This booic is dnc* ^^-^'^^^laat date PR -9 1925 Digitized by the Internet Archive in 2007 with funding from Microsoft Corporation http://www.archive.org/details/bearingstheirlubOOalforich BEARINGS AND THEIR LUBRICATION Bearings and Their Lubrication By L. P. Alford, M. E. EDITOR, AMERICAN MACHINIST; MEMBER, AMERICAN SOCIETY OF MECHANICAL ENGINEERS; MEMBER, AMERICAN SOCIETY FOR. TESTING MATERIALS. ALLEGE U. OFC. LIBRARY 19 1 1 Published by the American Machinist McGraw-Hill Book Company, Sole Selling Agents 239 West Thirty-ninth Street, New York London, E. C, 6 Bouverie Street Berlin, N. W. 7. Unter der Linden 71 Copyright, 191 i, by the American Machinist 7^t-^V£.wv- American Machinist, New York, U.S. A. PREFACE The aim of this book has been two fold; to present the underlying principles involved in the design of all classes of machinery bearings and to show modern practice in the construction and application of important commercial types. It is believed to be the first treatment of these subjects ever put into book form. The tendency of modern machine design in many fields is toward the use of high speed shafts and spindles and friction reducing forms of bearings. There- fore the present seems a peculiarly fitting time to present these data vdth the purpose of making them permanently useful to engineers, designers, draftsmen and machinists; in fact to any one who is interested in any way with machinery bearings and their proper lubrication and care. Many of the data have never before been published in any form. Many of the experiments and researches quoted are from European sources and official transactions of technical societies and are not generally known to American readers. It is a pleasure to acknowledge the many courtesies received from engineers and machinery manufacturers and the valuable information furnished by them. Without this hearty cooperation it would have been impossible to have produced this work. L. P. Alford. New York City. August, 191 1. CONTENTS Classification of Machinery Bearings PART I BEARINGS WITH SLIDING CONTACT Section. I. Sliding Friction 5 II. Coefficients OF Friction OF Journal, Collar, Step AND Guide Bearings . . 31 III. Materials for Bearings 41 IV. Allowable Pressures, Speeds and Temperatures 65 V. Design of Journal Bearings 86 VI. Lubricants m VII. Design of Flat Sliding Surfaces and Special Bearings 132 VIII. Three Important Bearing Inventions 139 IX. Typical Designs and Constructions 141 X. Hints on the Care of Bearings 166 PART II BEARINGS WITH ROLLING CONTACT I. Rolling Friction and Factors of Design 173 II. Construction of Ball Bearings 180 III. Typical Designs and Mountings for Ball Bearings 195 IV. Lubrication of Ball Bearings 204 V. Roller Bearings with Flexible Rollers 209 VI. Radial Roller Bearings with Solid Rollers 217 VII. Roller Thrust Bearings 221 vu BEARINGS AND THEIR LUBRICATION CLASSIFICATION OF MACHINERY BEARINGS Machinery bearings are the parts of the bed, frame or other members that constrain rotating parts, such as shafts and spindles. They are divided into two general classes, journal and thrust, depending upon the direction in which the load acts. In the journal bearing the load acts at right angles to the axis; in thrust bearings, parallel to the axis. The bearing surface of all journal bearings is necessarily circular in cross- section. In profile it is ordinarily cylindrical, but may be conical, spherical or even of a more complicated shape. The bearing surface of thrust bearings is ordinarily flat, but may be spherical, conical, or shaped to the curve of the tractrix. Bearings are also divided into two classes by the kind of contact between the surfaces; that is, bearings having sliding contact, or ordinary bearings, and bear- ings having rolling contact or ball and roller bearings. The bearing surfaces of ball bearings both radial (journal) and thrust are ordinarily curved races with a circular cross-section. Less commonly the sur- faces are flat. The bearing surfaces of journal roller and thrust bearings are circular in cross-section and either cylindrical or conical in profile. The accompanying table gives a classification of machinery bearings. Flat constraining surfaces for sliding machine parts having rectilinear move- ments have no common name, but are referred to as ways, guides, and the like. The relative motion of all these constraining members and the members constrained is resisted at their surfaces of contact by a force which is called the force of friction. Thus, all kinds of bearings must be designed uath particular reference to minimizing this force, and our starting-point must be a study of friction. BEARINGS AND THEIR LUBRICATION Bearing Surfaces Cylindrical. Journal. . . \ Conical. Spherical. Bearings with sliding contact. Thrust Flat. Conical. Spherical. Generated by curve of tractrix. Bearings with rolling contact. Ball bearings. Roller bearings. f Radial. Thrust. Journal. . Thrust , Spherical (of balls) with double curved surfaces (of races). Spherical (of balls) with conical (of races) . Spherical (of balls) with flat (of seats). Spherical (of balls) with double curved (of seats). Cylindrical with cylindrical. Conical with conical. Cylindrical with flat. Conical with conical. The cross-section, perpendicular to the axis of rotation, of all these bearing surfaces is necessarily circular. PARTI BEARINGS WITH SLIDING CONTACT SECTION I FRICTION OF BEARINGS WITH SLIDING CONTACT All sliding surfaces, no matter how carefully they have been prepared, are known to consist of minute humps and hollows; this is true even of the smooth- est surfaces that can be made although, of course, the height of the depressions and hollows varies with different kinds of materials and different degrees of finish. This is similar to, though much less in degree than the ''nap " of woolen cloth and the "pile " of velvet. It may also be likened in an exaggerated degree to the "fur" on a rough planed or milled surface of cast iron. When two solid surfaces are held in contact by an appreciable force, these minute parts of the surfaces interlock and resist relative motion. This resist- ing force acts tangentially to the surface of separation, and is the force of friction. In addition to this interlocking action of surfaces, there is another that takes place between surfaces that are very carefully fitted together and are in intimate contact. This action is called adhesion and still further resists relative motion. One of the best examples of this adhesion of accurately finished surfaces is the action of the Swedish gages that are now familiar to most engineers and machinists. The working surfaces of these gages have a splendid lapped finish. When two of these surfaces are rubbed into intimate contact they will adhere with a force having an intensity several times that of atmospheric pressure; that is, the force of adhesion between the surfaces tending to prevent separation by a direct pull is much greater than the force exerted by the pressure of the atmos- phere upon the exposed surfaces. Experiments witnessed by the author showed that the force required to slide one of the gages upon another was about 21/2 lb. when the two were in intimate contact. The surfaces were i 3/8 in. long by 3/8 in. wide, representing an area of about 1/2 sq. in. The upper gage was 1/2 in. thick, thus giving a weight of about i oz. This force of adhesion is but little understood, but this simple experiment shows how much it can be quantitatively between accurate surfaces. Just as the intimate contact of the surfaces tends to increase the resistance of the relative motion, a separation of these surfaces by oil, grease or other lubri- cant reduces this resistance to a considerable extent. This fact will lead us later on to a consideration of the lubrication of bearings. In some machine elements, as clutches and brakes, this resistance to relative 5 6 BEARINGS AND THEIR LUBRICATION motion is a desiraoie thing; but in bearings, on the contrary, it is undesirable as it requires power to overcome it and, if excessive, may cause the surfaces to cut and ruin the bearings. Thus a knowledge of the laws of friction is abso- lutely essential to the design of successful bearings. Unfortunately, these laws are not thoroughly understood, and it is now recognized that many of the older and accepted principles only hold true for a very limited range of conditions. FRICTION OF REST AND MOTION Experiment has shown that it requires a greater force to start one surface to sliding over another than to continue this sliding after motion has begun. Therefore, two forces of friction are recognized; the friction of rest (sometimes called stiction) and the friction of motion. A knowledge of the laws of the friction of motion is of more importance and for that reason the friction of rest will be treated briefly. Again, the laws of friction naturally divide themselves into two groups referred to the condition of the surfaces in contact; (a) for dry or unlubricated surfaces, (b) for lubricated surfaces. COEFFICIENT OF FRICTION In a preceding paragraph it is stated that the force of friction is the re- sisting force to relative motion when two solid surfaces are held in contact by an appreciable force. The ratio of the force required to produce motion to the force holding the surfaces in contact is called the coefficient of friction. For flat surfaces, if we let F equal the frictional resistance, \V the normal load, and / the coefficient of friction, then F=fW, or /= -. For flat surfaces where the top body is acted upon by its own weight only, the coefficient of friction is equal to the tangent of the angle of repose; that is, the angle of inclination with the horizontal of the plane of contact of the two surfaces where sliding will begin to take place. If this angle is indicated by a, the coefficient of friction equals the tangent of angle a. Or, /=tan. a. In a similar manner the coefficient of friction for cylindrical bearings is the ratio of the frictional resistance to the normalpressure and is again indicated by /. The character and fit of the surfaces, however, have a great deal to do with the value of the coefficient for such cases; the distribution of the normal pres- sure is variable, and the intensity of the normal pressure is a difficult quantity to SLIDING FRICTION 7 determine accurately. For convenience, it is customary to take as the normal pressure the intensity of pressure per unit of projected area of the bearing. Thus, if d is the diameter of the shaft and / its length, the intensity of pressure per unit of projected area is W 1)0= ---. dl The reason for the difficulty in determining accurately the intensity of normal pressure is evident to anyone familiar with bearings. It is perfectly possible to fit a journal into its bearing so that there shall be a complete arc of contact for 1 80 degrees of bearing circumference. In practice, however, a journal is made free in its bearing, the amount of freeing depending upon circumstances. Thus the arc of contact is decreased. In many cases this arc may not be over one- half of the total bearing depth of the journal, or say about 120 degrees. Again, journals and bearings tend to wear in use and thus produce a varying amount of contact depending upon the amount and location of the wear that has taken place. WORK OF FRICTION The work of friction Is an important point to consider in the design of bear- ings, for it gives a measure of the amount of energy transformed into heat by the frictional resistance. This quantity of heat has a direct bearing upon the per- formance and life of many types of bearings. In some cases, a circulation of water or oil is established to conduct away the heat thus liberated. For flat plates H the foot pounds of energy absorbed per minute equals / W V, where v is the velocity in feet per minute and W is expressed in pounds. For circular surfaces, if n is the number of revolutions per minute frrdnW H=- =o.26iSfdnW. ■ 12 Table i gives the moment of friction in inch pounds and the energy lost in friction in foot pounds per minute for the ordinary types of machinery bear- ing surfaces. The notation is as follows: / = Coefficient of friction. W =load on journals or pivots in pounds. r = radius in inches. r^ = inner radius of collar in inches. ^2 = outer radius of collar in inches. d = diameter in inches. n = number of revolutions per minute. b = half of the included angle of cone. BEARINGS AND THEIR LUBRICATION Kind of bearing surfaces Moment of friction in inch pounds Energy lost in friction in foot pounds per minute Journal bearings as ordinarily fitted 0.5/^1^ 0. 2618 fdWn. Journal bearings tightly fitted . . 0.62s fdW 0. 3:^ 24 fdWn. Journal bearings fitted to grasp the shaft and give a uniform pressure throughout. 0.78/^t^ 0.4112 fdlVn. Flat pivot bearings 0.66 frW . T,4g fWrn. Flat collar bearings "-KvIt;: 0.349/^l^w^; Conical pivot bearings 0.66 fWr cosec. b . T,4C) fWrn cosec. b. Conical journal bearings or journal bearings. tapered 0. 66 fWr sec. b 0.349/I'Fm sec. b. Truncated cone pivot bearings J, 3_^ 3 o.66fW ' ' ^2 sin. b m/ '"-'^^ 0.349/PFw . ^2 sm. b Hemispherical pivot bearings. . . frW o.26i8fdWn. (tractrix frW 0. 2618 fdWn. curve). Table i. — Moment of Friction and Work of Friction for Bearings with Sliding Surfaces. The mathematical work showing the derivation of these formulas of Table i can be found in Thurston's Friction and Lost Work, beginning on page 40. Attention should be directed to the three items in Table i dealing with journal bearings. The first is for bearings as ordinarily fitted; the second with bearings tightly fitted and shows that the loss of energy in friction is 1.27 times that of properly fitted bearings. The third deals with bearings that have been so fitted, or have so worn that the intensity of pressure is uniform throughout. Such a condition produces a loss of power 1.57 times that of properly fitted bearings. POWER LOST IN BEARINGS To obtain the horsepower lost in bearings, or the amount of power required to drive a journal or pivot against its own friction divide the quantities obtained SLIDING FRICTION 9 from the formulas in the last column of Table i by 33,000. Similarly, to obtain the number of British thermal units of heat liberated in a given bearing, divide the quantities obtained from the formulas in the last column of Table i by 778. FRICTION OF UNLUBRICATED SURFACES The early experimental work to determine the values of the coefficient of friction for various surfaces in contact v^as done by Coulomb, Morin and Rennie. The work of General Morin, a most original and scientific engineer, was the most extensive. From his experiments he deduced three laws, which are 1. The frictional resistance is proportional to the pressure. 2. The frictional resistance is independent of the speed. 3. The frictional resistance is independent of the extent of the surfaces.^ These laws have frequently been quoted and treated as if they were rigidly true. Such, however, is not the case. They are now known to be only ap- proximately true for a very limited range of conditions. In fact, General Morin himself considered them only as approximations as shown by this para- graph from a letter written by him to the secretary of the Institute of Mechanical Engineers, England, and printed in the proceedings of that Institution for the year of 1883, at page 663. "The results furnished by my experiments as to the relations between pressure, surface, and speed, on the one hand, and sliding friction on the other, have always been regarded by myself, not as mathematical laws, but as the close approximations to the truth vdthin the limits of the data of the experiments themselves. The same holds, in my opinion, for many other laws of practical mechanics; such as those of rolling resistance, etc." Table 2 is translated from the work of General Morin and gives the results of some of his experiments on plain surfaces. The values are for static friction, or friction at very low velocity, with light pressures. The surfaces were dry, or but very slightly lubricated. In Table 2 it will be noticed that the values in the last column are consider- ably greater than those in the first. This shows us the effect of leaving compara- tively soft bodies in contact for a considerable length of time. The irregular particles of the surfaces tend to compress and interlock to a greater degree than if the surfaces are merely brought in contact and motion at once takes place. The experiments of Rennie in 1829 were made in general with dry and un- lubricated surfaces. The principles that he deduced were that the friction of sliding surfaces differs with the character of the surfaces; that the friction of lubricated surfaces depends upon the lubricant, rather than upon the bodies themselves; that friction is least with hard materials and greatest with soft ones; ^See Annales des Mines, 3 serie, October, 1836, p. 27. lO BEARINGS AND THEIR LUBRICATION Mutual arrange- ment of the fibers Coefficient of friction Surfaces in contact When the body is in motion When the surfaces have been some time in contact Oak on oak, dry Oak on oak, dry Oak on oak, wet Elm on oak, dry Elm on oak, dry Ash on oak, dry Fir on oak, dry Beech on oak, dry Wild pear tree on oak, dry Wrought iron on oak, dry Yellow copper on oak, dry Black curried leather on oak, dry. . . Cowhide sole leather laid flat on oak, dry. Sole leather on edge on oak, dry. . . . Sole leather on edge on oak, wet. . . . Mat of small hempen cords on oak, dry. Parallel Perpendicular . . Perpendicular . . . Parallel Perpendicular . . . Parallel Parallel Parallel Parallel Parallel Parallel Parallel Parallel Parallel Parallel Parallel G.48 G.32 G.25 0.43 0.45 G.40 0.36 0.36 0.4G G.62 G.62 G.27 G.52 0.34 G. 29 0.32 o.6Ga G.65 0-54 G.71 0.69 0-57 0.50 G.52 0-53 0.44 G.62 G.62 0.74 G.61 0-43 G.79 0.50 Table 2. — Coefficients of Friction for Various Surfaces in Contact {Morin). that the limit of abrasion of two rubbing surfaces is determined by the hardness of the softer, and that with wood and metals friction varies with the pressure and is independent of the extent of the surface, time of contact and velocity. Table 3 is from Rennie's experiments. Pressure in pounds per square inch Brass on cast iron Wrought iron on wrought iron Wrought iron on cast iron Steel on cast iron 187 0.23 G.22 0.21 G.21 G.23 G.23 G.23 0.25 0.27 0.31 0.38 0.41 abraded abraded G.28 G.29 0-33 0-37 0-37 0.38 abraded 0.30 0-33 0-35 0.35 G.36 224 336 448 Kdo . 672 784 abraded Table 3. — CoEFFfciENXs of Friction of Unlubricated Surfaces {Rmnie). SLIDING FRICTION II Tables 4 to 7, inclusive, are also from General Morin's experiments and were made at low velocities, with surface either dry, oily, or greasy, as stated in the tables. In the original table of General Morin's from which Table 5 has been translated, there is a column giving the position of the fibres of the wooden test pieces with reference to the direction of motion, and in some cases with refer- ence to the supposed direction of the fibres of the metal test pieces. How- ever, the references are not entirely clear and have been omitted in the trans- lations; apparently all of the fibres of the wooden test pieces ran parallel with the direction of motion and parallel with the supposed direction of the fibers of the metal pieces. Similarly a column has been dropped in translating Table 6, with the belief that the values are not rendered less useful thereby. Inconsistencies that may be noted in the values of the coefficients of friction may be explained by variations in the conditions of the tests. Surfaces in contact Condition of the surfaces Mutual arrangement of the fibers Oak on oak ' Coated with dry soap ' Parallel. Oak on oak Coated with tallow Parallel. Oak on oak Coated with lard I Parallel. Oak on oak Oily Oak on oak Dry Oak on oak Coated with tallow. Oak on oak Coated with lard. . . Oak on oak Oily Beech on oak. ... Coated with tallow. Oily. Beech on oak. Elm on oak Coated with dry soap. Elm on oak Coated with tallow. . . Elm on oak. Elm on oak. Elm on elm. Elm on elm. Coated with lard Oily Coated with dry soap Oily Oak on elm Dry Oak on elm Coated with dry soap Oak on elm Coated with tallow Parallel Oak on elm Coated with lard | Parallel Oak on elm Oily Parallel Parallel Perpendicular . Perpendicular . Perpendicular . Perpendicular . Parallel Parallel Parallel Parallel Parallel Parallel Parallel Parallel Parallel Parallel Coefficient of friction o. 164 0.075 0.067 0.108 0.336 0.083 0.072 0.143 0-055 0153 0.137 0.070 0.060 o. 119 0.139 0.140 0.246 0.136 0.073 0.066 0.136 Table 4. — Coefficients of Friction for Wood on Wood, Dry or Scantily Lubricated Surfaces {Morin). 12 BEARINGS AND THEIR LUBRICATION Surfaces in contact Condition of the surfaces Coefficient of friction Wrought iron on oak Wrought iron on oak Wrought iron on oak Cast iron on oak .... Cast iron on oak Cast iron on oak .... Cast iron on oak Cast iron on oak Cast iron on oak Cast iron on oak Copper on oak Copper on oak Cast iron on elm .... Cast iron on elm .... Cast iron on elm .... Cast iron on elm .... Cast iron on elm Cast Iron on elm Moistened with water. Coated with dry soap . . Coated with tallow. . . . Dry Coated with dry soap . Moistened with water. Coated with tallow . . . Coated with lard Coated with olive oil. , Oily Coated with tallow . Oily Dry Wrought iron on elm Wrought iron on elm Wrought iron on elm Wrought iron on elm Wrought iron on elm .... Oak on cast iron Oak on cast iron Elm on cast iron Elm on cast iron Lignum-vitae on cast iron. Lignum-vitae on cast iron. Lignum-vitae on cast iron. Oak on wrought iron Oak on wrought iron Lignum-vitae on bronze. . , Lignum-vitae on bronze. . , Lignum-vitae on bronze. . Coated with tallow Coated with olive oil Coated with lard and plumbago. . . Greasy after coating with tallow. . .. Greasy after coating with lard and plumbago. Dry Coated with tallow. . Coated with lard .... Coated with olive oil . Oily Coated with tallow. . Oily Coated with tallow . . Oily Coated with tallow. . . Coated with olive oil. Oily Coated with tallow . . Oily Coated with tallow . . Coated with olive oil. Oilv 0.256 o. 214 0.085 0.490 0.189 0.218 0.078 0.075 0.075 0.107 0.069 o.ioo 0.195 0.077 0.061 0.091 0.125 0.137 0.252 0.078 0.076 0-055 0.138 0.080 0.168 0.066 0-135 0.074 0.076 O.I2I 0.098 0.149 0.082 0-053 0.146 Table 5. — Coefficients of Friction for Wood and Metal, Dry or Scantily Lubricated Surfaces (Morin). SLIDING FRICTION 13 Surfaces in contact Condition of the surfaces Cast iron on cast iron Cast iron on cast iron Cast iron on cast iron Cast iron on cast iron Cast iron on cast iron Cast iron on cast iron Cast iron on cast iron Cast iron on cast iron Wrought iron on cast iron . . . . Wrought iron on cast iron. . . . Wrought iron on cast iron. . . . Wrought iron on cast iron Steel on cast iron Steel on cast iron Steel on cast iron Steel on cast iron Steel on cast iron Yellow copper on cast iron . . . Yellow copper on cast iron. . . Yellow copper on cast iron . . . Yellow copper on cast iron . . . Yellow copper on cast iron. . . Bronze on cast iron Bronze on cast iron Bronze on cast iron Bronze on cast iron Cast iron on wrought iron. . . . Cast iron on wrought iron. . . . Cast iron on wrought iron .... Wrought iron on wrought iron Wrought iron on wrought iron Wrought iron on wrought iron Wrought iron on wrought iron Wrought iron on wrought iron Steel on wrought iron Steel on wrought iron Bronze on wrought iron Bronze on wrought iron Bronze on wrought iron Bronze on wrought iron Bronze on wrought iron Cast iron on bronze Cast iron on bronze Dry Moistened with water Coated with soap Coated with tallow Coated with lard Coated with olive oil Coated with lard and plumbago. Oily Dry Coated with tallow , Coated with lard , Coated on olive oil Dry Coated with tallow. Coated with lard. . . Coated on olive oil . Oily Dry Coated with tallow. . Coated with lard .... Coated with olive oil , Oily Dry Coated with tallow. . Coated with olive oil . Oily Coated with tallow. . Coated with lard Coated with olive oil . Dry Coated with tallow . . Coated with lard Coated with olive oil , Oily Coated with tallow. . Coated with lard .... Dry Coated with tallow Coated with lard and plumbago. Coated with olive oil Oily Dry Coated with tallow Coeflficient of friction 0.152 0.314 0.197 O.IOO 0.070 0.064 0-055 0.144 0.194 0.103 0.076 0.066 0.202 0.105 0.081 0.079 0.109 0.189 0.072 0.068 0.066 0.II5 0.217 0.086 0.077 0.107 0.098 0.053 0.063 0.138 0.082 0.081 0.070 0.177 0.093 0.076 0.I6I 0.081 0.089 0.072 0.166 0.147 0.085 Table. 6. — Coefficients of Friction for Metal on Metal, Dry or Scantily Lubricated Surfaces {Morin). ' 14 BEARINGS AND THEIR LUBRICATION Surfaces in contact Cast iron on bronze Cast iron on bronze Cast iron on bronze Wrought iron on bronze Wrought iron on bronze Wrought iron on bronze Wrought iron on bronze Wrought iron on bronze Steel on bronze Steel on bronze Steel on bronze Steel on bronze Bronze on bronze Bronze on bronze | Coated with ohve oil Bronze on bronze Oily Condition of the surfaces Coated with lard Coated with olive oil . . Oily Dry Coated with tallow Coated with lard Coated with olive oil Oily Dry Coated with tallow Coated with olive oil Coated with lard and plumbago.. Dry Coefficient of friction 0.070 0.067 0.132 0.172 0.103 0.075 0.078 0.160 0.152 0.056 0.053 0.067 0.201 0.058 0134 Table 6 (Cow^iwMsc?).— Coefficients of Friction for Metal on Metal, Scantily Lubricated Surfaces {Morin). Dry or Surfaces in contact Condition of the surfaces Arrangement of the surfaces Coefficient of friction Tanned cowhide on cast Dry j The leather laid flat. . . iron, j Tanned cowhide on cast Moistened a n d saturated The leather laid flat. . . iron. Tanned cowhide on cast iron. Tanned cowhide on cast iron. Tanned cowhide on cast iron. Tanned cowhide on cast iron. Tanned cowhide on cast iron. with water. | Coated with tallow The leather laid flat. . Coated with oHve oil The leather laid flat. . The leather oily, the cast The leather laid flat iron moistened with water Moistened with water 0.559 0.365 0.159 0.133 0.229 Moistened with olive oil. The leather placed 0.338 edgewise. The leather placed 0.135 edgewise. Table 7. — Coefficients of Friction for Leather on Cast Iron, Dry or Scantily Lubricated Surfaces {Morin). As speeds increase the value of the coefficient of friction decreases, but the amount of this decrease is unknown experimentally, except for certain com- binations of metals. In Elements of Machine Design by Kimball and Barr, SLIDING FRICTION 1 5 page 99, the coefficient of friction for cast iron or steel at a velocity of 440 ft. per minute is given as 0.32. At 2460 ft. per minute as 0.2, and at 5280 ft., or a mile a minute, as 0.006. It is also stated that these data for cast iron on steel will serve as a rough guide to what may be expected to occur for other com- bination of materials. A consideration of the values of the coefficient of friction in Table 2 show that a great decrease takes place as soon as motion begins, even although the velocity is very low. It might be supposed that the change is very abrupt from t"he friction of rest of the friction of motion, but it is now generally believed that this change is gradual, and that the value of the coefficient at rest is not far different from the value at a very slow speed. LAWS OF DRY OR UNLUBRICATED FRICTION John Goodman sums up the laws governing dry friction in this manner, Proc. Inst. C. E., Vol. LXXXV, page 391. 1. The friction between dry surfaces under moderate loads and low veloci- ties varies directly as the normal pressure between them. 2. The normal pressure remaining unchanged, the friction is independent of the area in contact. 3. The friction is always greater on the reversal of direction of sliding. 4. The friction sensibly diminishes with a rise of the temperature. These laws do not differ materially from those of General Morin, except that they are restricted to low velocities and moderate loads. As a matter of fact, these were the conditions under which General Morin experimented. FRICTION OF LUBRICATED SURFACES The laws for lubricated friction must be considered from an entirely different standpoint than those from dry friction, although it is very difficult, if not impossible, to define accurately where one condition begins and the other ends. In a journal bearing, properly lubricated, the friction resistance must obviously conform to the laws for lubricated surfaces. On the other hand, if the supply of lubricant is intermittent, as the quantity diminishes the resistance will tend to conform to the laws for dry or unlubricated surfaces. This latter condition is represented by the usual oil hole and squirt-can method of oiling. Other fami- liar forms of oiling are by means of sight feed and syphon lubricators, pad lubri- cators, oiling rings and chains, and oil baths. The efficiency of these devices increases in the order named. Lubricant is introduced between bearing surfaces in order to separate them by a film of lubricant and thereby change the frictional resistance from that 1 6 BEARINGS AND THEIR LUBRICATION of the solid surfaces in contact to the fluid- resistance of the film. Thus the nature and condition of the lubricant has more to do with this kind of friction than the bearing surfaces themselves assuming that they are well fitted. FRICTION OF REST OF LUBRICATED SURFACES It is frequently observed that after a machine has been idle for some time, it ''starts hard." The bearings are sometimes said to be ''stiff." The reason for this is that the pressure of the journals in the bearings tends to squeeze out the lubricant and allows the metallic surfaces to come into more or less close contact. Experience has shown that it is very difficult even with small areas and heavy loads to squeeze out all the lubricant, but its thickness will be very much reduced if the bearing is idle for any considerable length of time. Thus when a machine is started after a long shut down the frictional resistance in its bearings is more nearly that of dry or unlubricated surfaces than of well lu"bri- cated surfaces. This fact shows the necessity of carefully oiling a heavy machine before it is started. If this is not done the surface of journals and bearings may be in such close contact that abrasion will take place before the lubricant re- maining in the bearing can be distributed and an oil film established. The coefficients of rest or starting of motion as given by Professor Thurston in Friction and Lost Work, page 319, for sperm oil, lard oil and mineral oil are, respectively: 0.15; o.i; 0.15. These are very much higher than similar coeffi- cients for well-lubricated surfaces when in motion. These values were obtained from experiments made on a testing machine at pressures ranging from 75 to 500 lb. per square inch and with lubricant between metallic surfaces. FRICTION OF POORLY LUBRICATED SURFACES In the preceding paragraph is pointed out the fact that the friction of lubri- cated surfaces at rest is much greater than the friction of similar surfaces well lubricated when in motion. Between these two extremes is the condition of imperfect lubrication. This is the condition under which most machine bear- ings are run. Yet there are so many variables that it is impossible to establish any laws. The better the imperfect lubrication, the more nearly will the condi- tions conform to the laws for well lubricated surfaces; The poorer the lubrica- tion, the more nearly will these conditions conform to the laws of dry surfaces. The variables that govern the condition of poorly lubricated surfaces are the character of the surfaces themselves, their fit, the supply and nature of the lubri- cant, the pressure between them, the velocity of rubbing and the temperature as affecting the viscosity of the oil or grease. With so many variables it is easy to see the difficulties of obtaining results from even the most careful experi- mental work. SLIDING FRICTION 17 Efforts have been made to state the laws of imperfect friction, but with very little success. It is probably of more value to consider briefly some of the actions that take place in a bearing. The pressure tends to squeeze out the lubricant. If this pressure is very high it may be impossible to get a sufficient supply of oil between the surfaces to produce a proper film. Then the bearing would be practically in metallic contact with its journal, and the coefficient of friction would be nearly that of dry surfaces in contact. Again, if the velocity of rubbing is so very low that the lubricant cannot be drawn between the surfaces to form a film, there may be practically metallic contact. This same condition may be produced by an insufficient quantity of oil and this is the most general cause of imperfect lubrication. If the pressure and velocity of rubbing are normal the supply of oil or grease may be such as to give any condition of lubrication between that of unlubricated surfaces and well lubricated surfaces. The influence of temperature is important, for even with a suitable supply of lubricant at a given pressure, velocity of rubbing and ordinary temperature, the thinning of the oil as the temperature increases may be sufficient to break down the film and produce metallic contact. Excellent lubrication can only be obtained by a plentiful supply of oil, as by an oil bath, flooded lubrication, and the best arrangements of oiling rings. The greater number of machine bearings are oiled either by a squirt can through an oil hole, by sight-feed oil cups, syphon oil cups, pads, or compression grease cups. All of these devices as ordinarily used can be considered as producing nothing but imperfect lubrication, and thereby bear out the general statement that the great mass of machinery bearings are in this second indefinite condi- tion of poorly lubricated surfaces. The only way in which this subject can be satisfactorily considered is by studying its limiting factors. In a former paragraph, page 15, is discussed the friction of dry surfaces and the following treats of the friction of well lubri- cated surfaces. WELL LUBRICATED SURFACES The best condition for a bearing to be in, and unfortunately the rarest, is, of course, well lubricated. This means a plentiful supply of lubricant, as from an oil bath. When this condition prevails the surfaces of the bearings are defi- nitely separated by an oil film and the frictional resistance is the resistance to shearing of this film due to the rubbing action. The journal floats on the lubricant. As the pressure on the journal under ordinary conditions has a definite l8 BEARINGS AND THEIR LUBRICATION direction, the space between the journal and bearing is less at the point of nearest approach than at a point diametrically opposite. This means that the space between the journal and bearing is a wedge-shaped ring. The action of the journal as it rotates is to take oil from the region of least pressure and carry it into the region of greater pressure. This action is sufficient to suck oil against small heads. This continual drawing in of oil preserves the film and at the same time necessitates the doing of a certain amount of work in transferring the oil from one point to another in the bearing and in overcoming its own fluid resistance. We are fortunate in the amount of experimental work that has been done to determine the laws of friction for well lubricated surfaces. This subject is by no means in the haze that surrounds poorly lubricated surfaces, although much work still awaits any investigator who chooses to enter the field. THEORY OF BEARING LUBRICATION Professor Osborne Reynolds has given a most interesting and valuable theoretical discussion of bearing lubrication in a paper, ''Theory of Lubri- cation," Phil. Trans. (Royal Society, London), 1886. This should be read by everyone who wishes to grasp the fundamentals of the subject, yet the empiri- cal rules given in following paragraphs are more directly serviceable for designing purposes. Professor Kingsbury, Journal Am. Soc. Naval Engineers, Vol. IX, No. 2, (1897) gives six laws for perfect lubrication arranged from Reynold's paper. The reference is to Fig. i, reproduced from Reynolds and representing a right section of a cylindrical journal and its brass; the rotation of the journal is counter-clockwise. The condi- tions are such that the film is complete. I. The vertical components of the pressure and friction balance the load. 2. The horizontal component of the oil pressure balances the horizontal component of the friction. 3. When the brass is unloaded, its point of nearest approach will be its middle point. The pressures are then symmetrically disposed with reference to O, the positive on the right, the negati/e on the left. 4. As the load increases, the positive vertical component must overbalance the negative component. This requires that H should be at the left of O. 5. As the load increases, OH reaches a maximum value which places H nearly, but not quite as the left extremity of the brass, but still leaves // small Fig. I. DIAGRAM OF SECTION OF JOURNAL AND ITS MATING BEARING. SLIDING FRICTION 19 as compared with GH. For a further increase of the load, H moves back again toward O. 6. As the load increases, the distance between the axes of brass and journal increases. LAWS OF FRICTION FOR WELL LUBRICATED SURFACES The following five empirical laws of friction for well lubricated surfaces have been modified from four laws laid down by John Goodman, Proc. Inst. C. E., Vol. LXXXV, page 380, and from the results of Lasche's experiments. It is, of course, uncertain just how far they apply. But each law is amplified by the results of experiments. 1 . The coefficient of friction for well lubricated surfaces is practically inde- pendent of the materials composing the bearings and journals, but is dependent upon the nature of the lubricant. 2. The coefficient of friction for well lubricated surfaces is from i/6toi/io that for unlubricated or scantily lubricated surfaces. 3. The coefficient of friction for well lubricated surfaces and moderate pressures and speeds varies approximately inversely as the normal pressure; that is, the frictional resistance per unit of area is a constant assuming the speed constant. The total frictional resistance varies directly as the area. Or, •J=J ^nd U,=f,p, h Pi 4. The coefficient of friction for well lubricated surfaces for a given pressure is very high for low rubbing velocities, though decreasing rapidly as the veloc- ity increases. For velocities from 100 to 500 ft. per minute it decreases, vary- ing approximately as the square roots of the velocities ; for velocities from 500 to 1600 ft. per minute it decreases at a lower rate, varying approximately as the fifth roots of the velocities; for velocities above 1600 ft. per minute, it is practi- cally independent of the velocity. Or, for velocities from 100 to 500 ft. per minute ^ = -— ^ fi VV2 Or, for velocities from t;oo to 1600 ft. per minute y = ., — Or, for velocities over 1600 ft. per minute J^=f^ 5. The coefficient of friction varies approximately inversely as the tempera- ture up to a point just before abrasion takes place. DISCUSSION OF THE FIRST LAW That the metals forming the journal and bearing in well lubricated bearings have little influence on the coefficient of friction has been observed by many 20 BEARINGS AND THEIR LUBRICATION investigators. In Traction and Transmission for January, 1903, Lasche describ- ing some of his own experiments says: "The results of numerous experiments go to prove that the metal forming the journal — nickel steel, ordinary Siemens- Martin steel, or mild steel — has no marked influence on the degree of friction or friction work." And again: "The journal v/as placed in a gun-metal bushing; in one with white-metal lining and in one with a mercury amalgam lining. As in the case with the metal forming the journal the various kinds of bushings gave approxi- mately the same results." The reason for this is that the load is fluid-borne and so the frictional resistance is that of the oil film not of the metallic surfaces. DISCUSSION OF THE SECOND LAW The truth of the second law is shown by the following table taken from the work of Beauchamp Tower, Proc. Inst. M. E. for 1883, page 651. Method of lubrication Pad under journal Oil bath Syphon lubricator Coefficient of friction Comparative friction 6.48 I 7 .06 In another experiment made by Mr. Tower with a load of 293 lb. per square inch and a journal speed of 314 ft. per minute with an oil bath, a coefficient of friction of 0.0016 was obtained. Under the same conditions with a pad the coefficient was 0.0097 which is six times the amount of the former. Under the same condition Mr. Stroudley obtained a coefficient of 0.00961 which is in very close agreement. DISCUSSION OF THE THIRD LAW The following tabulation is from the experiments of Mr. Tower, Proc. Inst. M. E. for 1883, page 650. Load in lb. per sq. in., 520 468 415 363 310 258 205 153 100 Frictional resistance per sq. in., 0.416 0.514 0.498 0.472 0.464 0.438 0.43 0.458 0.45 This shows that the unit frictional resistance is sensibly constant with vary- ing loads. The frictional resistance per square inch is the product of the coefficient of friction times the load per square inch on the projected area of the bearing. Thus, if this product is a constant one factor must vary inversely as the other; SLIDING FRICTION 21 a high load will give a low coefficient and a low load a high coefficient. This is true for well lubricated surfaces. For ordinary lubrication the coefficient is more nearly constant under vary- ing loads; the friction resistance then varies directly as the load. This is shown by Table 8 taken from the results of the experimental work of Beauchamp Tower, Proc. Inst. M. E., 1883. Nominal load lb. per sq. in. Actual load lb. per sq. in. Temper- ature Fahr. 100 rev. I OS ft. per min. lb. ISO rev. IS7 ft. per min. lb. 200 rev. 209 ft. per min. lb. 250 rev. 262 ft. per min. lb. 300 rev. 314 ft. per min. lb. 3 so rev. 366 ft. per min. lb. 400 rev. 419 ft. per min. lb. 328 310 293 275 258 20S 153 100 582 551 S20 498 4S8 364 272 178 82° 76° 77° 78° 82° 74° 75° 3-5 3.06 3.06 2.49 2.44 1.78 1-473 1.093 3-35 2.84 2.84 2.62 2.28 I.73S I.S6 I. 225 3-21 3 06 2.84 2.84 2.17 1. 60s 1. 60s ..33 3.06 3.c6 2.49 2.89 2.X4S X.37 0.992 . 2.1 1.75 1. 81 1.44 2.13 2.04 1.89 1. 54 1.56 1 .0 The chord of the arc of contact of the brass = 2 1/4 in. The nominal load per square inch is the total load divided by 4X6. The actual load per square inch is the total load divided by 2 1/4X6. The results with the actual load of 582 lb. per square inch were obtained with difficulty, and the bearing seized with that load after running for a short time. The lubricating pad consisted of a piece of felt pressing against the journal, and resting on worsted immersed in a tin box full of oil. Table 8.- -Relationship of Load and Frictional Resistance for Ordinary Lubrication Mr. Stroudley with pad lubrication on- railroad car axles on the Brighton Railway in England obtained the result of Table 9 which still further supports the law. Load per square inch on brass in pounds Coefficient of friction 293 251 237 0.00792 o . 0080 0.0077 o . 007 2 Friction resistance per square inch in pounds 2.639 2.34 1-93 1.70 Table 9. — Relationship of Load and Frictional Resistance as Determined by Stroudley. 22 BEARINGS AND THEIR LUBRICATION DISCUSSION OF THE FOURTH LAW Turning to the fourth law, A. M. Wellington, Trans. A.S.C. E., Vol. XIII, found from experiment on journals revolving at very low velocities that the friction under those conditions was very great and nearly constant under widely varying conditions of lubrication, load and temperature. As the speed increased the friction fell slowly and regularly, returning again to the original amount when the velocity was reduced to the same rate. The tabulation below shows some of the results. Feet per minute, 2.16 3.33 4.86 8.82 21.42 35-37 53.oi 89.28 106.02 Coefficient of friction, 0.118 0.094 0.0705 0.0685 0.055 0.047 0.04 0.035 0.3 0.0255 This shows a regular decrease in the value of the coefficient of friction with increase of speed. The results of experiments made by Professor Kimball, see page 1904 Americin Journal of Science, March, 1878, show this same law of the decrease of the coefficient of friction with the increase of speed. The follov^ng gives some of Professor Kimball's results: Velocity in ft. i 3 5 7 10 15 20 30 40 60 80 100 per min., Coefficient, 0.15 0.122 0.114 0.093 0.079 0.066 0.058 0.054 0.053 0.052 0.051 0.05 These experiments were made with journal bearings and show that the friction was reduced 67 per cent, with an increase of velocity from i to 100 ft. per minute. After this limit of velocity 100 ft. per minute has been passed, the coefficient of friction varies approximately with the square root of the velocity. Origin of data Observed. . Calculated. Observed.. Calculated. Load lb. per sq. in. Velocity in feet per minute 209 262 314 366 419 520 468 0.0013 0.0014 0.0015 0.0017 0.00145 0.00159 0.00172 0.0015 0.00184 471 o.ooi 0.0012 [ 0.0013 0.0014 0.0015 0.0017 I 0.0II8 0.00123 0.00132 0.00I4I 0.0015 0.002 0.00195 Observed.. Calculated. 415 0.0014 ! 0.0015 0.0017 0.0019 0.0021 0.0024 0.00157 0.00172 ! 0.00188 I 0.00198 ! 0.0021 Table 10. — Comparison of Observed and Calculated Coefficients of Friction for Velocities over 100 Ft. per Minute. SLIDING FRICTION 23 Table 10 shows us this fact. The observed coefficients of friction for various velocities are taken from some of Mr. Tower's experiments. The comparable calculated coefficients were computed from the ratio of the square roots of the velocities. The agreement is very close. For the experimental work to determine the coefficient of friction at high speeds, we are indebted to Thurston, Stribeck and Lasche. In Traction and Transmission, January, 1903, Lasche thus sums up the results of these experi- ments: From 500 ft. per minute to 800 ft. per minute the experiments of Stri- beck show that the rise in the value of the coefficient of friction is slow, and approaches the result of the Thurston experiment expressed in the formula ~ ~ — -w Ste hit< 3land 1 IV"*' ,1 N ckelS teelar d lit! Utl Nickel Steel and ^ Mild Ste el a nd 1 0.020 V in\ e ^ let al. ^JjkelSt eel aie •5 0.010 1 jL ^ ^ ^ ^ ^ ;=:: zz = = ^^ == -=4 — 1 n — -^ ^^'. = — = — 1 r = — — — n ~ — ■ — 1000 2000 3000 4000 Circumferential Journal Speed in Feet per Minute. 5000 FIG. 2. CURVES FROM LASCHE'S EXPERIMENTS SHOWING THE VARIATION IN THE COEFFICIENT OF JOURNAL FRICTION WITH VELOCITY. f=\/V. At a still higher speed the influence of the velocity disappears and at 1600 ft. per minute the curve of the values of the coefficient is only slightly higher than at 800 ft. per minute. The experiments of Lasche further give the conclusion that for velocities of 2000 ft. per minute and over the coefficient of friction is practically independent of the velocity. There is a break in these results between a speed of 1600 ft. per minute and 2000 ft. per minute, but from a consideration of the plotted curve of Lasche's work we seem justified in assuming that for velocities of 1600 ft. per minute and over the coefficient of friction is practically independent of the velocity. Fig. 2 shows the general shape of the coefficient of friction curves plotted against velocities of rubbing as abscisses. DISCUSSION OF THE FIFTH LAW Law 5 states that the variation of friction with temperature is approximately in inverse ratio. The following taken from Mr. Tower's experiments at a velocity of 262 ft. per minute with an oil bath show this law: 24 BEARINGS AND THEIR LUBRICATION Temperature in Fahrenheit degrees Observed coefficient of friction Calculated coefhcient of friction no lOO 90 80 70 60 o . 0044 0.0051 0.006 0.0073 0.0092 O.OII9 0.00451 0.00518 o . 00608 0.00733 0.00964 0.01252 However, this law does not hold good for imperfect lubrication, that is, for pad or syphon lubrication. Then the coefficient of friction diminishes more rapidly for a given temperature according to a gradually decreasing scale, until a normal temperature is reached. This is shown by the result of Mr. Stroud- ley's experiments with a pad of rape-seed oil. Temperature Fahrenheit degrees Coefficient of friction Decrease of coeifjcient 10; 0.022 0.018 0.016 0.014 0.0125 0.0115 O.OII 0.0106 0.0102 0.004 0.002 no lie 0.002 120 0.0015 O.OOI i2t: 130 i^e 0.0005 . 0004 0002 140. . . 141; THICKNESS OF THE OH. FH^M A series of experiments made by Goodman to determine the thickness of the oil film showed that with pad lubrication on first starting the journal, the thickness of the film was 0.0013 i^- As the speed increased and the film became better developed its thickness increased to 0.0029 in. In the American Machinist for September 17, 1903, page 13 16, Herbert S. Moore, writing in regard to a film formed with a heavy engine oil, states: ''It was at first attempted to get a perfect film in the Olsen- Carpenter fric- tion machine, in which the test bearing was a snug fit all around the journal. It was found impossible to get such a film except at rare intervals of short dura- tion, whereas with the altered Thurston machine, in which the diameter of the SLIDING FRICTION 25 bearing was 0.003 i^- greater than that of the journal, this film was always readily formed and preserved." If the film was uniform its thickness would thus be 0.0015 in. BREAKING-DOWN POINT OF THE OIL FILM The difficulty of lubricating bearings under high pressures is well known. The reason is evident. As the pressures are increased the film of oil is estab- lished and maintained with increasing difficulty. If the pressures are increased to a sufl&cient amount the film of oil finally breaks down, and the journal and bearings come into metallic contact. Herbert S. Moore, on page 1283, American Machinist for September 10, 1903, gives a formula based on experimental investigations, to determine the pressure at which a film of machinery oil will break down. This formula is in which p is the intensity of the load in pounds per square inch, and v is the rubbing velocity of the journal in feet per minute. The experimental work consisted in connecting into an electric circuit the rotating journal and its bearing. As long as the oil film was properly main- tained the resistance of this circuit was very high. As soon, however, as the oil film broke down, giving metallic contact between journal and bearing, the resistance of the circuit fell to almost nothing. PRESSURE OF THE OIL FILM If a journal is completely fluid-borne it follows that the oil film must be under a considerable pressure in order to support the weight of the journal and its attached parts. That such is a fact was demonstrated by an experiment by Mr. Towers reported in the Proc. Inst. M. E. for 1885, P^g^ 5^- I^ these ex- periments a half bearing or brass was used, having a length of 6 in. and a chord length of the arc of contact of 3.9 in. Three longitudinal holes were drilled a little more than half way of the length of the bearing, and spaced as indicated by the letters a, d and g in Fig. 3. At 9 points, as indicated by the letters in the figure, holes were drilled from the inside of the bearing into these three longitudinal channels. Each channel was connected with a pressure gage, and one hole at a time was left open from the channel into the inside of the bearing. By this means pressure readings were taken at nine different points, three on the side of the bearing where the journal entered, three on the middle line, and three on the side of the bearing where the journal left. Furthermore, as six of these holes were between a middle transverse plane and the end, and since it is fair to assume that the 26 BEARINGS AND THEIR LUBRICATION same pressure would exist on the other side of this middle plane, it can be said that the experiments covered 15 points on the surface of the bearing. The lubrication was by means of an oil bath, the journal being about half submerged. The speed of rotation was kept uniform at 150 revolutions per minute, and the temperature was held constant at 90° F. The total load on the bearing was 8008 lb., and the observed oil pressure at the 9 points investigated were as shown by the following: For longitudinal plane at entering side of the journal a 310, h 335, c 370; for middle longitudinal plane d 565, e 615,/ 625; for longitudinal plane at leaving side of the journal g 430, h 480, i 500. A summation of the total pressure on the bearing from these observations gives 7988 lb. This compares very closely with the actual pressure, which was 8008 lb., the small difference being probably due to inaccuracies in observations. c»i i i ■ , III' s — E 1 1 1 1 1 1 [i \h [p" 1 ]—-■—■ 1 1 1 1 1 .1 1 1 1 FIG. 3 DIAGRAM SHOWING POINTS AT WHICH OIL PRESSURES WERE TAKEN. These results present several interesting facts. The length of the bearing was 6 in., and the length of its chord of contact 3.9 in., giving a projected area of 23.4 sq. in. The applied load was 8008 lb., which divided by the pro- jected area gives an intensity of pressure per square inch of about 311 lb. The highest observed pressure at the point/, or center of the bearing area, was 625 lb., or double the computed intensity of pressure. The observed pressures at the points g, h and i on the leaving side of the bear- ing are considerably greater than the corresponding pressures at the points a, h and c on the entering side. This shows the shifting of the point of maximum pressure in the bearing from a point diametrically below or above the axis of rotation, to a point with the direction of rotation. See the following section devoted to a discussion of the points of maximum and minimum pressure in an oil film. SLIDING FRICTION 27 The existence of these pressures in an oil film emphasizes the necessity of using good care and judgment in selecting the points where lubricant is to be introduced. Obviously oil cannot be introduced into a bearing unless it is under a greater pressure than the pressure in the oil film at the point of entry. Some failures in elaborate oiling systems can be traced to this very fact; the attempt to feed oil at a point in the bearing where the oil film was under a greater pressure than the head of the feed. Referring again to Mr. Tower's experiment, he found that by varying the load the oil pressures rose and fell in exact proportion. Putting on weight increased the oil pressures, and removing weight decreased them. To deter- mine the effect of speed on this oil pressure the journal speed was reduced from 150 revolutions per minute to 20, or as the diameter of the shaft was 4 in., from about 150 ft. per minute to about 20 ft. per minute. The gage pressures of the oil film were exactly the same at the lower speeds as at the higher. Reynolds* theory and Kingsbury's experiments show, however, that the distribution of the pressure does vary with speed. As there are points of maximum pressure in the oil film in a bearing, there are likewise points of minimum pressure, and these may be below the pressure of the atmosphere, or at a slight vacuum. This is true for arcs of contact from 360 to less than 180 degrees. At page 1282 of the American Machinist, September 30, 1903, Herbert F. Moore describes an experiment with a small dynamo bearing, in which the bearing sucked oil against a head of 6 in. We quote: *'By simply leading a pipe from a reservoir of oil to that part of the bearing where the pressure would be negative, if anywhere, it is found that the bearing would suck in oil from a reservoir 6 in. below and thus lubricate itself." POINTS OF MAXIMUM AND MINIMUM PRESSURE AND NEAREST APPROACH In the earlier discussions of the relative positioning of a journal and its mating bearing when the journal was in motion with ample lubrication present, it was stated that the journal rolled back so that the point of nearest approach was on the on side of the bearing and at an angular distance from the vertical — assuming the load to act vertically with the bearing supporting the journal — equal to the angle of repose of the system. This theory has been shown to be fallacious by the experimental work of Tower and analytical work of Reynolds. In an ordinary journal bearing where there is some looseness, when the journal is at rest it lies at the bottom of the bearing and the point of nearest ap- proach, in this case contact, lies in a vertical passing through the centers of both. The rolling action outlined in the paragraph above may take place during the 28 BEARINGS AND THEIR LUBRICATION 190 180 170 160 .150 |140 Kl30 S120 o ^ , — - — — — ' ^^• pr^ ^^r J/ k V / / / / / 80 70 60 50 40 30 ^l< k pr-1 ^>. *^ V. r-- *^> ki '^c^,V r^J r"-^r N: -^. ■ — .^ Th nnes tPo 1 1 int of Film ^ Oil "■^ ' ^^ =^ ^ ::^ ^^ N CLc ^ ;^ ^ -. . \ -.* ^fnt "'^- _^ 3 20 10 p^ tax. Pres surp o^- ~1 — > n • 0.0007 0.0006 0.0005 0.0004 0.0003 0.0002 0.0001 800 1200 IGOO 2000 2400 28U0 3;i00 Revolutions per Minute 3600 4000 4400 400 800 1200 1600 2000 2400 2800 3200 3600 4000 4400 Revolutions per Minute FIG. 4. RELATION BETWEEN POINTS OF MINIMUM AND MAXIMUM PRESSURE AND POINT OF NEAREST APPROACH. SLIDING FRICTION 29 period of starting, but if there is any appreciable amount of lubricant present, as soon as sliding takes place the point of nearest approach and the point of maximum pressure (which are not identical) are thrown to the off side of the vertical through the centers of journal and bearing. See the upper diagram in Fig. 4. The angular amount of this shifting is dependent on the fit and finish of the surfaces, the quality of lubrication, the speed, and the pressure. With a vari- able load and constant speed, as the load increases the point of nearest approach swings upward continuing until it reaches a position nearly, but not quite, at the extremity of the horizontal axis. With a further increase of load, it swings back to a position found by Reynolds to be about 40 degrees from the vertical. With a still further increase of load, the oil film is ruptured and the conditions are changed from that of perfect lubrication to that of imperfect lubrication. Quoting from Reynolds' paper, Phil. Trans. (Royal Society, London) 1886: ''The circumstances which determine the greatest load which a bearing will carry with complete lubrication, that is, with a film of oil extending between brass and journal throughout the entire area, are definitely shown in the theory. " The effect of increasing the load beyond a certain small value being to cause the brass to approach nearer to the journal at a point on the off side which moves toward the vertical as the load increases and when the load is such that the least separating distance is about half the difference of the radii, the angular position of the point of nearest approach is 40 degrees to the off side of the vertical through the middle of the brass. At this point the pressure in the oil film is everywhere greater than at the extremities of the brass, but when the load further increases the pressure toward the extremity on the off side becomes smaller or negative. This, when suflScient, will cause rupture in the oil film, which will then only extend between the brass and journal over a portion of the whole area and a smaller portion as the load increases. Thus since the amount of negative pressure which the oil will bear depends on circumstances which are uncertain, the limit of the safe load for complete lubrication is that which causes the least separating distance to be half the difference of the radii of brass and journal." An examination of Reynolds' equations shows that the load and speed are inversely related; that is, an increase of load has an effect upon the form of the film of the same kind as a decrease of speed, and vice versa. That practice agrees with this theory is shown by a phenomenon observed by Tower and by the position of the wear in car journal brasses. Tower men- tions that the journal having been run in one direction until the initial tendency to heat had entirely disappeared, on being reversed immediately began to heat again; but this effect stopped when the process had been often repeated. The fact being that running in one direction, the brass had been worn to the journal 30 BEARINGS AND THEIR LUBRICATION only on the off side for that direction, so that when the motion was reversed the new off side acted like a new brass. In car brasses the point of wear is always forward or on the off side. The relative relations of the points of maximum pressure, minimum pres- sure, and nearest approach, are shown by the curves of Fig. 4. These are taken from a paper of Professor Kingsbury, published in the Journal of the American Society of Naval Engineers, Vol. IX, No. 2, 1897, and apply to an experiment with an air-lubricated bearing, running under a constant load and at variable speed. SECTION II COEFFICIENTS OF FRICTION OF JOURNAL, COLLAR, STEP AND GUIDE BEARINGS From the discussion of imperfect lubrication, that is the ordinary lubrication to machinery journal bearings, it is evident that no fixed values can be given for the coefficient of friction. Any values that are given must be for definite lubri- cants, and not for surfaces in contact. With perfect lubrication and moderate velocity the coefficient may be as low as o.ooi. With very low velocities and high pressures the coefficient approaches that for greasy or unlubricated sur- faces, a minimum value for which may be taken as 0.15. Tests of new well- fitted bearings have given values of 0.30 and more. These values are the ones to use in estimating the starting moment for heavy machinery. On page 105 of Elements of Machine Design, by Kimball and Barr, the coeffi- cients of friction to be used in the design of ordinary machinery are given as follows: A fair average range for pressures from 50 to 500 lb. and velocities from 50 to 500 ft. per minute is from 0.02 to 0.008, and for purposes of design of ordinary machinery, may be taken at 0.015. Perhaps the lowest coefficient of friction that has been reported from authen- tic tests will be found on page 149, Trans. A.S. M. E., Vol. LXXIV. There Professor Kingsbury presents a table giving coefficients of friction for sperm oil under a pressure of 340 lb. per square inch at a temperature of 90° F., and varying rubbing speeds. 42 feet per minute 64 feet per minute loi feel, per minute . 000906 0.000915 . 0008S6 0.000915 0.001171 0.001158 0.001158 O.OOI 1 25 0.00146 0.00155 0.00151 On page 146 of the same paper a value of 0.00053 is given. The lowest coefficient of which the author has learned is 0.00046. This was also obtained by Professor Kingsbury in a test, the results of which have never 31 32 BEARINGS AND THEIR LUBRICATION before been published. The lubricant was spindle oil, pressure 340 lb. per square inch, temperature 135° F., and a rubbing speed of 42 4 ft. per minute. To give a definite idea of the value, the coefficients of friction for general bearings and the way in which they vary under different conditions. Tables 11 to 19, inclusive, are presented. They are all from the experiments of Mr. Tower, made under the auspices of the Institution of Mechanical Engineers. In using 0.03 .H 0.02 u I 0.01 , •- _ Imperial 0. Oil «k ■^ o- RaieS Sperm 3ed Oil. Dil. oil ^ ^^ - — G— _ -v>-^ ^ -o— — ♦« 68 101 140 Temperature of Bearing t in Fahrenheit Degrees. 176 212 Conditions: Pressure on bearing, Q2 pounds per square inch; rubbing speed, 195 feet per minute; Bearing flooded with oil, about 0.8 quart per minute; journal of nickel steel, bushes of white metal, diameter about 10 inches, length about 4 1/2 inches. riG. 5. COTZFFICIENTS OF FRICTION FOR DIFFERENT LUBRICANTS (Laschc). 0.05 0.04 g 0.03 o S 0.02 ..... i \\ ■\ ■'\ V \S • Imperia Rape S€ 1 0. Oil 'ed Oil ^ s Spe rm( 3il ^^ *^ =^ ^=7:^ e=rr ■ — ■ ■■ _..J| 0.01 71 142 213 Pounds per Square Inch. Conditions: Rubbing speed, IQ5 feet per minute; temperature, 122 degrees Fahrenheit; bearing flooded with oil, about 0.8 quart per minute; journal nickel steel, bushings white metal, diameter about 10 inches, length about 4 1/2 inches. FIG. 6. coefficients of friction for different lubricants (Lasche). any values from these tables it must be clearly recognized for what kind and degree of lubrication they are applicable. Figs. 5 and 6 are reproduced from Traction and Transmission, January, 1903, and show results of Lasche's experiments. The "imperial oil" there mentioned is a Russian mineral oil of good quality for lubricating purposes. COEFFICIENTS OF FRICTION 33 Nominal load lb. per sq. in. Speeds 100 rev. 105 ft per min. 150 rev. 200 rev. 157 ft. 20; ft. per min. per min. i 250 rev. 262 ft. per min. 300 rev. 314 ft. per min. 350 rev. 366 ft. per min. 400 rev. 419 ft. per rain. Lb. 625 520 415 310 205 100 0.0013 1 0.00139 0.00147 0.00157 0.00123 0.00139 0.0015 0.00161 0.00123 O.OOIA-2 0.0016 ' 0.00176 0.00165 0.0017 0.00178 0.0019 0.002 0.00178 0.00334 0.00142 0.00205 0.00415 0.0016 0.00235 . 00494 0.00184 0.00269 0.00557 0.00207 0.00298 0.0062 D. 00225 0.00328 0.00676 0.00241 0.0035 0.0073 Conditions: A 4-in, journal of steel 6 in. long fitted on its upper side with a gun metal brass., or half box, embracing somewhat less than one-half the journal circumfer- ence — the chord of the arc contact equal to 3.92 in. — and when running kept at a constant temperature of 90° F. Table ii. — Coefficients of Friction for A Bath of Mineral Oil- Constant {Tower). -Temperature Nominal load lb. per sq. in. Speeds 100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. 450 rev. 105 ft. per min. 157 ft. per min,. 209 ft. per min. 262 ft. per min. 314 ft. per min. 366 ft. per min. 419 ft. per min. 471 ft. per min. Lb. 625 O.OOI 0.0012 0.0014 0.0014 0.0016 0.0018 0.002 520 0.0014 0.0016 0.0018 0.0019 0.002 0.0021 0.0022 415 0.0016 0.0019 0.0021 0.0023 0.0025 0.0026 0.0027 310 0.002 0.0022 0.0026 0.0029 0.0032 0.0035 0.00^8 0.004 205 . 0026 0.0034 . 0040 0.0047 0.0053 0.0058 0.0062 0.0066 153 0.0028 0.0038 . 0048 0.0057 0.0065 0.0071 0.0077 0.0083 100 0054 0.0076 . 0094 0.0109 0.0123 0.0133 0.0142 0.0151 Conditions: A 4-in. journal of steel 6 in. long fitted on its upper side with a gun metal brass, or half box, embracing somewhat less than one-half the journal circumfer- ence — the chord of the arc contact equal to 3.92 in. — and when running kept at a constant temperature of 90° F. Table 12. — Coefficient of friction for a Bath of Mineral Grease- Constant {Tower). 3 -Temperature 34 BEARINGS AND THEIR LUBRICATION Nominal load lb. per sq. in. Speeds 100 rev. 105 ft. per min. 150 rev. 157 ft. per min. 200 rev. 2og ft. per min. 250 rev. 262 ft. per min. 300 rev. 314 ft. per min. 350 rev. 366 ft. per min. 400 rev. 419 ft- per min. 450 rev. 471 ft. per min. ,Lb. 'S20 415 310 205 153 100 Seized ! 0.0015 O.OOII 0.0016 o.ooig 0.003 0.0017 0.0012 0.0018 .0.0023 . 0038 0.0018 O.ooig 0.002 0.0014 0.0016 O.OOI7 0.0021 0.0018 0.0025 0.0035 0.0061 0.0021 o.ooig 0.0027 0.0037 . 0064 0.0013 0.0016 0.0025 0.0021 0.0028 0.0044 0.0023 0.0030 0.0051 0.0024 0.0033 0.0057 Conditions: A 4-in. journal of steel 6 in. long fitted on its upper side w^ith a gun metal brass, or half box, embracing somewhat less than one-half the journal circumfer- ence — the chord of the arc contact equal to 3.92 in. — and when running kept at a constant temperature go° F. Table 13. — Coefficient of Friction for a Bath of Sperm Oil — Temperature Con- stant {Tower). N^nm 1 n a 1 Speeds load lb. per sq. in. 100 rev. 105 ft. per min. 150 rev. 157 ft. per min. 200 rev. 2og ft. per min. 250 rev. 262 ft. per min. 300 rev. 314 ft. per min. 350 rev. 366 ft. per min. 400 rev. 4ig ft. per min. 4^0 rev. 471ft. per min. Lb. 415 • 310 205 153 100 . ooog . 001 2 0014 0.0020 0.0027 0.0042 O.OOI 0.0014 0.0017 0.0023 0.0032 0.005 O.OOII 0.0015 0.002 .0028 0.0037 0.006 0.0013 0.0016 0.0022 . 003 I . 004 I . 0067 0.0015 0.0018 0.0025 0.0034 0.005 0.0076 0.0015 o.ooig . 0026 o.oo3g 0.0051 0.0081 0.0017 0.0029 0.0042 0.0052 0.009 0.0017 0.0022 0.003s Conditions: A 4-in. journal of steel 6 in. long fitted on its upper side with a gun metal brass, or half box, embracing somewhat less than one-half the journal circumfer- ence — the chord of the arc contact equal to 3.92 in. — and when running kept at a constant temperature of 90° F. Table 14. — Coefficients of Friction for Lard Oil Bath — Temperature Constant {Tower). COEFFICIENTS OF FRICTION 35 Speeds Temper- ature 100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. 450 rev. Fahr. 105 ft. 157 ft. 209 ft. 262 ft. 314 ft. 366 ft. 419 ft. 471 ft. per min. per mm. per mm. per mm. per mm. per mm per mm. per mm. 120° 0.0024 0.0029 0.0035 0.004 . 0044 0.0047 . 005 I 0.0054 110° 0.0026 0.0032 0.0039 0.0044 0.005 0.0055 . 0059 0.0064 100° 0.0029 0.0037 . 0045 0.0051 0.0058 0.0065 . 007 1 0.0077 90° 0.0034 0.0043 0.0052 0.006 0.0069 0.0077 0.0085 0.0093 80° 0.004 0.0052 . 0063 0.0073 0.0083 - 0.0093 0.0102 0.0112 70- . 0048 0.0065 0.008 0.0092 0.0103 Q.0115 0.0124 0.0133 60° 0.0059 0.0084 0.0103 0.0119 0.013 0.014 0.0148 0.0156 Conditions: A 4-in. journal of steel 6 in. long fitted on its upper side with a gun metal brass, or half box, embracing somewhat less than one-half the journal circumfer- ence — the chord of the arc contact equal to 3.92 in. — and when running kept at a constant temperature of 90° F. Table 15. — Coefficients of Friction for Bath of Lard Oil — Load Constant at 100 LB. PER Square Inch {Tower). Nominal load lb. per sq. in. Lb. 520 468 415 363 310 258 205 153 100 100 rev. 105 ft. per min. Speeds 150 rev. 157 ft. per min. 200 rev. I 250 rev. 209 ft. I 262 ft. 300 rev. 314 ft. I o . 0008 O.OOII j 0.0012 0.0013 ' 0.0015 0.0014 0.0017 0.0018 0.0021 0.0023 0.003 0.0036 0.0045 per mm. per mm.- per mm. o.ooi 0.0013 0.0014 0.0016 0.0017 0.002' 0.0025 0.0035 0.0055 350 rev. 366 ft. per min, 400 rev. 419 ft- 450 rev. 471 ft. per mm.! per mm. 0.0012 0.0013 0.0014 0.0015 0.0014 0.0015 0.0017 0.0018 0.0015 0.0017 0.0019 0.0021 0.0017 0.0019 0.002 0.0022 0.0019 0.0021 0.0022 . 0024 0.0023 0.0025 0.0026 0.0029 0.0028 0.003 0.0033 . 0036 0.004 0.0044 0.0047 0.005 0.0063 . 0069 0.0077 0.0082 0.0017 0.002 0.0024 0.0025 0.0027 0.0031 0.004 0.0057 o . 0089 Conditions: A 4-in. journal of steel 6 in. long fitted on its upper side with a gun metal brass, or half box, embracing somewhat less than one-half the journal circumfer- ence — the chord of the arc contact equal to 3.92 in. — and when running kept at a constant temperature 90° F. Table 16. — Coefficients of Friction for Bath of Olive Oil — Temperature Con- stant {Tower). 36 BEARINGS AND THEIR LUBRICATION Speeds load lb. per sq. in. 100 rev. 105 ft. per min. 150 rev. 157 ft. per min. 200 rev. 209 ft. per min. 250 rev. 262 ft. per min. 300 rev. 314 ft. per min. 350 rev. 366 ft. per min. 400 rev. 419 ft. per min 450 rev. 471ft. per min. Lb. 573 520 415 363 258 153 100 j i 0.00102 1 0.00055 1 . 00093 . 000S4 0.00107 : 0.00139 0.00162 ; 0.0020 0.00277 0-00357 0.00108 0.00105 0.00107 . 0096 0.00162 0.00239 0.00423 0. 001 18 O.OOII 5 0.00II9 O.OOII 00178 0.00267 0.00503 0.00126 0.00132 0.00125 0.00133 0.0013 :o. 00140 0.00122 0.00134 0.00195 0.00213 0.003 ,0.00334 0.00576 0.00619 0.00139 0.00142 0.00149 0.00147 0.00227 0.00367 . 00663 0.00148 0.00158 0.00155 0.00243 0.00396 0.00714 Conditions: A 4-in journal of steel 6 in. long fitted on its upper side with a gun metal brass, or half box, embracing somewhat less than one-half the journal circumfer- ence — the chord of the arc contact equal to 3 . 92 in. — and when running kept at a constant temperature of 90° F. Table 17. — Coefficients of Friction for a Bath of Rape Oil — Temperature Con- stant {Tower). Nominal load lb, per sq. in. Speeds Actual load lb. per sq. in. Temper- ature Fahr. 100 rev. ISO rev. 1 200 rev. 105 ft. 157 ft. 209 ft. per min. per min. per min. I 250 rev. 300 rev. 262 ft. j 314 ft. per min. | per min. 350 rev. j 400 rev. 366 ft. I 419 ft. per min. ; per min. 328 310 293 275 2S8 20s 153 100 582 551 520 498 4S8 364 272 178 0.0102 o.oios 0.C099 0.0105 0.0091 0.0112 O.OIOS 0.009 0.0099 O.OT07 0.0102 0.0099 0.0092 O.OIOS 0.0097 .0091 0.0095 0.009S 0.0088 0.0087 0.0085 0.0096 0.0102 0.0109 0.0122 0.0098 0.0099 0.0097 0.0103 0.0084 0.0078 0.0105 0.0133 0082 0085 0II9 0144 0.0083 O.OI O.OI2S 0.0154 Conditions: A 4-in. journal of steel 6 in. long fitted on its upper side with a gun metal brass, or half box, embracing somewhat less than one-half the journal circumfer- ence — the chord of the arc of equal to contact of 2.25 in. — and when running kept at a constant temperature of 90° F. Table i8. — Coefficients of Friction of Rape Oil from a Pad Under the Journal {Tower). COEFFICIENTS OF FRICTION 37 Actual Speeds Nominal load load 100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. lb. lb. 105 ft. 157 ft. 209 ft. 262 ft. 314 ft. 366 ft. 419 ft. per sq. in. per sq. in. per mm. per mm. per min. per mm. per mm. per min. per mm. Lb. Lb. 258 317 0.0056 0.0057 0.0063 0.0068 205 252 0.0132 0.0098 0.007 0.0077 0.0082 0.00S7 100 123 0.0144 0.0125 0.0146 0.0152 0.0163 0.0171 0.0178 Conditions: Four-in. journal, 6 in. long; chord of arc-3 1/4 in. Table 19. — Coefficients of Friction for Rape Oil Fed by Syphon Lubricator {Tower). COEFEICIENT OF FRICTION OF COLLAR BEARINGS The coefficient of friction for a collar bearing is much higher than for a journal bearing running under similar conditions. The reason lies in the difficulty with which the collar bearing is lubricated. There is little or no action tending to introduce oil between the surfaces as in journal bearings. Load in lb. total Speed in revolutions per minute 50 70 90 no 130 600 1200 1800 2400 2700 3000 3300 ^600 . 0450 0.0375 0.0357 0.0286 0.0354 0.0347 0.0337 0.0312 . 0646 0.0481 0.0399 0.0375 0.0334 0.0341 0.0322 0.0444 0.0433 0.0496 0.0361 0.0361 . 0346 0.0348 0.0348 0-0537 0.0489 0.0357 0.0373 0.0361 0.0352 0.0642 0.0475 0.0371 0.0410 0.0378 0.0356 Conditions: Bearing faces 12 in. inside diameter, 14 in. outside diameter, one of soft steel, the other of gun metal, face grooved to distribute oil; lubrication sufficient to prevent seizing with a mineral oil. Table 20. — Coefficients of Friction for a Collar Bearing. The coefficient of friction for collar bearing approaches that for journal bearings running at a low velocity and with poor lubrication, and tends to follow the laws for friction of solids rather than for fluids. It may be tciken as from 38 BEARINGS AND THEIR LUBRICATION 0.04 to 0.05. Experimental data tend to show that the coefficient is independ- ent of the speed of rubbing and decreases with an increase of load. At the same time this type of bearing is only usable for comparatively low unit pres- sures and for comparatively low speeds, unless the bearing surfaces are separated by a film of oil maintained by a pressure, or forced lubrication. Experiments made by Mr. Tower and reported in the Proc. Inst. M. E., 1888, page 179, were for a collar bearing 14 in. outside diameter, 12 in. inside diameter, having a soft steel face opposed to a gun metal face. Table 20 has been compiled from these experimental results. FRICTION OF STEP BEARINGS The friction of step bearings is about the same as of collar bearings. Step bearings are, in fact, no better than collar bearings except that they are generally of smaller diameters, hence have a lower speed and are therefore less liable to overheating. They present the same difficulties as regard lubrication. It is very difficult to assign an average value of the coefficient for a step bearing. However, it may be taken as o.oi for ordinary machine design. In the Proc. Inst. M. E., 1891, page 114, Tower gives results of experi- ments made by him with a fiat pivot bearing having a steel step opposed to a manganese bronze-bearing surface. Table 21 gives some of the results of those experiments. Speed in revolutions per minute Load lb. per Sc 128 194 290 353 sq. in. 1 Oil drops Coeflf. Oil drops Coeflf. of Oil drops Coeflf. of Oil drops Coeflf. of Oil drops Coeflf. of per min. of fric. per min. fric. per min. fric. per min. fric. per min. fric. 20 20 0.0196 79 0.0080 196 0.0102 Con- 0.0178 Con- 0.0167 40 26 0.0147 82 0.0054 stream . 006 I tinuous 0.0107 tinuous 0.0096 60 40 0.0167 80 O.OOS3 200 0.0051 stream 0.0078 stream 0.0073 80 so 0.0181 83 0.0063 200 0.0045 0.0064 0.0063 100 54 0.0219 98 0.0077 200 • 0.0044 0.0056 O.OOS7 t20 56 0.0221 84 0.0083 168 0.0052 . 0048 O.OOS3 140 90 84 0.0093 0.0113 158 168 0.0062 0.0068 0.0046 . 0044 .00053 O.OOS4 160 200 260 300 seized ' Conditions: Steel footstep running with a manganese bronze bearing diameter 3 m.; lubrication with a mineral oi. as indicated. Table 21.— Coefficients of Friction of Pivot Step Bearing. COEFFICIFNTS OE FRICTION 39 Another series of experiments was made with white metal in place of the l^ronze bearing; the coefficient of friction in this case was a trifle greater than in the former, but the difference was so small that the results may be looked upon as practically identical. This is to be expected when the friction is the friction of the lubricant and not of the metallic surfaces. COEFFICIENT OF LATERAL FRICTION The endwise floating of a level shaft, provided there is end play between the bearings and journal shoulders, is something that is frequently noticed. In the construction of electric motors and generators it is very common to leave considerable end play as this lateral motion of the shaft aids in distribut- ing oil throughout the bearings, and thus improves the quality of the lubrication. Anyone who has had occasion to touch a shaft under such conditions has prob- ably been surprised at the very small amount of pressure that is required to make the shaft move; in fact, the change in the level of the floor by shifting of the weight of a person walking around the machine is often sufficient to make the shaft move endwise. This motion can be likened to that of a fine pitch screw rotating and moving in a fixed nut, with the mating thread formed by the oil film. Professor Sweet in the American Machinist for February 16, 191 1, page 315, gives this explanation for the smallness of the force required to move a leveled and well lubricated shaft in a lateral direction. ''The Straight Line engines were, before the direct-connected age, built with a quarter of an inch end play in the main bearings and crank. When running, say at 250 revolutions, with 2-ton flywheels and 4 3/4-in. shafts, it was easy to slide the wheels and shafts back and fourth with a match. My explanation was this: In making, say three revolutions, the journals were sliding about 4 ft., and the match moved them 1/4 in. Thus, what the match has to do was to make up the difference between the base and the hypotenuse, where the base was 4 ft. and the perpendicular 1/4 in., which was very little, and as small as the match was, it was equal to do that litde work. " John Goodman, in a paper in the Proc. Inst. C. E., Vol. LXXXIX, page 424, discusses this question of lateral friction in an experimental apparatus, where a bronze half box composed of an alloy of i part of tin to 7 parts of copper, having a diameter of 2 in. and a length of 4 in., was run in contact with a manganese steel shaft. At a circurnferential velocity of 200 ft. per minute, and a lateral velocity of 0.05 in. per minute, the lateral coefficient of friction was found to be 0.000021. This factor is given here more as a matter of interest than because of any practical value. In general the lateral coefficient of friction varies directly as the lateral velocity, and inversely as the circumferential velocity. 40 BEARINGS AND THEIR LUBRICATION COEFFICIENT OF FRICTION FOR SURFACES SLIDING IN STRAIGHT PATHS The coefficient of friction for surfaces sliding in straight paths is not of much importance, as ordinarily guides, ways and V's of machines are not sub- jected to great pressures and do not move with high velocities. The condition of lubrication is that of unlubricated or very poorly lubricated surfaces. For metal in contact with metal under such conditions a safe factor of design for coefficient of friction is 0.3. For a more lengthy discussion of the subject reference should be had to Section I, where the experimental results of General Morin are presented. John Aspinwall in Proc. Inst. C. E., Vol. CXXXIII, page 13, gives results of a long series of experiments to determine the coefficient of friction of locomo- tive slide valves. The factors are: For the ordinary i)-valve with its face vertical, 0.068; for a jD-phosphor bronze partially balanced valve in a horizontal position, 0.0919; for an unbalanced cast-iron valve, 0.0878. SECTION III MATERIALS FOR BEARINGS Probably nothing in connection with bearings has had a wider discussion or a greater amount of experimental investigation than the materials from which they are made. All of this implies that it is a question of importance; yet its degree of importance is determined to a great extent by the nature and quality of the lubrication that the bearings will receive in use. The first law of friction for perfectly lubricated surfaces, page 19, is that the coefficient of friction for well lubricated surfaces is practically independent of the nature of those sur- faces. Thus, if the bearing surfaces can be kept separated by a film of oil the wear is the wear of the oil particles and not of the surfaces of the bear- ing and its journal. The principal requisites for the materials are that they should have sufficient mechanical strength to sustain the pressures to which they are subjected, and that they should tend to remain smooth and to conform to each other under the rubbing action, thus maintaining the conditions neces- sary for the formation of the oil film. But the builder of machinery has no control over the lubrication that the bearings of his machine will receive after they have been put into service, and it is well known they are often abused and neglected. Thus the conditions of ordinary lubrication fall into the class of poorly lubricated surfaces and the question of what kind of material to use assumes considerable importance, Metals are, of course, the common materials; although wood and fiber have limited applications. C. B. Dudley, late chemist to the Pennsylvania Railroad, states that a good bearing metal should have five characteristics. See The Journal of the Franklin Institute, February, 1892, page 83. These are: 1. The metal must have strength enough to sustain its load. 2. It must not heat rapidly. 3. It must work well in the foundry. 4. It must show a small coefficient of friction. 5. It must give a large amount of service with a small loss of metal by wear. To these five we might add a sixth: In general it should be dissimilar from the metal of the journal with which it is to run. This is broadly true, except for hardened steel in contact with hardened steel, and cast iron in contact with cast iron. There seems to be no entirely satisfactory reason for this dissimi- larity except it is claimed that wear is more rapid and friction greater when the 41 42 BEARINGS AND THEIR LUBRICATION journal and its bearing are made from the same metal. It is possible that adhesion has something to do with it. On page 5 it is pointed out that two metallic surfaces in intimate contact may be held together by a considerable force that we call adhesion or molecular attraction. The ordinary combinations of metals used for spindles or shafts and their bearings are as follows: Hard steel journals in contact with hard steel bearings, used for high speeds and moderate pressures. It calls for excellent workmanship, as both journal and bearing must be accurately ground and carefully alined. Under such conditions the service rendered is good. Hard steel journalswith soft steel or wrought iron bearings. This combina- tion is found in some presses where a hardened steel toggle pin fits into soft steel seats. It is generally considered to give good service for high pressures and low speeds. Hard steel journals with bronze bearings. This is a very common com- bination in machine tools. Hard steel journals with cast iron bearings. This combination is found to a limited extent in machine tools. The use of cast iron as a bearing metal has had a peculiar history, with strong advocates and just as strong opponents. It is of such importance that is is treated of in detail on page 43. Soft steel journals with bronze bearings. This combination is the most commonly used of any among machine tools and high-grade machinery and gives good results. Soft steel journals with babbitt bearings. This combination is the one most commonly used in machinery in general. Soft steel journals with cast iron bearings. This combination is used to some extent in machine tools and to a much greater extent in medium and light weight machinery where the bearing pressures are small and the speeds low. Soft steel journals and lignum-vitae bearings. This combination has been used for line shaft bearings in the past. Now it is used for the steps of vertical water turbines. Tail shaft bearings of steamships are a bronze sleeve against lignum vitae. Cast iron with cast iron. This combination is a great exception to the disuse of like metals. It is employed for the step bearings of vertical steam turbines, to a limited extent for spindles and bearings of machine tools, is com- monly used for the piston rings and cylinders of all kinds of reciprocating engines, has been used in the crossheads and slides of steam engines and air compressers, and is the combination commonly found in rectilinear slides and guides of all kinds of machinery. MATERIALS FOR BEARINGS 43 WEAR OF BEARING METALS The wear of bearing metals is of importance in railroad practice for it represents an enormous loss of valuable metal each year in the amount rubbed off and lost. In electrical machinery, such as motors and generators, the wearing of the bearing tends to decenter the armatures and rotors and thereby unequalize the air gap to the disadvantage of the operating of the machine. In any class of machinery where the alinement of the shafts is of importance, the wear of the bearing is likewise of importance, for this determines the position of the shafts. Dr. Dudley in the Journal of the Franklin Institute^ March, 1892, page 169, lays down three elements in the wear of bearing metals. These are: 1. That metal which will suffer the most distortion without rupture will wear best. Or, stated a little differently: That metal which has the greatest percentage of elongation will wear the best. 2. With a satisfactory elongation an increase in tensile strength will add to the wearing property of the metal. 3. With equal percentages of elongation and equal tensile strengths the finer the granular structure of the metal the longer will it wear. The idea of wear is the rubbing off or rupturing of minute particles. Thus a metal which has the greatest tensile strength to resist rupture and the property of withstanding the greatest amount of distortion without rupture and a fine granular structure, so that the particles that are rubbed off are minute, can be considered as having the greatest resistance to wear. There are few experi- mental data to support the relationship between these variables, although the conclusion seems to be well founded that an increase of tensile strength at the expense of elongation is decidedly to the detriment of the wearing qualities. In following paragraphs dealing with bronzes this question of wear will be again considered in quantitative factors. CAST IRON It is emphasized by its advocates and conceded by its opponents that cast iron under certain conditions will take on a hard glaze that makes a splendid wearing surface. This glaze is so hard that it will resist the cutting edge of a scraper. Furthermore, cast iron is porous and readily absorbs oil, thus to a degree becoming self-lubricating. It is granular in structure, in wear it rubs off in minute particles. It can be obtained in varying degrees of hardness, even chilled. Other advantages are the ease with which it is handled in the foundry and machine shop, and its cheapness. The objection to it is summed up by saying that it is a 'treacherous" metal. After glazing the bearing is very permanent and has great resistance to cutting; 44 BEARINGS AND THEIR LUBRICATION but before glazing, if cutting is once started it takes but a very short time to ruin the bearing. With an assurance of adequate lubrication, the case against cast iron as a bearing metal would at once break down. Professor Sweet has long been an earnest advocate of the cast-iron box, as frequent articles and discussions in the American Machinist show. William Sellers & Company, of Philadelphia, has been furnishing cast-iron lineshaft bearings for years. R. K. LeBlond in the American Machinist, March 23, 191 1, page 537, gives experiences with four experimental lathes fitted with spindles and bearings as follows: 1. Hardened steel spindle with cast-iron boxes. 2. Soft steel spindle with babbitt boxes. 3. Hardened steel spindle with bronze boxes. . 4. Soft steel spindle with bronze boxes. The soft steel spindles were of 60-carbon crucible steel; the bronze was made to the specifications of the Pennsylvania Railroad Company. After the end of some 8 years' service and treatment as far as possible identical for all four lathes, it was found that their rating as regards absence of wear and general satisfaction was in the order as given above; that is, the hardened steel spindle with cast-iron boxes was the best combination. Both spindle and boxes were in as good condition as when placed in the lathe, and from all appearances and tests showed absolutely no wear. Mr. LeBlond says further: "The dry bronze box with soft spindle will stand more abuse than a dry cast-iron box and soft spindle; but a dry cast-iron box and hard spindle will withstand more abuse than the bronze." And again: ''The hardened steel spindle and cast-iron box will stand as much neglect as any combination of metals, has a much longer life, will retain its accuracy for an indefinite period, will stand intermittent cuts or series of blows which would peen out and loosen babbitt, and, from our experience, is positively the best and most lasting bearing ever put into a machine tool." The experience of the John Steptoe Shaper Company is shown by the follow- ing quotations taken from an article of Professor Sweet's printed in the American Machinist. " We have been using cast-iron bearings on all our machines since July, 1909. Since that time we have turned out over 500 machines, and we have yet to re- ceive the first complaint with regard to the bearings running hot; but we have made proper provision for the distribution of the oil over the bearings. We have chased spiral oil grooves in the shafts and have provided the bearings with ring oilers so that they are constantly flooded with oil. We believe that the cast-iron box is the best bearing that we have ever put in our machine." Referring to Professor Sweet's own work, he says: ''In the artisan lathe J MATERIALS FOR BEARINGS 45 put a ground steel shaft in a solid cast-iron box. The outside of the box fitted in a conical piece with a nut at each end, but was not split, so when they first became loose the boxes wdll compress enough without splitting, and when not enough they can be split." Cast iron bearings have been extensively used in shoe machinery for many years, and in textile machinery. One of the most troublesome places in the machine shop is in countershaft bearings. It has long been accepted practice in many places to bush countershaft hangers and loose pulleys with cast iron when the original bearings wore out. Coleman Sellers, Jr., writes of the practice of his firm in regard to the use of cast-iron bearings and spindles as follows: " William Sellers & Company have always believed in cast-iron bearings where the condi- tions of feed and pressure are not excessive and where efficient lubrication can be insured. Under such conditions we believe that a film of oil separates the metallic surfaces so that they do not come in contact. Where the load or speed are such that the film of oil is likely to be disturbed then it becomes necessary to use some material in the bearing not so likely to adhere or 'cut.' It has always been our practice to use cast-iron boxes in our own shafting, and we have had cases where boxes have run for 30 years, not only without cutting but without appreciable wear. We have built many machines with cast-iron spindles and cast-iron bearings and with steel spindles in cast-iron bearings both with excellent results. We realize that such bearings must be thoroughly lubricated. They will not run well if dry. We have carried heavy vertical shafts on cast-iron steps with perfect results; but care has been taken to make these steps so large that the unit pressure would be low and the oil circulation was made very effective. Boring mill tables, planer tables, boring mill center spindles and boring mill steps are all emaples of cast-iron running on cast-iron, and it is well understood that under proper condi- tions of load, speed and lubrication such bearings are perfectly satisfactory." In the Proc. Inst. C. E., Vol. LXXXV, in a discussion by Dr. Goodman, it is stated that on the Brighton Railroad in England cast iron eccentric straps working on cast-iron hubs have run at least 100,000 miles without having been taken up. These few quotations are sufficient to show the esteem in which cast iron is held as a bearing metal, provided it is used under advantageous conditions. The common prejudice against it is unfounded, but one must not lose sight of the fact that a cast iron bearing will not run dry. SOFT ALLOYS OR WHITE METALS Babbitts are white bearing metals whose most common constituents are tin, lead, and antimony. As distinguished from a brass or bronze, babbitt metal can be melted in an ordinary ladle. This is one of the governing rea- sons for its use as a lining for journal bearings. It is easily melted, easily 46 BEARINGS AND THEIR LUBRICATION poured and run, easily anchored in place in its shell, has good anti-friction properties, and may be poured around the journal with which it is to run (although this is not approved practice). Thus the babbitting process is the easiest method known for obtaining well alined bearings. B. R. Tompkins, writing in the Mechanical News for January, 1891, says: "For slow running journals, where the load is moderate, almost any metal that may be conveniently melted and will run free will answer the purpose. For wearing properties, with a moderate speed, there is probably nothing superior to pure zinc, but when not combined with some other metal it shrinks so much in cooling that it cannot be held firmly in the recess, and soon works loose; and it lacks those anti-friction properties which are necessary in ordar to stand high speed. ''For line shafting and all work where the speed is not over 300 or 400 r. p. m. an alloy of 8 parts zinc and 2 parts block tin will not only wear longer than any composition of its class, but will successfully resist a heavy load. The tin counteracts the shrinkage, so that the metal, if not overheated, will firmly adhere to the box until it is worn out. But this mixture does not possess sufficient anti-friction properties to warrant its use for fast running journals. "Among all the soft metals in use there is none that possesses greater anti- friction properties than pure lead; but lead alone is impracticable, for it is so soft that it cannot be retained in the recess. But when by any process lead can be sufficiently hardened to be retained in the boxes without materially injuring its anti-friction properties, there is no metal that will wear longer in light fast running journals. With most of the best and most popular anti-friction metals in use and sold under the name of 'babbitt metal,' the basis is lead. "Lead and antimony have the property of combining with each other in all proportions without impairing the anti-friction properties of either. The antimony hardens the lead, and when mixed in the proportion of 80 parts of lead by weight with 20 parts antimony, no other known composition of metals produces greater anti-friction or wearing properties, or will stand a higher speed without heat or abrasion. It runs free in its melted state, has no shrink- age and is better adapted to light high speed machinery than any other known metal. Care, however, should be manifested in using it, and it should never be heated beyond a temperature that would scorch a dry pine stick. "Many different compositions are sold under the name of babbitt metal. Some are good, but more are worthless. But very little genuine babbitt metal is sold that is made strictly according to the original formula, most of the metals sold under that name are the refuse of type foundries and other smelting works, melted and cast into fancy ingots with special brands, and sold under the name of babbitt metal, " It is difficult to determine the exact formula used by the original discoverer, MATERIALS FOR BEARINGS 47 Babbitt. Tin, copper and antimony were the ingredients, and from the best sources of information the original proportions in per cent, were as follows: Tin =89.3 or 83.3 or 89. 1. Copper= 3.6 or 8.3 or 3.7. Antimony= 7 . i or 8 . 3 or 7.4. This metal, when carefully prepared, is probably one of the best metals in use for lining boxes that are subjected to a heavy weight and wear; but for light fast running journals the copper renders it more susceptible to friction. BRONZE BEARING METALS Next in importance to the babbitt is the bronze series of bearing metals. A bronze is commonly taken to mean an alloy of copper and tin with small amounts of other constituents. A brass is an alloy in which copper and zinc are the principal constituents. Detail showing Joint Cut Oue Lug of Packing Ring. '" ««ch Seg- meut thus for Liftiug. FIG. 7. BEARING METAL APPLIED TO A STEAM ENGINE PISTON. In contrast to the softer metals, they must be melted in a crucible and cannot be manipulated in the easy manner in which the soft metal linings are made. As in the case of babbitt metals there are a multitude of bronzes on the market. In an intermediate position between the babbitt series and the bronze series is a metal made by A. Allan & Son, known as Allan "red metal." It is composed of 50 per cent, copper and 50 per cent. lead. It is a soft metal, red in color, will withstand a temperature of 575° F. without injury, may be poured 48 BEARINGS AND THEIR LUBRICATION into a lining shell, but not around a metal mandrel, and is extensively used for the bearing rings of reciprocating steam engine pistons. Fig. 7 shows the way in which it is used for this purpose. A similar metal is ''Plumbic Bronze." In the preceding paragraphs it was pointed out that lead is one of our most useful bearing metals when used with others to make an alloy. Two theories have been advanced as to the relative action of the constituents. The one is that the harder metal which forms a net work around particles of the softer metal is the important one. That is, it is the harder metal that takes the wear and determines the frictional value of the alloy; the softer metal merely fills up the spaces throughout the network of the harder matrix, and wears away more rapidly in order that the matrix may always be in contact with the journal. The other theory is that the softer metal is the one that determines the anti- friction qualities of the alloy. This is the explanation that is more generally accepted, and in view of the fact that we know that lead is a most valuable con- stituent in these alloys, we are still further led to believe that it is the soft metal that is determinative. Andrew Allan, of the firm of A. Allan & Son, was the first manufacturer of bronzes in America to point out the value of lead in large proportions as a valu- able constituent of copper-lead-tin bearing metal. He has made such alloys since 1876. He has standard bronzes with different proportions of cop- per, lead and tin, for different services, as indicated below: Number Copper in per cent. Lead in per cent. Tin in per cent. Uses 2 66 25 9 For severe service as mill pinion bearings. 4 59-5 35 5-5 For locomotive brasses. 5 50-5 45 4-5 For passenger and freight car brasses. 6 48.75 48.75 2-5 For service similar to that of hard babbitt. Name of metal Copper Tin Lead 1 Anti- mony Zinc Iron Camelia metal 70.2 1.6 4.2s 98.13 X4.75 10.2 0.55 White metal 87.92 ' 84.87 I.IS 12.08 15-1 Car brass lining trace 9.91 Salgee metal 4.01 85.57 ■ Table 22. — ^Analyses of Common' Bearing Metals {Dudley). MATERIALS FOR BEARINGS 49 Name of metal Copper i Tin 1 Lead | And- Zinc ! Iron 14-38 Graphite — 67-73 none 80.69 14-57 possible trace 12.4 — trace S-i 83-55. on, copper, 78.44 0.31 15. c6 12.52 16.73 18.83 tcrmined Carbon bronze 7 "C A^ 9.72 Carbon— 9.6 Phosphorus 2.37 77.83 92.39 trace Delta metal Llagnolia metal 16.45 zinc and pc 19.6 American anti-friction metal.... 59 7C.8 Traces of ir bismuth. 2.16 9.2 10.6 9.58 Manganese 10.98 Phosphorus ssibly 0.98 38.4 0.65 Graney bronze I.Ianganese bronze 90.52 81.24 Ajax metal —none 7.27 or arsenic 88.32 0.37 11.93 Harrington bronze 55-73 0.97 42.67 trace 0.68 84.33 94-4 9.61 — 0.94 1 5 14-38 6.03 Hard lead Phosphor bronze 79-17 76.8 10.22 Phosphorus 8 Phosphorus Ex. B metal . — 0. 2 Table 22 {Continued). — Analyses of Common Bearing Metals (Dudley). ANALYSES AND PHYSICAL PROPERTIES OF BEARING METALS Dr. Dudley, in The Journal of the Franklin Institute, February, 1892, page 87, gives a number of analyses of common bearing metals, mostly bronzes as determined in the laboratory of the Pennsylvania Railroad. Table 22 has been compiled from the results there given. G. H. Clamer also in The Journal of the Franklin Institute gives a large number of analyses of bearing metals from which Tables 23-26 have been compiled. These are of value as showing a wide variety of both American and European practice. 50 BEARINGS AND THEIR LUBRICATION Tin Copper Antimony References 96 4 8 Quoted by Thurston. Used for ordinary bearings. 90 2 8 Quoted by Thurston. Quoted by Hiorns for bearings heavily loaded , used by Russian railroads for car bearings. 83.8 3-7 7-4 Quoted by Thurston and BoUey, as Karmarsch metal. Used in France in naval constructions. 87.5 87 12 5 Quoted by Thurston, Karmarsch metal. 6 7 Quoted by Hiorns, for bearings heavily loaded. 8S 5 10 Quoted by Ledebur and Hiorns as Jacoby metal for light pressure. 83.33 S-SS II . II Used for car bearings, "Compagnies de I'Est, P. L. M., Quest," etc. 8] 6 II Quoted by Ledebur. Used by Berlin railroads. 82 6 12 Quoted by Ledebur. Used by Orleans and the Western Austrian railroads. 82 8 10 Bearings for valve rods and eccentric collars, " Compagnie du Nord." 81 5 14 Quoted by Hiorns, for very hard bearings. 80 10 10 Quoted by Thurston. Used by Swiss railroads. 78. s 10 II-5 Quoted by Thurston. Used by Russian railroads. 76.7 7.8 15. 5 Quoted by Ledebur and Thurston as English alloy. 76 7 17 Quoted by Hiorns, for bearings lightly loaded. Quoted by Thurston and BoUey, as Karmarsch metal. 73 9 18 Quoted by Thurston and Hiorns, for light pressures. 71.4 21.4 7.2 Quoted by Thurston. Karmarsch metal. 71 S 24 Thurston standard white metal. Used by the P. L. M. Company for packing of valves and eccentric collars. 67 22 II Quoted by Thurston. Used by the Great Western Railway (England) 67 n 22 French state railroads. 33-3 22.2 44.5 Quoted by Hiorns. Dewrance metal for locomotives. 12 4 82 Quoted by Hiorns, for very hard bearings. Table 23. — Alloys of Tin, Copper and Antimony (Clamcr). Lead Tin Antimony 85 .s 84 16 80 12 8 77.7 5-9 16.8 76 14 10 73 12 15 70 20 10 68 IS 17 60 20 20 42 46 12 42 42 16 37 38 25 References. Soft alloy quoted by Hiorns. Quoted also by Dudley. Quoted by Ledebur. For slow revolving pulleys. Used by Eastern Railroad (France) for metallic packings. Quoted by Thurston, as being the composition of Magnolia and Tandem metals. Used for metallic packings by the Orleans and P. L. M. Railroads Metallic packings of piston rods. Northern Company (France). Metallic packings of eccentric collars. French state railroads. Graphite ( ?) metal analyzed by Dudley. Quoted by Ledebur and used for railroad bearings. Quoted by Hiorns. Hoyles metal. Quoted by Ledebur. Used for journal boxes, French state railroads. Quoted by Thurston. Italian railroad companies. Table 24. — Alloys of Lead, Tin and Antimony (Clamer). MATERIALS FOR BEARINGS 51 Copper Tin Zinc References 86 14 84 14 ! 2 83 IS 3 82 x8 82 15 ! 3 82 16 2 80 18 ! 2 78.7 6 3 IS 78 20 2 77-4 15 6 7 S8 28 14 57 14 29 56 28 16 6 14 i 80 S-S 17 S 77 Quoted by Thurston for locomotive bearings. Quoted by Thurston. Italian railroads. Quoted by Bolley as a hard alloy. Quoted by Thurston for locomotive bearings. , Used by French State railroads for pieces subiectcd to alternative friction. "Lafond" alloy for heavily loaded bearings. Quoted by Ledcbur and Thurston. Car bearings of the " Compagnie du Nord." Used by Orleans Railroad for valve-rod bearings. Quoted by Ledebur for locomotive bearings. Used by French State railroads for pieces subjected to circular friction. Quoted by Thurston and Bolley as "Lafond" alloy. Quoted by Bolley for friction upon cast iron. Quoted by Bolley and Thurston for car bearings. Quoted by Haswell as a hard bronze for bearings. Quoted by Thurston as "Margraflf" alloy. Quoted by Hiorns for bearings for propclloi shafts. Quoted by Thurston, "Fenton" alloy. Quoted by Hiorns and Thurston. "Fenton" alloy for locomotive and car bearings. Quoted by Ledebur, for high-speed horizontal shafts. Table 25. — Alloys of Copper, Tin And Zinc {Clamer). Copper Tin Lead Zinc Iron Anti- mony Phos- phorus References 10 65 25 Bearings for locomotives and tenders 74 . 1 1 7 95 5 83.3 76.2 52 -■JO Quoted by Thurston and Ledebur Loco- 8 7.6 I7-S 46 15 32 8 3 0. T 25 60 motive bearings. Quoted by Bolley and Ledebur. Bearings for engines. Quoted by Ledebur as "Pierrot" metal. " Beugnot" white bronze used in France in naval constructions. "White bronze used for ship engines. Dunnlevic and Jones metal. Quoted by Ledebur, "Kniess" metal. Used by the western railroads (France) for piston rods and eccentric packings. 8.3 5.6 3.8 1.6 0.4 5 3 Table 26. — Miscellaneous Alloys {Clamer). An increasing amount of attention is being given to the physical properties of bearing metals; the tensile and compressive strengths, elongation, reduction of area and Brinell hardness number. It is recognized that a suitable chemical analysis alone is not sufficient to ensure physical characteristics that will give a good alloy in service. From the records of tests in the laboratory of the Pennsyl- 52 BEARINGS AND THEIR LUBRICATION vania Railroad Table 27 has been arranged. This gives the physical proper- ties of many commercial alloys. Table 28 give representative chemical analyses of many of these same alloys, but the two tables must not be connected directly. For small variations in the con- stituents of such metals are common, while the effect of such variations cannot be stated definitely it is possible that small amounts of some elements may have a very decided influence upon the physical properties of the alloy that they enter. This has led some firms to insist that nothing but ingot metal of known com- position shall be used either in melting babbitts or in charging crucibles for bronze mixtures. In no case can scrap metal or the fins, or sprues, or defective castings from previous pourings be remelted. All such metal is sent to a refining room where it is analyzed and used to form the standard mixtures that are issued in the form of ingots to the babbitting room and bronze foundry. This case is exercised to keep the mixtures uniform, standard and under control. In preparing the ingot metal, chemical tests alone are not relied upon, however, but the physical properties are likewise investigated. Considerable weight is given to the Brinell hardness number, although in some place the sclerescope hardness number is the one found and considered instead. In this connection it is interesting to note the same bronzes will give a sclere- scope hardness number as high as high-grade alloy steel. This is probably due to the elasticity of the metals, for the sclerescope measures the recupera- tive power of the specimen tested when a permanent deformation is caused by the impact of the hammer. Sclerescope readings for different metals should not be compared. Comparisons of Brinell and sclerescope hardness numbers for the same specimens indicate that there is some relationship between them, though just what has not been accurately determined. In testing babbitts for hardness, and in fact other physical properties, care must be exercised to use specimens of the same general size and cast under similar conditions. Babbitt poured in a very thin section against a metal shell will tend to be harder than if poured in a thicker section. Turning again to Table 27, two striking features are: first, the wide varia- tions of the hardness numbers, and second, the increase in hardness after a compress in of 1/8 inch of the specimen. White metals run from 18 to about 35 Brinell hardness number, the highest being for Souther babbitt. The bronzes run from 40 to 106 Brinell hardness number, the latter being for a maganese bronze. This seems to be an unusually high number for a second determination only gives 68. This latter is closely approached by two of the Cramp's mixtures and exceeded by two other bronzes. Plumbic bronze, which is a half-and-half mixture of lead and copper, has a Brinell hardness number of 21.8, thus being in the range of the babbitts. MATERIALS FOR BEARINGS 53 vo Tf ^ ro vj- 't (^ rr> P..S u JD -q :=! H rt rt S, ^ c Jf| K^ S o o -E m M Kl ' W : VD O O O 00 On Ov ro *0 NO >n lO \0 O vO O lO ro r^ M o o o o o O O vO to o NvOnOnO M vow cm toOO NO NO lO 00 O O O NO t^ ■O00MMON' Q "o J? iS rt .a 4^ a; .t:: •r; Ta "O -O X P c c c ^ O rt rt rt rt SQQQn { 5s 54 BEARINGS AND THEIR LUBRICATION Composition in per cent. Copper Lead Tin Antimony Nickel Phosphorus Sulphur Zinc 64 79.7 64.75 SO 2.25 60.67 3-7 2.5 7 5.55 30 95 30 SO O.IS 32.97 0.25 5 10 5 I Phosphor bronze Cyprus bronze 0.8 0.2s Parsons white brass 64.9 4.6 88.89 58.38 84 88.33 32.9 Demo bronze 2. 1 Standard babbitt 7.41 Shonberg M. M. metal. . . Souther babbitt 38.93 9 II . II Table 28. — Analysis of Bearing Metals {Pennsylvania Railroad Laboratory). The variation in babbitt metals purchased on brand only has been pointed out on page 46. Thus a careful purchaser will buy on specification, not on brand merely, and at the same time will watch the phyvsical properties of the metals furnished. A similar course will be followed by all who are careful about the bronzes used for bearings. Name of metal Composition in per cent. Copper Tin Lead ■ Phos- phorus Arsenic Phosphor bronze ' 79 . 7 10 Copper tin (ist experiment) ; 87 . 5 12.5 Copper tin (2d experiment) same Copper tin (3d experiment . ) same Arsenic bronze (ist experiment). . 89.2 10 Arsenic bronze (2d experiment) ... 79.2 10 Arsenic bronze (3d experiment). . . 79 . 7 10 "K" bronze (ist experiment). ... 77 10.5 "K" bronze (2d experiment) same Alloy ''B" 77 8 9-5 o» -. . . . o.o3 7 o. oS 9-5 ooS 12.5 15 Relative wear in per cent. 100 148 153 147 142 115 lOI 92 92. 86. The physical properties of the phosphor bronze are tensile strength 30,000 pounds per square inche, elongation 6 per cent. Physical properties of the alloy "B" are tensile strength 24,000 lb. per square inche, elongation 1 1 per cent. Table 29. — Relative Wear of Various Bronzes {Dudley). MATERIALS FOR BEARINGS 55 WEAR OF CAR BRASSES Dr. Dudley gives the wear of car brasses as a loss of i lb. of metal for every 18,000 to 25,000 miles run. From his experiments to determine the relative wear of different alloys in actual railroad service Table 29 has been arranged. Referring to the last column it is seen that the relative wear in per cent, decreased steadily with an increase in the proportion of lead, and that the wear of the alloys containing no lead was far greater in every case than those contain- ing that metal. Copper Tin Lead Wear in grams 8^76 14. no . 2800 90 95 90 6.7 01 9 4. 45 95 62 0.1768 0.0776 0.0542 82 4 4.82 85 12 4 64 10.64 0.0380 81 27 5 17 14.14 0.0327 68 71 5 24 26.67 0.0204 64 34 4 70 31.22 0.0130 Table 30. — Relative Wear of Copper-tin and Copper-tin-lead Alloys {Clamer). Copper Tin Lead Zinc Wear in grams 85.12 4.64 5-28 4-71 5.62 4.68 10.64 10.25 10.30 11.42 10. 6i 0.0380 0.0415 0.0466 . 047 2 0.0846 82.27 79.84 77-38 74.28 2.07 5-44 6.54 II .04 Table :elative Wear of Copper-tin-lead-zinc Alloys (Clamer). Tables 30 and 31 are from Clamer's experiments. The first shows plainly the decrease of wear with the increase of the proportion of lead, while the second shows the increase of wear with the increase of zinc. Recently compiled data of the wear of tender truck journal bearings are given in Table 32. S6 BEARINGS AND THEIR LUBRICATION Kind Size in. Load in lbs. per sq. in. projected area Wear per bearing in lbs. per 1000 miles Kind of tender Plain or filled bearings Composition Plastic bronze 4 1/4x8 t^t 0.01343 1 Plain PI a stir Rrnn7.f> Phosphor bronze. . 4 1/4x8 383 0.01700 Plain Copper 64% Lead 30% Tin 5% Nickel 1% 4 1/4x8 383 0.01200 Plain Phosphor bronze. . 4 1/4x8 383 0.01814 riain Cyprus bronze. . . . 4 1/4x8 383 0.01600 Plain Copper 79.70% Lead 9.50% Tin . 10 00% Phosphor bronze. . 4 1/4x8 383 0.02043 Plain Phosphor 0.80% Cyprus Bronze Plastic bronze. . . . 5 1/2XIC 395 0.02159 5500 gall. Plain Standard filled . . . S 1/2XIC 395 0.02348 5 5 00 gall. Filled Copper 64.75% Lead 30.00% Tin 5.00% Sulphur 0. 25% Standard filled . . . S 1/2XIC 395 0. 02431 5 5 00 gall. Filled Lightened filled. . . 5 T/2XTC 395 0.02437 5500 gall. Filled Standard filled . . . 5 I/2XIC 420 0.03542 7000 gall. Filled Filling Metal Lead 87% Antimony 13% Lightened fillod. . . S 1/2x10 429 0.03572 7000 gall. Filled 4 1/2x8 inches used in both passenger and freight service. 5 1/2x10 inches 5500 gallon tenders — passenger service. 5 1/2x10 inches 7000 gallon tenders — freight service. Table 32. — Wear of Tender Truck Journal Bearings (Pennsylvania Railroad Laboratory). REPRESENTATIVE PRACTICE IN THE USE OF BEARING METALS The bearing metal known as the standard of the Bureau of Steam Engineer- ing of the United States Navy, also called "anti-friction" or "anti- attrition" metal, has this composition: Best refined copper 3.7 per cent. Banca tin 88 . 8 per cent. Regulus of antimony 7.5 per cent. The percentages are by weight. The mixture must be well fluxed with borax and rosin in mixing. The mixing of this anti-friction metal is a trick which must be learned. The best practice is to melt the copper, tin and antimony separately, adding the tin to the copper and the antimony to this mixture, fluxing it with borax with the proportion of about i 1/2 lb. to 175 lb. of the mixture; but satisfactory MATERIALS FOR BEARINGS 57 results are obtained by melting the copper first, dropping the cold tin into the melted copper and adding the antimony, which has been separately melted. This metal is carefully skimmed before pouring, and is poured into pigs and carried into stock as it stands. The journal bronze used on battleships of the United States Navy has this composition: Copper 82 to 84, tin 12.5 to 14.5, zinc 2.5 to 4.5, iron (max.) 0.06, lead (max.) i.oo, all in per cent., with a normal of 83-13 1/2-3 1/2. It is used for bearings, bushings, sleeves, slides, guide gibs, wedges on water- tight doors and all parts subject to considerable wear. The anti-friction metals commonly specified by the United States War Department are Parson's White Brass, Genuine Babbitt, Magnolia Metal, Phoenix Metal, and Shonberg's "M.M." Parson's White Brass from one analysis is composed of: tin 64.90, lead 0.15, copper 2.25, zinc 32.93; Shonberg M.M. white bronze: tin 38.38, zinc 38.93, copper 2.5, lead 0.25. This comparison of the physical properties of three of these metals is taken from the Journal of the American Society of the Naval Engineers, Volume XIX. Naval white brass Parson's white brass Magnolia metal Original diameter, inches height, inches. Elastic limit (about) , lbs Final compression loads, lbs. . Final height, inches 0.7S 0.75 0.7S9 0.758 0.751 8,000 8,000 5.000 5,000 5,000 36,000 37.000 35. 000 36,800 37.000 0.2 0.192 0.180S 0..8 0. i6r 0.748 4,500 36,500 0.166 The Westinghouse Electric & Manufacturing Company has developed most careful practice in preparing and handling bearing metals. In general, two kinds of babbitt are used ; one having a tin base and the other a lead base. Not only are the proportions of the alloys carefully watched, but also the physical properties. A standard test block to determine hardness is in the form of a disk 2-1/4 in. in diameter, 3/4 in. thick, and surf aced . both sides. When tested by the Brinell method, the genuine babbitt shows a hardness number of about 25; similarly, the softer gives a number of about 21. Care is used not to overheat the metals, for the harder the pouring tempera- ture should not exceed 890° F.; the corresponding minimum temperature is about 660° F. In practice an effort is made to keep the pouring temperature about midway of there limits. For the softer metal the corresponding temper- atures are 1000 and 840° F. This alloy is more easily abused by overheating than the harder one. In general the softer metal is used for the regular run of work and the harder in railway motor bearings and bearings subjected to pound. Experience has 58 BEARINGS AND THEIR LUBRICATION shown that for conditions of nearly steady pressure the softer metal is in many instances preferable to the harder one, being less liable to overheating in service. Albert E. Guy gives the composition of the high-speed babbitt used in De Laval steam turbines as: copper lo, tin 80, and antimony 10 per cent. For low speeds the metal used is: lead 77, tin 6 and antimony 17 per cent. RAILROAD PRACTICE Representative practice of the Pennsylvania Railroad Company is as follows: For all axle bearings of passenger and freight cars, antimonial lead lining on ExB bronze shells; for locomotive driving wheel axle bearings, crank pin bear- ings, cross-head pin bearings, and small bearings generally, ExB, phosphor-, or plastic bronze; for cross-head shoes, block tin; for four-wheel engine truck boxes, standard babbitt strips set in bronze. The specifications for journal bearing lining metal calls for a homogeneous alloy of lead and anitmony as free as possible from every other substance, and of the following composition: Lead 87 per cent., antimony 13 per cent. The rejection limits by analysis are: Lead less than 86 and more than 88 per cent., tin and antimony less than 12 and more than 14 per cent., tin more than 2 per cent., copper more than 0.5 per cent., or if the sum of the amounts of lead, tin, antimony and copper is less than 99.5 per cent. The specifications for phosphor-bronze bearing metal call for a homo- geneous alloy of copper, tin, lead, phosphorus, as free as possible from every other substance, and of the following composition: Copper 79.7 per cent., tin 10 per cent., lead 9.5 per cent., phosphorus 0.8 per cent. The rejection limits are as follows: Lots will not be accepted if the analysis as above described gives results outside the following limits: Tin, below 9 per cent, or over 11 per cent.; lead, below 8 per cent, or over 11 per cent.; phosphorus, below 0.7 or over i per cent. ; nor if the metal contains a sum-total of any other substances than copper, tin, lead and phosphorus in greater quantity than i per cent. The specification for standard babbitt calls for a homogeneous alloy of tin, antimony and copper, having the following composition in per cent.: tin 88.89, antimony 7.41, copper 3.7. The ExB bronze analysis is: Copper 76.75 per cent., limits 75 -75 to 77 .75 Lead 15 per cent., limits 13.5 to 16 . 5 Tin S per cent., limits 7 to 9 Phosphorus 0.25 per cent., limit .' not below 0.20 Impurities, limit not above 0.75 Robert Garbe, in Die Danpflokomotiven der Gegenwart, page 345, gives the composition of a white bearing metal that has been used successfully for 50 years by the Prussian State Railways; during that time it has remained un- MATERIALS FOR BEARINGS " 59 changed. It is used not only for the axle bearings of locomotives, but also in both freight and passenger cars. It is made of copper, antimony and tin, in the following manner: i kg. (2.2 lb.) of copper is melted together with 2 kg. (4.4 lb.) antimony (regulus) and 6 kg. (13.2 lb.) of Strait's tin. The antimony is added when the copper is melted, and when both metals are fluid, the tin. This alloy is cast in thin plates and each 9 kg. (19.8 lb.) is remelted with 9 kg. (19.8 lb.) of pure tin. The whole is then cast in plates 15 mm. (0.6 in.) thick and is therewith ready for use. Larger quantities than those mentioned above ought not to be melted together at one time. The individual constituents of the alloy must be as pure as possible. Thus the antimony should contain at the most i per cent, of impurities and among them not more than o.i per cent, of arsenic. Tin, on the other hand, should contain not more than 0.2 per. cent, of impurities. Furthermore, especially, lead and zinc ought not to be mixed in. The constituents stated in per cent, are: tin 83.33, copper 5.55, antimony II. II. G. H. Clamer sums up American railroad practice in 1907 in the Proc. A. S. T. M., Vol. VII, page 302. Continued experiments have shown that tin could be reduced and lead increased beyond the amounts given by Dr. Dudley, and that a satisfactory bearing metal could be made composed of 65 per cent, copper, 5 per cent, tin and 30 per cent. lead. This alloy is largely sold under the name of 'Aplastic bronze." It has a compressing strength of about 15,000 lb. to the square inch and is found to operate without distortion in the bearings of the heaviest locomotives for the driving brasses, rod brasses and bushings, and for bearings of cars of 100,000 lb. capacity. Railroad specifications cover bearing alloys having a tin content of from 8 to 10 per cent, and lead from 10 to 15 per cent. As railroad practice yields a large amount of old material, or scrap, which must be used, many car brasses are made from old metal. These contain copper 65 to 76 per cent., tin 2 to 8 per cent., lead 10 to 18 per cent., and zinc from 5 to 20 per cent. It is estimated that such bearings constitute from 50 to 75 per cent, of all the car bearings in use. The Mesta Machine Company on rolling mill work uses two grades of babbitt and a bronze. For the general run of work, a lead babbitt is satis- factory having this composition in per cent., lead 75, tin 12.5, antimony 1^.$. For high rubbing speeds a mixture is made of i part of the above and 2 parts of genuine babbitt. This genuine babbitt, alone, is used on rolling mill engines and in bearings subjected to shock and pound. Its composition in per cent, is: tin 82, copper 5.4, antimony 12.6. The bronze is a tough copper-tin-lead alloy very similar to Pennsylvania Railroad metal. 6o BEARINGS AND THEIR LUBRICATION In the rolling mills for repairs any babbitt is considered satisfactory having about lo per cent, tin, 12 to 14 per cent, antimony, and the rest unspecified. " Copper alloy" is also used for roll bearings. AUTOMOBILE PRACTICE The alloys division of the Standards Committee of the Society of Auto- mobile Engineers in their report for June, 191 1, specifies three bearing metals as follows: Babbit Metal, Specification No. 24 Tin 84 per cent. Antimony 9 per cent. Copper 7 per cent. A variation of i per cent, either way will be permissible in the tin, and 0.5 per cent, either way will be permissible in the antim.ony and copper. The use of other than virgin metals is prohibited. No impurity will be permitted other than lead, and that not in excess of 0.25 per cent. Note: This grade of babbitt is special, owing to the large amount of copper contained therein. It is used for the connecting-rod bearings of gasoline motor bearings, locomotive work, or for any service where machinery designers are confronted with severe operating conditions. White Brass, Specification No. 25 Copper 3 . 00 to 6 . 00 per cent. Tin, not less than 65 .00 per cent. Zinc 28 . 00 to 30. 00 per cent. Metal containing more than 0.25 per cent, impurities may be rejected. Note: This alloy gives good results in automobile engines, but provision should be made to have it generously lubricated. Phosphor Bronze Bearing Metal, Specification No. 26 Copper 80 . 00 per cent. Tin 1 o. GO per cent. Lead i o. 00 per cent. Phosphorus o. 05 to o. 25 per cent. Impurities in excess of 0.25 per cent, will not be permitted. Note: This is a metal similar to that specified by many railroads for various purposes. It is an excellent composition where good anti-frictional qualities are desired, standing up exceedingly well under heavy loads and severe usage. It should be used only against hardened steel in automobile construction. MATERIALS FOR BEARINGS 6i Red Brass, Specification No. 27 Copper 85 .00 per cent. Tin i 5 . 00 per cent. Lead 5 .00 per cent. Zinc 5 . 00 per cent. A tolerance of i per cent, plus or minus will be allowed in the above percent- ages. Impurities in excess of 0.25 per cent, will not be permitted. Note: A high grade of composition metal, and an excellent bearing where speed and pressure are not excessive. Largely used for light castings, and possesses good machining qualities. ALLOYS FOR METALLIC PACKING . In metallic packing for piston rods, valve stems, and turbine shafts, we find a use of alloys similar to their use as bearing metals but with this important Plan Shaft Sleeve. Packing against Pressure. Springs placed •^ about 4"apart. — Packing against Vacuum. Driving Fit -^ |-1-^-'* 5C Y2 Metallic Back. f^»»^^ Notches in Back L _o7^'l _J to hold Metallic ' ' "^^ '^^ „ Packing. Section 6-6 Metallic Packing. Brass Band. FIG. 8. CUNNINGHAM METALLIC PACKING FOR VTCRTICAL STEAM TURBINE. 62 BEARINGS AND THEIR LUBRICATION difference: In packings the metal lining is not intended to carry any weight whatever but closes tightly around the shaft or rod and makes a steam-tight joint. However, as the rod and shaft are in motion, we have a condition of ordinary sliding friction, but differing again from bearing practice in that the parts are heated to a high temperature by steam. FIG. 9. SECTION OF LABYRINTH PACKING FOR STEAM TURBINE SHAFT. Fig. 8 shows the Cunningham turbine shaft packing as used on Curtis verticle turbines in place of carbon packing. It consists of cast-iron quadrants filled with Allan ''red metal," 50 per cent, lead and 50 per cent, copper, fitted together and surrounded by a brass band with helical tension springs. The General Electric Company uses a similar packing but arranged in the form of a labyrinth, as shown in Fig. 9. Both of these packings are run without lubrication and are peculiarly well adapted for high-speeds and high temper- atures. Babbitt Rings in Halves '/is Cut-out Br 133 Dowel Pin^ \' f^^^^^^m^''^^ Babbitt Rings in Halves Brass II— 4-4-J-^^ Multiangular PaclsmS—. ■^"^Single^ Angle f _ _ JPackiQg' I FIG. 10. ARRANGEMENT OF KING METALLIC PACKING FOR LOCOMOTIVE PISTON RODS. To show the practice in regard to piston rods and valve stem packings, Fig. 10 is given. These are the "single angle" and "multi- angular" packing of the United States Metallic Packing Company, as applied in locomotive service. The babbitt mixture used for the single angle packing is 100 parts tin, 9 parts copper, and 6 parts antimony. For the multi-angular packing the MATERIALS FOR BEARINGS 63 mixture is 83 1/3 parts lead, 8 1/3 parts tin, and 8 1/3 parts antimony. Allan "red metal" is also used. In addition to these solid metal packings, there are many soft packings composed of metallic wool or fibers pressed together into a cord or rope and used in place of ordinary fiber packings. They are composed largely of lead with the addition of a small amount of tin or copper. A representative analysis is 88 lead and 12 copper. VARIATIONS IN BEARING METALS In the discussion of babbitt metals it is pointed out that many such "bab- bitts," placed on the market in fancy shaped ingots and under attractive names, are nothing but waste metals from type foundries and "clean up" from smelters. Of course, such metals vary in composition. If a particular mixture is desired for bearing purposes, it should be purchased from a reputable dealer according to a fixed analysis. In the use of bronzes there is not only the variations in proportions of the constituent elements to be taken into consideration, but also the way in which the mixture is compounded. Under different foundry conditions the same constituent metals in identical proportions may give alloys with different physical properties. In adopting a bronze bearing metal one of the important points is to select an alloy that is easily handled under ordinary foundry con- ditions. If a precise mixture is desired, reputable metal dealers and foundry- men can be found who will guarantee to furnish either ingot metal or castings, in which none of the constituent metals will vary more than one half of i per cent, from the specification. WOOD AS A BEARING MATERIAL But little need be said in regard to the use of wood for bearings. In the early days of mill construction, lignum vitai was successfully used for the bearings of mill shafting. This wood has a peculiar greasy character which, coupled with its extreme hardness and density, rendered it successful for moderate pressure and comparatively low speed. Two other uses are deserving of mention: One is in the step bearings of water turbines where the step is constantly flooded with water. If the pres- sures and speed were not too high it has proved reasonably successful in this service. The second, and at the present time practically its only application, is in the tail shaft bearings of steam ships and submerged guide bearings of water wheels. Here it is commonly put in in the form of staves, and as the bearing is open to the water it is constantly lubricated and cooled. Other kinds of wood have 64 BEARINGS AND THEIR LUBRICATION been used to a limited extent, such as oak, wild pear, and apple tree wood. Several attempts have been made to put on the market wooden bushings soaked with a wax or grease to render it unnecessary to oil them. They are usually described as ''oilless bearings." For very light pressures they have had a certain amount of success, although their use is by no means general. GRA.PHITE AS A BEARING MATERIAL Graphite bearings usually consist of a brass or bronze shell having a series of openings or slots into which graphite in the form of a paste is forced under pressure and allowed to harden. Such bearings are intended to be run without oil, and are especially adapted for positions where oiling would be very difficult or under conditions of considerable heat. In operation the bronze shell supports the shaft and the graphite acts as a lubricant. Another form of graphite bearing consists in casting a piece of netting having a number of slugs of graphite attached thereto into the bearing lining if of babbitt, or into the box or brass if of bronze. This type is not intended to run without lubrication, but the addition of the graphite in the form of these slugs equally spaced throughout the bearing surface is intended to reduce friction by adding a valuable solid lubricant. MISCELLANEOUS MATERIALS Hard fiber is used for small thrust collars and with low pressures and ample lubrication is satisfactory. Knife-edge V's in weighing machines and computing scales are frequently made of agate with carefully polished surfaces. The most frictionless bearings that are produced — if bearings they can be called — are mercury baths. They have been used in French lighthouses and to support certain parts of astronomical instruments. They permit of the freest motion of the floating members. This is an example of perfect fluid support. As a metal that will not amalgamate must be used for the float and casing, these parts are made of cast-iron. The specific gravity of mercury is 13.58. SECTION IV ALLOWABLE BEARING PRESSURES, SPEEDS AND TEMPERATURES In Section II it is pointed out that values of the coefficient of friction for types of bearings and various kinds of lubrication are none too accurately known. The same condition exists to a certain extent for allowable bearing pressures. However, good practice is given in Table 33 for several kinds of machines and service. For light machinery the unit pressures are so small as to be negligible. Kind of bearing and condition of operation Allowable bearing pressure in pounds per square inch of projected area Bearings for very low speeds and intermittent service as in turntables and bridges. 7000 to 9000 Railroad Practice Locomotive cro?s-head pin bearings Locomotive crank pin bearings Locomotive driving wheel journal bearings . Car axles bearings Tender axle bearings 3000 to 4000 1500 to 1700 to 550^ 300 to 325 to 425 British Railway Practice Locomotive crank pin bearings Locomotive cross-head pin bearings. Locomotive driving axle bearings. . . Car axle bearings ... to 1400 ... to 2000 250 to 300 ... to 330 Table 32. — Allowable Bearing Pressures for Machinery Bearings. 1 Pennsylvania Railroad Class E6 locomotives have a bearing pressure on bearings of main drivers of 556 lb. per square inch. S 6s 66 BEARINGS AND THEIR LUBTICATION Eind of bearing and condition of operation Allowable bearing pressure in pounds per square inch of projected area United States Naval Practice Main engine bearings Main engine crank pin bearings . . Steam turbine bearings Thrust bearings for torpedo boats. 275 to 400 400 to 500 ...to 85 . . to 50 For weight alone. Merchant Marine Practice Main engine bearings 400 to 500 Main engine crank pin bearings ] 400 to 500 High-speed Stationary Engine Practice Main bearings , Main bearings Crank pin bearings, overhung cranl. Crank pin bearings, center crank. . Cross-head pin bearings , 60 to 120 For dead load. 150 to 250 900 to 1500 400 to 600 1000 to 1800 For steam load. Slow Speed Stationary Engine Practice Main bearings Main bearings Crank pin bearings Cross-head pin bearings. For dead load. For steam load. Air Compressor Practice ^ Straight line, steam driven, 100 lb. steam and air. Main bearings Crank pin bearings Cross-head pin bearings. 160 to 237 565 to 700 628 to 820 Table 33. — Continued. — Allowable Bearing Pressures for Machinery Bearings. 1 Canadian Rand Company. ALLOWABLE PRESSURES 67 Kind of bearing and condition of operation Allowable bearing pressure n pounds per square inch of projected area Straight line, belt driven, center crank, 100 lb. steam and air. Main bearings ! 122 to 220 Crank pin bearings I 244 to 402 Cross-head pin bearings | 400 to 785 Straight line, belt driven, side crank, 100 lb. steam and air. Main bearings | 178 to 227 Crank pin bearings . . . . | 628 to 825 Cross-head pin bearings [ 628 to 825 Straight line, steam driven, side crank, 100 lb. steam and air. Main bearings 198 to 227 Crank pin bearings 462 to 825 Cross-head pin bearings 462 to 825 i Duplex, Meyer cut-off, steam-driven, 100 lb. stearr and air. Main bearings Crank pin bearings Cross-head pin bearings. 157 to 200 644 to 855 850 to 1370 Duplex Corliss valve gear, steam driven, 100 lb. steam and air. Main bearings Crank pin bearings Cross-head pin bearings Direct-connected, motor driven main bearings Gas Engine Practice Main bearings Crank pin bearings Cross-head pin bearings. 500 to 700 1500 to 1800 1500 to 2000 Table s^. — Cw^/inwei.— Allowable Bearing Pressures for Machinery Bearings. 68 BEARINGS AND THEIR LUBRICATION Kind of bearing and condition of operation Allowable bearing pressure in pounds per square inch of projected area Electrical Machinery Practice Generator and motor bearings Main engine bearings, driving generators Horizontal steam turbine bearings Vertical steam turbine steps 30 to 80 40 to 80 400 to 60 00 to 1000 Rolling Mill Practice ^ Rubbing velocity in feet per minute Pinion housing bearings I 30 to 50 ^ Roll housing beanngs i 100 to 2000 ^ Table roller bearings j 30 to 50 Table line-shaft bearings 1 30 to 50 Main bearings of shears 1 1800 to 2500 350 to 600 350 to 600 150 150 50 to 65 Miscellaneous Practice Bearings for slow speed and intermittent load as in punch presses, shears, and the like. Main bearings of slow speed pumping engines Heavy line-shaft bearings, bronze or babbitt lined. . Light line-shaft bearings, cast-iron Heavy slow speed step bearings Drill press thrust collars Angular-thrust bearing tor boring mill tables 3000 to 4000 to 600 100 to 150 15 to 25 to 2000 to 325 — to 75^ Table 33. — Continued. — ^Allowable Bearing Pressures for Machinery Bearings. Tables 33A to 33F, inclusive, are from an article by G. W. Lewis and A. G. Kessler published the American Machinist. They give main bearing crank pin bearing, and wrist or piston pin bearing pressures for large stationary gas engines, both horizontal and vertical. Four maximum explosion pressures 1 Mesta Machine Company, Pittsburg, Pa. ^ These factors are of value as showing good practice, not for purposes of design. The diameters and lengths of the bearings are determined by the requirement of strength in the pinion and roll necks and their housings. ^ Practice of BuUard Machine Tool Company. ALLOWABLE PRESSURES 69 in pounds per square inch of piston face are assumed, namely, 250, 300, 350, and 400 lb. These tables give a portion of the results of an extensive inves- tigation of the designs of a large number of engines and thoroughly reflect modern practice. D = cylinder diameter in inches. Pm = maximum explosion pressure in pounds per square inch of piston face. Dcp = bearing diameter of crank pin in inches. Lcp = bearing length of crarik pin in inches. Kcp = maximum unit bearing pressure in pounds per square inch. Acp = DcpX Lcp in square inches. Max. Kcp = 71 Pm -^ (Dcp X Lcp) (A) 4 D 4 8 12 16 20 Assumed Dcp Lcp Acp I 1/2 I 5/8 2.44 3 1/8 3 1/4 10.15 4 3/4 4 7/8 .3- 63/8 6 9/16 41.75 8 1/8 8 3/8 68 DcpX Lcp = Acp Pm Assumed Kcp 1290 1240 1220 1150 From equation A Pm 300 Assumed Kcp 1550 1485 1450 1450 , 1390 From equation A Pm Kcp 350 1800 j 1730 1710 Assumed 1690 1620 ! From equation A Pm 400 Assumed Kcp 2060 1980 I ■ i 1950 1930 1850 I From equation A Table 33A. — Crank-pins Bearing Pressures for Horizontal Stationary Gas Engines. 70 BEARINGS AND THEIR LUBRICATION D = cylinder diameter in inches. Pm = maximum explosion pressure in pounds per square inch of piston face. Dcp = bearing diameter of crank pin in inches. Lcp = bearing length of crank pin in inches. Kcp = maximum unit bearing pressure in pounds per square inch. Acp =DcpX Lcp in square inches. Pm D 4 8 12 16 20 Assumed Dcp Lcp Acp I 5/8 I 5/8 2.64 3 1/4 3 5/8 II. 8 4 7/8 5 5/8 27.8 6 1/2 7 5/8 49-75 8 3/16 9 5/8 78.75 Dcp X Lcp 250 Assumed Kcp 1190 1065 1035 1015 995 Pm 30c ) Assumed Kcp 1430 1280 1240 1215 1200 Pm 35c ) Assumed Kcp 1660 1490 1440 1420 1400 Pm 40c Assumed Kcp 1920 1720 I 60 1620 1600 Table 2>:i B. — Crank-pin Bearing Pressures for Vertical, Stationary Gas Engines. ALLOWABLE PRESSURES n D = cylinder diameter in inches. Dmh = main bearing diameter in inches. Lmb = main bearing length in inches. Amb = projected area main bearing (one) = DmhX Lmb. Pm = maximum explosion pressure in pounds per square inch of piston face. Kmb = maximum unit bearing pressure in pounds per square inch considering explosion to occur on dead center. D i6 20 Assumed Dmb Lmb Amb 1-3 31 4.8s 2.6 6.75 10.8 3-4 20.9 52.5 6.6 14.9 98.5 8.4 19. 1 160.5 Dmb X Lmb Vm 25c > Assumed Kmb 462 300 270 25s 244 Pm 30c Assumed Kmb 553 360 324 307 293 Pm 350 . Assumed Kmb "' 420 377 358 342 Pm 40c ) Assumed Kmb 738 503 430 408 391 Table 33C. — Main Bearing Pressures for Horizontal, Stationary Gas Engines. 72 BEARINGS AND THEIR LUBRICATION D = cylinder diameter in inches Dmh = bearing diameter of main bearing in inches Lmh = bearing length of main bearing in inches Amh = DmbxLmb Pm = maximum explosion pressure in pounds per square inch of piston face. Pm Kmb 300 D 4 8 12 16 20 Assumed Dmb Lmb Amb I 1/2 3 1/2 5-25 3 1/2 63/4 23.6 5 1/2 10 55 7 1/2 13 97-5 9 1/2 16 152 DmbxLmb 267 250 258 Assumed 258 258 Pm Kmb 359 320 300 Assumed 310 310 310 Pm Kmb 415 373 350 Assumed 360 I 360 ! 360 Pm 400 Assumed Kmb I 481 414 414 414 t 414 Table 33D. — Main Bearing Pressures for Vertical Stationary Gas Engines. ALLOWABLE PRESSURES 73 D = cylinder diameter in inches Dwp = bearing diameter of piston pin in inches Lwp = bearing length of piston pin in inches Awp = projected area piston pin in square inches Pm = maximum unit explosion pressure Kwp = maximum unit bearing pressure in pounds per square inch JDwp =0.0143 D"+o. f (A) Lwp =i.'j^Dwp (B) D 4 8 12 16 20 Dwp •93 1.62 2.76 4.36 6.42 Lwp 1.6 2.8 4-77 7-52 II. 15 I Awp 1.49 4-54 13.2 32.8 71-5 From Equation (B) Dwp X Lwp Pm 250 Assumed Kwp 2100 2760 2145 1530 1 100 Pm 300 Assumed Kwp 2530 3320 2570 1840 1320 Pm 35° Assumed Kwp 2950 3880 3000 2150 1540 Pm 400 Assumed Kwp 3370 4425 3430 2455 1760 Table 33E. — ^\Vrist or Piston Pin Bearing Pressures for Horizontal, Stationary Gas Engines. 74 BEARINGS AND THEIR LUBRICATION D = cylinder diameter in inches Dwp = bearing diameter of piston pin Lwp = bearing length of piston pin Awp = projected area piston pin in square inches Kwp = maximum unit bearing pressure in pounds per square inch Pm = Maximum unit explosion pressure. Dwp =0.00795 ^'+ I 3/8'' (C) Lwp =1.82 Dwp (D) D Dwp Lwp Awp 15/8 27/8 4.67 16 I 7/8 I 2 1/2 j 3 3/8 3 3/8 I 4 1/2 I 6 1/8 6.33 I 11.25 i 20.65 4 1/2 Equation (C) 8 1/8 Equation (D) 36.6 Dwp X Lwp Pm 250 - Assumed Kwp 1510 1990 2520 2430 2145 Pm 300 Assumed Kwp 1810 2380 3020 2920 2580 Pm 350 Assumed Kwp 2120 2780 3520 3410 3010 Pm 400 Assumed Kwp 2420 3190 4030 3900 3440 Table 33F. — Wrist or Piston Pin Bearing Pressures for Vertical Stationary Gas Engines. Tables 33 G to 33L, inclusive, are from an article by G. W. Lewis and A. G. Kessler published in the American Machinist. They give the maximum unit pressures and rubbing speeds on the bearings of automobile engines ALLOWABLE PRESSURES 75 for various assumed explosion pressures and for various sizes of engine cylinders. They were developed from an investigation of some 30 Amer- ican automobile engines and represent average practice. D = cylinder diameter in inches, Z)c& = bearing diameter of center bearing in inches, Lcb == bearing length of center bearing in inches, Xc6 = maximum unit bearing pressure on center bearing in pounds per square inch, Pm = maximum explosion pressure in pounds per square inch of piston face. Dcb=o.32D+o.3'' (5) Lcb =2.^ Deb- 2.2" (6) D 4 1/2 5 1/2 Deb Lcb Acb 1 9/16^ I 3/4 2 3/16 2 3/4 342 4-3 I 7/8 2 1/16 3 1/16 3 5/8 5-75 I 7-5 From equation (5). From equation (6). Deb X Lcb. Pm 250 Assimied. Kcb 690 615 620 575 Pm 300 Assumed. Kcb 830 730 750 690 Pm 350 Assumed. Kcb 970 855 870 810 Pm 400 Assumed. Kcb HOC 980 1000 920 Table 33G. — Maximum Unit Bearing Pressure on Center Bearings of Automobile Engines. 76 BEARINGS AND THEIR LUBRICATION D = cylinder diameter in inches, Z)/6 = bearing diameter of front bearing in inches, Ljb = bearing length of front bearing in inches, Z/6= maximum unit bearing pressure on front bearing in pounds per square inch, Pm = Maximum explosion pressure in pounds per square inch of piston face. DJb=o.s2D+o.y' (7) Lfb = D/b+i i/S'' (8). D Dfb Lfb Afb Pm 4 i/2 5 i/2 1 9/i6 I 2 I1/16 4.2 1 3/4 2 7/8 5.02 I 7/8 ' 2 1/16 From equation (7). 3 3 3/16 From equation (8). 5.63 6.6 Dfb X Lfb. 250 Assumed. Kfb Pm Kfb Pm 560 665 595 640 670 300 Assumed. 710 775 800 350 Assumed. Kfb 780 830 900 930 Pm 400 Assumed. Kfb 890 945 1030 1075 Table 33H. — Maximum Unit Bearing Pressure on Front Bearings of Automobile Engines. ALLOWABLE PRESSURE 77 D = cylinder diameter in inches, Drh = bearing diameter of rear bearing in inches, Lrb = bearing length of rear bearing in inches, Arb = projected bearing area of rear bearing in square inches, Krb = maximum bearing pressure on rear bearing in pounds per square inch, Pm = maximum explosion pressure in pounds per square inch of piston face. Drb=o.s2D +0.3" (9) Lrb^S-2>Drb-S-Z" (10). 4 1/2 51/^ Drb Lrb Arb 1-9/ 16 3 4-7 1-3/4 1-7/8 2-1/16 4-5/8 5-5/8 8.67 II. 6 From equation (9) From equation (10) Drb X Lrb. Pm 250 Assumed. Krb 56s 495 495 465 Pm Krb 300 Assumed. Pm 350 Assumed. Krb 760 660 655 605 Pm 400 Assumed. Krb 855 735 735 675 Table 33I. — Maximum Unit Bearing Pressure on Rear Bearings of Automobile Engines. 78 BEARINGS AND THEIR LUBRICATION D = cylinder diameter in inches, Dwp = bearing diameter of wrist pin in inches, Lwp = bearing length of wrist pin in inches, Awp = projected bearing area of wrist pin in square inches, Pm = maximum unit explosion pressure in pounds per square inch. of piston face. Dwp =o.T,4D-o.sf (i) Lwp =2.2^ Dwp (2) Kwp = maximum unit bearing pressure in pounds per square inch. Pm J92 0.7854 Kwp= -— {a) Awp Pm Kwp D 4 4 1/2 5 5 1/2 Dwp Lwp Awp 7/8 2 1-75 I 2.25 2.25 1 3/16 2 5/8 3.12 I 3/8 3 1/16 4.2 From equation (i). From equation (2). Dwp X Lwp. 1800 250 1780 1570 1420 Assumed. Equation (a). Pm 300 Assumed. Kwp Pm 2150 2130 1890 j 1700 Equation (a). 350 Assumed. Kwp 2510 2480 2200 1980 j Equation (a). Pm 400 Assumed. Kwp 2870 2840 2510 2270 Equation (a). Table 33 J. — ^Maximum Unit Wrist Pin Bearing Pressures in Automobile Engines. ALLOWABLE PRESSURES 79 D = cylinder diameter in inches, Dcp = bearing diameter of crank pin in inches, Lcp = bearing length of crank pin in inches, Acp = projected bearing area of crank pin in square inches, Kcp = maximum unit bearing pressure on crank pin in pounds per square inch, Pm = maximum unit explosion pressure in pounds per square inch of piston face. Dcp= o.z2D + o.f (3) Lcp = 1.35 Dcp (4) F mD' .JS54 _ ^^^= A^p (') 4-1/2 S-1/2 Dcp Lcp Acp 1 9/16 I 3/4 2 1/18 2 3/8 3-32 4-17 1 7/8 2 1/2 4.68 2 1/16 2 3/4 5-68 From equation (3) From equation (4) Dcp X Lcp. Pm 250 Assumed. Kcp 945 960 1050 1050 From equation (b). Pm 300 Assumed. Kcp 1 130 1 1 50 1260 1260 From equation (b). Pm 350 Assumed. Kcp 1320 1350 1470 1470 From equation (b). Pm 400 Assumed. Kcp 1510 1540 1675 167s From equation (b). Table 33 K. — Maximum Unit Crank Pin Bearing Pressures for Automobile Engines. 8o BEARINGS AND THEIR LUBRICATION D = cylinder diameter in inches, L = stroke in inches, ■^mean "^ Hican total prcssuie on piston in pounds for entire cycle (assumed), RPM = revolutions per minute at looo ft. per minute piston speed, Km = mean unit bearing pressure on bearings in pounds per square inch, V = rubbing speed in feet per second, W = work of friction = {Km X V/j.) in foot pounds per second, L = 1.1 D. D 4-1/2 5-1/2 } Average 4-3/8 5-1/2 R.P.M. [370 1090 Rubbing speed in ft. per min. on 560 bearings listed below. 550 535 540 546 V 9-35 915 8.9 252 320 395 475 Crank-pin bearing. Center bearing Front bearing. Km 76 77 84 W 710 705 Km 55 49.2 W 515 450 745 49. 445 Km 44.7 47-5 52 W 418 435 462 84 755 46 415 53-4 480 80 729 50 456 49-4 449 Rear bearing. Km 56 45-5 43-5 38 [ W 523 415 388 342 45-7 417 Table 33L. — Average Rubbing Speed and Work of Friction for Automobile Engines. — Computed for a piston speed of 1000 ft. per minute and a mean pressure of 20 lb. per square inch of piston face. ALLOWABLE PRESSURES 8i — 1 r 1 _ \ \ \ \ 'i 1 5 \ J \ 1 C ^ s§ fi p la V \ — 1 ^^ \ s \ ^ •eajy ps^oarojjj qoui gjBnbs jad spunoj ui oanssaa j SuvvsdQ 82 BEARINGS AND THEIR LUBRICATION Fig. II gives maximum safe bearing pressures for rubbing speeds up to 5000 ft. per minute and for perfect film lubrication. It shows the practice of the (jeneral Electric Company for the bearings of electrical machinery and is representative of the best of American practice. REPRESENTATIVE GERMAN PRACTICE Representative German practice in regard to unit pressures for journal bearings as found on page 742 Des Ingenieurs Taschenbuch, Hiitte, Vol. i, is given below with the metric units transformed into corresponding English units in round numbers. Materials in contact Pressure in pounds per square inch of pro- jected area. Hardened crucible steel on hardened crucible steel 2130 1280 Unhardened crucible steel on bronze 850 425 355 355 Mild steel with a smooth compact surface on bronze Mild steel with not an entirely clear surface or cast-iron on bronze. Mild steel with not an entirely clear surface on cast iron Mild steel on lignum vita? with water lubrication In explanation of the above values, we read: ''If the total pressure even when the journal is at rest maintains approximately of the same magnitude and direction (for example, in heavily loaded shafts, heavy gears, etc.) the unit pressure should be taken smaller than above. ''For journals (or bearings) which oscillate only the unit pressure may be considerably higher. "For the bearings of rope and chain sheaves, etc., which turn intermittently and on which the wear is either small or unimportant, the above unit values, 850 to 355 lb., may be doubled or tripled. ''For crucible steel crank and cross-head pins running on bronze in ordinary steam engines the unit pressure may be taken from 710 to 995 and 1070 to 1130 lb., respectively; for similar purposes on locomotive 1420 and 21301b., respect- ively; on high speed steam engines 570 to 710 lb., respectively. For fly wheel bearings of steam engines use 213 to 217 lb. "For crank pin bearings of punches and shears the unit pressure may be as high as 2845 It), per square inch of projected area." PERMISSIBLE VELOCITY OF RUBBING There are no well defined limits for rubbing velocities. In many classes of machinery the speed is so low that the velocity is an unimportant factor. With ALLOWABLE TEMPERATURES 83 the exception of power generating machinery, particularly steam turbines, high speed is usually accompanied with very low pressure. For instance, in machine tools speeds are uniformly low with the exception of grinding machinery, and there the pressures are slight. In Power, February, 1893, C. J. Field gives 350 ft. per minute as the max- imum surface velocity for the main bearings of steam engines in elect'-ical service with grease lubrication. The corresponding maximum pressure per square inch of projected area of journal is 80 pounds. The ordinary range of velocities for electric motors and generators, as indicated by the curve in Fig. 11, is from 400 to 1200 ft. per minute. Steam turbine practice in America is from 2400 to 6000 ft. per minute. However, this higher figure has been exceeded. One case that came to the author's attention was of a horizontal steam turbine with a bearing rubbing velocity of 8000 ft. per minute. PERMISSIBLE RISE IN TEMPERATURE A bearing shows distress by heating — by becoming a ''hot box." Within limits there is no objection to the rise in temperature; in fact, it must take place. It is sometimes said that the hotter a bearing runs the hotter it can run. The reason for this statement lies in the fact that the higher the temperature the less the viscosity of the lubricant, therefore the less the work of friction and the less the amount of energy transformed into heat at a given speed. With soft metal bearings the limiting temperature is one at which the lining will melt. With other materials the limiting temperature is one at which lubrication becomes insufficient and the bearing seizes. In practice it is necessary to design bearings to run at a much lower tem- perature than will cause damage because of the requirements of the average customer. Such a maximum temperature is from 140 to 160° F. It is prob- ably true that the average bearing running at that temperature could just as well run at a temperature of 200° F. Occasionally manufacturers take advantage of this fact in designing a machine to be located in some out-of-the-way place where it will not have close inspection, or will not be approached by the ordinary passer-by who might place his hand upon a bearing, and a high temperature is allowed. There is no valid objection to running a bearing hot, provided the temper- ature reached is not sufficient to do harm and provided there is no progressive increase in heat indicating trouble. This is opposed to the ideas of some machinery users who look upon any bearing that warms up as improperly designed. There is no reason in expecting or demanding that a machinery bearing should run stone cold. In the very nature of the device, energy must be transformed into heat, and this heat will cause a rise in the temperature of 84 BEARINGS AND THEIR LUBRICATION the bearing until a condition is reached where the heat radiated and dissipated in a unit of time balances the amount liberated in the same time. In machine-tool practice low running temperatures are the rule. This indicates a very ample factor of safety. The Bullard Machine Tool Company has fixed the maximum running temperature under load for bearings of the ma- chines that they build as i io° F. Every machine is tested in this particular before shipment. The actual running temperatures are considerably lower as shown by the following tabulation, giving the log of tests of six machines. Machine 24-inch vertical turret lathe. 24-inch vertical turret lathe. 24-inch vertical turret lathe 36-inch vertical turret lathe. 36-inch vertical turret lathe. 36-inch vertical turret lathe. Running Temperature of Bearings in Degrees Fahrenheit Speed box Head- stock Table pinion f I- 70 65 J 2- 80 68 3-85 68 I- 79 61 2- 95 65 3- 91 67 I- 74 64 2- 96 68 3- 95 68 r I-- 75 62 1 2- 81 65 3-80 65 r I- 82 70 2-100 72 3- 97 72 r I- 80 64 2- 85 65 ^ 3- 86 66 76 79 80 Rail raising bracket 73 78 80 78 82 82 85 no 112^ Power travers- ing bracket Upper Lower 74 80 80 77 80 79 86 84 82 75 80 80 68 70 71 74 72 74 70 76 78 76 80 78 68 66 70 70 71 72 85 83 74 78 79 Feed works Upper Lower Driving pulley bracket 66 ' 65 70 68 69 80 68 1 69 I 78 The three sets of readings in each case were taken: i, at the end of the first hour's run without load; 2, at the end of the second hour's run with load; 3, at the end of the third hour's run without load. ^ After correction of this bearing the temperature dropped to 78° at the end of a 2-hr. run ALLOWABLE PRESSURES 85 PRODUCT OF UNIT PRESSURE AND VELOCITY In designing bearings neither the pressure nor the circumferential speed alone can serve as a standard for calculating dimensions. The product of the pressure by the speed and the coefficient of friction must be taken into account, that is, the work of friction. In Section II it is shown that the product of the unit pressure and the coefficient of friction remains almost a constant for a wide range of pressures. This has been taken advantage of by many designers to establish a limiting factor of design commonly referred to as C, which represents the product of the pressure in pounds per square inch and the velocity in feet per minute. However, the values of this factor differ so widely that no attempt will be made here to tabulate them for general types of bearings. To give an idea of the range, F. W. Taylor, Trans. A.S.M. E., Vol. XXVII, gives C for mill work as 24,000 and for cast-iron bearings with ordinary lubrication as 12,000. The same factor used in turbo-generator design for forced lubrication and water cooling is 360,000. It is probable that these may be taken as limiting values, but it is impossible to give consistent data for other kinds of bearings in order to complete the series. Tower mentions two cases in which his experimental bearing seized, the conditions being a bronze half box, a steel journal and bath lubrication. In one instance the bearing seized at a pressure of 573 lb. per square inch and a speed of 419 ft. per minute. This is equivalent to a factor of 240,000. In the other case the pressure was 625 lb. per square and the rubbing velocity 366 ft. per minute. The corresponding factor is 228,000. One instance that has come to the author's attention is of a large turbo- generator running with a bearing pressure of 41 Ib-per square inch and rubbing velocity of 8,000 ft. per minute. The value of C in this case is 328,000. Some of the greatest values for this factor that have come to the author's attention are from some experiments made by the General Electric Company. In one case a bearing was run at a pressure of 285 lb. per square inch and a rubbing velocity of 2,700 ft. per minute — C= 770,000. In the second case the pressure was 200 lb. and the velocity 5,000 ft. — C= 1,000,000. The Westinghouse Machine Company has carried this experimentally to 1,720,000; 400 lb. pressure, 4300 ft. per minute, and for long runs. In spite of the fact that it is impossible to give values for this factor that seem to be consistent with themselves and in sufficient number to cover the ordinary range of commercial bearings, pressures, and speeds, the designer should not think it is of no value. On the contrary, it is of considerable value as a rough check upon design. It is wise for any one having to do with bearing design to tabulate this factor carefully for all important bearings coming within his experience, to arrive at limiting values beyond which he will know that the running of the bearing is liable to be unsatisfactory if not impossible. SECTION V DESIGN OF JOURNAL BEARINGS The important points to consider in proportioning and designing journal bearings may be grouped under six heads: - a. Proportions and dimensions of diameter and length. b. Selection of the bearing material. c. Provisions for anchoring the bearing metal if a soft lining, and provisions or adjustment to preserve fit and alinement. d. Clearance between journal and bearing. e. Provisions for lubricating. f. Provisions for radiating and dissipating heat. RATIO OF BEARING LENGTH TO DIAMETER In proportioning bearings, the designer must harmonize conflicting con- ditions. On the one hand the shaft must be sufficiently strong and rigid to sustain its load without undue deflection (this calls for large diameters) ; again the bearing pressures must not exceed a certain amount in order that an oil film may be established and maintained (this calls for large diameters and long bearings). On the other hand the circumferential speed must be kept as low as possible, thus keeping down the work of friction and reducing the amount of heat liberated and the risk of seizing (this calls for small diameters). Again the bearing length should be short, thus rendering the surfaces more effective and minimizing the danger of a local concentration of pressure from the deflection of the journal (this calls for short bearings). In actual designing the total load to be carried by the bearing divided by the allowable unit pressure, gives at once the projected area of the bearing. The smallest permissible diameter of journal can then be found by deter- mining the shaft deflection under running conditions. This investigation can best be made by graphic statics. A valuable discussion of this subject will be found in Elements of Graphic Statics, by Cathcart and Chaffee, at page 269. The diameter thus determined can be checked by finding out if its circumfer- ential rubbing velocity is greater than allowable. If the first proportions selected are found to be beyond the limits of good practice, another set can be taken and the investigation repeated until the 86 DESIGN OF JOURNAL BEARINGS 87 proportions are deemed satisfactory. Section IV gives allowable bearing pressures and velocities. It is seen from the preceding principles that the diameter and length of a bearing should both be kept as small as possible, consistent with the necessary strength and stiffness; the same conclusion is reached from the viewpoint of economy of material. Practice has tended to standardize the ratio of bearing length to diameter and the following tabulation gives representative proportions. Here I is the bearing length and d the diameter. Type of bearing Values of -^ a I Marine engine main bearings ' i to i . 5 Marine engine crank pin bearings j i to i . 5 Stationary engine main bearings i . 5 to 2 . 5 Stationary engine crank pin bearings i Stationary engine cross-head pin bearings i to i . 5 Ordinary heavy shafting with fixed bearings 2 to 3 Ordinary shafting with self-adjusting bearings 3 to 4 Generator and motor bearings i 2 to 3* Machine tool bearings | 2 to 4 PROPORTIONS OF AUTOMOBILE CRANK SHAFT BEARINGS A. G. Kessler and G. W. Lewis, in the American Machinist of September 29, 1 9 10, give the results of an investigation of the proportions of some 30 repre- sentative American automobile engines. The diameter of the front crank shaft bearing was found to vary from i 1/2 to 2 1/4 in., and is found from the equation: where -S^ = diameter of the front bearing in inches, and D = diameter of the engine cylinder in inches. The length of the front bearing is expressed by the equations: Zi=5i+ I 1/8 in.f and L^= 1.67 ^1 + 0.95 in. The reason for two equations lies in the arrangement of the curves drawn through the points plotted from the dimensions of the engines investigated. It shows that there are two types of shaft in use. The diameter B2 of the rear crank shaft bearing is found from the equation: B2=o.S3D-s/Sin. This is identical with the equation for the diameter of the front bearing, as 88 BEARINGS AND THEIR LUBRICATION it is customary to make all bearing diameters the same on automobile engine crank shafts. The length of the rear bearing is found from the equation: ^2=5-33 A- 5-66 in. 2^2= Length of rear bearing in inches and D= diameter of engine cylinder in inches, as above. SELECTION OF THE BEARING METAL The bearing metal must be selected with reference to the metal in the journal with which it is to run, to the load which it is to carry — it must be strong enough to resist deformation — and in many classes of machinery it must be selected for its resistance to wear. See Section III, Bronze is ordinarily used in the form of a sleeve or bushing either in a single piece or split in half. If a soft metal lining (babbitt) is used, it may either be poured directly into the housing of the bearing or inserted in a removable shell. Such shells are commonly made of cast iron, although brass is used to a limited extent and cast steel still more rarely. Standard American railway practice consists in alloying the lining and its bronze shell at their joining surfaces. The bronze surface is carefully cleaned, coated with solder, and the lining poured against it. If the work is well done there is no danger of the lining loosening. There is no mechanical anchorage. Sometimes different metals are used for the lower half of a bearing and its cap. If the cap takes no load it can be lined with a cheaper, softer metal than the load-carrying box. ANCHORING OF THE METAL LINING Babbitt linings must be securely anchored to their shells for if they loosen they are apt to crack and be destroyed. If this does not happen, the slight motion between the lining and its shell may act as pump and take the oil out of the bearing. The devices for anchoring are simple. In a common form the shell has a series of ridges with dovetailed edges around which the metal is cast. Fig. 12 shows an anchoring arrangement using anchors having a circular cross-section from the practice of the Westinghouse Electric and Manufacturing Company. The patterns are made to leave a plain core in the green sand, to which are added baked anchor cores secured with brads as shown . There are two sizes, 1/2 and 3/4 in. In bearings that for any reason have to be bored before babbitting, the anchor holes are cast extra deep, so that after boring the holes will be standard depth. DESIGN OF JOURNAL BEARINGS 89 The illustration gives the spacing. Along the straight lips of each half, the anchors should be fairly close together, and as near the edge as they can be cast, so that when the casting is finished the anchor holes will break through. Baked Anchor Core Center Mark Standard Anchors. Far all Split Bearings less thau 9"in Dia. ?<' Anchor -^ j;^//////////////, For all Split Bearing g'or more in Dia. Extra Deep Anchors. i ^ i^/Ayy^^^'MX n^^ '^.'-v-? ]-,,,._ 1 -''^;»_", "-' 4' f '-'-'' -/'./'-'O b ^^^ ^^^ ^^^ These Holes (through which the Babbitt is poured) are Cast in the Bottom Half of the Bearing directly opposite the Oil Holes and Slots in the Top Half. 'vi under 5 Dia. ijafor 5"Dia. and Pitch P for Yi. Anchor about \M. Pitch P for ?4 "Anchor about \M'' Over the remainder of the Bear- ing the Holes should be more sparsely distributed. „ Pitch D for i^'Anchor about 2M. Pitch D for ?i"Anchor about 3." FIG. 12. SIZE AND PITCH OF BABBITT ANCHORS. ADJUSTING DEVICES TO PRESERVE ALINEMENT Various devices are used to preserve alinement, or to make bearings self alining within small limits. In some cases, as for instance the spindle bearings of machine tools, the accuracy of the seat controls the alinement. In small elec- tric motors and a large number of shoe machines bushings are forced into place in the frames or housings and are then carefully line reamed. In case a bushing has to be replaced for any cause whatever the process of line reaming must be repeated. In steam engine work the bearing brasses are usually adjustable by means of wedges or bolts. In electric motors, generators, horizontal turbines, and 90 BEARINGS AND THEIR LUBRICATION some gas engines, the bearings are made with a spherical seat, as shown by Fig. 13. Bearings for line shafting are usually held on a self-alining ball seat pro- vided with adjusting screws in a vertical direction and with elongated slots in the feet of the hanger for sidewise adjustment. (See Fig. 13.) The self alining feature lies in the ball seat of the box. Machine countershaft bushings and occasionally more important bearing linings are frequently placed in position on the shaft inside of the housings and '^ r Ball or Spherical Surface Electric Motor Bearing. Plain Ball and Socket Shafting Box (Sellers.) FIG. 13. TYPICAL SPHERICAL ADJUSTING SURFACES FOR BEARINGS. the space between bushing and housing run with babbitt or an alloy that ex- pands slightly in cooling. This is a simple, cheap, and satisfactory manner to aline and locate such bushings having all of the advantages in manipulation possessed by the babitted box. On some textile machinery provided with rigid bearings, a slight amount of self-adjustment is obtained by tapering the journal from full diameter at about its centre to a reduction of 0.003 or 0.004 in. at the ends. This double taper allows for a slight amount of swiveling. CLEARANCE BETWEEN BEARING AND JOURNAL The space or clearance between the rubbing surfaces is of importance. If too small the coefficient of friction will be very high with a tendency to heat. If too large it will be difficult to maintain a suitable film of oil. The shaft should be freely supported and freely guided in its bearings, without producing any local strain in the material and no local heating through shocks. As a rule the tendency is to make the fit too close. If for any reason a journal DESIGN OF JOURNAL BEARINGS 91 must be tightly fitted in its bearings, which means that the entire circumference of the surface is loaded approximately equally, oil should be supplied by forced lubrication. This condition gives a high amount of friction. Lasche found that the friction force decreased rapidly with an increasing clearance becoming practically constant when the clearance was 0.06 in. His experiment was with a bearing 10 in. in diameter and 4 in. long. This is an extreme case. Good operating practice for main engine bearings is a slackness of 0.005 in. for diameters from 6 to 12 in. and for cross-head pins and crank pins up to 6 in. in diameter 0.004 in. L. D. Burlingame, Trans. A.S. M. E., 1910, gives these limits used by Brown & Sharpe Mfg. Co. for ground cylindrical fits of small size. The tolerances are in decimals of an inch. Running Fits — Ordinary Speed To 1/2 -in. diameter, inc 0.00025 to 0.00075 Small To I -in. diameter, inc 0.00075 to 0.0015 Small To 2 -in. diameter, inc 0.0015 to 0.0025 Small To 3 1/2-in. diameter, inc 0.0025 to 0.0035 Small To 6 -in. diameter, inc 0.0035 to 0.005 Small Running Fits — High Speed, Heavy Pressure and Rocker Shafts To 1/2-in. diameter, inc o .0005 to o .001 Small To I -in. diameter, inc o.ooi to 0.002 Small To 2 -in. diameter, inc 0.002 to 0.003 Small To 3 1/2 in. diameter, inc 0.003 t^ 0.0045 Small To 6 -in. diameter, inc 0.0045 to 0.0065 Small Sliding Fits To 1/2-in. diameter, inc o .00025 to o .0005 Small To I -in. diameter, inc o . 0005 to o . 001 Small To 2 -in. diameter, inc o.ooi to 0.002 Small To 3 1/2-in. diameter, inc 0.002 to 0.0035 Small To 6 -in. diameter, inc o .003 to o .005 Small Grinding Limits for Holes To I -in. diameter, inc Standard to 0.00075 Large To 2 -in, diameter, inc Standard to o .001 Large To 3 1/2-in. diameter, inc Standard to 0.0015 Large To 6 -in. diameter, inc Standard to o .002 Large To 12 in. diameter, inc Standard to 0.0025 Large Table 34 gives the standard practice of the General Electric Company. 92 BEARINGS AND THEIR LUBRICATION A Journal Bearings Axle for ry linings -.M Horizontal Vertical Step . motors Allowable 1 •5 a ^ i ''5 Max. diam. in. variation below max. diameter Min. bore in. Allowable variation above min. bore Min. bore in. Allowable j variation above min. bore Min. bore in. Allowable variation above min. bore Min. bore m. Allowable variation above min. bore 3/8 .375 .0005 .377 .001 .376 .001 .3755 .0005 .380 .004 1/2 .500 0005 .502 .001 .501 .001 .5005 .0005 .505 .004 5/8 .625 .0005 .627 .001 .626 .001 -6255 .0005 .63c .004 3/4 .750 .0005 .752 .001 .751 .001 .7505 .0005 .755 .004 7/8 .875 .0005 .877 .001 .876 .001 .8755 .0005 .880 .004 I 1 .000 .0005 1.002 .001 1 .001 .001 I .0005 .0005 1.005 .004 I 1/8 1. 125 .0005 1. 128 .001 1. 127 .001 1 . 126 .0005 1 . 13c .004 I 1/4 1.250 .0005 • 1.253 .001 1.252 .001 I. 251 .0005 I-2S5 .004 I 1/2 1 .500 .0005 1.503 .001 1 .502 .001 1.501 .0005 I-S05 .004 I 3/4 1.750 .0005 1.753 .001 1.752 .001 I. 751 .C005 I. 755 .004 2 2.000 .0005 2.003 .001 2 .002 .001 2.001 .0005 2.005 .004 2 1/4 2.250 .0005 2.253 .001 2.252 .001 2.251 .0005 2-255 .004 2 1/2 2.500 .0005 2.503 .001 2.503 .001 2.501 .0005 } 2.505 .004 2 3/4 2.750 .0005 2.754 .002 2.753 .002 2.7515 .0005 2.75s .004 3 3.000 .0005 3.004 .002 3.003 .002 3.0015 .0005 3 -00s .004 3 1/2 3.500 .001 3-504 .002 3.504 002 3.5015 .0005 3.507 .004 4 4.000 .001 4.005 .002 4.004 .002 4.002 .001 4-007 ! .004 4 1/2: 4.500 .001 4-505 .002 4.504 .002 4.502 .001 4-509 .004 5 1 5.000 .001 5.006 .002 5.005 .002 5. 0025 .001 : 5-009 .004 5 1/2 5.S00 .001 5.507 .002 5 -505 .002 5.503 .001 5-511 .004 6 6 .000 .001 6.009 .002 6.005 .002 6.003 .001 j 6. on .004 7 7.000 .001 7.011 .002 7.006 .002 7.0035 .CCI i 7.012 .004 8 8.000 .001 8.012 .003 8.006 .003 8.004 .C02 8.013 .004 9 9.013 .004 y. / .004 9-0045 10.005 11.0055 12.006 12 12.000 .0015 12 .016 .005 12,008 .005 .002 1 13 :r3.ooo 14 14.000 15 It;. 000 .0015 13.016 14.016 . 005 13.009 14.009 15 .010 .005 13.0065 14.007 iS-0075 16.008 002 .0015 .005 . 005 002 .0015 IS .016 .005 .005 • . 002 1 16 16.000 .0015 16.016 .005 16.010 .005 .002 17 1 7 . 000 .0015 17.018 .005 17. on .005 17.008 18 18.000 .0015 18.018 .005 18. on .005 18.008 .002 19 20 19.000 20.000 .0015 .0015 19.018 20.018 .005 .005 19.012 20.012 .005 .005 19.008 20.008 002 .002 21 21 .000 .002 21.018 .005 21.013 .005 21.008 .002 22 22.000 .002 22.02c .008 22.013 .005 22.008 .002 i . - oon 23 .02c 24 .020 25.020 26.020 27.022 28.022 29.022 23-013 24-013 .005 .005 23 .008 24.008 ■^ ,--• .008 25 26 25.000 26.000 27.000 28.000 29.000 .003 .003 .003 .003 .003 .oc8 .008 .008 j 27 28 j ! .008 .008 29 30 31 30.000 31 .000 .003 .003 30.022 31 .022 008 .008 1 j 32 33 34 35 36 32.000 33.000 34.000 35- 000 36.000 .003 .003 .003 .003 .003 32.024 33.024 34-024 35.024 36.024 .010 ; 1 ... i i::::::::: i 1 .010 !■•;■:■ 1 1 Table 34. — Allowances and Limits for Journals and Bearings. DESIGN OF JOURNAL BEARINGS 93 PROVISIONS FOR LUBRICATING American lubricating oils and greases are by no means standardized as to quality and constituents. It is, therefore, necessary for the designer to con- sider the possibility of a very poor lubricant being used, and prepare for that condition. American machinery builders do not specify what oil shall be used except by the expression ''a good oil for the purpose." Therefore, at the outset, the designer must prepare for the contingency of a poor oil or grease. As to the method of applying the lubricant there is a wide choice and it is one of the most important features of design. At the same time it is one that is ordinarily neglected; even a superficial reading of the pages of technical peri- odicals will show many protests on the part of users of machinery against inadequate oiling devices; even the elementary principles of feeding oil to a bearing are ofttimes ignored. On page 1 8 it is pointed out that a properly fitted and designed bearing will suck oil against a considerable head, provided it is introduced at a proper point. It has also been pointed out that the oil film in a well lubricated bearing is under considerable pressure. Obviously oil by a gravity feed cannot enter a bearing at this latter point. As a general principle oil fed by gravity must be introduced into a bearing at about the point of least pressure. The common methods may classified as follows: From a squirt can through an open oil hole. From a squirt can upon a felt washer covering an oil hole. From a squirt can through an oil hole cover to the oil hole beneath. Oil "I From an oil cup through a wick (syphon feed). From a sight-feed oil cup. From a pad pressed against the journal. From an oil well by moving rings or chains or by vanes fastened to the journal. From an oil well by centrifugal means. Ordinary or imperfect lubrica- tion. {From a compression grease cup. From a stick of grease pressed against the journal. Perfect lubrication Oil. From a pump or gravity system under pres- sure (15 pounds up) (the oil may or may not be artificially cooled). From a bath in which the journal dips. From a reservoir in which the bearing is sub- merged or from which it receives a flood under very little pressure (2 to 5 pounds). 94 BEARINGS AND THEIR LUBRICATION The relative values of several methods of lubrication were found by Goodman, Proc. Inst. C. E., Vol. LXXXIX, page 447, to be: Bath Saturated pad Ordinary pad Syphon I 1.32 2.21 4. 2 The numerical quantities indicate the frictional resistances referred to the best — bath lubrication — as i. For all gravity feeds, the point of application should be at about the point of minimum pressure in the oil film. With forced feed, practice indicates that the point of application should be at an angular distance of 45 degrees from the vertical in a direction counter to the direction of rotation. This is indicated by the diagram Fig. 14. Point to Introduce Oil under Pressure ^ / FIG. 14. LOCATION TO INTRODUCE OIL WITH FORCED LUBRICATION. DISCUSSION OF THE METHODS The common open oil hole has a serious objection; it is an opening to admit dirt and grit into the bearing as well as oil. With the felt plug it is consider- ably improved as dirt is excluded and the felt acts as a small reservoir from which oil can drip. (See Fig. 15.) Many self-closing oil hole covers are on the market and serve their purpose until they are broken off or lost. On some kinds of machinery they are used as ornaments. The oil pad in contact with the journal is a universal method for lubricat- ing railroad car bearings in Europe, and is extensively used for line shafting. It is open to the objection that in time it will become glazed on the surface in contact with the moving journal and lose its capillarity and ability to transfer oil from the well below to the journal above. A remedy is to take it out and soak it in kerosene or gasoline to remove the accumulation of dirt. DESIGN OF JOURNAL BEARINGS 95 The pad itself is often made of a fabric having a face of wool velvet and a back of cotton. This is fastened to a block of wood, and pressed against the journal by springs which lead down to the oil. In some designs, narrow ledges of the wood back fit the journal to reduce the wear on the pad. The arrange- ments are satisfactory but are open to two objections in America, the expense and difficulty of adapting them to standard equipment. The standard method of lubricating American car axle journals is by wool waste packing. The wool is elastic even when oil-soaked. It has a disadvan- tage of less capillary power than cotton and of short length fibers. Cotton, however, loses its elasticity when wet. Oiling Hole %. Vz or % Inch in Diameter. / TTTI FIG. 15. OPEN OILING HOLE WITH FELT WASHER. Recent experiments have been made with a woven fabric supported close to the edges of the brass and hanging down in a catenary form into the oil. The pressure against the journal is very slight, and long wicks bring up the oil. The results are promising. Ring or chain lubrication is the preferred form for better classes of machin- ery where the heat from the bearing can be dissipated without artificial means as, for instance, in electric generators and motors. Fig. 16 is representative of the arrangement. The rings hang loosely on the journal, rotate with it, pass around the outside of the box, and dip into a reservoir of oil beneath; thus they move in the same direction as the journal and carry oil to its upper surface. A certain amount finds its way to the rubbing surfaces and lubricates the bear- ing. The rest is merely transferred from one side of the reservoir to the other. Lasche states that there is a limit of speed beyond which rings cannot be used successfully, commonly said to be the speed at which they begin to '' dance " on the shaft instead of rotate. In his experiments he found that, considering the amount of oil delivered to the bearings, this maximum speed was about 2,000 r. p. m., and at about 2,500 revolutions the amount of oil delivered decreased very rapidly with a further increase of speed. 96 BEARINGS AND THEIR LUBRICATION American practice, however, uses rings at journal speeds as high as 3,600 r. p. m. The "dancing" is stopped by iacreasing the weight of the ring and depth of submersion. After a journal speed of from 600 to 800 r. p. m. has been reached, the rings do not increase in speed for an increase of journal speed. Good proportions are a ring diameter equal to two journal diameters and a submersion in the oil equal to one-half the journal diameter. Allowing for a suitable clearance around the rings and the size and capacity of the oil well are determined. The curves of Fig. 1 7 are from Lasche's experiments and may be used as a guide in estimating the amount of oil delivered by rings to the rubbing surfaces of the bearings. Housing Conical Collar to return Oil to Reservoir FIG. 16. RING OILING BEARING. These curves are especially remarkable in that they show how widely differ-^ ent quantities of oil are delivered for similar bearing conditions. When the journal revolved in the direction of the hands of a watch, the quantity of oil flowing was only 1/2 to 2/3 of that which flowed when the journal revolved in the opposite direction. Further, in both directions of rotation the quantity of oil which flowed from the farther side was only about 1/3 of that which escaped from the side nearer the driving belt. When the journal revolved in the more favorable direction and the oil entered at the unloaded side, 0.19 pint per minute flowed through the bearing. When it revolved in the other direction the quantity was only 0.12 pint per minute. That the relationship of these quantities must not be taken as indicating a DESIGN OF JOURNAL BEARINGS 97 law is shown by the second bearing tested in which the quantities of oil, de- pending upon the direction of rotation, happened to be in a reverse order. To work well the rings should be smooth and free from sharp edges and 200 500 1000 1500 2000 2500 Eevolutions per Minute Direction of_ Rotation IIT Direction of Rotation L.~ 0.20 » 0.189 « 0.18 a 0.168 i^ O.IC O-cj 0.147 2«J0.14 c ^ 0.12G .S 1 0.12 '^^ 0.105 >.= 0.10 g^ 0.084 g 1, 0.08 3 o. 0.063 = P.O.OC " 0.042 - 0.04 5 0.021 b 0.02 Total iQuantityofI Oil Flowing th rough th(^ Bearing Su ^C^"-^ " fa ce. (Side 1&2), y^ . ^ ^^-^ y .Side I — / ^ 3^- ^Side 11 __ i y ^ ^=^»— — 1 -^^ N Direction of Rotation I." Direction of Rotation 11." 200 500 1000 1500 2000 2500 3000 3500 Eevolutions per Minute Conditions: Steel shaft, 3.6 inches in diameter; bearing I of gun metal in one piece 10.4 inches long; bearing II of white metal, split, good mineral oil, two rings immersed 1.6 inches; rings six inches in diameter. FIG. 17. OIL DELIVERY CAPACITY OF OIL RINGS AT VARIOUS SPEEDS. corners that may catch and interfere with their free motion. Also the passages in which they lie must be free from projections upon which they might catch. Many split steel rings have been used but are open to the objection of possible 7 98 BEARINGS AND THEIR LUBRICATION catching. Less commonly they have been die cast in halves from white metal and the parts fastened together by small screws through a halved joint. The preferred ring is continuous, and made either of a hard white metal or brass. Artificial cooling is seldom resorted to with ring oiling, as there is no good way in which the oil can be cooled. Bearings have been installed with a ser- pentine coil of pipe in the oil well through which cold water was circulated, but such cases are rare. There are no data available as to the amount of oil which oil ring reservoirs should contain. The quantity should be great enough so that all particles will have sufficient time to give up their heat to the surrounding metal of the reser- Drill and Counter 'bore for Screw. ^ Countersink sufficient to have Head of Screw clear Surface^ Prick.P.unch ^Round Edges,' FIG. l8. Detail at Joint on Split Rings. DESIGN OF SPLIT OIL RING. voir before they are again brought in contact with the shaft. In some instances, originally hot running bearings have been made safer by merely increasing the quantity of the oil. Good practice places a limit of 4 in. as the greatest distance that an oil ring can be expected to successfully distribute oil; that is, if a bearing is 8 in. long, one oil ring at the center is sufficient. If the bearing is longer than 8 in. and shorter than 16 in. two rings must be used, and so on in proportion. Fig. 18 shows an approved form of split ring. Centrifugal oilers are frequently used with step bearings. They consist of a rotating vane, or series of vanes, that pump the oil from a well into which it drains by gravity, through a system of pipes to the bearings. There is a growing tendency toward an increased use of greases. These are fed through cups of two general types. In one the cover of the cup is screwed down at intervals by hand upon the contained grease, thus putting it DESIGN OF JOURNAL BEARINGS 99 under pressure and forcing it into the bearings. In the second type the mass of grease is under constant pressure from a spiral spring, or from a weight. Bath lubrication is not very common. Here the journal dips into a well of FIG. ig. DEVELOPMENT OF WHITE-METAL LINING SHOWING THERMOMETER LOCATIONS. FIG. 20. DISTRIBUTION OF TEMPERATURE IN A BEARING — SEE FIG. 1 9 (Lasche). oil, and thus has an opportunity to take up all that it will. This is the method of lubrication used by Tower in his celebrated friction experiments. Flooded lubrication is the supplying of oil to bearings by means of a pump, but at no very great pressure. A typical example is the system used in oiling lOO BEARINGS AND THEIR LUBRICATION the interior bearings of shafts and gears in certain constant speed drive milling machines; also some forms of steam turbine bearings. Forced lubrication consists in supplying oil to bearings under pressure usually from 15 to 25 lb. per square inch for horizontal bearings. As a rule the oil is cooled by artificial means, although this is not inherent in the system. It is important that the oil shall pass through the bearing at such a rate of flow that it has time to absorb the heat of friction and take it away. This is the reason for introducing cool oil at a point not far distant from that of maximum pres- sure. In this way intense local heating is avoided. Fig. 19 is reproduced from Lasche's work and shows the development of a white metal lining. The numbered points show where thermometers were brought very close to the inner surface of the lining, and temperature readings taken. Fig. 20 shows in a solid diagram the observed temperatures. The mini- mum temperature was at 12, and the maximum at about 19, where the pres- sure was greatest. OIL GROOVES Oil grooves should be designed according to the following principles: A, The general direction of grooves should be at right angles to the direction of motion. B, They should stop short of the ends of the bearing, to prevent leak- age. C, For conditions of considerable speed no groove must pass length- wise of the bearing in the region of greatest pressure; even a diagonal or cir- cumferential groove should not be used on the pressure side as this will to some extent at least reduce the oil film thickness. It is quite possible that for very slow speed bearings grooves in the pressure area are best. D, The edges of all oil grooves and the meeting edges of split boxes must be carefully rounded over. Fig. 21 shows the oil grooving practice of the Westinghouse Electric and Manufacturing Company. The grooves vary from 3/32 to 1/4 in. in width and from 3/32 to 1/8 in. in depth. If grooves are cut clear through to the end, oil leakage will take place. At the same time, if a journal and its bearings are very closely fitted, it may be necessary to scratch or score the surface from the end of the groove out in order to provide an air vent. The grooves at the end of the lower half of the bushings are not brought together, but are separated by some little distance. This is in the region of the greatest oil pressure. If this area is cut across by oil grooves, it interferes with the formation of a proper film. At the meeting point of the halves of the bearing a large V-shaped groove is sometimes cut, which serves as a reservoir of oil, tends to relieve the side pressure on the journal, and aids materially in good lubrication. DESIGN OF JOURNAL BEARINGS lOI At the same time there is a wide diversity of opinion in regard to grooves in bearings. Tradition says put them in, and this is usually done. Yet some engineers claim that there should be no grooves in the half of the bearing that One Ring Type -Leugth of. Beariug- Waste Packed X ting purposea X ""^ i ,^. ^ Top of Bearing Omit Groove on this eud for Blct. Bearings When less than ?4"oinit end Grooved. X Details of Grooves Overflow for Automatic Feed. ■/ Special - Used when necessary to View both Oil Ringg /. from common sight Hole. ^ 1^ igth oi J3.eiulug ARRANGEMENT AND DETAILS OF OIL GROOVES. supports the load. Milling the box at the split is favored as this tends to form a small reservoir. Oil grooves are by no means confined to the bearings, but are often cut in the journals, as shown by Fig. 22 from the practice of the Step toe Shaper Companjv Journals thus made are run by them in cast iron bearings with success. It is the H Oil Grooves 't%UA^Il FIG. 22. ARRANGMENT OF OIL GROOVES IN A JONRNAL. practice of one English builder of heavy engines to mill a flat spot lengthwise of the journals to aid in their lubrication. This undoubtedly distributes the oil over the entire surface of the bearing, and while successful at low speed, is prob- ably questionable for higher speeds, as the coming of the flat spot to the region I02 BEARINGS AND THEIR LUBRICATION of maximum pressure must tend to break down the oil film at each revolution. In machine tools the Osgood system of oil grooving has been used to a consider- able extent. It consists in cutting closed spirals, either single or double, in the lining. PROVISIONS FOR RADIATING AND DISSIPATING HEAT The heat radiating and dissipating capacity of a bearing' must be considered if the engineer or designer is dealing with comparatively high pressures and high velocities. It has been pointed out that the customer usually determines the maximum temperature at which a bearing may run, 140° F. being a common limit. As we have seen, the coefficienj; of friction for all rough calculations may be taken as constant for the higher commercial speeds. This being so, the maximum limit of speed is determined by the amount of heat produced by friction in a unit of time, which must not be greater than the amount which can be dealt with by radiation and dissipation in the same unit of time. Thus an increase in the speed of the shaft under these conditions is only allowable when the excess of frictional heat resulting therefrom can be carried away by air drafts or oil or water cooling. Lasche conducted a long series of experiments to determine the heat dissipa- tion from journal bearings. From his results Fig. 23 has been plotted with a transformation of the factors into English units. Three curves are shown. The first is for a bearing surrounded by a thin cast-iron housing so that the bearing surface and the upper surface of the housing can be considered as at the same temperature. In this case it is better to assume that the radiation reaches its lowest limit. Curve 2 is for a standard type of bearing in still air and is the most useful of the three. It can be taken as representing an average for the radiating capacity of ordinary bearings and used as a basis from which to calcu- late friction work and to determine whether or not a bearing with its natural capacity of radiation can dissipate the heat liberated by friction or whether artificial cooling should be resorted to. Curve 3 shows radiation from well ventilated bearings with large masses of iron such as are found in turbo -generator construction. The unit work of friction or energy liberated in the form of heat, measured in foot pounds, is given by the formula fw V, in. which/ is the coefficient of friction, w the unit pressure in pounds per square inch of projected area, and V the rubbing velocity in feet per minute, or letting H equal the rate of radiation per square inch of projected area expressed in foot pounds per minute; H=fw V. If in the design of a journal bearing the value of the coefficient of friction,/ can be determined for the particular working conditions of lubrication, then from the above equation the unit of radiation can be determined by computation DfiSIGN OF JOURNAL BEARINGS 103 If we then compare the unit thus found with the radiation as traced from Fig. 23 for the determined permissible rise in temperature of the bearing, it can at once be seen whether the natural radiation is sufficient to maintain a suitable tem- perature, or some artificial means of cooling must be employed. c ^ il H m S 5 ;:zu 200 1 180 ; ^ ^ / y X 160 / / / / / / 140 / / / / 60 / ^ A w ^ 120 1 / •0 c <^ /' 100 1 /i Sy A y />s \^ ^ , 80 [5 / ^ J # 4^ ^ ■1 / / ^0^ ^' 60 1 / / ^ f -^ if / A' .'* 40 y / /' / / 20 V / / 120 240 360 480 600 720 840 960 Radiation in Foot Pounds per Minute per Square Inch of Projected Area of Bearing Surface. FIG 23. CHART SHOWING HEAT RADIATED PER UNIT OF TIME FOR THREE TYPES OF JOURNAL BEARINGS. Good judgment must be exercised in making use of these curves, for the conduction of heat to a bearing along its shaft, as in steam generating ma- chinery, may play an important part in the amount of heat that must be dissi- I04 BEARINGS AND THEIR LUBRICATION pated. Again, if the bearing is surrounded by large masses of iron, with a large amount of exposed radiating surface, the bearing will get rid of more heat than if the surrounding walls are comparatively thin and the radiating surface comparatively small It is sometimes necessary merely to increase the 126 108 90 P. s ^ H g fl \- Ti n •o to li w ^ > ?i 54 rt .^ g O U h bo cS ^ Pm u S 20 18 70 ^ j^^ ^^ 120 0— ^ ^ ^ ^ 60 y y ^ 1000 y / ^ ^ / / ^ / / ^ 800 / / / y' ^ ^ ^ / / y ^ / / ) / ^ / / / / ^ y 600 At\ / / / / y ^^ ^- -^ / / / y ^ -^ ^ / / / / y \^ r / / ^ y 30 / y — , 400 / / y ^ / / ^- ?C^ ^t^ Uxi- / ^ 'v^^ r^^ ?n / ^ ^ ^^^ ^ y' / / 10 10 90 100 !0 30 40 50 60 70 80 Pressure Lbs. per Sq. In. Proj. Area. FIG. 24. TEMPERATURE RISE OF OIL RING BEARING IN STILL AIR— ROOM TEM- PERATURE 25 DEGREES CENTIGRADE, 77 DEGREES FAHRENHEIT. size of the oil reservoir. Again, an increase in the extent of the surface of a bearing housing may be all that is required to cause a bearing to run within its permissible rise in temperature. Another factor is the proximity of rotating engine, generator, or motor parts, which set up windage and cause a current DESIGN OF JOURNAL BEARINGS 105 of air to pass over the bearing. All of these factors, as well as any others which directly affect the radiation or dissipation of heat must be taken into account in determining the temperature at which a given bearing will run in service. The underlying principle may be stated thus: The work of friction per unit of bearing area must not be greater than the rate of radiation for the same unit area, and for the determined working temperature. Lasche^ developed an equation expressing the relationship of the coefficient of friction, unit pressure, and rise in temperature, which transformed into eo 72 -o 54 « 36 bo P U ^ 18 en 1 120( Ft 40 ^ 100( Ft ^ •^ 30 ^ y 800 Ft. y y ^ ^ ^ ^ / ^ y 600 Ft. y ^ -^ ^ 20 / -X y -^ X ^ 400 Ft. / ^ ^ ^^ 7^ n. ,— - y ^ rsv bs f5 Tyt 10 ^ ^ -^ ^^ ^ r 10 20 90 100 40 50 60 70 8 Pressure. Lbs. per Sq. In. Proj. Area. FIG. 25. TEMPERATURE RISE OF OIL RING BEARING FOR WELL VENTILATED CONDITION ROOM TEMPERATURE 25 DEGREES CENTIGRADE, 77 DEGREES FAHRENHEIT. fw= English unit is: /w(/— 32) = 51.2, in which /is the coefficient of friction, w the pressure in pounds per square inch, and / the final bearing temperature. By transformation /-32 In the section dealing with values of the coefficient of friction it was shown that for conditions wherein a perfect oil film can be formed the f rictional resistance for a given velocity is practically a constant and independent of the pressure. Thus, iPor a valuable discussion of this equation see Elements of Machine Design, Kimball and Ban, page 247. lo6 BEARINGS AND THEIR LUBRICATION as far as ordinary designing is concerned, we may assume that the coefficient of friction for perfect lubrication and all velocities greater than 500 ft. per minute is a constant, and the equation given above may be used without serious error. By adding the factor V to each side the right hand member becomes equal to the frictional loss per unit of projected area expressed in foot pounds per minute, as below: fwV= t-32 This equation can now be used to compute the heat liberated by friction in a perfectly lubricated high speed journal bearing for a given velocity in feet per minute at a given working temperature in degrees Fahrenheit. This gives us a unit of radiation to compare with the curves of Fig. 23 to determine whether or not artificial cooling must be used. The General Electric Company has carried on an extensive series of tests to determine the rise in temperature for bearings of a type used by them in electrical machinery and under various conditions; the results of these experi- ments are given in the curves of Figs. 24, 25, and 26. The first is for a ring oiling bearing running in still air, as in a tightly closed room with an air tempera- ture of 77° F. Fig. 25 is for the same kind of bearing well ventilated. Fig. 26 is for a bearing running with forced lubrication and three different rates of oil feed. In each case the lubricating oil was what is known as ''dynamo and shafting oil," having a specific heat of 0.4, a weight per gallon of 7.27 lb., and a viscosity at 70° F. of 80.3 Doolittle. Curves 24 and 25 are for a range of speed from 400 to 1200 ft. per minute. The curves of Fig. 26 are for rubbing speeds up to 3,000 ft. per minute and for four different pressures per square inch of projected area, namely, 35, 50, 75, and 100 lb. For pressures and speeds beyond the limit of this chart it is wise to resort to water jacket cooling to remove the liberated heat. RULES FOR BEARING DESIGNS This section on bearing design can be closed in no better way than to give the rules for bearing designs as adopted by the Committee of Mechanical Design of the General Electric Company, issued in connection with the preceding charts. These rules are applicable to the design of the great majority of generator and motor bearings. The temperature of any bearing having a given radiating surface is deter- mined by the amount of work lost by friction. This work (in foot pounds per minute) =/ W V, where, /= coefficient of friction, W=\o2id on the bearing in pounds, V= surface speed in feet per minute. DESIGN OF JOURNAL BEARINGS 107 1 — \ — \ — m — r\i — 1 — 1 — \ — \ — \ — \ — \ — \ — \ — 1 — \ — \ — \ — 1 — \ — 1^ \l \l ^'' X \_ \ \ \ \_ \\ A g \ \ V \ ' ^ y \ \ \ \ A ^ A VI "4 "4 "^ 2 \% % \ ^ K %_ h 4 %. \ \ ^ \\ \ \ g A ^^ ^ J ^ A \ V \ g Aa^^\3 ^t^3 t A^a5 V-A ^^^-^y sn. 8 V\ W^t^ \F XASvlilli 1 \ \\i\V V V^ - \ \ Vk\K \ \ ^ ^ \ K N \ \ a V\ \V n~^L y H ~^~^ AVAV \ V ^ g ''^^ff-^AA ^ \ "^X i\ \ V \ \aJ^%_\ \ \\^-\*^^ 5 \ \ \ > _^%r «> s Q < ti il H W P M o '^ g o w w W M W W fU CM O w m Q •^uaQ _^g2 aAoqB •:}ua3 '3a(j 3si'y[ ajn^BJOduiaj, M ■* CO 00 t- »0 Ji^ rH Io8 BEARINGS AND THEIR LUBRICATION Assuming that the load remains fixed it is evident that the work lost and therefore the amount of heat generated becomes smaller as/ and V become smaller; or, in other words the smaller the bearing diameter for a given load the less it will heat. Another advantage of small diameter is that the coefficient of friction decreases with the surface speed. It is, therefore, very desirable in laying out bearings to keep the diameter as small as possible, consistent with sufficient strength of shaft and suitable deflection of the journals both inside and outside the bearings. It is also very desirable to so dimension bearings that they are fairly well loaded, in order to avoid bulky machines and also because the coefficient of friction rises quite rapidly when the load is less than 50 lb. per square inch of projected area. When calculating the projected area of any bearing, especially if it is to be heavily loaded, the amount of space lost through the drain grooves at both ends must be deducted. This is particularly important when the length of the bear- ing is small in proportion to the diameter. It is also necessary — and this applies to all forms of lubrication — that there be no sharp corners on the edges of the oil distributing grooves or channels, but that these be gradually cased off so that the oil can be drawn in between the journal and the bearing. Sharp corners are invariably oil wipers and often absolutely prevent proper lubrication. The heat generated in any bearing may be dissipated: 1. By radiation from the housings and conduction by the shaft. 2. By forcing cooled oil through the bearing. 3. By surrrounding the bearing by some form of water jacket. Other forms of artificial cooling are possible but these described under 2 and 3 are the most commonly used. I. — BEARINGS WITHOUT ARTIFICIAL COOLING Bearings of this type are usually lubricated by oil rings or similar devices or by gravity feed. It is essential that an abundant supply of oil be delivered to all parts of the bearing by suitably arranging the channels so that a perfect film will be maintained at all times between the journal and bearing, and that there is no opportunity for the oil forming this film to escape through openings or grooves at the points of greatest pressure and thus allow the metals t6 come in contact. The heat generated in bearings having no artificial cooling is conducted away and radiated by the housings. The great variation in the design of bearing housings and the different conditions of ventilation, etc., make it extremely difficult to predetermine the ultimate temperature of such bearings with any great accuracy, and it is always necessary to allow a considerable margin of safety. DESIGN OF JOURNAL BEARINGS 109 Fig. 1 1 covers the range of pressure and speed ordinarily permissible in this type of bearing, while Figs. 24 and 25 show the ultimate temperatures for differ- ent speeds and loads. These curves were made up from the readings obtained from special bearings and afterward checked by the test records of a large number of machines — both of the pillow block and shield types — which have gone through the testing department during the last few years. Fig. 24 shows the temperatures to be expected under the most unfavorable conditions, that is, of a bearing so situated that no current of air can circulate about it, and there- fore cooled by radiation only. There is, however, a considerable circulation of air about most machines, due to the fanning action of the revolving parts, and the ultimate temperatures to be expected in such cases are shown on the curves of Fig. 25. These curves apply to the great majority of open generators and motors, both of the pillow block and end shield types. When the machine is enclosed, or the free circulation of air in any way interrupted, higher tempera- tures will result, until finally the conditions of Fig. 24 are reached. A part of the heat of the bearings of motors and generators is usually con- ducted away by the shaft and radiated by the spider and other revolving parts. When machines are totally enclosed or are connected to other machinery whose temperature is high, heat may be transmitted through the shaft to the bearing, thus raising the latter's temperature, and due allowance must be made for this. 2. — COOLING BY FORCED LUBRICATION When bearing pressures and speeds are unusually high it is often necessary to force oil under pressure into the bearings and advantage is often taken of this to keep the heating down by artificially cooling the oil. The temperature of bearings is also sometimes kept down by forcing cooled oil through them, when otherwise forced lubrication is not necessary. This method, although used to a very considerable extent, is usually not as efficient as a water jacket. For all practical purposes, it may be considered that the entire heat generated is taken away by the oil, and it is therefore possible to predetermine the bearing temperature with considerable accuracy. Fig. 26 shows the ultimate temper- ature of bearings using the quantities of oil most commonly pumped through, and with the assistance of these curves the necessary amount can be determined. Intermediate speeds and pressures can be easily interpolated. For pressures and speeds beyond the limits of this table, it is advisable to resort to water jacket cooling. In arranging bearings for this form of lubrication, care must be taken to force the oil to the point where the work is being done, as otherwise the oil coming from the bearing may be comparatively cool, while the bearing itself no BEARINGS AND THEIR LUBRICATION is much too warm. In addition to this, it means that an excessive amount of oil must be pumped through the bearing requiring unnecessarily large pumps, pipings, etc. Experiments with such bearings show that when oil begins to run out of the ends quite freely, nothing is gained by forcing through a larger quantity. 3. — COOLING BY WATER JACKET A properly designed water jacket will carry away a very much larger amount of heat than will any form of forced oil lubrication, as the specific heat of water is very much higher. As is the case of forced oil lubrication the ultimate temperature of a well designed water jacket bearing can be very accurately determined. It is essential that the pipes or channels be located as close as possible to the surface of the lining where the work is being done, in order that the heat generated may be absorbed without danger of damage to the lining. Water circulated at some distance away from the lining surface is of compara- tively little assistance, as heat may be generated so rapidly that the lining will be destroyed before the heat strikes the jacket. The water passages must also be so arranged that an even and continuous circulation is kept up in all parts. With properly constructed passages, it is safe to assume that heat may be removed at the rate of from 3 to 5 horse power for each gallon of water circulated per minute, and where the conditions are unusually good and the jacketing carefully arranged from 10 to 12 horsepower per gallon can be dissipated. With water jackets any suitable method of lubrication may be used which will insure at all times a good film of oil between journal and bearing. Lubricating oils. SECTION VI Lubricants for Bearings The subject of lubricants for bearings is one of great complexity. The chemical analysis of oils and greases is very difficult and is work that can be successfully carried on only by a specially trained chemist. Many oils are mixed or blended and for the user there are no simple tests by means of which he can determine just what the constituents of a particular kind of oil are. At the best he can only make a few simple physical tests and then study the behavior of the oil in use. The following paragraphs will not touch upon the chemical side, but merely point out some of the important kinds of oils and greases, their applications, and some of the physical characteristics that they should have. CLASSIFICATION OF LUBRICANTS All lubricants fall into three classes: oils, greases, and solids. High pressure cylinder oil, Low pressure cylinder oil, Engine oil, Dynamo and shafting oil, Machinery ^oil for heavy duty, Machinery oil for light duty, Loom oil, Spindle oil. Non-fluid oil. Soft grease. Hard grease. f Graphite, Solid lubricants < Sulphur, [ Talc. The oils are again divided into three divisions as regards origin, mineral, vegetable and animal. The mineral oils are petroleum products and in this country are spoken of as of two kinds, oils with a paraffine base and oils with an asphalt base , the Pennsylvania product being representative of the first class and the Texas product of the second. Oils from certain localities have come to be recognized as standards for some purposes, as, for instance, the best cyl- inder oils are supposed to be made from Franklin stock. Among vegetable oils, colza, rape seed, olive, and palm oils are good lubri- cants. Castor oil has good lubricating qualities but oxidizes and gums. Among animal oils, lard, tallow, sperm, and whale are all used both pure and for compounding with mineral stock. Ill Lubricating greases. 112 BEARINGS AND THEIR LUBRICATION Graphite is the only solid lubricant that is used to any extent, and it is recog- nized as one of our most valuable lubricants. It is employed in a solid form fitted into recesses in bearing linings, in powdered form mixed with oil and in the form of deflocculated graphite mixed with oil or water. This last form is a special product of the Acheson Graphite Company. Sulphur was formerly employed for emergency use on over-heated bearings, prior to the common use of graphite. Ground talc or soapstone free from grit has had a limited use for the same purposes. Water has been used as a lubricant, particularly for the steps of vertical water turbines and vertical steam turbines. Its action, however, is not thai of a true lubricant. In the vertical steam turbine and some water turbines it serves as a means to separate the plates of the steps. It is pumped in under pressure and floats the load. In other water turbines having wood block and metal cup bearings it removes the heat liberated by friction and prevents the bearing from burning and cutting. Many other kinds of oils are used but they are often looked upon as adul- terants; for instance, cotton-seed oil is frequently mixed with lard oil. As mentioned above, their detection is very difficult and should only be undertaken by the trained chemist. Just what oils may be used as adulterants at a given time depends largely upon the relative market values of the oils. Turning to physical tests, the oil user is most largely interested in the qualities of viscosity and body. The viscosity of an oil is the property that determines the rate of free flow. The property, body, is difficult to define. The user gets his idea of it by dipping his thumb and forefinger into the oil and rubbing them together. The degree of oiliness or greasiness is an indication to him of its body. It has been defined as that property of an oil that influences the change in viscosity when the oil is under pressure; or as that property that influences the intensifying of the viscosity in that portion of the oil film within the region of attraction of the surface molecules of the metals of bearing and journal. Professor Kingsbury, Trans. A,S. M. E., 1903, page 147, gives these rela- tions of viscosity and body. With increase of Where the viscosity is effective, the coefficient of friction Where the body is effective, the coeflicient of friction Pressure. . . . Speed Temperature. Viscosity Body Decreases. Increases . Decreases. Increases . Increases. Decreases. Increases, Decreases. Decreases. LUBRICANTS 113 From this it is seen that the effects of body and viscosity are in nearly all respects diametrically opposite and it must necessarily be very difficult to derive reliable information regarding the lubricating values of oil from friction tests in which the effects of viscosity and body are not separately recognizable. Thus we have but very litde information that is of value on the body of oils. As a general rule, mineral oils have much less body than animal and vegetable oils. Thus of a mineral oil and lard oil of approximately the same viscosity the lard oil would have decidedly the greater body. On the other hand, some cylinder oils of mineral origin greatly exceeding lard oil in viscosity will also have a greater body. It is not uncommon to mix a small percentage of animal oil with mineral oil in order to increase the body. For illustration, turn to the latter part of this section where specifications for oils used by the United States War and Navy Departments are given. During the past few years there has been a tendency to use more and more grease for the lubrication of machinery bearings. In comparing the relative qualities of a grease and an oil, the former must be tested at the running tem- perature of the bearing where it is to be used. In general the lower the speed the thicker should be the lubricant. With heavy pressures and low speed it is difficult to form a film of oil and maintain it. This is specially true when the machine is started after being shut down. While idle the oil is squeezed out from between the journal and bearing surfaces and there is metallic contact which produces abrasion as soon as motion begins with an attending rise in temperature and thinning of the oil if that is used. For such work greases are especially suitable, as the solid or semisolid substance is not squeezed out from the journal in the way a liquid is acted upon. Professor Goodman, Proc. Inst. C. £., Vol. LXXXIX, page 432, gives com- parative results of three oils and one grease. The nominal pressure per square inch was 85 lb. Lubricants Natural coefficient of friction Coefficient of friction after standing 16 hours Machinery oil Valve oil J A Thick and viscous (^ B Hard solid grease o . 0084 0.0252 0.0329 0.0350 0.192 0.147 0.1715 o . 0(;o It will be noted that the coefficient of friction after standing is only about half as great for the grease as for the oils. 114 BEARINGS AND THEIR LUBRICATION On the other hand, a grease should not be used unless the conditions for feeding it to the bearing are such that it will reach all parts of the rubbing surfaces. An incident given by W. H. Booth, American Machinist^ Vol. XXXII part 2, page 790, bears on this point. *' A set of pumps driven by a worm and wheel were required to be tested for general sufficiency. The worm was below the wheel and there was a thrust bearing in a narrow cylindrical recess out of the oil casing which inclosed the worm and wheel. A thick oil was filled into the casing and the test was made. The thrust bearing soon grew unbearably hot, and it looked as if the pump would be condemned. Quite suddenly the box went comparatively cool and no further heating took place. The cooling coincided with the thinning or liquefying of the oil. At first this was so very thick that it could not freely enter the thrust chamber against the tendency of the worm to draw the oil away from the thrust recess and pile it up on the opposite side of the casing, but this piling up ceased when the oil became hot and thin, for it could then get well into the thrust collars, and these cooled at once, and the test was declared good. A thinner oil would have run from the start in all likelihood, and this is confirmed by the experience of many, that a thin oil is a better lubricant than a thick oil up to the point where the pressure begins to squeeze the oil from between the surfaces. " From Tower's experiments, Proc. Inst. M. E., 1883, the following comparison of five oils and one grease is taken: Lubricant Mean resistance, pounds Sperm oil Rape oil Mineral oil Lard oil Olive oil Mineral grease Per cent 100 106 129 135 135 217 The second column gives the actual frictional resistance at the surface of the journal per square inch of bearing area at a speed of 300 revolutions per minute, and for nominal loads from 100 to 310 lb. per square inch. The percentage figures may be taken to represent the relative thickness or body of the various oils, and in their order, if not in numerical proportion, the relative weight carrying ability. Thus it is seen that the sperm oil has by far the greatest lubricating value and the least weight carrying ability. It must be recognized that these values, however, hold only for the conditions LUBRICANTS I15 of investigation. For greater pressures or higher temperatures the relative order of excellence might easily be changed with some of the heavier oils leading. Another very important quality of any lubricating oil is that it shall be free from acids and alkalies. Animal and vegetable oils must be watched more closely in this particular than mineral oils. As a rule mineral oils have no action on metals unless they contain sulphur, and then there is no liability of harm except at high temperatures. Among the organic oils lard oil and pure animal oils in general have little action; on the other hand, olive and vegetable oils generally produce corrosion. Other points to be investigated in lubricating oils are the flash and fire points, the amount of contained water, the amount of tarry ancjl suspended matter, the volatility and action under saponification. For good practice in regard to these points see the United States War Department specifications in the latter part of this section. QUANTITIES OF OIL USED As is to be expected, there are few data available for the quantity of oil used for bearing lubrication. The variation in type of bearing, rubbing speed, pressures, quality of the oil and its method of application, all effect this question. As representative of American practice in the lubrication of steam engine cylinders. Table 35 is reproduced from Power, Feb. 15, 1910, page 303. A glance at the last column shows that the average is about one pint for every million square feet of cylinder surface rubbed over. For bearings lubricated by forced pressure and water cooled 1/50 gal. per square inch of projected area per minute is good practice. For bearings with flooded lubrication and cooling of the oil i/io gal. per square inch of projected area per minute is frequent practice. For the step bearings and guide bearings of Curtis vertical steam turbines Table 36 can be considered as presenting average practice. ii6 BEARINGS AND THEIR LUBRICATION 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 I 44 45 46 47 48 49 50 Esti- mated h. p. 1,350 67s 67s 975 975 650 650 575 875 450 650 290 290 200 200 400 450 290 200 225 225 225 450 200 180 160 160 70 450 4SO 290 290 290 200 160 290 200 130 250 160 160 130 130 ISO 130 120 160 130 130 100 Description of engine Cylin- ders Stroke I Corliss triple expansion ; 20-34-52 i Cross compound. I Cross comp. St. Louis Corliss. Corliss compound Hamilton Corliss cross comp. Hamilton Corliss cross comp . Corliss compound Corliss tandem compound. Russell cross compound. . . Cross comp. condensing. . . > Cross compound Harris standard cross comp . . Harris standard cross comp . . Cross comp. St. Louis Corliss. Corliss cross comp Phoenix Iron Works Russel tandem comp Corliss cross compound. Buckeye tandem comp IngersoU cross compound American ball compound Westinghouse automatic comp. Cross comp. Meyer valves Tandem comp. piston valve. . . IngersoU cross comp Piston valve Double eccentric corliss Corliss Harris Corliss Greene Sioux Corliss Harris Corliss Corliss New Brown Atlas single valve automatic. . . Ames Corliss High speed automatic High speed automatic Piston valve engine Slide valve throttling governor. Single valve auto. Fitchburg . . Atlas four valve . , Harris Corliss h. p. 26 1. p. 52 22-44 24-44 20—36 20-36 18-34 24-42 16-30 18-36 h. p. 18 1- p. 34 h. p. 16 1. p. 28 14-28 18-30 14-24 I3-20i h. p. 16 1. p. 30 12-21 18-30 102-20^- 11-19 11-18 iof-i8 * 8-12 30 30 24 24 24 20 18 24 20 16 22^ 18 18 16 16 I7i 16 15^ 18 16 16 14 60 48 48 48 60 44 44 48 48 24 42 36 36 18 18 42 36 18 20 36 36 21 24 12 1 1 14 18 12 48 48 48 48 42 44 42 48 36 36 30 16 36 16 16 16 30 14 15 36 38 Revolu- tions per minute 65 81 85 6S 100 100 85 6 5 150 72 86 86 200 200 8S 84 84 84 200 100 28s 300 16s IIS 150 95 84 100 105 6S 102 125 los 230 100 250 250 220 126 27S 220 100 92 280 Steam pressure 140 140 100 sup IIS 115 ISO 150 50 no no IIS 115 80 95 125 65 . Sq. ft. rubbed over per hour 1,084,000 265,800 531,600 706,500 694,500 644,400 644,400 555,900 539,600, 433,3CO 428,100 146,000 276,000 150,800 263,900 393, coo 380,000 376,100 368,200 126,800 237,500 363,100 302,000 277,600 259.300 175.400 172,100 94,3°° 357.800 318,000 308,100 302,000 245,500 230,200 207,900 196,200 196,200 188, 8co 183,500 173,700 169,800 167,800 167,800 161,500 158,200 156,200 iSS,9co 150,900 146,600 143,200 Table 35. — Steam Engine Cylinder Lubrication (Power) LUBRICANTS 117 No. Oil used Name or description Price per gal. Amount used per hr., pints _ . I Pmts per Cost per. I ^. ^ , , estimated hr., . 1000 h. p. cents , hours Cents per 1 Pints per estimated | million 1000 h. p. I sq. ft. rub- hours I bed over 12 I? 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 SO No. 725 cyl. comp Capitol oil No. 650 dark valve cyl, "600 W." Oil of beeswax; 600 fire test 1 cyl. stock and acidless tallow J Heavy body comp "600 W." Improved high pressure cyl . . . Best grade cyl Best grade cyl Harris h. p. valve Rarus No. 650 dark valve "600 W." "600 W." Best grade cyl. Best grade cyl. "600 W." "600 W." No. 725 cyl. comp Best high grade mineral. High grade mineral "600 W." "Cyl. oil No. 10." "Cyl. oil No. 10." 600 W. "Cyl. oil No. 10." Harris high-press valve. Buckeye cyl Capital oil High grade mineral . High grade mineral. Light colored oil ... . Capitol .28 .75 •35 .50 .28 .60 90 .28 .40 .60 •75 .60 .60 .60 0.69 0.466 0.666 o. 104 0.621 o. 21 0.25s 0.46s 0.2 0.25 O . 1 1 I 0.64s 0.323 o. 184 o. 131 0.09 0.417 0.167 o. 167 0.64s 0.323 0.2s o. 167 0.273 0.083 0.012 0.037 0.C83 0.417 0.5 o.cgi 0.091 0.369 0.91 o. 125 o. 19 o. 167 O. 22 Rarus cyl Harris std. grade. Compounded oil. . Model cyl 0. 1 o.ois 0.22 0.03 o. I 0.138 0.094 o.ts 0.15 0.04s 2.42 4.37 2.91 0.65 2.17 1-57 2.87 i^63 1 .00 1.87 4-34 2.42 1.38 0.56 3^13 0.90 1^25 4^34 2.42 1.87 I^2S O.C75 0.23 0.62 1.30 1.56 2.77 0.94 0.52 0.88 0.525 0.96 1 . 12 0.094 o. 14 0.44 0.48 I . 12 0.81 0.511 0.69 0.99 0. 107 0.639 0.323 0.392 0.81 0.229 0.555 0.I7I 2 . 22 I . II 0.92 0.65 0.225 0.927 0.58 0.835 2.86 1-43 I .III 0.371 1.365 0.461 0.075 0.231 1. 185 0.927 1 . Ill 0.314 0.314 1 . 272 0.45S 0.781 0.655 0.83s 0.908 0.600 1.375- 0.625 0.115 o. 169 0.2 0.769 1. 15 0.587 1. 153 1. 153 0.4s 1.79 6.46 4.33 0.67 2.23 2.42 2.84 1. 14 4.16 15.80 8.32 6.90 I .41 6 .95 3 .10 6 25 21 45 10 72 8 33 2 78 4 32 47 I 44 8 89 2 90 3 47 5.86 2 .61 6.81 2. 1 6.00 7.C4 0.72 1 .06 3 36 4 02 8 64 6 20 Table 35. — Steam Engine Lubrication {Power). ii8 BEARINGS AND THEIR LUBRICATION Esti- No. mated h.p. SI 240 52 160 53 130 54 100 55 130 56 100 57 200 58 160 59 100 60 85 61 no 62 100 6;j 1^0 64 70 6S 70 66 100 67 70 68 70 69 70 70 70 71 70 72 50 73 100 74 50 75 70 76 70 77 40 78 50 79 50 80 85 81 SO Description of engine Cylin- ders Revolu- Stroke tions per minute 23 106 42 70 36 90 14 260 42 75 36 100 30 80 30 85 36 90 18 190 14 212 14 223 32 85 12 300 21 i6s 36 80 18 180 36 90 36 88 36 8S 34 90 14 260 20 130 12 300 30 100 14 210 12 300 12 240 IS 170 17 los 12 250 Steam pressure Sq. ft. rubbed over per hour Wright Corliss Nordberg Corliss Robt. Armstrong automatic. . . Corliss Corliss Slide valve Greene Variable speed, St. Louis Corliss Atlas automatic Ames Automatic piston valve Nordberg Corliss Ideal Buckeye Corliss Atlas single valve automatic . . . Corliss Bates Corliss St. Louis Corliss New Brown Atlas McEwen Fitchburg Slide valve Ball Atlas Center crank Slide-valve 85 125 135-140 125 no 100 no very wet 100 135-140 80 80 80 90 150 100 100 140,500 138,800 135.800 133.500 132,100 132,000 125,800 120,200 118,800 116,800 116,600 114.500 113,800 113,100 108,900 105,800 101,800 101,800 99,500 96,350 96,200 9S.400 95.400 94,400 94,400 92,500 84.750 75,400 66,800 60,900 31.410 Table 35, — Continued. — Steam Engine Cylinder Lubrication {Power) LUBRICANTS 119 No. Oil used Name or description Price per gal. I Amount used per hr., pints Cost per hr.. cents Pints per estimated 1000 h. p. hours Cents per estimated 1000 h. p. hours Pints per million sq. ft. rub- bed over " Eureka cyl. Capitol Val valine. . . , "600 W."... "600 W." "600 W." No. 650 dark valve cyl " 600 W." vacuum cyl Capitol High grade heavy Dark heavy cyl. with graph. Capitol cjd Buckeye cyl. Capitol cyl No. 650 dark valve c^l. . Franklin oil Capitol cyl Capitol cyl W. P. Miller's cyl. ccrr.p. Flake graph, with eng. oil. Porno oil "600 W." Premium valve oil . .35 .50 • 75 • 35 •35 ■75 •45 O. IDS O.I2S 0.2 0.15 0.138 0.06 0.018 0.03 0.04s 067 II 125 115 083 0.049 0.6 0.0835 0.2 0.04s O.I 0.055 0.2 0.04 0.055 0.083 0.167 0.045 0.083 0.125 0.091 0.55 1.87 I . 12 0.48 0.45 O. 22 0.28 0.50 0.48 0.78 •36 0.87 0.28 0.94 o. 24 0.87 0.37 0.31 0.34 0.94 Averages 0.437 0.781 1.538 1-5 1 .061 0.6 0.09 0.187 0.45 0.788 i.o 1.25 0.885 1. 186 1.428 0.49 8.57 1. 19 2.857 0.643 1.428 1 .1 2.0 8 786 185 175 9 66 47 82 3.41 14.38 11.22 3-71 450 1 .46 2.82 S.90 4-37 7.8 5.18 30.0 12.48 4.01 13.38 4.81 8.74 7.50 4.42 6.75 6.19 0.75 0.90 1.47 1 . 12 1 .04 0.45 o. 14 0.2s 0.38 0.57 0.94 1.09 1 .01 0.73 0.92 0.47 5.94 0.82 2.01 0.47 1 .04 0.58 2. 10 0.42 0.58 0.90 1.97 0.60 1.24 2.0s 2.90 Table 35. — Continued. — Steam Engine Cylinder Lubrication {Power), I20 BEARINGS AND THEIR LUBRICATION Rating Gallons oil per minute Pressure per square inch Kilo- 1 !^, ^^ |R. p. m.ibtep watts Guide Gear Total Step Baffler drop Pump- less line Gear and guides Unbalanced rev. wt. Step block 0. D., I. D. 750 1500 3750 Sooo 9000 14000 15000 20000 1800 1500 900 750 750 750 7SO 750 3 4 9 12 IS 18 20 25 1,5 X.7S 2.0 3.0 4.0 4.0 4.0 4.0 2.5 2.25 50 6.0 8.0 10. X2.0 14.0 7 8 16 21 27 32 36 43 290 320 525 825 900 900 850 950 60 6s 100 120 .50 ISO 150 .so 350 385 625 945 1050 1050 1000 1 1 00 120 120 120 120 120 120 120 120 10800 19700 65000 lOIOOO 148000 190000 216000 233000 8x5 3/4 11x8 16x9 1/4 i6xro i8i X 103/4 20x13 1/2 21x15 21x15 Table 36. — ^Veetical Steam Turbine Lubrication. FILTRATION OF LUBRICATING OILS Oil after having passed through machinery bearings seems to lose a large proportion of its lubricating power, besides being contaminated with foreign particles, such as metallic dust, carbon, dirt and the like. It is a difficult matter to restore it to its original purity and value. The usual shop method is by filtration, but the filtered oil is seldom used alone as a lubricant. It is usually mixed with a certain amount that is fresh. The three common methods of filtration are by capillary action, by gravity filtering, or by pressure filtering. Of these the second is the most common. G. W. Bissell describes a capillary filter in the Trans. A . S. M. E., Vol. XVII, page 293. "The oil filter that I used. consists of two shallow rectangular tin trays and some wide lamp-wicks. One of the trays, slightly smaller than the other, is supported within it on two blocks, which raise it an inch or so from the bottom of the larger pan. The wicks are laid in the upper pan, so as to hang over its edges into the lower pan. The oil to be filtered is poured into the upper pan. A drip cock in the lower pan, and a cover for the whole, complete the apparatus. No quantitative tests have been made, but the results, as far as the eye can judge, are good.'* Gravity filtration consists in passing the used oil through mediums, such as mineral or animal wool, bone dust, vegetable or animal charcoal, hair felt, absorbent cotton and the like. Heat is usually applied, and the final condition of the filtered oil depends upon the cleaness of the material and the rate of its flow. It is doubtful if any of these materials can completely remove finely divided metallic particles. Pressure filtering is only used for very large quantities of oil. LUBRICANTS 121 SIMPLE TESTS FOR GREASES. Professor G. W. Lewis writing in the American Machinist, gives a few simple tests of greases as follows: There are two general types of greases; the so-called mineral, and the animal or tallow greases. Mineral greases should be made from a high grade refined petroleum oil and so treated that the least possible amount of foreign substances be used to bring it to a solid state. Lime soap is generally used and the mixture emulsified by mechanical sitrring or blowing air through the mixture. Some manufacturers use resin, resinous oils, graphite, soap-stone wax, talc, powdered mica, and asbestos, which are introduced into the grease for the purpose of creating an artificial body. Tallow greases should consist of some hard animal or vegetable fat, such as tallow, with a small amount of mineral oil. Here the mixture is solidified by the use of a soap, in this case a potash soap being usually used. The melting temperature of tallow grease is from 50 to 75 degrees lower than the melting temperature of the mineral grease. Some very inferior greases are made from oils which are the waste of soap factories and sewage disposal systems. All greases look good when the can is opened and nearly all will work well on a bearing for a short time, and then a poor grease will cause the bearing to heat, which decomposes the grease into oil and soap, especially if the per cent, of soap is high, and impurities have been used. Any adulterants used continu- ally accumulate in the bearings as the grease melts away, and cause gumming and sometimes scoring of the bearing. The consumer, to protect himself, should make the foUovdng simple tests on any grease he contemplates using, test for acids, alkalies, volatile matter, filling, water and gumming. The test for free acid is best made by melting a small quantity of grease and applying blue litmus paper. The test for alkali is made in the same way by applying red litmus paper. The test for volatile matter, such as naphtha or benzine, is made by heating a measured quantity to 150° to 200° F. and determining by weight the amount volatilized. The test for filling, or the amount of soap or other material, used to form a body, can be partially determined by melting a test-tube of the grease to a liquid state and noting the amount of residue or cloudiness in the bottom of the tube, which gives an idea of the amount of soap or other material used in bringing the grease to a solid state. A more accurate test is to take a small sample of grease and dissolve in gasoline, then add phosphoric acid in alcohol solution and if any soap is present, it will precipitate and give a relative idea of the amount in the grease To determine the amount of water in the grease, and to see if it will "wash 122 BEARINGS AND THEIR LUBRICATION out, " place a small amount in a linen cloth or a common handkerchief, fold in the form of a bag, and place in water and knead for five minutes. In this way the amount of water present can be determined, and also if the grease will wash out, which is very important, especially in paper making and laundry machines where water comes in contact with the bearings. The simplest method to determine the amount of gumming is to take a sherry glass, put in a few pieces of copper wire, and add concentrated nitric acid, then add the grease in finely divided particles to the solution, and stir until all the grease has been broken up. In the case of a tallow grease, the tallow will form a solid crsut on top of the liquid in about an hour; with a vegetable oil, the oil remaining will be about the consistency of butter, while with a mineral oil, the oil remains in a liquid state. The blackish substance that appears below the oil represents the impurities that cause gumming of the bearing. Next separate the acid from the oil and the black substance and add gasoline to the latter two. This washes out the oil and separates the gumming substances so that a fair idea of the quantity in the grease can be obtained. PENNSYLVANIA RAILROAD COMPANY SPECIFICATIONS FOR LARD OIL Two grades of lard oil, known in market as "Extra" and "Extra No. i," will be used, the former principally for burning and the latter as a lubricant. The material desired under this specification is oil from the lard of corn fed hogs, unmixed with other oils, and containing the least possible amount of free acid. Also from October i to May i it should show a cold test not higher than 40° F. on from lard of "mast" or distillery fed hogs does not give good results in service, and should never be sent. Also care should be taken to have the oil made from fresh lard. Old lard gives an oil that does not burn well, and also gums badly as a lubricant. The use of the so-called neatsfoot stock, either alone or as an admixture in making the "Extra No. i" grade, is not recommended. Neatsfoot oil is used by the Railroad Company when the price will admit, but it is preferred to have it unmixed. Shipments must be made as soon as possible after the order is placed. Also shipments received at any shop after October i will be subjected to cold test and rejected if they fail, unless it can be shown that the shipment has been more than a week in transit. Shipments of the "Extra" grade will not be accepted which I. Contain admixture of any other oils. II. Contain more free acid than is neutralized by 4 c. c. of alkali as described in the printed method. III. Show a cold test above 45° F. from October i to May i. IV. Show coloration when tested with nitrate of silver as described below. LUBRICANTS 123 Shipments of "Extra No. i" grade will not be accepted, which, I. Contain admixtures of any other oils. II. Contain more free acid than is neutralized by 20 c. c. of alkali as described in the printed method. III. Show a cold test above 45° F. from October i, to May, i. The cold test and the amount of free acid must be determined in accordance with P. R. R. standard methods, printed copies of which will be furnished upon application. The nitrate of silver test is as follows: Have ready a solution of nitrate of silver in alcohol and ether, made on the following formula: Nitrate of silver i gram. Alcohol 200 grams. Ether 40 grams. After the ingredients are mixed and dissolved, allow the solution to stand in the sun or in diffused light until it has become perfectly clear; it is then ready for use and should be kept in a dimly lighted place and tightly corked. Into a 50 c. c. test-tube, put 10 c. c. of the oil to be tested (which should have been previously filtered through washed filter paper), and 5 c. c. of the above solution, shake thoroughly and heat in a vessel of boiling water 15 min. with occasional shaking. Satisfactory oil shows no change of color under this test. UNITED STATES NAVY SPECIFICATIONS Following are the specifications for lubricating oil for marine machinery as issued by the United States Navy Department, September 13, 1906, and still in force. 1. Must be a properly compounded oil to form a homogeneous compound that will not separate under varying temperatures, and must consist of a pure mineral oil, and not more than 30 per cent, nor less than 20 per cent, of suitable non-drying fixed oils as may be best suited for lubrication. 2. The compounded oil must be free from rosin, tar, drying oils, sulphur, asphaltic or tarry bodies, soaps or oil thickners, water, grit, dirt or other sus- pended matter; and must be free from mineral acid, and must not contain more than 2 per cent, of free oleic acid. The speciiEic gravity to be between 0.915 and 0.927 at 60° F. ^. As a lubricant the oil when tested on an oil-testing machine, owned and operated by the Government, having a standard brass bearing of about 9 sq. in. projected area on a polished steel mandrel making about 160 revolutions per minute with a surface speed of about 250 ft. per minute must, to be satisfactory, perform as follows: 124 BEARINGS AND THEIR LUBRICATION (a) The temperature of bearing must at no time during the test be per- mitted to exceed 130° F. Only sufficient oil under test to be applied to prevent excessive friction and heating of bearing. (b) The average load on the bearing for two hours must be at least 300 lb. per square inch of projected area of bearing. (c) The quotient found by dividing the product of the average total pressure on journal and surface speed of journal in feet per minute by the weight of oil in grains used for lubrication in the test must not fall below 325,000. (d) The duration of test must not be less than two hours nor more than two and one-half hours, and but one test will be permitted for any one lot of oil. 4. To be purchased and inspected by weight, the number of pounds per gallon to be determined by the specific gravity of the oil at 60° F. multiplied by 8.33 lb., the weight of a gallon (231 cu. in.) of distilled water at the same temperature. 5. Flashing point must not be below 400° F. 6. Freedom from gumming. Using a half-pint brass oil cup maintained at about 140° F., practically equal quantities of oil must feed through the wick in equal intervals of time for three intervals of eight hours each; the wick to be of zephyr wool, four strands, doubled once. The oil in the cup will be brought to the original level at the beginning of each hour, and not less than 14 oz. avoirdupois must feed through the wick during the first period of eight hours. At the end of test the wick must be clean and the sides of oil cup bright and clean. 7. Cold Test. — The oil must flow at a temperature of 32° F. 8. Freedom Jrom Acid. — A small quantity when applied to a polished copper plate must not turn the surface of the metal green if allowed to stand exposed to the air for twenty-four hours. 9. Viscosity by Engler viscosimeter. At 90° F. must not be below 685; at 150° F. must not be below 155; at 225° F. must not be below 75, compared with distilled water (49) at 90° F. 10. Oil accepted and paid for under these specifications will be the subject of further tests in actual use on board ship, and any brand of oil failing to afford thoroughly efficient service in these conclusive tests shall not again be considered for naval use. UNITED STATES WAR DEPARTMENT SPECIFICATIONS No. I ENGINE OIL This oil is suitable for general lubrication of marine engines. All work except cylinder. LUBRICANTS 1 25 SPECIFICATIONS Must be a compounded oil composed of 25 per cent, pure acidless tallow oil and 75 per cent pure mineral oil. Must be free from acid, alkali, and suspended matter, and must satisfactorily pass the following tests: TESTS Specific Gravity. — Must not be below .920 nor above .923 at 60° F. Flash. — Must not flash below 415° F. Fire. — Must not burn below 470° F. Viscosity. — Must not be below 8. 11 at 50° C. (Engler.) Cold Test. — Must flow at a temperature of 35° F. Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Oil must not froth nor bump in flash cup when heated. Saponification — When oil is treated with alcoholic caustic potash must show presence of 25 per cent, tallow oil. No. 2 CYLINDER OIL (dARK) This oil is suitable where sieam pressures are high and lubricating conditions severe. SPECIFICATIONS Must be a compounded oil composed of 5 per cent, pure acidless tallow oil and 95 per cent, pure mineral oil. Must be free from acid, alkali, tarry or suspended matter, and must satisfactorily pass the following tests. TESTS Specific Gravity. — Must not be less than .900 nor more than .905 at 60° F Flash. — Must not flash below 540° F. Fire. — Must not burn below 600° F. Viscosity. — Must not be'below 3.83 at 100° C. (Engler.) Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Tarry and Suspended Matter. — Put 5 c.c. oil in 100 c.c. stoppered measuring cylinder and 95 c.c. of benzine. Shake well and allow to stand at least 10 minutes, and must be no precipitation. Volatility. — Heat small quantity of oil on watch glass for two hours at 400° F. There must not be a loss of more than 5 per cent, by weight. Saponification. — When oil is treated with alcoholic caustic potash must show presence of 5 per cent, tallow oil. 126 BEARINGS AND THEIR LUBRICATION No. 3 CYLINDER OIL (FILTERED) This oil is suitable where a light colored valve oil is desired. SPECIFICATIONS Must be a compounded oil composed of 15 per cent, pure acidless tallow oil, and 85 per cent, pure filtered mineral oil. Must be free from acid, alkali tarry or suspended matter, and must satisfactorily pass the following tests: TESTS Specific Gravity. — Must not be less than .891 nor more than .895 at 60° F. Flash. — Must not flash below 470° F. Fire. — Must not burn below 540° F. Viscosity. — Must not be less than 2.30 at 100° C. (Engler.) Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Tarry and Suspended Matter. — Put 5 c.c. oil in 100 c.c. stoppered measuring cylinder and 95 c.c. of benzine. Shake well and allow to stand at least 10 minutes. There must be no precipitation. Volatility. — Heat small quantity of oil on watch glass for 2 hours at 400° F. There must not be a loss of more than 5 per cent, by weight. Saponification. — When oil is treated with alcoholic caustic potash, must show a presence of 15 per cent, tallow oil. No. 4 GAS ENGINE OIL This oil is suitable for lubrication of gas engines that are water cooled, also for bearings. SPECIFICATIONS Must be a pure filtered mineral oil. Must be free from acid, alkali, suspended matter, and satisfactorily pass the following tests. TESTS Specific Gravity. — Must not be less than .896 nor more than .899 at 60° F. Flash. — Must not flash below 400° F. Fire. — Must not burn below 460° F. Viscosity. — Must not be less than 3.09 at 50° C. (Engler.) Cold Test. — Must flow at a temperature of 29° F. Acid. — Must not give an acid reaction on polished copper in 24 hours. LUBRICANTS 1 27 Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — Must be unaffected by an alcoholic solution of caustic potash. No. 5 GAS ENGINE OIL This oil is suitable for lubrication of gas engine cylinders that are water cooled, also bearings, where an oil of lighter body than No. 4 is desired. SPECIFICATIONS Must be a pure filtered mineral oil. Must be free from acid, alkali, and suspended matter, and must satisfactorily pass the following tests. TESTS Specific Gravity. — Must not be less than .874 nor more than .877 at 60° F. Flash.— Must not flash below 385° F. Fire. — Must not burn below 430° F. Viscosity. — Must not be less than 4.61 at 50° C. (Engler.) Cold Test. — Oil must flow at temperature of ^t,° F. Acid. — Must not give a reaction of acid on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — Must be unaffected by an alcoholic solution of caustic potash . No. 6 ENGINE OIL • , This oil is suitable for lubrication of high speed engines, dynamos, and high speed work generally. SPECIFICATIONS Must be a pure filtered mineral oil. Must be free from acid, alkali, and suspended matter, and pass satisfactorily the following tests. TESTS Specific Gravity. — Must not be less than .874 nor more than .878 at 60° F. Flash.— Musi not flash below 375° F. Fire. — Must not burn below 420° F. Viscosity. — Must not be less than 2.40 at 50° C. (Engler). Cold Test. — Must flow at a temperature of 14° F. 128 BEARINGS AND THEIR LUBRICATION Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — Must be unaffected by an alcoholic solution of caustic potash. No. 7 ENGINE OIL This oil is suitable for very troublesome bearings. SPECIFICATIONS Must be a pure filtered mineral oil. Must be free from acid, alkali, and suspended matter, and pass satisfactorily the following tests. TESTS Specific Gravity. — Must not be less than .877 nor more than .880 at 60° F. Flash. — Must not flash below 340° F. Fire. — Must not burn below 390° F. Viscosity. — Must not be less than 4.50 at 50° C. (Engler). Cold Test. — Must flow at a temperature of 32° F. Acid. — Must not give an acid reaction on poHshed copper in 24 hours. Alkali. — ^Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — Must be unaffected by an alcoholic solution of caustic potash. No. 8 ENGINE OIL This oil is suitable for engine bearings and difficult work where pressure is not extraordinarily high. SPECIFICATIONS Must be a pure filtered mineral oil. Must be free from acid, alkali, sus- pended matter, and pass satisfactorily the following tests. TESTS Specific Gravity. — Must not be less than .875 nor more than .879 at 60° F. Flash.— M.\isi not flash below 380° F. Fire. — Must not burn below 420° F. Viscosity. — Must not be less than 2.98 at 50° C. (Engler). LUBRICANTS 1 29 Cold Test. — Must flow at a temperature of 22° F. Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — Must be unaffected by an alcoholic solution of caustic potash. No. 9 MACHINE OIL This oil is suitable for shafting and ordinary lubricating duties on light running machinery. SPECIFICATIONS Must be a pure filtered mineral oil. Must be free from acid, alkali, and suspended matter, and pass satisfactorily the following tests. TESTS Specific Gravity. — Must not be less than .868 nor more than .872 at 60° F. Flash. — Must not flash below 370° F. Fire. — Must not burn below 420° F. Viscosity. — Must not be less than 2.38 at 50° G. (Engler). Cold Test. — Must flow at a temperature of 14° F. Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — Must be unaffected by an alcoholic solution of caustic potash. No. 10 MACHINE OIL This oil is suitable for ordinary machinery where heavy pressure and slow speed call for an oil of heavier body thau No. 9. SPECIFICATIONS Must be a pure filtered mineral oil. Must be free from acid, alkali, and suspended matter, and pass satisfactorily the following tests. TESTS Specific Gravity. — Must not be less than .890 nor more than .893 at 60° F. Flash. — Must not flash below 390° F. 9 I30 BEARINGS AND THEIR LUBRICATION Fire. — Must not burn below 440° F. Viscosity. — Must not be less than 4.29 at 50° C. (Engler). Cold Test. — Must flow at a temperature of 32° F. Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — Must be unafi"ected by an alcoholic solution of caustic potash. No. II ENGINE OIL This oil is suitable for use in cylinders of kerosene engines. SPECIFICATIONS Must be a compounded oil composed of 20 per cent, pure acidless tallow oil, and 80 per cent, pure filtered mineral oil. Must be free from acid, alkali, and suspended matter, and pass satisfactorily the following tests. TESTS Specific Gravity. — Must not be less than .884 nor more than .888 at 60° F. Flash. — Must not flash below 395° F. Fire. — Must not burn below 430° F. Viscosity.^yL\xsi not be less than 4.06 at 50° C. (Engler). Cold Test. — Must flow at a temperature of 30° F. Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — When oil is treated with alcoholic caustic potash must show presence of 20 per cent, tallow oil. No. 12 CYLINDER OIL This oil is suitable for use in ammonia cylinders of ice and refrigerating machinery. SPECIFICATIONS Must be a pure filtered mineral oil. Must be free from acid, alkali, and suspended matter, and satisfactorily pass the following tests. LUBRICANTS 131 TESTS Specific Gravity. — Must not be less than .870 nor more than .875 at 60° F. Flash. — Must not flash below 370° F. i^fre.— Must not burn below 420° F. Viscosity. — Must not be less than 2.38 at 50° C. (Engler). Cold Test. — Must flow at temperature of 10° F. Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali. — ^Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Saponification. — Must be unaffected by an alcoholic solution of caustic potash. No. 13 CYLINDER OIL This oil is suitable for use in cylinders of Westinghouse Engines. SPECIFICATIONS Must be a pure mineral oil, and pass satisfactorily the following tests. TESTS Specific Gravity. — Must not be less than .892 nor more than .896 at 60° F. Flash. — Must not flash below 450° F. Fire. — ^-Must not burn below 540° F. Viscosity. — Must not be less than 3. 11 at 100° C. (Engler). Acid. — Must not give an acid reaction on polished copper in 24 hours. Alkali — ^Ash must not show an alkaline reaction. Water. — Must not froth nor bump when heated in flash cup. Sponification. — Must be unaffected by an alcoholic sloution of caustic potash. Tarry and Suspended Matter. — ^Put 5 c.c. oil in 100 c.c. stoppered measuring cylinder and 95 c.c. of benzine. Shake well and allow to stand at least 10 minutes; must give only slight precipitation. Volatility. — Heat small quantity on watch glass for two hours at 400° F. Must not lose more than 5 per cent, by weight. SECTION VII Design of Sliding Surfaces and Special Bearings We are indebted to Professor Sweet for a most valuable discussion of the design of flat sliding surfaces in a little book Things That Are Usually Wrong. The following is adapted from that source. The underlying principal is to make the wearing surfaces of equal length. With this proportion experience has shown that even under varying loads and speeds the wear is very small and practically uniform over the entire extent of the contact surfaces. The reason is stated thus: ''Things that do not tend to wear out of true do not wear much." Professor Sweet has used this principle in the cross-head shoes and guides of steam engines, for the wearing surfaces between column and knee and hori- zontal slide and knee of a milling machine, for the ram and guides of a punch press and for the cross slide of a lathe. In each case the design was successful. FIG. 27. SHORT CROSS-HEAD GUIDE WITH RATCHET SURFACES. Fig. 27 shows a cross head and guide with equal length bearing surfaces made by reducing the length of the guide through cutting a ratchet surface at each end. This surface holds the oil pushed along by the end of the head and acts as a reservoir from which lubricant is taken on the return stroke. This construction is said to prevent the splashing and throwing of oil to a much greater extent than a full length guide. In some classes of machinery, particularly machine tools where the sliding members are short stroked much of the time, excessive local wear may take place either on the sliding member or on the guides. The shaper is an illus- tration. Here the contact surfaces of ram and guide are sure to wear most at the front end. As a partial remedy, Professor Sweet suggests to cut away the wearing surfaces at the rear increasing the proportion of gaps to lands as the ' 132 DESIGN OF FLAT SLIDING SURFACES 133 rear end is approached. Fig. 28 shows how this is done and gives a graphical method of determining the widths of gaps and lands. The shaded, portion represents the parts of the surface that have been cut away and the white por- tion those remaining to take the wear. The method of laying out is to draw a diagonal AB across the surface to be relieved ; then lay ofif the line BC at a convenient angle with the length of the guide, say 45 degrees. From C measure CD equal to the width of the guide C D O "^ FIG. 28. METHOD OF RELIEVING FLAT WEARING SURFACES. and draw DE parallel to BC and intersecting AB Sit F. The vertical distance from F to the upper edge of the slide is the width to be relieved in the next horizontal section of the guide having a length equal to the guide width. This process of laying out is repeated to the end, the widths of the gaps increasing. In a similar manner the surface of the sliding member should be relieved but sloped in the opposite direction. If the sliding member is very short in comparison to the length of the guide no relief will be necessary. With regard to the ways of lathes, Professor Sweet says: ''There has always been diverse practice and, from time to time, much discussion about the guiding inir r ^ s FIG. 29. A DESIGN FOR LATHE WAYS. of the slide-rests of lathes. The V has the merit of remaining free from lost motion however much worn, but nothing is more ridiculous than two F's, for the one at the back does no good and costs money. The common flat way is bad because the guiding surfaces are too far apart. The plan adopted by John Lang & Sons, Fig. 29, is much better, the guiding, though by flat sur- faces, is at the front where it ought to be, and when all the metal and work is concentrated on the one guide it can be twice as long and four times as efficient." LUBRICATION OF SLIDING SURFACES The lubrication of sliding surfaces is difficult owing to the wiping action of I he moving member. It is common to oil groove the surfaces and feed oil by 134 BEARINGS AND THEIR LUBRICATION some of the ordinary methods. Unless oil is introduced under pressure the grooves should run across the face at right angles to the direction of motion and should be stopped at some distance from the edges. If thus made they will aid in forming and maintaining a lubricating film. If forced lubrication is used the grooves can be cut along the diagonals of the surface and the lubricant introduced through a hole at their intersection. All of the edges of the oil grooves and the edges of both slides and guides should be chamfered or rounded over, to do away with scraping action. Planer ways are commonly oiled by a small roller dipping in an oil pocket, the number used depending upon the length of the bed. Professor Sweet says that it would be a great improvement 'Ho use deep pockets and wheels 6 inches or a foot in diameter, and then to pipe the oil from the end pockets to the bottom of the roller pockets and thus make the oiling continuous." INFLUENCE OF REDUCING THE ARC OF CONTACT The arc of contact between journal and brass in railroad car axle bearings is made considerably less than i8o degrees. This, of course, reduces the pro- jected area and is sometimes referred to as a peculiarity in design. John Goodman, Proc. Inst. C. E., Vol. LXXXIX, gives the results of experiments and an accompanying empirical mathematical investigation which shows that the force of friction for a given set of conditions can be reduced by decreasing the arc of journal contact within certain limits. The experiments were made with an alloy bearing composed of i part tin and 7 parts copper; it was 4 in. long, 2 in. in diameter, while its width or the length of the chord of contact was varied from 2 in. to 1/2 in. The journal with which it ran was made of manganese steel. Several forms of lubrication were tried. The results show that the oil film is probably at its best for an arc of contact subtended by a center angle of from 80 to no degrees. Table 37 gives the mean results of these experiments. The empirical formula which agrees very closely with the experimental results is: Log. R = log. i?i +0.00671 Co in which R = the frictional resistance to be determined when the width of the chord is equal to C. i?i = the frictional resistance found by experiment, when the width of the chord is C^. r = the frictional resistance when the width of the chord = i (or o.oi of the diameter). This is termed the base of the curve. Co=C-C,. REDUCING ARC OF CONTACT 13$ To show the use of this equation, two examples are given as follows: Given a shaft 5 in. in diameter fitted with a semi-circular brass, the total frictional resistance is found by experiment to be 8 lb., it is required to find how much the friction will be reduced by cutting down the chord of the arc of contact to 3 in., the total load is assumed to be unaltered. Here R^=S lb., log. 8=0.90309. C =;^ in., =60 units. 0^=5 in., = 100 units. Cq = 2 in., = 40 units. Then log. R =0.90309—0.00671X40: = 0.63469 i? =4.31 lb. Thus, by reducing the width of the brass from 5 in. to 3 in., the frictional resistance has been reduced from 8 lb. to 4.31 lb., or 46 per cent. Given an axle of 5 in. in diameter carrying a load of 7,500 lb. on a semi-circular bearing 10 in. long; the frictional resistance is 50 lb. Assuming that the intensity of the load must not exceed 300 lb. per square inch of projected area, how much can the frictional resistance be reduced by shortening the arc of contact? The equation for the shortest permissible chord is: Log. R minimum = log. R^ +0.00671 (x—C^) in which the notation is the same as given above, except that x is the length of the shortest per- missible chord. 7.500 Then x = — = 2.5 m. = 50 units 300X 10 and log. R minimum =log. 50+0.00671 (50— 100) = 1.3635 R minimum =23.1 lb. Thus the frictional resistance on this brass may be reduced from 50 to 23.1 lb. without exceeding the safe intensity of load. 136 BEARINGS AND THEIR LUBRICATION Length of Chord of Contact Total 2 in. 1.75 in. 1.5 in. i.o in. 0.5 inch load lb. Angle Subtended 180 degrees 122 degrees 97 degrees 60 degrees 29 degrees P f fP f fP f fP f fP f fP SO 100 ISO 200 250 300 350 400 4SO Soo SSO 0.0441 0.0288 0.0192 O.OIS9 0.0128 0.0103 0.0085 0.0071 0.0062 0.0058 0.0051 2. 20 2.88 2.88 3.18 3.20 3 -09 2.97 2.84 2.79 2.90 2.85 0.0458 0.0238 0.0160 0.0121 0.0097 . 009 1 0. 0070 0.0062 0.0055 0.0049 0.0045 2.29 2.38 2.40 2.42 2.42 2.43 2. 45 2.48 2.47 2.4s 2.47 0.0418 0.0209 0.0140 0.0106 0.0085 0.0071 0.0061 0.0054 0.0049 0.0044 . 0040 2 .09 2 . 09 2. 10 2. 12 2. 12 2.13 2.13 2.16 2. 20 2. 20 2.20 0.0210 0.0105 0.0070 0.0055 . 0044 0.0037 0.0033 0.0029 0.0028 0.0029 0.0027 1.05 1.05 1.05 1 . 10 1 . 10 I . II I-I5 1. 16 1.26 1.45 1.48 0.0126 0.0084 . 0056 0. 0042 0.0034 0.0028 0.0024 0.0021 0.0019 0.0017 0.0015 0.63 0.84 0.84 0.84 0.84 0.84 0.84 0.84 0.84 0.84 0.84 Mean 2.89 2.42 2.14 1. 18 Conditions: Bath lubricator at 104° F.. 233 r. p. m. Table 37.— Mean Results of Experiments with Differing Arcs of Contact. KNIFE-EDGE BEARINGS Knife-edge bearings are used in weighing machines, testing machines and scales. J. W. Bramwell, Engineering News, Vol. LV, page 653, discusses their design. From his article the following is condensed. Allowable loads on knife edges vary with the manner in which the pivots or knife edges are held in the levers and the pivot supports of scales, and secured to the bases of weighing machines. Very little care is exercised in most heavy weighing scalesas to the proportion of length of the pivot or the method employed in securing it to its lever. Fre- quently slender pivots are overhung from the side of the lever, and because of the deflection, even under a moderate load, the extension of the pivot beyond the solid support is practically worthless. The case of the pivot is much worse than the shaft because the local distortion is greater; hence the pivot would touch its support only over a fraction of its length. In testing machinery greater care is generally given to these details, although serious faults are still prevalent with some machines considered as standard. The testing machine is nothing but a weighing scale having a straining mechanism, which pulls upon a specimen; the resistance of this specimen is weighed precisely as upon a platform scale. KNIFE EDGE BEARINGS 137 A defective but cheap practice in fitting pivots to levers of testing machines, particularly those designed to carry a heavy load, is to cast the pivots in the body of the lever using square tool-steel pivots for cores. These pivots have a taper which allows them to be driven out when the casting is cool. In spite of the care used in placing these cores in the mold their correct position will be changed after the casting is poured. To bring them into proper alinement again, considerable skill and care is necessary. The greater fault of such practice, however, is in the shape of the hole in the casting. The hole being square and its diagonal necessarily vertical, a very bad breaking point is made by the sharp corner. The tendency to break is also increased by the chilled iron immediately in contact with the pivot. The initial strain due to driving the taper pivot in still further reduces the strength of the lever at this point. The hole is frequently enlarged by the wedging action of the pivot under load, and on account of the irregular surface at the upper sider of the pivot it will assume another position and change what should be a fixed ratio — the lever lengths. Allowable loads are variable because of these irregularities. Loads of 10,000 lb. per inch of length on knife edges are permissible, but the pivot must be flat at its upper portion, normal to the load and supported, its whole length with a minimum deflection of parts to ensure reasonable accuracy. The quality of steel used in both pivots and seats has an important bearing upon the aUowable load. In all cases it is essential that a high-grade uniform tool steel be used, having a carbon content of o. 90 to i per cent. Such steel will take a very high temper and yet have sufficient ductility to resist sharp blows without crumbling. The temper of the seats should be drawn to a very light straw color; that of the pivots should be slightly darker. An angle of 90 degrees for the knife edge has given good results for heavy loads. For ordinary weighing machinery and most testing machinery 5,000 lb. per inch of length will be found an ample allowance. In a recently built testing machine of 800,000 lb. capacity the two large levers had a load of 10,000 pounds per inch of length. For greater loads the sharp edge is rubbed with an oilstone so that a smoothness is just visible. A pronounced radius of knife edge will decrease the sensibility of the apparatus. The seat, if an angular one, is shaped with a small radius at the intersection of the sides of the angle. SPINDLE BEARINGS FOR TRAVERSE SPINDLE GRINDERS The bearings for the spindles of traverse spindle grinders, essentially watch factory tools, are an exception in that under proper conditions of fitting they are run without lubrication. Commercially they are usually run with a very 138 BEARINGS AND THEIR LUBRICATION little watch oil, benzine or kerosene. They show a practical application of the phenomenon of air lubrication/ In diameter, they are about 1/2 in. and run at speeds from 12,000 to 13,000 r. p. m. Both bearings and mating spindles are of hardened steel most care- fully lapped as to straightness and fit. W. H. Sawtell, in the American Machinist for August 5, 1909, page 248, writes of the conditions under which these bearings run: ''The traverse spindle when properly fitted will run without a particle of oil and it is common practice to use it in this way. However, it must be kept clean. Use clean tissue paper and alcohol to clean the spindle and bushings, being careful to keep oily fingers from, coming in contact with them. If the spindle shows a tendency to cloud up or stick, rub it with a little wad of paper, keeping the spindle revolving and traversing to reach as much of the surface as possible. ''Spindles which will run absolutely dry are rare, almost as rare as the men who understand how to make and use them. Most spindles made commer- cially are fitted a little free and require the least bit of lubricating. The best method is to wad up tightly a piece of clean waste about the size of a walnut and put on it a couple of drops of watch oil. Apply this to the revolving spindle at the same time traversing back and forth several times. If the spindle appears to gum or cloud, keep at it, and after a few cleanings and oilings it will settle down to business and give no further trouble. "The fit of these spindles in their bearings is very close, more like the fit of a plug gage and its mating ring than of sliding fits as we ordinarily know them. Under the best of conditions they will run with air lubrication and after the belt is thrown off will frequently run on for two or three minutes before stopping. At the same time when standing still it is difficult to slide the spindle through its bearings so close is the contact." ^For a discussion of air lubrication see paper of Professor Kingsburry, "Experiments with an Air-lubricated Journal," published in the Journal of the American Society of Naval Engineers, Vol. IX, No. 2, 1897. SECTION VIII Three Important Bearing Inventions Before taking up typical designs and construction for journal bearings, three important inventions must be presented. These three more than all the others have influenced bearing development not only in this country but also abroad. The reason for presenting them is all the stronger because in regard to at least two of them, there is a general misapprehension as to the real inventor. Taken in chronological order, they are the process of babbitting, the ball and socket seat, and ring oiling. The fame of Isaac Babbitt is generally supposed to rest upon his discovery of the white metal alloy called to-day babbitt metal. As a matter of fact his invention of the process of ''babbitting" has had a much wider influence on the construction of machinery bearings. While there are numberless lining REPRODUCTION OF DRAWING OF BALL-AND-SOCKET BEARING SEAT ACCOMPANYING BANCROFT OR SELLERS PATENT. metals, many of which are far different from his original formula, the majority of machinery bearings made to-day consist of a soft metal lining anchored by recesses into a shell. In a little booklet issued by the inventor in 1848 he says: "Isaac Babbitt would inform the public that his patent does not consist in the use of soft metal, simply, but in the mode of Us application and confinement in boxes prepared for the purpose." Babbitt's patent was No. 1252, issued July 13, 1839. I^ was apparently very profitable. The Baltimore & Susquehanna Railroad Company paid him $1,050 for the right to use it on their machinery, including locomotives and cars. 139 I40 BEARINGS AND THEIR LUBRICATION On September 20, 1842, the secretary of the United States Navy, acting under authority of Congress, paid him $20,000 for a license to use the patent in the construction of bearings for the machinery of war vessels. The second of these great inventions is the so-called Sellers' bearing, in- vented by Edward Bancroft. Fig. 30 is reproduced from the patent. The fol- lowing is taken from a letter written by Coleman Sellers, Jr. : "The Bancroft patent for this type of hanger was originally granted October 9, 184Q. After the death of Mr. Bancroft, it v/as surrendered and re-issued May 12, 1857, the re-issue number is 463. "Mr. Bancroft was in partnership with the late William Sellers under the firm name of Bancroft & Sellers. They were manufacturers of shafting and mill gearing, and also makers of lathes, planers and other machine tools. This patent covers the idea of forming a portion of each box of spherical form and supporting them in guiding sockets so 'as to allow the boxes thus held to play within the sockets in the manner of a universal joint substantially as described.' "Mr. Bancroft took out another patent May 22, 1849, '^^ which the box was supported on trunion screws in a vibrating yoke." The third of there inventions is that of ring oiling. The inventor is John E. Sweet, and the device was never patented, although it is universally used to-day in the bearings of high-grade machinery, particularly electric motors and generators. Professor Sweet writes as follows in regard to its early development: "The oil ring was first shown and described in Engineering (London) for January, 1868, page 44. The description is this: 'A self-lubricating journal box is shown in which it is pro- posed to return the oil from the drip cup to the shaft by means of two loose rings.' "About all that I can say is — that when I returned from England, I went into the machine shop at the Navy Yard and saw on the floor a half dozen or more hangers made from the drawing in Engineering, and I learned that they were to be sent to some other Navy Yard, possibly it was Beaufort. "When I reached home I found my brother was putting up a lineshaft and also using the same thing. When I built the first Straight Line engine and all afterward, it was one of the features, but so far as I know, no one else ever used it until we exhibited an engine at the Mechanics' Fair in New York, when Mr. Edison got on to it and began using them in his dynamos; after that it began to be used by others in various places. "Had I taken out a patent at first it would have expired before anyone else adopted it, and like many another patent it would not have been worth the paper it was written upon." It is not uncommon to find bearings using all three of these inventions. SECTION IX Typical Designs and Constructions This section presents, by line drawings and a few halftone engravings, a number of typical bearing designs, constructions and lubricating systems that are representative of American practice. In the space available, it is obviously impossible to present more than a very few such illustrations. ELECTRIC MOTOR AND GENERATOR BEARINGS Figs. 31 to 37, inclusive, show several types of electrical generator and motor bearings. Figs. 31, 32, and 33 are from the General Electric Company. The first shows a large spherical-seated, babbitt lined, oil-ring lubricated bear- .-->, Momiual Dimensions A B c D E F // J K L A^ P R S 7 21 H 22 U^ i'/i I VA 12^4 \Q% •lii 5% 6H 2}i 8 24 H 25 16>i e^4 Wa i% 14 12 >, 3 (iH 1H 2% 27 H 28 18}^ 7H i% SVa 15^ 14 M 7 8% 3 FIG. 31. DESIGN OF 3:1 SPLIT SPHERICAL SEAT BEARINGS. 142 BEARINGS AND THEIR LUBRICATION ing, for nominal diameters of from 7 to 9 in. This type is also used for much smaller bearings. Fig. 32 shows a pedestal box, while Fig. ;^t, shows a form of 3 to I bronze sleeve bearing for small journal diameters. Figs. 34 to 37, inclusive, are drawn from the practice of the Westinghouse Electric and Manufacturing Company, and need no detailed description. A size of Bearing 22H 27 H 2m 3>J^. 4»/l6 4% 29% 5% H 2h M N Dowel Pins Foot Bolts m-5 FIG. 32. DESIGN FOR 3:1 SPHERICAL SEAT PEDESTAL BEARINGS. feature worth considering in Fig. 37 is the cover of the bearing housing permit- ting inspection of the oil ring and the condition of the lubrication. This cover is under spring control, is locked by a winged nut and provided with packing to make a dust-proof joint. The oil reservoirs of all these bearings have drainage plugs. HEAVY GRINDER SPINDLE BEARINGS Fig. 38 shows a typical construction for heavy grinding spindle bearings from the Norton Grinding Company. The wipers are of felt. In more recent TYPICAL DESIGNS AND CONSTRUCTIONS 143 constructions wood has been substituted, as it is believed that fibers of the felt tend to be picked up and carried into the bearing. FLOODED LUBRICATION IN MACHINE TOOL PRACTICE The recent designs of vertical turret lathes built by the Bullard Machine Tool Company have a system of "continuous flow" or flooded lubrication. In the base of the machine is an oil reservoir to which all of the oil drains after passing through the various bearings and gear boxes. t) r »i'jf<-Uril Nominal Dia. of Bearing 6 Dimensions Dimensious 1 of Set Screwl A B c D E F G H J K L M N P R S IH IH A SH 6 % 2H % IH ''/n y* 'Ke IK, H H J«-ll H % 2 A a 6?4 % 3 %: \% 'h2 'A, '■v.. i\ Vfi Ji ^/^-ll % % 2}i A m -i^i % m %; \^-i 'h, K« 'A 1% %. >i 5i-ll H % 2ii A l\i 8^i % 3h % ' V4 ^'/32 M 1 2h, ^16 ^a 94-10 % « 2h A Wi 9 % 3% % 2 •Vsi ^ 1M« 2Ji ^6 '•Jf6 9i-10 % H 3 B 9 10 ^ 4V4 % m H 'Vn % \H 2^-lc ?ra 7i ?i-io % H ■m B iQ% U^ H 4H ^I'a VA H ''M % Wi 2% ^6 % ?4-10 % « 4 B 12 13 H 5'-^ H 2H I 'V^ 'ke m ^H y* 7^ ?i-io H « 4H B 13M UVz H 6^4 H 2% 1 '■y-n H V/i 3^ H 1 1-8 ''^i'« M FIG. 33. DESIGN FOR 3:1 BRONZE SLEEVE BEARINGS. Submerged therein and driven from the driving shaft is a geared pump which delivers the oil to a distributing reservoir located on the outside of the column at a height to give a slight head for the free flow of oil through pipes leading to the bearings and gear chambers. The excess oil pumped to the distributing reservoirs drains to the sump through an overflow. 144 BEARINGS AND THEIR LUBRICATION In the pipes leading out of the reservoir are sights so that the flow can be observed, and if a stoppage occurs the overflow from this point at once calls attention to it. FIG. 34. MOTOR RING OILING BEARING. FIG. 35. RING-OILING MOTOR BEARING SHOWING OIL RINGS AND GROOVES. FIG. 36. RING-OILING MOTOR BEARING SHOWING OIL RINGS AND GROOVES. The flow of oil is at the rate of 0.08 gal. per square inch of projected bearing area per minute. TYPICAL DESIGNS AND CONSTRUCTIONS 145 Fig. 39 shows the table spindle bearing and method of lubricating of one of these machines. MACHINE TOOL BEARINGS The illustrations from machine tool practice, Figs. 40 to 48, inclusive, are from the Pratt and Whitney Company and show bearings from a number of different machines. FIG. 37. RING-OILING MOTOR BEARING — RIGID SLEEVE TYPE, FIG. 38. BEARING CONSTRUCTION FOR HEAVY GRINDER SPINDLE. Fig 40 is a section of the front spindle bearing of a turn-table lathe. It is of bronze, tapered, split, with an adjusting nut on each end, an oil hole top and 10 146 BEARINGS AND THEIR LUBRICATION bottom with a longitudinal groove running through each. These holes connect with an annular chamber in the headstock casting. Practice limits the rubbing velocity for a bearing of this kind to 800 ft. per minute. Fig. 41 shows a spindle bearing of a 2 1/2X26 in. turret lathe. It is babbitted with a method of anchoring the babbitt which is used generally on FIG. 39. TABLE SPINDLE BEARING OF VERTICAL TURRET LATHE. Pratt and Whitney tools. This consists in turning a dovetailed groove at each end of the bearing housing and then drilling a number of holes starting in the surfaces where the bearing is split in such a manner that there holes break through into the lining space at the joining surface of the cap and bottom TYPICAL DESIGNS AND CONSTRUCTIONS 147 FIG. 40. FRONT SPINDLE BEARING OF TURN-TABLE LATHE. FIG. 41. FRONT SPINDLE BEARING OF TURRET LATHE. f^^^^^m^ FIG. 42. DETAIL OF ANCHORAGE FOR BABBITT LINING. FIG. 43. FRONT SPINDLE BEARING OF FIG. 44. FRONT SPINDLE BEARING OF 14-INCH TOOL-MAKER'S LATHE. HAND MILLER. 148 BEARINGS AND THEIR LUBRICATION half of the box. The babbitt enters these holes and is thereby held securely at the edges of the split, at the same time due to its contraction, it is drawn tightly into the dovetailed grooves at the ends of the box, see Fig. 42. The front spindle bearing of a 14 in. tool-maker's lathe is shown in section in Fig. 43. It is babbitted with the regular method of anchorage. The proportions are a little less than 1:2, the dimensions being 2 7/16X4 in. Babbitt Thrust Collar. Babbitt Lining:. Ball Thrust Bearing. FIG. 45. SPINDLE AND BEARING OF VERTICAL SURFACE GRINDER. Fig. 44 shows the construction of the front spindle bearing of a No. 10 hand miller. This is of bronze with the same device for taking up wear as in Fig. 40. The bearings are also made babbitted. The spindle and bearings of a vertical surface grinder are shown in Fig. 45. Here the lower bearing is babbitt lined with ball thrust collars, and the upper is also babbitt lined with a babbitt thrust collar. Instead of anchoring by TYPICAL DESIGNS AND CONSTRUCTIONS 149 drilling a number of small holes at the split surfaces, a number of l^rge holes are drilled through the cap and its mating half. These serve as anchorage with the dovetailed grooves at the ends. The front spindle bearing of a 5X48 in. cylindrical grinder is shown in V\"\\V\^V>J ' ^ FIG. 46. FRONT SPINDLE BEARING OF 5 X 4S-INCH FIG. 47. FRONT SPINDLE BEARING CYLINDERICAL GRINDER. ' OF SPLINE MILLER. section in Fig. 46. It is a bronze sleeve with oil grooves top and bottom. In the driving shaft bearings of this machine felt plugs are inserted in the bottom of the sleeve and the Osgood system of grooving is used. The front spindle bearing of a spline miller is shown in Fig. 47. It is of FIG. 48. SPINDLE BEARING OF CUTTING-OFF MACHINE. hardened and ground tool steel. The rear bearing is bronze. The spindle is likewise of hardened and ground tool steel with a clearance of 0.0005 in. on the diameter. The housing has a chamber for oil. A large cast-iron spindle bearing of a cutting-off machine is shown in I50 BEARINGS AND THEIR LUBRICATION Fig. 48, and is of interest as being an example of a cast-iron spindle running in a cast-iron box. The speed is slow. BEARINGS FOR STEAM TURBINES The step bearings of vertical Curtis turbines are the most prominent ex- ample of the use of a cast-iron step for heavy pressures and high speeds. Table 36 page 120 gives the range of pressures, speeds, and quantities of oil used. The bearing plate, or lower block, is of cast-iron rigidly held by the frame. See Fig. 49. The block is guided at the sides and^ carried on a large screw, passing through a steel nut and coming in contact with a steel block set in the FIG. 49. Oil or Water STEP BEARING FOR VERTICAL CURTIS TURBINE. bearing plate. It is essential that this plate should be rigidly held, that its upper face should be a true plane set at right angle to the shaft axis, and that the clearance should be so small that the relative alinement of blocks and shaft cannot vary appreciably. The step plate is likewise of cast-iron and keyed to the lower end of the shaft. Both plates are recessed so that the surfaces of contact are collars. Directly above the step plate is a cylindrical guide bearing. The contact faces of the blocks must be truly parallel, and the contact surfaces of the end of the TYPICAL DESIGNS AND CONSTRUCTIONS 151 screw beneath the lower block and its mate must be likewise true and free from convexity. Oil is introduced through the center of the screw, passes upward, enters the recess in the center of the plate, passes out between the contact surfaces, and ascends upward through the guide bearing, as indicated by the arrows in the illustration. These bearings can be run with either oil or water as the lubricant. In service the plates are actually separated by a lubricating film; some four or five times as much lubricant as is necessary is usually pumped as a safeguard. The greater the quantity the greater the separation between the plates and the Upper Half Section A-A iBearing Developed shovv- Oing Oil Grooves and Cooling Coils. FIG. 50. WATER-COOLED BEARING FOR HORIZONTAL TURBINE. less the danger of cutting out. The actual separation of the plate is, of course, only a few thousandths of an inch. From this fact it can be seen that this type of bearing calls for the best of workmanship. Fig. 50 shows a water-cooled bearing for Curtiss horizontal turbines including the arrangement of the cooling pipes and the oil grooving. The bearing is provided with a coil of thin copper tubing cast into the babbitt lining as close to the contact surface as possible, and having at the ends steel blocks securely brazed on. Water is led in through pipes passing through stuffing boxes in the bearing standard. In this way the heat of the bearing is taken up by the cooling water at the point where it is generated. There are no cooling coils in the oil tanks. The top half, or cap of the bearing, is cut away, except at the ends and center. This relieving is about 0.020 in., so that the oil fed to the bearing fills the space between the shaft and lining. The oil grooving in the lower half of the bearing is one straight groove into which oil is fed through a central hole as indicated in the cut, or, from several 152 BEARINGS AND THEIR LUBRICATION holes leading to a header hole, which in turn is fed from the pipe in the standard. This groove is well rounded over on the side toward which the shaft rotates; on the side toward the bottom of the bearing it is gradually tapered up to the bearing surface. This assists the action of the shaft in pulling oil into the region of greatest pressure. The ends of the main oil groove terminate in very small grooves which run through to the ends of the bearing. This allows for the washing out of small particles of grit and dirt, instead of letting them remain at the ends of the bear- ing and cut the shaft at that point. The oil is discharged from both ends around the entire circumference and particularly at the point where the ends and middle bands on the upper half of the bearing are cut away. FIG. 51. FLOATING SLEEVE TYPE OF HORIZONTAL TURBINE BEARING. At the ends of the bearing the casting is allowed to project slightly and a sheet brass plate is attached running down to a small clearance around the shaft. In this recess so formed a deflector or fan is provided, which throws off the oil from the ends of the bearings, preventing it from leaking out at the ends of the standard and from being slooped over the inside of the standard and fmally seeping out of the joints. Bearings of this type are designed to run with an oil pressure of from 15 to 25 per square inch. The water supply pipes are provided with stop cocks to regulate the amount delivered and the discharge is open. Water under low pressure as generally supplied from city mains is used. Most bearings of this kind are arranged so that the temperature of the oil can be taken and the amount of circulating water varied to meet the conditions of running. Fig. 51 shows a type of floating sleeve bearing used on smaller Westinghouse- Parsons turbines running at 3600 revolutions per minute. The ratio of diam- TYPICAL DESIGNS AND CONSTRUCTIONS 153 eter to length runs from 1:2 1/2 to 1:3. The bearing proper consists of a bronze sleeve provided with an oil groove at the. top and prevented from turn- ing by a tail projecting upward into a recess in the cap. Around this tube are three concentric sleeves, each diametral clearance being from 0.004 to 0.006 in. Each has four axial rows of holes to distribute oil and thus maintain a film between each pair of mating surfaces. This construction cushions the shaft and allows it to rotate around its gravity axis rather than its geometric axis if the two are not coincident. The sleeves are held in place by a nut which in turn is checked by a small set screw. The tube with its nest of sleeves is held in a cast-iron shell having m ^(F~~3)^ FIG. 52. BABBITT LINED HORIZONTAL TURBINE BEARING ARRANGED FOR FLOODED LUBRICATION. steel blocks resting in a cast-iron spherical seating ring in the pedestal. Under the blocks are thin liners used to aline the bearing. The adjustment is in increments of 0.005 ^^ch. Oil under a slight pump pressure, 2 to 5 lb., enters the bearing at the end, passes through it, and drains into a well in the pedestal. In some cases the bearing tube has been lined with babbitt. In the larger machines running at 1800 revolutions per minute or less a cast-iron shell, babbitt lined bearing is used without any sleeves. The ratio of diameter to length is from 1:2 to 1:2 1/2. See Fig. 52. 154 BEARINGS AND THEIR LUBRICATION It is split longitudinally in the usual manner and is provided with spherical seating blocks as described above. 3'i Stop Check Valve- m Stop Valve. 3H Stop Valve. Relief Dis- charge to Suction. Air Chamber Strainer' FIG 53. TYPE PLAN OF FORCED LUBRICATION FOR RECIPROCATING ENGINES OF UNITED STATES BATTLESHIPS AND ARMORED CRUISERS. At the bottom oil enters a copper tube that is set in a recess in the shell and held by the lining; it divides into two streams, passes around the bearing and TYPICAL DESIGNS AND CONSTRUCTIONS 155 into a longitudinal groove in the top of the cap cut to within about 3/8 in. of the ends. It escapes at the ends of the bearing and drains to a well in the pedestal. The babbitt is bored out so that the horizontal dimension of the bore is greater than the vertical and thus greater than the shaft diameter. The fit is over about two- thirds of the bottom half of the circumference. This FIG. 54. BALL-AND-SOCKET RING-OILING LINE SHAFT BEARING. FIG. 55. STEP AND GUIDE BEARINGS FOR VERTICAL SHAFT. prevents all side binding on the journal, does away with any edges to scrape off the oil and provides a small reservoir of oil within the bearing. Flooded lubrication is used at low pressures as in the other type. The oil is cooled by passing it through pipes submerged in water or through a tank containing cooling coils. At the open ends of the bearing housings are baffles that prevent leakage. 156 BEARINGS AND THEIR LUBRICATION The clearance between the shaft and the knife edges of these thin collars is from 0.020 to 0.030 in. Albert E. Guy gives both the European and standard practice in regard to the design of De Laval steam turbine bearings in Tables 38 and 39. This is of interset and value because of the extremely high speeds. The bearings are made of brass shells lined with babbitt. The pinion shaft bearings are lubricated by sight feed, or wick feed. The power shaft bearings are ring oiled with one ring in the middle of each bearing. H. r. R.P.M. d h h d Pi Ai I2 I2 P2 A2 V 30 20700 0.788" 4.56- 5.75 17.8 1.73 3 . 94" s. 13.36 1.3 ft. per sec. 71. 1 55 16600 I . 102" 5.9." 5.36 19.25 2. 10 5.67" 5. 15 13.36 1.45 79.7 7 5 16700 1.181" 8.27" 7. 17.64 2.05 7.29" 6.17 13.36 ..S6 85.9 110 13000 1.575" 9.05" S.75 .7.8 2.17 8. 11" 5.15 13.36 1.62 89.2 150 13000 1.575" 11.42" 7.25 19.2 ..34 10.24" 6.5 14.22 i.8s 89.2 225 1 1000 1.968" 12.21" 6.2 19.2 2.47 10.83" 5.5 14.22 1.85 ,4.S 300 ro6oo 2.16s" 14-57" 6.72 19.2 2.63 13- 11" 6.0s 14. 22 1.94 100.4 450 10600 2.16s 21.66 10. 19.3s 2.64 17.91" 8.27 15.65 2.13 100.4 d = diameter of journal in inches. 1 = length of journal in inches. pi and P2 = pressure per square inch of projection dl. A I and A2 = coefficients V= peripheral velocity of journal in feet per second. When V = qo ft. per second forced lubrication must be used— [oil brought to bearing under pressure]. P I and P2 = total pressure on bearing in lb. = p X d X 1. Load on middle bearing assumed to be i 1/2 times that on either of the others. PX R.P.M A. = -. A should ordinarily not exceed 2. 168,000X1 Table 38— European Bearing Practice with DeLaval Single Gear Turbine: One; Pinion Shaft, One Power Shaft. TYPICAL DESIGNS AND CONSTRUCTIONS 157 Pinion Shaft. Power Shaft. T . t2 ' H. P. d. ., I2 d R.P.M. Peripheral velocity ds I3 d3 R.P.M. Peripheral velocity 7 0.395" 31" ^!? s.s.-^ 7.2 3000c ft. per sec. SI. 7 § ll a i" 3tV 3.06 3000 ft. per sec 13. 1 10 0.395" 33" ^H" 8.55 7-2 24000 41.35 U' 3tV 3.06 2400 10.47s IS 0.532" 4|" sU" 7.75 6.8s 2400c 55,7 1.3" 3if 2.936 2400 13.62 20 0.708" 411" 4f" 6.62 6 2000c 61.8 1-31" 4r 2.435 20G0 11.43 30 ! 0.708" , si" 5" 7.77 7.06 2000c 61.8 1.46" StV 3.S55 2000 12.75 S"> 0.871" 6i" 6^ 7.89 7.89 20000 76 2" 6i" 4.2s iSoo 13.1 75 t" 5i" 4i" 5.5 4.875 16400 71.5 11 1 c 1 If" 6" 3.43 1500 11.45 1 10 1.25" 7r St" 6 4-3 13000 70.9 Nr 6" 2-4 1200 1 13. 1 150 1.25" j 7V sr 6 4-3 13000 70.9 2i" 6i" 2 6 1200 j 13. 1 225 1.6" 8" si" S 3.67s 1 1 06c 77.2 1 3 8" 2.667! 900 11.78 300 1.6" 9i" 7i" 6.17 4. S3 10500 73.3 3^ lO-i" 3 900 j 13.74 p = pressure per square inch on diametrical plane of bearing, p = 7 to 15 pounds per square inch up to 55 h. p. = 15 to 23 pounds per square inch from 55 to no, h. p. \ for pinion bearings. = 23 pounds per square inch from no to 300 h. p p = 1 5 to 40 pounds on power shaft beanngs. Spiral grooves in bearing to distribute the oil are 1/16 inch wide and 3/64 inch deep, with holes at the ends to watch the conditions of lubrication. Table 39. — Standard Bearing Practice of DeLaval Steam Turbines. FORCED LUBRICATION FOR HORIZONTAL ENGINES Fig. 53 is a type plan of forced lubrication for reciprocating engines of battleships and armored cruisers as worked out by the Bureau of Steam Engineering of the United States Navy. The location and arrangement of tanks, strainers, valves, and piping is shown. BALL-AND-SOCKET SHAFT BEARINGS Fig. 54 shows the ring-oiling line shafting box, and Fig. 55 a usual form of step bearing for vertical shafts as made by Wm. Sellers & Company, Inc. The horizontal box may be oiled by a syphon oil cup through the center or 158 BEARINGS AND THEIR LUBRICATION oil may be placed in the reservoir around the top ball and allowed to find its way through the small holes to the shaft. It is customary to fill the grease cups with tallow or other grease, which will melt if the box becomes at all warm. LOCOMOTIVE AND CAR BEARINGS The eleven illustrations, Figs. 56 to 66, inclusive, are reproduced from designs of the Baldwin Locomotive Works. Two are of locomotive driving rn T~l U'ln^^^rJ 7IG. 56. LOCO] ^ ^> ->l fn:^ % Brass Plugs . K -1 -^>i C^D LOCOMOTIVE DRIVING BOX. /^%- i /iM=n :-Hl»A pi ^ 1 r Hfl -f II -P -1-2— Journal- T1- n Wn\ Detail of Babbitt. FIG. 57. LOCOMOTIVE DRIVING BOX. boxes, two of tender boxes, one of an engine truck box, one of a crosshead, one of an eccentric strap and four of rod stubs. A few dimensions are included in each case to give an idea of proportions. TYPICAL DESIGNS AND CONSTRUCTIONS 159 As a matter of interest, Fig. 67 is reproduced. It is of a ring oiling car box invented and developed by W. O. Dunbar, and at one time experimented with on the Pennsylvania Railroad. For Journal BVz x 10 Inches. FIG. 58. LOCOMOTIVE TENDER BOX. Via S'teel Inserts FIG. 59. L0C0M0TI\rE TENDER BOX. KINGSBURY THRUST BEARING Fig. 68 shows in partial elevation and section, a thrust bearing developed and patented by Professor Kingsbury. i6o BEARINGS AND THEIR LUBRICATION The design differs from the ordinary collar bearing in that one of the pair of collars is divided into several sectors or '"shoes" which are individually supported on spherical seats whereby each shoe is enabled to adjust itself to ( y i I XT -J II Ij Ll 1 1-1 1 = mi ST I uu. -M- Vs Oil Holes 1 2a T-i^' -Oil-Groov iZTTTT , i-M-X—pellaT fj FIG. 6o. LOCOMOTIVE TRUCK BOX. FIG. 6l. LOCOMOTIVE CROSSHEAD. the best alinement and pressure distribution at the bearing face. The concave seats for the shoes are formed in a ring which has a spherical seat in the housing, for maintaining group alinement of the shoes. Lubrication is effected by a TYPICAL DESIGNS AND CONSTRUCTIONS l6l bath of oil in the case of vertical shafts, and by flood lubrication with horizontal shafts. The considerations leading to the adoption of the construction are as follows: Both theory and experiment have shown that in shaft journals in which perfect lubrication is maintained, the film of oil is invariably wedge-shaped, being '1'2 Iron Pipe — i. I, XlUll J. li^c I 1^ 111 Tin M K^M m- FIG. 62. LOCOMOTIVE ECCENTRIC STRAP. w ndleU Spi Style No. 1 Spindle Style No. 2 Oil Cup S tyle No. 2 Oil Cup Style No. 3 erg t^Jt^j^i FIG. 63. LOCOMOTIVE ROD STUB. thicker at the edge where the oil enters and thinner toward the leaving edge. In the ordinary collar bearing, the bearing surfaces remain strictly parallel, hence no wedge-shaped film of oil can form, and even with bath lubrication the action of the lubricant is imperfect, thus the specific loads are necessarily relatively low and the coefficients of friction high. The shoe construction on l62 BEARINGS AND THEIR LUBRICATION the other hand enables the wedge-shaped films to form and thus perfect lubri- cation is maintained. The leading edge of each shoe is depressed by the enter- ing oil, giving automatically the proper slight inclination between the bearing L'-j-Pj ii FIG. 64. LOCOMOTIVE ROD STUB. /^/32 Bushing Bronze ■I A_S ^^n nz J) FIG. 65. LOCOMOTIVE ROD STUB. surfaces. That this actually takes place is shown by the fact that such wear as occurs on the shoes is first indicated at the leaving edge. The shoe type bearings have shown good results under tests and in service. TYPICAL DESIGNS AND CONSTRUCTIONS 163 Pressures from 250 to 1000 lb. per square inch are borne under ordinary con- ditions with coefficients of friction as low as 0.0022. Test runs have been made FIG. 66. LOCOMOTIVE ROD STUB. FIG. 67. RING-OILING CAR BOX. at a circumferential speed of 4500 ft. per minute with pressures up to 2600 lb. per square inch, without causing failure of the oil film; and at lower speeds a pressure of 5000 lb. has been reached experimentally. 164 BEARINGS AND THEIR LUBRICATION TEXTILE MACHINERY BEARINGS W. S. South wick sums up American practice in the design and application of the bearings of textile machinery, thus: ■ "Among the kinds of bearings generally used on textile machinery, there are two or three of interest. The bearing most generally used is a simple cast-iron box mating a steel journal. Many of the caps for larger sized bearings have a recess cast in their upper side to form an oil ^Pr3 Drain \^^_,^^^<^m FIG. 68. KINGSBURY THRUST BEARING. well; the smaller bearings simply have a drilled and countersunk oil hole. A good grade of machinery oil is used for lubrication supplied by a squirt can. "There are no commonly accepted factors for design. In most cases the design is governed largely by experience and varies with conditions. A device that has been developed to a very high state of perfection, from a mechanical standpoint, is the spindle. Over 500 patents have been issued on it. As now built, it runs at from 8000 to 12,000 revolutions per minute ver}- smoothly, with little wear, and requiring but a small amount of power. The blade which is made of hardened steel, accurately ground to a slight taper, runs in a cast-iron bolster partly covered with cotton packing to reduce vibration. In some constructions the bolster is adjust- able to compensate for wear. The cast-iron base in which the bolster is fitted has on one side a projection adapted to receive the oil and convey it to the revolving parts which are surrounded by oil. A light clean spindle oil is used, generally supplied by an oil can. TYPICAL DESIGNS AND CONSTRUCTIONS 165 "Large vertical twister rings are often equipped with a self-oiling device. This consists of a wick which lubricates the traveler at a point below its contact with the yarn. The bearings on the parts which operate the harnesses and shuttle-boxes of fancy worsted and woolen looms are malleable iron and steel case-hardened. These bearings are about 3/4 inch in diameter and 5/32 inch wide. They are lubricated with an ordinary oil cup. In a few cases, chilled cast-iron journals and cast-iron boxes are used. They are oiled in the same manner. The temple roll is an illustration of a slow running bearing where a brass stud forms the journal and a wooden bushing previously boiled in linseed oil the lining shell. This is inserted into the roll. Albany grease is used for lubricating, being placed in the bushing before assembling." SECTION X Hints on the Care of Bearings Never start a new engine or machine without examining the fit of the bear- ings and knowing it to be right. Never start up any machine without knowing that its bearings are properly- oiled. A good way to adjust new engine main bearings is to set the boxes up tight and slack back the adjusting screws one face of the head. Then start the engine, watch it carefully and tighten up the screws gradually until the pound is all taken out. Remember that the clearance in a cold bearing is greater than in the same bearing at running temperature. When warming up the journal grows larger faster than the bore of the bearing. Remember that after a machine has been idle for some time the oil is probably squeezed out from between the bearing surfaces. A reasonable slackness for engine bearings from 6 to 12 in. in diameter is about 1/200 in., for crank pin and cross-head pin bearings up to 4 in. in diameter about 1/250 in. These dimensions can easily be approximated by setting the bearing up tight, then slacking back a certain number of faces on the wedge bolt -head depending upon a number of threads and the taper of the wedges. A bearing set up too slack is nearly as bad as one too tight. Pounding will make a box run hot and is liable to peen the brass or babbitt and destroy the fit. Don't squirt oil into an oil hole and hope that it will reach the journal. See that the hole is open and not filled with dirt. Don't put the end of the spout of an oil can into an oil hole and pump the bottom unless there is oil in the can. Squirt oil into the oil holes and not onto the floor— in the latter place it may lubricate your shoes but it does not help the bearings. In cold weather a little heat will help the oil to run out of your oil can — the top of the steam-chest is a favorite place for this warming up. Don't neglect a bearing because it is under a bench, or in a dark corner, or in some out-of-the-way place. "Out of sight, out of mind" is poor policy to apply to bearings. Keep an eye on the oil tanks and do not let them get too low. Good oil can- not always be gotten in a hurry and some bearings may suffer, 166 HINTS ON THE CARE OF BEARINGS 167 In general, the best oil to use on bearings is the lightest oil with which they can be run safely. Kerosene is successfully used as a lubricant for high-speed, hardened steel spindles and bearings, were the pressures are very light. Don't depend entirely on the sight glass to tell you how much oil is in the well of an important bearing. When babbitting a box, a piece of paper wrapped around the shaft or mandrel will keep the hot metal from chilling and reduce the amount of subsequent scraping. If possible it is a good plan to heat a bearing shell before babbitting. Never pour babbitt into a shell that contains any water or is not perfectly dry. Many serious burns have come from a failure to observe this rule. If several small bearings of the same size are to be babbitted, it is better to use a babbitting arbor rather than a finished shaft. If the arbor is sprung a little from the heat, no harm is done. For severe service it is a good plan to peen the surface of a babbitted box before it is bored or reamed to size. Don't try to run babbitt in bearings subjected to heavy shock and pound. Don't forget to cut oil grooves when you rebabbitt a box. Don't forget to drill the oil hole when you rebabbitt a box. Don't cut oil grooves through the end of a box. The oil will run out. In general, oil grooves should be cut at riglit angles to the direction of motion of the journal. If oil will not follow the grooves of a tightly fitted bearing, try venting by a scratch with a file or scraper through to the end in order to let out the air. Be sure that both ends of oil tubes are securely fastened so that they cannot jar loose. Any self -oiling bearing with an oil well should be washed out at intervals. Be sure the well is refilled before the machine is started. Be sure that there are no points in oil tubes higher than the point where oil is introduced. Oil will not run up hill. When making new oiling rings be sure that they are smoothly finished and without any corners or edges to catch when in use and stop the rings from revolving. When putting in new chains for self-oiling bearings, see that they do not drag on the bottom of the oil well. If they do hit the bottom, a hole will be worn through in time, letting out the oil and probably ruining the bearing. Don't let induction motor bearings wear sufficiently to drop the rotor against the stator. Put in new bearings. Felt wipers should be taken out at intervals and soaked in kerosene to soften them and remove the glaze. 1 68 BEARINGS AND THEIR LUBRICATION Don't let a hot box begin to smoke before you attend to it. You should smell it long before you see it. The rapid heating of a bearing is a danger sign. If a bearing takes two or three hours to reach a temperature uncomfortably hot for the hand, it is prob- ably safe, but if that same temperature is reached in lo or 15 minutes there is liable to be trouble. If you have a hot box on a main engine bearing, first loosen it up a little all around and then give it a generous dose of cylinder oil through the feel holes, but don't try to feed it through the ordinary oil channels; they are too small. If this oil will stick to the shaft it is apt to cure the trouble. Another good remedy is a mixture of graphite and oil — one part of graphite to 10 of oil. But do not feed this until the graphite has been thoroughly wetted. Dry graphite will not help a hot box. If there is no graphite handy use flowers of sulphur in the same way; or failing in that, try ground talc. These should be mixed with oil in the same proportions, one of the powder to 10 of oil, and the dry part of the mixture must be thoroughly wetted before it is applied. In aggravated cases a stream of water from a hose will accomplish a good deal in bringing down bearings temperatures, or ice may be used packed around the bearings. Hot crank pin boxes may be cooled by fixing a hose in such a position that the crank will strike through the running water at about the upper part of its swing. The same scheme may be used with hot cross- head pin boxes. Small bearings that heat can sometimes be helped by flooding with kerosene or gasoline to wash out any dirt, grit or metal particles that may be in them. However, this method must be used with caution for the lubricating oil is also washed away. A large supply of regular lubricating oil must be at once poured in. If on hand the kerosene should be used in preference to gasoline as the latter cuts the lubricating oil more quickly. Sometimes a small bearing that heats can be helped by feeding all of the oil that can be run through it — flooding. This tends to bring down the temper- ature and may wash away some foreign particles. If a babbitted bearing heats pull it to pieces and look for black spots. Those are the places to scrape. Bearings can often be run much hotter than is generally thought. A bearing hot enough to blister the hand will run without injury. All important bearings, even of the self-oiling type, should be examined at frequent intervals. In many shops it is the practice to inspect all motor and generator bearings, at least once each day, and to thoroughly overhaul all of the line shaft bearings either once every six months or once a year, depending HINTS ON THE CARE OF BEARINGS 169 upon the nature of work. This means taking the bearings to pieces, thoroughly washing all parts, cleaning out the oil wells and refilling with fresh oil. Where large quantities of lubricating oil are used an oil filter will save money. As far as possible, do not let any oil from your bearings escape either by dripping or through atomizing. This is one of the most difficult points to care for in practice, particularly with high speed machinery. If oil is thrown out in quantities as in large drops or small streams, it is comparatively easy to stop it. Felt wiping rings at the end of the box are usually effectual but the material must not be too porous, or it may act as a sucking medium and instead of being a remedy be anoiher source of trouble. It is not unusual to fmd a central station where the walls of the room are covered with a coating of oil and dirt; around the generators and motors the atomized oil gathers particles of dust, settles in the air passages and may eventually shut off the ventilation. It is most difficult to stop this atomizing action. What must be done is to prevent any current of air passing through the bearing itself or across the surface of any body of oil. If this takes place oil will be vaporized and diffused throughout the entire room. The points where oil escapes can usually be determined by pieces of white paper or pieces of sheet steel painted a light color. The oil stains will show. The only remedy is to entirely shut off all air currents so that they cannot reach the oil. At the same time this is often a baffling problem to solve successfully. PART 11. BEARINGS WITH ROLLING CONTACT SECTION I Rolling Friction and Factors or Design for Ball Bearings The exact nature of rolling friction is not accurately known, but much splendid experimental work has been done with bearings that involve this princi- ple and we have empirical laws to govern designs. However, there is no theore- tical investigation giving results that can be applied to ball and roller bearings in the way that the work of Professor Osborne Reynolds can be applied to journal bearings with sliding contact. A consideration of the properties of ordinary journal bearings — bearings with sliding contacts — and ball and roller bearings presents a series of contrasts. At the outset we have sliding contact contrasted with rolling contact. In the journal bearing the friction is governed by the quality of the lubrica- tion. The better the lubrication the less the friction. In ball and roller bear- ings the reverse is true. It has been demonstrated by experiment that a well designed ball or roller bearing may give a lower coefficient friction when run- ning clean and dry than when lubricated with even a thin oil. This must not be taken to mean that such bearings should be run without oil or grease, for the reverse is necessary. ' In perfectly lubricated journal bearings there is no metallic contact between the surfaces. In ball and roller bearings there is always metallic contact. In journal bearings many different kinds of materials are used, running from wood and soft babbitt to hardened steel. With a few rare exceptions, the only material used for ball and roller bearings is steel, and in general it is hardened. The coefficient of friction for sliding contact is much higher than for rolling contact. The coefficient of rest and slow motion for sliding contact is very much greater than for motion at ordinary speeds. The coefficient of friction for the beginning of motion with rolling contact is not much, if any, greater than for motion at ordinary speeds. For this reason ball and roller bearings are peculiarly adapted for machinery having a high starting torque. In journal bearings running at a high rubbing speed the dissipation of heat becomes an important problem. With properly designed and operated ball and roller bearings the amount of energy liberated in heat is so slight that it is a negligible factor in design. Cooling devices are not used. With journal bearings practice indicates that a limiting speed is soon reached. For illustration, internal grinding fixtures with such bearings are seldom run 173 174 BEARINGS AND THEIR LUBRICATION over 12,000 to 15,000 revolutions per minute. By contrast, internal grinding spindles mounted on ball bearings are commonly run at speeds of 30,000 revolu- tions per minute, are reported to have been run commercially at a speed of 100,000 revolutions per minute,' and experimentally at a speed of 120,000 revo- lutions per minute. For heavy loads supported on vertical step bearings the ordinary step type with sliding contact has frequently been a source of trouble. On the other hand, there are ball and roller thrust bearings in use carrying loads up to 400,000 lb. and in one installation a load of over 2,000,000 lb. is sustained on a plain roller thrust bearing. FORMULAS FOR UNIT LOADS The formulas for specific load, or load per unit of carrying element, are as follows: W p= for roller journal bearings. ""id c W p= for radial ball bearings. c W p= — for roller and ball thrust bearings. n In these formulas Pr=the total load, />=the unit load, or load per carrying element, Z=the length of the journal bearings, c^=the diamenter of the rollers in a roller bearing, Z) = the diameter of the balls in a ball bearing, w=the number of rollers or balls, as the case may be and c=a constant = 5 for conditions recited below. It will be noted that the entire number of balls or rollers is not considered as carrying load, but the total number divided by c. It is clear that not more than half of the balls (or rollers) can be in the loaded side of the journal. It is also clear that that ball immediately in line with the direction of the load will carry most, while the balls (or rollers) at either side will carry consecutively less and less. For bearings that would take between 15 and 30 balls or rollers it will do to take ^=5. For other conditions the proportionate load imposed on each ball (or roller) must be figured; this is readily done mathematically or graphically. In many forms of ball and roller bearings the balls (or rollers) are separated, thus their number is less than the possible maximum that might be inserted did they lie in actual contact; n in the above formula is to be taken always as this possible maximum. ROLLING FRICTION AND FACTORS OF DESIGN 175 Professor Stribeck gives values of p for roller bearings and continuous ser- vice from 85 to 155 lb. Fig. 69 is taken from Stribeck's experiments and shows the relationship .uou 11 i i .040 1 1 1! I 1 1 i i i .030 |! 11 1 il ll 11 \ "i! .020 \ ■Mil \ MM \ 1 l\V \ \ D 1 \ \ 4\ \ \ .010 \v\ \ \ "-. 1^, \^ \ s^ •-^ ^^^ 1' % ^''- Sx W E^^^ h-: ^2 \ ^-0. ^^^ ■0— T*-*^ ^ tS_ — S:,_ -^ — IP— s:~ 10 20 30 40 50 60 70 80 Specific Load in Kilograms, (Kilograms and C.G.S. Units) 90 100 FIG. 69. RELATIONS OF COEFFICIENTS OF FRICTION FOR PLAIN, ROLLER AND BALL BEARINGS (Stribeck). between the coefficient of friction and unit load for journal bearings, several types of roller bearings and one type of ball bearing. Referring to the specific curves: Curves S are for a babbitted journal — S^ at iioo revolutions per 176 BEARINGS AND THEIR LUBRICATION minute, 6*2 at 380, ^3 at 64 and S^ at 12. Curves A are for rollers in a bronze cage, circumferentially fitting and supporting each roller throughout its length. A sleeve on the shaft and a liner in the box carry the rolls. Hard and soft rollers were tested; A, at 11 00 revolutions per minute, A2 Sit 285, ^3 at 190. Curves B are for hollow rollers on loosely fitted pins that connect the end cages. No liners are used on the shaft nor in the box. The speeds are 190 to 760 revo- lutions per minute. Curves D are for hollow, flexible, short, sheet steel rollers on loosely fitted pins that connect the end cages. No liners were used on the shaft or in the box. The speed was iioo revolutions per minute. Curves E are for flexible spiral rollers, placed loosely in the box and alined by three longi- tudinal bars connecting the end cages. No liner for the shaft or box are em- ployed. The speeds are 56 to iioo revolutions per minute. Curve F is for a ball bearing of the two-point, curved ball-race type (with race curvature but slightly in excess of ball curvature) at 65 to 11 50 revolutions per minute. The range of permissible specific load, as indicated by the appearance of the bearing after long continued runs, is shown by smafl circles on the curves. The above curves and description are taken from a translation and amplifi- cation of the work of Professor Stribeck, made by Henry Hess, and found in the Trans. A.S. M. E., Vol. XXVII. That celebrated work lays the foundation for the intelligent design of ball and roller bearings. As ball and roller bearings are made by only a few firms and as many of the devices are patented it will be necessary to refer frequently to specific bear- ings in the following pages. PROFESSOR STRIBECK'S INVESTIGATION Ball bearings first came into general use in the bicycle. They were of the so-called cone-bearing type and in general had an adjustable cup and cone and two-point contact. They were designed by rule of thumb and the basic principles were in obscurity. This was the condition of affairs in 1898, when the problem of determining the carrying capacity of balls and the laws for the design of ball bearings was attacked by Professor Stribeck, then director of the Scientific Technical Experi- mental Laboratories at Neu-Bablesberg, Germany. His solution set a thorough precedent. He first determined the load sustaining capacity of balls and found that it is proportional to the square of the ball diameter, and very much dependent upon the shape of the surfaces or grooves between which the ball is held. CURVATURE OF THE RACE GROOVES The governing principle in regard to the shape of the races is that the ball shall roll over them and not slide. ' Theoretically balls touch their contact sur- ROLLING FRICTION AND FACTORS OF DESIGN 177 faces in points only. Practically, even with balls of hardened steel of high elasticity they touch over surfaces of appreciable area instead of at points. Theory and experiments both prove that the two-point contact is the best. To- day all of the better known ball bearings and most of the ball-thrusts have two-point contact. If it were possible to so shape the groove that all parts would sustain load equally we would have a condition of maximum sustaining capacity of the ball. Such a groove would be represented by a profile semi-circular in section with radius of curvature equal to one-half the ball diameter. Because of excessive friction this form is not practicable. It is evidently one limit of the curvature of the race. Turning to the other limit, if we begin with a cylindrical race the sustaining power increases without a proportinate increase of friction as the curvature is in- creased up to the point of equality of curvature, when sliding sets in and with that a change from rolling to sliding friction. The best results come from proportions in which ^1 = 9/16 D and ^2= 25/48 D. Here ri = the radius of curvature of the outer race groove, r2 = the radius of curvature of the inner race groove, and Z) = the ball diameter. The radius for the outer groove will be seen to be slightly greater than the other, the reason for this being that the concavity of the outer race tends to embrace more of the ball than the inner race, which is convex. Even greater approximation to equality of race and ball curvature is desirable, but is impracticable owing to workshop difficulties. LOAD CARRYING CAPACITY OF RADIAL BEARINGS Early investigators of ball bearings took into consideration the crushing strength of the balls. This is an erroneous starting-point. The capacity of a ball in a ball bearing is not its crushing strength by any means, but it is its resistance to permanent deformation; some deformations take place with any appreciable load no matter how high grade the material from which the balls are made. The load carrying capacity of a ball bearing is directiy proportional to the number of balls and to the square of the ball diameter. According to Stribeck this capacity P is expressed by the equation P=k n U^ in which n= the number of balls, D the ball diameter, and k a cofficient depending upon the type of bearing, the material from which it is made and the speed at which it is run. Mr. Hess gives values of this factor k for a special steel used by his firm as follows, provided that D is expressed in units of an eighth of an inch: For uninterrupted race grooves with cross-sectional curvature as mentioned above, separated balls, uniformly distributed load and a uniform speed up to 3,000 revolutions per minute, ^ = 9. 178 BEARINGS AND THEIR LUBRICATION For full type bearings with the filling opening in one race at the unloaded side, otherwise as above, ^ = 5. For both ball tracks interrupted by filling openings, inelastic cage separators for the balls, or for full ball type; and speeds not over 2,000 revolutions per minute with a uniform distributed load, k =2.5. For thrust load on a radial bearing of the first type in this tabulation k =0.9. In general, the larger the number of balls, the smaller the value of k. The tadial load bearing is up to high speeds, practically unaffected by the speed as ro its carrying capacity. LOAD CARRYING CAPACITY OF THRUST BEARINGS Ball thrust bearings are made in three general types. In the first both races are flat, in the second one race is flat, the other grooved. In the third both races are grooved. The load carrying equation given by Mr. Hess is: P= — Vs In which F= the load capacity in pounds, k^ = Si factor determined by exper- iment and varying with the material and the shape of the ball track, n=the number of balls, D=the ball diameter and 5= the number of revolutions per minute of the bearing. Up to about 3000 revolutions per minute the effect of speed varies with its cube root. Mr. Hess gives the following values for k^. For material such as used by his firm and for races having grooves with a cross-sectional radius equal approximately to 0.82 D, k^ = 2^ to 40. For unhardened steel, such as is occasionally used for very large races and where there is no hammering or no sharp blows, ^^ = 0.5. When one or both races are flat k^ should be reduced to one-fourth the above value. The Standard Roller Bearing Company gives the following load capacity formulas for ball thrust bearings having a groove in each washer, stating that the ratings obtained from their use are very conservative and give a condition of loading under which a bearing can be guaranteed if properly installed and lubricated. The formula for the light type bearing with 17 balls is: P=32,ooo : in which P = the load capacity in pounds, D=the ball diameter in inches, n =the number of balls, a = the pitch diameter of the ball grooves in inches, S = the speed of the shaft in revolutions per minute. ROLLING FRICTION AND FACTORS OF DESIGN 179 The formula for the medium and heavy bearings having 11 and 9 balls respectively is: P= 10, 200 /- a's/S The notation is the same as given above. Ball thrust bearings are commonly of the two-point type to which all of the preceding formulas apply. However, a line of thrust bearings having four points of contact is made by the Auburn Ball Bearing Company. SECTION II CONSTRUCTION OF BALL BEARINGS From the fact that the load carrying capacity of a ball bearing is directly proportional to the number of balls, early designs were of the full type with an opening to admit the balls. Fig. 70 shows an entering notch at the side cutting through into the groove in the upper race. This notch was closed by means of a block, as shown at the left in Fig. 71. A modification of the same idea was FIG. 70. RADIAL BALL BEARING WITH BALL ENTERING NOTCH IN UPPER RACE. a hole through the outer race closed by a screw plug as shown at the right in Fig. 71. In both of these constructions the closing piece was located at the point of least pressure. A modification of the first device used in small bearings consisted in stopping up the notch at the side by means of a small screw. These early devices are credited to Riebe, Bierschenk and Bruhl. FIG. 71. DEVICES FOR CLOSING THE ENTERING NOTCH IN BALL-BEARING RACES. All of these constructions reduce the load carrying capacity of the bearing because of the weakened race section where the notch or opening is made and because after slight wear there develops an interference between the balls and the edges of the opening. Furthermore this type of bearing cannot take any thrust load, and in practically all constructions radial bearings are required to 180 CONSTRUCTION OF BALL BEARINGS l8l take a limited amount of thrust because of the end play of the shaits that they support. To overcome these obvious defects the partly filled type was developed. Here the balls are separated by means of spacing blocks or cages. As it is desirable to enter as many balls as possible two general constructions have grown up. The first is known as the eccentrically filled bearing, in which the inner race is brought in contact with the outer race, leaving a crescent shaped opening between them into which all of the balls but one are entered. They are then rolled around until the outer and inner races are nearly concentric and the last ball is snapped into place under a slight pressure. The races are per- fectly continuous and there is nothing structurally to interfere with the free, perfect rolling of the balls. This is the D. W. F. and Hess-Bright Construction. HofTmann also patented a process in which the outer ring is heated to expand it, or the inner cooled to contract it, or both operations performed simultane- ously. With this process it is not possible to introduce more than two or at the most three more balls than if no such device is resorted to. The other type is made with two modifications. In the first there are two filling notches, one in the outer and one in the inner race, sometimes introduced at an angle with the circumference, in which case they are inclined in opposite directions. The balls are filled through these notches. The other modification consists in grinding, on one side, the inner face of the outer ring eccentric with its circumference, so that this ground surface is tangent, or nearly tangent, to the bottom of the groove. This forms an opening through which the balls can be entered. The first construction is used by Fichtel and Sachs, the second by the Rhinelander Machine Works. THE BALL CAGE With the use of the pardy filled type it became necessary to hold the balls in positions equally spaced around the race. In fact, it is practically impossible so to proportion races and balls, even of the full-ball type, that the latter exactly fit the circumference. Furthermore, the cage has two other important uses. The one to reduce noise, the second to prevent wear. If a full ball type of bearing is set in motion under the usual vertical load- ing, whenever a ball passes from the region of load to the region where it is free from pressure it is snapped forward with a motion analogous to the snap- ping of an apple seed from between a boy's fingers. It strikes forcibly against the adjacent ball and inner race, bounds back against the outer race and contin- ues to vibrate. Before it comes to rest it is struck by the following ball which is then being subjected to the same action. This is a chief cause of noise in ball bearings. 1 82 BEARINGS AND THEIR LUBRICATION The second trouble that the cage overcomes is wear of one ball on another. With full type bearings, it is not uncommon to find balls each of which has a groove worn around its circumference with a radius of curvature equal to that of the ball diameter. One cause of this is the deflection of the shaft or journal, as in the case of a large gear which bends its shaft when its normal load or an overload is suddenly applied. The inner race of the bearing is turned obliquely with reference to the outer race. If the axis of the outer and inner races are co-incident, the path of the ball is a true circle. If, however, we twist one race into an oblique position with reference to the other and plot the ball path, we will find that it is no longer a circle, but a compound curve with two cusps. Thus in one revolution of such a bearing, the balls approach and recede from each other twice. This action interferes with the rolling action and causes sliding and pressure between balls and balls in full bearings; and between balls and cages where the latter are used. As the cages are stationary with reference to the balls and in the best forms also yield slightly this "binding" tendency is practically eliminated. It is important to keep all ball bearings properly alined. Cages were devised to keep the balls properly separated in a partly filled tjrpe bearing, to reduce noise and to prevent certain forms of ball wear. One of the most perfect theoretical forms is that of Conrad which consisted in separating the balls by a short helical spring with suitable bearing plates. This form was generally used for several years by the Hess-Bright and D. W. F. Companies and is still used as a special form for certain conditions. Other satisfactory forms consist of a cage made of brass or some form of bearing metal and so constructed as to be somewhat elastic. Many ball bearing failures can be traced to improperly designed and constructed cages, in fact it is not unusual in a failed bearing to find a cage literally ground to pieces. It is not a difficult matter to construct a cage which shall be satisfactory for a given set of conditions, but ball bearings are nowadays manufactured for stock in many thousands per day; it is therefore impossible to know whether a given bearing will be used at a few or many thousands of revolutions per minute, under uniform or variable loads or speeds, under no or great shock, etc. No cage has yet been found that will respond with equal satisfaction to all of these many conditions. Many attempts have been made to introduce a rolling member between the balls of ball bearings and the rollers of roller bearings. These are used instead of cages and are based on the conception that there will thus be only rolling contact between the various rolling elements and that pressure between them will then be rendered harmless. As a matter of fact there can be no pressure contact between balls or balls and separators in a correctly acting bearing. Such pressure is due either to faulty original design or to faulty action CONSTRUCTION OF BALL BEARINGS 183 from deflections, etc. In either case spinning or sliding will be caused and rolling prevented. It is clear that an element that must roll to fulfil its mission cannot do so when not allowed to roll at the critical periods. Both balls and rollers have been used for this purpose. FRICTION OF RADIAL BALL BEARINGS A great deal has been written in regard to the smallness of the coefficient of friction in ball and roller bearings. Undoubtedly this feature has been given much more importance than it deserves. No argument or reiteration is necessary to prove that bearings with rolling contact have less friction than bearings with sliding contact. It is impossible to generalize in regard to such coefficients of friction In a radial bearing the nature of the loading has a direct influence upon the shape of the friction curve. Figure 72 reproduced from the American Machinist ■ i 1.5 1.4 1.3 1.2 ^1.1 c 1.0 > N ™% ^s.oV .&>"1.^ / X^"^^ .2 0.5 >*"" ' ^^ c" i f^ k - - ^ ri^"" S' f\ . L >a r ^ piwau n uai, 0.4 0.3 , 0.2 — , , -Varfab" e ^ Radial .Ji^ — — 1 — 0.1 1 1 1 1" 1 1 r 1 1 - 400 800 1200 1600 2000 2400 2800 3200 Speed in R.P.M. FIG. 72. TYPICAL CURVES FOR A NO. 308 RADIAL BALL BEARING UNDER VARIOUS CONDITIONS OF LOADING. of March 4, 1909, shows three curves. The first is for a variable radial load, the second for a variable thrust load, the third for a combination of con- stant radial load and variable thrust load all on a radial ball bearing. Tables 40, 41, and 42 show the logs of the tests from which the curves were plotted. These tests were made on a special ball bearing testing machine, designed by Mr. Hess and built by Riehle Brothers Testing Machine Company. These coefficients and curves are characteristic of this type of bearing. Under a constant radial load the coefficient of friction for a given speed is nearly constant. With a continued constant radial load and variable thrust load, the coefficient increases in value and the curve rises rayidly. This con- dition is caused by a sliding action which increases in degree with an increase in the thrust load. With a variable thrust load only, the curve starts high, drops sharply and then continues at nearly a constant value. i84 BEARINGS AND THEIR LUBRICATION It must not be argued from this one set of curves that radial bearings should not be used under thrust. For speeds roughly below 1500 revolutions per minute the thrust type of bearing is best; at about this speed (the exact point varies with changes in size) the curve of the radial type under thrust crosses and falls below that of the pure thrust type. It will thus be good engineering practice to use a single radial bearing to carry both radial and thrust load, or to use one radial bearing to carry radial load and a second one to carry thrust load only, which it is made to do by free- ing its outer race circumferential! y and allowing it to rest against its radial faces only, or to use instead of this second radial bearing a separate bearing of the pure thrust type. Radial Coefficient of Oil tempera- load Vernier Vernier Difference Average friction in ture, degrees lb. per cent. Fahrenheit 400 44875 44661 0.00214 0.00107 0.1633 600 44851 44652 0.00199 0.000995 0.1517 800 44663 44489 0.00174 0.000870 0-1325 79.1 1000 44581 44390 0.0017 1 0.000855 0.1302 1200 44602 44439 0.00163 0.000815 0. 1240 1400 44524 44366 0.00158 0.00079 0.1200 1600 44506 44349 0.00157 0.000785 0.1195 82.5 1800 44443 44276 0.00167 0.000835 0.1272 2000 44467 44281 0.00186 . 00093 0.1416 84.0 2400 44452 44231 0.00221 0.001105 0.1686 2800 44469 44229 0.00240 0.00120 0.1830 85-5 Room temperature, 76° F. Speed, 325 r. p. m. Bearing, No. 308. Bore 1.5748 in. Rated radial load, 1450 lb. Table 40. — Log or Test on a No. 308 Radial Bearing under a Variable Radial Load. CONSTRUCTION OF BALL BEARINGS I8S Room temperature, 73°. F. Speed 325 r. p. m. Bearing, No. 308. Table 41.— Log of Test of a No. 308 Radial Bearing under a Constant Radial and Variable Thrust Load. Thrust Balance Weight Coefficient of Oil temperature load lb. Weight Average Torque friction in per cent. degrees Fahren- heit 2 CO 0.132 0.125 0.128 1 .262 0.803 73 400 0.158 0.14S 0153 1 . 509 0.479 600 0.218 0.201 0.209 2.061 0.436 76 800 0.288 0.267 0.277 2.732 0.434 1000 0-355 0-343 0.340 . 3.440 0.437 80 1200 0.412 0.413 0.412 4.061 0.431 1350 0.446 0.485 0.466 4.600 0-433 85 * Room temperature, 74° F. Speed, 325 p. m. Bearing, No. 308. Table 42. — Log of Test of a No. 308 Radial Bearing under a Variable Thrust Load. Certain forms have also been devised carrying two rows of balls, so arranged that original radial load develops thrust components and vice versa. Such a bearing is thus capable of carrying also simultaneous thrust and radial loads. While ball bearings all demand a very high degree of accuracy, this need is greatly accentuated for such two row construction. A location which demands a bearing of this kind is in the hubs of automobile wheels. A word should be said about the mounting of radial bearings. Good prac- i86 BEARINGS AND THEIR LUBRICATION tice calls for a fit of the inner race upon the shaft, such that it will not slide but rotate with it as a single member. Keying is objectionable because of the weakening of the race cross-section. Approved practice consists in making the inner race a pressing fit on the shaft and binding it laterally by means of jam nuts or some equi.valent device. A hammer should never be directly employed in contact with a ball bearing. Nor should the inner race ever be FIG. 73. HESS-B RIGHT DOUBLE-ROW BALL BEARING. forced home by pressure, or blows applied to the outer one, acting through the balls. If the outer race slides within its housing no harm is done. For that reason such a fit is usually made a light pressing fit or a "sucking fit" with a little freedom sidewise. In fact such sliding is beneficial and is generally insisted in by ball bearing makers as the slow creep gradually brings all portions of this outer race around under the load and so distributes its effect over the whole of the outer race track. CONSTRUCTION OF BALL BEARINGS TWO ROW RADIAL BALL BEARINGS 187 For reasons cited on page 185 several developments of double row or two-row bearings have appeared. Fig. 73 shows such a bearing by Hess- B right. This consists of an inner race, an outer race, two rows of balls and a ball separator. The inner ball track is concaved, the outer one flat, but inclined at an angle of 30° to the axis. The inner race radius center lies in the line passing through the ball center normally to the outer race face. The flat inner race allows an even distribution of load between the two rows. Fig. 74 shows a bearing of this kind made by the New Departure Manu- facturing Company. The balls are staggered, fit into a two-piece outer race, one-piece inner race, and the parts are assembled by a soft outer steel shell. FIG. 74. TWO-ROW BALL BEARING OF NEW DEPARTURE MANUFACTURING COMPANY. Fig. 75 shows a Swedish bearing of the S. K. F. Ball Bearing Company. The novel feature is the spherical bearing surface of the outer race which makes the bearing self alining, for the center of this spherical surface lies on the axis of the bearing. The method of caging the balls is shown as well as the method of assembling. It must not be assumed that carrying capacity is doubled by doubling the number of balls. Workmanship of the necessary accuracy is not feasible. When accuracy of workmanship is depended upon solely for load distribution, a liberal factor must be employed in reduction; when in addition races of small cur- vature as compared with the ball grooves are used this factor must be still further liberalized. See Section I, page 176, relating to curvature of race grooves. i88 BEARINGS AND THEIR LUBRICATION CONE-TYPE BALL BEARINGS Another type that is used in small sizes, as for magnetos, is a modification of the cone-type bearing of bicycle days. It is particularly serviceable where assembling is difficult, for the inner race can be slipped into its seat, the ball filled cage then inserted and the outer race added. This type must be adjusted with care, but it is a fallacy to think that adjust- ment can compensate for wear. CAGES FOR THRUST BEARINGS Experiment has shown that the balls in thrust bearings will run against each other in a manner similar to radial bearings if they are free, thus it is customary to cage the balls. . Mechanically such a cage has another use in that it assembles the balls into a unit, permitting them to be easily handled in the operations of shipping and inserting in the bearing housing. In order that each ball will take its proportionate share of the load, it is essential that at least one of the races should be provided with a spherical groove. FIG. 75. S. K. F. TWO-ROW RADIAL BALL BEARING. That the bearing may adjust itself to its load, it is common practice to make the lower race or ring with a spherical surface entering a spherical seat in the machine housing, or in an adaptor ring. The advantage of the latter construction lies in the difficulty with which a spherical seat is machined in large castings. Experiments have shown that the theoretical advantages of such a con- struction are realized in practice. CONSTRUCTION OF BALL BEARINGS 189 FRICTION OF THRUST BALL BEARINGS Fig. 76 shows a typical curve for a thrust ball bearing under a variable thrust load. Figs. 77 and 78 show a two-row thrust bearing made by the S. K. F.Bali Bearing Company. Because o| the spherical casing it is self alining. It will take thrust in either direction. ACCURACY OF BALL BEARINGS There is no other machine element produced in large numbers which calls for such accurate workmanship as do high-grade ball bearings. The accuracy is almost painful. No details can be overlooked or slighted. At the outset the material must be suitable in both balls and races, the parts must be properly hardened and tempered, and the grinding of the race grooves and balls must be accurate to within o.oooi in. It is a practical impossibility 2^ •2 .e 0.11 a 0.10 o.c 0.08 0.07 0.06 0.05 " ~ ~ ~ ~ "" ~ ~ - \ \ s V ^a nab Jp rr rhl, st ji 21 iJ L-J ' — 1 p- - J 200 400 FIG. 76. TYPICAL CURVES /OR A 600 800 1000 1200 1400 1600 Speed in R. P.M. SMALL THRUST BEARING UNDER VARLABLE LOAD. commercially to grind all balls to a limit of o.oooi in. In practice, therefore, they are sorted into grades and the allowable limit of variation in each grade is O.OOOI in. When a bearing is assembled, care is taken not to mix balls of different grades. Only by adopting this course is it possible to produce a bear- ing in which the load will be properly distributed between the balls with an assurance that one ball if large may not at times take all of the load, or, if small, receive none, allowing its neighbors to carry its share. Furthermore, it is the only way in which to produce a ball bearing that will run quietly. Irregularity among the balls themselves is a fruitful source of noise; possibly second only to inaccuracies in the races. As a general rule, however, the greater the inaccuracies in both races and balls, the greater will be the noise in running. Only by having balls of practically equal diameter can a bearing be made that will run quietly. IQO BEARINGS AND THEIR LUBRICATION This indicates the degree of accuracy required of the bearing manufacturer. There is another side to the question, however, and that is the accuracy of the parts of the machine in which the bearing is fitted. If the two-part housing in which the outer race is held does not fit, it is more than likely that if tightly clamped, it distorts the race and presses upon the FIG. 77. S. K. F. DOUBLE THRUST BALL BEARING. balls, to the detriment of the running of the bearing and perhaps causing its failure. Again, if the shaft is too much larger in diameter than the bore of the inner race, the latter may be expanded slightly and the balls pinched. These are two common causes of failures in ball bearings. Another trouble is overloading. This is the chief cause of abrasions. If the abrasions are in the middle of the groove of either a radial or thrust bearing, /%JI FIG. 78. SECTION OF S. K. F. DOUBLE THRUST BALL BEARING. they probably indicate that the load in a normal direction is too great. If abra- sions are on one side of the groove in a radial bearing, they indicate that the thrust load is too great. If the track of the balls in a radial bearing swerves from side to side, it indicates deflection of shaft or bearing, or original mis- alinement. CONSTRUCTION OF BALL BEARINGS 191 MATERIALS FOR BALL BEARINGS The material used for ball bearings for machinery is steel. In the higher duty bearings, both balls and races are made from special alloy (tool) steel that has a very high elastic limit after hardening and drawing. In some bearings, both balls and races are made of a low carbon steel with case-hardened surfaces. For severe service such bearings have not proved successful. The elastic limit of the material is much lower than of the alloy (tool) steel,' and in the ball there is a tendency for the hard shell to separate from the softer core; this is in evidence also in the races. In very large thrust ball bearings a high carbon unannealed tool steel is sometimes used without hardening. This is because of the difficulty and expense of hardening and grinding such large rings without the accompanying losses from warping and checking and the difficulty of getting perfect uniformity of surface hardness. Hess-Bright (D. W. F.), however, produce races up to 54 in. diameter of a special alloy steel that are hardened throughout, as are their smaller ones. PROPORTIONS OF BALL BEARINGS The proportions of ball bearings are well standardized. Oddly enough, ball diameters are universally expressed in inches and fractions thereof, while the dimensions of the races are given in either millimeters or English units. German builders use millimeters with a single excej^tion where a firm has developed a series in English units, adapting them for the British trade, though that also uses chiefly ball bearings in millimeters. Even in England most ball bearings are made to millimeters as is also the more general practice of American manufacturers who have followed the German example. This gen- eral adoption of the millimeter dimensions is due to the fact that early German makers rehabilitated the ball bearing by the development of the principles and construction data of the modern type and secured a wide vogue for their pro- ducts that made the sizes standard. As a rule each manufacturer makes a wide and narrow type of radial bearing, with three series for each; namely, light, medium, and heavy. In some cases a fourth series has been standardized, known as '* extra heavy. " Thrust bearings are usually made in three series, light, medium, and heavy. The heavier the bearing the larger the balls for a given strength. At the 191 1 Spring Meeting of the Society of Automobile Engineers, the standards committee rendered a report containing Tables 43, 44, and 45, giving recommended standard dimensions for the light, medium, and heavy series, respectively, for radial ball bearings. These standards are referred to as **Ball Bearings Standards A." It will be noted that the last column gives a 192 BEARINGS AND THEIR LUBRICATION radial load rating in pounds for each bearing. In this connection the committee reported: ''Attention is called to the fact that the capacities given in the tables are based upon ball bearings manufactured of suitable workmanship and of suitable material and running at uniform speed and uniform radial load, the speed not exceeding 500 revolutions per minute." " It is further suggested in explanation of the load standards that it cannot be expected that all conditions will be covered by the loads given. ^ For con- ditions of shock, actual thrust, and a combination of the two, greater factors of safety will have to be used." Table 46 is taken from this same report and shows the tolerances or limits worked to by many bearing manufacturers as embodied in statements made to the committee. The table applies to outer race diameter, bore and width. Other standard dimensions are "not given here for they can be found in manufacturers' catalogues. Table 43. — rEoroRTiONs of Radial Ball Bearings — Light Series CONSTRUCTION OF BALL BEARINGS 193 No. of bearing Bore Diameter Width Comer at bore of inner race Radial load in lbs. Mm. In. Mm. In. Mm. In. Mm. In. 300 10 0.39370 35 1.3779S II 0.43307 0.04 200 301 12 0.47244 37 1.45669 12 0.47244 0.04 240 302 15 0.5905s 42 1.65355 13 0.51181 0.04 280 303 17 0.66929 47 I .85040 14 0.55118 0.04 370 304 20 0.78740 52 2.04725 15 0.59055 0.04 440 305 25 0.98425 62 2.4409s 17 0.66929 0.04 620 306 30 1 .18110 72 2.83465 19 0.74803 3 0.08 860 307 35 I. 37795 80 3.14962 21 0.82677 2 0.08 IIOO 308 40 1.57481 90 3.54332 23 0.90551 2 0.08 1450 309 45 I. 77166 100 3.93702 25 0.9842s 2 0.08 1750 310 SO I. 96851 no 4.33072 27 I .06299 2 0.08 2100 3" 55 2. 16536 120 4.72443 29 1.14173 2 0.08 2400 312 60 2.36221 130 5.11813 31 1.22047 2 0.08 2800 313 65 2.55906 140 5.51183 33 I .29921 3 0. 12 3300 314 70 2.75591 150 5.90554 35 1.37795 3 0. 12 4000 315 75 2.95277 160 6.29924 37 1.45669 3 0. 12 4400 316 80 3.14962 170 6.69294 39 1.53544 3 0.12 5000 317 85 3.34647 180 7.08664 41 I .61418 3 0. 12 5 700 3i8 90 3.54332 190 7.48035 43 I .69292 3 0. 12 6400 319 95 3.74017 2PO 7.87405 45 I .77166 3 0. 12 7000 320 100 3.93702 215 8.46460 47 1.85040 3 0.12 7700 321 los 4.13387 225 8.85830 49 I. 92914 3 0. 12 8400 322 no 4.33072 240 9.44886 SO I. 96851 3 0. 12 1 0000 Table 44. — Proportions of Radial Ball Bearings — Medium Series Bore Diameter Width No. of bearing Comer at bore of inner race Mm. In. Mm. In. Mm. In. Mm. In. Radial load in lbs. 40, 17 0.66929 62 2.44095 17 0.66929 I 0.04 850 404 20 0.78740 72 2.83465 19 0.74803 2 0.08 1050 405 25 0.98425 80 3.14962 21 0.82677 2 0.08 1320 406 30 I . 18110 90 3-54332 23 0.90551 2 0.08 1600 407 35 1.37799 100 3-93702 25 0.98425 2 0.08 1900 408 40 I. 57481 no 4-33072 27 I .06299 2 0.08 2200 409 45 I .77166 120 4.72443 29 1.14173 2 0.08 2500 410 so 1. 9685 1 130 5.11813 31 1.22047 3 0.08 3400 411 55 2. 16536 140 5.S1183 33 I . 29921 3 0. 12 ^900 412 60 2.36221 ISO 5-905S4 35 1.37795 3 0. 12 4400 413 6S 2.55906 160 6.29924 37 1.45669 3 0.12 4900 414 70 2.75591 180 7.08664 42 1.6535s 3 0. 12 6200 41s 75 2.95277 190 7.4803s 45 I. 77166 3 0. 13 6600 416 80 3.14962 200 7.87405 48 1.88977 3 0. 12 7300 417 85 3-34647 210 8.26775 52 2.04725 3 0. 13 8580 418 90 3-54332 22s 8.85830 54 2.12599 3 O.I3 1 0000 419 95 3-74017 250 9.84256 SS 2.16536 3 0.12 11880 420 100 3.93702 26s 10.43311 60 2.36221 3 O.I3 14000 Table 45. — Proportions of Radial Ball Bearings — ^Heavy Series. 13 194 BEARINGS AND THEIR LUBRICATION Recommended by users Rhineland Warner Gear Co. S. R. Shepard C. E. Davis H ■ l.ltUli,l I '0 I. It U lit I g i OOOOhhmW OOOOOhwmP* C III 000 ■5 H N H H Tt- « C« ■ M 5 ■* SECTION III Typical Designs' and Mountings for Ball Bearings In this section are given a number of typical designs and mountings for ball bearings drawn from the practice of a few manufacturers, both foreign and domestic, and applying to various kinds of machines. Obviously other firms could have furnished similar data had there been space to include them. The first examples are taken from machine tools and indicate the interest that their builders have felt in adapting friction reducing bearings to their product. Figs. 79 to 82, inclusive, are of grinder spindles, ball bearing mounted. Fig. 79 is from an article by W. M. Byorkman, printed in the American Machinist of February 9, 191 1, page 242. m^mmmmmm^mmm. | FIG. 79. MOUNTING OF GRINDER SPINDLE ON BALL BEARINGS AND FLOATING BUSHINGS. It shows a grinder spindle mounted on radial ball bearings with floating bushings. These bushings are used to give steadiness to the spindle as in work where the dimension tolerances are small. They are located close to the ball bearings and have inside and outside clearances of about 0.0002 in. They have end clearances only sufficient to permit of appreciable movement. The bushings carry no load, but the clearance spaces are filled with oil which acts like a dash pot to suppress the minute vibrations which would otherwise arise in the ball bearings. The final effect is to do away with the "waves"' frequently noticed on ground work where the grinding spindle is ball bearing mounted 195 196 BEARINGS AND THEIR LUBRICATION The endwise movement of the shaft is prevented by a pair of thrust ball bearings seen at the right hand end. This particular spindle arrangement v^^as applied to a grinding machine for sizing to thickness type bars of a typewriter. Adjusting Screw FIG. 80. SECTION OF SPINDLE BEARINGS OF RIVETT GRINDER. Fig. 80 shows in section the spindle and bearings of a Rivett grinder. From a test made by A. Morin in Paris, France, a speed of 120,480 revolutions per minute has been obtained with a spindle of one of these machines. This is FIG. 81. BALL-BEARING MOUNTINGS FOR HIGH SPEED SPINDLE AND ITS PULLEY. undoubtedly the highest speed ever recorded for a rotating machine element driven by mechanical power. Figs. 81 and 82 show two recommended grinder spindle mountings developed Endclosure Lips Bored Vei Inch larger than Shaft. Inner Races of Radial Bearings Light Drive Fit on Shaft. Outer Races of Radial Bearings Sucking Fit in Housing. Washers for Adjustment of Thrust Bearings. FIG. 82. BALL-BEARING MOUNTINGS FOR BUFFING OR GRINDING HEAD. by the Hess-Bright Manufacturing Company. In Fig. 81 it will be noticed that the driving pulley is carried on two radial bearings supported by a sleeve inde- pendent from the spindle itself, thus the stress of the belt pull does not reach the TYPICAL DESIGNS AND MOUNTINGS 197. spindle. Another feature is in the right hand radial bearing of the spindle itself. The outer race instead of having a groove for the balls has a plain cylin- drical surface. Thus the balls have a chance to move in the direction of the length of the spindle if the parts lengthen from heat. c — ^sSS^ 83. BALL-BEARING THRUST MOUNT INGS FOR DRILLER SPINDLES. BALL-BEARING MOUNTING FOR A VERTICAL SPINDLE. Fig. 82 needs but little comment except to point out the grooves at the ends of the bearing housings provided to prevent the escape of oil and the entrance of dust. Top Bearing if Pulley is Placed on Bottom Top Bearing .^rrrm if Pulley ' ^^ is Placed ,,-- on Top Bottom Bearing if Pulley is Placed ou Bottom Bottom Bearing if Pulley is Placed, on Top 85. BALL-BEARING MOUNTING FOR ELECTRIC MOTOR SHAFT. FIG. 86. BALL-BEARING MOUNTINGS FOR VERTICAL ELECTRIC MOTOR. Fig. 83 shows two mountings for driller spindles and is likewise recommended by the Hess-Bright Manufacturing Company. Fig. 84 shows a vertical spindle resting on a ball thrust at the bottom and guided by two radial bearings. It is from designs of the Rhineland Machine 198 BEARINGS AND THEIR LUBRICATION FIG. 87. BALL-BEARING MOUNTING FOR STREET RAILWAY MOTOR SHAFT. FIG. 88. BALL-BEARING MOUNT- INGS FOR TRUCK WHEEL. FIG. 89. BALL-BEARING MOUNTING FOR FIG. 90. CENTRIFUGAL OILING DEVICE AND WORM AND WORM WHEEL. BALL-BEARING MOUNTING FOR A SHAFT STEP. FIG. 91. BALL-BEARING SHAFT MOUNTING WITH SELF-ALINING SEAT. TYPICAL DESIGNS AND MOUNTINGS 199 Works Company. An important point in installing thrust bearings is this: If the upper race is tight on the shaft, the lower leveling washer must be free to a slight extent to allow for movement in preserving alinement. On the other hand, if the upper washer is made with a slight clearance over the shaft diameter the lower ring need not be free. FIG. 92. BALL-BEARING PILLOW BLOCK. FIG. 93. TWO FORMS OF SHAFT ADAPTERS FOR BALL BEARINGS. MOTOR BALL BEARINGS Figs. 85, 86 and 87 show three applications of ball bearings to the shafts of electric motors. Fig. 85 is a Ivhineland design for a small sized motor. The clearance in an axially direction of the bearing at the right hand is very plainly FIG. 94. BALL-BEARING SWIVEL FOR A LIFTING EYE. shown. Fig. 86 shows two mountings for the shaft of a 45-h. p. vertical motor involving two different combinations of radial end thrust bearings. It is from the practice of the Standard Roller Bearing Company. Mountings of ball bearings for street railway motors are shown in Fig. 87. A large number of motors with this type of mounting have been in use on some of the street railway lines in New York City. The design is Hess-B right. 200 BEARINGS AND THEIR LUBRICATION FIG. 95. BALL-BEARING MARINE PROPELLER BLOCK. Oil Returning Grooves A Screw for Locating Leveling Washer so that Grooves A are kept Coincident with Sloping Channels in Casing Walls. Required for Returning of Oil to Center of Beariag Gouge Slot Across Oil Retaining Grooves at Convenient ,ce to Afford Sharp Edge fur Wiping Oil from Shaft FIG. 96. COMBINATION RADIAL AND THRUST BALL-BEARING MOUNTING. FIG. 97. BALL-BEARING MOUNTING FOR THREE-BEARING, FOUR-CYLINDER AUTOMOBILE CRANK SHAFT. TYPICAL DESIGNS AND MOUNTINGS 201 GENERAL APPLICATIONS Figs. 88 to 96, inclusive, show a number of ball bearing mountings applied to machinery in general. Of these Figs. 88 to 94, inclusive, are Rhineland designs; Fig. 95 is a Hess-Bright design and Fig. 96 a Standard Roller Bearing design. FIG. 98. BALL-BEARING MOUNTING FOR TWO-BEARING, FOUR-CYLINDER AUTOMOBILE CRANK SHAFT. Fig. 88 shows two radial bearings applied to a heavy truck wheel. Fig. 89 presents the mounting of a worm and worm wheel with both radial and thrust bearings. The thrust bearings are of the Rhineland double thrust type. Fig. 90 is of interest for the design of a centrifugal oiler in a chamber below the step whereby oil is pumped from the lower level of this chamber upward through FIG. 99. BALL-BEARING AUTOMOBILE mJB MOUNTING — FLOATING TYPE. FIG. 100. BALL- BEARING AUTOMOBILE HUB MOUNTING — FIXED TYPE. the center of the shaft and then outward through channels into the spaces sur- rounding the balls. The oil finds its way back to the reservoir through an annular space around the bearing. Figs. 91 and 92 are pedestal box mountings. In Fig. 91 the housing for the bearings has a spherical mid-section fitting a spherical seat in the pedestal frame. This allows for self alinement. Fig. 92 shows the outer races of the 202 BEARINGS AND THEIR LUBRICATION ball bearings provided with spherical surfaces resting against cylindrical sur- faces inside of the pedestal frame. It will be noted that the inner surface of the inner ball races is conical and is fitted to the cylindrical shaft by means of an adapter bushing. Fig. 93 shows iwo adapter constructions of Hess-Bright design. The one at the right has a combination of two tapered bushings, one surface of each fitting the cylindrical shaft and the cylindrical surface of the inside of the inner race, respectively. FIG. lOI. BALL-BEARING MOUNTING FOR AUTOMOBILE REAR AXLE DIFFERENTIAL AND HUB, Fig. 94 shows a thrust bearing mounted in the swivel of a large lifting eye. The following figure, Fig. 95, illustrates a marine propeller box fitted with two thrust bearings and two radial bearings to carry the load of the shaft. Further description is unnecessary. The design of Fig. 96 is a large thrust bearing for a 2 1/2 in. shaft combined with two radial hearings to carry vertical load. A FEW AUTOMOBILE APPLICATIONS Figs. 97 to 1 01 inclusive, show a few automobile ball bearing mountings. Of these, .Fig. 97 is from the Rhineland Company and the others from Hess- Bright. Fig. 97 shows three bearings for the crankshaft of a four-cylinder automobile. Each bearing consists of two radial bearings placed side by side and having the TYPICAL DESIGNS AND MOUNTINGS 203 external surface of the outer races of each pair ground to the same radius of curvature. This makes each pair self alining and is a form of two-row ball bearing not mentioned in the previous discussion of this subject. The other figures in this group do not need particular mention. FIG. 102. BALL THRUST BEARING WITH FLAT WASHERS. FIG. 103. BALL JOURNAL BEARING FOR FLOATING SHAFT. The final construction, Figs. 102 and 103, are from the Standard Roller Bearing Company and show, respectively, a plain ball thrust having flat washers — a type previously described — and a special ball journal bearing developed for use on a spindle which had considerable endwise travel. SECTION IV Lubrication of Ball Bearings No better discussion of the lubrication of ball bearings can be found than is given in an article in the American Machinist, Vol. XXXIV, page 87, by W. L. Batt. It is given herewith. Oil is essential for the operation of plain bearings, but it is not so well known that ball bearings likewise require a certain amount of lubrication; it is all too frequently assumed that the latter may run entirely dry. Contrary to this assumption lubrication has a very positive action in a ball bearing. This action is many fold, existing both between the ball and the path upon which it rolls and between the ball and the separator. Though the principle of the ball bearing is a rolling one, there is in the most perfectly designed bearing a sliding action which, however slight, must yet be provided for. If balls and raceways were absolutely incompressible there would in practice be solely point contact and, therefore, a pure rolling action, but as far as present day mechanical processes go even the most perfect of materials are somewhat elastic and there is an actual area of deformation between the ball and the raceway under load. At the extremities of this area, vs^hose diameter is, of course, very small, sliding must exist, but the presence of a lubricant at this point renders what small amount of friction there is almost negligible. Improbable as it may seem, there is an actual film of lubricant maintained between these surfaces of contact, however minute they are, providing the lubricant be one of sufficient viscosity. PROTECTIVE VALUE OF THE LUBRICANT Aside from its lubricating quality, oil or grease acts as a protecting agent for the ball bearing. The finely polished surfaces of balls and raceways -are subject to attack by rust through atmospheric action and may be damaged by the entrance of foreign matter from the outside, such as grit and dust. Lubri- cant effectually surrounding the bearing will not be penetrated by this atmos- pheric moisture and thus the surfaces of action are preserved in their original finely polished condition. If, however, rust or grit should be allowed to attack the carrying surfaces these are quickly destroyed and an earlier failure of the bearing may be expected. In order that the lubricant may in itself be no source of danger, it must 204 LUBRICATION OF BALL BEARINGS 205 respond to certain requirements; incidentally requirements exacted of the ball- bearing lubricant are thoroughly desirable in one to be used in plain bearings as well. REQUIREMENTS OF A BALL BEARING LUBRICANT First and most vital is the requirement that the lubricant of itself shall do no damage to the bearing, neither originally nor through deterioration. The most common fault in oils is the presence of free acid or its development. Free acid is never found in properly refined mineral or hydro-carbon oils, nor will these deteriorate due to the action of the atmosphere, and therefore the only kind of oil satisfactory for ball bearing lubrication is mineral oil. Vegetable oils, such as castor, cotton-seed, rape, linseed and the like are barred, principally because of their tendency to gum up, become rancid and develop acid. The animal oils are objectionable for the same reason. A simple test for the presence of objectionable elements in an oil or grease is to coat a finely pol- ished surface of steel or brass or the surface of a ball bearing, with the oil or grease in question and expose it to the sunlight for a few weeks; at the end of that time the steel should show its original polish. Any discoloration or drying up of the lubricant on the steel is sufficient to bar it from consideration. The so-called petroleum greases, among which are vaseline and cosmoline, have no detri- mental action in themselves, since they are derivatives of mineral oil ; they have, however, low viscosity and a low melting-point — 100 to 125° F.; they pound out thin in action and have little lubricating value. Their use is limited to very slow speeds and their chief advantage is low cost. Just as acid is a thing to be guarded against in oils, so is tree alkali the most common enemy to ball bearings, among the greases. The familiar yellow cup grease is usually a combination of a mineral oil and some vegetable or mineral oil or fat which latter is saponified by the addition of a caustic; the result is a lubricant having body and stiffness The saponifying material should be small in quantity and very carefully compounded, else free alkali, having a detrimental action on the steel may result The action of an alkali, as of an acid, is to pit or etch the surfaces upon which it is deposited. The addition of mica, ground cork, wood and such substances, frequently added to overcome noise in gear cases of automobiles, is a positive menace to the ball bearing, since this foreign matter opposes free ball rotation; if it be present in large enough amount, the result may easily be that the balls are wedged between the raceways and actual fracture may result. Certainly the free rolling quality of the ball bearing will be lost. The question of the beneficial effect of graphite in ball bearing lubrication is one often asked. The answer is simply that graphite in any shape or form that will settle and pack with time when quiescent cannot be of assistance to 2o6 BEARINGS AND THEIR LUBRICATION the ball bearing itself. Whether it is a detriment of enough importance to offset its undoubted advantage to gears or whatever the mounting may be, depends entirely on the kind and .fineness of the graphite and the proportions in which it is added to the lubricant. It may readily be seen that any considerable amount of graphite settling in the bottom of the raceway will interfere with the free rolling of the balls in the same manner as ground cork or any other substance. The claim made for graphite, and one certainly proved in the case of the plain bearing, is that it deposits itself in a fine film on the surface upon which the shaft may revolve. A microscopic examination of the balls and grooves of a carefully made ball bearing shows almost no scratches or inequalities; after such a bearing has been run in oil and graphite of proper proportions, examination again will show that absolutely no graphite has been deposited. Frequent observations carefully made show a slight increase in the friction when graphite is present over that when oil alone is used. Such results clearly show that graphite has no value in so far as reduction of ball bearing friction may be concerned. Over the usual life of a ball bearing, however, probably no actual damage will result and considerable good may have accrued to the elements carried upon the ball bearings; such use would, of course, be justifiable. The damage, however, from graphites that settle and accumulate in the bearings cannot be overlooked. EFFECT OF SPEED UPON CHOICE OF LUBRICANT The effect of speed upon the choice of lubricant must likewise be considered. As a general statement, greases are desirable for use up to just as high a speed as is possible. An undue increase of temperature and a thinning out of grease will indicate the limit of its usefulness; when this is noted oil must be employed. Oil has a higher viscosity than grease and theoretically is a better lubricant, but its advantage in this regard is more than offset by the difficulty of its proper retention, in the bearing case. Such case may be solidly packed with grease that will thus not only lubricate the bearing, but will also prevent foreign substances from entry. If oil be used the ball will merely dip into the oil, splash- ing it over the entire bearing, but the bearing is then more open to the entry of foreign matter. There will be more internal friction in the grease than in the oil and it is this that at high speed produces excessive temperature. As a general thing greases may be used up to about 2000 to 2500 revolutions per minute. MOUNTINGS MADE TO RETAIN LUBRICANT In order that the lubricant may be effectually retained, various arrange- ments are used, depending upon the conditions surrounding .the bearing. In the simplest sort of mounting for a radial bearing the shaft projects through LUBRICATION OF BALL BEARINGS 207 the casing, and the casing itself is provided with two lips, between which is a space for lubricant. If the condition be such that additional protection is needed, one additional groove is provided and this may be fitted with a cup of some sort from which grease will be steadily fed to the groove to keep that filled. This makes a definite frictionless packing. For still more severe con- ditions a third groove is added; in this latter groove felt is occasionally placed, whose adherence to the shaft is guaranteed by some sort of spring tension. Unfortunately, this is subject to drying out and thus loses its efficiency; when that occurs it is a positive detriment. The single or multiple groove arrangement empty or with only grease filling is the most effective; but it is essential that the bore of the lips be not more than 1/64 in. larger than the shaft in diameter, that the lip edges be sharp instead of rounded over, and that the lips be at least 3/32 in. wide. The grooves may also be cut in the shaft, leaving bands between with sharp edges. The only objection to this is the weakness of the shaft. HINTS ON THE CARE OF AUTOMOBILE BALL BEARINGS The preceding paragraphs point out the need of lubrication and the necessity of keeping ball bearings free from all foreign particles. V. W Page, writing in the American Machinist , gives a number of hints on the care of ball bearings in automobiles. He says: "Many cases of trouble in rear axle and transmission case bearings have been traced to the presence of minute particles of metal ground off from the gears, or to particles of sand loosened from the interior of the gear case or rear axle housing castings. In sliding gears, especially when operated by an inexperienced person, there is a constant clashing of the pinions in changing speeds, which tends to loosen particles of metal from the teeth. These fall into the lubricants and ofttimes find their way into the ball races. The rapid failure of the bearing is the inevitable consequence. Likewise failures in ball bearings on automobile engine crankshafts have been traced to the pres- ence of particles of carbon and other foreign matter in the oil. Another source of trouble is the washing of rust and dirt into the bearings through carelessness in washing the automobile. This is particularly true of wheel bearings, where many failures have been traced directly to the presence of rust and dirt caused by the indiscriminate use when washing with a stream of water under pressure of 40 to 50 lb. per square inch; that is, the normal pressure of the city water mains. Another practice that is detrimental to ball bearings is used in many motor car repair shops. It consists in dipping the bearings for the purpose of cleansing into dirty gasoline in which perhaps gears and other parts covered with metallic particles have already been washed. 2o8 BEARINGS AND THEIR LUBRICATION Still another practice prevalent not only in garages and repair shops, but in too many machine shops is the employment of a hammer, particularly a hard hammer, in assembling or dismounting ball bearings. It will be remembered that these are made of hardened steel. The man who hammers a reamer or cutter is immediately discharged, but gets only a reprimand when he similarly abases a ball bearing. SECTION V Roller Bearings with Flexible Rollers Roller bearings with hollow, cylindrical, helical, flexible rollers made by the Hyatt Roller Bearing Company are used in machinery in general, on line shafting and in automobiles. The rollers are wound from flat strip steel into a closed helical coil. Thus they are flexible and can adapt themselves to slight irregularities in either journal or box without causing excessive pressure. The cylindrical hollows in rollers serve as storage space for lubricant and the helical FIG. 104. FLEXIBLE ROLLER BEARING MOUNTINGS FOR TRANSMISSION SHAFTS OF MULTIPLE DRILLER. interstices distribute the oil the entire length of the box. One-half of the rollers in a box have a right-hand helix, the other half left-hand. In the form of bushings, two types are made, known as the standard type and the high duty type. In the standard type the rollers are of carbon steel with 14 209 2IO BEARINGS AND THEIR LUBRICATION an outer shell or lining of special analysis sheet steel. The rollers run in con- tact with the shaft or journal. Where the bearing surface is generous it is satisfactory to operate the rollers direct on soft steel surfaces. In the second type the rollers are of nickel steel, 31/2 per cent, nickel, and FIG. 105. FLEXIBLE ROLLER BEARINGS MOUNTING FOR A CIRCULAR SAW ARBOR. heat treated. The lining is tubular and a tubular sleeve is provided to slip over the shaft or journal. Both of these parts are also heat treated. Several devices are used to cage the rollers, which are squared on the ends to thrust against the ends of the box. As the allowable unit bearing pressures of the high duty rollers are higher than for the standard rollers the high-duty bearings m. ^W FIG. 106. FLEXIBLE ROLLER BEARING APPLIED TO COUNTERSHAFT BEARING AND LOOSE PULLEY. have the practical advantage of being shorter for the same load than the com- mercial bearings. The shafting bearing is of the standard type with a horizontally split box so that it can be put on a shaft anywhere and does not have to be slipped on over the end. ROLLEI^ BEARINGS WITH FLEXIBLE ROLLERS 211 FIG. 107. FLEXIBLE ROLLER BEARING MOUNTING FOR CLUTCH PULLEY. FIG. 108. FLEXIBLE ROLLER BEARING FOR FLOUR MILL. 212 BEARINGS AND THEIR LUBRICATION COEFFICIENTS OF FRICTION In the Trans. A.S.M.E., Vol. XXVII, page 504, Prof. A. L Williston gives the results of various experiments with Hyatt bearings. From that source, the following coefficients of friction are taken: Average for various loads at 130 r. p. m., 0.011-4; at 302 r. p. m., 0.0099; ^t 585 r. p. m., 0.0147. The average for very heavy loads at 185 r. p. m. was o.oii; at 215 r. p. m. 0.0093. A test made by the author comparing Hyatt standard lineshaft bearings with babbitted bearings gave a saving for the roller bearings of 64.9 per cent. The shaft was 152 ft. long, 2 15/16 in. in diameter and was supported by 20 hangers. ALLOWABLE PRESSURES AND SPEEDS The allowable pressures for the standard or commercial type are fixed by the quality of the shaft or journal against which the rollers bear. For low speeds riG. log. FLEXIBLE ROLLER BEARING FOR PIG BREAKER. up to 50 r. p. m. the maximum limit lies between 400 to 500 lb. per square inch of projected journal area. The method of housing, particularly in its relation to the distribution of load, and quality of lubrication have an in- fluence in determining these limits. With an increase of speed the allowable pressure decreases. For lineshaft bearings up to 3 15/16 in. in diameter running up to 600 r. p. m. 30 lb. per square inch is considered good practice. For larger shafts the same factor is allowable up to speeds of 400 r. p. m. The high-duty bearings carry much greater unit loads. A rating of 750 lb. per square inch at 1000 r. p. m. is conservative. A limiting maximum speed for the standard or commercial bearings is about 1500 r. p. m.; for the high-duty bearings, 3000 r. p. m. ROLLER BEARINGS WITH FLEXIBLE ROLLERS 213 LUBRICATION These bearings must be well lubricated. If the speed is medium or high, a good body machinery oil is suitable. If the speed is slow a heavy body machinery oil or grease is better. FIG. no. FLEXIBLE ROLLER BEARINGS MOUNTING FOR SHAFTS OF A VERTICAL PUMP. TYPICAL APPLICATIONS IN MACHINERY Several typical designs and machinery applications are shown in Figs. 104 to no, inclusive. The first, Fig. 104, is of shaft mountings, for a multiple- spindle driller; the second, Fig. 105, is a circular saw arbor on two bearings. 214 BEARINGS AND THEIR LUBRICATION FIG. III. STANDARD FLEXIBLE ROLLER BEARING APPLIED TO BEVEL GEAR REAR AXLE WITH BEARING ON DIFFERENTIAL HUB. FIG. 112. STANDARD FLEXIBLE ROLLER BEARING APPLIED TO BEVEL GEAR REAR AXLE WITH BEARING MOUNTED ON SOFT AXLE. FIG. 113. HIGH DUTY FLEXIBLE ROLLER BEARING MOUNTING ON BEVEL GEAR REAR AXLE WITH BEARING ON TUBE. ROLLER BEARINGS WITH FLEXIBLE ROLLERS 215 riG. 114. HIGH DUTY FLEXIBLE ROLLER BEARING MOUNTING FOR BEVEL GEAR REAR AXLE WITH SLEEVE AND NO THRUST BEARING, ^>//////^//////////////, Y/////////////////////A FIG. 115. HIGH DUTY FLEXIBLE ROLLER BEARING MOUNTING FOR INNER END OF BEVEL PINION SHAFT — MATING GEAR ON REAR AXLE. FIG. 116. HIGH-DUTY FLEXIBLE ROLLER BEARINGS IN SELECTIVE TRANSMISSION MOUNTED INDEPENDENTLY. 2l6 BEARINGS AND THEIR LUBRICATION Figs. io6 and 107 show a countershaft bearing and loose pulley mounting, and two bearing applied to a friction clutch. In this latter example the short high-duty bearing carries the load and the longer standard bearing is a steady- ing bearing at the end of the hub. Fig. 108 is a flour mill construction and is of interest as showing the device used to prevent the escape of oil. Fig. 109 is the design of an 8-in. bearing for a pig breaker. Because of the length — 22 in. — two sets of rollers are used. Fig. no shows the mount- ing of an 8-in. vertical spiral pump. Roller bearings are used for the guide bearings and ball bearings for the thrusts. FIG. 117. HIGH-DUTY FLEXIBLE ROLLER BEARINGS IN SELECTIVE TRANSMISSION ON REAR AXLE. IN AUTOMOBILE PRACTICE These bearings have had an extensive use in automobiles. The seven engravings, Figs. 1 1 1 to 117 inclusive, show a few typical arrangements. Figs. Ill and 112 are of standard bearings applied to bevel gear rear axles. Figs. 113 and 114 are of high duty bearings also applied to bevel gear rear axles. Fig. 115 shows the inner bearing of a bevel pinion with a ball thrust, and Figs. 116 and 117 show two mountings for selective transmissions. SECTION VI Radial Roller Bearings with Solid Rollers ♦ Solid roller journal bearings divide into two classes, the first using cylindrical rollers, and the second conical rollers. Both have their extensive applications In general, the cylindrical roller is long and small in diameter compared with its length, while the conical roller is usually short and with a lesser difference between its length and diameter. The cylindrical roller is used on machinery in general, while the conical roller has had its special extensive application in automobile practice A newer development in the use of the cylindrical roller is for the bearings of street railway car axles. Figs. ii8 and 119 show typical designs for cylindrical roller journal bearings, as made by the Standard Roller Bearing Company. The first is a FIG. 118. ROLLER JOURNAL PEDESTAL BOX. Oil Drain. Note:- Space Beariiigs as far apart as possible aud adjust Collars to allow desired Amount of End Play of Shaft. FIG. 119. ROLLER JOURNAL BEARING FOR ELECTRIC MOTOR. design for a pedestal box. It will be noted that the rollers are comparatively small in diameter, are long, have a ball at each end to reduce the friction from thrust pressure, are held together in a cage and the bearing shell is fitted into the housing by a spherical seat to permit of self alinement. Another feature worth commenting upon is the felt rings at the end to exclude dirt and keep in lubricant. Fig. 119 shows a motor bearing having two rows of rollers. This bearing is distinguished from the one that precedes it by the fact that the rollers are shorter, larger in diameter, and that the ends have angular faces to take up a certain amount of end thrust. These oblique faces fit between corresponding 217 2l8 BEARINGS AND THEIR LUBRICATION shoulders, so that the motor or armature shaft can float endwise a short distance and be checked by the meeting of these inclined surfaces. It is evident that no great amount of end thrust can be successfully carried. The ends of the faces are provided with a felt washer and grooves to check the outward flow of oil and the entrance of dust. FIG. 1 20. ROLLER BEARING FOR ELECTRIC CAR TRUCK AXLE. Fig. 120 shows a roller bearing designed for the axles of electric cars, this particular one being for a Halsey truck. The journal is 4 1/4 in. in diameter and 9 in. long. It is covered by a steel sleeve, upon which the rollers bear. The details of design, including provision to exclude dust and provide means for introducing lubricant and storage place for the lubricant are plainly shown. FIG. 121. ANOTHER FORM OF ROLLER BEARING FOR ELECTRIC CAR TRUCK AXLE. At the end, in a vertical position, is a small plain roller thrust bearing to take the end thrust of the axle. Another design of a roller bearing for this purpose is shown by Fig. 121, in which there are two sets of rollers instead of one, otherwise the general details of the bearing design are not different, except that this has a large pocket at the center for oil, with a draining plug at the bottom. This design makes RADIAL ROLLER BEARINGS WITH SOLID ROLLERS 219 the roller bearing interchangeable with plain brass lined bearings and applicable to the standard size of thrusts in which the latter are customarily used. Bearings of this kind have been used experimentally to quite an extent in this country and as the experimental stage can be assumed to have been passed, their adoption by railways may be looked for as a development of the next few years. FORMULA FOR CAPACITY FOR CYLINDRICAL JOURNAL ROLLER BEARINGS The formula for capacity for solid roller journal bearings used by the Standard Roller Bearing Company is as follows: d'^nl 130,000 = capacity in pounds. Here d = the diameter of the rollers in inches, n = the number of rollers, Z = the length of each roller in inches, s = the circumferential speed of each roller in feet per minute. FIG. 122. CONICAL ROLLER JOURNAL BEARINGS APPLIED TO AUTOMOBILE HUBS. FIG. 123. CONICAL ROLLER BEAR- ING AS DIFFERENTIAL BEARING ON AN AUTOMOBILE REAR AXLE. CONICAL ROLLER JOURNAL BEARINGS Conical roller journal bearings have had their most extensive use in auto- mobile practice. Figs. 122, 123 and 124 are from designs of the Standard Roller Bearing Company. Fig. 122 showing two constructions for automo- bile hubs having two single rows of rollers. Fig. 123 shows a bearing of this type, used as a differential bearing on a rear axle. Fig. 124 shows in section a double taper roller bearing. 220 BEARINGS AND THEIR LUBRICATION FORMULA FOR CAPACITIES FOR TAPERED OR CONICAL JOURNAL BEARINGS The load carrying capacity in pounds is given by the following formula, which is from the practice of the Standard Roller Bearing Company: 130,000 = capacity in pounds. In which d = the mid-diameter of the rollers in inches, n = the number of rollers, I = the contact length of a roller with its bearing washer in inches, 5 = the mean circumferential speed of each roller in feet per minute. FIG. 124. DOUBLE CONICAL JOURNAL BEARING. SECTION VII Roller Thrust Bearings The Standard Roller Bearing Company is the builder of a patented type of plain roller thrust bearings that has been adapted to a wide field including situations where very heavy loads are to be sustained. It has had an extensive application for water turbine generators; in fact, its first use for heavy loads was in connection with the turbines at Niagara. The design consists essentially of two flat plates or washers, between which is a bronze cage containing a number of cylindrical rollers. These rollers are PLAIN ROLLER STEP BEARING APPLIED TO A CHEMICAL FURNACE. comparatively short for their diameters, have spherical ends and are fitted into radial slots in the cage. The outermost roller of each row is recessed to take a ball, which is held in a hole in a steel band surrounding the cage. A second band then surrounds the first and keeps the balls in position. These balls are used to reduce the friction due to end thrust from centrifugal force. The washers and cages are made either split or solid according to conditions. 221 222 BEARINGS AND THEIR LUBRICATION It is evident that the motion of the roller is a combination of rolling and sliding, for cylindrical rollers are traveling in a circular pitch. These bearings are made for all kinds of load and service, from a small bear- ing adapted to a 7/8 in. shaft up to a single bearing supporting a load of over 2,000,000 lb. Figs. 125, 126 and 127 show three of these bearings. The first is for a 6 in. shaft and was designed for a revolving chemical furnace. Figs. 126 and 127 are of larger bearings, the latter being for the turbo-generators at McCall's FIG. 126. PLAIN ROLLER THRUST BEARING FOR NINE-INCH SHAFT, THROUGH BEARING PROPER. WITH SHAFT PASSING Ferry on the Susquehanna River. Five bearings of this design have been installed there on the main turbine generators, three sustaining a load each of 330,000 lb., and two a load each of 410,000 lb. The size of the bearing can be realized when the shaft diameter is seen to be 21 in. The thrust is transmitted to the bearing through a collar 45 1/2 in. in di- ameter forged integral with the shaft. In every bearing two kinds of rollers are used — short and long. These are so arranged that they do not track, but cover the working surfaces of the washers so that the latter are stressed uniformly. To insure perfect alinement of the ROLLER THRUST BEARINGS 223 two washers it is customary to make the bottom of the lower one hemispherical in shape and seated correspondingly or to use a pair of leveling plates. Lubrication is an important point. In the larger bearings flooded lubrica- tion is maintained. The smaller sixes run in a bath of oil. . Drain Piping Arrangemeut. FIG. 127. PLAIN ROLLER THRUST BEARING — TYPE INSTALLED AT MCCALL'S FERRY. FORMULA FOR CAPACITY The formula for capacity of these bearings, based finally on a careful analysis of a large number of bearings in successful use, is dependent upon the roller diameter, the mean circumferential speed of the roller and the sum of their lengths. This formula is: 100,000— —- = load capacity in pounds. 2 24 BEARINGS AND THEIR LUBRICATION n which D = the diameter of the rollers in inches; M = the sum of the lengths of all the rollers in inches; 5 = the mean circumferential speed of the rollers in feet per minute. The value 100,000 of the constant factor is a minimum and thereby gives a minimum value for the capacity. Owing to advantageous conditions, this numerical factor may be considerably larger in some designs. Experience has shown that good practice sets a maximum limit for roller speeds of about 3000 r. p. m. Similarly the maximum circumferential speed of the roller is fixed at about 3000 ft. per minute. The factors of a uniform or variable speed and a uniform or variable load, enter to modify the given numerical constant. TEST OF BEARINGS AT McCALL'S FERRY An extensive series of tests has been run on the bearings at McCall's Ferry to determine the friction loss. Valuable data have been obtained. A word of explanation in regard to this installation is necessary. At the time of testing five units were in operation, three of 7500 and two of 10,000 kilovolt amperes rating. Each turbine has a rating of 13,500 h. p., with a head of 53 ft. and consists of two Francis wheels, located in tandem on the same shaft, the lower one being below the tail race level. The generators are three-phase, 25-cycle, 11,000 volts at a speed of 94 r. p. m. The first unit at the time of this writing has been in operation for over one year. Each bearing is lubricated by flood lubrication, circulation being maintained by gravity with the supply taken from an elevated tank. The overflow from each bearing is to a filter, from whence it is pumped into the storage tank. The filters are provided with cooling coils for circulating water, but these are seldom used except in the very hottest summer weather. Oil enters the bearing through the lower leveling washer into the space outside of the vertical oil guard. Thence it passes upward and by means of centrifugal action of the rollers passes outward through the openings in the cage and between the rollers themselves, and overflows at a level about midway of the height of the upper bearing washer. In case the oil supply should be cut off for any reason whatever, circulation will still be automatically maintained by the pumping action of the rollers and through oil grooves in the lower leveling washer. Each turbo-generator has three babbitted guide bearings and two lignum vitae guide bearings, in addition to the roller thrust. The method of test was to measure accurately the rise in temperature of the oil as it passed out of the bearing, to determine carefully the amount of oil in the bearing, to find the unit heat capacity of the oil, and to determine the heat loss ROLLER THRUST BEARINGS 225 226 BEARINGS AND THEIR LUBRICATION from the bearing per unit of time, due to radiation, dissipation and convection of heat. A proper handling of these data gave the amount of heat liberated in each bearing during each test. Thus it involved the friction of the rollers rolling and sliding on their bearing plates, the friction of the rollers between themselves and between the band; in fact, all of the frictional losses in the thrust bearing itself. Any losses taking place in any other bearing in the machine in no wise affected the observations or results. Fig. 128 shows the graphic log of a test of one of the more heavily loaded bearings. It will be seen that the speed was varied in six steps, that the tempera- ture of the entering oil and its quantity were kept practically uniform. The temperature of the issuing oil is seen to rise with the various steps of the increas- S.9 I r/ / A / / / / / ener Odd itorC "apac nrinn ity = l =4.in 0,000 100 Ih K.V.A. / Normal Speed = 94 f\ Friction Loss in Perce "ev.perMin. ntageof ^afed Generator Out}. 7ut=C .09% FIG. 129. 10 20 .30 40 50 60 70 80 90 100 Revolutions per Minute HORSEPOWER LOSS AT VARIOUS SPEEDS IN A LARGE PLAIN ROLLER BEARING. SEE FIG. 128. ing speed. At the extreme right is the temperature drop per unit of time after the turbine was shut down At the left hand end are data giving the specific gravity and specific heat of the lubricating oil. Fig. 129 shows the horsepower loss curve, corresponding with the test, graphically shown in Fig. 128. It will be noticed that at normal speed the total loss is slightly greater than 12 h. p. Similarly Fig. 130 shows the power loss curve for the same kind and size of bearing, but with a load of 330,000 lb., or 80,000 lb. less than for Fig. 129. Here will be noticed that the horsepower loss at normal speed is slightly greater than lo h. p. Fig. 131 similarly shows a horsepower loss curve for a much smaller bearing, a bearing used on an exciter. Here the load is 22,300 pounds. A consideration of these curves shows that in every case the curve is a straight line, indicating that the loss is directly proportional to the speed. Further a comparison of Figs. 129 and 130 shows that the loss is roughly pro- portional to the load. ROLLER THRUST BEARINGS 227 The loss for the two larger bearings is approximately o. i per cent, of the rated generator output. It is only fair to state that the bearing tested for the exciter set was originally y ' / / / / / / ' / / / c 7ener afor Capacity- 7500K.V.A. / ■Normal Speed =^94 Rev.perMin. / 7fGe aerator Output =O.IC z / 10 20 30 40 50 60 70 80 Revolutions per Minute 90 100 FIG. 130. HORSEPOWER LOSS AT VARIOUS SPEEDS IN A LARGE PLAIN ROLLER BEARING SUPPORTING A LOAD OF 330,000 POUNDS. designed for a load of 40,000 lb., and thus when tested was working at only 56 per cent, load, therefore, under somewhat disadvantageous conditions. A calculation has been made to determine the coefficient of friction for the 3 X 7 — / \' y V 7 / g- / / 1' / Lo nera ad 0! tore 1 Ron opac rinn WOK nnih Y. / Nc Fr )rmal 5peed=240 Rev.perMin. iction Loss in Percentaqe^ / of Generator Output =-0.B3 Z 111111 FIG. I3T. 25 50 75 100 125 150 175 200 225 250 Revolutions per Minute HORSEPOWER LOSS FOR VARIOUS SPEEDS IN A PLAIN ROLLER BEARING APPLIED TO AN EXCITER SET. two larger bearings. The method used was as follows: The total load was divided by the number of rollers proportional to their length. This gave a unit load for each roller in the bearing. The distance from the mid-point of the 228 BEARINGS AND THEIR LUBRICATION length of each roller to th« center of the shaft was determined, and from this was calculated the moment for its load. These moments were then summed for the entire bearing, multiplied by the average angular velocity and equated to the known friction loss divided by/ the coefficient of friction. Solving this equation gave a value for the coefficient of combined rolling and sliding friction of 0.0012. CONICAL ROLLER THRUST BEARINGS Table 47 is taken from the American Machinist for February 13, 1908, page 239, and gives the proportions for conical roller thrust bearings as made by the Standard Machinery Company. The design of these bearings consists in making the roller cone apex at the center on the shaft axis and the angle of the cone does not exceed 6 to 7 degrees. If the angles are larger than these, the outward thrust in a radial direction is excessive and hence causes destruction of the outer ends of the rollers. Two collars form the rolling surfaces for the bearing, one being stationary and the other fastened to the shaft and revolving with it. The inner surfaces of the collars are made conical to correspond with the angle of the rollers. A ring fits tightly over the bearing and serves to hold the rollers in position. The collars, as well as the ring, are of high carbon tool steel, hardened and ground. The rollers are of medium carbon steel and spring tempered. Diameter of Number Area of bear Safe pressure on Safe pressure on shaft of rolls ing plate bear.ng, speed 50 rev. bearing, speed 100 rev. 2-1/16 to 2-1/4 30 10.137 19,000 9,500 3-1/ 1 6 to 3-1/4 30 20.862 40,000 20,000 4-1/16 to 4-1/4 30 35 -^^97 70,000 35,000 5-1/16 to 5-1/4 30 54-2 108 000 56,000 6-1/16 to 6-1/2 30 78.017 125 000 62 000 8-1/16 to 8-1/2 32 132.06 200,000 100 000 9-T/16 to g-1/2 32 162.98 3^0,000 150,000 Table 47. — ^Proportion or Mossberg Conical Thrust Bearings. INDEX Accuracy of ball bearings, 189 Acids in oils, 115 Adhesion of Swedish gages, 5 Air compressor practice in bearing pressures, 66 lubrication, 138 Ajax metal, 49 Alining devices for bearings, 90 Alkalies in oils, 115 Allan red metal, 47 Allowable bearing pressures, 65 Albys, composition of for bearings, 49, 50, 51 of copper, tin and zinc, 51 of lead, tin and antimony, 50 for metallic packing, 61 miscellaneous, 51 physical properties of, 53 soft, for bearings, 45 of tin, copper and antimony, 50 Analyses of bearing metals, 49, 50, 51, 54 Anchoring of babbitt lining, 88, 147 Animal oils, 11 1 Anti-attrition metal, 56 friction metal, 48, 56 Antimonial lead, 48, 58 Arc of contact, effect of reducing, 134 Atomizing of oil, how to prevent, 169 Automobile ball bearings, 201, 202 bearing metals, 60 engine bearing pressures, 75 engine bearing proportions, 87 engine bearing rubbing speeds, 80 roller bearings, 216, 219 Babbitting, invention of, 139 kinks, 167 Babbitts, 45, 48, 49, 53, 54, 55 Babbitt, anchoring of, 88, 147 Genuine, 47 German, 53, 54, 58 Souther, 53, 54 standard, 53, 54, 58 Ball-and-socket bearing, typical design of, 155 Ball bearings, accuracy of, 189 for automobiles, 201, 202 cages for, 181, 188 classification of, i cone type, 188 construction of, 180—194 double-row radial, 186 for electric motors, 199 formulas for, 177 general applications, 201 limits of accuracy for, 194 lubrication of, 204—207 for machine tools, 195 materials for, 191 mountings for, 195—203 proportions of, 191 radial, capacity of, 177 thrust, capacity of, 178 two-row radial, 185, 187 Ball race grooves, curvature of, 176 Balls, formula for unit load on, 174 Bath lubrication, coefficients of friction for, 33, 34, 35, 36 Bearing inventions, 139 metals, See Metals. pressures. See Pressures. Bearings with rolling contact, classification of, 2 sliding contact, classification of, 2 Body of oils, 112 Bores of bearings, permissible variations in, 90 Brasses, car, wear of, 30, 55 Breaking-down point of oil film, 25 Bronzes, Allan, 48 composition of, 48, 49, 50, 50, 51, 54 physical properties of, 53 wear of, 54, 55, 56 Cages far ball bearings, 181, li Camelia metal, 48 229 230 INDEX Capacity of conical roller, journal bearings, 220 conical roller thrust bearings, 228 cylindrical roller, journal bearings, 219 flexible roller bearings, 212 plain roller thrust bearings, 223 radial ball bearings, 177 rings to deliver oil, 97 roller, journal bearings, 219 thrust ball bearings, 178 Carbon bronze, 48 Car box metal, 49 ring-oiling design, 163 brass lining, 48 brasses, wear of, 55 truck bearings, 158 roller bearings, 218 Care of automobile ball bearings, 207 bearings, 166 Cast iron as bearing metal, 43 Castor oil, in Centrifugal oilers, 98 Characteristics of a good bearing metal, 41 Classification of bearings, i lubricating devices, 93 lubricants, in Clearances in journal bearings, 90 Coefficient of friction, definition of, 6 lateral, 39 Morin, 10, n, 12, 13, 14 Stribeck, 175 Coefficients of friction of ball bearings, 175 collar bearings, 37 flexible roller bearings, 212 journal bearings minimum, 31 leather on cast iron, 14 locomotive values, 40 metal on metal, 14 oils and greases, comparative, 113, 114 plain roller thrust bearings, 228 radial ball bearings, 183 roller bearings, 175 sliding surfaces, 40 step bearings, 38 thrust ball bearings, 189 Tov^er for journal bearings. 33, 34, 35, 36 unlubricated surfaces, Rennie, 10 various substances, 10 Coefficients of wood and metal, 12, 13 wood on wood, 11 Collar bearings, coefficients of friction of, 37 moment of friction of, 8 work of friction of, 8 Combination of metals for journals and bearings, 42 Comparison of lubricating methods, 20, 94 Cone-type ball bearings, 188 Conical roller journal bearings, 219 thrust bearings, 228 Construction of ball bearings, 180-195 journal bearings, 141— 165 Cooling bearings by water jacket, no Cornish bronze, 48 Curvature of ball race grooves, 176 Cylinder lubrication, quantity of oil for, 116 oil, specifications for, 130, 131 (dark) specification for, 125 (filtered) specification for, 126 Cyprus bronze, 53, 54, 56 Damascus bronze, 49 DeLaval steam turbine bearings, 156 Delta metal, 48 Demo bronze, 53, 54 Design of ball bearings, 174 journal bearings, 86—110 journal bearings cooled by forced lubrica- tion 109 without artificial cooling, 108 with water jacket for cooling, no knife-edge bearings, 136 oil grooves, loi oil rings, 98 sliding surfaces, 132 Designs, typical for journal bearings, 141— 166 Devices for anchoring babbitt, 89, 147 lubricating, 93 Diameter and length of bearings, ratio of, 86 Distribution of temperature in a bearing, 99 Dissipation of heat, 102 Double-row radial ball bearing, 186 Driller ball bearing mountings, 197 Dry surfaces, laws of friction for, 15 Ex B. metal, 49, 58 Electrical machinery practice in bearing pressures, 68 INDEX 231 Electric generator bearings, 141 Electric motor ball bearing mounting, 197 journal bearings, 141 roller bearings, 219 ". ngi e oil, specifications of, 124, 127, 128, 130 E: gines, reciprocating, type plan for forced lubrication, 154 steam, lubrication of cylinders of, 116 Fiber for bearings, 64 Filling metal, 56 Film of oil, breaking down point of, 25 pressure of, 25 suction of, 27 thickness of, 24 Filtration of oil, 120 Fits of bearings, 91, 92 Flat wearing surfaces, 133 Flexible roller bearings, 209 Floating sleeve steam turbine bearing, 152 Flooded lubrication for machine tools, 146 quantity of oil used, 115 steam turbine bearing arranged for, 153 Forced lubrication, 94 chart of, 107 design of bearings for, 109 point to introduce oil, 94 quantity of oil used, 115 type plan for reciprocating engines, 154 vertical steam turbine bearing ar- ranged for, 150 Formulas for moment of friction, 8 work of friction, 8 Friction, coefficients for journal bearings, 33. 34, 35, 36 of collar bearings, 37 of step bearings, 38 coefficient of, lateral, 39 definition of coefficient of, 6 formulas for moment of, 8 work of, 8 laws of, for dry surfaces, 15 for well lubricated surfaces, 19 lowest coefficients for journal bearings, 31 of lubricated surfaces, 15 of poorly lubricated surfaces, 16 Friction of radial ball bearings, 183 of rest, 6 rolling, 173 of sliding surfaces, 40 of thrust ball bearings, 189 of unlubricated surfaces, 9 of well lubricated surfaces, 17 Fractional loss in bearings, formula for, 106 resistance and load, relationship of, 21 Gas engine oil, specification for, 126, 127 practice in bearing pressures, 69—73 Generator, electric, ball bearings, 199 journal bearings, 141 Genuine babbitt, 47 German babbitt, 53, 54, 58 practice in bearing pressures, 82 railroad practice in use of bearing metals, 58 Graney metal, 49 Graphite bearing metal, 48 as lubricant, iii, 112 in bearings, 64 Grease and oil, comparison of friction of, 113, 114 Greases, tests for, 121 Grinder spindle ball bearing mounting, 195 196 bearings, 137, 142, 148 Grooves for oil, 100 in ball bearing races, curvature of, 176 Hard lead, 49 Harrington bronze, 49 Heat liberated, formulas for, 8 radiation of, 102 unit factors for, 103 Horizontal steam turbine bearings, 151, 156 Hot box, how to cure, 168 Hyatt roller bearings, 209 Influence of bearing metals on friction, 19 Inventions, three important ones, 139 Journal bearings, classification of, 2 clearances of, 90 coefficients of friction for, 33, 34, 35, 36 design of, 106 heat radiation of, 103 moment of friction of, 8 232 INDEX Journal bearings, tests of, 8i, 103, 104, 105, 107 wear of, 56 work of friction of, 8 design of, 86 diameters, permissible variations in, 90 oil grooves, loi Kingsbury thrust bearing, 159 Knife-edge bearings, 136 Lard oil, coefficients of friction for, 34, 35, 114 specifications for, 122 test for, 1 23 Lateral friction, coefficient of, 39 Laws of dry friction, 15 friction, Morin, 9 for well lubricated surfaces, 19 Length and diameter of bearings, ratio of. 86 Limits for bearings, 91, 92 journals, 91, 92 radial ball bearings, 194 Line shaft bearing, typical design of, 155 Lining metal, anchoring of, 89, 147 Load and f rictional resistancfe, relationship of, 21 carrying capacity. See Capacity. Locomotive eccentric strap, 161 rod stubs, 161, 162 slide valves, coefficient of friction for, 40 truck bearings, 158 Loss of power in bearings, 8 Lubricated surfaces, friction of, 15 Lubricants for ball bearings, 204—207 classification of, in Lubricating devices, classification of, 93 greases, classification of, in methods, comparison of, 20, 94 oils, classification of, in filtration of, 1 20 specifications for, 123 Lubrication, air, 138 forced, chart of, 107 of ball bearings, 204—207 of flexible roller bearings, 213 of sliding surf aces, 133 of steam engine cylinders, 116 of steam turbines, 120 theory of, 18 Machine oil, specifications for, 129 tool ball bearings, 195 bearings, 145 tool, temperature of bearings of, 84 Magnolia metal, 48, 57 Manganese bronze, 49, 53 Materials of ball bearings, 191 Mercury bearings, 64 Metal, anchoring of, 89, 147 on metal, coefficients of friction for, 14 Metal and wood, coefiicients of friction for, 12, 13 Metallic packing, 61 Metals, See Babbitts and Bronzes. Metals for automobile bearings, 60 bearings, analyses of, 48, 49, 50, 51, 54 characteristics of, 41 physical properties of, 53 selection of, 88 white, 45 combination of, for bearings and jour- nal, 42 composition of, 48, 49, 50, 51, 54 experiment with, for bearings, 44 influence of, on friction, 19 used in DeLaval turbine, 58 by Pennsylvania R. R. Co., 58 by U. S. Navy, 56, 57 by U. S. War Department, 56 variations in, for bearings, 63 wear of, in bearings, 54 Mineral grease, coefficients of friction for, 2>3 oil, coefficient of friction for, 33, 114 Moment of friction, 8 Morin's friction experiments, 9 laws of friction, 9 Motion, friction of, 6 Motor ball bearing mountings, 197, 198, 199 journal bearings, 141 Mountings of ball bearings, 195—203 Naval practice in bearing pressures, 66 white brass, 57 Navy bearing metals, 56 specifications for oils, 123 Oil, coefficients of friction for various kinds, 2>2>, 34, 35, 36 delivery capacity of rings, 97 INDEX 233 Oil, feed, rates of, for flooded lubrication, 115 forced lubrication, 107, 115 steam engine cylinders, 116 steam turbine steps, 120 film, breaking down point, 25 point of maximum pressure, 27 minimum pressure, 27 nearest approach, 27 pressure of, 25 suction of, 27 thickness of, film of, 24 filtration of, 120 and grease, comparison of friction of , 1 13, 114 grooves, 100 how to prevent atomizing of, 169 hole with left washer, 95 point to introduce in forced lubrication 94 quantities used, 107, 115, 116, 120 ring bearing, temperature rise of, 104, 105 rings, design of, 97 Oils, properties of, 112 specifications for, 122 Olive oil, coefficient of, friction for, 35, 114 Practice in use of bearing metals, 45, 56 Pressure and velocity, product of, 85 maximum point in film, 27 minimum point in film, 27 in oil film, 25 Pressures in automobile engine bearings, 75— 79 bearings, air compressor practice, 66 electrical machinery practice, 68 gas engine practice, 69, 74 German practice, 82 knife-edge bearings, 137 maximum safe for perfect film lubri- cation, 81 naval practice, 66 railroad practice, 65 rolling mill practice, 68 steam engine practice, 66 flexible roller bearings, 212 Product of pressure and velocity, 85 Proportions of automobile engine bearings 87 of journal bearings, 86 radial ball bearings, 191 Provisions for dissipating heat, 102 lubricating, 93 Pad lubrication, coefficients of friction for, 20, 36, 94 Parson's white brass, 54, 57 Pedestal bearings, typical design of, 141 Pennsylvania Railroad Company's bearing metals, 58 Permissible rise in temperature, 83 Phosphor bronze, 49, 53, 54, 56, 58, 60 Physical properties of bearing metals, 53 Pivot bearings, moment of friction of, 8 work of friction of, 8 step bearings, coefficient3 of friction of, 38 Plain roller thrust bearings, 221—228 test of, 224 Plastic bronze, 53, 54, 56 Plumbic bronze, 53, 54 Point to introduce oil with forced lubrication, 94 of nearest approach of bearing and jour- nal, 27 Power lost in friction, 8 Quantities of oil used for lubricating, 115 Race grooves in ball bearings, curvature of 176 Radial ball' bearings, cages for, 181 capacity of, 177 doubje-row, 186 friction of, 183 limits, 194 proportion of, 192 tolerances, 194 tests of, 184 two-row, 185, 187 types of, 180 Radiation of heat from bearings, 102 Railroad practice in bearing pressures, 65 use of bearing metals, 58, 59 Rape oil, coefficients of friction for, 36, 37 114 Ratio of bearing length to diameter, 86 Reciprocating engines, type plan for forced lubrication of, 154 234 INDEX Red brass, 6i metal, Allan, 48 Reducing arc of contact, effect of, 134 Rest, friction of, 6, 16 Ring oiling, 95 car box, 163 invention of, 140 Rise in temperature, permissible, 83 Roller bearings, classification of, 2 flexible rollers, 209-216 journal bearings, conical rollers, 219 cylindrical rollers, 217 plain thrust bearings, 221 thrust bearings with conical rollers, 228 Rollers, unit loads on, 174 Rolling friction, 173 mill, practice in bearing pressures, 68 Rubbing speed in automobile engine bear- ings, 80 relation between and rise in tempera ture, 104, 105, 107 with relation to pressure, 81 velocity, permissible, 82 Rules for bearing design, 106 Running fits for shafts, 91 Salgee metal, 48 Scale bearings, 136 Selection of bearing metal, 88 Seller's bearing, invention of, 140 Shonberg M. M. metal, 54, 57 Slack in bearing, how to take up, 166 Sleeve bearings, typical design of, 143 Sliding fits for shafts, 91 friction, theory of, 5 surfaces, design of, 132 lubrication of, 133 Solid lubricants, iii roller bearings, 217—228 Souther babbitt, 53, 54 Specifications for lard oil, 122 oils, U. S. Navy, 123 U. S. War Department, 124 Sperm oil, coefficients of friction for, 34 114 Spindle bearings for traverse grinder, 137 Standard babbitt, 53, 54, 58 Steam engine bearing pressures, 66 lubrication of cylinders of, 116 piston bearing metal, 47 Steam turbine bearings, 150 quantity of oil used in, 120 Step bearings, coefficients of friction for, 38 for vertical shaft, typical design of, 155 quantity of oil used in, 120 Suction of oil film, 27 Sulphur as lubricant, in, 168 Swedish gages, adhesion of, 5 Syphon lubrication, coefficients of friction for, 20, 37 Talc as lubricant, in, 168 Tallow oil, in Temperature distribution in a bearing, 99 rise of journal bearings, 104, 105, 107 permissible, 83 Temperatures in machine tool bearings, 84 Tender boxes, 159 ball bearings, 183, 189 flexible roller bearings, 212 journal bearings, 56, 81, 103, 104, 105, 107 Test of plain roller thrust bearings, 224 Testing machine bearings, 136 Tests for greases, 121 lard oil, 123 Textile machinery bearings, 164 Theory of lubrication, 18 Thickness of oil film, 24 Thrust ball bearings, cages for, 188 capacity of, 178 friction of, 189 bearings, conical roller, 228 Kingsbury, 159 plain roller, 221 steam turbine, 150 Tobin bronze, 49 Tolerances in journal bearings, 90 radial ball bearings, 194 Traverse grinder spindle bearings, 137 Turbines, steam, bearing metals of, 58 bearings of, 150 quantity of oil used in, 1 20 Two-row radial ball bearings, 185-187 Typical constructions of journal bearings 141-165 designs of journal bearings, 141-165 Unit pressure, formula for, 7 INDEX 235 Unlubricated friction, laws of, 15 surfaces, friction of, 9 Valves, coefficient of friction for locomotive, 40 Variations in bearing metals, 6^ Vegetable oilS; 1 1 1 Velocity and pressure, product of, 35 of rubbing, 82 in automobile engine bearings, 80 in relation to coefficient of friction, 22 Vertical steam turbine bearings, 150 Viscosity of oils, 112 War Department bearing metals, 57 specifications for oils, 1 24 Water as lubricant, 112 cooled steam turbine bearing, 151 jacket for cooling, no Wear of bronzes, 54, 55, 56 filling metal, 56 Wearing surfaces, fiat, design of, 133 Weighing machine bearings, 136 Whale oil as lubricant, in White brass, 60 metal, 48 metals for bearings, 45 Wood for bearings, 63 and metal, coefficients of friction for, 1 2, 13 on wood, coefficients of friction for, 11 Work of friction, 7, 8. THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO 50 CENTS ON THE FOURTH DAY AND TO $1.00 ON THE SEVENTH DAY OVERDUE. JUN5 "S* SSNOV'SOLf sgNeveoLU Book Slip-10m-8,'51(6813s4)458 ^^ 402235 Alford Bearings lubrication JUNS '5« H. iA% 119^3 and their TJ1061 JL53 402233 UNIVERSITY OF CALIFORNIA LIBRARY