LIBRARY OF CONGRESS, /It Shelf ..Hi^- UNITED STATES OF AMERICA. ""Mlly W- %.l. ' -Ai CAR LUBRICATION. W. E;%ALL, B.S., M.E. SECOND EDITION, REVISED. FIRST THOUSAND. -^s:,^i)^i:^^ ri,' NEW YORK : JOHN WILEY & SONS. London : CHAPMAN & HALL, Limited. T895. f^ Copyright, 189s, BY JOHN WILEY & SONS. BOBEPT DRUMMOND. ELECTROTYPER AJJP PRINTER, NEW YORIf, PREFACE. Some years ago the subject of car lubrication became one of much interest to the writer, and, when attempt- ing to acquaint himself with the laws influencing suc- cessful practice, he was surprised to find how very little information, either of a theoretical or practical nature, could be obtained. The accompanying pages are not presented as a solution of the question, or as contain- ing any important original research, but rather in the hope that some of the knotty problems which were then presented for solution may be made clearer. The laws of friction, as accepted until recently, have by later experiments been limited in their application, if not confined entirely, to solids in contact. It is at least certain that they will not apply in any way as the laws governing the friction of solids separated by a lubricant. Prof. Thurston's " Friction and Lost Work," and the experiments of Mr. Woodbury with those of Mr. Tower, are quoted, and several others to a limited extent — to whom, it is hoped, proper credit has been given. It is desired that these few words may stimulate further thought, and in that way result in a more satis- factory solution of the problem. The Author. Altoona Machine Shops, May, 1891, PREFACE TO SECOND EDITION. Since the first edition of " Car Lubrication " was pub- lished the subject has received some attention by the railroads, as will be noted by the " Report of the Committee on Lubrication of Cars," contained in the Proceedings of the Master Car-Builders for 1894, and the article of Dr. Dudley and Mr. Pease in the Febru- ary number of The Railroad and Engineering Journal, 1892, all of which confirms the conclusions that had been reached by the investigations of the author and set forth in the first edition. The subject, however, has not, by any means, received the consideration it deserves. Attention is drawn to the continual drift towards the softer metals, and it now looks as though a com- paratively hard white metal in shells of iron or hard brass would be finally adopted as the best bearing metal. In the second edition a number of typographical errors that accidentally appeared in the first edition have been corrected. The Author. May, 1895. CAR LUBRICATION. CHAPTER I. INTRODUCTION. In taking up the study of the subject of car lubrica- tion we shall find it necessary to consider the features in the construction of the car-box and the proportion of the different parts, in addition to the care that should be given to lubrication, so that the journal will be properly supplied with oil and kept in a well-lubri- cated condition. While there is no reason why it should be so, yet it is a fact that there is probably no other branch of railroad engineering that has been so dependent upon empirical laws and where the prac- tice has been at such variance with the heretofore ac- cepted laws governing it. The widely taught law that friction is independent of the extent of surface in contact, but varies only with the pressure, is about ready to be placed among the archives of ancient scientists. The pressure inferred in this relationship is that exerted over the whole surface, and not per square inch — that is, a surface one square foot in area under a pressure of one pound per square inch would require the same force to move it over a resisting surface as it would if made one square inch in area under a pressure of 144 pounds per square inch. I 2 CAI? LUBRICATION. The extent of surface in contact was supposed to have no effect upon the force or work of friction necessary to move one body upon another, and consequently required no increased effort to produce motion, pro- vided the same total pressure was exerted although the area of the surfaces in contact might be at variance. Recent experiments made with various grades of lubricants, to determine the coefficient of friction of lubricated surfaces under varying conditions, prove conclusively that the amount of surface in contact materially influences the work of friction. If the rela- tionship of the/' resistance of friction as independent of the area of surfaces in contact, but dependent upon the pressure," were true, the temptation would be to reduce the work of friction and the abrasion of the materials by increasing the area of the surfaces in con- tact, which would allow the use of a lighter oil by re- ducing the pressure per square inch without increasing the abrasion. Practical demonstration, however, has proven the necessity of avoiding long journals, and, with the friction of rotation, an increase in the diameter of the journal means a corresponding increase in the work of friction. While recent investigations have not by any means furnished data that enable the subject of car lubrica- tion to be reduced to the desired degree of efficiency, they give information that is of much value for guid- ance in the design, construction, and management of lubricated surfaces, and results that will be found to accord quite closely with those obtained from practice. They indicate, and quite conclusively, that when the rubbing surfaces are kept well separated by the lubri- IN TROD UCTION. 3 cant the friction is more dependent upon the nature and fluidity of the lubricant than upon the nature of the sohds carrying the load. There seems to be a combined friction consisting of that inherent in the particles forming the lubricant and of the moving surface in contact with it. With con- stant pressure and temperature, it is dependent upon the extent of surface in contact and varies directly with it. It is also influenced by the unit pressure, and varies in some ratio with the change in the load, but not in the same ratio as had been previously supposed. As the resistance of lubricated surfaces is made up of the resistance of the particles of the lubricant, it is evident that any influence that will change the fluidity or density of the lubricant will also affect the frictional resistance. Increase of temperature, increasing the fluidity, causes a decrease in the coefficient of friction ; while increase in unit pressure causes an increase in the den- sity of the fluid, and, necessarily, an increase in the friction when motion is produced. The ideal condition of lubrication is attained when the viscosity of the lubricant at the working tem- perature is sufficient, and no more, to keep the surfaces of the solids apart under the maximum pressure they may have to sustain. We should always bear in mind that frictional resist- ance and the abrasion of the surfaces represents an ex- penditure of money, and, in the aggregate, is of greater moment than is generally supposed. The subject will be treated in the following chapters by taking up the different parts in detail and then con- 4 CAR LUBRICATION. sidering the relationship that must exist to give the most economical results. It will be considered under the two following general heads* 1st. The proportions and materials that are required to meet the demands of the service. 2d. The most economical way in which these may be applied. THEORETICAL RELATIONS. CHAPTER II. THEORETICAL RELATIONS. The resistance of friction in car lubrication is that which is generally known as "sliding friction of rota- tion." It is similar to linear motion, but, as it is an arc of contact, it differs in the distribution of the load per unit of surface. When the bearing is first placed upon the journal the arc of contact is small, and it is only after wear has taken place that the whole arc included within the bearing is in contact with the journal. The amount of wear which is necessary to produce this condition is comparatively small unless the radius of the bearing is made much larger than that of the journal, which must be classed as bad practice. It will be found that the larger part of the pressure is taken at the top of the journal, and decreases in a determinable ratio from that point to the horizontal axis through the centre of the journal. The work of friction is then but a question of the space which is passed over against the frictional resistance which is offered to the rotation. The space in this case is a function of the circumference, and varies as the diam- eter of the journal. The law of the distribution of the pressure is as fol- lows ; • CAE LUBRICATION. Let (see Fig. i) P z=. vertical pressure on unit surface ; P^ =. pressure in a radial direction on unit surface ; R = radius of the journal ; GO = angle made by the radius from a point O with a vertical line through the centre of the journal ; /= length of bearing; L = total load carried. For any point, O, P. = P cos a). The pressure upon a surface Rdo) is PJRdao. THEORETICAL RELATIONS. 7 The summation of the pressure should be for an equal arc on each side of the vertical, or L= r PJRdoo ; and by inserting the value of PJ ^— j PIR cos GO doD, From this, the load carried by various subtended arcs can readily be obtained. The accompanying table in- dicates the relative loads for several values of oo. Value of (0 Average pres- sure carried per square inch. Square inches of surface when radius equal unity. Percentage car- ried by the first 10° of the arc. IO° 20° 30° 40° 50° 0.34729 0.68404 I. 00000 1.28558 1.53209 0.1745 0,3490 0.5235 0.6980 0.8725 100.00 50.78 34-73 27.01 22.67 It is then evident that a higher percentage of the pressure is taken by a small arc of the journal, and that the lower surface of the arc of contact is the least im- portant part of the distributing surface. It will be noticed, too, that as regards the distribution of the load there is no serious objection to giving the bearings a surface contact that is less than the width of the bearing, for instance a width of bearing surface of three (3) inches on a journal of four inches diameter. It also follows that the practice of boring out the bear- ing to a greater radius than the journal is open to no serious objection, but, on the other hand, must be done to prevent the bearing seizing the journal, providing CAR LUBRICATION. the difference in the radii is not made too great. When this difference is excessive a condition is ob- tained similar to that shown in Fig. 2, which will result in very poor lubrication and likely produce a heated journal, the sharp corners at A having a tendency to scrape the oil from the journal. This effect will be made more apparent further on. In practice it is found that the best results are obtained when the radius to Fig. 2. which the bearing is bored is about one thirty-second (sV) ^^ ^" hyoh greater than the radius of the journal. This gives a safe working margin for a new bearing and journal, and yet does not give too great a difference or opening at the sides of the bearing when a new one is placed on a journal that is worn to the minimum di- ameter. In cases where journals of different diameters^ are used bearings corresponding to these different sizes should be kept on hand. In any case the possibility of any heating of the journal on this score can be obvi- ated by using the lead-lined bearings. THEORETICAL RELATIONS. 9 Before stating the elements determining the work of friction, it is necessary to review in a general way the results of recent experiments which were made to de- termine the laws governing the resistance to motion of bodies when separated by a lubricant. Reference is made particularly to the experiments by Woodbury and Tower, the results from both of which, while under different conditions, are corroborative. Those by Woodbury were made with low pressures, and the curves obtained from his results give a relationship of pressure and coefficient of friction, as shown in the diagram, marked as Fig. 3. They show conclusively that frictional resistance, with an intervening lubricant, is not a direct ratio factor of the pressure, as it is with unlubricated surfaces ; but, on the contrary, the laws seem to follow those of fluid friction more closely than those of solids. The result produced by the mo- tion of two solids underpressure is a more or less rapid abrasion of the metals in contact. With a lubricant interposed the conditions are quite changed, and follow more closely the resistance which the fluid would offer by its own friction. Whether or no this be a motion of the particles of the lubricant or of the solid upon the surface of the fluid does not concern us. The impor- tant consideration is the extent of surface in contact which should enter as an element in the calculation of the work done. The best way for our purpose is to ob- tain the frictional resistance of the lubricant at the pressure carried, reduced to the resistance under these conditions for a unit surface. The resistance of friction would then consist of two elements : the coefficient of friction per unit of surface for the working pressure 10 CAR LUBRICATION. and temperature, and the number of units of surface in contact. Representing these by/" and a respectively, would give as the relationship of the total work of fric- tion W = f.a.p . 27tR X n, where n = number of revolutions per unit of time ; /f = pressure per square inch. With given pressure and temperature the minimum value of the function/", a is determinable. First, the pressure on the journal and the tempera- ture attained under the maximum speed will indicate the density of the lubricant which it is necessary to use to prevent the surfaces coming in contact, and in the case of car lubrication will vary with the seasons of the year, causing a grading of the oils into those for sum- mer, and lighter ones for winter use. Second, the value of a will depend npon the available space allowed for the journal. It will be seen further on that the results of the experiments indicate that the most economical conditions are obtained by increasing the area, within practical limits, and using a correspond- ingly lighter body oil for the lubricant. The value of a for the most economical results is where any further decrease in the resistance by the use of a more fluid oil is counteracted by the increased resistance resulting from the larger surface in contact. More explicitly the experimental results indicate that, within practical limits, a lubricant of greater fluidity and correspondingly lower coefficient of friction can be used by increasing the area, and in that way reducing the unit pressure until the limits are exceeded, when THEORETICAL RELATIONS. II any further increase of the contact surfaces produces a reverse effect. It will be remembered we found that the additional surface obtained by increasing the arc of contact does not produce a proportionate decrease in the pressure per unit of surface, so that the oil that is selected for the purpose must depend upon the pres- sure that exists at the top of the bearing. Practically, an arc of contact of some magnitude is necessary for strength and stability, and to give a fairly large area to accommodate for the abrasion which also takes place. Any increase beyond this arc is economical only so long as the increased surface decreases the pressure carried at the centre of the arc to an extent that the lighter oil will, by the consequent reduction in the co- efficient of friction, overbalance the increased resistance produced by a greater area. Assuming c as the con- stant and necessary arc of contact and w as the desired angle, it is not economical to increase go when the ex- pression is greater than /—- X 6.282. 360 /'4-x 6.282. 360 /and/' indicate the coefficients of friction per unit of area for the lubricant which must be used to overcome the maximum pressure existing in the two cases which, in one sense, measures the fluidity of the lubricant. This is on the basis of the diameter and length of the journal remaining constant. An increase in the length of the journal seems to be advantageous, provided it is not carried beyond the 12 CAR LUBRICATION. limits placed upon it by practice. The diameter of the journal is dependent upon its length and the load to be carried. Taking the usual expression for a beam sup- ported at one end and uniformly loaded, we have _ TTtR' U where T ^= safe ultimate load for the metal, and the other symbols indicate the same as in previous formulae. For the deflection we have 27tR'' where d represents the deflection. All are indicated in pounds and inches. The formula for the variation of the diameter for changes in the length will be used again. COEFFICIENT OF FRICTION: 13 CHAPTER III. COEFFICIENT OF FRICTION. With the exception of the method of lubrication, there is no other element in connection with the sub- ject under consideration that has received more atten- tion than that of the coefificient of friction, and yet there is no other that is in as crude and indeterminable a state. As investigation progresses, the subject seems surrounded with more and more variables of a com- plicated nature, which indicate the importance, if not necessity, of the utmost care when the best results from lubrication are desired. The latest study of the subject has brought out some very interesting results, and has conclusively shown that it is now necessary to at least limit the old laws of friction to dry surfaces in contact, if not exclude them totally. The resistance of friction, when a medium is introduced between the so-called rubbing surfaces, fol- lows laws quite different and more intricate than those determined by Morin, which were to the effect that " friction was independent of the surface in contact, but directly dependent upon the pressure keeping the surfaces together." Where friction is produced it is important to distin guish between the two conditions to which the two sets of laws apply ; in one it is a solid against a solid, the particles of each interlapping and causing resistance by 14 CAR LUBRICATION'. the efforts of the particles of one metal to tear away those of the other. Where lubrication is introduced it is intended that the two solids shall be separated by a film of the lubricant, generally a liquid. In this latter case the resistance assumes the nature of the laws of fluids, and consists of the friction of the particles of the lubricant and that of the solid against the fluid, form- ing a combined resistance, the percentage of each to the whole retardation depending upon the nature of the lubricant and the metal surfaces. As long as the metals are prevented by the lubricant from coming in contact, it is found that the friction is dependent upon the fluidity of the lubricant, and varies with changes of this fluid condition, decreasing with a higher tempera- ture and increasing with a less degree of heat. We will assume, first, that the lubricant prevents any contact of the metal surfaces. The condition then stands between the laws of solid friction on the one hand — that is, independent of the surfaces in contact, but dependent upon the total pressure, — and the laws of friction of liquids on the other, where it is inde- pendent of the pressure per unit of surface, but is directly dependent upon the area and increases as the square of the velocity. From most recent investigation this intermediate condition has been found to be, when stated in a general way, that the coefficient of friction decreases with an increase of the pressure, although the total resistance rises directly but not proportionately with the higher unit pressures and increases with the velocity, although not as rapidly as its square. It is also found to be dependent upon the extent of surface in contact. An exact relation between these varying Coefficient of FRictioht. i^ conditions has not yet been obtained, evidently because they vary so materially with any slight variation in the method used of lubricating the surfaces. As, for in- stance, when the oil-bath is used the laws of lubricated surfaces, especially as regards surface and pressure, fol- low those of liquid friction very closely ; while with less efficient means of lubrication the results show a condition between solid and liquid friction. This mat- ter will be brought out more prominently in the chap- ter on the methods of lubrication. With the surfaces in good condition and the oil-bath method of supplying the oil, which may be considered as practically perfect lubrication, it was found that the mean resistance per square inch of surface, with pres- sures varying from lOO to 310 pounds per square inch, was as follows : Lubricant. Mean resistance in pounds. Sperm-oil 0.484 Rape-oil 0.5^2 Mineral-oil... 0.623 Lard-oil 0.652 Olive-oil 0.654 Mineral grease 1.048 [Results obtained by Tower. — See Engineering for November 16, 1883, and Feb. 6, 1885.] The speed was 300 revolutions per minute, and jour- nal four inches diameter and six inches long, while the temperature was maintained at 90° Fahrenheit. A constant temperature is essential for a proper com- parison, for in one case the coefficient of friction of lard oil decreased to one third (^) its value at 60° by increas- ing the temperature of the lubricant to 120° Fahren- heit. l6 CAR LUBRICATION. Probably the most accurate laboratory experiments that have been conducted for the determination of the resistance of lubricating oils were those made by Woodbury for the North-Eastern Cotton Manufac- turers' Association, as published in their proceedings of April 28, 1880, and in the proceedings of the American Society of Mechanical Engineers, as contained in volume VI. They were made with the object of ap- propriating the results to cotton and woollen machinery where low pressures are used, and to that extent are not well adapted to the lubrication of car journals ex- cepting as showing the action of lubricants under vary- ing conditions of temperature and a limited range of pressure. They were presented about the same time as the results of Mr. Tower's experiments, the latter, however, under heavier pressures, but both clearly showing the different conditions under which friction must be studied when solid surfaces are lubricated by such bodies as the mineral and animal oils. It was found in the tests that uniform results could not be looked for unless constant temperature, velocity, press- ure, area of surface in contact, and thickness of the film of the oil between the surfaces were maintained, the latter depending somewhat upon the method of lu- brication, the results indicating clearly that the resist- ance of friction was dependent upon and changed with a variation in any of the above conditions. The metallic surfaces were cast iron and bronze, the latter composed of copper 32, tin 2, lead 2, and zinc I. In one case all conditions were kept constant excepting that of pressure, the diagram represented as Fig. 3 in- dicating a decrease in the coefficient of friction, with COEFFICIENT OF FRICTION. 17 increased pressure per unit of surface, but an increase of total resistance. Similar tests were made, keeping the pressure constant and varying the temperature, which gave the eJEfect of the variation of the tem- perature upon the coefficient of friction. The two diagrams, Figs. 3 and 4, indicate by the two curves the variation which takes place in the coefficient of friction under the conditions named. In Fig. 4 it will be noticed that the variation in the 1 Coefficient 130 -110 poo 1 90 ig 80 70 \ \ \ \ Te of nperat Frictio andC i.varyi oefficii at ProBaviie and \ \" pom lonatau da per t=2a iquare ad 6 nch 1 \ \ .ao \ Is OfF mperal oie.coi slant =100 F. g'4 -\ I \ \ 1 \ \ \ .s2 £ r \ \ V \ ■ \ i 10 . c JO . oef. of 30 .'■ Friction 10 . If 50 ( .1 c :i oot-of U .4 Frictio .b 1. Fig. 3. Fig. 4. coefficient of friction due to changes in temperature follows closely the laws of the straight line, indicating a proportionate decrease with the increase in tempera- ture ; the angle of the line with the abscissae depending upon the pressure per square inch. Combining these two diagrams gives a curve, as shown in Fig. 5, for a coefficient of friction where the two elements, pressure and temperature, vary ; and it is this relationship which most concerns the lubrication of surfaces such as car journals, as in this case the tern- lo CAR LUBRICATION^ perature is subject to variations arising from changes of seasons and weather, while the pressure carried per square inch is dependent upon how long the bearing has been subjected to wear and attrition, the unit press- sure decreasing with increase of service. While the results obtained by Woodbury are the most accurate that have been published and probably ever made, both as regards design of apparatus as well as its manipu- lation, there still lacks sufficient uniformity for the derivation of a definite law as to the variation of fric- tion with changes in temperature and pressure. For instance, the decrease in the coefficient of friction for pressures of from one (i) to five (5) pounds per square inch is as follows: Pressure per square inch. Pounds. Coefficient of Friction. Decrease in Coefficient of Friction, Difference in the amount of decrease. I 2 3 4 5 0.3818 0.2686 O.2171 0.1849 0.1743 0.0000 O.II32 0.0515 0.0322 0.0106 0.0000 0.0000 0.0617 0.0193 0.0216 The decrease in the coefficient of friction from an in- crease of pressure of one (i) to two (2) pounds was 0.0617 more than that from two (2) to three (3) pounds, while the increase from three (3) to four (4) pounds was 0.0193, and nearly the same as took place when the pressure was increased from four (4) to five (5) pounds. The above is cited more to prevent a deduction of too wide a nature rather than to deter from the grati- tude which the engineering profession must feel for the derivation of the general law of the variation of COEFFICIENT OF FRICTION. 19 friction with lubricated surfaces when the temperature and pressure are varied. The care which it was neces- sary for Mr. Woodbury to exercise to obtain these results can be appreciated when it was found essential to run the apparatus, using gasoline or its equivalent, to clean an oil from the surfaces after testing. It re- quired a travel of one surface over the other equivalent 12 3 4 5 •Fxesaure per sc^uare-inoh. Fig. 5. Note. — The coefBcient of friction in the three cases is represented in actual pounds resistance. to about forty (40) miles before it was advisable to commence the trial of the succeeding oil, and even then indications could be noticed in the test following of the properties of the oil previously tried. This is con- sidered further evidence that the friction of lubricated surfaces is made up of the friction of the fluid, and tends to prove that the lubricant imbeds itself into the surface of the metal, producing a fluid resistance rather than a resistance due to the rubbing of the sur- face of the metal upon the surface of the lubricant. This effect will be taken up again in the chapter on Bearing Metals, 20 CAR LUBRICATION. The variation in the coefficient of friction with changes of temperature can readily be carried to an extreme, as it has been found that while the resistance decreases as the temperature is raised, there is a point, depending upon the unit pressure and viscosity of the lubricant, where the coefficient starts to increase very rapidly with increase of temperature. The same holds true with a variation in the pressure ; and while the laws of changes are true as stated, in a general way, they depend and are limited by the viscosity of the lubri- cant used, and also the pressure which it is necessary to carry. The rapid increase, when the limits of tem- perature and pressure are exceeded, is due to the solids coming in contact and causing increased friction by the abrasion of the surfaces, reducing the condition from friction of fluids to that of solids. An attempt has been made to prove a positive rela- tionship between the viscosity of an oil and its coeffi- cient of friction ; and while they are, no doubt, more or less dependent, there is hardly sufficient data at hand to resolve this to a definite basis. In addition to that of viscosity lubricants possess a property designated as unctuousness, which seems to influence the coefficient of friction as much, if not more, than the viscosity. It appears, and there seems sufficient information at hand to anticipate it, that a relationship of a positive and determinable nature between the three elements, co- efficient of friction, viscosity, and unction, is obtainable. The results of the experiments made by Mr. Tower, as presented before the British Institute of Mechanical Engineers, are of such a nature that they can readily be converted into practice with a resulting profitable C0EFFICIEN7' OF FRICTION. 21 application. They go to corroborate, to a close degree, the results of Mr. Woodbury, although they have the advantage of having been made with higher pressures. It should be remembered that with high pressures, such as those obtained before seizure takes place, the film of oil separating the bearing and journal has been found to be very thin, and shows the result that maybe expected where the irregularities or projections on the surface of the bearing are greater than the thickness of the film of oil used to separate them, producing when in motion a rapid and detrimental abrasion of the metals with a marked increase in the friction. It would not be safe to allow the irregularities to project from the surface more than the thickness of the film of oil, unless, of course, the prevailing area is of this height. The other extreme, of having the surfaces too highly polished, must also be avoided, as it has been found that a moderately rough machined surface will carry something like seven (7) times more pressure before seizing than can be obtained from highly polished sur- faces. The proper condition would seem to be about that produced by a boring tool, except with the soft metals, which, from their-nature, are incapable of tak- ing a high polish. For the purpose of comparison let us consider briefly some of the results obtained by Mr. Tower where the journal was lubricated by the oil-bath method and the surfaces were in good condition. Assuming a loaded car of total weight of 80,000 pounds would give 10.000 pounds per journal. At a pressure of 300 pounds per square inch this would require a bearing area of 33.33 square inches to carry the load. With 22 CAR LUBRICATION. the resistance given for mineral oil of 0.623 pounds per square inch on a journal 4 inches in diameter and a wheel 33 inches in diameter, a tractive resistance of 2.52 pounds per journal, or of 0.504 pounds per ton, would be required after motion had been produced. With the latest dynamometer readings the resistance of journal friction for loaded cars will probably reach as low a figure as 2 pounds per ton on level tangent when running at a speed of 15 miles per hour. Retracing, this figure gives a journal resistance of 82.5 pounds, and as high a figure as 2.48 pounds per square inch of bear- ing contact. It will be seen how low an efficiency is obtained in practice, but it should be remembered that the above laboratory tests were with the oil-bath method of lubrication, which has proven, when so tried, to be far superior to any other method that is at pres- ent known for lubricating surfaces. In most cases it has been found that the resistance of friction is a direct proportionate function of the area of surfaces in con- tact ; twice the bearing surface, all other conditions remaining the same, will give, approximately, twice the resistance from friction. The information as presented by the theoretical tests of oils is of much importance in the selection of the one best suited for the conditions of any particular service. In car lubrication, however, a constant rela- tionship between bearing surface, lubrication, tempera- ture, and pressure does not exist, and for that reason a large factor of safety must be used to cover the varia- tions and the extreme conditions that exist in practice. For instance, when starting with a new bearing, the surface in contact is much less than when it has wori) COEFFICIENT OF FRICTION. 23 down to a point where the whole arc of the bearing comes in contact with the journal. This is one of the conditions which must be met, for with the irregularities of the parts accompanying the distribution of the load it is found, excepting with the so-called soft-bearing metals, that heating will almost invariably result if the bearing is fitted to the journal throughout the whole arc which it is capable of including. There seems to be a binding action on the journal. If the bearing is so fitted as to allow a small amount of motion of the bearing on the journal, the wear will take place in a manner consistent with the alignment of the journal- box. The variation in the unit pressure is not, how- ever, as wide as would at first be supposed, as will be seen on reference to Chapter II, where the variation in unit pressure due to changes in arc of contact is given. The viscosity of the oil selected should, then, be such that it will keep the surfaces apart under the conditions of minimum arc of contact, and at the highest tem- perature that will be met. This temperature is not de- pendent altogether upon that of the atmosphere, but, on the contrary, will vary much with the nature of the service. For instance, with long continuous fast runs the temperature of the journal will be consider- ably above that of the atmosphere. With this service the heat arising from the work of friction will be such as to raise the temperature of the journal and bearing before a constant condition is reached, and the conduc- tivity and radiation of the heat through and from the surfaces is not sufficiently rapid to accommodate for all the heat generated, and keep the temperature down to that of the atmosphere. It will not be found un- 24 CAR LUBRICATION. common for journals in severe service to reach a tem- perature of a hundred and fifteen (115) degrees Fahr. with the atmosphere only 50 to 60 degrees. This results in better lubrication, provided an oil has been selected with sufificient body to meet the conditions. If such is not used, the parts are reduced to such a sen- sitive state that the slightest cutting from the entrance of foreign matter between the bearing surfaces is apt to result in an overheated journal, or what is generally known as a hot box. The. maximum presure per square inch that must be sustained without seizure at the highest temperature that will be reached will deter- mine the grade of the oil which it will be necessary to use. The resistance is a minimum when the product of the coefficient of friction and the area in contact is a minimum. When the limitations of the case require the use of high unit pressures, correspondingly heavier oils must be used to prevent the bearing seizing the journal ; but that oil, all other variables remaining the same, which will give the lowest coefficient of friction and prevent the surfaces coming in contact is the one to be used. The work done is dependent upon the circumference of the journal, so that any change in its diameter af- fects correspondingly the work of friction. The diam- eter of the journal varies closely as 1/ l^. The work of friction is dependent upon the coefficient of friction per unit of surface, the area in contact, and the distance travelled ; or, depends upon W=^7tafnWT,, Assuming a constant deflection, the variation in the COEFFICIJSJVT OF FRICTiON'. 2$ diameter is that necessary to maintain strength for the changes in the length. By referring to the table on page 15 it will be seen that we can determine the in- trinsic values of the heavy and light oils on the basis of the work of friction, resulting from the resistance which each offers to motion. For instance, Tower found that with rape-seed oil the pressure which it would resist up to the point of seizure was 573 pounds, and with min- eral oil a pressure of 625 pounds per square inch. To make a comparison between sperm oil, which is of still lighter body, and mineral oil we would have, assuming a proportionate power of resisting pressure, 541 pounds pressure as the capacity of the sperm oil. Taking this oil, and with a bearing surface of 1^ by 8 inches, we - , ., , , . f . , ., li X 8 625 find it would require for mineral oil = — ~ or X 540 10.4 square inches of surface to sustain the load. The corresponding length would be 6.9 against 8 inches with sperm. The expressions for the work of friction in the two cases would be W = Ttafn \/X', W, = na'f'n VJJ-, and their ratio fa' f // Sperm oil =/ = 0.484, a =. \2, /, = 8 ; Mineral oil =/' = 0.623, a' = 10.4, //= 6.9 ; 112.73 r = = o.Qii. 123.75 26 CAJ? LUBRlCATiOl^. While some of the figures are approximate only, they are sufficiently close to show that the heavy oils, even with the decrease of bearing surface which they allow are not as economical as the lighter ones, so that it may be stated in a general way that the work of fric- tion is least where the conditions are such that they will allow an increase in the length of journal, result- ing in an increase of the bearing surface under the same load, when an oil of lighter body may be used. This conclusion has, of course, its practical limits, and it would probably require modifTcation if experiments of a more detailed nature were made on the same lines as started by Mr. Tower, but would not, in all proba- bility, change the general result. It clearly shows the importance and value which must be attached to fur- ther investigation in this direction. BEARING METALS. 2 J CHAPTER IV. BEARING METALS. Oil and bearing metal, in their relation and applica- tion to car lubrication, are capable of almost unlimited treatment, so much so that it has now become the work of a specialist to properly follow each and advise as to their efficiency. The easy adulteration of oils make them a subject of suspicion and necessitate rigid speci- fications and inspection to eliminate the chance of such deterioration. This having been successfully accom- plished, through proper specification and rigid inspec- tion, the next point is the selection of the oil best suited for the journals to be lubricated. This requires the consideration of properties, in addition to the co- efficient of friction. Those referred to are the rate of evaporation at the working temperature, the tendency of spontaneous combustion from the evaporation, the decomposition of the oil by the atmosphere, and, still further, that of the chemical action of the acid, which animal and vegetable oils contain to a greater or less extent, on the metal used for the bearing. For instance, an oil, whose exposed surface gave an evaporation of 20 per cent would be far inferior to one which gave but 10 per cent evaporation at the same temperature and in the same time. It is also objec- tionable on the ground that the oil giving the higher 28 CAR LUBRICATION. evaporation would also be more liable to give trouble from combustion arising from the rapidly vaporized oil. Care must be taken not to conflict the flashing point with the rate of evaporation, as they are not in anyway related, for it has been found * that, in one case, two oils having the same flashing point gave the rate of evaporation of 9.4 and 24.6 per cent respectively. When determining the percentage of evaporation of oils, the surface, time, and temperature should be the same in all cases, as otherwise it would not be a true comparison. The mineral oils have a low evaporation, and when mixed with those of an animal and vegetable nature prevent, to a large extent, the spontaneous combustion which is apt to result and give trouble when the animal or vegetable oils are used alone. The chemical effect arising from exposure to the atmos- phere is of much importance in its influence upon the lubricating value. This action, with the fine particles of dust or foreign matter which enter through the front and back of the box, reduces the oil in the top of the waste to a pasty condition which materially depreciates it as a lubricant. It should be remembered that the result obtained by the use of an admixture of the mineral and animal oils is dependent upon their relative proportions and the temperature to which the mixture is subjected. The use of the petroleum products has a remarkable effect toward reducing the tendency to inflame, so common when the animal and vegetable oils are used alone. The relative value of oils on the basis * See Prof. Ordway, in Proceedings of Semi-annual Meeting of N. E. Cotton Manufacturers' Association, held in Boston Oct, 30, 1878. BEARING METALS. 29 of percentage of evaporation and inflammability is un- known ; in fact, it is as yet an undeveloped field, but represents properties which must sooner or later enter as factors in the efificiency of an oil for lunricating pur- poses. The importance of these will be appreciated from the results which Ordway found, where with one oil the evaporation in twelve hours, at a temperature 140 degrees (Fahr.), was 24.6 per cent. As well, too, should the question of the chemical ef- fect of the acids in the oil be taken into consideration. For these two reasons, lower evaporation and freedom from acid, the mineral oils are coming into general use for car-lubricating purposes, while they also give, from their lubricating qualities, as low a coefificient of fric- tion as any of the animal or vegetable oils. They can be obtained of almost any desired gravity and fire test,- and, when clean, are particularly well adapted to the service in question. The brass-foundry practice of to-day is still so much dependent upon empirical laws that it is impossible to reach any definite or concise conclusion as to the exact nature of the alloy, all things considered, which gives the best results for car-bearings. So much depends upon the foundry treatment that a chemical analysis is of very little value from which to draw any definite conclusions as to the nature of the service which a known mixture of metals will give. The same ingre- dients differently treated will give alloys of marked variation in their physical properties, and until the foundry working can be reduced to a more accurate science we must be subjected to the so-called " kinks " which have in some cases produced metals of remark- 30 CAR LUBRICATION. ably good wearing qualities. This is illustrated in phosphor bronze, where the metal as produced contains about 0.75 per cent of phosphorus, while about one (i) per cent is used during the treatment. The effect of the phosphorus is to produce a more solid casting by reducing the amount of oxidation which takes place during the mixing of the metals, but the presence of the phosphorus in the metal after melting has no effect upon the quality. The metals used for bearings may be classed about as follows : phosphor bronze, brass, and the so-called white metals, the latter containing a large percentage of lead, zinc, tin, or antimony, with but little or no copper. Each has a wide range of hardness, but from all that can now be gathered, the white metals give excellent service and wear less than the harder alloys. In 1883 the writer had an excellent opportunity to compare, in a general way, the service of a hard bearing with one composed of antimony and lead, — the latter material was run into an iron shell. The two roads using the metals were located in the same part of the country, started from the same place, and had the same destina- tion ; in fact, they paralleled each other for a large part of the distance, so that the service was as nearly alike as it is possible to obtain for a comparison. The bronze required, on an average, a consumption of oil of 0.945 of a pound, and the white metal an oil consumption of 0.3075 pound, each, per car per 100 miles. Where the bronze was used it was necessary to resort to lard oil, making the cost per car per 100 miles in the two cases BEARING METALS. 31 6.3 and 0.88 cents respectively, nothing but common black oil being used with the white metal. The white alloy was remarkably free from heating, while with the bronze bearings hot journals were giving continual annoyance. As regards the wear of the soft and hard metals, the experience with bearings lined with lead alone indicates the remarkably long service which can be obtained from even a lining but -^-^ of an inch thick. Experi- ments as given in the Railroad Gazette for March 5, 1886, are much in favor of the white metals. There is some difference of opinion as to the resist- ance of friction, as well as their effect upon axle wear, with the red brass or equally hard-bearing metal, and the so-called white metal, but, as far as is known, no experiments have been made giving results by which a comparison can be drawn to determine the efificiency of these two features of the metals. Research with this object would be of much value. As regards axle wear, however, it will be seen, by reference to the chap- ter on the cost of lubrication, that bearing metal and axle wear are almost equal in value for the loss result- ing from abrasion per 1000 miles. With soft and hard metals moving together the result is always a more rapid abrasion of the harder one. This is where the surfaces are separated by a grinding material ; but when properly lubricated the condition is quite different, as then the separating material is fluid and slow in its wearing action, and, from the experience of those who have the white metal in general use, the wear is not in- creased by the particles of dust which its opponents 32 CAR LUBRICAl^ION. claim become imbedded in the bearing, and in that way exercising an additional grinding action upon the jour- nal and increasing the wear over that produced by the harder metals. While there seems no reason to expect a more rapid wear of the journal from the softer metals, yet it is a subject seriously affecting the cost of lubrication, and one upon which there is lacking sufficient information to warrant the positive assertion that the softer metals are superior to the harder ones for bearings. Ex- perience so far, however, seems to be in favor of the white metals. This much can certainly be said that a large part of the cost of lubrication of cars where the harder bearing metals are used is due to the loss result- ing from heated journals, and the white or softer metals invariably give less trouble from this cause than any other alloy that has yet been made. There has been a tendency to attribute the metal con- tained in an oil that had been in service solely to the wearing of the bearing, but the experiments of Volney will show the relative action of different oils upon the decomposition of brass. The figures also represent the value of the oils in this respect, and, together with the oxidation which results from exposure to the atmos- phere, indicate the influence which these properties have in producing the pasty condition of the waste. The dissolving power should be considered in its rel- ative importance in the selection of the oil that is to be used. It also shows that the bearing metal found in the oil of a journal-box is not all due to abrasion. Attention is drawn to the low dissolving power of the crude petroleum oils, another property that should fa' BEARING METALS. 33 vor the use of these oils for car lubricating purposes. On the whole they will maintain a more uniform con- dition, and they have fewer detrimental properties than oils of either vegetable or animal origin. Name of Oil. Relative Dissolving Power. Menhaden oil 0.51 1 Neatsfoot " 0.505 Olive oil o. 504 Crude cotton-seed oil. .. ., 0.348 Lard oil 0.131 Crude petroleum from Scio 0.000 Note. — For general results of tests of bearing metals and oils made on the Paris-Lyons-Mediterranean Railway, see appendix containing translation from Revue Gdn&ale des Chemins de Fer. 34 CAJi lubrication: CHAPTER V. METHODS OF LUBRICATION. The devices that have been designed and the so- called inventions that have been patented to lubricate car journals are innumerable, and yet there is not one at present in use that can be said to be in such a stage of development as to promise superiority over the method of using cotton or woollen waste when this is properly arranged and manipulated. The writer has had experience with numerous de- vices, most of which were arranged to lubricate from the under side of the journal. The nature of such devices was various ; some were made up of a revolving cylinder in contact with the under surface of the jour- nal, while the lubricating mechanism ran in oil. The roller, when such is used, receives its motion from the journal in which it is kept in contact by a spring or some similar arrangement. Devices of this general na- ture have been made in numerous quantities, all differ- ing only in some minor detail. None of them have been known to produce satisfactory results. Mechani- cal methods in the nature of pads kept in contact with the journal by springs have been tried, but from a thorough test the results seem to indicate that the elas- ticity of the waste commonly used is superior to the mechanical devices, not only in the quality of the lubri- cation, but also in the mileage rendered. One case is METHODS OF LUBRICATION. 35 known where an attempt was made to lubricate by feeding oil through the top of the bearing, and by- waste at the bottom of the journal, similar to boxes used on foreign roads ; and although the trial was of short duration, no apparent advantage over the method now generally used was noticeable. Devices have also been attached to the front of the journal for lifting the oil to the bearing, some of which have proven fairly satisfactory. In fact, the possible methods of lubri- cating car journals are innumerable, but with any mechanical device the objection which can be predicted with a fair degree of certainty is the annoyance and ex- pense that would result from even a small percentage of breakages. The number of hot boxes, when figured on a percentage basis, is exceedingly small and less, it is believed, than can be obtained by any mechanical de- vice, however simple its construction may be. We may except the so-called roller bearings, which have been tried with more or less satisfaction in an experimental way. The theoretical advantage arising by resolving the friction from that of sliding to that of rolling would appear to be a large gain ; and yet from the experiments of Wellington (see Proceedings of American Society of Civil Engineers) it would appear that this advantage is indicated only during starting, but is not so large a per- centage gain after the velocity is increased. Roller bearings have been known to run successfully for loo,- ooo miles, but it is not known that they have been sub- jected, by a more general introduction, to an extensive trial to determine their mechanical efficiency, such as wear and tear, and comparison with sliding friction with the surfaces separated by a lubricant. 36 CAR LUBRICATION. The results of Tower's experiments, previously re- ferred to, give a close idea as to what is to be expected of the different methods of lubrication. It was found with three (3) methods of lubricating journals, feeding the oil from below and from above the journal, that the following ratios of their efficiencies will result : Method. Actual Load. Pounds per sq. inch. Coefficient of Friction. Comparative Friction. Oil Bath 262 252 272 0.00139 0.00980 O.ocgoo 1. 00 Siphon Lubricator 7.06 6.48 Pad under journal The siphon lubricator was placed on the top of the bearing. The tests were under the same conditions. Rape-seed oil used. Journal, 4 inches diameter, run- ning at a speed of 150 revolutions per minute. Tem- perature 100 degrees (Fahr.). There can be no question, then, as to their relative efficiencies. The oil-bath rnethod has had numerous trials upon car journals, but has always proved a fail- ure from the difficulty of obtaining a mechanical means that would retain a tight joint at the back of the box unless it be made of such a complicated nature as to outweigh, on account of repairs, the advantages accru- ing from the method. The siphon method has objec- tions on account of the high resistance offered, the cause for which will appear further on. It can safely be concluded that the use of a pad under the journal gives a higher resistance than would be obtained with what is known as waste, due to the closer texture which its name imphes. This gives it less power to M&THdDS OF LUBRiCATIOl^. 37 absorb the oil, the importance of which is evident from the high efificiency obtained with the oil-bath method of lubrication. Two surfaces that fit tightly wnen dry can be made to move easily on one another oy interposing a lubricant. It would seem, with the resist- ance arising from the tight fit when dry, that the intro- duction of additional material between their surfaces would offer still more resistance to their motion. The application is so common that the reason why it should be so is generally lost sight of. With this exceedingly thin layer of oil, whether in the nature of globules which have penetrated the pores of the metal, or a continuous layer of the lubricant between the surfaces, the result indicates the strength of the wedging action between the metals in contact which enables the lubri- cant to be introduced. This is quite the same action as when lubricating by means of an absorbent material which is saturated with the lubricant and placed in con- tact with the lower side of the journal, unless the bear- ing grips the journal and scrapes the oil from the sur- face, in which case the object is defeated. With journal friction, from the nature of the distribution of the load, it will be noticed, by referring to Chapter I, that the radial pressure of the journal increases from zero, when the bearing includes a semicircle, to a maximum which is on a vertical line through the centre of the journal ; that is to say, the nature of the distribution is such as to make the action the same as that of a wedge even when the bearing is in contact throughout the whole of its arc. It is an increase from a minimum to a maximum pressure per unit of surface. The results of Tower are practical demonstrations of this effect. 3S Car LuBRiCATioi^. During the progress of the experiments with the oil- bath method of lubrication he had occasion to remove the bearing. It was then decided to insert a lubricator in the top of the bearing, for which a |^-inch hole was drilled. After re-starting the experiments and before the cups were inserted in the top, oil was observed to rise in the hole which had been drilled, and was noticed to exude at considerable pressure, which, when indi- cated on the guage attached to the top of the bearing, was found to be more than 200 pounds per square inch, while the average pressure (vertical) upon the bearing was 100 pounds per square inch. It was further found, when a groove was cut the whole length of the bearing and a lubricator attached to feed oil to this groove, that even with a pressure of seven (7) inches head of oil, it would not feed to the bearing, but, on the con- trary, it appeared to be the means of escape for the film of oil between the bearing and the journal. When the cup lubricator was the only feeder, the bearing would not run cool with the pressure as low as 100 pounds per square inch. In this case, care was taken to chamfer the edges of the groove to prevent any scrap- ing action. As the point of application of the lubricant was moved from a vertical towards a horizontal posi- tion, the friction decreased and the bearing was found capable of carrying greater pressure before seizure. The experiments by Tower to determine the pressures at different parts of the bearing are so indicative of the wedging action which takes place that they are referred to somewhat in detail. The bearing was divided into three (3) vertical planes lengthwise of the bearing, and METHODS OP LUBRICATION. 39 each half into three (3) planes at right angles to the first ones. The lubricator was placed on the intersection of the planes passing through No. o, No. i, and No. 2, and N6. Ni. 1' 'No. 2 I No.8A,l