that no brake test can be made in the final place of erection. 39 The number of diagrams to be taken during any given test cannot be definitely specified. Much depends upon the length of test, and the decision may be left to the judgment of the experimenter. It is, however, always recommended that a bundle of diagrams, instead of only one, be taken on every card. Thus a series of diagrams are obtained, while, if only a single diagram is taken, it is possible to hit upon the same diagram in the series a number of times. (See under extract 8.) The work of fluid friction, that is, the lower-loop diagram, cannot be determined with certainty from the full indicator cards. It is best for that reason to ignore the loop when deter- mining the positive work and to find the negative work from special weak spring diagrams. 48 The measurements of the quantity of lubricating oil used is of importance in smaller engines, because the fuel consumption can be favorably influenced by a copious supply of the lubricant. 49 If under low loads, only one end of the cylinder is allowed to work, the fuel consumption would be much lower. But since this is not generally done in operation, the results would be erroneous. If, however, the governor during operation shuts off the individual cylinders or cylinder ends, as the load drops, this is of course also permissible during a test. CHAPTER XVII THE PERFORMANCE OF GAS ENGINES AND GAS PRODUCERS 1. As indicated in the rules for testing, the very great ma- jority of tests of engines are made to determine capacity and fuel consumption. In some special cases, as with engines driving generators, tests are also sometimes made of the regulation. These three tests together take care of what may be termed the commercial side of testing. All other tests are special in that they are not often executed in an acceptance test, but form in most cases the object of scientific or laboratory investigation. Such investigations are in many instances very valuable, and have served to throw a flood of light on the somewhat complex cylinder actions of the internal combustion engine. It was thus found that the temperature of the cylinder walls, i.e., the cooling water conditions, piston speed, ignition, proportion of mixture, compression, etc., all had a more or less marked effect upon engine performance. In the following the results obtained and the conclusions drawn by various experimenters concerning the effect of these various factors are briefly set forth. The main bulk of this work has been done by Witz in France, Slaby and E. Meyer in Germany, and Burstall and others in Eng- land. In spite of the fact that none of these investigations are open to serious objection on the score of inaccuracy, the con- clusions arrived at are not always in accord. This is undoubtedly due to the complexity of the cylinder actions, and the inter- dependence of the various factors entering the problem. 2. Cooling Water Conditions and Piston Speed. It is rea- sonable to suppose that the higher the wall temperature of the cylinder, i.e., the smaller the temperature difference between mixture and wall and the greater the piston speed, cutting down the time of exposure, the smaller the loss to the jacket. The heat thus saved, however, may go in two directions: either 530 PERFORMANCE OF GAS ENGINES 531 a greater thermal efficiency is shown, resulting in greater power developed for the same expenditure of heat, or the heat saved from going into the jacket is lost in the exhaust. Witz, upon the basis of his experiments, comes to the former conclusion, and says " The action of the jacket is the great regu- lator of combustion phenomena." He summarizes his results as follows : 1. The efficiency increases with the piston speed and with the temperature of the surrounding walls. 2. The combustion of explosive mixtures is the more rapid the greater the speed of expansion and the hotter the cylinder walls. The work of Slaby and of Meyer, however, seems to contro- vert these conclusions. Some of Slaby's tests show that while the gas consumption per horse-power hour decreases somewhat with an increase in the piston speed, there is an increase in the consumption with a rise of jacket water temperature. The following table of figures, quoted by Schottler from some of Meyer's tests, illustrates the point that the heat saved from the jacket by higher piston speed may go to the exhaust, leaving the ther- mal return practically unaffected. Ratio of Com- R. P. M. MEAN EFFECT- IVE PRESSURE Ratio HEATING VALUE OF CHARGE WORK DONE BY 1 B.T.U. EXHAUST TEMP. HEAT DISTRIBUTION IN% pression No. 1 sq. inch B.T.U. Ft. Lbs. F Work Jacket Water Exhaust 2.67 187 54.3 7.11 18.5 140 1022 18.0 51.2 30.8 2.67 247 51.7 7.35 17.4 141 1137 18.1 45.6 36.3 4.32 187 69.3 7.43 17.0 190 867 24.4 53.8 21.8 4.32 247 65.2 7.40 16.8 184 992 23.7 49.5 26.8 In tests of this kind there are one or two simultaneous actions, not directly under control, which may serve to modify the final result and account in a measure for the discrepancy appearing in the results above discussed. An increase in the temperature of the walls or an increase in the piston speed both cause a de- crease in the charge volume, the former by heating the incoming charge and decreasing the density of the mixture, the latter by increased friction loss in pipes and ports. The direct result of 532 INTERNAL COMBUSTION ENGINES this is that the effect of the action of the cylinder wall upon the leaner charge is proportionately greater. Thus the beneficial effect of greater piston speed may be partly annulled by the relatively stronger action of the walls. Further it is true that a smaller charge weight means a lower compression pressure, and the comparatively greater admixture of burned gases at high speeds causes a less rapid combustion. Both of these actions tend to decrease the efficiency. There are thus several antago- nistic actions, and the final result is consequently in many cases quite problematical. The net result of an increase in the cylinder wall temperature or of the piston speed, or both, in an existing machine, is cer- tainly a decrease of maximum capacity for reasons already pointed out. Further, the effect of a variation in the temperature of the jacket water, while perhaps not quite so marked in engines using gas, is certainly quite noticeable in liquid fuel engines, especially those using kerosene or alcohol. It is quite possible in kerosene engines, by running the jacket too cold, to increase the oil consumption seriously by condensation of the oil vapor on the comparatively cold cylinder surfaces. The same holds true of alcohol. Thus hot walls are in such cases of undoubted benefit. The limits to temperature are, of course, decrease of engine capacity and danger of pre-ignition. 3. Compression. The theoretical effect of increasing the compression, and the commercial limits to this increase have already been discussed in Chapter III. Much of the increased efficiency of blast furnace and producer gas engines as compared with illuminating gas and liquid fuel engines is directly due to the greater compressions that the former fuels can stand. The above table of Meyer's results gives some idea of the gain that can be made with illuminating gas by increasing the compression. With a compression ratio of 2.67, the average thermal efficiency was 18.05 per cent, with a ratio of 4.32 the average was 24.05 per . 24.05 - 18.05 cent, a gain of - - - = 33 per cent. lo.Uo Another test made by E. Meyer * on a 10 horse-power engine, which was operated with illuminating and with producer gas, gave the following results: * E. Meyer, Z. d. V. d. I., July 5, 1902. PERFORMANCE OF GAS ENGINES 533 INDICATED THERMAL EFFICIENCY COMPRESSION RATIO With Illuminating Gas With Producer Gas 4.98 27.1 24.4 4.59 26.5 23.2 3.84 24.8 21.5 Here again the beneficial influence of the higher compression is marked, although the gain is not so great as in the former case, owing to the smaller change in the compression ratio. Other instances pointing to the same result can be adduced without difficulty. See the next table below, also by E. Meyer, who has done an immense amount of work in the investigation of gas engines. Banki took advantage of the principle in his gasoline engine, in which, by . using water injection, he could employ compression ratios similar to those used in producer gas work and realized thermal efficiencies fully equal to those ob- tained with the leaner power gases. 4. The Mixture. The inherent advantage of the use of lean mixtures has already been shown in Chapter III. Burstall * on the basis of his tests for the British Institution of Mechanical Engineers, concludes that the thermal efficiency depends upon the correct choice of the mixture, and that the ratio of air to gas should increase with the compression. His results, however, do not definitely warrant the latter part of this deduction, although it finds some support in the above-mentioned tests by Meyer, f as shown by the following table: Test No. Ratio of Compression Ratio Air to Gas Gas Consumption Cu. Ft. per I. H. P.-hour U;,. 2.67 6.41 8.08 27.1 25.4 !}- 3.23 6.38 8.07 22.6 21.6 8}-- 3.87 5.93 8.29 20.6 18.5 II-- 4.32 6.00 8.35 19.4 17.9 * Proceedings, 1898, p. 209. f E. Meyer, Z. d. V. d. I., 1899, p. 361. 534 INTERNAL COMBUSTION ENGINES This table shows that, whatever the ratio of compression, the gas consumption is less with the leaner mixtures. The cause for this, besides the theoretical reason, may possibly be found in the fact that with leaner mixtures the maximum temperatures in the cycle are lower than with rich mixtures, always assuming, of course, that the mixture contains no excess gas. How the efficiency of an engine may be affected by careless setting of the fuel valve is well shown in some results quoted by Lucke.* The test cited is on a 10 I. H. P. Otto engine governed by hit and miss. The fuel used was carbureted water gas. Gas Valve Number Efficiency 9 16.0 8 16.5 7 18.0 6 19.0 5 10.0 It is quite evident from the figures that the last setting wasted a lot of unburned gas, while the leaner mixtures were inefficient, probably due to sluggish combustion. The same thing was noticed in a series of tests on German alcohol engines, in which it was found that the setting of the fuel needle valve had a pro- nounced effect upon the economy. 5. Variation of Point of Ignition. The effect on the appear- ance of the diagram of varying the point of ignition has already been discussed. To get some idea of the influence of varying time of ignition on engine capacity and efficiency, the following- figures are given. The first set f was obtained in determining the range of adjustment of the igniter gear on an 8" x 12" hori- zontal hit-and miss-engine, running 265 r.p.m. on natural gas. * Lucke, Gas Engine Design. t Obtained through the courtesy of Mr. A. B. Gould, of the Wellman, Seaver, Morgan Co., Cleveland. PERFORMANCE OF GAS ENGINES 535 Card No. Crank Angle below horizontal at time of ignition Max. brake load possible at 265 R. P. M. 8 35 47 Ibs. net 9 33 47 ' 10 32 47i ' ' 11 24 47 t 12 23 44J ' ' 13 15 44 ' The accompanying indicator cards are shown, much reduced, in Fig. 17-1. The scale of spring was 160 pounds. FIG. 17-1. The second set of figures is due to Mr. J. R. Bibbins, and was published by him in the Michigan Technic for February, 1907. They are here reproduced by permission of the author, obtained through Mr. R. D. Day of the Westinghouse Machine Co. The tests were made on a 9 x 11", 2-cylinder Westinghouse gas engine. The load was kept constant at about 70 B. H. P., the speed was held constant at about 300. r.p.m. The gas used was natural 536 INTERNAL COMBUSTION ENGINES gas with a constant lower heating value of 934 B. T. U. The point of ignition was changed by steps from dead center to 55 de- gree crank angle ahead of the center, that is, the spark was ad- vanced to that extent. The following table shows the results: Point of Ignition Degrees Load B.H.P. R.P.M. Gas per B. H. P. per hour B.T. U. per B. H. P.- hour Thermal Efficiency on Brake Relative Efficiency PRESSURES LBS. per sq. inch early cu. ft. % Max. Release 70.0 292 14.38 13410 19.0 .815 151 36 8 70.8 295 13.34 12470 20.4 .875 168 36 20 71.0 296 12.36 11530 22.1 .947 177.5 33.6 25 71.0 296 12.3 11490 22.2 .951 220.5 31.2 30 71.3 297 11.71 10940 23.3 1.000 252 31.2 35 71.3 297 12.03 11230 22.7 .972 252 28.8 45 71.2 296.5 12.40 11590 21.9 .942 379 28.8 55 70.0 292 15.74 14700 17.3 .742 437 24.0 The results of the test are shown graphically in Fig. 17-2. The best lead angle for the sparking gear appears to be between 30 and 35 degrees, in which this test agrees closely with the re- Cf/ect of I/triable Ig Point on Gas Ehsine Point of ft 'nit ion- Degrees ar W PERFORMANCE OF GAS ENGINES 537 suits obtained by Mr. Gould on a similar engine. Fig. 17-3 shows the typical indicator card accompanying each igniter position. 6. Engine Economy depending upon Load. As is the case in the steam engine, the efficiency of a gas engine decreases as the load decreases below the normal. The amount of this de- crease varies in different engines, depending mainly upon the IGNITION I9NITIQN IGNITION 20 IGNITION 25 * IGNITION J0* IGNITION IGNITION 43 IGNITION J5* FIG. 17-3. kind of fuel used and the system of governing employed. The following set of curves, Fig. 17-4, makes this clear. Regarding the range of load above normal, however, it is found that, while a steam .engine generally shows a decrease in efficiency for over- loads, the gas engine usually shows a greater efficiency at the maximum load than at the normal; in other words, just as long as a gas engine keeps up the normal speed under an increase in 538 INTERNAL COMBUSTION ENGINES PERFORMANCE OF GAS ENGINES 539 load the thermal efficiency will rise with the load. A few typical efficiency curves are shown in the figure. The data for these has been collected from various sources, as shown in the accompany- ing table: Curve No. Type of Engine B.H.P. R.P.M. Fuel Governing Reference 1 Deutz, Single 50 200 111. Gas Throttling E. Meyer, Cylinder Gas. Z. d. V. d. I., 1898. 2 Westinghouse 3-cycl. Vert. 100 270 Natural Gas. Throttling Mix. Robertson, 1899. 3 Deutz. 450 Producer Guarantee Gas. Fig. for en- tire plant. Josse, 1904. 4 Giildner, Single- 35 220 Producer Throttling Giildner, 1906, cylinder. Gas. for plant in- cluding generator. 5 Niirnberg. 1200 106 Blast- Throttling Riedler, Furnace Gas. Gross-Gas- Gas. Maschinen, 1905. 6 Swiderski, Single- 15 235 Alcohol. Hit& Miss. E. Meyer, cylinder. Z.d.V.d.L, 1903. 7 Deutz, Single- 12 285 Alcohol. Throttling. cylinder. 8 Diesel. 70 158 Russian Cut-off. tt Kerosene. 9 Diesel. 8 275 a u 10 Hornsby-Akroyd 25 202 Kerosene. Regulating Robinson, Oil. 1898. 11 Bdnki. 25 210 Gasoline Hit& Miss. Jonas and with water Taborsky injection. Z. d. Oest. Arch.& Ing V., 1900. 7. The Heat Balance. Accounting for the heat furnished to a heat engine is called constructing the Heat Balance. In an internal combustion engine the heat supplied to the engine is that contained in the fuel furnished to the engine in a given time. For convenience all heat calculations are referred to some standard temperature, usually room temperature. It is usual to account for the heat in four separate items: 1. Heat represented in indicated work. 2. Heat carried off in the jacket water. 540 INTERNAL COMBUSTION ENGINES 3. Heat lost in exhaust. 4. Heat loss due to radiation, conduction, etc. Of these the first and second items admit of accurate determina- tion, the fourth is nearly always found by difference between the sum of the other three items and the heat supplied. Item three is much more difficult of exact determination. Its calculation involves the determination of the weight of exhaust gases and of the specific heat of these gases at exhaust temperature. The only accurate way to find the weight of the exhaust gases is by metering or otherwise determining the air supply to the engine. The weight of the exhaust gases is in all cases the weight of air plus the weight of the fuel. There are two other ways sometimes employed, but either one can only give approximate results. One of these determines the charge weight from the piston displacement. This involves (a) The volumetric efficiency of the suction stroke, and (6) Some assumption as to the temperature of the charge at the end of the suction stroke. It is possible to determine the volumetric efficiency with fair accuracy by means of a weak spring card, but the second point offers 'much more difficulty. The charge at the end of the suction stroke, taking the case of a gas engine, consists of a certain amount of air, of fresh gas, and of burned gases from the clearance. The entering temperature of air and of gas can be accurately deter- mined, but the temperature of the mixture, as it enters the cylin- der, changes, due to contact with the hot walls and to mixing with the clearance gases. Nothing is definitely known of the weight of the clearance gases, for although their pressure and volume are known, nothing is known of the temperature. Hence neither the wall effect nor that due to the clearance gases can be definitely gaged and all temperature computations therefore become approximate. The only positive way of determining the temperature at the end of the suction stroke is by actual measurement. This has been successfully done, but the apparatus is not such as could well be employed in ordinary testing. It must be evident, therefore, that piston displacement meas- urement of the weight of the exhaust gases can be approximate only. To cite a case in point, the test of Brooks and Stewart, mentioned in Chapter V, showed an actual ratio of air to gas on PERFORMANCE OF GAS ENGINES 541 test of 6.63, while piston displacement computation showed 8.32. The second approximate computation for the weight of the exhaust gases is based upon the exhaust gas analysis. The method of doing this has been explained in detail in Chapter VI. The trouble with this scheme lies in the difficulty of obtaining representative gas samples. But granting even that these are obtained, it is often found that computations based upon the analyses show an excess coefficient smaller than that really used. That is, the weight of exhaust gases so determined is less than the actual amount. Thus, Schottler, in computation on some of Slaby's tests, shows in one case that, based upon analysis, the excess coefficient was 5.52, while in reality it was 6.2. Schottler attributes this discrepancy to a change in the analysis due to a burning up of the lubricating oil, which is apt to increase the CO 2 content of the exhaust gas at the expense of the percentage of O. It has been shown that a variation in the supply of lubri- cating oil may change the fuel consumption under circumstances quite materially, and Schottler's surmise is therefore probably correct. At the same time, however, any such effect must be more marked in the smaller machines, and the writer believes that, given representative samples of exhaust gas from any- thing but the small machines, in conjunction with accurate fuel analysis, a very close approximation to the actual weight of the exhaust gases can be obtained. At any rate the method should be more accurate than that based on piston displace- ment. Finally, it should not be forgotten that even an accurate determination of the air supply still leaves open the question of the specific heat of the exhaust gases. The heat balance of the four items above outlined is sometimes shortened to three by combining items 3 and 4 and determining their sum by difference. On the other hand, a balance of many more items can be drawn up. Thus each event of the cycle may be examined by itself and the heat and energy interchanges be determined. A very detailed balance of this kind is given in Schottler's "Die" Gasmaschine," p. 321. How far this question needs to be entered into depends altogether upon the importance of the test, but an effort should be made in every case to so arrange 542 INTERNAL COMBUSTION ENGINES the apparatus that at least a heat balance of the kind first dis- cussed can be drawn up from the results of the test. 8. Results of Tests of Engines and Gas Producers. By this term is here meant the results shown by engines or producers when operated under normal conditions and at or near normal load. The number of tests from which this data can be obtained is at the present writing quite large, and some notable collections of test data have been made. The mcst extensive is perhaps that contained in Appendix A of Brian Donkin's "Gas, Oil and Air Engines." This collection consists of eleven tables, arranged according to fuel used, containing in all some 280 tests. Another large collection is that found appended to Witz's " Moteurs a Gas et a Pet role. " Such collections are valuable as showing what has been done, and serve as a guide as to what may be expected from an engine under design or construction. The greatest care, however, should be exercised to see that only reliable data is incorporated. ENGINE TESTS. Table I (pp. 544 and 545) contains a series of engine test data taken from various sources. The tests are arranged according to the kind of fuel used, this being the most logical way. In some cases not all the data is given in the original report. Wherever possible computations have been made to make the items complete. In many cases, however, not enough information is given to permit of this, and the record necessarily remains incomplete. TESTS OF PRODUCERS AND PRODUCER PLANTS. Table II, p. 546, gives some of the results obtained on tests of producers. In some cases the tests refer to producers only, giving no data regarding the engine used; in others the data is fairly complete for the entire plant. PERFORMANCE OF GAS ENGINES 543 In conclusion, Fig. 17-5 shows a set of curves drawn by C. H. Day in 1905 during an investigation on the economy of producer gas. The curves show the relation between cubic feet of gas used arid brake horse-power for various kinds of gas. While it must be understood that they are approximate only, they represent a fair general average of what is accomplished to-day. Many plants show much better fuel consumption, but others show FIG. 17-5. correspondingly worse. The curves also show in a measure the sizes up to which the various engines are built. Thus an illu- minating gas engine of 150 horse-power is a large engine of its type. Producer plants in 1905 apparently were not built much larger than 500 horse-power, while natural gas and blast furnace gas engines were built exceeding 1000 horse-power, and for the latter gas 2000 and 3000 horse-power is not now out of ordinary. TABLE I. DIMENSIONS ANU OTHER DATA Name Rated Kind of Fuel of B.H.P R.p.m. Engine Dia.of Cyl. Inch. Stroke Inch. No. of Cyls. 2 or 4-cycle Single or Oouble Act. Illuminating Gas Koerting 4 s 161 Illuminating Gas Deutz 4.96 5.92 1 4 S 2 260 Illuminating Gas Giildner 1 4 S 15 Illuminating Gas Giildner 10 15.7 1 4 S 20 210.7 Illuminating Gas Banki 1 4 S 16 255 Illuminating Gas Illuminating Gas Illuminating Gas Tangye Crossley Westinghouse 10 7 19 15 1 3 4 4 4 S S S 193.6 200 235 Natural Gas Westinghouse 25 30 3 4 S 550 150 Natural Gas Snow 25 48 4 4 Natural Gas Otto 11.25 18 1 4 S 36 220.4 Natural Gas Walrath 13 14 3 4 s 75 253.3 Producer Gas R. D. Wood 25 30 2 (Tandem) 4 s 300 149 Producer Gas Westinghouse 3 4 s 235 Producer Gas Crossley 26 36 4 s 152.4 Producer Gas Koerting 21.6 37.7 1 2 D 101 Producer Gas Deutz _ 4 , 161.6 Coke Oven Gas Oechelhauser 65 37.5 1 2 2 pistons in 500 110.6 (Borsig) 1 cylinder Mond Gas Mond Gas Blast Furnace Gas Crossley Crossley Nurnberg 16.9 26 33.5 24 36 43.4 2 (Opposed) 2 (Opposed) 2 (Tandem) 4 4 4 S S D 400 162 148.5 105.6 Blast Furnace Gas Berlin-Anhalt 16.95 27.5 1 4 S 60 160.6 Blast Furnace Gas Cockerill 51.2 55.2 1 4 D 94.57 Gasoline Fairbanks 6.5 9 1 4 S 7 300 Gasoline Springfield 6.5 12 1 4 S 6 "230.4 Gasoline Lozier 5 6 1 2 S 5 513 Gasoline Westinghouse 5.75 8 2 4 s 10 289.7 Gasoline Banki . . 1 4 s 209.1 Gasoline Daimler 3.56 5.11 4 4 s 16-20 400 Gasoline Daimler 3.56 5.11 4 4 s 16-20 600 Gasoline Daimler 3.56 5.11 4 4 s 16-20 1000 Kerosene Diesel 10.23 16.16 1 4 s 186.6 Kerosene, Russian Diesel 15.75 23.65 1 4 s 70 158.8 Kerosene, Russian Diesel 6.65 10.60 1 4 s 8 270.3 Kerosene Priestman 10.9 14.1 1 4 s 172 Kerosene Grob & Co. 9.07 9.07 1 , 4 s 8 266 Kerosene Swiderski 10 10 1 4 s 249 Kerosene Koerting 6.9 10.82 4 s 222 Kerosene Blackstone 7 12 1 4 s 240 Kerosene Stevenson 9.5 18 1 4 s 200 Kerosene Hornsby 8.2 14 1 4 s 213.8 Kerosene Hornsby 14.5 17 1 4 s 25 202.6 Crude Oil Diesel 4 s 225 169.1 Alcohol, 86.1 vol. % Deutz 8.35 11.8 1 4 s 12 276.9 Alcohol, 86.1 vol.% Marienfelde 9.95 15.75 1 4 s 14 197.6 Alcohol, 86.1 vol.% Alcohol, 87 vol. % Koerting Banki 6.16 9.95 4 4 s s 6 20 307.3 225 544 ENGINE TESTS LOWER HEATING . H.P. ON TEST VALUE OF HEAT DISTRIBUTION. % tesS IU^L B.T.U. Mech. fit PER Eff. gsd References and Remarks -*cu % j8 I.H.P. B.H.P. Lb. Cu. ft. Is Jacket Exhaust Rest s 108.1 500 28.22 De La Vergne Machine Co., Cat. 2.30 1.72 572 21.5 50.4 25.0 4.1 75 16.1 Wimplinger, Zeitschrift d. V. d. I. Sept. 8, 1906. , 580 40.5 . Dubbel Z d V d I Nov. 3, 1906. 35.9 42.7 33.2 24.1 Test by Schroter, Z. d. V. d. I., June, 1904. 17.03 27.97 31.0 Schimanek, Z. d. V. d. I., Jan. 17, 1903. 28.6 25.4 609 85 24.5 Witz, 1902. 14 12 680 86 22 Clerk, 1894. 196.5 147.5 597.4 35 76 26.7 Ballinger & Hunt, Sibley College Thesis, 1904. 618 550 1000 27.1 89 24.1 J. R. Bibbins, A.I.E.E., Dec. 1903. 736.7 594.5 1175(?) 29.4 ' 80.7 23.7 Hastings & Parker, Sib. Coll. Thesis, 1901. 36.3 28.8 . . 1086 20.2 79.5 16.1 Hunting, Sib. Coll. Thesis, 1902. 86.7 78.7 1041 27.1 49.5 23.4 78.7 21.3 Geer & Vanelain, Sib. Coll. Thesis, 1902. 242 169 145 24.4 25.0 50.6 70 17.1 Goldsmith & Hartwig, Sib. Coll. Thesis, 1905. 173.6 124.8 123.8 33 71.8 23.7 Ballinger & Hunt, Sib. Coll. Thesis, 1904. 377.9 313 90.13 83 21.78 Humphrey, Inst. Mech. Engs., 1900. 481 341 129.5 71 24.1 E. Meyer, Jour, fur Gas Belenchtung, 1900. 81 70.4 113 87 29 E. Meyer, 1903. 765 (net) 628 376 33.4 82.1 27.5 E. Meyer, Z. d. V. d. I., (Blowing Feb. 1905. Test No. Vllb. Cyl). 141.6 135 29.1 | Humphrey's Test, Schott- 440 364 _ 138 28.6 83 23.7 ler, Z. Jan. 1902. 1427 1186 88 33.9 83.1 28.2 Riedler, Gross-gasmaschi- nen, p. 158. 79.5 54.5 102 . 26.2 E. Meyer. Z. 1899, p. 448. 786.16 575 99.5 20.4 Hubert, March, 1900. 11.28 7.23 18200 20.5 64 13.15 Elwood, Ford, Garrow, Sib. Coll. Thesis, 1907. 9.63 6.16 20078 15.8 64 10.2 Keeley & Spier, Sib. Coll. Thesis, 1900. 6.28 4.92 18520 19.8 78.9 15.7 Bayne & Speiden, Sib. Coll. Thesis, 1902. 9.19 5.58 28.5 60.7 17.3 Glasgow & Powley, Sib. Coll. Thesis, 1902. 26.4 28 Jonas & Taborsky, Z. Oest. Arch.&Ing. Verein, p. 512, 1900. = - 17500 17500 17500 = 19.3 22.0 24.2 5 = = 86.1 85.4 79.8 16.6 18.8 19.3 I Prof. B. Hopkinson, J Cambridge, 1906. 30.46 20.81 18604 37.7 . 68.3 25.8 Cat. American .Diesel En- gine Co. 88 69.6 18610 40.3 79.1 31.9 I E. Meyer, Z. d. V. d. I., 11.19 8.6 18610 35.8 77.0 27.6 > Mav, 1903. 10.69 10.18 19800 95.2(?) 13.4 Hartman, Z. d. V. d. I., 1895. p. 586. 11.68 9.22 19800 79 13.6 Hartman, Z. d. V. d. I., 1895, p. 616. 10 4.15 19600 19600 15.8 9.9 | Schottler, die Gasma- schine, p. 207. 3.05 18000 19.7 * Engineering, Vol. 88, 12.44 18000 9.2 1899. 6.07 4.95 18600 16 29 55 81.5 13 Robinson, 1893, Gas & Pet. Engines, p. 702. 31 26.74 18870 21 50 29 86.0 18 Robinson, 1898, Gas & 249.7 19460 28.1 Pet. Engines, p. 710. Kimberley & Clark Paper Co., Kimberley, Wis., Power, Oct. 1906. 16.8 19.77 7.39 9900 9900 9900 = = = = = = 31.6 32.7 21.8 1 E. Meyer, Z. d. V. d. I., April, 1903. 32-13 9700 32.1 30.18 Schimanek, Z. d. V. d. I., Jan. 17, 1903. 545 how ~ us t => d fc o fc iffe *S fc * * III s .li llll H I iI W H > \rira C< "5 O (M O * J-J Jj - O5 * i-< OJ t^; O5 C gs 1 8 I il ^^ >o co 1 1 1 1 1 1 1 1 1 1 lO Sg g | 1 ^ ^ 3 eS 'Srf "I oc to bC" bfl_ w ^ E U O Q b w fe| W E! w|^ W|OM ^ E Builder c a s. V ^3 ""! *! ^E ffl "^ Effi ji 3 "? 1 bo'tj ^1 fM ifils l.p 111 111 "a I ^ X C/3 U ^ ' ^ f 100 200 250 300 300 300 600 600 600 600 600 600 750 1200 1200 1200 1200 150 105 150 120 120 140 80 130 110 130 130 110 90 80 130 120 110 Cockerill. . .. Cockerill. . .. Cockerill. . .. Deutz Deutz Deutz Cockerill Cockerill Oechelhauser . . Deutz Deutz Koerting Niirnberg Cockerill Deutz . . Nurnberg Oechelhauser . . 1 1 1 2 4 1 2 1 2 4 1 1 2 4 4 4 4 4 4 4 4 4 4 2 4 4 4 4 4 4 S. S. S. s. s. s. s. s. s. s. d. s. s. s. s. s. sc. sc. td. sc. tw. d. tw. sc. td. sc. tw. d. tw. sc. sc. td. d. tw. d. tw. tw. 45,000 83,000 65,000 83,500 101,000 110,000 207,000 185,000 143,000 158,000 189,000 136,500 297,000 365,000 354,000 280,000 260,000 9,000 25,000 10,000 35,000 14,000 3,500 100,000 46,000 48,000 28,000 7,000 18,000 115,000 ' 95,000 14,000 16,000 16,000 21,100 58,500 23,400 81,800 32,800 8,200 234,000 107,500 112,000 65,500 16,400 42,200 26,900 222,000 32,800 37,400 37,400 540 540 300 295 383 484 512 386 318 310 327 258 560 383 307 246 230 661 706 353 551 447 394 734 487 425 371 342 297 538 488 322 264 248 2.05 1.81 1.24 2.07 1.52 1.32 0.99 1.13 1.23 1.67 1.08 1.11 1.03 0.68 1.01 0.94 0.9 1200 1400 110 110 Koerting Cockerill 2 2 2 4 d. d. tw. td. 250,000 374,000 4,500 8,600 10,500 20,000 212 170 218 164 0.9 0.42 It will be noted from column 6 that there is only one double- acting four-cycle engine in the list. The large engines cited are all single-acting twin or tandem, or double engines. It would naturally be assumed that the later types of large four-cycle engines, which are nearly always double-acting tandem, or double-acting twin tandem machines, should show a decrease in the floor space required as compared with the figures in column 13. An indication of the saving in both weight and floor space that may be effected by adopting the double-acting principle is given by the figures for the 1400 B. H. P. double- acting tandem Cockerill engine in the last line of the table. The saving there shown is remarkable and the writer has not been able to check it in the case of other engines the design of which has been changed from the twin or double-twin single-acting to the double-acting tandem. The reason for this probably is that opposed single-acting engines usually work without a cross-head, COST OF INSTALLATION AND OF OPERATION 553 while double-acting cylinders demand the use of the same. Hence the total length of engine is not changed materially. 5. Cost of Operation. The total cost of operation consists of the following items: (a) Interest on capital. (6) Depreciation of Plant and Buildings. (c) Insurance. (d) Fuel Cost. (e) Cost of Cooling Water. (/) Lubricating Oil and Waste. (g) Attendance. (h) Maintenance and Repairs. Of these, the first three items are usually called the fixed charges, and the last five the operating or works cost. The sum of fixed charges and works cost is the total operating cost. (a) INTEREST ON CAPITAL. The usual allowance in this country for interest on capital is 6 per cent. (6) DEPRECIATION OF PLANT AND BUILDINGS. There is little doubt that as far as gas producers are concerned, the wear and tear on this part of the plant is much less than it is on a boiler plant of the same capacity. On the other hand, the stresses in a gas engine are generally higher throughout than those in a steam engine of the same power, and this naturally leads to a shorter life and consequently somewhat higher allowance for deprecia- tion. Taking it altogether, an allowance of from 7 to 10 per cent of the capital outlay for the power plant should cover deprecia- tion, the lower figure for a producer plant, the higher where engines alone are concerned. Depreciation on the building should not exceed 2 to 3 per cent of the building cost. In case the building is rented, the rent instead of depreciation should be charged against the plant. (c) INSURANCE. This item is usually very small as compared with the rest and is therefore in most cases neglected. (d) FUEL COSTS. This item is very often used as the sole criterion of the economic status of a plant, but in many cases this is not true. Where the cost of fuel is very high, as for instance where illuminating gas or gasoline are used, the fuel cost usually forms the major part of the operating expenses. But where the cheaper fuels are used or the plant is of high efficiency, as is often 554 INTERNAL COMBUSTION ENGINES the case in producer plants, an analysis will show that what are usually considered the " incidental" expenses in many cases far outweigh the fuel cost. The fuel cost varies with the load on the engine. The figures given by manufacturers usually represent the best figures obtained at maximum load, but such conditions rarely obtain in practice. At any rate it would not be safe to base computations upon such " parade" figures. Haeder * considers that the economy figure given by an engine at T 7 ^ maximum load, which is equal to 85 per cent of rated load if an over-capacity of 20 per cent is assumed, is the best figure on which to base computations. The same writer gives the following table for the variation of the fuel consumption with load. The third line in this table gives the same information regarding a steam plant: VARIATION OF FUEL CONSUMPTION WITH LOAD Gas Engine .1 Max. Load .9 .8 .7 .6 .5 .4 .3 .2 of Max. Load Hit-and-miss regulation 1 1.03 1.08 1.14 1.23 1.35 1.50 1.8 2.5 3.0 Throttling regulation . Steam plant 1 1.1 1.06 1.05 1.13 1.02 1.21 1.0 1.33 .96 1.50 1.02 1.75 1.07 2.2 1.15 3.0 1.30 5.0 1.60 The table is based upon the assumption that the gas engine operates normally on T 7 ^ of its maximum load, and this is put equal to the normal load on the steam engine. It is interesting to note the superiority of hit-and-miss regulation, as far as econ- omy is concerned, over the other method of governing, and the superiority of the steam engine over both as regards small varia- tion in economy over a wide range of load. Gulderf makes a similar estimate with the following results: INCREASE OF FUEL CONSUMPTION WITH DECREASING LOAD Approximate Load 75 .66 .50 .33 Illuminating Gas Engine . 10 20 35 60 Suction Gas Engine 20 30 50 75 Diesel Oil Engine 10 20 30 55 .25 of Normal Load. 90 ) per cent higher fuel 100 > consumption than 80 ) at full load. * Haeder, Die Gasmotoren. f Guldner, Verbrennungsmotoren, p. 427. COST OF INSTALLATION AND OF OPERATION 555 Thus, for example, if an illuminating gas engine uses 20 cu* ft. of gas per B. H. P. hour at full load, we may except it to use 1.35 X 20 = 27 cu. ft. per B. H. P. hour at half load, and 1.9 X 20 = 38 cu. ft. per B. H. P. hour at quarter load. The anthracite coal consumption of suction gas plants for various sizes of plants and for different loads is given by Haeder in the following table: CONSUMPTION OF ANTHRACITE IN SUCTION GAS PLANTS IN POUNDS PER B. H. P.-hour Max. Capacity B. H. P. ... 14 40 70 100 140 210 420 Normal Load B. H .P. 10 30 50 70 100 150 300 Consumption pounds per H. P.-hour B. .7B. H.P H.P max. max. 1.21 1.45 1.0 1.19 .93 1.12 1 90 10 .88 1.06 .86 1.03 .84 1.01 based on . . .5B. H.P max. 1.80 1.47 1.38 1. 34 1.32 1.28 1.25 It is evident from a study of the figures in the above tables that the probable average load at which a prospective plant will operate plays an important part in the estimation of fuel costs, a point which should not be lost sight of. We find in the engineering literature of the day considerable information regarding fuel consumption and fuel costs. The cost figures given belong to one of two classes: either they are based upon assumptions regarding both consumption of fuel and cost of unit weight of the fuel, or the figures are obtained in actual operation. It goes without saying that the latter class is much more valuable than the former. In the case of assumed costs and consumptions, comparisons are usually made between steam and gas power under various conditions of operation. Such computations are interesting because they show what might be realized in each case, but in attempting to practically apply the information the greatest attention should be paid to the conditions assumed. As regards actual consumption figures, the tables of the previous chapter give considerable information regarding efficiency of operation. From this data it should be easy to compute the fuel costs as soon as the local cost of unit weight is known. Below are given a few hypothetical computations. In some cases both the fuel consumption and the fuel cost are assumed 1 outright, in others an attempt has been made to determine the 556 INTERNAL COMBUSTION ENGINES fuel consumption in relation to the size of the engine. These figures are followed by a few fuel cost figures from actual practice. Additional data for fuel cost from actual operation will be found in the data on total operating costs at the end of this chapter. The following table shows an estimate taken from the cata- logue of a well-known manufacturer. RELATIVE COSTS OF FUEL WITH DIFFERENT TYPES OF ENGINES Type of Engine Class of Fuel Price of Fuel Fuel Con- sumed per B. H. P. per hour Cost in cents per B. H. P. per hour Cost per annum 3000 hours per 100 B. H. P. Simple n on- con- densing slide valve Compound con- densing Steam Turbine Oil Engine Bituminous Coal Bituminous Coal Bituminous Coal Gasoline Crude Oil IlluminatingGas Natural Gas . . . Anthracite Coal Coke $3.00 per gross ton $3.00 $3.00 14 cents per gal. 4 cents per gal-, 75 cents per 1000 cu. ft. 20 cents per 1000 cu. ft. $3.00 per gross ton $3.00 per gross ton 51bs. 3 It*. 3 Ibs. 0.125 gal. 0.10 gal. 19 cu. ft. 13 cu. ft. 1 Ib. 1.251b. .669 .402 .402 1.75 .40 1.425 .26 .134 .167 $2,007.00 1,205.00 1,205.00 5,250.00 1,200.00 4,275.00 780.00 402.00 502.00 Oil Engine Gas Engine. . . Gas Engine Gas Engine with suction producer. . Gas Engine with suction producer. . The next table, compiled by J. I. Wile, is similar to the above, except that the assumptions differ somewhat. A comparison of the corresponding items in the two tables serves to show the variation in the final results arrived at by computations of this kind. COST OF INSTALLATION AND OF OPERATION 557 STATISTICS OF FUEL CONSUMPTION AND COST PER ANNUM OF PRODUCER GAS POWER AND OTHER POWERS Type of Engine Kind of Fuel Price per Ton Fuel consump- tion in Ibs. per B. H. P. per Cost in cents per B. H. P. Cost in dollars per 100 B. H. P. Cost per 100B.H. P. per Annum hour per per 3000 hour hour hours O P Oil Engine Gasoline 12 cents per 1 Pint 1.50 $1.50 $4,500 1 Gallon i Gas Engine Illuminating 75c per 1000 18 Cubic Feet 1.35 1.35 4,050 w Gas Cubic Feet per B. H. P. Gas Engine Natural Gas 30c per 1000 13 Cubic Feet .39 .39 1,170 -*- a Cubic Feet per B. H. P B Simple Steam Engine . . . Bituminous Coal $3.00 8 Ibs. 1.2 1.20 3,600 | V C/3 I Compound Steam Engine Non-condensing Bituminous Coal 3.00 5 .75 .75 2,250 3 M Triple Expanding Steam Engine, condensing . . . Bituminous Coal 3.00 2 .3 .30 900 Producer Gas Engine . . . Anthracite 5.00 1 .25 .25 750 1 "^ Coal 3 Producer Gas Engine . . . Gas Coke 3.00 1.25 .1875 18| 565 Producer Gas Engine . . . Bituminous 2.50 1.25 .1565 !l565 470 Coal ! Producer Gas Engine . . . Anthracite 3.00 1 .15 .15 450 E? Coal 1 C. H. Day,* in an investigation on the economy of gas pro- ducer engines, compiled the following fuel cost data for the various kinds of prime movers mentioned. The fuel consump- tion figures were obtained in each case by collaborating the results of a considerable number of tests. The costs per annum were then computed by assuming the cost of unit, weight of the fuel. The total time of operation per year is taken at 3000 hours. ANNUAL FUEL COSTS FOR STEAM ENGINES. COST OF COAL ASSUMED AT $3.00 PER. TON OF 2000 LBS. Type of Engine B. H. P. Lbs. Coal per B. H. P. per hour Cost of Coal per B. H. P. per year Simple, non-condensing Simple, condensing Compound, condensing Compound, condensing Compound, condensing Triple exp. condensing 50 100 200 600 1000 2000 5.75 4.45 2.74 1.97 1.90 1.87 $25.90 20.00 12.70 8.85 8.55 , 8.40 The data available for steam turbines is not as extensive as *Sibley College Thesis, 1905. 558 INTERNAL COMBUSTION ENGINES that for steam engines, and the range covered is not so wide as regards capacity. The following- figures, also compiled by Mr. Day, show some fuel costs for prime movers of this kind. The working time has again been assumed at 3000 hours per annum. ANNUAL FUEL COSTS FOR STEAM TURBINES. COST OF COAL ASSUMED AT $3.00 PER TON OF 2000 LBS. Number of Tests Average B. H. P. Lbs. Coal per B. H. P.-hour Cost of Coal per B. H. P. per year 27 10 5 25 616 1085 1359 1739 1.740 1.735 1.725 1.655 $7.83 7.80 7.75 7.44 The computations made on illuminating and natural gas engines show the following results. Time of operation per year 3000 hours, cost of illuminating gas 75 cents per 1000 cu. ft., cost of natural gas 50 cents per 1000 cu. ft. The latter assumption is high, since in many localities natural gas is sold at 30 or even 20 cents per 1000 cu. ft. It is a simple matter, however, to reduce the figures in the table in the corresponding ratio. ANNUAL FUEL COST FOR ILLUMINATING AND NATURAL GAS ENGINES. COST OF ILLUMINATING GAS ASSUMED AT 75 CENTS, THAT OF NATURAL GAS AT 50 CENTS, PER 1000 CU. FT. Cu. ft. per B. H. P.-hour Annual Cost of Gas per B. H. P. per year of 3000 hours B. H. P. Illuminating Gas Natural Gas Illuminating Gas Engines Natural Gas Engines 10 22.0 $49.50 $24.00 20 21.5 16.0 48.30 23.50 30 21.0 15.5 47.20 23.30 40 20.5 15.0 46.00 22.60 50 20.0 14.5 44.80 21.80 75 19.1 13.8 42.80 20.80 100 18.3 13.0 41.10 19.60 200 11.4 17.10 300 10.2 15.40 400 9.7 14.60 500 9.4 14.10 COST OF INSTALLATION AND OF OPERATION 559 For producer gas engines Day found the following figures for the consumption of gas and of coal, from which the fuel costs per year of 3000 hours are computed, assuming coal to cost $4 per ton of 2000 pounds. ANNUAL FUEL COST OF PRODUCER GAS ENGINES B. H. P. Cu. ft. of Gas per B. H. P.-hour Lbs. of Coal per B. H. P.-hour Annual Cost of Coal per B. H. P. 50 105 .45 $8.70 100 96 .35 8.10 150 89 .23 7.33 200 83 .17 7.03 250 79 .12 6.72 300 76 1.08 6.48 400 73 1.05 6.30 500 72 1.03 6.18 . The cost of the gas for blast furnace gas engines has in many computations been neglected on the assumption that this gas, if not used in engines, is a mere waste product. Lately, however, it has come to be recognized that some value must be assigned to this fuel since money was expended on it in every. case for clean- ing it preparatory to making it fit for use in engines. The ordinary method of evaluating this gas is to compare its heat content with that of steam and to assign a value to the gas corresponding to the cost of steam. Thus the cost of the gas will in every locality vary with the cost of coal. As an example, H. Freyn * makes the following computation: "Let us assume that the price of coal delivered into bins at the plant be $2.75 per ton, that the coal have a heat value of 13,000 B. T. U. per pound, and, further, that steam of 150 pounds boiler pressure or about 165 pounds absolute pressure be raised by burning this coal under boilers. One pound of steam will then contain 1225 B. T. U. from zero degrees Fahrenheit. Assum- ing feed water at 70 degrees, there would be required 1 155 B. T. U. to generate 1 pound of steam at 150 pounds boiler pressure. In a boiler plant having 65 per cent efficiency, 1000 pounds of coal * H. Freyn, Available Power and Cost of Operation of a, Power Station for Waste Gases from a Blast-furnace Plant. Journal Western Society of Engineers, February, 1906. 560 INTERNAL COMBUSTION ENGINES could raise (0.65 X 1000 X 13,000) -i- 1155 = 7300 pounds of steam. The value of 1000 pounds of steam would be 2.75 -r- (2 X 7.3) .= $0.188, or 18.8 c. To this must be added for labor and maintenance approximately 1 c. per 1000 pounds of steam, making the total value of 1000 pounds = 19.8 c. 1000 cu. ft. of blast furnace gas have a heat value of 1000 X 90 = 90,000 B. T. U. and are equivalent to (0.65 X 90,000) H- 1155 = 51 pounds of steam, which in turn are worth (51 -r- 1000) X 19.8 = 1 c. "The value of 1000 cu. ft. of blast furnace gas would, therefore, be 1 c." L. Eberhardt, in an article in the Zeitschrift d. V. d. I,* makes a similar computation, arriving at a somewhat different result. Starting with the assumption that 1000 pounds of steam will cost from 25 to 32 c., according to the price and quality of coal, he finds that 1000 cu. ft. of cleaned blast furnace gas should have a value of from 1.43 to 1.84 c., the gas having a heating value of 102 B. T. U. per cu. ft. This is considerably higher than the value found by Freyn, which is mainly due to the lower grade of coal (11,650 B. T. U. per pound) and the higher grade of gas assumed. To get some idea of the fuel cost of blast furnace gas power as compared with steam power, Eberhardt, in the article men- tioned, gives the following tables. The original tables give the cost per B. H. P. hour, but these have been recomputed to the basis of a year of 3000 working hours, to make them directly com- parable with the figures in previous tables. ANNUAL FUEL COST OF BLAST-FURNACE GAS ENGINES. HEATING VALUE OF COAL TAKEN AT 11650 B. T. U. PER LB. HEATING VALUE OF GAS TAKEN AT 102 B. T. U. PER cu. FT. Cost of coal per ton of 2000 Ibs dollars 2.34 2.75 3.18 Cost of 1000 Ibs. of steam cents 24.8 28.4 31.8 Cost of 1000 cu. ft. of gas cents 1.43 1.64 1.84 A fair blast furnace gas engine will show the following fuel consumption : At full load 99 cu. ft. of gas per B. H. P. hour. At T 9 ^ load 106 " " " At T \ load 113 " " At load 122 " " At i load 131 " " * June 3, 1905. COST OF INSTALLATION AND OF OPERATION 561 With these assumptions of gas consumption per B. H. P., and of local cost of coal, the fuel costs per B. H. P. per year of 3000 hours will then be as per the following table: Cost of Coal per Ton Per Cent full Load on Engine 100 90 80 66 50 Annual Cost of Blast Furnace Gas per B. H. P. Dollars $2.34 2.75 3.18 4.26 4.85 5.46 4.56 5.18 5.83 4.87 5.55 6.23 5.25 5.97 6.72 5.63 6.42 7.20 The fuel cost of operating on gasoline as fuel is of course con- siderably higher than any of the figures above quoted. The usual assumption made in regard to gasoline engines is that they will require 1 pint of gasoline per B. H. P. hour. This corresponds to a thermal efficiency of about 18 per cent on the brake, a figure which should be reached by a fair-sized engine in good condition. For the smaller machines, however, one pint is perhaps some- what low, and for say a 2 horse-power machine a consumption of 2 pints per B. H. P. hour is probably a safer assumption. It should also be borne in mind that the purchase price of gasoline varies somewhat with the quantity bought. The following table of fuel costs takes these various points into account. FUEL COST FOR GASOLINE ENGINES PER B. H. P. PER YEAR OF 3000 WORKING HOURS Size of Engine, B. H. P 2 4 6 10 20 Cost of gasoline per gal., cents 20 18 16 14 13 Consumption of gasoline per B. H. P. hr., gallons 25 .20 .18 .15 .12 Fuel cost per B. H. P. per year, dollars 150.00 108.00 86.40 63.00 46.80 (e) COST OF WATER FOR COOLING AND WASHING. Many esti- mates of total operating costs totally neglect the cost of water for cooling and washing purposes required by the plant, although in many localities this may amount to a considerable item of expense. In general terms, where water has to be brought from city mains, it pays to install cooling apparatus of some kind. Of course for small installations this may be simply a tank from 562 INTERNAL COMBUSTION ENGINES the surface of which the heat is radiated. The tank is connected at top and bottom with the water jacket of the engine, and the water circulates by convection. Where such a cooling tank can- not be placed in the immediate vicinity of the engine, and con- nected to the jacket by short straight pipes, a circulating pump of some kind must be used. As the size of the installation grows, cooling towers or cooling ponds have to be resorted to, unless an abundant source of clean water is available without cost, except that of pumping. Under such conditions the cost of cooling water is very materially reduced. With any kind of cooling system the cost then consists of the cost of pumping plus the cost of any clean water that must be supplied from time to time to replace that lost by evaporation. Of course it is not possible to give any definite figures for this cost item, because it depends upon the cost of water in the particular locality, upon the total amount of water circulated, and upon the kind of pumps used. A few figures for the amount of water required or in circulation per B. H. P. hour are given below. In a producer plant an additional amount of water is required for the producers and scrubbers. The water vaporized for the producers of course is not again available. That used for the washing plant is in general subject to the same considerations as that used for cooling the engines. But where a cooling system is used for both, they should never be combined, because the re- quirements for clean cooling water for the engines are much more strict. The scrubber water usually carries considerable quantities of sediment and is contaminated with ammonia and sulfur com- pounds, so strongly in some cases as to give it a very noxious odor. In such cases, settling tanks and ponds, together with a considerable addition of clean water, seem to be the only remedies to recover at least a part of the water. The approximate quantity of cooling water required may be figured from the following considerations: One horse-power, assuming an engine efficiency of say 25 per cent at the cylinder, requires the expenditure of - = 10,200 .25 B. T. U. per I. H. P. hour. The jacket loss approximates about 40 per cent of the heat expended, that is, in this case the jacket water must carry off .40 X 10,200 = 4080 B. T. U. per hour. COST OF INSTALLATION AND OF OPERATION 563 Assuming the inlet temperature at 70 degrees, the outlet at 150 degrees, we find the number of pounds of water required = 4080 approx. - - =51 pounds =say 6 gallons per I. H. P. hour 1 oU 7 u at full load. The following figures from various sources show how this estimate agrees with others and with some data from actual practice. W. Heym, in the Gasmotorentechnik* states that a producer plant requires per horse-power hour 6.5 gallons for the engine, 4 gallons for the scrubber, and .13 gallons for the producer vapo- rizer. Freyn, in the paper already mentioned, makes the following estimates for a blast furnace gas plant of 10,500 B. H. P. Theissen washers are used to clean the gas. GALLONS OF COOLING WATER REQUIRED Ixxul on Engines For the Engines per B. H. P.-hour For the Washers per 1000 cu. ft. of gas cleaned per minute Full I 8.5 10.5 13.0 12.0 14.0 16.0 Freyn also estimates that if the plant is located near a stream of water, the cost of pumping would probably be in the neighbor- hood of 2 c. per 1000 gallons. J. R. Bibbins f on a 51-hour test of a 500 horse-power Westing- house horizontal engine, reported a water consumption for the engines only of 9.4 gallons per B. H. P. hour at full load. The same authority reports a figure of 5.65 gallons per B. H. P. hourt found on similiar engines in another plant. In conclusion it should be said that the water-consumption of gas engines depends somewhat upon the size of the cylinder. Thus a single-acting cylinder of very large diameter usually * November, 1907. f J. R. Bibbins, Proceedings A. S. M. E., Mid-November, 1907. t J. R. Bibbins, Gas Driven Electric Power Station,Proc. of the Eng. Soc. of W. Pa. 564 INTERNAL COMBUSTION ENGINES requires more cooling water than two single-acting or one double- acting cylinder of the same capacity. Skilful attendance also is a considerable factor in the amount of cooling water used, a point which is very often neglected. (/) OIL AND WASTE. The consumption of lubricating oil per B. H. P. hour depends upon the size of the engine and, as in the case of cooling water, largely also upon the care of attendants. It is also true that new machines may for a time require two or three times the ordinary amount of oil until all the parts have become adjusted to their service. In large and medium sized plants it is usual to employ a gravity or circulating system of oiling, in which case the oil is recovered, filtered, and again used. In such systems the cost of oil is a small item and consists mostly of the cost of oil necessary to replace that unavoidably lost. Of course none of the cylinder oil used is recovered. Giildner estimates that the consumption of lubricating oil usually varies from .006 to .008 pints of oil per B. H. P.-hour, and that under favorable conditions .003 pints per B. H. P.-hour may be reached. L. L. Brewer finds that the consumption for the large engines quoted in the table, page 552, varies from .0045 to .0055 pints per B. H. P.-hour, and that in double-acting two-cycle engines the consumption may be as low as .0035 pints. The following table compiled by J. R. Bibbins and published in the paper before the Society of Engineers of West Pennsylvania, already quoted, shows the actual oil consumption in a plant con- taining two horizontal Westinghouse engines of 500 horse-power each, direct connected to 300 K. W. generators. The figures cover a period of four months and the results should therefore be very reliable, - The system of oiling used is of the continuous circulating and filtering type. The consumption amounts to .00202 pints of cylinder oil and .00286 pints of engine oil per horse-power hour, which agrees well with the figures given by Brewer. COST OF INSTALLATION AND OF OPERATION 565 OIL CONSUMPTION 4 months ending Sunday, May 23, 1906 CYLINDER OIL * ENGINE OIL f Kind Imperial Cylinder Oil Special Gas Engine Oil Maker Union Petroleum Co., Philadelphia Price (barrel lots) 32 cents per Gal. 18 cents per Gal. Quantity used 561 Gals. 765 Gals. Quantity used per month 140 Gals. 191J Gals. Quantity used per (opertg.) day . . . 4.68 Gals. 6.38 Gals. Quantity used per full day 6.07 Gals. 8.27 Gals. Engine Hours per day 37 37 Engine H.P. Hours per day 18500 18500 Oil per Engine Hour 0.127 Gals. 0.172 Gals. Oil per H.P. Hour 0.000253 Gals. 0.000345 Gals. Cost per Engine Hour 4.05 cents 3.1 cents. Cost per H.P. Hour 0.00809 0.00621 Total Cost 0.0143 cents per H.P. Hour The cost of cotton waste or other cleaning material is an item very hard to estimate, and in any case of little influence on the final result. It is usually combined with the cost of oil, and some figures for this combined cost will be found in the estimates of total operating costs at the end of this chapter. (g) ATTENDANCE. The statement very often carelessly made with regard to small gas engine installations, that they require no waiting on, is of course not quite the truth. The fact is that a small gas or gasoline engine, if in good order, does not require, after starting, much attending except the proper handling of the lubricators. In this respect the gas power installation has an advantage over the small steam power plant, where at least one man is generally required all the time. For small natural gas or illuminating gas engines, therefore, it is possible to employ the " engineer," to use a familiar term, somewhere else at least part of the time, but naturally, as the size of engine increases, this spare time decreases, and it is doubtful if suction gas plants, no matter how small, can get along with less than one man's entire time. The general methods of taking care of medium sized and large * Includes crank case oil for exciter engines. Drainage from glands of main engines collected, mixed with old engine oil and used in crank case. No crank case oil purchased. f Includes oil consumption o.f auxiliaries. Sufficient quantity old engine oil drawn from circulating system to supply auxiliaries replaced by fresh oil. 566 INTERNAL COMBUSTION ENGINES gas engines are perhaps not far different from those used in steam engine practice. It cannot be denied, however, that gas engines have much more opportunity, so so speak, to go wrong. That is, jacket water, lubrication, igniters, valves, etc., all need careful looking after, and any of these things, if neglected, may cause a shut down. While, therefore, steam engine attendants generally accustom themselves to the new service very quickly, it would be wrong to assume that they can in general take care of a gas engine plant without some instruction and practice. Of course the larger the plant, the more important this point becomes. There seems to be nothing in English engineering literature to give an approximate idea of the time actually required in attending gas engines. Giildner proposes the following equations, apparently based on practical experience: For illuminating gas, natural gas and oil engines, n hours. For suction gas plants W = 1.25 vA/ hours where W = time of attendance required in hours per day of 10 hours and N n = rated capacity of plant. Thus, for instance, a 100 horse-power natural gas engine would actually require about .25 V 100 = 2.5 hours in a 10-hour shift. Hence the attendant's time could be largely used somewhere else. On the other hand, a 100 horse-power suction gas plant would actually require 1.25 V 100 = 12.5 hours, which means that two men will have to be employed. Where a station is made up of several smaller units and pro- ducers, the cost of attendance naturally increases. Thus if there are n units in a plant of N n horse-power total, Giildner states that the actual time required per operating day of the plant may be expressed by W = 1.25 VnAT hours. Thus if we take a 1200 horse-power suction gas plant made up of 4 units of 300 horse-power each, W 1.25 V4 x 1200 = 86.5 hours, which would mean the service of 9 men in the 10-hour shift. If COST OF INSTALLATION AND OF OPERATION 567 this capacity had been put in one unit, the time would have been W = 1.25 VT200 = 43.5 hours, which would have called for only 5 men per shift. If in any of the above equations we introduce the hourly wage scale, we may express the cost of attendance per B. H. P. per hour. Thus if the scale should be 20 c. per hour, we would have for illuminating gas, natural gas, or oil engines, for the 10-hour shift, Cost per B. H. P.-hour - 20 X . 10 N For suction gas plants, Cost per B. H. P.-hour = 20 X 1.25 = _^ 10 N n VN n and for large suction gas plants of n units, Cost per B. H. P.-hour -^Xi^ cents. There is little data available to check the accuracy of these formulae, but the two or three instances cited by various authori- ties for the labor cost in suction gas plants check very well with the results of the formula. (h) MAINTENANCE AND REPAIRS. The expenditures for maintenance and repairs of a gas engine installation should not exceed 3 per cent of the purchase price of the engines and pro- ducers. There is no doubt that this item may easily run to 10 per cent and over, especially if in a producer plant insufficient attention is paid to cleaning the gas. But such conditions are not normal and should not occur in good practice. Freyn states that the following are average figures for repair accounts for a blast furnace gas station : 2 per cent per year of the price of the engines and electric generators. 7 per cent per year of the price of the cleaning plant. 5 per cent per year of the price of the air compressor used for starting. 2 per cent per year of the price of piping, etc. To this may be added from 1 to 2 per cent of the cost of the building as maintenance for the building. 568 INTERNAL COMBUSTION ENGINES Total Operating Costs and Cost as Compared with other Prime Movers As in the case of fuel costs, the information available in engineering literature on total operating costs is of two kinds. In the first kind the estimates are based entirely upon hypothetical assumptions, in the second the results are those obtained in actual operation. Again, the latter informa- tion is of course of much greater value, but except for large in- stallations little of this class has been made public. Especially as regards small installations of gasoline, oil and illuminating gas engines there seems to be an almost absolute lack of data referring to total operating costs. For that reason, in order to get some idea of costs, it will be necessary for the time being to depend upon cost computations based on assumptions. In any concrete case these assumptions should be carefully scanned to see how they agree with actually existing conditions before forming a final idea of operating cost. The data available from actual practice is mainly due to Mr. J. R. Bibbins * and others, of the Westinghouse Machine Com- pany, and of course relates to Westinghouse engines. This, how- ever, does not preclude their applicability to other similar plants. This information, combined with that from a few English plants, is about all that can be cited at the present writing. Hypothetical computations are nearly all made on a compara- tive basis. Of -this kind are the tables published by W. O. Webber in the Engineering News, August 15, 1907. It is not easy to give a concise abstract of the various factors assumed in the compu- tation of these tables, and for that reason they are given in full, except for electric power which does not especially interest us here. * The following is a partial list of the valuable papers published : J. R. Bibbins, Gas Driven Electric Power Systems, Warren & Jamestown Street Railway, Eng. Soc. of W. Pa.; Gas Power for Central Stations, Am. Inst. of E. E., New York, December 18, 1903; Producer Gas Power Plant, Proc. A. S. M. E., December, 1906; Application of Gas Power to Central Station Work, National Elec. Light Assoc., Washington, D. C., June, 1907; Duty Test on Gas Power Plant, Proc. A. S. M. E., Mid-November, 1907; and A. West, Gas Power in Electric Railway Work, American Street & Interurban Railway Assoc., Philadelphia Convention, 1905. E. E. Arnold, The High Power Internal Combustion Engine and its Fit- ness for Central Station Work, New England Street Railway Club, March, 1904. Reprint from Power, December, 1903, A Gas Engine Pumping Station. COST OF INSTALLATION AND OF OPERATION 569 COST OF GASOLINE POWER Size of plant in H.P 2 Price of engine in place $150.00 Gasoline per B. H. P. per hour. . gal. Cost per gallon $0.22 = cost per 3,080 hours $451.53 Attendance at $1 per day 308.00 Interest, 5 per cent 7.50 Depreciation, 5 per cent 7.50 Repairs, 10 per cent 15.00 Supplies, 20 per cent 30.00 Insurance, 2 per cent 3.00 Taxes, 1 per cent 1.50 Power cost $824.03 $1,371.75 $1,498.13 $2~01&50 To these figures should be added charges on space occupied as follows: Value of space occupied $100.00 $150.00 $200.00 $300.00 Interest, 5 per cent $5.00 Repairs, 2 per cent 2.00 Insurance, 1 per cent 1.00 Taxes, 1 per cent 1.00 Total annual charge for space $9.00 Total cost per annum $833.03 Cost of 1 H.P. per annum, 10 hour basis 416.51 Cost of 1 H.P. per hour $0.1352 6 $325.00 igal. $0.20 10 $500.00 fgai. $0.19 20 $750.00 igal. $0.18 $924.00 308.00 16.25 16.25 32.50 65.00 6.50 3.25 $975.13 308.00 25.00 25.00 50.00 100.00 10.00 5.00 $1,386.00 308.00 37.50 37.50 75.00 150.00 15.00 7.50 $7.50 3.00 1.50 1.50 $10.00 4.00 2.00 2.00 $15.00 6.00 3.00 3.00 $13.50 $18.00 $27.00 $1,385.25 $1,516.13 $2,043.30 239.87 $0.0780 151.61 $0.0492 102.17 $0.0331 COST OF GAS POWER $1.50 per 1000 cubic feet of gas less 20 per cent, 'if paid in 10 days = $1.20 net, gas 760 B. T. U. Size of plant in H.P 2 6 10 20 Engine cost in place $200.00 $375.00 $550.00 $1,050.00 Gas per H.P. -hour in cubic feet. 30 Value of gas consumed, 3080 hours Attendance, $1 per day Interest, 5 per cent Depreciation, 5 per cent Repairs, 10 per cent Supplies, 20 per cent Insurance, 2 per cent rri i $221.76 308.00 10.00 10.00 20.00 40.00 4.00 2.00 $615.76 Annual charge for space 9.00 Total cost per annum $624.76 Cost of 1 H.P. per annum, 10 hour basis 312.38 Cost of 1 H.P. per hour $0.1014 Insurance, 2 per cent Taxes, 1 per cent . Power cost 25 $554.40 308.00 18.75 18.75 37.50 75.00 7.50 3.75 $1,023.65 13.50 172.86 $0.0561 22 $843.12 308.00 27.50 27.50 55.00 110.00 11.00 5.50 $1,387.62 18.00 110.56 $0.0456 20 $1,478.00 308.00 52.50 52.50 105.00 210.00 21.00 10.50 $2,237.50 27.00 $1,037.15 $1,405.62 $2,264.50 143.22 $003.67 570 INTERNAL COMBUSTION ENGINES COST OF STEAM POWER Size of plant in H.P 6 Cost of plant per H.P . $250.00 Fixed charge, 14 per cent $35.00 Coal per H.P.-hour, in pounds 20 Cost of coal at $5 per ton $1 512? Attendance, 3080 hours 75.00 Oil, waste and supplies lo.OU Cost 1 H.P. per annum, 10-hour basis $279.00 Cost of 1 H.P. per hour $0.0906 10 20 $220.00 $30.80 15 $103.00 50.00 10.00 $194.80 $0.0832 $200.00 $28.00 12 $82.50 30.00 6.00 $146.50 $0.0475 ANNUAL COST OF POWER PER BRAKE HORSE-POWER * B. H. P. of Unit 1 $600.00 500.00 3 " 437.50 4 ' 375.00 5 ' 320.00 6 ' 280.00 7 250.00 8 230.00 9 210.00 10 195.00 12 175.00 14 165.00 16 157.50 18 : 150.00 20 146.00 22 140.00 24 137.50 26 133.00 28 130.00 30 127.50 35 124.00 40 120.00 50 112.50 60 105.00 70 100.00 80 95.00 90 90.50 100 86.40 Steam Gas $380.00 312.50 260.00 220.00 192.50 172.50 160.00 152.50 145.00 140.00 132.50 126.00 120.00 116.50 113.00 110.00 107.50 105.00 102.50 102.00 100.00 98.00 96.00 94.00 92.00 90.00 88.00 86.00 Gasoline $487.50 416.00 350.00 300.00 262.50 240.00 210.00 182.50 165.00 152.00 137.50 122.00 112.50 107.50 102.00 98.00 95.00 92.50 90.00 87.50 85.00 82.50 80.00 78.00 76.00 74.00 72.00 70.00 Attention is called in these tables to the high allowance for repairs in the case of the gasoline and gas engine and the high coal consumption assumed for the steam engine. Another interesting comparison is that made by H. A. Clark *Unit costs: Coal, $5 per ton; gas $1.20 per 1000 cubic feet, at 760 B. T. U. ; gasoline, $0.20 per gallon. COST OF INSTALLATION AND OF OPERATION 571 in a paper on the Diesel Engine.* In this case the computations are carried out for three sizes of Diesel engine, Crossley Gas Engine and Dowson producer, and high-speed compound con- densing steam engine. The conditions assumed are partly as follows, the remainder being given in the table itself. Cost of fuel delivered : Diesel engine, crude oil at $10.90 per ton = appr. 3.8 c. per gallon. Crossley engine, anthracite at $5.70 per ton. Steam engine, coal at $3 per ton. The fuel cost is based on the following assumption: For the Diesel, the consumption has been taken in each case as the mean between full and half-load rates. These rates were actually determined by tests. For the Crossley, 1.5 to 1.25 pounds of coal per B. H. P. hour plus an allowance for stand-by losses. For the steam engine, 4 to 3.5 pounds of coal per B. H. P. hour for the lowest and highest powers, assuming an evaporation of 8 pounds of water per pound of coal. The cost figures in the original table have all been transposed to dollars and cents. The computations are based on a year of 2700 working hours. In the discussion on Clark's paper, the cost computations were rather severely criticised, mainly on account of the fuel costs assumed. The claim was made that both the coal consump- tion for the steam engine and the cost of the steam coal was assumed too high, and that the cost of oil at 3.8 c. per gallon could only apply to seaboard towns. As far as American conditions are concerned, the assumption regarding the steam engine would seem to be about right, while the cost of anthracite at $5.70 per ton for the gas engine is certainly not too low. On the other hand, the price of 3.8 c. per gallon for the crude oil would seem to favor the Diesel engine. It all comes to the point that applies to all computations of this kind, and that is that the results are strictly applicable only to the locality for which they are com- puted. Allowing for possible difference in the labor costs, how- ever, it seems to the writer that the comparison is quite fair as between the steam and the 'gas engine. * Proc. of M. E./ 1903, II. 572 , INTERNAL COMBUSTION ENGINES f 88 -8 8 88 8 8 8 8 '88;888 S < ! W o^ o J 8 8 8 8^3 g I o 10 10 o 10 10 --H oo t^ CO CO CO CD X OCO (N C01> 00 (N I 88 8 8 S 8 8 o ^ x 88 8 8 888 8 g 8 o o o oo ^ i to O 'O CO O 88 8 8 8 828 [ S ^ 8O O O O5 00 _ O CO rH t>. t^ C 10 CO rH t^ t^ CD O5 CM "tl rH CO CD 8 8 8 g g g S * W bC j a- g'SS W S C3 r-T O ^^^ o S fi w) C^l * c3 C b g "E'3) s e 2 O C 01 H O^ 03 > S L3 S OJ fe-al Iff 111 ||1 vr ^ -M 1 88 . .8 rf CO ' ' -l S .8 , . . S(M (N '1 T *! Tji rH O 8S . .8 O . O . CO CO O5 CO o O Orf t^ (N lO LO O5 O5 ^ TH oc i-3 d - oq ci d co -H TJ? r-4 d ci p ip co c oiddd O O . >O iO CO rH CO 1C i-5 d 8888 . QO CO CO O ' S2 CI^|. i i d - i co d 't^r-5 ic'cor-Jd cu o> P ^ 4 . . . er ted hoi iner er ow ra ed ch 2 'li p per rat ma isr:3i^l* o o o o o CO CB OD 03 O3 OOOOO 0000^ < i - 00 O> O 'I-H'C^ CO Tf ic CO t^'(X) COST OF INSTALLATION AND OF OPERATION 575 d!(NrH' 00 10 t^ CO O O 00 t>. to 00 CO s O5 00 CM tO b- Tf< rH p oqoq o 06 ^ co CM CO O O t>- CM CM O5 IO rH|> t>. rH CO CM 00 r-Hp O^ ^f d d O5 co t-~ to t>. O5 t^ O5 rH rH "H^ CO O CO 05 00 rH 00 O to d d O CM rH CO !> 00 tot> doo rH CM 00 CO o co Tji O **& CM O tO O rH 8 CO 00 CO to rH tO CO S rH .CO rH CM CO O .- O Ob- 00 Tfl tO IO * CO O O5 00 rHCM rH CM O O l^- to CO CM rH CO o oo rH tO I s * CM rH O iO OO p t^ >O rH OO co co O to t^ to r-5 o d d tot^ co O5 00 tO d c1 S O CM OOO co CM p to 55 (M CO" CO' CO CO O tO CO rH CO 00 O to 05 CO rH O rH CO rH rH CO O5 p l>; tq CO rH O '^ K i 35_ rHO COCO CM 00 i 1|> rH rH CM O5 COO T 1 '-' t^ O5 CM rH & CO rH Tf CM .8 CO CM to CM rH O a i to O CO O5 CM Tt^ 10 TJV CM rH O O CM CO q Tt? co 06 rH Tfl CM 7373 7373 CM CO tq rn' 00 tO O 00 l> p t> CM CM CM d d d CM rH CM rH 4H "V o : ,, ^^ -ft 51;- H: : PQ pe '^ g 2 S^ g 2^ OO^OJ^CJ^^ ^'o^'B'o'flCM ^ 66 < < o PH J :* ' ffi ^ : "^ : ;-|j| o o> b <5 :f|^|? 3 S" 53 C O* C 3 rH C/3 g, -0-^ 'CMO ' .a i " O e ss OJ CO O, S CO T-5 O5 prH rH CM CM CM CM CO CO CO CO Tfl IO COCO t>- 00 COCO O5 CO 576 INTERNAL COMBUSTION ENGINES The second and third tables following are due to Mr. J. R. Bibbins of the Westinghouse Machine Company and were pub- lished in a paper on the Application of Gas Power to Central Station Work, read before the National Electric Light Associa- tion at Washington, June, 1907. Both of the stations mentioned use natural gas for fuel, but the Bradford plant is equipped with rather small Westinghouse vertical engines, while the Warren and Jamestown station contains two large direct connected units of 500 horse-power each. In both cases the load factor is con- siderably less than 50 per cent; in the case of the Bradford plant it is only about 19 per cent.' The figures for the latter plant are remarkable for the length of time covered, 8J years. WORKS COST OF OPERATING GAS PLANTS PU int Midland Railway Co., Leicester Urban District Station, Walthamstow Engine: Capacity tested 300 1500 Number 6 7 Make Length of run Load factor K. W.-hour generated Remarks service . . 5 months 200497 Arc and incandescent 12 months 15.25 659756 Duty of plant: * Los. per K. W.-hour . Lbs. per B. H. P. . lighting, Dowsongas. 3 2 25 lighting, Dowson gas. - 2 1 C Cost: Coal per ton Fuel per K. W.-hour Oil, waste, water . Wages $3.75 .123c. .020c. oqop $6.50 .17c. .095c. f)1 C~ Total works cost power, cents per K. W.-hour .... Authority: .473c. R. M. DEELY, Supt. Locomotives. .445c. F. A. WILKINSON, Elec. Engineer in charge. WORKS COST OF POWER GAS DRIVEN CENTRAL STATION BRADFORD ELECTRIC LIGHT & POWER Co., BRADFORD, PA. Station capacity (five engines) . . 470 kilowatts .Number of years in operation 85 Average station load factor ('03-'06) .''..... '.'. 19^09 per cent Average gas consumption per kilowatt-hour ('03-'06) 25 5 cu. ft. Average heat efficiency at switchboard 14.14 per cent * Including "stand-by" losses. COST OF INSTALLATION AND OF OPERATION 577 Works Cost of Power Dollars per year. Average 8 years Cents per Kilowatt-hour. Average 4 years Fuel, gas, including heating Station wages $3,121 3,068 0.307 0.410 Oil, waste and supplies 454 0.054 Running repairs (total) 848 0.085 Running repairs gas engines 285 0.034 Total operating cost $7,776 0.890 Engine repairs $0.36 per horse-power per year. 57.00 per engine per year. .75 per cent on investment. Fuel Bradford natural gas. WORKS COSTS GAS POWER RAILWAY PLANT WARREN & JAMESTOWN STREET RAILWAY COMPANY Capacity of plant (two units) 600 kilowatts Average time operated per day 18.5 hours Average output per day 4115 kilowatt hours Average load, per cent rating 37 per cent Works Cost of Power Dollars per day Cents per Kilowatt-hour Average First Four Months, 1906: Fuel gas $12.97 0.315 ' Wages 12 37 0300 Oil 2.63 0.064 Repairs and miscellaneous supplies . . 3.29 0.080 Total $31.26 0.759 Fuel " Bradford Sand " natural gas. In conclusion, to make the figures in the above tables readily comparable among themselves, the cost data has been recom- puted to the basis of cost in cents per B. H. P. hour in all cases. Table I gives the total operating costs for small and medium sized plants as computed from the data of Webber & Clark. The load factor is in all cases 100 per cent. The difference in the cost of steam power arrived at by the two authorities is remark- able, Clark's figures being less than 50 per cent those of Webber. The reason for this is found in two directions, different assump- tions as to cost of coal, $5. per ton as against $3, and different assumptions as to the coal consumption per horse-power. Webber 578 INTERNAL COMBUSTION ENGINES assumes as high as 20 pounds of coal per horse-power for a 6 horse-power plant, and charges a 20 horse-power plant up with 12 pounds per horse-power-hour. In fact Webber's figures are very liberal throughout for all three kinds of power. In Table II, the works and total operating costs per B. H. P.- hour are shown side by side as computed from Findlay's table. TABLE I Total Operating Costs per B. H. P.-hour Small and Medium Sized Plants E Size of Plant o >, TCinH O (-< 53 Kind of Power of ?D 8 10 | 16 20 24 30 35 50 80 100 100 J3 Fuel 2 p I 3 < Cost in cents per B. H. P.-hour Steam Engine . Coal 100 3080 13.9 6.3 5.1 4.7 4.5 4.1 4.0 3.7 3.1 2.8 Webber Gas Engine . . 111. Gas 100 3080 6.2 4.6 3.9 3.6 3.5 3.3 3.2 3.1 2.9 2.8 Webber Gas Engine . . Gasoline 100 3080 8.5 4.9 3.7 3.3 3.1 2.8 2.7 2.6 2.4 2.3 Webber Steam Engine . Coal 100 2700 1.78 1.26 .98 Clark Gas Engine . . Prod. Gas 100 2700 1.38 1.04 .80 Clark Oil Engine, Diesel .... Crude Oil 100 2700 1.08 .78 .64 Clark TABLE II Total Operating and Works Cost per B. H. P.-hour. Computed from Table by Mr. L. G. Findlay. Works costs in brackets. Kind of Power Hours in Operation per day Load Factor Size of Plant B. H. P. Steam Bituminous Producer Gas Anthracite Producer Gas Cents per B. H. P.-hour 24 100 ' 1000 .39 ( .30) .31 (.21) .39 (.30) 24 50 1000 .69 ( .41)' .54 (.33) .67 (.49) 10.25 100 1000 .53 ( .33) .46 (.22) .55 (.33) , 10.25 50 1000 .88 ( .47) .84 (.36) .99 (.56) 24 100 200 .76 ( .69) .40 (.29) .43 (.38) 24 50 200 1.23 (1.10) .71 (.49) .86 (.66) 10.25 100 200 .88 ( .73) .59 (.33) .69 (.45) 10.25 50 . 200 .1.44 (1.15) 1.11 (.59) 1.26 (.78) COST OF INSTALLATION AND OF OPERATION 579 The data from the Bradford and Jamestown plants and from the two English plants shows the following results for the works cost per B. H. P.-hour. Fuel Load Factor B. H. P. rated Works Cost per B. H. P.- hour cents Plant Natural gas 19.09 640 .67 Bradford Natural gas Dowson gas Dowson gas 37.0 15.25 800 400 2000 .57 .35 .33 Jamestown Midland Railway Walthamstow From these figures, and those of Table II, the conclusion seems justified that it should be easily possible to produce one brake horse-power in a fair sized plant, say not under 200 horse- power, for a works cost of from .5 to .75 cents per hour, depend- ing upon the conditions involved. This excludes fixed charges and assumes a load factor in the neighborhood of 50 per cent. INDEX PAGE Abeille carbureter 189 Absolute temperature 6 Accumulators or storage batter- ies 411 Acetylene, constants for 211 Adiabatic and isothermal changes, graphical ex- pressions for . . , 55 Adiabatic change 16 expansion, work per- formed in 53 line, equation of 52 Admixture of benzol to alcohol 184 After-burning 220 Air gas, production of 147 Air required for combustion of alcohol 183 Air required for combustion. . . . 136 Air thermometer 8 Alcohol, commercial, table of specific gravities and heating values 183 Alcohol, air required for com- bustion of 183 Alcohol, composition and heat- ing value of ( 181 Alcohol engine 390 vapor air mixtures .... 203 Alcohol, denatured 183 Alcohol vaporizer, Altman .... 198 Deutz 198 Dresden 201 Duerr 201 S\viderski-Longuemarre . . . 199 PAGE Alcohol, vaporizing devices for. . 196 Allis-Chalmers gas engine 345 Altman alcohol vaporizer 198 American Crossley suction pro- ducer 167 Atomizing or spraying carburet- ers 187 Atomizer, Hornsby-Akroyd .... 193 Attendance, cost of 565 Automatic cut-off engine, Jacob- son 271 Automobile engine, Continental 373 Franklin 372,374 Moore 374 Horch 373 horse-power rating of 483 Automobile gasoline engine .... 372 Auto-sparker 416 Auxiliaries and piping, cost of. . 548 Auxiliary spark gap 409 Barnett's engine 236 ignition cock 392 Barsanti and Matteucci free pis- ton engine 239 Batteries, primary and second- ary, method of connecting 420 Beau de Rochas 243 Beau de Rochas or Otto cycle, theoretical 65 Benzol 184 admixture of to alcohol . . . 184 Blast furnace gas, composition and heating value of .... 209 581 582 INDEX PAGE Blast furnace gas engines, fuel costs for 561 Blast furnace gas, preparation of 210 value of 560 Bomb calorimeter, Mahler's ... 129 Brayton cycle, theoretical 69 engine 248 engine diagram 250 oil engine, test of 250 Brake horse-power, definition of 38 Brake, Prony 39 British Thermal Unit 14 Brown, engine of 234 Bruce-Merriam-Abbott engine . 275 Buckeye two-cycle engine, gov- erning of 467 Buckeye two-cycle gas engine . . 286 Buffalo tandem engine 282 Buildings and floor-space, cost of 549 Calorie 14 Calorific intensity 144 Calorimeter, Carpenter's coal . . 130 Junker's gas 131 Mahler's bomb 129 Calorimetric thermometers 11 Campbell oil-engine, governor for 456 Carbon, heating value of 132 Carbon monoxide, heating value of 132 Carbureter, atomizing or spray- ing 187 Carbureter, bubbling, type of . . 185 Daimler 188 DeDion 190 Gautier .....' 191 Sintz 188 surface 186 Carbureters 185 Carnot cycle 62 Carnot or reversible engine 54 Cells, wet and dry 410 PAGE Centrifugal governors for hit-and- miss regulation 454 Charging of storage batteries . . 414 Classification of fuels 146 heat engines . . 17 internal combus- tion engines . 27 Classification of producers 158 Cleaning of blast furnace gas for use in engines 210 Clearances and compression pres- sures for various fuels, Otto cycle, table of 90 Clerk's engine 255 Clerk engine diagram 257 engines, tests on 257 Clerk's experiments on specific heat 224 Clerk-Lanchester starter 432 Closed cycle, definition of 61 Cockerill engines 336 Code of German Society of En- gineers, for testing gas producers and gas engines , 511 Coefficient, excess, definition of, 137 Coefficient of fly-wheel regula- tion 439 Coefficient of governor regulation 440 Coke oven gas, constants for. . . 208 Cold gas efficiency 151 Cooling water conditions affect- ing economy 530 Combining weights and volumes, for gases 127 Combination gasoline and alco- hol vaporizer 202 Combination producers 173 Combination producer, Crossley 173 Deutz double zone 173 Loomis-Pettibone 174 Combination systems of govern- ing 447 Combustion, air required for ... 136 Combustion line, Otto cycle ... 90 Combustion of alcohol, air re- quired 183 INDEX 583 PAGE Combustion, pressure, due to . . 220 products of 136 Comparative cost of power for various prime movers, 569-579 Comparison of theoretical and actual heat engines 61 Comparison of various theoret- ical cycles 73 Composition and heating value of alcohol 181 Composition and heating value blast furnace gas 209 Composition and heating value of gasoline 179 Composition and heating value of kerosene 179 Composition of oil gas 207 Composition of most common commercial gases, graph- ical representation 214 Composition of producer gas by volume per pound of car- bon gasified 152 Composition of producer gas by weight per pound of car- bon gasified 151 Compression, effect on economy 532 Compression pressures and tem- peratures for Otto cycle, table of 88 Compression pressures and clear- ances for various fuels, Otto cycle, table of 90 Compression stroke, Otto cycle 86 Compression, temperature due to, 220 ' Condensers in spark coils 404 , Conditions required for the prop- er formation of alcohol vapor-air mixtures 203 Constant pressure engines 30 cycle 104 Constant pressure, transfer of heat at 50 Constant temperature engines . '31 Constant volume or explosion engines 28 | PAGE Constant volume, transfer of heat at 50 Constants for acetylene 211 coke oven gas . . . 208 Constants, gas, table of 126 Constants for gas engine fuel gases, table of 213 Constants for illuminating gas . . 206 natural gas 212 water-gas 212 Conversion of solid fuels to gas 146 Coal calorimeter, Carpenter .... 130 Cost of attendance 565 Cost of erection 548 Cost of floor-space and build- ings 549 Cost of fuel for illuminating gas and natural gas engines 558 Cost of power for various prime movers 569-579 Cost of producers and engines . 547 oil and waste 564 operation 533 piping and auxiliaries . . 548 Cost of water for cooling and washing 561 Costs, fuel 553 operating 568 Crosby indicator 34 Crossley combination producer . 173 gas engine 324 Crossley hit-and-miss governor. 452 vaporizer 194 Grouse-Hinds distributor 425 double-ball timer. 408 Crude oil 178 Crude oil distillates 180 Crude oils, table of heating value and composition 178 Current, sources of 410 Cycle, Brayton, theoretical .... 69 Carnot . . : 62 choice of best 79 closed, definition of 61 constant pressure 104 definition of . 3 584 INDEX PAGE Cycle, Diesel, theoretical 71 Otto, combustion line 90 Otto, compression stroke . . 86 Otto, cyclic efficiency of ... 68 Otto, exhaust stroke 100 Otto, expansion line 97 Otto, pressure ratio in ... 92 Cycle, Otto, requirements for best efficiency in 100 Cycle, Otto, suction stroke .... 84 Cycle, Otto, table of allowable compression pressures and clearances for various fuels 90 Cycle, Otto, table of compres- sion pressures and tem- peratures 88 Cycle, Otto, table of cyclic effi- ciencies 69 Cycle, Otto, table of volumetric efficiencies and suction pressures 86 Cycle, Otto, typical lower loop diagrams 101 Cycle, Theoretical Beau de Ro- chas or Otto 65 Cycle, two-stroke . . 102 Cycles, theoretical, comparison . 73 Cyclic efficiency, definition of . . 62 Cyclic efficiency of theoretical Otto cycle 68 Cyclic efficiencies for Otto cycle, table of . 69 D Daimler carbureter 188 Dashboard spark coil 406 DeDion carbureter 190 Delamar hit-and-miss governor. 454 De La Vergne two-cycle oil en- gine 380 Denatured alcohol 183 Denaturizing agents for alcohol, table of 184 Depreciation of plant and build- ings 553 PAGE Design, general features of .... 263 Determination of excess coeffi- cient from exhaust gas- analyses 143 Deutz alcohol vaporizer 198 Deutz double-zone combination producer 173 Deutz gas engine 351 pressure producer 162 suction producer 165 Development of the Diesel en- gine 259 Development of the gas engine industry 258 Diagram, Clerk engine 257 Diagram, entropy, graphical construction of 120 Diagram, entropy, interpretation of 119 Diagram, entropy, mathematical construction of 112 Diagram of Lenoir engine 242 Diagram pressure volume, de- fined 3 Diagrams, Diesel engine 106 indicator, forms of 40 lower loop, Otto cycle 101 Otto cycle 94 Diesel cycle, theoretical 71 engine 385 Diesel engine diagrams 106 Diesel engine, development of . 259 governor for 459 Diesel engines, early, tests on. . 262 Dissociation, theory of 98 Distributor, Grouse-Hinds 425 high tension 423 Leavitt 426 Dresden alcohol vaporizer 201 Diirr alcohol vaporizer 201 Dynamos and magnetos 415 E Economist crude oil vaporizer . 195 INDEX 585 PAGE Efficiency, cyclic, definition of . . 62 Efficiency, cyclic, of theoretical Otto cycle 68 Efficiency, hot gas 151 cold gas 151 Efficiencies for Otto cycle, table of 69 Ehrhardt and Sehmer engine, governor for 461 Ejector muffler 428 Electrical thermometers 8 starters 436 Energy, kinetic 2 Engine, Barnett 236 Brayton 248 Brayton oil, test of 250 Brown 234 Carnot or reversible 54 Cayley's 26 Clerk 255 Clerk, tests on 257 Engine diagram, Clerk 257 Brayton 250 Engine diagrams, Diesel 106 Engine, Diesel, development of 259 furnace gas 25 Engine, free piston, Otto and Langen 244 Engine, gas, method of opera- tion 26 Engines, gunpowder 233 Engine, Hugon 242 Engine indicator 32 Engine, Lebon 234 Lenoir 239 Lenoir, diagram of 242 Engine, Lenoir, gas consump- tion of 242 Engine, Otto 251 Engines, Otto, tests on 254 Engine, Papin 232 Perry 238 Robert Street 233 Stirling hot-air 24 Wright :.. 235 Engine economy as affected by compression 532 PAGE Engine economy, as affected by cooling water conditions 530 Engine economy as affected by piston speed . 530 Engine economy as affected by variation in fuel mix- ture 533 Engine economy depending upon load 537 Engine economy depending upon point of ignition 534 Engines, constant pressure .... 30 constant temperature 31 Engines, constant volume or ex- plosion 28 Engines, Diesel, tests on 262 heat, classification of 17 Engines, heat, comparison of theoretical and actual . . 61 Engines, hot-air 22 Engines, internal combustion, classification of 27 Entropy 15, 107 Entropy diagram, graphical con- struction of 120 Entropy diagram, interpretation of 119 Entropy diagram, mathematical construction of 112 Entropy relations, general .... 109 Erection, cost, of 548 Ericsson hot-air engine 22 Excess coefficient, definition of . 137 Excess coefficient from exhaust analysis 143 Expansion line, Otto cycle 97 Experiments on specific heat by Clerk 224 Experiments on specific heat by Langen 220 Experiments on specific heat by Mallard and LeChatelier 220 Exhaust gas analysis, excess co- efficient from 143 Exhaust gas, computations on . 138 stroke, Otto cycle 100 586 INDEX PAGE Explosibility of fuel mix- tures 215,218 Explosion or constant volume engines 28 Explosion recorder, Mathot .... 500 Explosion, time of 227 Explosive mixture 215 Fairbanks engine 288 Fairbanks, Morse & Co., engine. 277 Fairbanks-Morse crude oil vapor- izer 389 Fairbanks-Morse suction pro- ducer 168 Fay and Bowen make-and -break igniter 400 Felten and Guilleaume electric starter 436 Fixation of tar-forming gases in producer gas 157 Flame propagation, velocity of . 227 Floor-space and buildings, cost of 549 Fly-wheel regulation, coefficient of 439 Fly-wheel weights, table of .... 440 Foos gasoline engine 358 Forms of indicator diagrams ... 40 Formula for mean effective pres- sure, Grover's 472 Four-cycle gas engine, Koerting . 280 Four-cycle gasoline engine, Stre- linger 362 Four-terminal spark coil 405 Franklin automobile engine, 372, 374 Free-piston engine, Barsanti and Matteucci 239 Free-piston engine, Otto and Langen 244 Fuel costs 553 Fuel costs for blast furnace gas engines 561 Fuel costs for gasoline engines. . 561 PAGE Fuel costs for illuminating gas and natural gas engines . . 558 Fuel costs for producer gas en- gines 559 Fuel costs for steam engines . . . 557 steam turbines . . 558 Fuel gases, constants for 213 Fuel mixture, computations on. 138 Fuel mixture, variation in, af- fecting economy 533 Fuel mixtures, explosibility of . . 215 Fuels, classification of 146 liquid, heating value of ... 134 solid, heating value of .... 135 Furnace gas engine 25 Fusion thermometers . 11 Gas and oil engines, tests of, Code of A.S.M.E 487 Gas calorimeter, Junker's 131 Gas, coke oven, constants for . . 208 Gas constants R for perfect gases, table of 48 Gas constants, table of 126 Gas consumption of Lenoir en- gine 242 Gas, blast furnace, composition and heating value of .... 209 Gas, illuminating, constants for, 206 natural, constants for 212 oil, composition of 207 oil, heating value of 208 perfect, specific heat of ... 12 Gas producers and engines, tests of 542 Gas producers in practice 156 Gas, waver, constants for 212 Gases, combining weights and volumes 127 Gases, commercial, composition of 214 Gases, fuel, constants for 213 Gases, illumination, table of composition 207 INDEX 587 PAGE Gases, perfect, characteristics of. 45 Gases, perfect, laws of 40 Gases, perfect, .table of specific volumes 47 Gases, specific heat of 48 Gas engine, Allis-Chalmers .... 345 Bruce-Merriam-Abbott .... 275 Buckeye two-cycle 286 Buffalo tandem 282 Crossley 324 Deutz 351 Diesel 385 Fairbanks, 288 Fairbanks, Morse & Co. . . . 277 furnace 25 Hautefeuille 232 Gas engines, Jacobson 267 Gas engine, Koerting four- cycle 280 Koerting two-cycle 313 Nurnberg 339 Gas engine, Niirnberg, table of standard sizes 346 Gas engine, Oechelhauser 355 Olds 292 Philadelphia Otto 291 Premier 347 Riverside 320 Sargent 310 Gas engines, Snow 329 Gas engine, Tod 306 Warren hit-and-miss 295 Westinghouse 266 Westinghouse horizontal. . . 304 Gas engine, Westinghouse verti- cal single-acting tandem, 306 Gas engine governing 439 Gas engine industry, develop- ment of 258 Gas engine regulation 439 tests, tables of . .544, 545 Gas engines and producers, cost of 547 Gas engines and gas producers, results of tests 542 Gas engines, Cockerill 336 PAGE Gas engines, blast furnace, fuel costs for 561 Gas engines, illuminating and natural, fuel costs for . . . 558 Gas engines, method of opera- tion 26 Gas engines, methods of testing 486 producer, fuel costs for . . . 559 small and. medium size .... '265 Gasification in pressure produc- ers, rate of 177 Gasoline, composition and heat- ing value of 179 Gasoline, mixing devices for ... 185 Gasoline engine, automobile .... 372 Foos 358 Lozier two-cycle 364 Olds 360 Standard marine 367 Strelinger four-cycle 362 Gasoline engines 358 fuel costs for 561 marine 361 Gautier carbureter 191 Generators, current, mechanical forms of 415 German Society of Engineers, Code for testing gas pro- ducers and gas engines 511 Gibbon kerosene vaporizer .... 193 Governor, Campbell oil engine . . 456 Crossley hit-and-miss 452 Delamare hit-and-miss .... 454 Diesel engine 459 Governor, Ehrhardt and Sehmer engine 461 Gov^nor, Hornsby-Akroyd en- gine 458 Governor, Koerting four-cycle engine 460 Governor, Nurnberg engine .... 458 Governor regulation, coefficient of 440 Governor, Robey 455 Springfield hit-and-miss . . . 453 .Governor, Westinghouse vertical engine . . 460 588 INDEX PAGE Governing Buckeye two-cycle engine 467 Governing of gas engines 439 Governing, hit-and-miss system 444 Governing of Koerting engine . . 468 Governing of Oechelhauser en- gine 467 Governing of two-cycle engines 449 Governing, Letombe system of . 465 Reinhardt's method. ...... 464 Reichenbach engine 463 Governing small two-cycle en- gine 467 Governing systems 441 Governing by varying time of ignition 449 Governing by varying the qual- ity of the fuel mixture . . 444 Governing by varying the quan- tity of fuel mixture 446 Governors, pendulum, for hit- and-miss regulation .... 450 Governors, centrifugal, for hit- and-miss regulation .... 454 Governors, mechanical details of, 450 Graphical construction of the entropy diagram . ...... 120 Graphical expressions for adia- batic and isothermal changes . . 55 Graphical representation of com- position of most common commercial gases 214 Grover's formula for mean effect- ive pressure 472 Giildner's method of determin- ing horse-power 477 Gunpowder engines 233 H Hammer break ignition 398 Hautefeuille, Abbe, gas engine of 232 Hay vaporizer 190 Heat at constant pressure, trans- fer of . 50 PAGE Heat at constant volume, trans- fer of 50 Heat balance 539 Heat, definition of 3 Heat engines, classification of . . 17 Heat engines, theoretical and actual, comparison of . . 61 Heat, mechanical equivalent of 15 relation of to entropy 54 Heat unit 14 Heating value and composition for true explosive mix- tures from liquid fuels, table of 217 Heating value and composition of alcohol 181 Heating value and composition of blast furnace gas .... 209 Heating value, definition of ... 129 Heating value of carbon 132 Heating value of carbon mon- oxide 132 Heating value of crude oils, table of 178 Heating value of gasoline 179 hydrogen .... 132 kerosene 179 liquid fuels . . 134 oil gas 208 solid fuels ... 135 Heating values and specific grav- ities of commercial alco- hol, table of 183 Heating values of hydrocarbons, table of * . 132 Heating values of true explosive mixtures, table of 216 High tension distributor 423 jump-spark system 423 Hit-and-miss engine, Jacobson . . 268 Warren 295 Hit-and-miss governor, Crossley 452 Delamare 454 Springfield 453 Hit-and-miss system of govern- ing , 442 INDEX 589 PAGE Horch automobile engines 373 Horizontal gas engine, Westing- house 304 Horsnby-Akroyd atomizer 193 Horrisby-Akroyd engine, gov- ernor for 458 Hornsby-Akroyd oil engine .... 376 Horse-power, brake, definition of 38 Horse-power defined 1 Horse-power, determination from mean effective pressure. . 472 Horse-power from standard air reference diagram 476 Horse-power, Giildner's method 477 Horse-power, indicated, defini- tion of 38 Horse-power rating of automo- bile engines 483 Hot-air engine, Ericsson 22 Stirling 24 Hot-air-engines 22 Hot gas efficiency 1.51 Hot tube igniter, Koerting .... 395 Newton's 238 Hot tube ignition 394 with timing valve .... 395 Hugon engine 242 Hydrocarbons, table of heating values of 132 Hydrogen, lower and higher heating value of 132 Hyperbola, methods of drawing 43 Ignition 392 Ignition by electric spark 397 heat of compression 396 hot tube 394 open flame 392 Ignition cock, Harriett's 392 Ignition, hammer break 398 jump-spark 401 make-and-break ....:.... 397 Ignition, variation in, affecting economy 534 PAGE Igniter, Fay and Bowen make- and-break 400 Igniter, hot tube, Newton's .... 238 Koerting, hot tube. 395 Koerting open flame 393 Illuminating gas, constants for. 206 Illuminating gases, table of com- position of 207 Index for expansion and com- pression lines, method of finding 115 Indicated horse-power, defini- tion of 38 Indicators, engine 32 Indicator, Crosby .- 34 optical 35 Tabor 34 Thompson 33 Indicator diagrams, forms of . . . 40 Inertia governors for hit-and- miss regulation 450 Internal combustion engines, classification of 27 Isothermal line, equation of. ... 52 Isothermal and adiabatic changes, graphical ex- pressions for 55 Isothermal change 16 Isothermal expansion, work per- formed in . 53 Jacobson automatic cut-off en- gine 271 Jacobson gas engines 267 hit-and-miss engine . . . 268 throttling engine .... 272 Jump-spark ignition 401 Jump-spark and make-and-break systems compared 410 Jump-spark system, high ten- sion 423 Junker's gas calorimeter 131 590 INDEX K PAGE Kerosene, composition and heat- ing value of 179 Kerosene vaporizer, Gibbon . . . 193 Kinetic energy 2 Koerting engine, governing of . 468 four-cycle gas engine . 280 Koerting four-cycle engine, gov- ernor for 460 Koerting hot tube igniter 395 open flame igniter . . . 393 suction producer 166 pressure producer .... 161 Koerting suction producer for peat 172 Koerting two-cycle engine 313 Lacoste timer 407 Langen's experiments on specific heat 220 Laws of perfect gases . . 46 Leavitt distributor 426 Lencauchez double-zone suction producer 170 Lencauchez suction producer . . 170 Lenoir engine 239 diagram of 242 gas consumption of 242 Letombe system of governing. . 465 Lebon's engine . . 234 Limits of explosibility for fuel mixtures made from dif- ferent fuels 218 Limit of piston speed 471 Liquid fuel engines 358 Liquid fuels, heating value of . . . 134 Liquid fuels, mixing devices for. 185 Loomis-Pettibone combination producer 174 Lowe system for making oil gas. 196 Low tension system of wiring. . 421 Lower and higher heating value of hydrogen 132 PAGE Lower loop diagrams, Otto cycle 101 Lozier two-cycle marine gasoline engine 364 Lunkenheimer mixing valves . . 187 M Magneto, action of 416 Magnetos and dynamos 415 Magnetos, systems of wiring em- ploying 422 Mahler's bomb calorimeter .... 129 Maintenance and repairs 567 Make-and -break and jump-spark systems compared 410 Make-and-break ignition 397 Mallard and LeChatelier's ex- periments on specific heat 220 Manograph 35 Marine gasoline engine, Lozier two-cycle 364 Marine gasoline engine, Standard 367 Marine gasoline engines 361 Mathematical construction of the entropy diagram 112 Mathot explosion recorder 500 Matteucci and Barsanti free pis- ton engine 239 Mean effective pressure, Grover's formula 472 Mean effective pressure, tables for determining 473 Mechanical details of governors 450 equivalent of heat . 15 forms of generators 415 Method of connecting up pri- mary and secondary bat- teries ^120 Method of drawing hyperbola . . 43 Method of finding index for ex- pansion arid compression lines 115 Method of governing, Rein- hardt's 464 Method of operating gas en- gines 20 INDEX 591 PAGE Method of test, Code of 1901, brake horse-power 497 Methods of test, Code of 1901, calibration of instru- ments 488 Methods of test, Code of 1901, computation of temper- atures 503 Methods of test, Code of 1901, duration of tests 492 Methods of test, Code of 1901, heat balance 502 Methods of test, Code of 1901, heat units consumed . . . 493 Methods of test, Code of 1901, indicated horse-power. . .. 494 Methods of test, Code of 1901, indicator diagrams 501 Methods of test, Code of 1901, measurement of fuel .... 493 Methods of test, Code of 1901, measurement of jacket water , 494 Methods of test, Code of 1901, speed determination .... 498 Methods of test, Code of 1901, standards of economy. . . 501 Methods of test, Code of 1901, starting and stopping tests 493 Mietz and Weiss oil engine 383 Mixing devices for gasoline .... 185 liquid fuels . 185 Mixing valves, Lunkenheimer . . 187 Mond pressure producer 163 Mond process for making pro- ducer gas 163 Moore automobile engine 374 Morgan pressure producer 160 Muffler, ejector 428 Powell 428 Mufflers . , .426 X Natural gas, constants for 212 PAGE Newton's hot tube igniter 238 Non-trembler spark coil 402 Nurnberg engine 339 governor for 458 Nurnberg gas engine, table of standard sizes 346 O Oechelhauser engine, governing of 467 Oechelhauser gas engine 355 Oil and gas engines, tests, of, Code of A.S.M.E. 488 Oil and waste, costs of 564 Oil, crude, distillates 180 crude 178 Oils, crude, table of heating value and composition . . 178 Oil engine, Brayton, test of. ... 250 De La Vergne two-cycle ... 380 Hornsby-Akroyd 376 Mietz and Weiss 383 Priestman 388 Oil engines 375 Oil gas, composition of 207 heating value of 208 Lowe system for making . . 196 Olds gas engine 292 Olds gasoline engine 360 Open flame igniter, Koerting . . 393 ignition 392 Operating costs 568 Operation, cost of 533 Optical indicator 35 pyrometers 9 Otto or Beau de Rochas cycle, theoretical 65 Otto cycle, combustion line ... 90 compression stroke 86 diagrams, typical 94 exhaust stroke 100 expansion line 97 pressure ratio in 92 592 INDEX PAGE Otto cycle, requirements for best efficiency in 100 Otto cycle, suction stroke 84 Otto cycle, table of allowable compression pressures and clearances for various fuels 90 Otto cycle, table of compression pressures and tempera- tures 88 Otto cycle, table of cyclic effi- ciencies 69 Otto cycle, table of volumetric efficiencies and suction pressures 86 Otto cycle, theoretical, cyclic efficiency of 68 Otto cycle, two-stroke 102 Otto cycle, typical lower loop diagrams 101 Otto engine 251 engines, early, tests on .... 254 Otto and Langen free piston en- gine 244 Otto-Langen free piston engine, table of tests : . . . 247 Otto-Langen free piston engine, typical diagram 248 Otto gas engine, Philadelphia . . 291 Papin's engine 232 Pendulum or inertia governors for hit-and-miss regula- tion 450 Perfect gases, characteristics of 45 laws of . . . . 46 Perfect gas, specific heat of .... 12 Perfect gases, table of specific volumes 47 Perry's engine 238 Petreano surface carbureter. . . . 187 Philadelphia Otto engine 291 Piping and auxiliaries, cost of . 54 S PAGE Piston speed, effect on economy, 530 limit of 471 Pittsfield timer '407 Poetter pressure producer 162 Powell muffler 428 Power of gas engines 471 Premier gas engine 347 Pressure after combustion 220 producers 159 producer capacities ... 176 Pressure producers, rate of gasi- fication 177 Pressure producer, Deutz 162 Koerting 161 Mond 163 Morgan 160 Poetter 162 Taylor 159 Wile 160 Pressure ratio in the Otto cycle 92 volume diagram defined 3 Priestman oil engine 388 vaporizer 194 Prime movers, comparative cost of power for 569-579 Producers and gas engines, cost of 547 Producers, classification of .... 158 combination 173 Producer, combination, Crossley 173 Producer, combination, Deutz double-zone 173 Producer, combination, Loomis- Pettibone 174 Producer details 175 gas : 149 Producer gas, composition of by volume per pound of car- bon gasified 152 Producer gas, determination of weight and volume per pound of carbon 151 Producer gas engines, fuel costs for 559 Producer gas, fixation of ....... 157 gas, Mond process ... 163 593 PAGE Producer plants, tests of 546 Producers, pressure 159 Producer, pressure, Deutz 162 Koerting 161 Mond 163 Morgan 160 Poetter 162 Taylor 159 Wile 160 Producers, suction 165 Producer, suction, American Crossley 167 Producer, suction, Deutz 165 Fairbanks-Morse 168 Koerting . 166 Koerting for peat 172 Lencauchez 170 Riche 169 Producer yield, theoretical .... 153 Production of air gas 147 water gas 147 Products of combustion 136 Prony brake 39 Pyrometers 7 optical 9 Quality governing 444 Quantity governing 446 R Rate of gasification in pressure producers 177 Rate of work 1 Rating of storage batteries 413 Regulation, coefficient of fly- wheel 439 Regulation, governor 440 Regulation of gas engines 439 Reichenbach engine, governing of 463 Reinhardt's method of governing 464 Relation of heat to entropy ... 54 PAGE Repairs and maintenance 567 Requirements for best efficiency in Otto cycle 100 Requirements for proper scav- enging of cylinder, two- cycle engine 103 Reversible engine, Carnot 54 Riche suction producer 169 Riverside gas engine 320 Robey governor 455 S Sargent gas engine 310 Scavenging of cylinder, two- cycle engine, require- ments for 103 Second law of thermodynamics 55 Sintz carbureter 188 timer 406 Snow gas engines 329 Solid fuels, heating value of. ... 135 Sources of current 410 Spark coil, action of 402 dashboard 406 four-terminal 405 three-terminal 405 Non-trembler 402 trembler 403 Spark gap, auxiliary 409 coils, condensers in' 404 coils, types of 405 plugs 4C6 Specific gravities and heating values of commercial al- cohol, table of 183 Specific heat at constant pres- sure 13 Specific heat at constant volume 13 Specific heat experiments by Clerk 224 Specific heat experiments by L:inen 220 Specific heat experiments by Mallard and LeChatelier. 220 594 INDEX PAGE Specific heat, definition of 12 Specific heat of perfect gas .... 12, 48 Specific heat variation with temperature 220 Specific heats, table of 13 Specific volumes of perfect gases, table of 47 Spraying or atomizing carburet- ers 187 Springfield hit-and-miss gover- nor 453 Standard marine gasoline engine 367 Starter, Clerk-Lanchester 432 Felten arid Guilleaume .... 436 Starting apparatus 428 Starting by auxiliary source of power 431 Starting by compressed air . . . 433 electricity 436 fuel mixture 431 means of crank .... 429 Steam engines, fuel costs for . . . .557 turbines, fuel costs for . . 558 Stirling hot-air engine 24 Storage batteries or accumu- lators 411 Storage batteries, charging of . . 414 rating of 413 testing of 414 Stratification, theory of ....... 97 Strelinger four-cycle gasoline en- gine 362 Street, Robert, engine of 233 Suction plants, volume of scrub- ber 176 Suction pressures and volumetric efficiencies for Otto cycle, table of 86 Suction producers 165 Suction producer, American Crossley 167 Suction producer, Deutz 165 Fairbanks-Morse 168 Lencauchez 170 Suction producer, Lencauchez, double-zone . .170 PAGE Suction producer, Koerting .... 166 Koerting, for peat 172 Riche 169 Suction stroke of Otto cycle ... 84 Surface carbureter 186 Swiderski-Longuemarre alcohol vaporizer 199 System of governing, Letombe. 465 System of wiring employing magnetos 422 System of wiring, low tension. . 421 Systems of governing 441 combination 447 System of governing, hit-and- miss 442 Systems of wiring 420 Tables for determining mean effective pressure 473 Table of allowable compression pressures and clearances for various fuels, Otto cycle 90 Table of combining weights and volumes for gases 128 Table of composition and heating value for true explosive mixtures from liquid fuels 217 Table of composition of typical illuminating gases 207 Table of compression pressures and temperatures for Otto cycle 88 Table of constants for gas-engine fuel gases 213 Table of cyclic efficiencies for Otto cycle 69 Table of denaturizing agents for alcohol 184 Table of engine tests 544, 545 Table of gas constants R for per- fect gases 48 Table of heating value and com- position of crude oils. ... 178 INDEX 595 PAGE Table oT heating values of hydro- carbons 132 Table of heating values of true explosive mixtures 216 Table of relative fly-wheel weights 440 Table of specific gravities and heating values of com- mercial alcohol 183 Table of specific heats 13 Table of specific heats for per- fect gases 48 Table of specific volumes of per- fect gases 47 Table of standard sizes Niirn- berg gas engine ' 346 Table of tests on producer plants, 546 thermometric scales . . 5 Table of true explosive mix- tures, for commercial gases 216 Table of vapor tension for alco- hol and water 204 Table of various gas constants. . 126 Table of volumetric efficiencies and suction pressures for Otto cycle 86 Tabor indicator 34 Tandem .gas engine, Buffalo . . . 282 Westinghouse 306 Tar-forming gases, fixation of. . 157 Taylor pressure producer 159 Temperature, absolute 6 Temperature after combustion . 220 Temperature, definition of 3 Temperatures and compression pressures for Otto cycle, table of 88 Temperature scales 5 Test on Brayton oil engine 250 Tests of Clerk engines 257 Tests of engines and gas produc- ers, results of 542 Tests of gas engines 544, 545 on early Diesel engines . . 262 on early Otto engines .... 254 PAGE Tests of gas and oil engines, Code of A.S.M.E. for 1901 .... 487 Tests on producer plants, table. 546 Testing of gas engines, methods for ^186 Testing of storage batteries .... 414 Theoretical and actual heat en- engines, comparison of . 61 Theoretical Brayton cycle 69 cycles, comparison of 73 Diesel, cycle 71 Theoretical Otto cycle, cyclic efficiency of 68 Theoretical Otto or Beau de Rochas cycle 65 Theoretical yield of producer . . 153 Theory of dissociation 98 stratification 97 Thermal unit, British 14 Thermo dynamics, second law of 55 Thermo-element 8 Thermometer, air 8 Thermometers, : . . 6 calorimetric 11 electrical 8 fusion 11 vapor 11 Thermometric scales, table of . . 5 Thermopyle 8 Thompson indicator 33 Three-port two-cycle engine .... 366 Three-terminal spark coil 405 Throttling engine, Jacobson . . . 272 engines, Warren .... 295 Time of explosion 227 Timer, Grouse-Hinds double-ball 408 Lacoste 407 Pittsfield 407 Sintz "... 406 Timing valve 395 Timers . 405 Tod gas engine 306 Total operating costs 568 Transfer of heat at constant pres- sure . 50 596 INDEX PAGE Transfer of heat at constant volume 50 Trembler spark coil 403 True explosive mixture 215 True explosive mixtures, for commercial gases, table of 216 Two-cycle engine, governing of 449, 467 Koerting 313 three-port 366 Two-cycle gas-engine, Buckeye 286 Two-cycle marine gasoline en- gine, Lozier 364 Two-cycle oil engine, De La Vergne 380 Two-stroke Otto cycle 102 Typical computations on fuel mixtures and exhaust gases. . 138 Typical diagram, Otto-Langen free piston engine 248 Typical Diesel engine diagrams. 106 Typical lower loop diagrams, Otto cycle 101 Typical Otto cycle diagrams ... 94 r Unit, British thermal 14 Value of blast furnace gas 560 Vapor tension for alcohol and water, table of 204 Vapor thermometers 11 Vaporizer, alcohol, Diirr 201 Vaporizer, alcohol, Swiderski- Longuemarre 199 Vaporizer, Altman alcohol 198 Vaporizer, combination alcohol and gasoline 202 PAGE Vaporizer, Crossley 194 crude oil, Economist 195 Vaporizer, crude oil, Fairbanks- Morse 389 Vaporizer, Deutz alcohol 198 Dresden alcohol 201 Gibbon kerosene 193 Priestman 194 W. Hay 190 Vaporizing devices for alcohol . . 196 Vaporizing devices for crude oil and kerosene 192 Variation of fuel consumption with load 554 Variation of specific heat with temperature 220 Velocity of flame propagation. . 227 Vertical gas engine, Westing- house 306 Volume of scrubber in suction plants 176 Volumetric efficiencies and suc- tion pressures for Otto cycle, table of 86 W Warren hit-and-miss engine . . . 295 throttling engines 295 Waste and oil, cost of 564 Water gas, constants for 212 production of 147 Wiring, low tension system of . . 421 Weights, fly-wheel, table of. ... 440 Westinghouse gas engine 266 Westinghouse horizontal gas en- gine 304 W^estinghouse vertical engine, governor for 460 Westinghouse vertical single-act- ing tandem gas engine . . 306 Wet and dry cells 410 Wile pressure producer 160 Wiring employing magnetos, sys- tems of . . . 422 INDEX 597 PAGE Wiring, systems of 420 Work performed in adiabatic ex- pansion 53 Work performed in isothermal expansion 53 Work, rate of 1 PAGE Wright's engine 235 Yield of producer, theoretical . . 153 UNIVERSITY OF CALIFORNIA LIBRARY BERKELEY Return to desk from which borrowed. This book is DUE on the last date stamped below. REC'D LO NOV 19 1957 REC'D LD MAY 2 9 1962 LD 21-100m-7,'52(A2528sl6)476 TJ-7TS 3, UNIVERSITY OF CALIFORNIA LIBRARY