EFFECT OF BRAKES UPON RAILWAY TRAINS Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924004136432 Cornell University Library TF 415.G18 The effect of brakes upon railway trains 3 1924 004 136 432 THE EFFECT OF BRAKES UPON RAILWAY TRAINS CAPTAIN DOLGLAS GALTON, C.B., HON. D.C.L., F.R.S., MKM. INST. M. E. REPRINTED WITH A PREFACE JiY THE WESTINGHOUSE AIR BRAKE COMPANY PITTSBURGH, PA. I8 94 PREFACE. In reprinting the reports of Captain Douglas Galton, to the Insti- tution of Mechanical Engineers of London, of experiments conducted by him, in the year 1878, with apparatus constructed by George Westinghouse, Jr., there are two purposes ; one is to preserve an in- teresting portion of the history of the development of power brakes for railway trains, and the other is to call attention to the pertinence of the results, found at so remote a date, in the conditions of the present day. These experiments originated under the following circumstances : A short time previous to these experiments, a paper relating to brakes was read before the Institution of Mechanical Engineers, and during the discussion of this paper, Mr. Westinghouse called atten- tion to the fact that, in testing the action of various kinds of brake shoes, he had observed a very marked difference in the friction of the shoes upon the wheels at high speeds and at low speeds. He believed that a determination of the facts was of great importance and volunteered to design and construct the necessary automatic recording apparatus and to carry out a system of experiments under the direction of any person who should be appointed by the President of the Society to supervise the tests and report to the Society. The Society immediately took advantage of this offer of Mr. Westinghouse and designated Captain Douglas Galton, who, on behalf of the Society, personally directed the experiments. The success of the project became assured when the London Brighton and South Coast Railway placed a locomotive and brake van at the disposal of Captain Galton and Mr. Westinghouse, and offered every facility for conducting the experiments ; to the great interest manifested by this railway, the valuable information secured from these experiments is largely due. It is not a little interesting to observe that the results obtained in these experiments, which were the first investigations of this character to be carried out upon any practical scale, may be considered the most, if not the only, reliable data in existence upon this subject ; furthermore, subsequent investigations have tended to confirm the KJ-FKCT OF RAILWAY liRAKES. general character of the results, if not the entire accuracy of the data, thus presented. During the time which has elapsecls>ince these experiments were conducted, there appears to have been, until recently, little dis- position to enlarge upon the relations between the friction of brake shoes upon car wheels and the materials, speed, pressure and tem- perature of the surfaces in contact. Until recently, also, little effort has been directed toward a practical application of the knowledge of the fluctuation of the coefficient of friction with the speed, so definitely discovered in these experiments. This apparent indifference, how- ever, is largely explained by the reflection that the practical working requirements of the times have been sufficiently fulfilled, without the use of the necessary additional appliances to attain the highest pos- sible efficiency, and by the fact that the development and perfection of the mechanical apparatus, by which the pressure is applied to the brake shoes, have presented much more immediate and urgent de- mands for consideration. At the time when these experiments were conducted, short trains and moderate speeds represented the conditions of railroad service which power brakes must meet. Since that time there has been a growing tendency to increase the speed of passenger trains toward the maximum limit and, with the advent of the use of power brakes upon freight trains, where the number of cars far exceeds that of any pas- senger service, the conditions of the operation of power brakes have become so entirely changed that very great modifications in the method of applying the power to the brake shoes have resulted. It has thus occurred that, even if, during the past fifteen years, little has been done in pursuing the line of investigation inaugurated by Captain Galton, very great strides have been made in the perfection of the mechanical apparatus of power brakes for railroad service. At that time, while the automatic air brake apparatus was beginning to receive general and favorable consideration by the railroads, the vacuum and the straight air brake apparatus were still very largely in use and had many devotees. Since then both of these forms of brake have given way almost entirely to the automatic air brake, which, since the elab- orate and remarkable experiments of the Master Car Builders' Asso- EFFECT OF RAILWAY BRAKES. Ill ciation, at Burlington, Iowa, in 1886-7, nas > m turn, been almost completely replaced by the quick-acting automatic air brake. Suc- cessive comparisons of the efficiencies of the vacuum, straight air, plain automatic air and quick-acting automatic air brake systems, will furnish most convincing proof of the steady development of this branch of the subject of power brakes for railway service. The high speeds recently employed in passenger service in some localities, have called attention to the great desirability, if not absolute necessity, of compensating for the reduced friction between the brake shoe and wheel at high speeds, by an increased pressure at such speeds, which sh?U be gradually reduced as the speed becomes re- duced and the coefficient of friction increases. Probably no more 'accurate or efficient method of automatically governing the pressure of the brake shoes could be employed than that designed by Mr, Westinghouse and described in the reports of Captain Galton ; but there are certain mechanical disadvantages in such a method, when applied to the four or six-wheeled trucks used in passenger service of the present day. These experiments, however, furnish the general information upon which the utilization of a varying pressure between the brake shoe and wheel must be based. Briefly stated, the suc- cessful operation of a variable pressure brake requires that, at high speed, the maximum practicable pressure between the brake shoe and wheel shall be immediately and fully applied and that this pressure shall be gradually, continuously and automatically reduced, until it reaches a point at which the wheels will not slide upon the rails in coming to a stop. In these reports of Captain Galton, the results of a large number of experiments are tabulated in such a manner as to indicate quite plainly the variableness of the coefficient of friction with different speeds, and also its variation with the time of continued application of the brake shoe to the wheel when the same speed is uniformly maintained. No attempt, however, seems to have been made to gen- eralize from these results, and, as preparations are now being made by the Master Car Builders' Association to experimentally determine the brake-shoe friction under a greater variety of conditions, it will be useful to have the results of these experiments presented in a more general form. EFFECT OF RAILWAY BRAKES. Table IX, on page 172 of the third paper, shows the results of a large number of experiments and gives the mean coefficient of friction observed, at the initial application of the brake shoes to the wheels, for different speeds. The striking characteristic of these results is that the friction is greatest when the wheels are just revolving, and that at consecutively increased speeds the friction becomes constantly dimin- ished, but at a less rapid rate as the speeds become greater. It is interesting here to note that, at least between dry, metallic surfaces, the coefficient of friction follows a uniform, unbroken law, from the point of absolute rest up to high velocities, and the old theory of the association of friction of rest with adhesion appears to be without foundation. If the mean coefficients of friction of Table IX be plotted, it will be seen that there is a regular gradation from the highest down to the lowest speed, when motion is just perceptible. It will also be observed that but one of these mean coefficients of fric- tion, ?'. e., that for sixty miles per hour, deviates, to a marked degree, from a smooth, regular curve. The value given for sixty miles per hour is .074. which is probably somewhat too low. The maximum and minimum observations are given, as well as the mean of all the observations, for each speed, and it will be found that, in most cases, the mean of the maximum and minimum is not far from the mean of all. The mean coefficient for a speed of sixty miles, however, is notably less than the mean of the maximum and minimum observa- tions ; it is also the mean of a much smaller number of observations than at most of the other speeds. It is highly probable, also, that at such high speeds greater difficulty would be encountered in making correct observations and in maintaining the adjustment of the record- ing apparatus. The mean value recorded for the speed of sixty miles may therefore be regarded with some distrust and it has been omitted in determining the following expression for the relation between the coefficient of friction and the speed. In determining the law governing this variation of the coefficient of friction, the controlling considerations are that as the speed in- creases the coefficient of friction decreases, but with gradually less rapidity, and, as it is manifest that the friction between the brake shoe and the wheel could not become absolutely zero at any finite speed, >F RAILWAY liRAKKS. however great, the coefficient of friction varies from that of rest, when the speed is zero, to zero, when the speed is infinite. The mathe- matical form of expression fulfilling these conditions is / = -^-jjr in which / denotes the coefficient of friction, V is the speed in miles per hour, and a and b are constants to be determined from the mean values of the coefficient of friction, given in Table IX. By the application of the method of the least squares to the mean observed values of Table IX, it is found that the most probable values of these constants are a = .326 and b = .03532, so that the relation between the coefficient of friction and the speed in miles per hour, under the conditions of these experiments, is expressed by the equa- tion .326 f = 1 + -03532/' ' The following table shows the values of the coefficient of friction calculated from this formula, at speeds varying from o to 100 miles per hour, and also the means of the observed values in Table IX. Coefficient of Friction, f. Coefficient of Friction, f. V V Calculated. Observed. 45 Calculated. Observed. .326 ■330 .126 .127 s ?77 2 73 5° .118 .116 10 ^4' 242 55 .III .III r S ~ [ 3 223 60 .105 .074 20 191 192 65 .099 2 5 '73 166 70 .094 30 158 164 80 .085 35 146 142 90 .078 40 135 .140 100 .072 In making use of values of the coefficient of friction calculated from this formula, it should always be borne in mind that they are only known to be applicable to conditions similar to those under which these experiments were conducted, that is, with cast iron brake shoes and steel tired wheels ; also, that the values obtained for speeds above sixty miles per hour might become somewhat modified, if the experiments from which the formula is deduced had been extended to cover observations at higher speeds. VI EI'KEl'l OF KAII.WVY BRAKES. Xo reliable data exists, at the present time, to show the exact effect of speed upon the coefficient of friction between shoes and wheels of other materials than those described in these experiments ; such information, however, as can be secured, indicates that, while the values of the constants a and b, in the above formula, may vary for brake shoes and wheels of other materials, the general effect of speed upon the coefficient of friction is the same ; that is, the friction is greatest at very low speeds, and decreases, by the same general law, as the speed increases. The results shown in Table X, on page 172 of the third paper, uniformly indicate that, at any constant speed, the friction also dimin- ishes, according to some law, as the time of rubbing of the brake shoe upon the wheel is extended. The extent of the experiments in this direction is somewhat too limited to determine, as fully as is to be desired, the effect of the time of continued rubbing upon the coefficient of friction. Yet it is apparent that, however long the brake shoe continues to rub upon the wheel, the friction can never become reduced to zero, and the general relation between the coefficient of friction and the time of continued rubbing must therefore be expressed by an equation similar to that for the coefficient of friction as effected by speed. Lety represent the coefficient of friction for a given speed, at the time when the brake shoe is first applied to the wheel, as found from the above table ; then, at t seconds afterward, the coefficient of friction will be /"'= — f — , in which 1 is a constant to be deter- I -\- ct mined. The most probable value of c, determined by the method of least squares from the observed values of/' and t for the different speeds in Table X, is found to be different for each speed. This indicates that the rate of reduction of the coefficient of friction, as the time of rubbing of the brake shoe upon the wheel is extended, is governed by a different law for each different speed ; or, in other words, the quantity c is not a constant but a function of the speed. From the values of c obtained for different speeds, it is found that c = .0022 V, so that /' = — x— 002^/7 If v re P resents tne velocity in feet per second at the speed of I' miles per hour, r =-— v and 44 .0022 It = .00157'/. The product vt of the feet per second by EFFKCT OK RAILWIY IIKAKKb. Vll the number of seconds is, of course, the distance in feet through which the shoe has rubbed upon the wheel, or which the train has traveled since the shoe was first applied to the wheel. Representing this distance by s, /'= — As / = — -,- ' J -- ,-.- , the i + .0015,. 1 + -03532/ coefficient of friction, at an}' instant during the stopping of a train, is / =--— ■-"- — — in which V is the speed, in miles I 1 + -°353 2f ) (1 + -oor5j) ' per hour, at that instant, and s is the distance in feet traveled since the brake shoes were first applied to the wheels. It has already been suggested that the relation between the coeffi- cient of friction and the speed, when the brake shoe is first applied to the wheel, seems to follow the same general law, /. c, f= — A- 7-=^, re- gardless of the materials of which the shoe and wheel are constructed ; but, from such investigations and experiments as tend to throw any light upon the subject, the relation between the coefficient of friction and the time of continued rubbing of the shoe upon the wheel appears to be very different for the use of different materials in the brake shoes and wheel tires. While the coefficient of friction was found to be reduced through continued rubbing in the experiments of Captain Galton, with cast iron shoes rubbing upon steel tired wheels, in other cases, where chilled cast iron wheels were used, it was found that the coefficient of friction increased through a certain period of continued rubbing and gradually approached a point at which it remained constant thereafter. It has even been found that, with the same chilled cast iron wheels and with cast iron brake shoes, the relation between the coefficient of friction and the time of rubbing followed laws of a wholly different character where the shoes dif- fered in hardness and quality of cast iron. It has indeed been observed that wheels have been caused to slide upon ordinarily dry rails, when it seemed quite certain that the pressure of the brake shoes upon the wheels was considerably less than the pressure between the wheels and the rails ; so that the coefficient of friction between the brake shoes and the wheels, even at no insignificant speeds, must have exceeded the coefficient of friction of rest between the wheels and the rails. In these cases, the brake shoes had been applied to the wheels for a number of seconds, in reducing the speed EFFECT OF RAILWAY BRAKES. of the train, before the wheels commenced to slide, and it is reasonably inferred that the effect of continued rubbing of the shoes upon the wheels was to considerably augment the coefficient of friction. It was decided by the Master Car Builders' Association, after the experiments at Burlington in 1886-7, that, for freight cars, a pressure of only 70 per cent, of that which the wheel brings upon the rail should be applied to the wheel by the brake shoe as the maximum braking pressure. This seemingly low limit was found by experience to be necessary to avoid sliding the wheels upon the rails. Such a necessity might, in some degree, be traced to an increase of the coefficient of friction at low speeds, from the continued rubbing of the brake shoe upon the wheel during the earlier period of the stop. The increase in the coefficient of friction at low speeds, from this cause, would manifestly be greater if the train be running at a high speed when the brakes are first applied to the wheels. The results of some experiments have also seemed to indicate that, at least with the use of certain materials, the coefficient of friction may depend, in some degree, upon the pressure, per square inch of surface, with which the brake shoe is applied to the wheel. This should also be a subject for further investigation. It becomes evident, therefore, that the friction developed by the application of brakes to car wheels is a complicated matter, in which the use of different materials for brake shoes plays a most important part. It will be quite clear that the condition most to be desired would be one in which the friction between the brake shoe and wheel should remain constant during the whole time of application of the brake ; this would permit the attainment of the highest efficiency of the brake with a constant pressure of the shoe upon the wheel. It is equally clear that, with a constant pressure of the shoe upon the wheel, the least desirable condition is that in which the friction varies the most considerably ; this condition requires the pressure of the shoe upon the wheel to be sufficiently low that the wheel shall not slide upon the rail when the coefficient of friction is at its maximum and the efficiency of the brake is reduced at all other times during the stop. It is therefore entirely reasonable to conclude that a material for brake shoes which causes the friction to decrease as the time of con- EFFECT OF RAILWAY BRAKES. IX tinned rubbing is extended, so offsetting in some degree the increase of friction as the speed declines, is more desirable than a material by the use of which the friction increases with both the time of continued rubbing and the decline of the speed. While it is desirable to select a material for brake shoes which shall produce a high average coefficient of friction, it might easily occur that the expense of renewals, through the wearing away of the shoes, would be altogether out of proportion to the increased friction obtained, in comparison with the use of other materials. There could be little objection to the use of a material for brake shoes which should cause an inferior coefficient of friction, if, thereby, the cost of mainten- ance could be considerably reduced ; but it is certain that, with a decreased coefficient of friction, to maintain the highest efficiency of the brakes, the pressure between the brake shoe and the wheel should be proportionately increased, so that the actual retarding force may be maintained at as high a point as is consistent with a reasonable assurance of not sliding the wheels. To intelligently accomplish this result, it will be apparent that it is needful to secure much more accurate information, concerning the laws governing the friction between brake shoes and wheels of different materials, than is now possessed. It might even be wisely suggested that, with the present information concerning power brakes, the further development of their practical efficiency seems more likely to follow careful investi- gation in this direction than through attempts to improve the mechanical apparatus in use. The Westinghouse Air Brake Co. Pittsburgh, May, 1894. < )N THE EFFECT OF BRAKES UPON RAILWAY TRAINS CAPTAIN DOUGLAS GALTON, C.B., HON. D.C.L., F.R.S., MEM. INST. M. E. EXCERPT MINUTES OF PROCEEDINGS OF THE MEETING OF THE INSTITUTION OF MECHANICAL ENGINEERS IN PARIS, 13TH, JUNE, 1S7S JOHN ROBINSON, ESQ., PRESIDENT IN THE CHAIR BY AUTHORITY OF THE COUNCIL Inst. M. E., 10 Victoria Chambers, Victoria St., Westminster June 1S78. 467 ON THE EFFECT OF BRAKES UPON RAILWAY TRAINS. By Captain DOUGLAS GALTON, C.B., Hon. D.C.L., F.R.S. The following paper is an account of experiments upon the coefficient of friction between the brake-blocks and the wheels, and between the wheels and the rails, at different velocities, both when the wheels are revolving and when skidded. These experiments form the first instalment of a series which it is intended to make, in order to ascertain, 1 st, the actual pressure which it is necessary to exert on the wheels of a train in order to produce a maximum retardation at different velocities ; 2nd, the actual pressure exerted on the wheels with the several kinds of continuous brakes now in use ; 3rd, the time required to bring the brake-blocks into operation in different parts of a train with the several kinds of continuous brakes ; 4th, the retarding power of the different kinds of continuous brakes now in use on trains under similar conditions of equal weight and running at the same speed. This paper includes the first series of experiments only. The author was enabled to make this series through the courtesy of the London Brighton and South Coast Railway Company, and of their locomotive superintendent, Mr. Stroudley, who provided a van and other facilities for making the experiments ; and through the courtesy and assistance of Mr. Westinghouse, by whom the recording apparatus was designed. The author was assisted in making the experiments, and in their reduction, by Mr. Horace Darwin. The experiments described in this paper were made on the Brighton Railway, with a special van constructed for the purpose ; it was attached to an engine, and was run at various speeds, during which time various forces were measured by self-recording dynamometers. The principle of these dynamometers is that the force to be measured acts on a piston fitting in a cylinder full of water, and the pressure '5 4 6S EFFECT OF RAILWAY BRAKES. JUNE 1878. of the water is measured by a Richards indicator connected by a pipe to the cylinder; thus, as the drum of the indicator revolves, diagrams are obtained giving the force acting on the piston. The advantages of this method are obvious, as the indicator can be placed at any convenient point, and the inertia of the water tends to make the pencil keep a position corresponding to the mean force. The construction of the Dynamometers is represented in Figs. 3 to 5, Plate 60. Each consists of a piston, and what answers to the cylinder, but would be better described as a cylindrical box with a ring fastened upon its upper edge. A is the rod by which the thrust to be measured is transmitted to the piston B. This piston merely consists of a cast-iron disc, having on one side a central cavity in which rests the rounded end of the rod A, and on the other side a central projection which acts as a guide. The ring C, resting on the edge of the cylindrical box D and bolted to it, is of the same thickness as the piston, which fits within it, the ring and piston thus forming a cover to the box I). The piston fits so as to slide easily within the ring, with but little friction ; and is made water-tight by placing below it a disc of india-rubber, which is fastened to the centre of the piston by a brass collar, and has its outer edge clamped in between the ring and the edge of the cylindrical box D. The dynamometer has thus a perfectly water-tight piston, which will move with very little friction ; and as its movement is very small, the disturbing effect of the india-rubber at its edge may be neglected ; consequently the forces acting on the piston through the rod A will be registered by the indicator by means of the pressure of the water. F is the pipe leading to the indicator. We will neglect the valve E for the present, and explain its use a little further on. Supposing the whole apparatus to be filled with water, and that a force were applied to the piston by the rod A, it would force some of the water out of the box D, through the pipe F, into the indicator cylinder. The area of the indicator piston is 0.5 square inch, and its maximum range 0.8 inch, therefore the quantity of water required to make a maximum movement of the pencil is 0,4 cubic inch ; and as the area of the piston B is 30 square inches, its movement would only be 0.013 or 16 June 1S7S. Ei'i'ii t nr railway iirvkks. 469 7 '. inch ; which is such a small movement that the india-rubber will introduce no appreciable error. If the indicator piston did not leak, and if it were possible to keep exactly the right quantity of water in the apparatus, nothing more would be required to make it work properly; but as this is evidently impossible, the self-acting supply valve K, opening outwards, becomes necessary. A small pipe G leads from an accumulator H, Fig. 1, Plate 58, which is loaded to a greater pressure than can ever arise in the box 1); the excess of pressure from the accumulator tends to close the valve E, and there is also a spring which forces the valve on to its seat. The valve is seated with india-rubber, and is made perfectly water-tight ; its spindle passes up so as very nearly to touch the brass collar on the inner side of the piston. Suppose the whole apparatus to be filled with water when there is no force acting or. the piston ; then if a sufficient force is applied by the rod A, this will move the piston inwards so as to send some water into the indicator, and raise the pencil, and will also open the valve E ; and, as the pressure in the accumulator is in excess of that in the box 1), water will enter the box, and will go on entering till the piston is raised again so as no longer to open the valve. Now if the force on the piston be removed, the indicator spring will force the quantity of water received, which is less than 0.4 cubic inch, back into the box D, and will thereby raise the piston, but through a space less than J. inch ; and thus the piston will never move more than J s inch above the position in which it touched the valve E. But if a smaller force be applied to the piston, as in practice, it will not be pushed in so far, unless sufficient leakage has meantime taken place ; in that case the piston will move inwards through its full distance, and will then open the valve. Thus the valve always keeps the right quantity of water in the apparatus to make it work properly, by occasionally opening and letting in enough water to make up for leakage. In Figs. 1 and 2, Plates 58 and 59, is shown in elevation and plan the general arrangement of the apparatus in the special brake-van built by the London Brighton and South Coast Railway Company for 17 470 EFFECT or RAILWAY BRAKES. JUNE 187S. these experiments. To this van the Westinghouse automatic brake was applied, having four dynamometers attached to it, like the one described. The dynamometers Nos. i and 2, situated as shown in Plates 58 and 59, measure the retarding force which the friction of the brake-blocks exerts on the wheels ; No. 3 measures the force with which the blocks press against the wheels ; No. 4, the force required to drag the van. The arrangement of the levers for applying the brake is not the same as that used on the ordinary rolling stock of the Brighton Railway, but has been slightly modified by Mr. Westinghouse in order to make the pressure equal on both sides of the wheels, and to provide for the application of the dynamometers. M is the cylinder belonging to the Westinghouse brake apparatus ; into this the compressed air flows from the reservoir N when the brake is applied, and forces the two pistons apart, thus moving the two rods P P outwards, and by means of the levers pressing the brake-blocks against the wheels. It is evident from the arrangement here shown that the pressure must be equal on each side of the wheels ; and that the pressure on dynamometer No. 3 must be equal to the thrust on the rod P, and hence proportional to the pressure on the wheels. The lever Q, pivoted at its centre, will evidently tend to turn with a moment equal to the retarding moment exerted by the friction of the brake-blocks on the wheels ; and hence dynamometers Nos. 1 and -> will register forces proportional to this moment. The brake could be applied to all the wheels of the van, but during the experiments it was only applied to the pair of wheels to the levers of which the dynamometers Nos. 1, 2 and 3 were attached. Dynamometer No. 4 is connected to the draw-bar by a lever, as shown in the plan, Plate 59, and thus registers the force required to draw the van. A self-recording speed-indicator was used, designed by Mr. Westinghouse. This instrument has been repeatedly tested, and was used at the brake trials on the North British Railway, and on the German State Railway. It consists of a small dynamometer made on the same principle as that just described ; it measures the centrifugal force of two weights, which are made to revolve by a strap from a pulley on a shaft driven by friction-gear from the pair of wheels to which the brake was applied ; a Richards indicator is used, as with the other dynamometers. As the centrifugal force 18 June 187s. effect of railway drakes. 471 varies as the square of the velocity, the speed is got by taking the square root of the ordinate at any point of the indicator diagram. The diagrams from the speed indicator show the speed of the pair of wheels to which the brake was applied, and therefore the velocity of the train at the moment of applying the brake and subsequently, provided there is no slipping. Any variation in the speed diagram is due to the wheels slipping, and shows to what extent and in what way the brake stops the wheels. Two of Mr. Stroudley's speed indicators were fixed side by side in the van ; one attached to the axle belonging to the braked wheels, the other to the axle which was running free. The difference of these indicators showed if slipping took place. They did not record any diagrams, but were read by means of a Bourdon gauge attached to them, with the face divided in such a way that the hand shows the speed in miles per hour. A similar gauge was attached, for convenience, to the ^Yestingl^ouse speed-indicator described above. The indicators were all placed on a table T in the centre of the van, as shown in Plates 58 and 59 ; and their drums were made to revolve by the cords being wound up on pulleys on the shaft S. which is turned at a uniform rate by a water clock U. This clock merely consists of a plunger sliding in a cylinder through a water-tight packing, and loaded with a heavy weight ; it is wound up by connecting it with the accumulator H, and at the beginning of each experiment a small cock is opened which allows the water to run out and the weight to fall, thereby turning the indicators round at an ascertained uniform speed. Thus while the ordinates of the diagrams taken from these several indicators show the various forces, the abscissae show the time occupied in the experiments. In these experiments the tires were of steel, and the brake- blocks of cast-iron. The difficulties attendant upon the preparation and adjustment of this delicate apparatus consumed so much more time than had been anticipated, that it was only on the 27th, 28th, and 29th of May that a series of experiments could be made. These took place in the vicinity of Brighton. The first day was dry ; the second stormy ; the third fine, with showers. '9 47- EFFECT OF RAILWAY liKAKES. Jl'NE 1S7S. Numerous diagrams were taken by the apparatus, which have required very careful reduction. Unfortunately the date at which it was necessary to send in the paper to the Institution has not afforded time for a complete collation of the results : the author therefore submits this as a preliminary paper; and limits himself to exhibiting in Figs. 6 to 14, Plates 61 to 63, a few of the diagrams which were taken, and which illustrate the more striking results. In these diagrams, the line P P represents the Pressure applied to the brake-blocks ; its ordinates multiplied by 2 40 give the total pressure in lbs., collectively exerted upon the four brake-blocks of the braked pair of wheels. The line F 1 F 1 represents the corresponding Friction between the brake-blocks and the wheels, as measured by dynamometer No. 1 ; its ordinates multiplied by 45 give the total friction in lbs., collectively produced between the two brake-blocks, to which No. 1 was attached, and the wheel. The line F 2 F 2 represents the same for dynamometer No. 2. The line T T represents the Traction ; its ordinates multiplied by 60 give the absolute strain upon the draw-bar in lbs. The line S S represents the linear speed of the circumference of the braked wheels in miles per hour, which, when there is no slipping, is equal to the velocity of the train. The length of each diagram shows the duration of the experiment in seconds, according to the scale marked along the base line. No. 1. Fig. 6, Plate 61. (Exp. n. 29th May.) — In this case the speed remained nearly constant, varying from 41 miles per hour at the beginning to 40 miles per hour at the end. The compressed air was allowed to escape from the brake-cylinder through a small aperture, and thus the pressure between the blocks and the wheels diminished, as shown by line P P. The line F F shows that the friction between the brake-blocks and the wheels diminished more rapidly than the pressure. The speed indicated by the rotation of the wheels to which the brakes were applied was the same as that of the wheels running free ; as shown by the Stroudley indicators. Jl'N'E 1S7S. KKI-'KA'l' ill' RAILWAY BRAKES. X O. 2 . Fig. 7, Plate 61. (Exp. 11. 27th May.) — This experiment was commenced when the van was moving at a speed of 25 miles an hour. The application of the brake slackened the speed to 20 miles an hour in 10 seconds, when the wheels skidded, as shown by line S S; and the experiment terminated in 22 seconds, when the speed had been reduced to 17 miles an hour. In this case the line S S shows no diminution in the speed of rotation of the wheels below the speed due to the velocity of the train, until the skidding took place. In the 10 seconds which elapsed from the time when the brakes were fully on till the wheels skidded, the speed of the train was reduced from 25 to 20 miles an hour, as shown by the line S S ; whereas in the period of 1 1 seconds which elapsed between the skidding of the wheels and the cessation of the experiment the train's velocity was only reduced from 20 miles to 17 miles per hour, as shown by the Stroudley indicators. The skidding was only momentary, the wheels beginning to revolve again almost instantly ; but the grip of the brake-blocks was then almost entirely on one wheel, as shown by the wide divergence between the lines F 1 F 1 and F- F-; even in the case of the higher line the friction was much less than before the skidding ; so that the total retarding effect on the train was greatly diminished. Xo. 3. Fig. 8, Plate 61. (Exp. 24. 28th May.) — In this experiment the velocity of the train was 2 1 miles an hour when the brakes were applied to the wheels, and was reduced by the action of the brakes to 18 miles, when the wheels immediately skidded. The line F F shows that the coefficient of friction between the brake-blocks and the wheels gradually increased as the speed diminished, until the skidding point was reached ; and line T T shows that the tractive force exerted on the draw-bar was suddenly greatly diminished after the skidding took place. 474 EFFECT OF RAILWAY BRAKES. JUNE 1878. Xo. 4. Fig. 9, Plate 62. (Exp. 15. 28th May.) — In this case the brake- van was detached from the engine by means of a slip coupling, when travelling at a speed of 40 miles an hour. The pressure of air in the brake-cylinder, and consequently the pressure on the brake-blocks, remained nearly constant during the experiment, as shown by line P P. The pressure being greater than that required by the coefficient of friction between the brake-blocks and wheels corresponding to the velocity, the friction increased so rapidly as to cause the wheels to skid immediately, as shown by the line S S. After the skidding, the friction at once decreased rapidly, as shown by F F ; but rose again as the train's velocity diminished, and attained its maximum when the train came to rest, which occurred with many jerks in 12^2 seconds. No. S . Fig. 10, Plate 62. (Exp. 16. 28th May.) — In this case also the brake-van was detached from the engine by means of a slip coupling, when travelling at a speed of 46 miles per hour. The pressure of the air in the brake-cylinder was less than in the preceding case, and it was gradually diminished during the experiment ; consequently the force with which the blocks pressed on the wheels diminished to the same slight extent, as shown by P P. At first the friction between the brake-blocks and wheels also diminished very slightly ; but when the velocity of the van had greatly decreased, as shown by S S, the friction, as shown by F F, increased rapidly. The van came to rest in 12 seconds, without any jerk, before this friction had risen to a point sufficiently high to produce skidding. These two diagrams, Figs. 9 and 10, afford a comparison of a stop with skidding and one without skidding. The latter stop was effected, with a uniform motion without jerks, from a speed of 46 miles an hour, in 12 seconds ; while the former required i2}4 seconds June 1S7S. effect of railway brakes. 475 to stop from a speed of 40 miles an hour, with a series of unpleasant jerks. No. 6. Fig. 11, Plate 62. (Exp. 3. 28th May.) — In this experiment the train's velocity was uniform throughout at 44)2 miles an hour. The pressure applied to the brake-bljcks sufficed to skid the wheels at once, as shown by S S. The line F F shows that the coefficient of friction between the brake-blocks and the wheels decreased immediately after the skidding, and only rose at the end of the experiment. Line T T shows that the tractive force on the draw-bar increased with the act of skidding, but largely decreased as soon as the skidding was effected. Xo. 7 . Fig. 12, Plate 63. (Exp. 21. 28th May.) — In this experiment the speed was 45 miles an hour at the beginning, and decreased to 42 l 2 miles at the end. The pressure was slightly decreased during the experiment ; it did not suffice to skid the wheels ; and the wheels with brakes and those without brakes revolved at the same rate, as shown by the Stroudley indicators. The tractive force on the draw-bar, shown by T T, follows a nearly uniform line. A comparison between this diagram, Fig. 12, and the preceding one, Fig. 11, shows that, although the pressure on the brake-blocks in the former case, in which the wheels were skidded, was greater than in this case, where the wheels were not skidded, yet the piactical effect of the brakes, as shown by the tractive force on the draw-bar, was much greater with the wheels braked but not skidded (Fig. 12), than with the skidded wheels (Fig. 11). Xo. 8. Fig. 13, Plate 63. (Exp. 9. 29th May.) — In this experiment the van and the engine were brought to rest by means of the brake from a speed of 40 miles per hour. 476 EFFECT OF RAILWAY BRAKES. JUNE iS/S. The wheels skidded very quickly after the brake was applied, as shown by the line S S ; the retarding force, shown by F F, rose greatly at the moment of skidding, and then fell considerably below the original amount. The wheels remained skidded to the end of the experiment. The diagram shows an increase in the coefficient of friction, measured by the rise in the friction line F F, and fall in the pressure line P P, as the train's velocity diminished ; this increase was slight at first, but more rapid as the velocity became reduced, and was very great at the moment of stopping. Xo. 9. Fig. 14, Plate 63. (Exp. 3. 29th May.) — In this experiment the van and the engine were brought to rest from 39 miles an hour by means of the brakes upon both. The compressed air in the brake-cylinder was allowed to escape through a small aperture after the brake was applied ; and thus its pressure, and consequently the force with which the blocks pressed against the wheels, diminished during the experiment, as shown by P P. The line F F shows that the retarding force due to the pressure of the brake-blocks on the wheels at first diminished, until the reduction of velocity, shown by S S, reached the point where the increase in the coefficient of friction was sufficient to overcome the effect of the diminished pressure applied to the brake-blocks. From this point the retarding effect of the brakes increased rapidly, and before long the wheels were skidded. Up to this point the two pairs of wheels of the van, to one of which brakes were applied whilst the other was running freely, had revolved approximately at the same rate, as shown by Mr. Stroudley's speed indicators. At the moment of skidding the friction curve F F rose in a nearly vertical line, thus showing that the coefficient of friction became very great as the wheels came to rest ; and the time during which the wheels were partly rotating and partly slipping was almost inappreciable. Immediately after this rise the curve F F fell to a point far below Junk i S 7 S . ekkect uk railw.w iirakes. 477 its former level, thus showing a great diminution in the retarding effect of the brakes as soon as the wheels were skidded. After this point the curve rose again while the velocity continued to decrease; and thus showed that the coefficient of friction between the rails and the skidded wheels increased as the velocity of the train diminished. At the moment when the van came to rest the coefficient of friction became very great, as shown by the final rise in F F. The traction line T T varies greatly, according to the working of the brakes upon the engine and van respectively. It is unnecessary in this preliminary paper to give the particulars of the reduction of the diagrams ; but the principal results shown by them may be summed up as follows : — 1. The application of brakes to the wheels, when skidding is not produced, does not appear to retard the rapidity of rotation of the wheels. 2. When the rotation of the wheels falls below that due to the speed at which the train is moving, skidding appears to follow immediately. 3. The resistance which results from the application of brakes without skidding is greater than that caused by skidded wheels. 4. Just at the moment of skidding, the retarding force increases to an amount much beyond that which prevailed before the skidding took place ; but immediately after the complete skidding has taken place, the retarding force falls down again to much below what it was before the skidding. 5. The pressure required to skid the wheels is much higher than that required to hold them skidded ; and appears to bear a relation to the weight on the wheels themselves, as well as to their adhesion and velocity. It would seem that the great increase in the frictional resistance of the blocks on the wheels just before and at the moment of skidding, due to the increase in the coefficient of friction when the relative motion of the blocks and the wheels becomes small, is what destroys the rotating momentum of the wheels so quickly. 2 5 478 EFFECT OF RAILWAY KRAKES. JUNE 1878. With constant pressures, the friction between the blocks and the wheels, and consequently the retarding force, increases as the velocity decreases. In order to obtain the maximum retarding power on the train, the wheels ought never to skid ; but the pressure of the brake-blocks on the wheels ought just to stop short of the skidding point. In order that this may be the case, the pressure between the blocks and the wheels ought to be very great when the brakes are first applied, and ought gradually to diminish until the train comes to rest. There are other points of interest indicated by the diagrams, which require further elucidation ; among which may be mentioned the question whether the coefficient of friction diminishes when the pressure increases and the velocity remains constant. Another question is as to the practical effect on the wheels themselves resulting from the greater amount of work done by retarding the wheels without skidding, as compared with the effect of skidding. The general conclusion which would appear to follow from the results of these preliminary experiments is that none of the hand brakes, and only some of the continuous brakes now in use, have been designed with a clear knowledge of the most essential conditions required in a perfect brake. Experiments connected with the action of brakes on railway trains require very delicate apparatus ; and the author in conclusion wishes to explain that the credit of the design of the apparatus used in these experiments, and of the successful manner in which the apparatus was applied, belongs entirely to Mr. Westinghouse. The efficiency of the arrangements for making the experiments which contributed to the successful results obtained is due to the London Brighton and South Coast Railway Company, as represented by their locomotive engineer, Mr. Stroudley, who gave much personal attention to the work, and by Mr. Knight, the general manager, who afforded every facility for the use of the line. 26 June 1S7S. effect of railway i;kakes. 479 Captain Dduclas Galton said that before his paper was discussed he wished to say a few words in order to apologize to the Institution for having presented it in so incomplete a condition. The fact was, that when he originally undertook to read a paper he had hoped that it would be possible to complete the experiments which were to be made in sufficient time to have them all reduced in a proper form by the date of the meeting ; but unfortunately so many delays had taken place in the making of the apparatus, which was of the most complicated description, that the experiments were not able to be made till about ten clays previously ; and the reduction of those experiments was of so intricate a nature that it was quite impossible for him to place the results in a finished condition before the Institution ; but he hoped in a future paper to be able to present them in a more complete form, together with many further points of interest. He also wished to draw particular attention to the line of speed in the accompanying diagrams, which indicated not the general motion of the train, but the rotary motion of the wheels to which the brakes were applied. In No. 1 (Plate 61, Fig 6) this line, it would be seen, continued straight on, and in that case the wheels continued to revolve at the same speed ; in other experiments the line came gradually down to the bottom, the wheels stopping as the train stopped. In No. 6 experiment (Plate 62, Fig. n) the line of speed ran straight down ; but what this showed was that the wheels to which the brakes were applied skidded, and therefore their speed of rotation was destroyed, whereas the train itself continued to run for some distance further. The President asked whether General Morin would not be willing to make some observations on the paper, as his experience on such subjects was very considerable. General Morin said he had followed with great interest the reading of the paper and the conclusions of the author, and the latter appeared to him to be in complete accordance with the principles of mechanics ; but he desired to have opportunity for reflection before expressing any distinct opinion. The matter was one of very great 27 4S0 EFFECT OF RAILWAY I1RAKES. J11NE1S7.S. importance, and the author had given a promise of presenting a subsequent paper. When this appeared, he (Gen. Morin) should have pleasure in taking the whole subject into consideration, and expressing his opinion upon it. Mr. G. Westix<;house, Jun., said he had a little to add to what had been stated in the paper. Just at the end of each stop there had appeared a very considerable rise in the friction, as shown by the line F F. Now he thought that all who had travelled by train must have noticed the unpleasant jerk that occurred just at the end of a stop. It had been thought that this was due more to the destruction of the momentum of the train than to anything else ; but it now appeared to be largely due to a sudden rise in the friction of the brake-blocks. It had thus been concluded that if, at the moment of coming to rest, the brake force were to be greatly reduced, the momentum would still be destroyed, but in a manner that would remove all objection. The '-brake pressure line" P P in the diagrams indicated the force that was applied to the brake-blocks. Now it would be observed that in some of the diagrams, especially in experiments 9 and 7 (Figs. 14 and 12, Plate 63), the brake force decreased considerably; but during the first period of application the friction of the blocks (as shown by the " friction line " F F) decreased to a greater extent than the decrease in the brake force ; this indicated that the brake force should be applied at first with a nearly uniform pressure, and then decreased more and more rapidly as the train came to rest. A theoretical line might be constructed, to indicate the pressure that should be applied. At the point of skidding, as in experiment No. 2 (Plate 61, Fig. 7), it was found that the "friction line " rose almost perpendicularly, and went in some instances quite beyond the range of the indicator. They had two indicators placed side by side for giving the speed of rotation of the two pairs of wheels, and one of the observers watched these speed indicators carefully, but failed to find, except in one case, that there was any such thing as partial skidding. 'When the wheels once began to revolve less rapidly than was clue to the velocity of the train, their rotation was immediatelv destroyed, complete skidding taking place in about a second in almost June 1S7S. effect ok railway iieakes. 481 every instance. No. 4 (Plate 62, Fig. 9) was the only one out of nearly 100 diagrams that showed any partial skidding. The brake had been specially arranged to go on instantly with its maximum force : and it was found that in this sudden application of the brake there was no inconvenience whatever. It was only just after the wheels skidded, or at the final stop, that any inconvenient jerking was felt. He was clearly of opinion from what he had seen that brakes could be made to operate with much greater force than had yet been applied ; for it had been demonstrated by the experiments that the brake pressure on each wheel should in some cases equal 2 l /z times the weight of the wheel on the rail. At the end of the stop the force should however be reduced to something below the weight of the wheel on the rail, say 75 per cent, of that weight. None of the diagrams shown -were produced when the train was running at high speed ; yet in some cases they had a pressure equal to twice, and in others to one and a half times the weight of the wheel on the rail, without producing skidding even at the speeds shown. In experiment No. 2 (Fig. 7)" the "line of pressure" was seen to be very high, indicating that there was immense pressure, which did not instantly produce skidding. This was one of the first experiments, and whether this was due to grease upon the blocks, or to any imperfection in the surface of the block casting, it was difficult to say. Again, with a wet rail it was found a great disadvantage to use too high a pressure, skidding taking place with a very small amount of pressure in some instances. A very wet rail produced good results ; but, as before stated, with a moist rail the results were not at all satisfactory, the effect of a skidded wheel on a moist rail being much less than on a dry rail. One very curious thing was noticed during the experiments. The two pairs of brake-blocks were attached to two different recording apparatus, and in some instances with skidded wheels the brake-blocks on one wheel alone would record the entire force applied to the brakes. On one of the dynamometers there would be a long zero line, fluctuating a little ; and it only began to record pressure as the wheels began to revolve, or the train came almost to rest. This was clearly shown by the two lines F 1 F 1 and F- F 2 in Figs. 7 and 13, Plates 61 and 63. 482 EFFECT OF RAILWAY BRAKES. JUNE 1878. Since making the experiments it had been decided to change the arrangement of dynamometers Nos. i and 2. The pressures would be all recorded upon one diagram, one indicator (No. 1 ) recording the mean pressure of the four brake-blocks. As the number of persons in the van might vary, it became necessary to have an automatic recording apparatus for the actual weight upon the two pairs of wheels. It was proposed therefore to change the use of dynamometer No. 2, and arrange it for automatically recording the weight of the wheels upon the rails. This would be accomplished in practice by recording the weight upon the pair of wheels having no brake applied ; then the subtraction of this from the total weight would give the weight upon the wheels that had brakes applied. He thought in conclusion that gentlemen who were not pecuniarily interested in brakes should take part in these experiments. He was perhaps too one-sided in his business interests to be of the same service as gentlemen would be who had no interest in the matter. The experiments ought to be taken in hand in the interests of railways generally. He believed it would be quite possible for an engineer to take a train, to measure the levers and the amount of force they brought to bear, and also the length of time it took to bring the brake force into operation, and thus to tell the exact amount of work that they would do in stopping the train, in the same way as the power of guns could be measured. The first thing to do was to settle the exact diagram of brake-block pressure that ought to be used under various circumstances. The pressure should be high at first, and then the greatest amount of work would be done in the beginning of the stop ; but if the line were to run on too nearly horizontal, as might be the case in some instances, the wheels would be skidded, and the retarding effect would not be more than half what it actually ought to be. The essential point was to find out what this pressure line ought to be theoretically at various speeds and under various conditions ; and then to make the brakes such that they would bring out that line in practice. Mr. J. A. Haswell observed that the result mentioned in the paper as to skidding was certainly somewhat surprising, after the 3° Jl'NE tSyS. EFFECT OF RAILWAY 11KAKKS. 483 deductions drawn by the Royal Commission on railway accidents to the effect that when wheels were skidded they retarded the force of the train more than when revolving. Xow that Capt. Galton's experiments had shown the opposite, it was clear how very careful they ought to be in expressing opinions without having thoroughly examined the question. Mr. J. Tomlinson was not at all surprised to find the results of the experiments were such as they had been summed up by the author of the paper, and was glad that some practical trials could now be brought forward to support the views which he had held for the last thirty years, though in opposition, he was aware, to those of many railway men. At the same time, he thought it had been well known by every practical engine-driver for the last twenty-five years that the skidding of wheels was a great mistake. He remembered that in 1S46, when he joined Mr. J. V. Gooch on the London and South Western Railway, the orders to the drivers were that they were on no account to skid the wheels ; and that, the moment the wheels were skidded by accident, they were slightly to release the brake and let the wheels revolve. That was the result of the practical experience of Mr. Gooch and others at that time, and it accorded with his own experience. Thus the experiments that had been made simply bore out that which practical men already knew. He had at the time disputed the correctness of the conclusions of the Royal Commission, that a skidded wheel must do more than a revolving wheel ; and he thought that almost any ordinary engine-man, if asked the question, would say that if he " picked up ' the wheels of his engine she would slide on the rail like sliding on a pair of skates, and that no retarding force, comparatively speaking, would result. In many cases on the Metropolitan Railway he had known slight accidents take place from running against stops ; and the excuse of the driver had always been, " I picked up my wheels just as I was entering the siding." The experiments of Capt. Galton confirmed the correctness of his views, and proved that the excuses of engine- men for mishaps were not always wrong, though they were often considered so by those who had to deal with the cases ; and they also 4S4 EFFECT OF RAILWAY BRAKES. JUNE 187S. proved conclusively that the deductions of the Royal Commission on the brake trials at Newark were not correct. Mr. Charles Brown wished to call attention to a very remarkable accident that took place on the Waldensweil-Einsiedlen line in Switzerland. A part of this line consisted of a continuous incline of about 10 kilometres long (6J4 miles), with gradients varying from 4 to 5 per cent., and it was intended to work this incline on YVetli's spiral rack-rail system. The experiment however was abruptly brought to an end by a very serious accident, caused probably by the loss of the retarding effect of the brakes, through the skidding of the wheels. The engine, with one loaded wagon, had ascended the incline without anything special having been remarked ; the descent was then commenced, when, from some unascertained cause, the spiral rack wheel failed to keep in gear with the rack rail, and the further descent had to be controlled by the common brakes ; but, although these were applied with great energy, and the engine reversed, the speed of the descending train kept continually increasing, the wheels being skidded throughout ; until at a curve the engine and wagon capsized and became a total wreck. Since the accident the line had been worked exclusively by common adhesion locomotives, fitted with air-brakes after the Rigi type, which could be so accurately controlled that no skidding ever took place. From that time the trains had worked with a regularity and security that left nothing to be desired. This furnished a further proof that so soon as the wheels began to skid the retarding effect became very much less. M. Ernest Pontzen said he wished to call attention to the duration of the skidding, for he believed that was an item to be noticed. If the brake was released almost immediately after skidding, none of the drawbacks of the skidding would, he believed, be found to occur. Mr. D. M. Yeomans said that he did not think there was much that was new in the information given. When he first brought the Smith vacuum brake under the notice of the Metropolitan Railway, Mr. Tomlinson had distinctly told him that if the brake went on, as 3= June 1S78. kki-ect of railway iihakes. 4S5 he expressed it, "with a bang" against the wheels, he would not adopt it ; that the application must be quick, but gradual, and be entirely under the control of the driver, so as not to skid the wheels. When the first trial was made, it was found that with 15 in. vacuum the wheels were rapidly skidded, but with 12 or 13 in. they were not skidded, and at the same time made better stops. That was of course a matter easily controlled by the driver ; and after adopting the brake and fitting all his trains with it, Mr. Tomlinson issued printed instructions to drivers not to exceed in ordinary stops 12 in. vacuum. They had carried out those instructions for years very successfully, and without any damage to engines or carriages. It had been stated that the greatest pressure possible ought to be applied to the wheels at the commencement ; but he differed from that view. Every child knew that it was much easier to break a piece of string by a sudden jerk than by gradual strain ; and in the same way there must be much greater likelihood of injury being done to rolling stock by applying a brake with its greatest pressure when a train was at its greatest speed, than by an application which, though quick, was gradual. Railway companies could not afford to use brake power on any system that would increase to a great extent the cost of maintaining their rolling stock, because their passengers were also to a great extent shareholders, and rather than have their dividends diminished would take the chance of collisions. The desired information, however, was only to be obtained by the experience of years in actual service, and not by isolated trials. Mr. Charles Cochrane said he had always understood the grave objection to skidding to be that a flat surface was worn on the wheels, the whole destruction of material taking place at that one spot. He agreed with what had fallen from Captain Galton and Mr. Westinghouse as to the importance of having an average record of the total loads upon the four wheels ; for he could not help feeling that a sudden stop would tend to create a momentum, the line of action of which would be above the journals of the carriage, and which would tend to lift it bodily off the hind wheels, so that there would be less weight on those wheels and a greater weight on the front wheels. 33 486 EFFECT OF RAILWAY 11RAKES. Jl'NE 1878. Mr. Horace Darwin said that the question of the effect of putting a brake upon one pair of wheels alone, in altering the distribution of pressure on the rails between the two pairs of wheels, had been considered before the experiments by himself and Professor Stuart of Cambridge, to whom the suggestion was due ; and they found that, though this difference of pressure was not very considerable, vet when the pull on the draw-bar was great, or the velocity of the van diminished rapidly, it ought to be taken into account. He had not however been able to do so in the present calculations, which were only preliminary. Mr. A. Paget said it would be interesting if Capt. Galton could set at rest the doubts that remained in the minds of a few of the members as to the question of the skidded wheel giving actually- less resistance than the wheel not skidded. Could Capt. Galton give any information as to whether the difference varied in any way according to the nature of the material, or the hardness of the wheel and rail as compared together ? It did seem possible that a skidded wheel, where there was friction between two hard surfaces, might give much less resistance in proportion to the unskidded wheel, than where the wheel and the rail were both of a softer material. Possibly Capt. Galton had experimented upon steel and iron rails, and might be able to give some information on that point. In any case he quite agreed that the experiments should be conducted with careful reference both to the material and the condition of the brake-blocks and wheels ; and perhaps it might be in Capt. Galton's power, with the help of the railway companies, to experiment also with reference to the material and condition of the wheel and the rail as affecting the skidded friction. Mr. Charles Hawksley wished to make a suggestion as to a possible means of obviating the skidding of the wheels of an experimental brake van. It appeared to him that if a governor were attached to the axle of the wheels to which the brakes were applied, and another similar governor to the axle of the wheels that were allowed to revolve free, a differential motion might by that means be obtained, June 1878. effect of railway brakes. 487 by which the brakes might be applied to the wheels with just such a pressure as would produce a maximum effect of retardation without skidding the wheels completely. Of course this was a mere crude suggestion, and he could see that many difficulties would arise in carrying it out'; but he imagined that those difficulties might be overcome. He thought that the experiments described showed very conclusively the correctness of what had been observed by Mr. William Bouch some time ago (as referred to by himself at a previous meeting* of the Institution), namely that skidding scaled off the iron both from the wheels and rails : so much so that on the Stockton and Darlington line, where there was an incline of about 1 in 50 for five miles in length, the rails on the line used in descending the gradient were soon worn away, and scales of iron were deposited along the line by the side of the track, which was not the case on the line used in ascending the gradient. Professor A. B. W. Kennedy said that Capt. Galton had been good enough to describe his experiments to him a short time ago, so that he had had some opportunity of considering them beforehand. He thought that the real cause of the extremely diminished resistance with the skidded wheels might be thus explained: as long as the wheels were revolving against the brakes, the surface under friction was continually changing, the wheels both shifting sideways and moving round against the brakes, and it was not possible for the surface to get polished ; but directly the wheels were skidded, there was theoretically merely a line, and practically merely a very small surface, under friction, which never changed, and therefore must be almost instantaneously polished to a high degree. That might account to a great extent for the extraordinary fall shown in the friction line on the diagrams. The increased resistance at slow speeds seemed to corroborate on a large scale the results obtained in the recent experiments of Professor Jenkin on friction at slow speeds, made upon a very small scale : namely as to the coefficient of friction increasing very considerably when the materials were nearly at rest. *See Proceedings Inst. M. E., January 1878, page 93. 35 _)SS EFFECT OF RAILWAY BRAKES. JUNE 1878. The President said he was glad that Professor Kennedy had called attention to that point, because, looking at the diagrams, there were two things that were somewhat surprising. The first was that there was no such thing as diminishing the velocity of rotation of the wheels by applying a brake ; it had always been thought that a great retardation took place when a brake was applied, the wheels still revolving, but more slowly than before. The second was the enormous increase in the retarding force, just when the speed of rotation was diminished to nothing, which would seem to make it desirable, when an accident was actually impending, to apply the brakes, after the speed of the train had already been reduced, with sufficient rapidity and force to cause skidding, which would produce immense retardation, as shown by the friction line in the diagram. Captain Galton said in reply that, although no doubt a large number of persons had been aware of the fact that skidding did not produce the same degree of retardation as was produced by the application of brakes without skidding, yet it was clear that that fact had not been thoroughly realised by all ; and therefore he considered that the experiments on this point were certainly needed. But even if the question as to skidding had been realised theoretically, it had certainly not been brought into practice by railway authorities ; for they had been content during twenty-five years to go on without endeavouring to make a brake which should give the best results that could be obtained. The experiments showed that it was a very difficult matter so to apply the brake that it should never skid the wheel. What would skid the wheel at one velocity would not skid it at another ; and what would skid the wheel when the rails were wet would not skid it when they were dry. It was therefore a perpetually varying coefficient, and it was an extremely difficult matter to meet all the cases. With regard to the question of wear and tear of the wheels, there was no doubt that the skidding of the wheel had a tendency to wear a flat face on it ; but they must also consider what was the effect on the surface of the wheel of applying a brake without skidding. The heat which was generated during the continued revolution of the wheel was a matter for consideration, and one which 16 June iS;S. r.ri-Ecr ok railway brakes. 4S9 seemed to require more elucidation. The question of the relative effect of brakes with different sorts of material, c. ,^., steel and iron rails, was also a question that required further experiments. They had been much hurried with the experiments ; and as the results on iron rails were obtained on a day when it rained, and those on steel rails on a day when it did not rain, he did not feel they could be properly compared together ; and he was therefore unable to separate the diagrams for the two materials so as to present them in any reliable form. On all these grounds he was anxious to continue the experiments ; and he hoped that by the further experiments, to be made with the assistance of the railway companies, results would be obtained that would be worthy of the attention of the Institution of Mechanical Engineers. The Prksidext said he would ask the members to pass a vote of thanks to Captain Galton for his interesting paper. They ought not to lose sight of the fact that, although the paper was short (which in some respects was desirable ), it was of great value. It was not the length of a paper which indicated the amount of labour that had been spent upon it ; and every member present would understand the amount of time, expense, and patience required to get all these facts placed before them in the way that Captain Galton had done. He asked the members to thank him, not only for what they had then received, but for the good things he had promised in the future. The vote of thanks was passed by acclamation. Plate 58. ^i < 1 C 5| ; < : v. DC in Li_ ^ fe( o =• H r o ~ LU & LU c LU PI life .50. EFFECT OF RAILWAY BRAKES. Construction of Jiipiti motiictei-s. Fig. 3. Side XCh ■cation , Plate GO. 9 10 11 12 Inches {Proceedings Inst. JI. E. 1878.) EFFECT OF RAILWAY BRAKES. Plate 61. Miles p.Jioiir Imlieator Diagrams from JSralte Experiments, Fig. 6. Experiment Xo. 1. S ji ;!/'"" t I SC*C T \ .\ V.\\ p.sq. in. ^ Fig. 7. // if/'' ii' F- ,« if p.liour ; : _-m .- ) i/ ii Ls 'X _. F 1 fj Experiment Xo. ,?, ii ii Ii ii ii !_■_ ■■-fi / i '' i i ! >-, F 2 v- F 1 Lbs. p sq.in. 100- 110 SO 70- \ 00 \ 50- \ *o- \ 30- \ -.'I i -A 18- Fig. 8. Experiment -To. 3. i! / ii i }}.Sq 70- 00- 1 v . -;i p\ j0- 1/ F' J \ Jlilc* \\s- | ^i \ ■M 10- 30- SO 10- ■it v>-. — „, / v- \y « — — \\f 2 ..- •, / \\ ' j. / K '•■-.-' v^ ■-r ■VTN X*P Pressure applied to JSrahe-Dlocks ', e'« Ms. I? 1 Friction between Brake- Jit oehs and Wheels, Xo. I Dynamometer. IT 2 Friction between, JiraUe-JBloeUs and Wlteels, JY*».3 Dynamometer. TT Traction upon Draw-Ear, in lbs. SS Speed of Eralied Wheels, in miles per Hour. (Proceedings Inst. M.E. 1878.) EFFECT OF RAILWAY BRAKES. i] Indicator Diagrams from Brake Experiments^. ill JTig.,0. I&r per intent No. 4 I \ jj Plate 02. Lbs. p.sq. in. 5 10 Sec. pp Pressure applied to Brake- Blocks , in lbs. F 1 Friction between Brake-Blocks and IVJieels, No. 1 Dynamometer F 2 Friction between Brake-Blocks and Wheels, No. 2 Dynamometer TT Traction upon Draw-Bar, in lbs. SS Speed of Braked IFJieels, in mile* per hour. {Proceedings Inst. M.E. 1878.) EFFECT OF RAILWAY BRAKES. Indicator Diagrams from Brake Experiments. Fig. 13. Experiment No. 7. — V- Mites I p. hour! r45 (T -10 I | -JEl-'--- r :r •«,. \ V Plate US. Mm. jt.sq.in. 701 50 lo- se 20- *V i! /I /'! rh— ~ v / / | Miles j J jl,-i p.hour Fig. 13. Experiment No. 8. H>8. p.tsq.in. I'l 1/ 11 ill in 'IV /: \ / I / /! T— Fig. 14 PP Pressure applied to Brake- Blocks , in lbs. F" friction between Drake- Blocks and Wlicels, No. 1 Di/namomctej'. F 2 Eriction between Brake- Blocks and TFIieels, No. 2 Dynamometer. TT Traction upon Draw-Bar, in lbs. SS Speed of Braked IVIiecls, in miles per hour, (Proceedings Inst. M.E. 1878.) ON THE EFFECT OF BRAKES UPON RAILWAY TRAINS (SECOA'D PAPER) CAPTAIN DOUGLAS GALTON, C.B., HON. D.C.L.. F.R.S., MEM. INST. M. E. EXCERPT MINUTES OF PROCEEDINGS OF THE MEETING OF THE INSTITUTION OF MECHANICAL ENGINEERS IN MANCHESTER, 24th OCTOBER, 1S7S JOHN ROBINSON, ESQ., PRESIDENT IN THE CHAIR BY AUTHORITY OF THE COUNCIL Inst. M. E., 10 Victoria Chambers, Victoria St., Westminster 5go Oct. 1878. ON THE EFFECT OF BRAKES UPON RAILWAY TRAINS. (Second Paper.) By Captain DOUGLAS GALTOX, C.B., Hon. D.C.L., F.R.S., of London. The experiments which the author brought to the notice of the Institution at the Paris meeting have since been continued, and the results then obtained have been more completely investigated. The apparatus used was substantially the same as was described in the first paper. But for the new experiments, it has been somewhat altered by a rearrangement of the levers. The altered arrangement is shown in Figs. 1 and 2, Plates 83 and 84. In the first experiments the friction of each pair of brake-blocks upon the wheel was recorded on a separate diagram. By the rearrangement the levers from all the brake-blocks act on one dynamometer, and the friction of all the four brake-blocks applied to the pair of braked wheels is recorded on one diagram. The description of the manner in which these levers act, given in the first paper, applies generally to the altered arrangement. An addition was also made to the apparatus, for the purpose of obtaining the proportion of the weight of the van which rested on the braked wheels. To effect this, a dynamometer (No. 2), of similar construction to the others already described, was fixed to levers LL, as shown in Plates 83 and 84, these levers being connected with the ends of the springs which support the body of the van above the unbraked wheels. From the diagrams furnished by the indicator for this dynamometer, the proportion of the weight of the body of the van resting on this pair of wheels was obtained. The weight of the body of the van with the apparatus, etc., when stationary, was found to be as follows : — On wheels not braked . 8,764 lb. On braked wheels 9.436 lb. Total . 18,200 lb. Oct. iS;S. effect of railway brakes. 591 To this had to be added the weight of the persons in the van, and that of the wheels, axles, axle-boxes and springs. The former was assumed at 160 lbs. per person ; an experiment made by weighing 13 persons, who had travelled in the van on one of the days of the experiments, having given an average weight of 159 lbs. per person. The latter was obtained as nearly as possible from the records of the locomotive department of the Jirighton Railway, and formed a constant quantity, one-half of which, with the allowance for the number of persons in the van, was added to the difference between the weight shown by the diagrams and the total weight of the body of the van, to obtain the weight on the braked wheels. This weight varied at almost every moment during the experiments, and the actual weight at the moment has in all cases been taken for calculating the adhesion.* Preliminary Observations on the E.xperim:nts. — The author would in the first place observe that most of the conclusions given in the first paper will be found to be generally borne out by a further * The mode of action of the dynamometers used for measuring the various forces was described in the first paper (Proceedings 1878, p. 46S) ; but as many questions have been asked of the author about the action of the supply valve E (PI. 60, Fig. 5), the following remarks are here added as a note. In order to make the action of the valve more clear, let it be supposed that there is a constant force, such as a weight, acting on the piston B B, and that the whole apparatus is full of water. As the water slowly leaks past the piston of the indicator, the piston B B will slowly descend until the brass collar touches the top of the valve spindle. It will take some force to open the valve, on account of the excess pressure of water on its other side, and also on account of the spring which forces it on its seat ; and this will introduce a small error in the reading of the indicator, as the whole weight will not be supported by the pressure of the water, and consequently the pressure will fall and the indicator pencil will stand too low. As the leakage continues, the brass collar will force open the valve, and it will open it just so much as to allow the same quantity of water to enter as leaks past the indicator piston ; for if more comes in than leaks out, the piston B B will be raised, and the valve will close slightly and thus diminish the quantity of water which enters ; and if less comes in, the opposite takes place. In passing through the small opening of the valve the water is wire-drawn, and its excess pressure destroyed. 53 592 EFFECT OF RAILWAY BRAKES. OCT. 1878. study of the diagrams ; but some modification is required in the opinion expressed that, when the friction of the brake-blocks is sufficient to check the rotation of the wheel, the rotation is immediately stopped altogether. On the contrary, a closer examination of the diagrams shows that in every case where the wheels become locked and slide on the rails, there occurs, just before this takes place, an appreciable interval of time during which the speed of rotation is gradually, though rapidly, diminished. Genera! Description of t/ie Experiments. — The brake-blocks were applied to both sides of the wheels ; that is to say, there were four brake-blocks to each pair of wheels. This arrangement prevents a strain being brought upon the axle. The blocks were made of cast-iron — wooden blocks being too soft, and wrought-iron blocks proving irregular in their action, apparently from a change in condition, owing to the high temperature evolved by friction. It would indeed seem probable that wrought-iron as well as steel blocks would prove injurious to the tires. The experiments on the Brighton Railway, which were made with the experimental van and an engine, have been supplemented by some further experiments made on the North Eastern Railway with a train of twelve vehicles fitted with the YVestinghouse brake and a similar train fitted with the Smith-Hardy vacuum brake. The experiments on the Brighton Railway may be divided into two classes : (i) those made whilst the speed of the engine was kept up at an ascertained rate; (2) those made by slipping the van when the required speed had been obtained, and allowing the van to come to rest by the application of the brakes. In the first class of experiments the recording apparatus was set in motion by hand at a convenient time before the brake was applied ; in the second class the recording apparatus was set in motion by the automatic application of the brake at the moment of separation from the engine. The general action of railway brakes may be thus described : When a train is moving at a given velocity, the adhesion of the wheels on the rails causes them to revolve ; every point on the surface of the tire moves round at the same rate as that at which the train itself is moving forward ; but every such point, in relation 54 Oct. 1S7S. effect of railway brakes. 593 to the forward movement of the train, comes successively to rest at the moment when it comes in contact with the rail. Now, when the brake is applied with a slight pressure only, the wheel continues to move round at the same rate as the train is moving, but it moves with more difficulty, and this increased difficulty in moving is shown either by an increase in the tractive force required to keep up the forward motion, or, in cases where the accelerating force is not kept up, by the tendency of the moving mass to come to rest in a shorter time than would otherwise be the case. But if the pressure with which the brake is applied be increased, a point is reached when the friction between the brake-block and the wheel first approaches, then equals and finally exceeds the adhesion of the wheel on the rail, which adhesion corresponds with the static friction between the surfaces, because the part of the tire in momentary contact with the rail during its rotation, is for that moment at rest in relation to the forward movement of the train. When this happens, the wheel first begins to revolve more slowly, and then ceases to revolve and slides along the rail, or, as it is usually termed, is skidded. In this case the retardation is no longer due to the pressure upon the brake-block and consequent friction between the brake-block and the tire of the wheel ; but the vehicle is transformed for the time from a vehicle on wheels into a sledge, and the retardation due to the brakes is thus the excess of resistance which is produced by making the vehicle slide along the rails, over that produced by making the vehicle move forward on wheels revolving freely. It is therefore necessary to consider the experiments under these two different conditions of retardation, and we arrive at the two following conclusions : — (A) So long as the wheels continue to revolve, the measure of retardation is the friction between the brake-blocks and the wheels, and this is represented by : — The coefficient of friction between the brake-blocks and the wheels x the pressure applied to force the blocks against the wheels. (B) As soon as the wheel begins to slide on the rail, the measure of retardation is the friction between the wheel and the rail (which 55 594 EFFECT OF RAILWAY BRAKES. OCT. 1S7S. is equal to the strain exerted by the brake-blocks to hold the wheels in their fixed position), and this is represented by : — The coefficient of friction between the rails and the wheels x the weight upon the wheels. In considering these experiments, it will therefore be convenient to class the results obtained under the following heads : — 1. The coefficient of friction between the brake-blocks and the wheels. 2. The coefficient of friction between the wheels and the rails. 3. The general effect of the application of the brakes in retarding the train, as shown by the strain on the draw-bar. 4. The proportion which the pressure applied to the brake-blocks should bear to the weight on the wheels at different velocities. 5. The effect of the time expended in bringing the pressure to bear on the wheels. 1 . Coefficient of friction between the brake-blocks and the wheels. The diagrams* shown in Figs. 3 to 5, Plate 85, will conveniently illustrate this question. During each of these experiments the velocity of the braked wheels was uniform, as shown by the line SS. In Fig. 3 the pressure on the brake-blocks was also uniform, as shown by the line PP. In Fig. 5 the pressure P was made to vary. It appears from Fig. 3, where the pressure and velocity were practically uniform, that the friction diminishes as the time of application of the brakes continues. This is shown by the fall in the friction line FF. It appears from Fig. 5, Plate 85, that the friction varies with the pressure, the friction line FF following the rise and fall of the * The multipliers required in the present series of diagrams for reducing the ordinates to actual lbs. of force are somewhat different from those in the former series (see Proceedings 1878, p. 472). They are here as follows : — ( 60 before skidding. Line FF, No. 1. Dynamometer (total friction on blocks) -: ( 67.5 during " Line PP, No. 3. •' (total pressure on blocks) 240 Line TT, No. 4. " (total traction on draw-bar) 30 Line SS, No. 5. " (speed of braked wheels) — to scale. 56 OCT. 1878. EFFECT OF RAILWAY MAKES. 595 pressure line PP ; and it appears by comparing Fig. 3, where the velocity was 55 miles an hour, and Fig. 4, where the velocity was 40 miles an hour, with Fig. 7, Plate 8f>, where the velocity was only 15 miles an hour, that the proportion of friction to pressure was greater at the lower than at the higher velocity. This latter fact is further shown in Fig. 6, where the experiment commenced at a speed of 52 miles an hour, and ended with the train coming to rest. In this case the effect of time on the application of the brakes came into play at first, and the friction became somewhat reduced ; but when the velocity had seriously diminished, the friction rapidly increased until it nearly approached the value of static friction. The proportion which exists between friction and pressure is the coefficient of friction, and it thus appears that the coefficient of friction increases as the Te/ocitv diminishes ; ami diminishes, at any rate when the surfaces are steel and cast iron, as the time increases during which the pressure has been applied. Some special experiments were made with blocks of small area. The brake-blocks generally used in these experiments were 1 2 inches long bv 3 inches wide, giving a surface of 36 square inches; the small brake-blocks were made so as to afford a surface of pressure against the wheel of only one-third of this amount, or 1 2 square inches, thus making the pressure per square inch three times as great as before. The diminution of surface was obtained by casting projections upon the face of the block. The author is not prepared to say that any greater coefficient of friction was obtained by the extra pressure per square inch, although in one of the experiments, Fig. S, Plate 86, at a velocity of 60 miles an hour, the rotation of the wheels was arrested by these blocks, whilst this effect had not been produced at that speed in other experiments. The experiments on this form of block were stopped because the blocks were entirely worn down in the course of about twelve experiments. Mr. Rennie showed* that high pressure per square inch produced a greater coefficient of friction between surfaces either moving very slowly or nearly at rest ; but it must be borne in mind that the author's experiments were made with high velocities, whereby a serious element of disturbance is introduced, viz., the grinding away * Phil. Trans, for 1S29, p. 159. 57 596 EFFECT OF RAILWAY BRAKES. OCT. 1878. of the surface ; and it is therefore probable that the increase in the coefficient of friction due to increased pressure may have been neutralised by the lubricating effect of the fine particles ground off the surfaces. While the author refrains from expressing any certain opinion as to the relation which the coefficient of friction bears to pressure, so far as these experiments are concerned it is quite clear that in proportion as the pressure is increased or diminished so will the actual friction obtained be increased or diminished. When the friction which exists between the brake-blocks and the wheel thus reaches a certain point, the wheel ceases to rotate and becomes fixed. This point is reached when the frictional resistance of the blocks exceeds the adhesion between the wheel and the rail if the speed is kept up ; or, if the speed is slackening, when it exceeds the adhesion between the wheel and the rail, plus the effort required to retard the rotation of the wheel equally with the retardation of the train ; and the excess of resistance then acts as an unbalanced force, tending to destroy the momentum of the wheel. As far as can be judged from the experiments, it would seem that, whether the speed be high or low, nearly the same absolute amount of frictional resistance is required to skid the wheel, other things being equal ; but that the time during which the rotation of the wheel is slackening, before it is completely stopped, varies with the speed. Thus it would appear from a comparison of Fig. 7 with Fig. 8, Plate 86, that whilst at 15 miles an hour the wheel skidded in about 3<£ second, at 60 miles an hour it required about 3 seconds to skid the wheel. The amount of frictional resistance which determines the point at which the rotation of the wheels is checked varies, it is true, in the different experiments. The ratio which it bears to the weight upon the braked wheels is in some cases as low as .19 ; and in some cases as high as .35 or even somewhat more, the average being about .25 of the weight on the wheels. But it clearly represents simply the adhesion between the wheel and the rail, and varies only with this, and not with the speed. Thus in Fig. 8, Plate 86, where the velocity was 60 miles an hour, the amount of actual frictional resistance which checked the rotation of the wheels was about 5S OCT. 1878. EFFECT OF RAILWAY BRAKES. 597 2000 lb., exhibiting an adhesion of about ig.i per cent. Again, in Fig. 7, where the velocity was 15 miles an hour, the actual amount of friction appears to have been about 2160 lb., exhibiting an adhesion of about 19.6 per cent.* As these two values are so nearly equivalent, it would thus appear that the effort retarding rotation is much the same at all speeds ; and therefore that the point at which the retardation of the rotation of the wheels commences does not vary with the momentum of the wheels, but depends entirely upon the adhesion between the wheel and the rail. It will be observed in Figs. 7 and 8, Plate 86, where the wheel was skidded, while the train went on, as well as in Fig. 6, in which the experiment commenced at a high velocity and ended with the stopping of the train, that just before rest, at the moment when the rotation of the wheel was stopped, a considerable increase in the amount of friction took place. At this point the coefficient of friction will be found to correspond with the coefficient which has been noted by former observers as that of static friction, or of friction between surfaces moving at very slow velocities. The sudden jump in the diagrams shows the sudden change from the one coefficient of friction to the other, at the moment of the wheels ceasing to rotate. In some instances the friction thus developed amounted to .30 and even .325 of the pressure, which is about equivalent to the coefficient of static friction between steel and cast iron, as given in Rennie's experiments. It has been shown above that, whether the velocity be high or low, the period at which the rotation of the wheel is stopped, so that it slides on the rail, depends upon the amount of friction between the brake-block and the wheel, the weight upon the wheel, and the condition of the rails, which governs the adhesion between the wheel and the rail. Therefore, with the same weight on the wheels, and the rails in a similar condition, the same amount of brake friction will stop the rotation of the wheels whatever be the speed. The manner in which speed affects this question is clue to the fact that in order to * In this and other cases the diagrams exhibited are merely specimens of a very large number in the author's possession, all leading to the same conclusion. 59 59S EFFECT OF RAILWAY BRAKES. OCT. 1S78. obtain the same absolute amount of frictional resistance at a high speed as at a low speed a greater pressure is required. For instance, selecting at random two experiments (Nos. 16 and 13, of the 22d July), in the first, with a speed of about 16 miles an hour, a pressure of 8169 lb. applied to the brake-blocks produced a friction of 1560 lb; whilst in the second, with a speed of about 50 miles per hour, a pressure of 13.900 lb. applied to the brake- blocks produced a friction of only 1400 lb. There is much difficulty, however, in satisfactorily establishing the coefficient of friction which obtains at different velocities, owing to the fact that the time during which the pressure is continued to be applied enters so largely into the amount of friction produced by a given pressure. Thus in Fig. 3, Plate 85, with a speed of from 55 miles falling to about 53 miles per hour, a pressure of 35,000 lb. at the commencement of the experiment produced a frictional resistance of about 2000 lb., whilst after 10 seconds the amount of frictional resistance had diminished to 1400 lb., although the pressure maintained was the same. Similarly, in experiment No. 31 of 23d August, at a speed of 30 miles an hour, a pressure of 12,000 lb. produced at first a frictional resistance equivalent to i860 lb.; but after 10 seconds this amount had fallen to 1260 lb., although the pressure had been raised to 13,440 lb. It thus appears that the amount of friction is greatly diminished as the surfaces continue in contact. But the full pressure cannot be applied absolutely instantaneously; and however short the interval of time between the commencement of the experiment and the point at which the friction and pressure are measured, that interval is sufficient to affect the proportion which the friction bears to the pressure. The experiments with which the author has had to deal are moreover so numerous, and their reduction has been a work of so much labour, that he is not yet prepared to say that the coefficient of friction which he is about to give may not be liable to some modification when he has had time for further collation of the results. With this reservation he appends the following table, which gives, it is believed, a fair approximation to the coefficient of friction at different velocities between cast-iron brake-blocks and steel tires as used in these experiments. 60 Oct. 1S7S. EFFECT OF RAILWAY BRAKES. 599- Table I. — Static and Dynamic Friction. VELOCITY Feet i Miles COE1- MCIENT OF FRICTION. per Second. per Hour. Noted by Former I Observers. See Kleeming At Com- Jenkin, Phil, i mence- ! Trans, for nient of 1877. Experi- ment Obtained Approximately from Recent Experiments. After ! After Sec- I Sec- onds. ! onds. After 15 Sec- onds. Static Friction. Morin. — Tron on Iron f S t e e 1 o 11 ] J C ast-I ron Ren.\ie '0 15 16 Qrs. 56 / 'armim Train. — ( Leading Engine 1 Driving ( Trailing Tender 2 Vans Carriages — 3 Composite 7 Third-class. Experimental Van . Total Vacuum Train '74 12 10 14 I 2 15 26 4 's 16 2 2S 2 1 60 12 9 12 179 34 12 2 9 12 163 4 2 '4 12 15 26 4 '5 16 2 2S 2 1 60 12 9 12 167 1 3 69 60S EFFECT OF RAILWAY BRAKES. OCT. 1878. Therefore, in the case of the Westinghouse train, 93.7 per cent, of the weight, and in the case of the Vacuum, 94.4 per cent, of the weight, was on the braked wheels. The weight of the engine and tender resting on braked wheels was, in the case of the Westinghouse, 26.7 per cent., and in the case of the Vacuum, 29.4 per cent., of the whole weight of the train. The recording apparatus in the van was set in motion by means of an electric connection with the brake lever attached to the engine. The author had intended this to come into operation at the first movement of the brake lever ; but by the arrangement made the lever could move through two-thirds of its stroke before setting the apparatus in motion ; so that, when the handle of the lever was moved slowly, any action on the brakes which might take place before this point had been reached would not be recorded. The Westinghouse train was a train which had been working in regular traffic, and was run in practically the same condition as when running on the line, except that it was put in good order. The Vacuum carriages and engine were altered from what had been their condition on the North Eastern Railway, by the substitution of the Hardy cylinders for the Smith Vacuum sack, and by alterations in the ejector on the engine. On the ordinary passenger engines of the North Eastern Railway the bottom nozzles of the ejector are 2 in. diameter, and the top nozzles 2/8 in., and the steam pipe ij^ in. In the experimental engine one of the bottom nozzles was 2?/% in. diameter, and the other 2 in. diameter : the top nozzles 2^3 in. diameter and 2 3/s in. respectively, and the steam pipe 2 in. inside diameter. The leverage was also much more powerful than that in use on the other passenger engines of the North Eastern Railway, so that a very high pressure could be applied to the engine wheels. These alterations, it was stated, were made in order to place the train in the same condition as trains recently fitted up on other lines. It was intended that the brakes should be so arranged that each brake-block, when in a state of inaction, should be removed a full quarter of an inch from the surface of the wheel, so as to ensure that there should be no dragging. This condition was fulfilled in the Westinghouse train ; but in the Vacuum train the blocks, except in two or three instances, were much closer, indeed often less than 7° Oct. 1S7S. effect of railway urakes. 609 }i inch, and in some cases less than ,',.- inch from the tire of the wheel. The Westinghouse train, as well as the experimental van, was fitted with continuous draw-bars. The Vacuum carriages had not continuous draw-bars. The experiments were commenced by putting the van next the engine, and so running from York to Knaresboro' ; the van was then turned and placed at the tail of the train on the return journey. After the Vacuum train had made six stops with the van next the engine, during the journey to Knaresboro', while making a stop on an ascending gradient of 1 in 120, a coupling-hook broke in the leading van, next behind the experimental van. A fresh connection was made, but it was found, on returning to York, that the vacuum cylinder on the tender, for working the engine brakes, had become detached ; this was apparently caused by a want of sufficient play in the lever connecting the Hardy cylinder with the engine brake. The experiments subsequent to the fracture rest under a doubt, therefore, as to representing truly the performance of the Vacuum train. On the following day the van was placed in the middle of the train. In this case the first experiment was made with a speed of 50 miles per hour. There was a violent jerk, and the draw-bar of the leading van was torn out, and the train separated from the engine. As the Vacuum is not an automatic brake, the fracture caused the brakes to be taken off the train ; the rapid and judicious action taken by those in charge of the engine prevented, however, any serious collision between the engine and the train. These fractures of couplings appear to have been clue to the more powerful and immediate application of the brakes on the engine in these cases as compared with the rate at which the brake-blocks came on at the rear of the train ; the buffers were consequently driven home for a time, and then, at the moment when the reaction of the buffer-springs was commencing to have its effect, the brakes at the rear of the train also came into operation ; thus a violent, and in the last-mentioned case a successful, effort at separation occurred between the engine and train. After this accident the experiments were continued with eleven carriages instead of twelve, the van being fifth instead of sixth from the engine. The brake-lever was moved with caution ; the electrical 71 6io EFFECT OF RAILWAY liRAKKS. Oct. 187S. connection for starting the recording apparatus also acted irregularly, so that the diagrams of these further experiments present great dis- crepancies between themselves as to the rate at which the pressure came on. The author on this account obtained the permission of the North Eastern Railway to repeat these experiments, and the renewed trials took place on the 18th October. On this occasion the recording apparatus was set in motion by the first movement of the brake lever. The experiments on the Westinghouse train were partly made by slipping the experimental van and eleven carriages, and allowing the slipped portion to come to rest by the automatic action of the brakes. The experiments on the Vacuum train were a repetition of those of the second day ; in the course of the ninth experiment the draw-bar of the second vehicle from the engine gave way. Figs. 17 and 18, Plate 89, show the results obtained from the two first days' experiments on the North Eastern Railway as to the rates at which the pressures and frictional resistances come into operation in each position of the van in the train ; and the accompanying table shows, as nearly as can be obtained from the diagrams, the time which was required after moving the brake-handle to set the brakes with various degrees of force in different parts of the train. The last experiment, where the van was the 2 1 st vehicle from the engine, was made with a stationary train of 20 carriages, there being only 12 carriages available on the experimental train. Table VII. — Time Expended in Putting ox Brakes. Vacuum Brake. Westinghouse Automatic Brake. Place of Experimental Commence- Three- Commence- Three - Van ment of Half Full ment of 1 Half Full from Engine. Movement ( )n. Oa On. Movement [ On. On. of Blocks. of Blocks. Seconds. Seconds Seconds Seconds Seconds. 'Secotids Seconds Seconds 1st Vehicle * 3 7 II i i i I* 7 th do. 2 6+ H M I | if 2 2* 13th do. 3i 7i 9* 14 If -* 3* 1+ 21st do. 5* 17 3° 3 ■ii 5 5i Oct. 1S7S. >F RAILWAY IIKAKLS. 6ll The short time which has elapsed since the above experiments were made has not allowed the author to analyse the results fully for the present meeting. They give, however, a fail- approximate indication of the rate at which the pressure comes on, with these forms of brake, in different parts of a train. It was evident in the course of these experiments that the difference in the rapidity of application in the rear as compared with the front of the train gives rise to jerks and unpleasant motion in the process of stopping, if not to actual clanger. They clearly show ( as do also the experiments already mentioned on the Brighton Railway) that the perfection of a brake would consist in its application being simulta- neous on all the wheels of a train. The following table gives the particulars of the stops which took place in the experiments of all three days. On the first day the weather was very rainy and disagreeable ; on the second day the weather was fine ; on the third day the morning was foggy and damp, and the rails very greasy ; the afternoon was dry and fine. Table VIII. — S'rors On North Eastern Railway. / acuitm Train. — First Day. (Van next Engine.) Distance Speed. Distance of Time of Gradient. Reduced to No. Miles per Stop. Stop. R=Rise. F=Fall. 50 Miles per Hour. Remarks. Hour. Yards. 40 21 I Yards. I R 1 in 130 3-9 Heavy rain. 2 35 154 14 R 1 in 130 3'4 Do. 3 1 40 220 17 F 1 in 1 200 343 Do. 4 43 22S 17 Level 308 Do. 5 49 294 2o£ Do. 306 Do. 6 35 162 '5 Do. 33° Do. 7 35 '54 14* R 1 in 120 3M Broke coupling. 8 49* 316 21 Level 322 9 53 3 2 5 1 Si Do. 289 Heavy rain. 10 ! 50* 334 21 Do. 3 2 7 Do. 1 1 35 162 I4f R 1 in 1200 33° Do. 12 47i 281 20 F 1 in 130 3 11 Do. I 3 3 2 136 '3* F 1 in 130 33 2 Do. 14 28 114 i=5 F 1 in 130 3/6 Do. '5 3' 136 14 F 1 in 130 353 Do. 73 6l2 EFFECT OF RAILWAY BRAKES. Oct. 1878. 2 3 26 27 28 31 Vu -uui/i Train. — Second Day. (Van in middle of train.) Speed. Distance Time Gradient. Distance Reduced to Miles of of 50 Miles per Stop. Stop. R=Rise. per Hour. Hour. Yards. Seconds. F=Fall. Yards. j Broke draw-bar of 49 440 64i Level 458 j leading van. 39 *I5S <3i F 1 ill 1200 259 Fine. Good rail. 43 *l84 Mi R 1 in 130 24S Do. 37 *I32 I2i Level -41 Do. 41 *l62 '3* R 1 in 1 20 240 Do. 5° *246 i6i Level 246 Do. 55i *299 1 S* Do. 242 Do. 36 *I40 '3 F 1 in 130 270 Do. 35 * I23 12 F 1 in 130 251 Do. *These distances are not reliable, owing to slow action of recording apparatus. Vacuum Train. — Third Day (repeated experiments). (Van in middle of train.) Speed. Distance Time Reduced to No. Miles per Stop. Stop. Gradient. 50 Miles per Hour. Remarks. Hour. Yards. Seconds. Yards. 7 53 308 19 Level 274 "■ 8 4ii '9.3 Mi Do. 280 9 49f 2 55 i6i Do. 258 10 35 M5 I2f Do. 296 Fine. Very good rail. n 57i 356 20 Do. 269 12 57 334 ■9i Do. 2 57 T 3 17 35 2 21 Do. . 271 M 53 312 19I Do. . 278 r 5 35 '91 22 F 1 in 130 . 394 Draw-bar broke. JVest/iig/t onsc Train . — Second Day. Speed. Distance Time Distance Reduced to No. Miles per of Stop. of Stop. Gradient. 50 Miles per Hour. Remarks. Hour. Yards. iSn Seconds. M Yards. 243 .8 43 Level Damp, without rain '9 30 114 12* Do. 3'6 Do. 20 46 21 I ISi Do. 249 Do. 21 37 140 12+ Do. -55 Do. 22 35 ' 2 3 11 + R 1 in 120 251 Do. =4 S6 3.6 >9i F 1 in 130 251 Do. 2 5 35 <3 2 12 Level 289 Do. 26 40 1(12 M R 1 in 1200 2 55 Do. 27 <;o 246 J7i Level . 246 Do. 28 4'i 171 M Do.orF 1 in 130 248 Do. 29 35 123 12 F 1 in 130 251 Do. 3° 2S SS TO F 1 in 130 . 2S0 Do. 74 Oct. 187S. EFFECT OF RAILWAY BRAKES. ''" 3 1 Time w Train. — Second Day. Speed Distance Distance 1 of of Gradient. K educed to Xo. Miles I per Stop. Stop. R=Rise. 50 Miles per Hour. Remarks. Hour. Yards. Seconds. F = Fall. Yards. I 5' 281 1 »i Level 270 Fine. Greasy rail 2 4f l«4 '5 F I in 1 200 -13 Do. 3 48 215 '54 Level 2 33 Do. 4 3° S8 L A Do. 244 Do. 5 36 '49 '34 Do. 287 Do. 6 5° 264 .Si Do. 264 Do. 7 374 140 >?i Do. 2 4 s Do. S 56 334 20f Do. 266 Do. 9 56 299 '9 Do. 238 ] >o. 10 3 ? 149 134 Do. 2 57 Do. 1 1 36 132 12 Do. 2 54 Do. i j 3'4 no "i Do. 277 Do. ! 3 5° 2 55 I7i R 1 in 1 200 . 2 55 Good rail. f4 40 '59 14 Lev. or F 130 24S Do. '5 35 l 3 2 I2i F 1 in 130 269 Do. 16 3^4 96 io£ F 1 in 130 227 • Do. iS I 594 404 2 3 Level 285 Do. r 9 59 378 22| Do. 271 Do. 20 45 202 15 Do. 249 j Signal from driver. | Good rail. 21 55 290 '9 Do. 239 Good rail. "■ 5° 246 ■7i F 1 in 130 246 Do. No diagram. ir. Speed. Miles per stii/^hons ' Train.- — Third Day. ( Van next E ngine.) No. Distance of Stop. Time of Stop. Gradient. Distance Reduced to 50 Miles per Hour. Remarks. Hour. 43 Yards. Seconds, 16 R 1 in 130 Yards. I 228 308 Damp and very greasy rail. 2 53 294 18 Level 2l8 Eleven carriages and experimental 3 57 259 17 Do. 199 van let slip from J engine. 4 59 506 27J F r in 130 363 Very greasy rail. 5 58 378 Level 2SO j>Stops made with j engine and train. 6 55 316 1 20 Do. 261 J 75 614 EFFECT OF RAILWAY BRAKES. OCT. 1878. Conclusions. — In conclusion the author would recapitulate what appear from these experiments to be the essential conditions of a good brake, in addition to other matters not coming immediately within the scope of this inquiry. 1 st. The skidding of the wheel, so that it slides on the rail, is altogether a mistake, so far as rapid stopping is concerned. 2nd. The pressure with which the brake-blocks are applied to the wheels should be as high as possible, short of the point which would cause the wheels to be skidded and to slide on the rails. 3rd. The rotation of the wheel is arrested as soon as the friction between the brake-block and the wheel exceeds the adhesion between the wheel and the rail ; and therefore the amount of pressure which should be applied to the wheel is a function of the weight which the wheel brings upon the rail. The value of this function varies with the adhesion ; hence, with a high adhesion a greater pressure can be applied, and a greater measure of retardation obtained, than with a low one. 4th. In practice and as a question of safety it is of the greatest importance that, in the case of a train travelling at a high speed, that speed should be reduced as rapidly as possible on the first application of the brakes. For instance, a brake which reduces the speed from 60 miles an hour to 20 miles an hour, in say 6 seconds, has a great advantage as regards safety over a brake which would only reduce the speed from 60 to 40 miles an hour in the same time. 5 th. The friction produced by the pressure of the brake-block on the wheel is less as the speed of the train is greater ; to produce the maximum retardation so far as speed is concerned, the pressure should thus be greatest on first application ; and should be diminished as the speed decreases, in order to prevent the wheels from being skidded (or sliding on the rails) in making a stop. It should be added that the coefficient of friction decreases as the time increases during which the brakes are kept on ; but this decrease is slower than the increase of the same coefficient due to the decrease of speed ; it has therefore little influence in the case of quick stops. 76 Oi'T. 1S7S. EFI'ECT OF RAILWAY 1IRAKES. 615 6th. The maximum pressure should be applied to the wheels as rapidly as possible, and uniformly in all parts of the train. 7th. To prevent retardation from the dragging of the brake-blocks against the wheels when the brakes are not in use, care should be taken that the brake-blocks are kept well clear of the wheels (say half an inch) when in a state of inaction. There are various mechanical questions connected with brakes, such as the desirability of automatic action, and other considerations, which do not enter into the scope of the present enquiry : the special object of which was to ascertain by direct experiment the forces brought into action in applying the brake-blocks to the wheels. Railway companies, in considering what form of brake is best suited for traffic, must, whilst they give full weight to the mechanical conditions discussed in this paper, also consider the question of the convenience of any particular form of brake, and ascertain its durability and facility of maintenance and repair. It is further clear from the present series of experiments that the universal application of continuous brakes will raise many questions as to the strength of the rolling stock now in use, much of which was constructed originallv to meet other conditions of traffic. In concluding this paper, the author would again apologise to the Institution for its incomplete character : the fact being that the enormous mass of information which has been collected has entailed so much detailed study that he has only been able to bring before the meeting on this occasion the present very incomplete sketch. He hopes on the next occasion to be able to complete his contribution upon this important subject. He has to repeat his thanks to Mr. Westinghouse for the beautiful apparatus contrived by him ; and for the very valuable assistance he has rendered in carrying out these experiments. He would further beg to tender his best thanks to the Directors of the Brighton Railway Company, and to their able and energetic Manager, Mr. Knight, and their able Locomotive Superintendent, Mr. Stroudley, for their co-operation and assistance in the enquiry, without which it could neither have been commenced nor carried out. 77 6l6 EFFECT OF RAILWAY BRAKES. OCT. 1878. He has also to thank the Directors of the North Eastern Railway Company, and their able Manager, Mr. Tennant, for their courteous assistance in elucidating this difficult subject. Captain Don .las Galtox wished to explain that the results of the second day's experiments with the Vacuum train, given in Table VIII., were practically superseded by those of the third day's experiments, given just below. The distances given in the former case were subject to some question, owing to discrepancies which he had found in the diagrams ; and therefore the second table of results should stand instead of the first. He had had the opportunity of reading some interesting observations drawn up by the Secretary on the theoretical questions raised by the paper ; and if the President thought right, he suggested that they should be read to the meeting. The President said that, as it was not usual for any of the officers of the Institution to make any remarks upon the papers read, he would take the sense of the meeting as to whether the Secretary's observations should be read, in order to secure their insertion in the Proceedings for the information of the members. The motion having been approved by the meeting, the following observations were read : — 78 Oct. 187S. effect of railway iirvkes. 617 ON THE THKORY OF THE ACTION OF BRAKES UPON THE WHEELS OF A TRAIN. The conclusions derived by Captain Galton from his experiments may be easily shown to be in accordance with the theoretical principles of the action of brakes. These principles ma)- be thus explained. Let us confine ourselves for the sake of simplicity to a single wheel of a railway train moving at a uniform speed. This wheel, while rolling upon the rail, revolves round its own centre with an angular velocity such that the linear velocity of its circumference is equal to the speed of the train. Any point of the wheel will thus have a motion compounded of the general horizontal motion of the train and of this rotary motion round the centre. To get rid of this compound motion it will be well to suppose a velocity equal and opposite to that of the train to be impressed upon even- point of the wheel and of the rail. By a well known principle of mechanics the relative action will not be interfered with. On this supposition the centre of the wheel will be stationary in space ; every other point of the wheel will revolve round the centre with the same velocity as before ; and the rail will move with a motion the same as that of the train, but in the opposite direction. The effect is in fact the same as if we supposed the engine to be employed not to pull the wheel over a stationary rail, but to pull the rail from under a stationary wheel. Let us now consider what will happen. If we neglect all friction of journals, &c, and suppose that the brake is not applied, the power required to keep up the motion will be nil. Let us now suppose the brake applied with a pressure P. This will produce by friction a force (say f P, where f is the coefficient of friction) tangential to the wheel and tending to stop its rotation. This force, transmitted through the frictional resistance or adhesion between the wheel and the rail, will act upon the rail, and tend to stop its motion. If this motion is to be kept up as before, a force equal and opposite \o f P must be applied by the engine to the rail. Hence (A) the additional tractive force required when the brake is 79 6lS EFFECT OF RAILWAY BRAKES. OCT. 1S7S. applied (or in other words the retarding effect of the brake) is equal to the tangential frictional strain of the brake upon the circumference of the wheel. This conclusion is not strictly true except where the speed is kept uniform. When the train is stopping under the friction of the brake, a part of this friction is employed in checking the rotation of the wheel to correspond with the checking of the train. This part of the frictional resistance is thus wasted, as far as stopping the train is concerned : but it is always in practice a small fraction only of the total resistance ; and it is constant at all speeds, since the rate at which the velocity of a body is destroyed depends only on the mass of the body and the amount of force applied to it, and not on the initial velocity of the body. If the adhesion of the rail and wheel were unlimited, this would be a complete account of the whole matter. But this adhesion has a limit, say F'\Y, where F is the coefficient of static friction between the wheel and the rail ; F will always be much greater than f, which is the coefficient of dynamic friction between the wheel and the brake-block. Let the pressure however be so much increased that the brake friction f P is greater than the rail friction FYV. Then, since the force transmitted from the rail to the wheel cannot be greater than F\Y, it follows that, whatever be the pull of the engine, the difference between these two frictional resistances, acting in opposite directions, will remain as an unbalanced force, tending to stop the rotation of the wheel. Consequently this rotation will be checked, and if the wheel had no inertia, it would be stopped instantaneously. As the wheel has inertia, it will be stopped, not instantaneously, but after an interval of time, which will be greater as the mass and velocity of the wheel are greater, and less as the difference between f P and F\Y, the two frictional resistances, is greater. In all cases however this interval will be very short ; because, as soon as the rotation is materially checked, so that the wheel is slipping over the rail with an appreciable velocity, the coefficient of friction between wheel and rail will change from its original value for static friction to the much smaller value for dynamic friction. Hence, whatever may have So OCT. 1S7S. EFFECT OF RAILWAY I'.RAKKS. 619 been the original difference between the two frictional resistances, it will now be largely increased ; and as it is this difference which tends to stop the rotation, this stoppage will be completed in a very short time. Hence (B) the wheel will always continue to rotate at the train speed, until the frictional resistance between wheel and brake-block becomes greater than that between wheel and rail ; but as soon as this takes place the rotation will be checked, and will be stopped completely in an interval of time which in practice will always be very small. If the coefficient f of brake-block friction were the same at all speeds, the pressure which would produce skidding would also be the same at all speeds ; although the time it would take to skid the wheel completelv would be greater at a high speed than at a low one. As the coefficient of friction is less as the speed is greater, this is not the case ; and it requires a greater pressure to skid the wheel at a high speed than at a low one. But in all cases the amount of tangential brake friction which will skid the wheel is independent of the speed, that is unless the rail friction varies with the speed, which there is no reason to suppose is the case. Let us next consider what will happen when the wheel is skidded. Just at the moment when it is coming to rest, and the motion of the wheel under the brake-block is therefore very small, the coefficient of friction between these two will change from its value for dynamic friction to its much higher value for static friction. Consequently the frictional resistance of the brake-block will show a large increase ; but this will not be transmitted to the rail, because the adhesion between the wheel and the rail is already transmitting all it can, and hence this increase in the brake-block friction will not be accompanied by any increase in the tractive force. This is in accordance with the diagrams shown in Figs. 7,8, 12 and 13. Plates 86 and 88, where the friction line is seen always to rise sharply at the moment of skidding, while the tractive force remains stationary, or even falls, at the same moment. Although in stopping the rotation of the wheel a certain amount of vis vii'ii is destroyed, yet this has no retarding effect on the train ; 81 620 EFFECT OF RAILWAY URAKES. OCT. lS/S. since the motion lost is only the rotatory motion of the wheel round its stationary axis, and has therefore no component in any one direction in space. As soon as the wheel is completely skidded, the brake-block friction becomes reduced to a mere mechanical means for holding the wheel fixed, and has no longer any direct effect in stopping the train. The whole of the stopping is thenceforth done by the rail friction. But this has now only its low value for dynamic friction, and not its high value (commonly called "adhesion") for static friction, which gave the measure of the retarding force when the wheel was in motion. Hence (C) the retarding force available for stopping a train is greater when the wheels revolve than when they are skidded, in about the proportion of the value of the static to that of the dynamic friction between wheel and rail. It may be added that (D) this dynamic friction of the skidded wheel will be still further diminished as the time of skidding goes on by the polishing of the small surface resting on the rail. This effect is illustrated by the gradual fall of the traction line in experiment No. 3 (ist Paper, Proceedings 1878, PI. 61, Fig. 8); it was mentioned by Prof. Kennedy (p. 487) as probably a chief cause of the lower retarding effect of the skidded wheels ; but it is clear from the diagrams and from the above considerations that its influence is comparatively small. Let us now consider what will happen if the pressure is taken off the brake-block, and the wheel released. It is clear that as soon as the brake-block friction falls below the rail friction the difference between the two becomes an unbalanced force tending to turn the wheel, and it will begin to rotate. The vis viva thus imparted to the wheel must of course be given to it by the rail, and will produce an increased pull on the rail, in other words an increase in the tractive force. Moreover just when the wheel is coming to its full speed the rail and the tire will be coming to rest relatively to each other ; consequently the friction will change from its dynamic to its static value, the pull it is capable of exerting on the wheel will be greatly increased, and the vis viva still wanting will be imparted to the wheel very rapidly, with of course a corresponding rapid rise in S2 OCT. 1878. EFFECT OF RAILWAY BRAKES. 62 I the tractive force. This will be made clear by supposing that the coefficient of friction suddenly became infinite ; the wheel would then be i/isfaiita/uviis/y brought up to its full speed, which could not take place without a violent impulsive reaction upon the rail. Hence (E) when the brake is slackened on a skidded wheel, the effect will be a rise in the tractive force, gradual at first, and then very rapid, as the wheel assumes its full speed of rotation. When this is completed, the tractive force will at once fall again to the value due to the remaining friction of the brake-block and wheel, or, if the brake is taken off altogether, to zero. These conclusions are illustrated by the rise and fall of the line of traction when the wheel is released, in the diagrams shown in Figs. 7, S, 13 and 14, Plates 86 and 88. The above considerations cover, it is believed, the whole of C'apt. Galton's conclusions, except the two deduced as to dynamical friction, viz., that it decreases with the increasing speed of the moving surfaces, and also with the increasing time during which it is exerted. These are of course empirical facts, established by his experiments. The first however is fully in accordance with the theory that the work done by friction is expended chiefly in shearing off the small inequalities of the rubbing surfaces : following on the well-known principle that all such operations are performed with a much smaller expenditure of work at a high speed than at a low one. The second may be referred to the polishing action, which must of course take place to some extent, combined possibly with the effect of the great heat developed between the surfaces, in diminishing the shearing resistance of the inequalities. Mr. Charlks Sacre observed that the question of continuous brakes was one which now occupied the attention of all railway companies, and its importance could not in any degree be overrated. Independently of the points treated in the paper, it was very desirable to know the cost of working the various brakes, as a great deal must depend in the long run upon the first cost and the number of failures. He suggested that if some commission 622 EFFFCF OF RAILWVV BRAKES. OCT. 1878. could be appointed, or some other means devised whereby all the information on this head could be collected and brought before the members at some future period (as it appeared that Capt. Galton proposed to read a further paper on the subject), it would assist materially the object in view. At present engineers were all in the dark, and had nothing before them on which dependence could be placed to show which system of brake should be adopted. Mr, F. T. Haggard observed that in Table YIL, page 610, the Westinghouse brake was said to be applied to the first carriage in i^4 seconds, and to the thirteenth carriage in 3^2 seconds; whereas the Vacuum brake was said to be applied to the first carriage in 1 1 seconds, and to the seventh and thirteenth carriages in 1 4 seconds. If that were true, he hoped Capt. Galton would explain how, with such a difference in the times required for putting the brakes full on, stop No. 32 made by the Vacuum train on the second day (see Table VIII., page 612) could be reconciled with stop No. 25 made by the Westinghouse train on the first day : since on these occasions the two trains pulled up from 35 miles an hour in exactly the same time, viz., 12 seconds. There seemed therefore to be no difference practically in the length of the stops, although there was such a wonderful difference in the times taken to bring the brakes into operation. Even if stop No. 32 was doubtful as belonging to the second day, he might refer to stop No. 10 on the third day, as showing that there did not seem to be much discrepancy in the stops. Again, there was nothing said in the paper as to the pressure used ; he understood there had been 1 1 lb. vacuum in the Vacuum train, and 85 lb. pressure in the Westinghouse train. Surely those points ought to be taken into consideration from a scientific point of view, so as to ascertain what was the actual pressure on the wheels at each second, and what was the power used to put that pressure on. Mr. E. A. Cowper mentioned that various observations which he had himself made were in exact accord with the present experiments, which gave absolute facts, put in admirable form, and proved by considerable experience. Capt. Galton had very wisely S 4 Oct. 1S78. effect of railway hrakes. 623 thrown out an experiment here and there, where it was a faulty one. and had founded his tables on the average of a considerable number. The increase of friction between surfaces when moving very slowly, or only just beginning to move, might be particularly noticed with regard to the driving wheels of a locomotive : it was there so palpable that no one could overlook it. The driving wheels would pull the train from rest, and start it slowly ; but all of a sudden they would slip, and would then fly round at great speed, because the friction was thereby immediately reduced. Directly the rubbing surfaces were moving quickly over each other, they moved with much less friction. He had himself tried experiments on the amount of friction between dry wheels and rails, using 7 ft. wheels and pulling them until they slid upon the rails ; the friction then amounted to as much as 35 per cent, of the weight on the rails, which accorded with Capt. Galton's experiments ; but as soon as the wheels were sliding, the friction became much less. The theory of this he thought had hardly yet been explained, but yet was very simple. Every surface was rough, having a number of little roughnesses like hills and valleys. When two surfaces in contact were at rest, the hills and valleys fitted into one another, and if it was tried to draw one surface over the other it would be found necessary to lift the hills of one surface out of the valleys of the other, and make them run over the tops of the other hills. It was true that the tops of the hills were more easily knocked off at a high speed than at a low speed, but in any case they had to be knocked off ; whereas the surfaces could not fit into each other, and therefore had not the same adhesion one to the other, if they were moving quickly, as when stationary or only moving very slowly. Similarly in stepping upon a stone flag sprinkled with sand, as long as the foot was stationary there was considerable resistance to sliding : but when once the foot moved, the particles of sand ran over the stone, and the resistance decreased. Again, it was found (as Capt. Galton's diagrams showed) that a train was arrested with much more retarding force just at the moment of stopping, because the wheels seized the rails or the brake seized the wheels just as the two rubbing surfaces were coming to rest relatively to each other. 85 624 EFFECT OF RAILWAY 11RAKES. OCT. 1S7S. Mr. \Y. S. Hall remarked that, when a wheel skidded, the first result was that a very high temperature was communicated to the one point of the tire in contact with the rail ; and the second result was that a flat place was worn down on the tire, which made the wheel very liable to skid at that same place on the next occasion. If alternate heating and cooling was thus produced over and over again at two or three points of the tire, especially in frosty weather, the result with steel tires might be serious ; and it became a question whether it was desirable to use steel tires for braked wheels. If however these and other experiments led to the invention and adoption of a brake by which the maximum retarding pressure could be put on without skidding, then any such objection to steel tires would be removed. Mr. D. M. Veomans wished to correct a few slight errors made in the paper with respect to the Vacuum brake. With regard to the experiment made with a train of twenty carriages, it was only after the run on the second day of the experiments that he was informed that Capt. Galton wished to make this experiment, if some additional carriages could be found to make up the train of twenty, there being only twelve in the experimental Vacuum train. As he had not prepared any additional carriages and could not furnish them, some Great Northern carriages then at York were put into the train ; but although some of these were fitted in the manner in which trains were now fitted with the Vacuum brake, others were fitted in the old fashioned manner, and were not in a fit condition for making the experiment. However he made no objection to the trial, only withholding his own concurrence in it, because it had not been notified to him beforehand. In reference to the statement on p. 608, " the Westinghouse train was a train which had been working in regular traffic, and was run in practically the same condition as when running on the line, except that it was put in good order," he did not know exactly how long that train had been running with traffic, but it was not very long ; in any case it was fitted with the latest improvements of the Westinghouse brake ; in fact he had been informed by the officers of the North Eastern Railwav that the latest of all — what was termed the reducing Si, OCT. 1878. EFIKl'l' OF RAILWAY IIKVKKs. 625 valve — was applied only a few days before the trials. Therefore this was practically a new train, and the Vacuum Brake Co. had the same right to fit their train, which had been in service for eighteen months, with every new improvement or modification they had made ; no alteration however was made in the leverage upon engine or carriages. Xext he wished to refer to the further statement on p. 608, that " it was intended that the brakes should be so arranged that each brake-block, when in a state of inaction, should be removed a full quarter of an inch from the surface of the wheel, so as to ensure that there should be no dragging." Capt. Galton asked him to place the blocks at that distance, and he had complied with the request ; the blocks were placed as nearly as possible at that distance, were tested and found to be in a proper position ; and the statement that some of them were within 1 in., and some within t l in., of the wheel, he believed was not correct. It should have been stated that the question arose as to whether the blocks were dragging in the Vacuum train ; and that Capt. Galton then proposed that an experiment should be made by getting up a speed of 50 miles an hour, and then letting the train run until it stopped. That experiment was accordingly made, and the train ran 3J2 miles after shutting off steam, and was then still running at 15 miles an hour, when Capt. Galton requested that it should be stopped, as he was satisfied that there was very little dragging. With reference to Table VII., page 610, he enquired what was meant by " half on " and "full on : " — whether it was merely a question of the number of inches of vacuum at the moment. With the Vacuum brake, under ordinary conditions, the engine drivers were directed to use only 10 inches of vacuum. The train in question was arranged according to his calculations, so that this 10 inches of vacuum gave a pressure of 12,320 lb. on the wheels of each carriage. But of course in any particular experiment there might be only 6000 lb. or 5 in. of vacuum, if it were wanted to stop the train with that amount of power : and then the brake would really be "full on" for that occasion, whereas Capt. Galton he supposed would only call it "half on." In Table VIII., page 613, with reference to the second day's experiments with the Westinghouse 87 626 EFFECT OF RAILWAY BRAKES. OCT. 1 878. brake, the remark was, " Fine. Greasy rail." In the case of the two last stops so marked, however, the rails were dry and rusty ; he was himself upon the engine, and got down together with two or three others to examine the rails. Mr. A. Paget remarked that, on hearing the wish expressed by Mr. Sacre that some means could be found of collating and comparing the various experiments which had been made with different brakes, and the information available as to their durability and cost of maintenance, he had felt inclined to rise and to point out that this was a function which certainly came within the scope of the Institution. But he confessed that, after listening to the present discussion, and observing the difficulties that arose from the animosity with which the advocates of any particular brake seemed to regard even any mere abstract fact that might possibly tell against it, he did not feel inclined to propose that the Institution should undertake such a task, or should request any one else to undertake it. The Institution owed a great deal to Capt. Galton for having, as regarded the purely scientific part of the question, thrown his energies and his great care and experience into the subject. He had done incalculable good ; and an)' hard facts which might show that, in one point or another, one brake was better than another brake, should be considered quietlv, always remembering that Capt. Galton had treated the subject, not as a question of comparison between rival brakes, but for the purpose of ascertaining facts in the interests of science. Dr. J. Hopkjxmjx wished to suggest a very probable partial cause of the smaller amount of friction between two surfaces when they were in motion relatively to each other, than when they were at rest. This appeared to him to be due to the fact that, when two surfaces were in contact, their pressure upon each other, while at rest, squeezed out the whole of the air from between them, so that they were in real, not merely nominal, contact. That being so, the coefficient obtained would be the true coefficient of the friction between the two surfaces. But when they were in motion at a high speed, that was no longer the case : the probability was that during the whole SS < >CT. iS-S. K F I- K- I < - < : I- 5 O ? it s LU Plate Si. EFFECT OF RAILWAY BRAKES. Plate 85. Indicator Diagrams from Brake K.eperi inputs. / v p. ttr/. in. _ 150-i 110- ! Fig. 3. Experiment No. 10. J 130- 1 Tfnifm v m Speed of Braked Wheels, 120- Miles per hour t-60 / and Uniform Pressure on Brake-Blocks. 1 Gradient rising 1 in 17C 110- -50 S i 100- ____S 90- so- 70- \ 00- < / 10 // If 50- 10- -30 1 .'l F 30- . /'. 30 _T Ji 111 -1" M_ r" Miles i perJiouv Fig 1 . 4. Experiment JVo, 11 30 Sec. •- ^ TJju. B0-! r ..■I TV N \ w • Ji Uniform Speed of Braked Wlieels, and Uniform Pressure on Brake -Blocks. Gradient level. \\ 10- -in Mile* per hour 1 /..■■-*■ Fig. 5. Ejrprrimrut No. IS. v \ f \ I \i p - \ TJis. '--—3- i 30 j: - i: 10 i' Uniform Speed of Braked WJteels, and Varying Pressure on Brake-Blocks. Gradient rising 1 in 220. 50- 40- 30 20 10- P Pressure on Brake-Blocks, in lbs. T Traction upon Draw-Bar, in lbs. F Friction between Brake-Blocks and Wheels. S Speed of Braked Wheels, miles p. hour. (Proceedings Inst. M.JS. 1878.) EFFECT OF RAILWAY BRAKES Indicator Diagrams from Brake JS.rperimentti, Fig". 6. Hxper 'intent No. 13. Plate 86. Eig. 8. JZjperi merit No, 15, Brake-Blocks of small area. P Pressure on Brake -Blocks, in lbs, T Traction ttpon Draw-Par, in lbs F Friction between Brake- Blocks and Wlieels. S Speed of Braked WJieelsj miles p. (Proceedings Inst. M.E. 1878.) Sec. Iioav, Lbs. p. sq. in, 110 iao uo 100 90 so -ro 60 50 -to 30 EFFECT OF RAILWAY BRAKES. Mate 87. Indicator Diagrams from JBrake Exjieriments. Fig. 10. Experiment No. 17. Slip Stop without Skidding. Gradient falling 1 in 1056. Fig. 9. JlSxperiment No. 1G, Van Starting from rest. Sand on rails. Miles per hour. 1-60 A n f\ w I I I I I I Lbs. p. sq. in. 80-1 -50 / r- — --... *-*.? 70- / ""•^.■i— 60- 10 'E : . 50 10- -30 30- 20- 20 1 \s 10- 10 =o-i — i— !■■ -i o- A^~v ^SP I I ' p ^^(N i i / T / Xbs. p. sq. in 50' 10- 30 ^^^- / , 10 15 ill Fig. 11. Experiment Xo, IS. Slip Stop with Skidding. Gradient level. lbs. p. sq. in. 100-1 ^Z*. 60- 10- 30- 20- 10- F,' * ' ' ' ' 26 ' ' ' h dec' ' "" T Traction upon Draw-Mar. in lbs. S Speed of JBraked Wheels, in miles per liour. P "Pressure on. Urake- Blocks, in lbs. F Friction between Braise- Blocks and Wheels. (Proceedings Inst. M.JS. X878.) EFFECT OF RAILWAY BRAKES. Indicator Diagrams t\ from Drake Kjt peri meats. Fig- 13. y' r Experiment -Xo, ID v .'*^\ / s' 1 y T 1 .' I: 1 |: per hou (Proceedings Inst. 3I.TS. 1H78.) ia ' ' ' as 1 ' ' ' 5 sleo FFTSra in preceding Plate 87. Miles per hjo.it )• (-35 Fig. 15. Experiment- No. 22. (1 ^ Slip Stop in 151 yards. Httff.r\ T-KL't <""/■. p. sq. in -110 / /p.. 110- Lhs. -100 / / / / Pig. 17. Experiment Xo. 24. 100- -00 / / / .'/ Smith-Hixvdy^Vacutnii Bralte 90- -so / 1 _,_■ 80- / 1 / ■'/ .^ — -70 / 1 \ ^" P 21 70- / I J --*' -60 / * ' 1: A-' " '! 60- -50 p,/ / •'" , '' 50- .•/ pV yl /' -10 ••/ / / !'! 10- -30 -20 J / i ! 1 i 1 30- 20- / 'Ex/' 1 10- 1 -0 r ' ' -i — r — i^i — i i i ' ' i ' x i i i 5 l!o 15 1 u'o 1 ' ' ' h ' 3*0 See. Fig. 18. Experiment No. 25. Wesfingliouse UraTce. Mbs, : in. 80-i P Pressure on Erahe- Blocks ', in lbs. {Proceedings Inst. 3I.E. 1878.) XJie figures suffixed to fJie letters P and F t •toi K'hjs. Yi and 18 denote tlie-place of tin- Experimental Van in the train, from, the engine. F Friction between Brake-Blocks and Wheels.. {3 Speed of Braized Wheels, in miles per Jiour. ON RECENT BRAKE EXPERIMENTS UPON THE LYONS RAILWAY liY M. GEORGE MARIE, MEM. INST. M. E. AND ON THE EFFECT OF BRAKES UPON RAILWAY TRAINS (THIRD PAPER) BY CAPTAIN DOUGLAS GALTON, C.B., HON. D.C.L., F.R.S., MEM. INST. M. E. EXCERPT MINUTES OF PROCEEDINGS OF THE MEETING OF THE INSTITUTION OF MECHANICAL ENGINEERS IN LONDON, 24th APRIL, 1879 JOHN ROBINSON, ESQ., PRESIDENT IN THE CHAIR BY AUTHORITY OF THE COUNCIL Inst. M. E., 10 Victoria Chamisers, Victoria St., Westminster April 1879. '57 ON RECENT BRAKE EXPERIMENTS UPON THE LYONS RAILWAY. Iiv M. GEORGE MARIE, of Paris. I. Particulars of Apparatus. The Paris and Lyons Railway has lately experimented on two trains, one fitted with the Westinghouse brake, and one with the Smith Yacuum brake. Both trains were alike, and were composed as follows : — Vehicle. Axles. Load per Axle. Lbs. Total Load. Lbs. I engine, 4 axles (in work- ing order) j 2 driving axles, together j 2 loose axles, together 55-125 > 44,100 j 99,225 t tender, 3 axles (in work- ing order) I 1st axle, with brake < 2nd axle, without brake ' 3rd axle, with brake 25.357 j 17,640 < 25.357 ) 68,354 25 carriages (empty), 3 axles C 1st axle, with brake ' 2nd axle, without brake ' 3rd axle, with brake 7,276 ) 6,394 20,946 7,276 ) Same carriages (loaded), 3 axles ( 1st axle, with brake ' 2nd axle, without brake ( 3rd axle, with brake 8,676 j 7,594 1 24,946 8,676 ) The engine is not fitted with any brake, except the " Le Chatelier " counterpressure apparatus. This apparatus,* as improved by the chief engineer, M. Marie, has worked excellently for fifteen years, in daily use, both in all stops at stations, and also on long inclines. Some remarks will be made hereafter on the power of the counterpressure apparatus compared with a brake on the driving wheels. * See Proceedings 1870, p. 21. i 5 S 3RAKE EXPERIMENTS, LYONS RAILWAY. April 1879 Each carriage had only two axles fitted with brakes, the middle axle being left free ; the fear of the complication arising from a six-wheel brake was the reason of this ; but, in order to have power enough, a great brake-block pressure was provided. The following Table shows the brake-block pressure on every axle of the train. The pressure in the brake cylinder is here supposed to be 37 lbs. per square inch; and the vacuum in the " Hardy " sack to be 16 inches of mercury. These pressures must be considered as an average of the ordinary pressures in practice. Air Pressure, or Vacuum. Pressure on the Outside Blocks. Pressure on the Inside Blocks. Westing house. Tender (est and 3rd axle) Empty carriage {1st and 31-d axle braked) Lbs. per Sq. In. 37 37 37 Inches Vacuum. 16 16 16 Lbs. 6,666 4,I9 T 4.I9 1 Lbs. 15,020 8,342 8,342 Lbs. IO,Ol6 6,3°3 6,3°3 Vacuum. Tender (1st and 3rd axle) Empty carriage (1st and 3rd axle braked) Lbs. 9,874 5-°45 5>°45 Total Brake- block Pressure per Axle. Load on Rail for each Braked Axle. Lbs. Lbs. 16,682 10,494 10,494 25,357 7,276 8,676 Lbs. Lbs. 24,894 13,387 13,387 25,357 7,276 8,676 Ratio of Brake- block Pressure to Load. Per Cent. 66 144 121 IS4 ■54 The pressures were not large enough to skid the wheels at high speeds, but at low speeds the wheels were skidded, especially with empty carriages. This latter is no real disadvantage, because in ordinary stops the driver can moderate the brake power. Mr. Westinghouse however, in order to avoid skidding under any circumstances, fitted, as a trial, twelve carriages with reducing valves, operated by the friction of the brake-blocks. The arrangement of these was the same as is given in Captain Galton's paper (Figs. 1-3, Plates 19 and 20). It will be seen that, in the case of the Westinghouse brake, the pressure on the outside blocks (or those furthest from the centre of the vehicle) is less than on the inside blocks, which arrangement has the disadvantage of creating a strain on the horn plates, equal to the difference of these two pressures ; but has, on the other hand, the following advantage. When the brake is applied to the whole train, each carriage is subject to a strain, created by inertia, equal to the Al'RII. 1879. liRAKE EXI'ERIMEM'S, LYONS RAILWAY. 1 59 mass of the carriage, multiplied by the rate of retardation of the train. This strain equals, on an average, about 1 , , r of the weight of the carriage ; it acts as a forward push applied to the centre of gravity of the carriage, and changes the distribution of the weight on the axles, the front axle becoming more loaded than usual, and the rear axle less. This is an important matter, because the springs of the carriages are sometimes damaged by the action of this extra load on the front axle. Mr. Westinghouse has therefore arranged his brake gear in such a way as to create a vertical force, counteracting partially this extra load, by giving more pressure on the inside blocks than on the outside ones. These pressures being different, the frictional resistances are different also, and the difference forms a vertical force acting upwards on the springs of the front axle, thereby balancing so far the action of inertia. In the case of the Vacuum brake gear the pressure on the outside blocks is greater than on the inside ones, which causes a reverse action, so as to add to the extra load created by inertia upon the front axle. From this it follows that the compression of the front springs is somewhat greater in the case of the Vacuum than of the Westinghouse brake. In the Westinghouse break gear this difference of pressures is equal to 6303 — 419 1 = 2112 lbs. for each axle ; and the difference of frictional resistances is equal to "- - = 422 lbs. (with a coefficient of friction of .20). In the Vacuum brake gear the difference of pressures is equal to 8342 — 5°45 = 3 2 97 lbs. ; the difference of resistances is therefore equal to = 659 lbs. Thus with the Westinghouse brake there is a counteracting force, as described, of 420 lbs. ; whilst with the Vacuum brake there is an augmenting force of 660 lbs. From this cause the front springs are more loaded with the Vacuum brake than with the Westinghouse ; and this strain is still greater towards the end of the stop, because the friction increases considerably at that moment, as may be seen in the friction diagrams of the Brighton experiments. The practical result of this during the trials was that in the Westinghouse train only one front spring was flattened down, while in the Vacaum train all were. The author does not understand the object of the latter type of brake gear, which is generally applied with the Smith brake, and is also extensively used by the Great Northern Railway in England. "3 l6o DRAKE EXPERIMENTS, LYONS RAILWAY. Al'RIL 1S79 It is impossible to avoid entirely the above-mentioned strain on the horn plates ; and in addition, the horizontal force retarding the train must produce another strain on the horn plates, which is equal to _ .494 __ 2Q g^ lbs. per axle for the Westinghouse brake. Thus with this brake the horn plates of the front axle have to support two strains in opposite directions, one equal to the difference of the brake- block pressures, or 21 12 lbs., and one equal to the retarding force, or 2099 lbs. ; the resultant strain is therefore only 21 12 — 2099 = 13 lbs. But on the horn plates of the rear axle the forces act in the same direction, and therefore the total strain is equal to 2 112 ~\- 2099 = 42 1 1 lbs. From this point of view it would be better to have the same brake-block pressures on both sides of the wheel, so that the strain on the horn plates should be only 2099 lbs. at each end of the carriage ; the advantage of unequal pressures, as above described, would then be lost, but still the author would prefer this arrangement. The stroke of the piston, both in the air cylinders and in the Hardy sacks, was calculated so as to have a theoretical clearance of 3/4 inch between the blocks and the wheels, when the brakes were off. In the trials, the distance between each block and its wheel was 3/ 8 inch when the brakes were off ; the difference, 3/£ inch, was allowed for the bending and springing of the brake gear. II. Practical Working of the two Brakes. Trains fitted with a Westinghouse and with a Vacuum brake have been running for two months from Paris to Montereau (49 miles) and back, and from Paris to Corbeil (20 miles) and back. The Westinghouse brake has worked very well when there has been no leakage in the pipes, but it is necessary to watch all the apparatus with the greatest attention ; sometimes one or two of the triple valves get out of order, and disturb the action of the brake, causing very severe shocks to the couplings, especially with trains of twenty-four carriages. The Vacuum train has also given good results, but the couplings are often damaged ; the operation of putting the couplings together is much more difficult than with the Westinghouse brake. With both brakes the passengers generally feel no shock, 114 April 1879. brake experiments, lyons railway. 101 provided the driver releases the brake a few yards before the end of the stop. The practical trial of both these trains has as yet been too short for giving any definite conclusion. III. Experiments of the ist and 2Ni> April. The London Brighton and South Coast Railway, as represented by their general manager, Mr. Knight, and locomotive superintendent, Mr. Stroudley, sent over to the Lyons Railway the experimental van already used during Captain Douglas Galton's experiments at Brighton and York. The author takes this opportunity of thanking those gentlemen, and also Captain Galton, Mr. Westinghouse, and Mr. Kapteyn, of Paris, for their assistance in these experiments. A description of the apparatus has already been given in Captain Galton's papers on the subject. In the present experiments two diagrams only were taken for each stop ; first, the diagram from the speed indicator, giving the square of the speed of the train at each point of the distance run ; second, the diagram giving the brake-block pressure on the front axle of the van. This axle had been fitted with a Westinghouse brake and with a Vacuum brake, which could be used independently. In Captain Galton's experiments the abscissae were made proportional to the time, the recording cylinders being connected with a water clock, and so turned round with a uniform speed. In the Lyons experiments, the abscissa- were made proportional to the distance. run by the train ; for this purpose the recording cylinders were connected with the rear axle of the van ; that axle had no brake, and therefore gave a distance exactly proportional to the distance run by the train. When the train was run without brakes being applied, this connection between the recording cylinders and the rear axle was broken : when the driver applied the brake, the connection was immediately and automatically established by means of friction gearing, actuated by electricity. The brake-block pressure is exactly proportional to the air pressure in the brake cylinder, or to the vacuum in the Hardy sack. On the diagrams, Plate 18, are two scales, adapted for measuring the air pressure and the vacuum from the diagram of the brake- block pressures. The lines marked W denote the Westinghouse brake, and the lines marked V the Vacuum. The full lines show "5 I 62 BRAKE EXPERIMENTS, LYONS RAILWAY. April 1879. the speeds, and the dotted lines give the air pressure in the brake cylinder in lbs. per square inch, or the vacuum in the Hardy sack in inches of mercury. It will be seen that in Captain Galton's diagrams the lines of speed are convex towards the base line, while in these they are concave ; the reason of this is the difference already explained between the abscissa; used in each case. On 1 st April experiments were made with the Vacuum brake. From Paris to Montargis (73 miles) the train was composed of engine, tender, ordinary van, experimental van, twenty-two carriages. From Montargis to Paris the train was much shorter, comprising engine, tender, ordinary van, experimental van, six carriages. On 2nd April the same experiments were made with the Westinghouse brake, and with the same number of vehicles. A great many diagrams were taken, from which five are chosen, representing the average of all the cases. By the aid of those five diagrams, Figs. 1-5, Plate 18, the results may now be described. Fig. 1 represents two stops with twenty-four carriages (including the two vans), one stop with the Westinghouse, the other with the Vacuum brake. The weights of the trains were as follows : — Weight on Unbraked Weight on Braked Wheels. Wheels. Lbs. Lbs. Engine 99-2-5 Tender 17.040 5 .7i4 Carriages and van, empty 147,062 334>696 Experimental van, loaded 1 0,000 10,400 Total 273,927 395,810 Total weight of the train, 669,737 lbs. ; proportion on braked wheels, 59 per cent. In order to be exact there ought to be added to the total weight of the train the weight of all the unbraked wheels, because the brake is obliged to destroy the rotating momentum of those wheels. But, as this error is very small, and the same for both brakes, it has not 116 April 1S79. BRAKE EXPERIMENTS, LYONS RAILWAY. '63 been taken into consideration in the calculations. The results of Fiir. 1 are therefore as follows : — S !; eud " f Length of 1 ra,n Stop. Miles per yrf Hour. Westinghouse Vacuum . . 35 217 2S0 Duration of Stop. 20| -Mi Gradient. ( Rising I I 1 in 200 ( Level State of Rail. Dry. Rather wet. The diagram shows the variation of the air pressure in the brake cylinder, and of the vacuum in the Hardy sack. The full pressure was almost instantaneously applied with the Westinghouse brake ; its variation shows the action of the regulating valve in the van. The vacuum increases slowly, and reaches 13 }j inches at the end of the stop ; in the rear of the train the vacuum was no doubt created more slowly still, but no experiments with recording apparatus were made to ascertain this point. This trial was the most unfavorable for the Vacuum brake, the number of carriages being large and the speed small. Figs. 2 and 3 show a few stops with the return train of eight carriages (including the two vans). The weights of the train were as follows : — Engine Tender Seven carriages, empty Experimental van, loaded Weight on Unbraked Wheels. Weight on Braked Wheels. Lbs. Lbs. 99,225 17,640 S vI4 44.7 5 8 101,864 10,000 IO.4OO Total 171,623 162,97s Total weight of the train, 334,601 lbs. ; proportion on braked wheels, 49 per cent. 117 i6 4 11 RAKE EXPERIMENTS, LYONS RAILWAY. April 1879. The particulars of the stops were as follows :- Brake. Speed. Miles per Hour. Length of Stop. Yards. Duration of Stop. Seconds. "Westinghouse, Fig. 2 Vacuum, Fig. 2 63} 6o£ 55° 57° 32 37 Gradient. Falling j 1 in 200 ( Do. State of Rail. YVestinghouse, Fig. 3 I Vacuum, Fig. 3 37 37 203 2I 3 19! ( Falling j ) 1 in 1000 \ Do. Dry. Quite wet. Dry. Quite wet. The two first stops (Fig. 2 ) were made with the aid of the counterpressure of the engine ; the two last ( Fig. 3 ) were made without counterpressure. The line C of Fig. 3 shows a stop made with counterpressure alone. As may be seen, the results are not very different, the trains having very few carriages. Fig. 4 shows a stop obtained with the Westinghouse brake, worked from the rear van of the train. The following are the particulars : — Speed, 40 miles per hour. Length of stop, 248 yards. Time, not observed. Gradient, level. Rail, dry. The train had exactly the same composition as that in the experiments with eight vehicles ; the regulator of the engine was kept full open until the train had come to rest. Fig. 5 shows a slip experiment with the 'Westinghouse brake. The part of the train which was slipped was composed of the experimental van and six carriages. The weights were as follows : — Experimental van, loaded Six carriages, empty Total Weight on Unbraked Wheels. Lbs Weight on Braked Wheels. Lbs. 10,000 38,364 10,400 S7,3 12 48,364 97,712 April 1879. brake experiments, lyons railway. 165 Total weight of the train, 146,076 lbs.; proportion on baked wheels, 67 per cent. The following are the particulars of the stop, which was a very good one : — Speed, 41 miles per hour. Length of stop, 150 yards. Duration of stop, 12 l A seconds. Gradient, level. Rail, dry. In this stop, as in the last, the recording instruments were started by the air pressure in the brake cylinder of the van, instead of being started by electricity as before. There was in consequence a slight loss of time, especially in the stop from the rear van, Fig. 4. IV. Comparison of the Retarding Forces in different Stops. It is easy to compare stops with each other, although the speed, the gradient, and the proportion of weight on braked wheels be different in each. This can be done by calculating the retarding forces of the brake for each stop. If we called S the speed in miles per hour, / the length of the stop in yards, and \V the weight of the train, then the retarding force F is given by the formula : — w = 0.0 1 1 x ~t~- The diagrams give with precision the values of S and / in each stop ; hence it is easy to calculate the retarding force in each case. On the diagrams, the ordinates are proportional to the square of the *Let s be the speed of the train in yards per second, t the number of seconds of the stop, and / the length of the stop in yards. Then, if we suppose the retardation /• to be constant during the stop, s = kt. Also / = - kf l ; hence /= -k — = ; thus, if we take the value of /• from this formula, we 2 k* 2 t have /• = — 2/ Now let g be the acceleration of gravity in yards per second ; let F be the retarding force of the brake in the whole train, and let W be the total weight of the train. Then, because forces are proportional to their accelerations, we have : — ¥ k s- . , S '* — = - = = 0.011 X — . W g 2/g I II 9 1 66 BRAKE EXPERIMENTS, LYONS RAILWAY. April 1879. speed, and the abscissa; proportional to the length of stop ; thus the average retarding force is proportional to the inclination of the straight line drawn between the extremities of the speed curve, while the exact retarding force at each point of the diagram is given by the inclination of the tangent to the curve at that point. We can now compare stops made at different speeds. To compare stops made on different gradients, we must add the F gradient to - , with the sign -f- if it is falling, and with the sign — Y' if it is rising. We will call - this corrected retarding force. To compare stops in which the percentage of weight on braked P" wheels is different, let A be that percentage. If we multiply - - by - , we have a number . which would be the retarding force for a train stopped on a level line with the whole weight of the train braked. F" If we calculate the values of — for all the stops, the numbers obtained compare exactly the powers of the brakes, whatever may be the circumstances of the stop. The following Table gives this comparison for the present diagrams : — Percentage F" (See Plate iS.) Speed s. 37 Length 217 F VV Gradient. F' w 0.070 of Weight on Braked Wheels. 59 W Power of the Brake. Westinghouse, alone : 24 carriages, Fig, 1 0.075 ( Rising 1 i 1 in 200 1 0.119 Vacuum, alone; 24 car- riages, Fig. 1 35 2 So 0.057 Level 0.057 59 0.096 Westinghouse, alone: Slip experiment, Fig. 5 4' 150 0.130 Do. 0-13° 67 0.200 Wes-iinghouse, alone: Brake from rear van, Fig. 4 40 24S 0.0S0 Do. 0.080 49 0. 160 Westinghouse, alone : S car- riages, Fig. 3 Vacuum, alone : 8 carriages, 37 203 0.0S5 1 Falling j 1 1 in 1000 | 0.086 49 0. 170 Fig- 3 37 213 0.0S0 Do. 0.081 49 0. 160 Comiterpressure, alone : S carriages, Fig. 3 35 5S3 0.024 Level 0.024 16 0. 150 The stops in Fig. 2 have not been worked out because they made with the brake and the counterpressure together. April 1S79. brake experiments, lvons kaiiway. 167 The retarding; force ought to be diminished by the ordinary resistances of the train, that is to say, by 0.003 to 0.008. This makes but a small difference when the proportion of braked weight is large ; but in the last stop, with counterpressure alone, this F" correction gives to a value 0.125 instead of 0.150. V. General Conclusions. We have seen that the value of is the best comparison for the power of the brakes in all cases. Now with the Westinghou.se F" brake -— is generally between 0.120 and 0.200, while with the F" Vacuum brake is generally between 0.100 and 0.160. This shows a slight advantage for the Westinghouse brake, but it must be remarked that the rail was better during the trial of the Westinghouse than of the Vacuum brake. We have also seen that the value of F" -^7- for the counterpressure is 0.125, while the average of the Westinghouse trials gives 0.160, with a brake-block pressure of about 140 per cent, of the weight on the braked wheels. Thus the counterpressure on the driving wheels is as powerful as a Westinghouse brake which has a pressure of 140 x -^-- = 109 per cent, of the weight on those wheels. It may be suggested that in a quick stop the time occupied in reversing the steam on the engine must practically diminish the power of the counterpressure ; this is true, but on the other hand the brake-block pressure on the driving wheels with the Westinghouse brake is generally less than 109 per cent., and often less than 80 per cent. Thus the counterpressure is as powerful as the Westinghouse brakes generally in use on the driving wheels. On the Lyons railway the drivers have been accustomed to the counterpressure for fifteen years ; hence, whatever may be the final choice of brake, the Company will put no brake on the driving wheels. The diagrams show that the Westinghouse brake comes on almost instantly over the whole train ; it is not the same with the Vacuum brake. There the brake goes on first upon the front carriages, and l68 BRAKE EXPERIMENTS, LYONS RAILWAY. APRIL 1879. the buffers are consequently compressed ; afterwards the carriages all become braked, and the buffers return to their first position, but not without oscillations which are serious, both for the passengers and for the couplings themselves. In the Westinghouse brake this action is much slighter. In all the trials the wheels skidded only at a very low speed; hence the regulating valves did not show any great advantage in the length of the stop. They save the tires by preventing the wheels from skidding, and they save also the springs of the carriages by diminishing the shock felt at the last moment of a stop. Though a very ingenious improvement, yet, if the driver can really graduate the strength of the brake, it may be doubtful whether they are necessary in continuous brakes ; thus one more complication may perhaps be avoided. The Lyons Railway had never previously had an accident caused by the breaking of draw-bars in a passenger train, but several were broken in the trains fitted with the Westinghouse and Vacuum brakes. Thus, if there has been no such breaking of draw-bars in past times, yet these are almost sure to occur frequently with the use of continuous brakes : and as there are on the railway several very steep gradients, it would seem that a non-automatic brake would on this line at least be dangerous, unless, to avoid accidents, the ordinary hand brakes are still kept in use on the trains. The arrangements of the Westinghouse brake are certainly rather complicated, especially the air pump and the triple valves. As regards the pump the author thinks that it would be difficult to make an air pump more simple and at the same time able to satisfy all conditions. And as regards the triple valve, he thinks it is impossible to simplify it without losing one of the two following advantages — the quickness of application of the brake, or its graduation in the stops. It might however be rendered less delicate by making it larger and stronger. All the other parts of the brake are quite satisfactory. All the parts of the Vacuum brake are very simple and strong, but the leather of the Hardy sack seems to be a bad material for practical work. Some arrangement of metallic pistons and cylinders April 1879. brake experiments, lyons railway. 169 would seem preferable, but practice alone can decide this question. The arrangement of the brake gear is not quite satisfactory, but it would be easy to improve it. The above opinions must be taken as those of the author alone, and as requiring to be checked by longer experience. Both brakes will be tried on long inclines, and a uniform speed maintained with them if possible. Until these trials have taken place, no decisive opinion can be given by the railway company. I2 3 170 April 1S79. OX THE EFFECT OF BRAKES UPON RAILWAY TRAINS. (Third Paper.) By Caitain DOUGLAS GALTON, C.B., Hon. D.C.L., F.R.S. In the previous papers upon this subject which the author brought to the notice of the members of the Institution in June and October 1878, it has been explained that it was anticipated, when the experiments were begun on the Brighton Railway with the van fitted with the self-recording apparatus, that the results would enable a rule to be laid down determining the amount of brake-block pressure, in proportion to the weight of the vehicles, which should be applied to the wheels of railway trains. Effect of Skidding. It was immediately discovered, however, that the retarding effect of a wheel sliding upon a rail was much less than when braked with such a force as would just allow it to continue to revolve. The annexed copies of two sets of diagrams, Figs. 4 and 5, Plate 21, taken during the experiments, show, more clearly than can be explained, the difference in the retarding force, before the wheels begin to slide upon the rails, and after. These two experiments were made with a single van slipped from the engine, the brakes going on automatically at the moment when separation from the engine took place. These and all the other diagrams illustrating this paper are similar to those given in the second Paper (Proceedings 1878, p. 590). The line SS shows the speed of the braked wheels at each instant, the scale for which is given in miles per hour. The line PP represents the pressure against four blocks acting upon one pair of wheels: the vertical height of P on the scale headed "force," when multiplied by 240, gives the total pressure in pounds on the four blocks collectively. The line FF shows the retarding effect of the four blocks upon the one pair of wheels before the wheels began to slide upon the rails ; and ff shows the effect while the wheels were sliding upon the rails. The vertical height of F or / on 124 April 1S79. effkct ok railway brakes. 171 the scale of force, multiplied by 60, gives the total retardation in pounds upon that pair of wheels. The calculations show that the friction between the wheel and the rail, when the wheel is sliding on the rail, is less than one-third of the friction produced between the brake-blocks and the wheel, when the brake-blocks are so applied as to allow the wheel to continue revolving. Coefficient of Friction as affected by Speed. The next important discovery was, that the coefficient of friction between the brake-blocks and the wheels varied inversely according to the speed of the train, a higher proportion of brake-block pressure to weight being required at high speeds, and a lower pressure for a lower speed. This was illustrated by the diagrams shown in Figs. 3, 4, 6, 7 of the second Paper (Proceedings 1878, Plates 85, 86). Table I. given in the second Paper (Proceedings 1878, p. 599) for the values of this coefficient at different speeds has since then been somewhat altered by obtaining a mean from a considerably larger number of experiments. The accompanying Table IX. (p. 172) shows the altered result. Coefficient of Friction as affected by Time. If the friction of the brake-blocks were always the same at the same speed, some simple rule might still be deduced which would give the pressure required at each speed for obtaining a certain amount of retardation ; but when the speed of the van was kept nearly uniform by the effort of the engine, the friction of the blocks decreased ; and this occurred notwithstanding a continued increase of the brake-block pressure : showing that, through some cause not yet fully determined, the holding power of brake-blocks at all speeds is considerably less after some seconds of application than when first applied. This peculiarity is illustrated by Figs. 3, 4, 5 of the second Paper (Proceedings 1878, Plate 85). Hence the question of the proper amount of brake force needed at each instant, during the time required to stop a train, is still further complicated by this decrease which occurs in the coefficient of friction after the brakes have been applied, and which results from the time during which they are kept applied, irrespective '-5 172 EFFECT OF RAILWAY BRAKES. April 1879. Taisle IX. — Coefficients of Friction at Varying Speeds. Cast-Iron Brake-Blocks on Steel Tires. VELOCITY. JOEFFICIENT OF FRICTION. Number of Experiments from which the ..., 1 F f Extremes Observed. Mean Is Taken. M H"P" ' If*^/ Mean. Ma ximum. 123 Minimum. 12 60 88 .058 074 67 55 1 8l 136 .060 1 1 1 55 50 73 '53 .050 116 77 45 66 179 .os 3 127 70 4° ; 59 i 194 .088 140 So 35 5' !97 .0S7 142 94 3° 44 196 .09s 164 70 25 36* 205 .108 166 6q 20 29 240 ■^3 192 7 S , 15 22 2S0 ■'3 r 223 54 J o Hi 281 .161 242 28 7* 3^5 • I2 3 244 20 Under 5 Under 7 340 .156 2 73 Just moving. Just moving. 33° Fleeming Jenkin (steel on steel .0002 to .00S6 337 •365 35 1 Rennie. Static Friction under 1 Pressure of 180 lbs. per square inch 300 Pressure of ^^6 lbs. per S( )uare inch 347 of any change in speed. This decrease in the coefficient of friction was shown in Table I. of the second Paper (p. 599), from which the following figures are taken : Table X. — Coeff ICIENT OF Friction as Affected by T IME. SPEED. COEFFICIENT OF FRICTION. Miles Commencement ' of Experiment.* , After After After After per Hour. 5 Seconds. 10 Seconds. 15 Seconds. 20 Seconds 20 1S2 .152 •133 .116 .099 27 .171 .130 .119 .081 .072 37 ■'5- .096 •0S 3 .069 47 •J3 2 .080 .070 60 .072 .063 .058 * The figures in this column are somewhat different from those that have just been given in the altered Table IX., because they resulted from the average of fewer experiments ; but the effect of time in reducing the coefficient of friction may be accepted as correct. 126 April 1S71). ei-tkct of railway brakes. 173 The decrease in the coefficient of friction arising from time sometimes overcomes the increase in the coefficient of friction arising from a decrease in speed, especially when, either from the stop being on a descending gradient or from a small proportion of the train only being fitted with brake power, the train takes a considerable time in coming to rest. Therefore a higher brake pressure is required in such cases than when the stop is made in a short time. The accompanying diagram, Fig. 6, Plate 21, shows a uniform coefficient of friction with a practically uniform speed, as obtained by means of an increasing brake-block pressure. The line P P shows the pressure, F F the friction (the coefficient of which is given for several ordinates), and S S the speed, which latter decreased slightly during the experiment, and would have caused an increase in the coefficient of friction had it not been counteracted by the element of time. Coefficient of Friction affected by Material ami Weather. On experimenting further, it was found that the coefficient of friction was also influenced by the kind of metal in the blocks, and by the state of the weather. Experiments on a Train cannot be free from Disturbing Elements. The experiments were made upon trains travelling under conditions which of necessity were continually varying, and therefore presented many elements of disturbance beyond the reach of calculation. The time during which the pressure was applied has been shown to enter largely into the question ; and this element of disturbance is in operation during the interval of time (however short), which necessarily occurs between the moment when the block first touches the wheel and the moment when the full pressure is obtained. Under these circumstances the author has limited himself to stating the facts obtained in the experiments, and has refrained from endeavouring to lay down the law of decrease in the coefficient of friction according to velocity ; as he believes that any law which could be laid down would only tend to mislead, owing to the continually varying conditions which occur in practice. The Institution of Mechanical Engineers has decided to carry on 127 174 EFFECT UF RAILWAY ]IR\KES. APRIL 1S79. through the medium of their Research Committee further experiments on this very interesting subject ; and if these new experiments can be arranged so as to be free from the disturbing elements incidental to those which he has had the opportunity of making, the author trusts that the question will soon receive fuller elucidation. Adhesion as affecting the Maximum Retardation. Notwithstanding the variations in the coefficient of friction between the blocks and wheels, it was found that under similar circumstances the adhesion of the wheels to the rails was practically constant, but that it varied according to the material — that is, whether the train was travelling upon iron or steel rails ; and according to the state of the rails, whether dry, wet, or sanded. ( )n dry rails it was found that the coefficient of adhesion of the wheels was generally over .20. In some cases it rose to .25 or even higher. On wet or greasv rails, without sand, it fell as low as .15 in one experiment, but averaged about .18. With the use of sand on wet rails it was above .20 at all times; and when the sand was applied at the moment of starting, so that the wind of the rotating wheels did not tend to blow it away, it rose up to .35, and even above .40. The retarding force which causes a train to be stopped by the application of brakes is limited to the adhesion or resistance obtained between the wheel and the rail ; therefore the greatest effect in stopping a train is produced when the friction between the brake- blocks and the wheel amounts to a quantity just short of the adhesion ; because as soon as the brake-block friction exceeds the adhesion, the wheel becomes fixed and begins to slide. If a certain amount of brake force or brake-block pressure would produce an equal amount of friction at all speeds, then the greatest possible amount of retardation during the time required to make a stop could easily be obtained ; but, as has already been proved, the brake pressure at high speeds must be much greater than at low speeds, in order to produce the same amount of retardation. The friction increases as the speed decreases, according to some law, which is complicated, as shown above, by the time during which the brakes have been applied. If therefore the pressure necessary to cause sufficient friction to balance the adhesion at a high speed be 128 April 1S79. Ei-i-Kcr of railway brakes. 175 continued till the train comes to rest, the friction will increase and gradually overcome the adhesion, and the wheels will become fixed. This is illustrated by Fig. 6, Plate 86 (Proceedings 1878). Therefore, in order to secure the best results in stopping, it is obviously necessary that the brake-block pressure should be so regulated as to give a friction about equal to the adhesion of the wheels at every stage during the process of stopping. Regulator for Brake-block Pressure. In some of the former experiments on the Brighton Railway a pressure-reducing valve was introduced by Mr. Westinghouse, with the view of reducing the brake-block pressure as the speed was reduced ; and excellent results were obtained. And it will be recollected that the author, in his last paper, pointed out the possibility of devising some self-acting apparatus whereby the friction between the blocks and wheels should actually regulate the pressure put upon the blocks, and keep it at the precise amount required. Mr. Westinghouse has since devised a new valve, so arranged in connection with one of the brake-blocks that the friction between the block and wheel regulates the pressure upon the blocks. The author had the opportunity of making some experiments on the Brighton Railway on 20th January r879, f° r the purpose of testing this new valve. These experiments are very interesting, as tending to elucidate this part of the subject. The valve, as then used for the first time, was found to regulate the pressure of the brake-blocks against the wheels, and to reduce the pressure as required, except at the last moment, when the escape port was incapable in some cases of discharging the air fast enough to prevent skidding for a very short distance. It was also found that there was a needless waste of air from the reservoir. From the experience gained, a slight alteration has been made, which obviates the above difficulties. The general arrangement of the regulating valve as applied to railway carriages is shown in Fig. 1, Plate 19 ; Figs. 2 and 3, Plate 20, are enlarged sections of the valve itself, the first in the normal condition when it is closed, and the second when it is opened to reduce the pressure. In order to understand the action of the valve, suppose that four blocks act against a pair of wheels, 129 I76 EFFECT OF RAILWAY HRAKES. Al'RIL 1S79. the load at A (Fig. i) on the two wheels being 10,000 lbs. The lever B is of such a proportion that the block C has 3 lbs. pressure upon it for every 2 lbs. on the block D. If the coefficient of adhesion at A is .20, the total adhesion is equal to 2000 lbs. for the two wheels, or 1000 lbs. for each wheel. If the rotating momentum of the wheels be taken as equivalent to an increase of t l of the weight at A, then the four brake-blocks will have to offer a resistance of 2200 lbs. to the two wheels, or 1100 lbs. to one wheel, in order to obtain a result equal to .20 of the load at A. If the friction of the blocks (_' and D on one wheel together equals 1 100 lbs., then, owing to the proportion of pressure upon them, the resistance offered bv C will be 660 lbs. and by D 440 lbs. The block C is suspended from one end of a lever E, the opposite end of which, acting through a link L, tends to move one or other arm of the double bell-crank lever F. The form of this lever F provides for the rotation of the wheel in either direction. The lever E has a proportion of 6% to 1, and consequently requires 97.7 lbs. at its long end, to equal 660 lbs., the assumed friction of the block C. The motion of the lever F, Figs. 2 and 3, Plate 20, is resisted by the bolt G and spring H. When, however, the force on the link L is sufficient to move the bolt G and compress the spring H, a plug valve J in a case K is moved by the bolt. This valve communicates with the triple valve and air-reservoir by the pipe P, and with the brake-cylinder by the pipe M ; while NN are openings to the external air. When the regulating valve J is closed ( Fig. 2 ) these openings are shut, and the passage from the reservoir to the brake-cylinder is open ; but when the valve J is pushed inwards by the bolt G, it first closes the passage of air from the reservoir to the brake-cylinder, and then, if moved slightly further, it opens the passage from the brake-cylinder to the atmosphere, thereby reducing the pressure of the air in the brake-cylinder, according to the area of the passage and the time during which it is kept open. The spring H alone offers 75^4 lbs. resistance to compression, which is equivalent to a coefficient of adhesion of .15 between the wheel and the rail. The regulating valve J has slightly over .3 sq. in. area, and if acted upon by an air pressure from the reservoir of 75 lbs. per sq. in. it gives an additional resistance of 22 Vn lbs., making a total resistance of 98 lbs. '3° Al'RIL 1S79. EFFECT < >F RAILWAY BRAKES. 1 77 equivalent to a coefficient of adhesion of .20. If a pressure of only 3 7 l /i lbs. per sq. in. be admitted to act upon the valve J, the total resistance against the lever E will be only 86.6 lbs., equivalent to a coefficient of adhesion of .175. Thus it will be seen that by simply regulating the air pressure in the reservoirs under the carriages, a considerable variation in the resistance can be made according to the state of the rails. One of these regulating valves is put on each carriage, and the spring H and valve J are alike in all. The lever E must be proportioned according to the load at A : and with the arrangement of brake lexers and blocks shown in Fig. 1, Plate 19, the long end will be to the short end as ijA to 1 for a load of 1 ton on each pair of wheels, or as j}4 to t for a load of 5 tons. If the brake levers and blocks were not arranged as shown, the proportion of the lever would have to be altered, so that the actual amount of resistance on the short end should give 98 lbs. pressure on the long end. It is essential in the use of this regulating valve, if one valve is to operate for the whole carriage, that the brake gear should be so arranged that any one block can act as a fulcrum for all the others. By this arrangement it becomes a simple matter so to regulate the brake pressure as to produce a definite maximum amount of brake resistance ; the amount thus fixed should equal the highest average amount required. The regulating valve can thus be arranged so as to prevent the skidding of the wheels : and by the fact that it closes the passage for the air from the reservoir to the brake-cylinder before it allows any escape of air from that cylinder to the atmosphere, it possesses the property of permitting a high working pressure to be constantly maintained in the reservoir, without the danger of ever getting too high a pressure in the brake cylinder. At the same time, if, after the valve has been opened to the atmosphere, an increased pressure on the brake blocks is again required in order to compensate for the diminution of friction owing to lapse of time or to other causes, the decreasing friction will allow the valve to close the escape orifice from the brake-cylinder and reopen the connection to the air reservoir ; it will thus replenish the cylinder with the original high pressure of air, which ma)' be again reduced and again restored by the valve as may be required. i 7 8 EFFECT OF RAILWAY BRAKES. April 1879. 1 -o ^ T-i W id en r ; f- 5 P ro n v ' 5h„^ <^ a£ 1 pWc'C 1 °5iC « W " H .1 Si! ^3 c^ b" i-l 2* ra^Q b "£ i> Mini mum at En of Expei ment g^ 'u'S> .± 5- £ oj'o-c: ^ i^fisz b &H~ rt "S S EM £ ro ^O O -s. i ~CL. JJ H - O Jj HI — S s V p) cc \D O Q " rr., "*■ tei ■** r -r I>. r-» IN. *0 \f- O in b^ O O O O O f-) >-0 w r. (i\ -f- f") W « rr> O i ) \D pi pi PI - ci M ii -r 1 - O O O O ■^ a kJ PI CO On LO a\ c VO *q W PI ro r^> fl PI pi ri p) - _, . O O n O O l~- in. \D in. m; Tf "-O -1- LT) <-Tl r^. '^ ro v: CO pi pi PI pi ►H April 1879. efflct oi- railway brakes. 179 Experiments made with the Friction Regulating Valve. The preceding Table XI. shows several of the experiments made on 20th January, 1879, by slipping the experimental van from the engine and bringing it to rest by means of the brake applied to all four wheels. The diagrams, Figs. 7 to n, Plates 21 and 22, which have been selected from the experiments shown in Table XL, sufficiently illustrate the action of the regulating valve. These diagrams show also very clearly the variation in the coefficient of friction according to speed. In the experiment, Fig. 7, Plate 21, the van was stopped from a speed of 60 miles per hour in 12 seconds on a rising gradient of 1 in 264. The maximum brake-block pressure on all the four wheels was 160 per cent, of the weight on the wheels at the beginning, and was reduced to ri4 per cent, at the end. The friction increased towards the end of the experiment so much as to cause the wheels to skid just at the end. The adhesion shown was about .25. In this case the pressure was not reduced sufficiently to keep the friction uniform. Had this stop been made on a steep descent, or had the brake- blocks been acting on only one pair of wheels, the time required to effect the stop would have been greater ; and consequently the brake- block pressure, instead of being reduced, must have been increased, so as to overcome the decrease in the holding power of the blocks which results from the length of time of application irrespective of the speed. In the experiment, Fig. 8, Plate 21, the van was stopped from a speed of 57 miles an hour on a rising gradient of 1 in 264 in 15 seconds. In this case the brake-block pressure was 114 per cent, of weight on wheels at the beginning of the experiment, and was reduced to 54 per cent, towards the end. The total friction of the brake-blocks on the four wheels may be estimated, from the actual friction obtained for one pair of wheels, at 3244 lbs. at the beginning of the experiment, and 3 1 44 lbs. in the middle, thus remaining very nearly constant ; and it slightly increased to 3400 lbs. towards the end. There was '33 iSo E1F1-.CT OF RAILWAY URAKES. Al'RIL 1879- no skidding ; but the greater length of time occupied in the stop shows that the pressure was not sufficiently high at the beginning. Fig. 9, Plate 22, shows a stop from a speed of 55 miles an hour in 12 seconds, on a falling gradient of 1 in 264. The brake- block pressure at the beginning of the experiment was 143 per cent, of weight on wheels, and was reduced to 106 per cent, at the end. The resulting friction as estimated for the brake-blocks on the four wheels is 4473 lbs. at the beginning, and 4574 lbs. in the middle of the experiment, after which, as the pressure was not reduced with sufficient rapidity, it rose rapidly, and caused a skid at the end. The adhesion shown was about .25. In this case the stop made was much better than that shown in Fig. 8, because of the greater initial pressure and greater consequent friction. Fig. 10, Plate 22, also illustrates this point by showing a stop from a speed of 55 miles an hour on a level in 18 seconds. The brake-block pressure was 87 per cent, of weight on wheels at the beginning, and 40 per cent, at the end ; and the consequent estimated friction was only 2825 lbs. at the beginning, and 2530 lbs. at the end ; consequently a longer time was required for making the stop. Fig. n, Plate 22, further illustrates the necessity of a high pressure on the first application of the brakes, if a rapid stop is to be effected. The heavy lines show the speed S, pressure P, and friction F, in a stop made from 60 miles an hour on a rising gradient of 1 in 311. The brake-block pressure was 162 per cent., or nearly two-thirds more than the weight on the wheels at the beginning of the experiment, but it was not reduced with sufficient rapidity ; hence one pair of wheels skidded at the end of 9 seconds, at which time the speed was reduced to about 17 miles per hour; and notwithstanding this skidding, the van came to rest in 167 yards and in 12^ seconds. The light lines give the speed s, pressure /, and friction f of the stop shown in Fig. 1 o ; and it will be seen that from an initial velocity of 55 miles an hour, and with a maximum brake-block pressure of about 87 per cent, of weight on wheels, the speed at the end of nine seconds had only been reduced to about 27 miles 134 April 1879. effect of railway brakes. 181 per hour, and that the van came to rest in 227 yards and in 18 seconds. The object of the regulating valve was to obtain a uniform brake- block friction during the whole progress of the stop, and to give to this friction the highest admissible value, /. c, a value as nearly as possible equal to the adhesion of the wheels upon the rails, and therefore just short of that which could cause the wheels to skid. It will be seen from the diagrams that the rapidity of the stop varied according to the greater or less approach made towards the attainment of this object, the resistance of the valve itself being purposely altered during the progress of the experiments. The conditions for these stops were very favourable, and indicate an adhesion of the wheels upon the rails in excess of the average obtainable ; which average, throughout 300 experiments, slightly exceeded .18 of the weight on the wheels. These experiments were made with the one van alone, without any carriages attached to it. Since making them the author has had the opportunity of making slip experiments on the Paris Lyons and Mediterranean Railway with twelve carriages attached to the van. The average of seven stops reduced to 50 miles an hour was 203 yards, with only 63 per cent, of the weight of the train braked. If brakes had been applied to all the wheels of the train, as was the case in the experiments with the single van, the result would have been 128 yards from 50 miles an hour, or a very close approach to the best results obtained with a single vehicle. The author has not had time to analyse these latter experiments fully ; but he is able to state that they demonstrate that the stops which have been obtained with a single vehicle may also be obtained with a train of several vehicles. Regulation of Pressure necessary for a perfect Brake. The regulating valve here described is an outcome of the former experiments on this subject ; and is proposed for the purpose of preventing the sliding of the wheels on the rails. Some such device is a necessary adjunct to a perfect brake, because it is only by the '.35 lt>2 EM-1CT OF RAILWAY ]i RAKES. APRIL 1879. prevention of skidding that the maximum of efficiency can be obtained ; whilst, in addition, skidding damages both wheels and rails, and increases the risk of accident. But the previous illustrations show that, however perfect any apparatus of this description may be, and however certainly it may act to prevent skidding, yet, owing to the verv numerous conditions which affect the application of brakes, it is necessary, if at the same time the maximum allowable friction is always to be exerted on the wheels so as to ensure the best result in stopping, that the action of the apparatus should be capable of being regulated from time to time, so as to meet the varying conditions as to adhesion &c. of the line on which it is travelling : unless indeed some arrangement could be made by which the actual adhesion at the moment could be brought into play to regulate the pressure. Moment u in of Wheels due to Rotation. In dealing with this subject the author has not hitherto directed attention to the question of the influence of the rotating momentum of the wheels ; but he now wishes to state what he has observed on this point. Usually there are in a train a certain number of vehicles braked and a certain number unbraked. When the brakes act on all the wheels, then the rotating momentum of the wheels does not add to the distance in stopping the train, because that momentum can be acted upon by the brakes directly, without in any way making use of the adhesion of the wheels upon the rails. It simply requires therefore an additional amount of brake-block pressure, and if a regulating valve be used, an increase in the resistance of the regulating valve to compensate for this rotating momentum. With the unbraked portion of a train, the rotating momentum of the wheels is an addition to the momentum due to the weight of the train (including therein the actual weight of the wheels), which addition cannot be utilised for retardation ; and it therefore seems the more important that there should be brakes on ever)' wheel of a train. Relation between the Retardation applieil and the Weight of Train. The following Table XII. shows the distances run in stopping a train on a level from a speed of 50 miles per hour, with a retarding i<6 April i S 7 9 . El'TECT OK 183 force varying from 5 to 30 per cent, of the total weight of the train : — T vble XII. — Retarding Forces and Si Length of Stop Retarding Force in H'S. Retarding Force n Length of Stop Proportion to from Proportion to from Weight of Train 50 Miles per Hour. Weight of Train. 50 Miles per Hou Per Cent. Yards. 555s- Per Cent Yards. 5 iS 154J 6 463 19 u,6i 7 3 & yH 20 r 39 8 347* 21 '32* 9 30« 22 '29-S 10 2 77J? 2 3 I 20J 11 2 --,21; 24 1 1 s;i 12 23^ -5 1 1 1 '3 -13* 2b 107 14 19s*- 27 <°3 '5 IS 5 28 99+ 16 1 7 3 H 29 95S 17 ">3i 3° 9*ir If the brakes act upon every wheel, then a retardation of 10 per cent, of the load carried by each wheel — counting the rotating momentum as part of the weight — will stop a train in 27 7 J?' 3 yards. If the brakes act upon only half the weight of a train, a retardation of 20 per cent, would have to be exerted upon the braked half to produce the same result. As already pointed out, 20 per cent, adhesion is rather above the average obtainable, while 24*2 per cent, is the highest result obtained under the most favourable circumstances at any considerable speed, or except when sand was applied to wheels moving slowly. The above Table XII. should be carefully noted, for it will be seen that, even when brakes act upon all the wheels, 24-% per cent, retardation will only give 26 yards better result than 20 per cent., or 52 yards if only half the train be braked. If compared with 18 per cent., the average adhesion obtainable, the advantage will be only 41 yards for the train braked throughout, and 82 yards for the train having brakes acting upon half of the weight. A consideration of this feature of the brake problem points out — 1st, that the advantage to be gained by trying to obtain above 20 per cent, retardation on each wheel is greatly overbalanced by the risk i37 1S4 EFFECT OF RULWAY BRAKES. Al'RILlS79. of skidding; and 2d, that it is far easier and safer to make a stop in 250 yards from 50 miles per hour with the whole train braked, than with brakes upon only half of the train. All of this points to the fact that in arranging valves for regulating the brake-block friction care should be taken not to exceed a safe limit of adhesion ; for in the effort to get more work, less may be the result. Too much stress cannot be laid upon the importance of immediately applying the full pressure of the brake-blocks against the wheels, and of making the application simultaneous against all the wheels of the train ; for any loss of time seriously impairs the efficiency of the brakes in several ways, as has been already explained, independently of the actual increase of distance run in the stop. In illustration of this point, the diagrams shown in Figs. 12 and 13, Plate 22, are added. Fig. 12 shows the result of an experiment made on 23rd August 1878, in which the application of the pressure was gradual, so as to represent the effect of a slowly-acting brake ; it furnished a diagram of a stop nearly identical with one of the best stops made by the Vacuum experimental train on the North Eastern Railway in October last. The curved lines S and F represent the speed and retardation obtained in the experiment ; and the straight lines S, and Fj the comparative results which would have been obtained if the full pressure had been applied at once, and if the consequent friction had been generated at once between the brake-blocks and wheels, and if this friction had been maintained at a uniform amount. In the latter case the stop would have been made in 125 yards, instead of 287 yards, the actual distance. The straight line S„ shows a stop which might have been made in the same distance of 287 yards, if the very moderate retardation indicated by the dotted line F 2 had been applied at once. It will be noticed that this stop is much better than the actual stop, though no shorter in distance, because at any intermediate point the speed is much lower; hence at 100 yards, for instance, the energy left in the train, as shown by the straight line S„, is only three-fifths of that shown by the curved line S representing the actual stop. Fig. 13 illustrates the advantage of applying the brakes to every wheel of a train. nS April 1S79. effect of railway brakes. 1S5 The diagonal line AB indicates the stop which a train could make from 50 miles an hour with the retardation of .20, shown by the horizontal line CD, if applied to every wheel in the train. The shaded area below CD shows the extra retardation consumed in overcoming the momentum of the braked wheels. The diagonal line AE shows the stop which a train could make from the same speed with the same retardation of .20 applied to only half the wheels and half the weight of the train, as indicated by the horizontal line FG. The shaded area below FG shows the extra retardation consumed in overcoming the momentum of the braked wheels ; and the shaded area below AE shows the extra distance run by the train in consequence of the momentum of the unbraked wheels. The diagonal line AH shows the stop which a train could make from the same speed with the same retardation of .20 applied to only one-fourth of the wheels and one-fourth of the weight of the train, as indicated by the horizontal line JK. The thickness of the line JK and the diagonal shaded area below AH show respectively the extra retardation consumed in overcoming the momentum of the braked wheels, and the extra distance run by the train in consequence of the momentum of the unbraked wheels. Requirements of (' As the speed was greater, the distance through which O would fall in its passage would be less ; consequently the distance O'C would be less, and the work of dragging (.) up to the vertex C would be less also. Hence it might be seen how the actual work done per foot run of surface, or in other words the apparent coefficient of friction, would be less as the speed was greater. He was indebted for this illustration to a letter by Mr. T. V. Krajewski, of Chicago, in the Railroad Gazette of nth April 1879. There was another point to which he desired to refer — the reason why the wheels when sliding on the rails produced a less retarding effect than that of the brake-blocks acting against the wheels when the latter were not skidded. When a wheel was rolling upon the rail, the point of the wheel which was in contact with the rail at each moment was for that moment in a state of rest, and therefore the adhesion between the wheel and the rail at that point corresponded with the coefficient of statical friction ; and the whole of that adhesion could be utilised by means of the brake-blocks for retardation of the train. Buc when the wheel slid on the rail, the surfaces were of course in the condition of dynamical friction, the value of which was much smaller than that of statical friction ; and for that reason the 14S April 1879. railway brakes. 195 retarding effect was much greater in the former case than in the latter. Mr. D. M. Ykomaxs wished to say a few words with reference to M. Marie's paper. Referring" to the question of brake-gear, it was said (page 159) : " The author does not understand the object of the latter type of brake-gear, which is generally applied with the Smith brake, and is also extensiyely used by the Great Northern Railway in England." Now the brake-gear used in the trials described in the paper was selected by M. Marie, Sen., locomotiye superintendent of the Lyons Railway. In no case did he ( Mr. Veomans) recommend any special form of brake-gear in connection with the Vacuum brake. He had that day looked oyer the list of English railways using the brake, and there were only one or two cases in which the gears were alike. The particular gear in question was submitted with three other arrangements to M. Marie', and was selected by him as the most proper to use in connection with the brake. It was an arrangement designed and used by Mr. Stirling of the Great Northern Railway, and, so far as he knew, it was perfectly satisfactory to him. With reference to the breaking of draw-bars, it was stated in M. Marie's paper (page 167) that in the case of the 'Westinghouse brake the application was practically instantaneous throughout the train, and that with the Vacuum it was much slower ; but in both cases, according to the statement, draw-bars were broken. If the application was instantaneous, so far as his (Mr. Yeomans') experience went, there should be no breaking of draw-bars, because each carriage would haye its brake power applied at the same moment ; there would therefore be no strain brought upon the draw- bars anywhere. This was a question that he had thought of and studied a great deal, and he had made experiments with an automatic brake of his own, with a view of ascertaining as nearly as possible the effect that would be obtained by instantaneous application. He had used electric arrangements for applying the brake throughout the train as nearly instantaneously as possible ; and he had found that when this was done there was practically no strain brought upon the draw-bars, because there was no compression of the buffers. He "49 196 RAILWAY BRAKES. Al'RIL 1879. thought there was something in that question which needed to be explained further. It was an important point, because in both cases it appeared that draw-bars were broken, although the nature of the brake was different. There was one other point to which he would refer, because it bore upon the Vacuum brake. The paper stated (page 168): "All the parts of the Vacuum brake are very simple and strong, but the leather of the Hardy sack seems to be a bad material for practical work." That, he felt sure, was only a speculative opinion on M. Marie"s part. He had had leathers in use upwards of two years, and out of more than two thousand which had been in use for that period not twenty had failed. M. Marie would endeavour to show how it was that on the Lyons Railway draw-bars were broken with continuous brakes more easily than was generally the case. PI. 23, Fig. 17, was a sketch in plan of the buffing and draw-gear of two carriages coupled together, as used on that railway. It would be seen that the buffers were worked by the ordinary laminated springs, and that the draw-gear was also worked by similar laminated springs, but much shorter and stiffer. The carriages in ordinary working were coupled up so as to bring a considerable compression on the buffers, whatever the tractive force might be ; the buffer springs could turn slightly round the point M, so as to work easily round curves. This compression of the buffer springs had a very good effect in preventing lateral oscillation ; but it was a disadvantage in the use of continuous brakes. For suppose the buffers to be violently compressed by the unequal action of the brakes ; then of course they would open out again as violently, and tend to push the carriages much further apart than they were in ordinary working. This of course brought a strain upon the draw-springs ; and these, being short and stiff, took up this strain very rapidly, producing in consequence a very severe tension on the draw-bars. That was the reason why, with their form of couplings, there were more breakages than with the ordinary form. Generally there was not so much breakage of couplings with the Westinghouse brake as with the 150 April 1S79. railway urakrn. 197 Vacuum brake, because the action of the brake was more instantaneous, so that each of the carriages would be stopped separately. But if the Westinghouse brake was not in good order — if there was a little leak in the pipe, or something of that kind — it gave a difference of action in the different carriages, and so might cause a breakage in the couplings. On one occasion there was a breakage of two couplings in a Westinghouse train, owing to the existence of a slight leakage. With regard to the brake-gear, the choice of it had been left by the chief engineer to the Westinghouse and the Vacuum companies ; what he had wished to point out in the paper was that the brake-gear of the Vacuum brake on the Lyons line did not give an equal pressure on the two blocks ; but it was obvious that this was not a defect in the brake itself, and it would be easy to remedy it. He had seen a recent design of the Vacuum company, showing a brake-gear which would give an equal pressure on both blocks ; that in his opinion would be the best arrangement of all. He could not give any definite results as to the Hardy sack, because it had only been in use on the Lyons Railway for a few months ; but he still thought it would be better to use metallic cylinders than the leather sack. M. D. Baxderali agreed in principle with M. Marie as to the material of the Hardy sack. It was possible, he supposed, to get a better material than leather for those sacks, and probably that would come to pass some day; but up to the present time they had found the leather much better than the old material, india-rubber, as used in the original Smith sacks, and he had no doubt that it would last longer. The use of the leather sack had of course some inconveniencies : the leather was not always very good ; it was sometimes porous, and then it was difficult to keep it air-tight. It was also necessary in constructing the sack to make holes in the leather, in order to attach the two central plates of iron which were connected with the break-gear. Sometimes there was a leakage caused by those holes ; but by care, and by putting proper packing between the lips of the two plates and the leather, that could be obviated. Up to the present time the leather certainly seemed to be I98 RAILWAY BRAKES. APRIL 1879. the best thing to use. It required to be greased and looked over occasionally, and then it would be found to answer well until something better was discovered. At page 159 of M. Marie's paper reference was made to the question of the change in the distribution of the weight of the carriages when the brake was applied. For three years the Smith Vacuum brake had been working on the Northern Railway of France, and his attention had been given to the subject ; but he had not found with that brake so great a difference in the distribution of the weight on applying the brake as appeared in the experiments carried out on the Lyons Railway. There was a certain effect of that kind, but it was not so important as it was found to be on the Lyons Railway and on some of the English railways. He attributed this to the fact that in the brake-gear generally adopted in England, and on the Lyons line, the blocks were hung direct upon the body of the carriage, which was supported by the springs, so that the friction of the wheels against the olocks produced a depressing action on the front part of the body, pressing it down on the springs. On the Northern Railway that was not observed, because the blocks were not hung direct upon the frame or body of the carriage, but upon a lower and separate frame supported by the axle-boxes (PL 24, Fig. 18). This arrangement had been found much better for many purposes, and had always been used on that railway. Thus when the friction of the wheels against the blocks began, all the vertical strain was taken by the second frame independently of the springs, and a great deal of the disturbing action that had been referred to was avoided ; the roughness or .shaking referred to by M. Marie' was also avoided. Another very important point to which he would refer was the cause of the breakage of draw-bars. From his own experience, and from the experience of many English engineers who had continuous brakes on their lines, he thought that the only plan of getting rid of the difficulty described by M. Marie would be to use continuous draw-bars, and to get rid of the stiff draw-springs as much as possible. On the Northern Railway a sort of continuous draw-bar was employed, but not quite so complete as the continuous draw-bar used on some lines in England ; and as a matter of fact they had no breakage of '5- April 1N79. railway israkes. 199 draw-bars in their service at all. Last summer express trains had frequently been run consisting of twenty-two or twenty-three carriages fitted with the Vacuum brake, and sometimes not completely fitted, which was the worst possible state of things ; and yet there had been no breakage, though of course there were sometimes slight shocks. M. Marie rightly observed (page 169) that it would be very interesting to make experiments on long inclines with these brakes, to see if a uniform speed could be maintained with them. That had been done on the Southern Railway of Austria by Herr Gottschalk, formerly the locomotive superintendent ; by whom the Smith Vacuum brake had been applied on the trains running down the Semmering incline, which was 1 in 40 for a long distance. The brake was now in use on that steep incline, and succeeded very well. Professor A. B. \Y. Kennedy desired to say a few words with reference to the connection of friction with velocity and with time, as investigated in Capt. Galton's paper. First, with reference to the diminution in the coefficient of friction as the velocity increased. In Table IX. Captain Galton had given the mean results of a great many experiments which seemed to point most distinctly and indubitably to that conclusion ; he had also given the maximum and minimum values of his coefficients, as well as the mean. The maximum values increased almost regularly as the speed decreased ; the minimum values also, but less regularly ; and the mean with very fair regularity. He had examined some other experiments on the same question, including some of the old French experiments made by Poire'e in 1851 (Societe des Inge'nieurs Civils, 185,2), and by Eochet about 1857 (Comptes Rendus, 1858 and i860), not with brake-blocks, but with wheels skidding, or with skid-brakes on the rails. In all cases the results were exactly similar. Those papers gave only the equations deduced from the experiments, while Captain Galton had given the experiments without the equations, which was an important advantage on his side. The only case in which he (Professor Kennedy) had found a contrary result was in experiments on the friction of journals by Professor Thurston (Journal of the Franklin Institute, Nov. 1878), in which the conclusion arrived at was that '53 200 RAILWAY BRAKES. APRIL 1879. the friction decreased up to a moderate velocity, /. e., from 50 feet per minute to 300 or 350 feet, and then began very slowly to increase up to 1200 feet per minute, the maximum velocity tested. He was not able to account for this result, unless it was due to the method of conducting the experiments (which unfortunately was not described), or to the particular lubricator used, those experiments having been made with lubricated journals revolving in bearings. As to the effect of time on friction, Captain Galton had noted that the coefficient of friction decreased as time went on. He (Professor Kennedy) had plotted out the figures of Table X. (which showed the diminution in the coefficient of friction after 5, 10, 15, and 20 seconds) ; and he found from the curve that the coefficient apparently became constant after about a minute to a minute and a half ;■ its ultimate value was in most cases very small and was not always the same, but the curve in all cases became horizontal. He did not know whether Captain Galton had made any experiments — by keeping the brake on for a considerable time — such as would show whether the coefficient of friction did actually come down to a limit after the braking had lasted about a minute. It would be very interesting to know whether all the curves came down to the same minimum value or not. As far as he had plotted them out, as shown in clotted lines on Fig. 14, PL 23, they did not. The higher velocities still retained the lower coefficients, even at the very lowest value, which was about .03 or .04. The polishing of the surfaces he suggested might sufficiently account for the diminution of friction, even in the very short time, about 20 seconds, shown on the diagrams. He had also worked out, as far as he could from Captain Galton's published diagrams, the relation between the friction and the intcnsitv of the pressure, that is, the pressure per square inch. He had found that the coefficient of friction between the wheel and the rail, during skidding, was (roughly) less than one-third the coefficient of friction between the wheel and the brake-block; and this conclusion was directly borne out by a statement in Captain Galton's paper. In reference to the real cause of this very greatly reduced coefficient of friction when the wheel was sliding on the rail, as compared with '54 April 1X79. railway brakes. 201 that when the wheel was revolving (and thus sliding) against the block, he had suggested at the Paris meeting that the pressure per square inch was so very much greater in the former case (the wheel surface in contact with the rail being so small compared with that in contact with the block), that the surface became immediately polished by the high pressure ; and to this view he still adhered. From Captain Galton's paper it would be seen that a heavy brake- block pressure might be taken as being 35,000 lbs. divided over 8 brake-blocks, each of 36 sq. in. area, which gave about 120 lbs. per sq. in. On the other hand, supposing the wheels to be flattened to the extent of % in. so that they rested on the rail across its whole breadth for a space ^ in. long, there would be about 10,000 lbs. per sq. in. pressure on that little surface ; and it was no wonder, comparing 1:0 lbs. with 10,000 lbs. per sq. in., that a somewhat different result was obtained. Captain Galton had however experimented with greatly differing brake-block pressures, varying so much (in different experiments) that this theory might certainly be expected to apply to them if it applied at all. He had therefore worked out a number of Captain Galton's diagrams to see if he could discover any diminution in the coefficient of friction at the higher brake-block pressures. He had found this to exist in a very marked manner indeed, if his working out was correct ; and so far of course the views he had just expressed were corroborated. All engineers must feel extremely indebted to Capt. Galton for the results he had given, for the clear way in which they were stated, and for the very careful manner in which they had been worked out. He thought also from an engineering point of view it was only just to recognise the way in which Mr. Westinghouse had set to work, immediately that an imperfection was pointed out, to remedy it to the best of his ability and with his accustomed ingenuity. Whether or not his regulating valve would ultimately become universally used, whether or not further experience would show it to be thoroughly durable, the way in which Mr. Westinghouse had gone to work, directly he found that something was wanted, to design precisely the thing that was wanted, was as good an illustration of the spirit in which engineers ought to work as could be found anywhere. '55 202 RAILWAY BRAKES. Al'RII. 1879. Mr. A. Pai;et wished to ask M. Marie, with the regard to Mr. Yeomans' remarks on the flattening of springs, whether on the Lyons Railway the suspending rods of the brake-blocks were attached to the main frame supported by the springs or not. M. Marie replied that the brake-blocks were suspended directly from the frame of the carriage, much as shown in Plate 19, Fig. 1, of Captain Galton's paper. Comparing that arrangement with one in which the blocks were hung from a separate frame supported by the axle-boxes, it was true that the front springs would be a little less flattened in the latter case than in the former. But with the Westinghouse arrangement of brake-gear, the pressure of the inside blocks being greater than the pressure of the outside blocks, the moments of their frictional resistances, taken round the centre of gravity of the carriage, were almost the same ; and in that case there was little or no extra load on the springs. As regarded the easy motion of the carriages during the application of the brakes, he would show that it was no disadvantage with the air or vacuum brakes to suspend the brake-blocks from the frame of the carriage ; it would however be a disadvantage with the chain brake or with the screw brake. In the former brakes the blocks were always applied by an apparatus which was elastic, and gave way when needed, viz., the air cylinder in the one case and the vacuum sack in the other. It was therefore easy to push the two blocks outwards from the wheels at any time, by applying an outward pressure equal to the pressure put upon the blocks by the brake. The contrary however was the case with the screw brake or chain brake : it was there impossible to move the blocks outwards from the wheels, because this movement was resisted by the screw in the one case, and by the dead pull of the chain in the other. Hence with an elastic system it was easy for the wheel to move vertically between the horn plates of the carriage frame, and so to rise between the brake-blocks, if required ; and as the brake-blocks were but little above the level of the wheel centre, a rise of 1 inch would only push the blocks say T l „ in, outwards on each side. Hence, when the wheels were not skidded, the springs were almost as flexible when the brakes were on as when the)' were off ; ,56 Al'RIL lN;<). RAILWAY BRAKES. 203 in that case the friction of the brake-blocks did not have any effect in preventing the wheel from rising, because while the wheel was rising no appreciable change took place in the relative velocity of the tire and the brake-blocks, and therefore there was no variation in the friction. But when the wheel was skidded the case was different. The wheel could not rise without at the same time moving from rest in contact with the brake-blocks, and thus overcoming the statical brake-block friction, which at that time was very large. Hence it was always felt that, when the wheels were skidded, the carriage went much harder and rougher than at other times. ( hi the whole it appeared that the springs were not more flattened, and were quite as elastic, with the Westinghouse arrangement as with blocks hung from a frame supported bv the axle-boxes. In the former case a slight vibration was indeed felt, caused bv the rigid connection between the blocks and carriage frame ; but that was a very small disadvantage compared with the great simplicity of having the brake-blocks suspended direct from the frame. Mr. A. Facet wished to ask Captain Galton for further information. He had stated two very interesting facts. The first was that the statical friction was very much greater in the particular case of brakes than the friction after movement had begun. This point had been exemplified and remarked upon very strongly by Mr. Hawksley in a former discussion on friction brakes (Proceedings 1876, p. 241). He (Mr. Paget) had also stumbled upon the same fact, not as to the friction between wheels and brakes, but between grooved wheels and ropes passing round them. He had made a great many experiments some years ago, in order to ascertain a rule by which the friction of a rope passing round half the circumference of a grooved wheel could be determined ; and he had found the ratio between the force required to start the rope, and the force required to keep up the motion when it was started, to be as three to one in some cases, while in other cases it was much less. Again, Captain Galton had stated that the dynamical friction of the brake-blocks very much diminished when it was continued for any time. He had found the same thing with ropes moving at high speeds, and '57 204 RAILWAY BRAKES. Al'RIL 1 879. he thought that it might arise from a species of lubrication going on, in consequence of a change in the nature of the material, caused by the heat that was generated. He observed that Captain Galton used the words, "through some cause not yet fully determined," in speaking of this decrease in the coefficient of friction ; but possibly since writing his paper Captain Galton might have found himself in a condition to give them information as to some principle by means of which the cause might eventually be fully determined. Mr. T. Hurry Riches said the Taff Yale Railway had started working with the Smith Vacuum brake in June 1876, and since that time, exclusive of shunting and other stops of that kind, they had made 79,100 stops with it; and, as the Board of Trade return would show, they had met with only two slight mishaps, neither of which had been serious to anybody. With reference to the question of the breaking of draw-bars, he might state that they had not had a single accident of that kind. On one of their branches, on which the brake was used regularly, the ruling gradient was 1 in 50, and the average gradient for five miles was 1 in 66 ; so that it would be seen the brake had as much work to do in a given time as on any line in the kingdom. Mr. R. D. Sanders noticed that upon the first page of Capt. Galton' s paper they were reminded that "the retarding effect of a wheel sliding upon a rail was much less than when braked with such a force as would just allow it to continue to revolve." Now he himself had had experience in India in working very heavy inclines — some of the heaviest in the world. There the brakes were worked by native Indian brakesmen, and they naturally, from want of experience, thought that the harder they screwed the brake on the slower the train would go ; but the contrary was the fact. Many locomotive superintendents besides himself had found, in turning up the wheels, that a hard place appeared here and there, and just upon that hard place the wheels seemed to have slid along the rails ; thus the friction was considerably reduced from the hardness of the two surfaces which were in contact. That, he thought, partly accounted for the reduced friction of the skidded wheel. '58 April 1S79. railway brakes. 205 He wished also to offer a remark on Captain Gallon's statement that the experiments were influenced by many circumstances which he could not examine or control. Now if it was not possible to examine all the circumstances which were connected with the braking of trains, and to put them in a practical form, was it possible to put any of them in such a form ? — because each one affected all the others. He did not wish in any way to depreciate the value of Captain Galton's experiments ; on the contrary, he thought that much praise and credit were due to him for the great amount of trouble he had taken in working them out ; but he thought it was well to be cautious in accepting his figures as data for working out the continuous brake of the future. Would those experiments enable a locomotive superintendent to say, " I must put on two brake-blocks," or " I must put on one ; " and " I must hang my blocks in this way, or in another way;" and so on? He thought that the details of the brake gear might well be left in the hands of the locomotive superintendents, who would adjust the blocks as the}' liked, and hang them in the way they thought best for their own requirements. He therefore asked what really were the practical results to be derived from the paper. In the appendix the question was asked (p. 193) — no doubt with the idea of elucidating that very point — " Has the brake any self-acting means of indicating when it is out of order ?" Now he thought that point was the most important to be considered in connection with continuous brakes. It was very evident that if a brake would not of itself indicate when it was out of order, it was little better than a non-automatic brake which was not in order when it was wanted. The result would in fact be worse, he contended, if an automatic brake did not indicate when it was out of order, than in the case of a non-automatic brake : because the driver, from the habit of seeing the gauge in front of him actually registering the power, would naturally have more confidence in the brake, and might thus be led into danger. It seemed to him that it was a matter of absolute necessity not only that the gauge should indicate what the pressure was in the continuous pipes along the train, but also that the gauges, both on the engine and in the guards' vans, should indicate the condition of the brakes themselves, as they would '59 206 • RAILWAY HRAKKS. APRIL 1879. be when the driver put them on. The question was, how was that to be effected ? If in any given case there was a pipe communicating from one point to another, and through that pipe it was desired to convey communication of any sort, an engineer would naturally refrain from putting cocks, or valves, or anything which might cause an obstruction, in the pipe ; for if this were done, it would probably be found that, through some failure of the apparatus or some mistake in leaving the cocks closed, the obstruction would prevent the passage of the compressed air or other medium through the pipe to the point where it was wanted. Therefore he contended that, from the gauges to the brake-pistons on the carriages, there should be no cock, no valve, nothing whatever which, either from negligence or by getting out of order, could possibly create an obstruction in the continuous pipe. If such obstructions existed, the driver might have a certain pressure registered by the gauge in front of him, and so might go running into a station at 60 miles an hour, thinking that all was right, but on applying the brake might find himself unable to stop : on examining the cause it would be found that somewhere or other there was some little cock or valve that had been shut, and thus the gauge on the engine only registered the condition of the pipe up to the point of obstruction. It appeared to him that the only way to overcome this difficulty was to reduce the brake arrangements to the utmost possible simplicity ; and he would confine himself to that point. First, with regard to the question of material, he was glad to hear india-rubber condemned. He was at present devoting his attention to cast-iron cylinders and pistons ; he believed that everything necessary to make a useful and perfect brake could be provided by means of a simple piston in a cylinder, with the pressure upon both sides of the piston ; no valves were then wanted between the continuous pipe and the reservoir, and no valves to regulate the flow of pressure from the reservoir to the cylinder ; thus the gauges, both in the guards' vans and on the engine, would at all times register the pressure that would be on the pistons at the moment the brake was required. Unless this condition were fulfilled to the utmost, he maintained that a brake was not absolutely reliable. There was another 160 April 1879. railway brakes. 207 point bearing upon this subject, which had been touched upon by M. Marie in speaking of broken draw-bars. If the brakes were worked by means of valves between the reservoir and the brake-cylinders, it was very evident that their effect depended, not merely upon the pressure in the reservoir, but also upon the action of those valves. If they all worked uniformly, the effect upon all the vehicles would be the same ; but if the valve upon the first carriage did not work, and the valve on the second carriage did, and that on the third did not, and so on throughout the train, there would then be set up that very motion which M. Marie had alluded to, and which was the cause of the breakage of draw-bars. By taking away the whole of those valves and cocks, and converting the whole apparatus, from the engine gauge to the guards' gauges, and from the reservoir to the cylinder, into one chamber as it were, what happened in one vehicle must then happen in all, and whatever pressure was registered upon the engine-gauge must also be the pressure on every piston when the brakes were applied. With such certainty and uniformity in the action of the brake, broken draw-bars would, he thought, be got rid of entirely. So far as the experience with his own brake upon the Great Western and the Midland Railway went, no such breakage had to his knowledge occurred : simply because the brake went on with a uniform pressure throughout the whole train. He thought that these remarks on the importance of simplicity were borne out by the observations in M. Marie's paper as to the triple valves. As he had already stated, if there were no complicated parts in a brake, there was nothing to get out of order ; but brakes having delicate pieces of mechanism, which in the case of a sharp frost might solidify, would be of little use. Mr. John F. Haskins was glad to hear gentlemen defending the vacuum side of the brake question, as being himself the only representative present of the vacuum system of his friend, Mr. Fred. W. Eames. He desired to thank Mr. Westinghouse, and also Captain Galton, for the assistance they had rendered in elucidating the subject of brakes. He also wished to say a word on the question of the durability of india-rubber sacks and diaphragms. There must 161 208 RAILWAY BRAKES. APRIL 1S79 be many engineers present who had used india-rubber in various forms for many years without failure ; and there were experiences on record which were decidedly in favour of india-rubber diaphragms. In Eames's brake (as to which some trials* had been made that very week on the Great Northern Railway) the diaphragms were of india-rubber. That brake was used extensively in the United States, and especially on the Metropolitan Elevated Railway and the New York Elevated Railway, both in New York. The trains on these railways ran at two minutes' interval for about twenty hours each day ; and the engineers' reports showed that on each line about 13,000 stops were made daily. With this aggregate of 26,000 stops made daily with the Eames form of india-rubber diaphragm, it seemed to him that there must be more value in india-rubber sacks and diaphragms than would appear from the remarks previously made. Sir Henry Tyler would like to take the opportunity, as an old member of the Institution, of expressing the satisfaction with which he saw so many foreign gentlemen amongst them, taking an active part in their discussions. He hoped that more and more of those gentlemen would come from all parts of the world, so that they might indulge in mutual exchanges of courtesy and information. Mr. R. Price Williams said that, far from sharing the views of a former speaker in questioning the value of the contents of Capt. Galton's paper, he had hesitated as to whether he was justified in offering any remarks on the subject, without having previously had the opportunity of carefully studying the mass of valuable data they had now before them. Professor Kennedy had referred to some experiments made long ago by Poire'e with regard to the coefficient of friction. He had himself given a good deal of consideration to the subject, and had been very much struck, before Prof. Kennedy rose, on comparing the results obtained by Poiree, a copy of which he held in his hand, with those given in Table IX. of Capt. Galton's paper, to find how very closely they agreed. He might state that * The particulars of these trials were published in " Iron " of 25 April, 1879. 162 April 1S79. RAILWAY BRAKES. 209 Poire'e's coefficients of " sledging friction," as he termed it, were given in Mr. L). K. Clark's Manual, p. 724. The experiments were made at different speeds, varying from 10 up to 49 miles an hour, and under different conditions of the rails, such as dry, very dry, wet, &c. He had taken the means of the coefficients in Poiree's Table for such speeds, and found, as in the case of the coefficients given in Capt. Galton's Table, that the results showed in a very marked manner that the adhesion decreased as the speed increased ; indeed he found TABLE A. — Coefficients of Friction at Varying Steeds. M. POIREE'S Experiments* with Skidded Wheels. Wrought-iron Tires and Rails. Four-wheeled Wagon Weighing 3.35 Tons. CAPTAIN GALTON'S Experiments, Table IX Cast-iron Brake-blocks on Steel Tires. Wheels Not Skidded. COEFFICIENT OF FRICTION SPEED. Miles per Hour. Rails Dry. .208 Rails Very Dry. Rails Dry. Springs Blocked. Rails Wet. SPEED. Miles per Hour. MEAN COEFFICIENT OF • FRICTION. 10 10 .242 1 3 .201 16 .200 15 • 22 3 18 .ISO .246 20 •'75 .110 20 .192 22 .167 24 .172 25 .166 27 ,162 29 .222 3° .164 32 .144 35 .154 35 .142 40 .202 40 .140 45 .136 .132 45 .127 47 .083 49 .187 1 50 .116 - -- . , ' * Memoires de la Societe des Ingenieurs Civils, 1852, pp. 110-115. '63 RAILWAY BRAKES. April 1879. in his own note-book a remark of his own to that effect. The comparison was given in the accompanying Table A (page 209). With reference to the diagrams, he could not help thinking that the mode of representing the results in which the abscissa showed the distances was very much more striking than that in which they showed the times. He might state that he had been at the trouble of showing the results given in Capt. Galton's previous papers in this way. He had also endeavoured to show in the same diagram the outline which, according to his notion, would afford the best possible result that could theoretically be obtained. Fig. 15, Plate 23, gave these results for one of Capt. Galton's experiments (Proceedings 1878, Plate 62, Fig. ro). The abscissae of the curve AB gave the distance run, and also the position of the brake-van after the end of each second ; while the ordinates gave the velocity of the van at each instant. The straight line AC showed what would be the velocity at each second, on the assumption that a retarding force equal to one-fifth of the weight was applied instantaneously and maintained uniform throughout the stop. It gave in fact the theoretical limit to which they should strive to approximate in practice as far as possible. The exact figures corresponding to the diagram he had annexed in Table B. TABLE B. — Captain Galton's Experiment No. 5. (Proceedings Inst. M. E., June 1878, Plate 62, Fig. 10.) Seconds of Stop. Mean Velocity During Each Second. From Diagram. Distance Run During Each Second. Total Distance Run at End of Each Second After Applying Brake. Miles per Hour. Feet. ISt 46 67 67 2nd 45 66 '33 3 rd 43 63 196 4th 41 61 257 5 th 38 56 3'3 6th 34 50 363 7 th 29 43 406 8th 24 35 441 9th 20 29 470 10th 16 2 3 493 nth 12 18 5" 12th 4 6 5i7 Total 517 feet or 172 yards. 164 Al'RIl. 1S79. RAILWAY U14AKI.S. 2 I I With regard to the allusion made in the paper as to the necessity or desirability of appointing a Government Commission to report on the brake question, he had hoped that before the discussion closed some one would have alluded to this matter. He considered that it was at the present moment of the greatest importance that this proposition should have the consideration it deserved. In his opinion a Government Commission, such as Capt. Galton had suggested, alone was capable of grappling with this most important but difficult subject. Xo one could have attended the interesting discussions which had occurred on these papers of Capt. Galton's without coming to the conclusion that the time had arrived when a Commission should be appointed, and that where matters of such gravity and importance were concerned, all personal elements should as far as possible be eliminated. He agreed entirely with Capt. Galton that the only way of arriving at that result would be by some such procedure as he had suggested ; and he hoped that the president of the Institution and the members of the Council would do all in their power to further this object. Mr. E. A. Cowper said it had been mentioned in M. Marie's paper that since the centre of gravity of a carriage was some little distance above the sole-bar, while the brake-blocks were applied much lower down, the carriage tended to sink down on the front springs, from the inertia at the centre of gravity pushing it over the point of resistance. But that point of resistance was in fact on the surface of the rails, viz., where the friction occurred that caused the retardation of the carriage, and any tendency in the carriage to rotate, or become depressed in front and elevated behind, was in fact caused by the centre of gravity being above the rails ; the gripping of the rotating wheels by the brake-blocks was simply the means of bringing into play the friction of the rails. With regard to friction, he partly agreed with Professor Kennedy as to the polishing, as it was usually called, or, as he preferred to call it, the burnishing of the surfaces. When a brake-block had been at work two or three minutes, and was then taken off, a quantity of dust always fell out, which had of course been rubbed off the block and 165 212 RAILWAY BRAKES. APRIL 1879. the tire. Now these surfaces, as Capt. Galton had mentioned, were no doubt covered with minute pointed hills and hollows, partly entering into each other. It seems very likely that, after the brake had been on a little while, the points of these hills got rubbed off, and then there were flattened surfaces riding over one another, and never fairly entering. In addition, the hollows got filled up more or less with dust ; hence there were no longer a number of hills and hollows fitting into one another, but a number of hollows partly filled with dust and powder, and a number of hills with their tops partly rubbed off riding over them ; so that there was a much more regular surface than there had been at starting. That would necessarily allow the two surfaces to be made to slide over each other with much less application of force than at the commencement of their motion. A certain amount of burnishing had taken place, which prevented the adhesion of the two surfaces to anything like the same extent as at first. The Hon. R. C. Parsons mentioned that a few years ago he had made some experiments with wrought-iron brake-straps applied to cast- iron wheels, with the view of taking up the power of a steam engine. The speed was very high, and the amount of dust that fell out when the brake was taken off was considerable. He had estimated the amount which could be between the wheel and the brake at any one time by the amount worn off the strap and the surface of the wheel, and he found that there was a very decided film between the strap of the brake-wheel and the wheel itself. The same idea was borne out by the facts connected with the metal used by Mr. Perkins for the piston-rings of his engines. A considerable amount of metal was always being worn off, and thus there was always a film of dust between the piston-ring and the cylinder ; and the friction was there diminished in a great degree. That was also the case where oil or any other lubricant was placed between the surfaces, especially where a current of oil was kept continually running through : he had tried experiments on this point, and found that the friction diminished wonderfully. 166 April iN;c). uaij.w.u brakes 213 Mr, Charles Hawksley observed that, on the occasion of Captain Galton's first paper being read at the Paris meeting, he had ventured to suggest that the skidding of wheels might be prevented by some automatic arrangement which would enable only so much pressure to be applied to the wheels as would just fall short of skidding them. He was very much pleased with the arrangement carried out by Mr. Westinghouse for that purpose. It appeared to him however that two other elements at least must enter into the consideration of the question : one was the weight upon the wheels, and the other the state of the rails. The last could not he thought be easily provided for ; but the first by the introduction of a lever, might be brought into combination with the lexers already forming part of Mr. Westinghouse's apparatus, and so produce a more perfect instrument. M. Marie wished to say a word with regard to the coefficient of friction diminishing with the time. Two causes might contribute to that effect: first the polishing of the surfaces, and secondly the heat evolved. The polishing of the surfaces did not seem to be the real cause, because the coefficient of friction always took exactly the same value at the beginning of a second experiment as it had at the beginning of a first experiment. ( )n the other hand it was well known that the condition of substances changed a great deal with heat ; also that the coefficient of friction changed when the condition of the substances was not the same. It was mainly the heat therefore, he thought, which was the cause of that diminution in the coefficient of friction, as the time went on. Mr. George Westinghouse, Jun., had long ago observed a plain- edged disc cutting through large iron beams in a rolling mill, and had seen that, without getting heated itself, it would cut through 1 2 or 1 4 inches of iron without difficulty. That had led him to the opinion that there was a difference between friction at high and at low speeds : for such a disc, running at a circumferential speed of c 000 ft. per minute, would scarcely cut anything, but at 12,000 ft. per minute it would cut hard steel. He had seen hard steel thus cut 167 214 RAILWAY BRAKES. APRIL 1879. with a piece of very soft wrought iron. His own idea in regard to the decrease in friction owing to time was that, instead of acting as a lubricant, the metal that was cut off might act as little rollers, in that way reducing the friction. It was very difficult however to ascertain the true cause, because it was impossible actually to see what took place. ^*ith regard to the friction regulating valve described in the paper, there was of course considerable doubt whether there would be a necessity for its use with continuous brakes, as the conditions varied to a great extent ; but he was satisfied from what he had seen that if the friction on the blocks was limited to about .18 of the load on the wheels, skidding would be avoided altogether ninety-nine times out of a hundred, and the average results would be much better. But at the end of the stop, where the friction was immensely increased by changing to static friction, the use of the regulating valve would prevent almost altogether the unpleasant shock so often noticed. It was especially observed on the Paris and Lyons line, by the people who were in the carriages fitted with the valves, that there was no shock whatever at the end of the stop. Some further trials had been made by M. Marie a few days after those alluded to in his paper. In these later trials the valves had been discontinued for three stops, and then used for four other stops; and the result was rather in favour of the apparatus without the valves, the stops being made from 50 miles an hour in about 10 or 12 yards less distance. That was accounted for however by the brakes going on more slowly when the valves were used, and by the fact that the valves did not open except during the last two or three seconds, the train coming to a stop without the slightest shock. That was a very important matter he thought, because it would diminish the strain upon the couplings and the brake-gear, so that all those parts might be made much lighter. Apart from that, it was simply a question of instantaneous action and of a proper degree of pressure. In practice however it was found that, without a regulator of some kind, if there was a very high pressure and a long leverage, the wheels were skidded so easily at low speeds that for practical work the leverage had to be reduced ; and then at high speed there was not sufficient 168 April 1S71J railway hrakkn. 215 power. If the brake-power could lie regulated by the valve, the little additional expense, of £\ per carriage at most, would be worth incurring ; for if it only saved the wheels from skidding on one occasion when there was risk of an accident, it would pay for the whole cost. Mr. Samuel Firth asked if Captain Galton was of opinion that in all cases an automatic brake should be used with its full power for every stop, whether a train was running into a station at 3 miles an hour, or stopping for danger from 60 miles an hour. He thought himself there were many cases when a brake should be put on gently bv hand, and that the automatic arrangement should always be kept in reserve for an emergencv ; and that the brake should at the same time show if it was out of order. The wear and tear of the rolling stock must be very great if a great many stoppages were made at intervals of only a few minutes, and if the full power were ahvays applied. Mr. T. Hurry Riches asked whether the results obtained by Captain Galton were upon steel or iron rails. It appeared to him that greater brake-block pressure might be used if the resistance of the rail to the wheel were greater, as would be the case with iron rails, practically tending to make the wheel rotate more persistently by the increase of friction between the wheel and rail. Captain Douglas Galton said in reply to Professor Kennedy's question as to the diminution in the coefficient of friction with time, that the experiments were necessarily limited, from the construction of the dynamometers, to about 30 seconds; and he could onlv state therefore what the diminution in the coefficient of friction was during that short time. Of course a portion of that time was consumed in putting on the brake-block pressure to begin with. As far as his experience went, he thought it was highly probable that the law suggested by Professor Kennedy did prevail, namely that after a certain time the coefficient of friction would become uniform. Some of his results seemed to tend very much in that direction, as shown 1 69 2l6 RAILWAY liRAKES. APRIL 1S79. by the plotting out of a few of his experiments in full lines on Fig. 14, PL 23. As to the reason why the coefficient of friction diminished with time, he confessed he was not quite clear. He was very much disposed to think that it was from the heat ; because if the brake-block pressure were taken off, and suddenly reapplied, exactly the same coefficient of friction was again obtained as at the beginning. It seemed questionable whether it could be from the abraded particles of iron getting between the brake-blocks and the wheel. It was very difficult to make delicate experiments of that nature with a train moving at 60 miles an hour. If those particulars were to be ^ascertained, it would have to be by means of experiments on a smaller scale, such as might be carried out by the Research Committee of the Institution. As to the question whether automatic brakes should always be applied with their full power, he had endeavoured to explain that at a high speed a very high brake-block pressure should be applied •instantaneously, and that this pressure should be reduced as the train came to rest ; but of course in approaching a station he assumed that a driver would not be running his train at a very high speed, but would previously have shut off steam and diminished the speed in preparation for putting on the brake for stopping, and that therefore such a very high pressure would not be required. He certainly thought it would be desirable that no higher pressure than that required for the particular speed should ever be put on ; but he also thought that with continuous brakes it was more than ever necessary that the driver should have the means of knowing what was the actual speed at which he was travelling at the moment, and should not be obliged to depend upon his judgment. On the Brighton line he found that all the engines were fitted with a speed indicator of Mr. Stroudley's design, which, according to the measurements obtained from Mr. Westinghouse's indicator, measured the speed very accurately indeed. It was of the greatest simplicity, indicating the speed by means of water raised in a tube by centrifugal action ; and as it did not record against the driver the speed at which he was running he had no inducement to put it out of order. 170 April 1S79. railway brakes. 217 In regard to Mr. Riches' question as to the proportion between the brake-block pressure and the weight on the wheels, — whether a smaller proportion would be required in the case of steel rails, — the rails on the Brighton Railway, upon which the experiments specially alluded to in this paper were made, were all steel rails, so that he had no data for answering that question. The President asked the members to accord their thanks to M. Marie and Captain Galton : to the former for the trouble he had taken, not only in trying the experiments, but also in coming so long a distance to give the members the results of them. With regard to Captain Galton, he was sure that the public of England, as well as the Institution, was very much obliged to him for his devotion to the question ; and if there should be any enquiry by a Government Commission, the best wish he could form for the public was that Captain Galton should be one of those appointed to act upon it. Several papers had now been read bearing upon the pressure-brake, and he thought it was quite proper that they should have the opportunity of hearing about the other system. He hoped therefore, after what had passed, that they would soon have an opportunity of hearing a paper on the Vacuum brake ; and at the same time he trusted they would take up the great question whether an automatic continuous brake was desirable, and was desired by the railway authorities. If they could make up their minds on that point, it would clear a great many cobwebs out of the way. A more simple brake might then be made, instead of adding parts to be used sometimes for automatic and sometimes for manual application. The vote of thanks was passed by acclamation. M. Marie wished to thank the President and the members of the Institution for the welcome given to French engineers : and to take that opportunity of mentioning that Captain Galton's experiments were already considered in France as among the most useful investigations which had been made for a long time in matters of railway engineering. 171 o © "- ~ ~ i. © c 3 S ' « ?o '■'■' / S \ \ \ ?■ \ •" / \ / / y f z I f / X \ "> / ; \/ / / V ^ // = ^ // -^ < // o £ J / v/ - r / A _ s / / x !M s/ /f \ rt / »/ N - l /•J <^ - ■J iHf— """"N - — -^-o "£ ? i r =„ = p/nrfe is ~V-\A..V\, Mate 19. CO < DC m < < DC I- t O ? lu a u. LL LU 4 £ in s EFFECT OF RAILWAY BRAKES. IZlllfl rgefl See/ions of l'l-ietion-lteijitlut! mj f'ul re. Fig. 3. Air Outlet clo.ii (I. Fig. 3. Mr Outlet open ^ \ ,„ K 51 J ~Iua&SSSL '■y M From BraAc Vylinilev Incites i 8 (Proceedit>{/n Inst. M.li. 1870.) Smlf '/3 a Miles 2>. hour -m EFFECT OF RAILWAY BRAKES. IntUctttor DUigrn his from Rrake JE.r per intents. Vlate 21. 10 Sec. For exi*l ft nation see jtatje 170. (Froceediii(/s Inst. M.JV. 1879.) EFFECT OF RAILWAY BRAKES. Plate :.':.'. Imlhfttfir Diagrams from Brake Experiments. For explanation ace page 170. !<"'''*■ {Proceedings Inst. 3I.J1. 187!).) Miles per hour. RAILWAY BRAKES. Fig. 14. Coefficient of Friction as affected by Th, Plate .20 -- <" - — h«i X - X - \" s i v - _• .HI ; \ \ _| >-, — 1 - •J *\\ \ - -~ H~ .; :ia \K\ ^- -~. -. - - i ^ \ -^ -~^ "- ~-- — i — — ' - J- lr> 02 -- ^_ \ v- h - 1 •J \ \ -. ~~~. ... ?5 ^)ii \ ^ ~~ -.^ 1 i ~_ .or. "" ~~ — — j ' i 1 - i 5 10 15 30 ^J Figures at com in aire me nt of eurre rep resent Miles per hour. Full lines represent single uperi tnents. Dotted lines represent averriye results. Fig. 13. Stop fallen from Proc. 1S7S. Plate <;•». Fir/. 10. lGornr<;sB Fig. 17. Plan of Buffing and Draw Springs, Lyons Railway. ^-- — _ N , /< Of f»7 {Proceedings Inst. 7,1.11. 1STJ.) CO < DC > < .J < DC u_ O h- o UJ u. LL ill Plate 14. f.artlett &. company, I>e^h,\'H!-:s, Engravers and Printer 21-33 Rose St., New York. »iH