chine bearings, is not directly proportional to the pressure, is not independent of the speed, and that the coefficients of Morin and Webber are about tenfold too great for modern journals. Prof. Denton offers an explanation of this apparent contradiction of authorities by showing, with laboratory testing machine data, that Moriu's laws hold for bearings lubricated by a restricted feed of lubricant, such as is afforded by the oil-cups common to machinery; whereas the modern experiments have been made with a surplus feed or superabundance of lubri cant, such as is provided only in railroad car journals, and a few special cases of practice. That the low coefficients of friction obtained under the latter conditions are realized in the case of car-journals, is proved by the fact that the temperature of car-boxes remains at 100° at high velocities; and experiment shows that this temperature is consistent only with a coefficient of friction of a fraction of one per cent. Deductions from experiments on train resistance also indicate the same low degree of friction. But these low co-efficients do not account for the internal friction of steam-engines as well as do the coefficients of Morin and Webber. In American Machinist, Oct. 23, 1890, Prof. Denton says: Morin's measurement of friction of lubricated journals did not extend to light pressures. They apply only to the conditions of general shafting and engine work. He clearly understood that there was a frictional resistance, due solely to the viscosity of the oil, and that therefore, for very light pressures, the laws which he enunciated did not prevail. He applied his dynamometers to ordinary shaft-journals without special preparation of the rubbing-surfaces, and without resorting to artificial methods of supplying the oil. Later experimenters have with few exceptions devoted themselves exclusively to the measurement of resistance practically due to viscosity alone. They have eliminated the resistance to which Morin confined his measurements, namely, the friction due to such contact of the rubbing-surfaces as prevail with a very thin film of lubricant between comparatively rough surfaces. Prof. Denton also says (Trans. A. S. M. E., x. 518): "I do not believe there is a particle of proof in any investigation of friction ever made, that Morin's laws do not hold for ordinary practical oil-cups or restricted rates of feed." Laws of Friction of well-lubricated Journals.-John Goodman (Trans. Inst. C. E. 1886, Eng'g News, Apr. 7 and 14, 1888), reviewing the results obtained from the testing-machines of Thurston, Tower, and Stroudley, arrives at the following laws: LAWS OF FRICTION: WELL-LUBRICATED SURFACES. (Oil-bath.) 1. The coefficient of friction with the surfaces efficiently lubricated is from 1/6 to 1/10 that for dry or scantily lubricated surfaces. 2. The coefficient of friction for moderate pressures and speeds varies approximately inversely as the normal pressure: the frictional resistance varies as the area in contact, the normal pressure remaining constant. 3 At very low journal speeds the coefficient of friction is abnormally high; but as the speed of sliding increases from about 10 to 100 ft. per min the friction diminishes, and again rises when that speed is exceeded, varying approximately as the square root of the speed. 4. The coefficient of friction varies approximately inversely as the temperature, within certain limits, namely, just before abrasion takes place. The evidence upon which these laws are based is taken from various modern experiments. That relating to Law 1 is derived from the "First Report on Friction Experiments," by Mr. Beauchamp Tower. With a load of 293 lbs per sq. in, and a journal speed of 314 ft. per mir. Mr. Tower found the coefficient of friction to be .0016 with an oil-bath, and 0097, or six times as much, with a pad. The very low coefficients obtained by Mr. Tower will be accounted for by Law 2, as he found that the frictional resistance per square inch under varying loads is nearly constant, as below: 205 Load in lbs. per sq. in..... 529 468 415 363 310 258 153 100 .458 .45 The frictional resistance per square inch is the product of the coefficient of friction into the load per square inch on horizontal sections of the brass. Hence, if this product be a constant, the one factor must vary inversely as the other, or a high load will give a low coefficient, and vice versa. For ordinary lubrication, the coefficient is more constant under varying loads; the frictional resistance then varies directly as the load, as shown by Mr. Tower in Table VIII of his report (Proc. Inst. M. E. 1883). With respect to Law 3, A. M. Wellington (Trans. A. S. C. E. 1884), in experiments on journals revolving at very low velocities, found that the friction was then very great, and nearly constant under varying conditions of the lubrication, load, and temperature. But as the speed increased the friction fell slowly and regularly, and again returned to the original amount when the velocity was reduced to the same rate. This is shown in the following table: Speed, feet per minute: 0+ 2.16 3.33 4.86 8.82 21.42 35.37 53.01 89.28 106.02 Coefficient of friction: .118 .094 .070 .069 .055 .047 .040 It was also found by Prof. Kimball that when the journal velocity was increased from 6 to 110 ft. per minute, the friction was reduced 70%; in another case the friction was reduced 67% when the velocity was increased from 1 to 100 ft. per minute; but after that point was reached the coefficient varied approximately with the square root of the velocity. The following results were obtained by Mr. Tower: Feet per minute... 209 262 314 366 419 471 Nominal Load per sq. in. The variation of friction with temperature is approximately in the inverse ratio, Law 4. Take, for example, Mr. Tower's results, at 262 ft. per minute: This law does not hold good for pad or siphon lubrication, as then the coefficient of friction diminishes more rapidly for given increments of temperature, but on a gradually decreasing scale, until the normal temperature has been reached; this normal temperature increases directly as the load per sq. in. This is shown in the following table taken from Mr. Stroudley's experiments with a pad of rape oil: Temp. F Coefficient. Decrease of coeff.. 105° 110° 115° 120° 125° 130° 135° 140° 145° .022 .0180 .0160 .0140 .0125 .0115 .0110 .0106 .0102 .0040 .0020 0020 .0015 .0010 .0005] .00041 .0002 In the Galton-Westinghouse experiments it was found that with velocities below 100 ft. per min., and with low pressures, the frictional resistance varied directly as the normal pressure; but when a velocity of 100 ft. per min. was exceeded, the coefficient of friction greatly diminished; from the same experiments Prof. Kennedy found that the coefficient of friction for high pressures was sensibly less than for low. Allowable Pressures on Bearing-surfaces. (Proc. Inst. M. E., May, 1888.)-The Committee on Friction experimented with a steel ring of rectangular section, pressed between two cast-iron disks, the annular bear ing-surfaces of which were covered with gun-metal, and were 12 in. inside diameter and 14 in. outside. The two disks were rotated together, and the steel ring was prevented from rotating by means of a lever, the holding force of which was measured. When oiled through grooves cut in eac. face of the ring and tested at from 50 to 130 revs. per min., it was found that a pressure of 75 lbs. per sq. in. of bearing-surface was as much as it would bear safely at the highest speed without seizing, although it carried 90 lbs. per sq. in. at the lowest speed. The coefficient of friction is also much higher than for a cylindrical bearing, and the friction follows the law of the friction of solids much more nearly than that of liquids. This is doubtless due to the much less perfect lubrication applicable to this form of bearing compared with a cylindrical one. The coefficient of friction appears to be about the same with the same load at all speeds, or, in other words, to be independent of the speed; but it seems to diminish somewhat as the load is increased, and may be stated approximately as 1/20 at 15 lbs. per sq. in., diminishing to 1/30 at 75 lbs. per sq. in. The high coefficients of friction are explained by the difficulty of lubricating a collar-bearing. It is similar to the slide-block of an engine, which can carry only about one tenth the load per sq. in. that can be carried by the crank-pins. In experiments on cylindrical journals it has been shown that when a cylindrical journal was lubricated from the side on which the pressure bore, 100 lbs. per sq. in. was the limit of pressure that it would carry; but when it came to be lubricated on the lower side and was allowed to drag the oil in with it, 600 lbs. per sq in. was reached with impunity; and if the 600 lbs. per sq. in., which was reckoned upon the full diameter of the bearing, came to be reckoned on the sixth part of the circle that was taking the greater proportion of the load, it followed that the pressure upon that part of the circle amounted to about 1200 lbs. per sq. in. In connection with these experiments Mr. Wicksteed states that in drilling-machines the pressure on the collars is frequently as high as 336 lbs. per sq. in., but the speed of rubbing in this case is lower than it was in any of the experiments of the Research Committee. In machines working very slowly and intermittently, as in testing-machines, very much higher pres sures are admissible. Mr. Adamson mentions the case of a heavy upright shaft carried upon a small footstep-bearing, where a weight of at least 20 tons was carried on a shaft of 5 in. diameter, or, say, 20 sq. in, area, giving a pressure of 1 ton per sq. in. The speed was 190 to 200 revs. per min. It was necessary to force the oil under the bearing by means of a pump. For heavy horizontal shafts, such as a fly-wheel shaft, carrying 100 tons on two journals, his practice for getting oil into the bearings was to flatten the journal along one side throughout its whole length to the extent of about an eighth of an inch in width for each inch in diameter up to 8 in. diameter; above that size rather less flat in proportion to the diameter. At first sight it appeared alarming to get a continuous flat place coming round in every revolution of a heavily loaded shaft; yet it carried the oil effectually into the bearing, which ran much better in consequence than a truly cylindrical journal without a flat side. In thrust-bearings on torpedo-boats Mr. Thornycroft allows a pressure of never more than 50 lbs. per sq. in. Prof. Thurston (Friction and Lost Work, p. 240) says 7000 to 9000 lbs. pressure per square inch is reached on the slow-working and rarely moved pivots of swing bridges. Mr. Tower says (Proc. Inst M E., Jan. 1884): In eccentric-pins of punching and shearing machines very high pressures are sometimes used without seizing. In addition to the alternation in the direction, the pressure is ap plied for only a very short space of time in these machines, so that the oil has no time to be squeezed out. In the discussion on Mr. Tower's paper (Proc. Inst. M. E. 1885) it was stated that it is well known from practical experience that with a constant load on an ordinary journal it is difficult and almost impossible to have more than 200 ibs. per square inch, otherwise the bearing would get hot and the oil go out of it; but when the motion was reciprocating, so that the load was alternately relieved from the journal, as with crank-pins and similar journals, much higher loads might be applied than even 700 or 800 lbs. per square inch. Mr. Goodman (Proc. Inst. C. E. 1886) found that the total frictional resistance is materially reduced by diminishing the width of the brass. The lubrication is most efficient in reducing the friction when the brass subtends an angle of from 120° to 60°. The film is probably at its best between the angles 80° and 110°. In the case of a brass of a railway axle-bearing where an oil-groove is cut along its crown and an oil-hole is drilled through the top of the brass into it, the wear is invariably on the off side, which is probably due to the oil escap ing as soon as it reaches the crown of the brass, and so leaving the off side almost dry, where the wear consequently ensues. In railway axles the brass wears always on the forward side. The same observation has been made in marine engine journals, which always wear in exactly the reverse way to what they might be expected. Mr. Stroudley thinks this peculiarity is due to a film of lubricant being drawn in from the under side of the journal to the aft part of the brass, which effectually lubricates and prevents wear on that side; and that when the lubricant reaches the forward side of the brass it is so attenuated down to a wedge shape that there is insufficient lubrication, and greater wear consequently follows. Prof. J. E. Denton (Am. Mach., Oct. 30, 1890) says: Regarding the pressure to which oil is subjected in railroad car-service, it is probably more severe than in any other class of practice. Car brasses, when used bare, are so imperfectly fitted to the journal, that during the early stages of their use the area of bearing may be but about one square inch. In this case the pressure per square inch is upwards of 6000 lbs. But at the slowest speeds of freight service the wear of a brass is so rapid that, within about thirty minutes the area is either increased to about three inches, and is thereby able to relieve the oil so that the latter can successfully prevent overheating of the journal, or else overheating takes place with any oil, and measures of relief must be taken which eliminate the question of differences of lubricating power among the different lubricants available. A brass which has been run about fifty miles under 5000 lbs. load may have extended the area of bearing-surface to about three square inches. The pressure is then about 1700 lbs. per square inch. It may be assumed that this is an average minimum area for car-service where no violent and unmanageable overheating has occurred during the use of a brass for a short time. This area will very slowly increase with any lubricant. C. J. Field (Power, Feb. 1893) says: One of the most vital points of an engine for electrical service is that of main bearings. They should have a surface velocity of not exceeding 350 feet per minute, with a mean bearingpressure per square inch of projected area of journal of not more than 80 lbs. This is considerably within the safe limit of cool performance and easy operation. If the bearings are designed in this way, it would admit the use of grease on all the main wearing-surface, which in a large type of engines for this class of work we think advisable. Oil-pressure in a Bearing.-Mr. Beauchamp Tower (Proc. Inst. M. E. Jan. 1885) made experiments with a brass bearing 4 inches diameter by 6 inches long, to determine the pressure of the oil between the brass and the journal. The bearing was half immersed in oil, and had a total load of 8008 lbs. upon it. The journal rotated 150 revolutions per minute. The pressure of the oil was determined by drilling small holes in the bearing at different points and connecting them by tubes to a Bourdon gauge. It was found that the pressure varied from 310 to 625 lbs. per square inch, the greatest pressure being a little to the "off" side of the centre line of the top of the bearing, in the direction of motion of the journal. The sum of the upward force exerted by these pressures for the whole lubricated area was nearly equal to the total pressure on the bearing. The speed was reduced from 150 to 20 revolutions, but the oil-pressure remained the same, showing that the brass was as completely oil-borne at the lower speed as at the higher. The following was the observed friction at the lower speed: 443 333 211 89 .00132 .00168 .00247 .0044 Nominal load, lbs. per square inch... The nominal load per square inch is the total load divided by the product of the diameter and length of the journal. At the same low speed of 20 revolutions per minute it was increased to 676 lbs. per square inch without any signs of heating or seizing. Friction of Car-journal Brasses. (J. E. Denton, Trans. A. S. M. E., xii. 405.)—A new brass dressed with an emery-wheel, loaded with 5000 lbs. may have an actual bearing-surface on the journal, as shown by the polic of a portion of the surface, of only 1 square inch. With this pressure of 5000 lbs. per square inch, the coefficient of friction may be 6%, and the brass may be overheated, scarred and cut but, on the contrary, it may wear down evenly to a smooth bearing, giving a highly polished area of contact of 3 square inches. or more, inside of two hours of running, gradually decreasing the pressure per square inch of contact, and a coefficient of friction of less than 0.5%. A reciprocating motion in the direction of the axis is of importance in reducing the friction. With such polished surfaces any oil will lubricate, and the coefficient of friction then depends on the viscosity of the oil. With a pressure of 1000 lbs per square inch, revolutions from 170 to 320 per minute, and temperatures of 75° to 113° F. with both sperm and parraffine oils, a coefficient of as low as 0.11% has been obtained, the oil being fed continuously by a pad. Experiments on Overheating of Bearings.-Hot Boxes. (Denton.)-Tests with car brasses loaded from 1100 to 4500 lbs. per square inch gave 7 cases of overheating out of 32 trials. The tests show how purely a matter of chance is the overheating, as a brass which ran hot at 5000 lbs. load on one day would run cool on a later date at the same or higher pressure. The explanation of this apparently arbitrary difference of behavior is that the accidental variations of the smoothness of the surfaces, almost infinitesimal in their magnitude, cause variations of friction which are always tending to produce overheating, and it is solely a matter of chance when these tendencies preponderate over the lubricating influence of the oil. There is no appreciable advantage shown by sperm-oil, when there is no tendency to overheat-that is, paraffine can lubricate under the highest pressures which occur, as well as sperm, when the surfaces are within the conditions affording the minimum coefficients of friction. Sperm and other oils of high heat-resisting qualities, like vegetable oil and petroleum cylinder stocks. only differ from the more volatile lubricants, like paraffine, in their ability to reduce the chances of the continual accidental infinitesimal abrasion producing overheating. The effect of emery or other gritty substance in reducing overheating of a bearing is thus explained: The effect of the emery upon the surfaces of the bearings is to cover the latter with a series of parallel grooves, and apparently after such grooves are made the presence of the emery does not practically increase the friction over the amount of the latter when pure oil only is between the surfaces. The infinite number of grooves constitute a very perfect means of insuring a uniform oil supply at every point of the bearings. As long as grooves in the journal match with those in the brasses the friction appears to amount to only about 10% to 15% of the pressure. But if a smooth journal is placed between a set of brasses which are grooved, and pressure be applied, the journal crushes the grooves and becomes brazed or coated with brass, and then the coefficient of friction becomes upward of 40%. If then emery is applied, the friction is made very much less by its presence, because the grooves are made to match each other, and a uniform oil supply prevails at every point of the bearings, whereas before the application of the emery many spots of the latter receive no oil between them. Moment of Friction and Work of Friction of Slidingsurfaces, etc. |