From the Rolls-Royce experimental archive: a quarter of a million communications from Rolls-Royce, 1906 to 1960's. Documents from the Sir Henry Royce Memorial Foundation (SHRMF).
Study of valve-mechanism idiosyncrasies, focusing on spring vibration, surge, and methods of measurement.
| Identifier | ExFiles\Box 56\2\ Scan081 | |
| Date | 15th January 1929 guessed | |
| 4 VALVE-MECHANISM IDIOSYNCRASIES changes the length of the slot and of the image on the screen accordingly. The shutter is attached to the spring coil by means of a piece of stiff wire, G, soldered to the coil and guided through a small hole, H.{Arthur M. Hanbury - Head Complaints} The valve lens, I, and its support, J, are the same as those designated by the same letters in Figs. 5 and 6, and K is the valve. The combination of the change in image length with the motion across the screen produced by the rotating mirror throws an enlarged picture of the coil motion on the screen. Since this picture is repeated by every face of the octagonal mirror, the persistence of vision makes it appear stationary, and the vibration can be studied carefully. For permanent record, this picture is thrown on a sensitized film, as in the valve-lift indicator. It is realized that this method of studying spring vibrations has some disadvantages. The shutter wire which is attached to the spring coil has some weight and is subject to some friction in its guides, therefore it damps the spring-vibration slightly. This may lower the frequency of vibration, but a measurement showed the change in frequency to be not greater than 5 per cent. It is also possible for the wire to superimpose a motion of its own on that of the spring coil and thus distort the resulting picture. Both these difficulties can be minimized by careful selection and installation of the wire. On the other hand, this instrument shows and records the actual vibration of the spring coil on a scale large enough so that each movement of the coil can be studied. It shows the number of vibrations per cycle of valve lift, the intensity of vibration, the wave form and direction in relation to the valve lift, the dying out of the vibration during the valve-closed period, and, in general, the actual mode of vibration of the spring as a check against theory. Fig. 9 is a photograph showing the complete assembly of the valve-lift indicator and the spring-vibration indicator, ready for use. STUDIES WITH VALVE-LIFT INDICATORS The stroboscopic projector and the telescopic point-by-point indicator were always used together, the former for determining the speeds at which valve performance was not good and the latter for studying the diagram at these speeds. Comparisons were always made with the lift curve at a camshaft speed of 200 r.p.m. Fig. 10 shows a curve plotted from the readings taken with the telescope. The fine line and open circles represent the 200 r.p.m. camshaft-speed. The heavy line and filled-in circles were taken from observations at a camshaft speed of 1130 r.p.m. A study of these two lift-diagrams shows that valve performance at this higher speed is very poor. Bouncing occurs on the cam as well as on the seat. Results are much more easily obtained with the valve-lift indicator, and photographic records of the lift curve can be taken at a great variety of speeds. Fig. 11 shows the valve-lift curve at three different speeds, each superimposed on a lift curve for 200 r.p.m. of the camshaft. The upper record was taken at a camshaft speed of 1370 r.p.m. Severe bouncing on the cam and on the seat is seen. The measurements show that the height of the first bounce is approximately 0.027 in., and it extends over an angle of approximately 19 deg. The second bounce is not so high but is almost as long, and the fluttering of the valve does not cease until approximately 68 deg. after the valve should be seated. The middle record was taken at a camshaft speed of 1400 r.p.m. The curve for this speed does not depart appreciably from the 200-r.p.m. base-line. It represents a reasonably good valve-lift diagram. The lower record was taken at a camshaft speed of 1430 r.p.m., and again bad bounces are visible. The seat bounces are not so severe at this speed as those recorded in the upper record, but the bounces on the cam itself are worse. Bouncing of any exhaust-valve to the extent shown in the upper and lower curves of Fig. 11 causes a serious overlapping, which results in a loss of power. When this power-loss occurs in the neighborhood of the maximum road-speed, it is more noticeable than when it occurs at a lower speed, although the resultant damage may be greater at the lower speed because it is within the range of normal driving. Fig. 12 shows the valve-lift curves at speeds of 1200, 1000, and 780 r.p.m. of the camshaft, each superimposed on a 200-r.p.m. base-line. These are rather good valve-lift curves and are, of course, at speeds at which the company is principally interested. The causes of bad valve-performance seem to depend upon the spring characteristics in their relation to the cam characteristics. Whenever a valve bounces, either on the cam or on the seat, the spring is too weak for the condition under which it is operating. If the spring is simply too weak to oppose the acceleration force such as is calculated from the ordinary valve-lift diagram, the bouncing will become worse as the speed increases. From Fig. 11 it is seen that a valve-lift curve may be better at a higher speed. The spring force in this case is, therefore, high enough to meet the ordinary cam-acceleration requirements, but for some reason is reduced temporarily. What is commonly known as spring surge seems to be the cause. Spring surge—or, better, spring vibration—is a torsion wave transmitted through the wire so that the wave travels up and down the spring. To explain why a spring surges is more difficult. Authorities agree that the spring gets into resonance with a periodic motion of the engine, and the most likely of these periodic motions is that imparted by the camshaft to the other parts of the valve mechanism. There is no doubt that a state of resonance must exist; or, in other words, that the spring vibrates sympathetically. A simple explanation of the conditions required for a state of resonance, or sympathetic vibrations, can be found in any elementary textbook.¹ Before violent surging develops, the spring must be subjected to a periodic force having a period equal to the time taken for a wave to travel up and down the spring. This is the simplest type of spring vibration and causes the most severe surging. If the frequency of the driving force is twice as great, then the first overtone of the spring is excited, and a condition arises that is similar to a propagation of waves from both ends of the spring simultaneously. This results in the phenomenon of a “standing wave,” with a node or stationary point at the center of the spring. In other words, the spring appears to vibrate in halves. In like manner, standing waves of higher frequencies can be produced, but they are of progressively less importance. Calculation of the frequency of the spring seems, therefore, to be of the utmost importance in the design of a valve mechanism. A useful formula for determining the approximate frequency of a steel compression spring in terms of its dimensions, derived from a formula in an article on Valve Springs, by Andrew Swan², is F = 250 (d × √G) / (D² × N) (1) in which F is the number of complete vibrations per minute; d is the diameter of the spring wire, in inches; G is the modulus of rigidity, 11,500,000 lb. per sq. in.; D is the mean diameter of the spring coil, in inches; and N is the number of active coils. DISTURBING FORCE The only periodic force acting on the spring is that due to the periodic motion of the valve, that is, the valve lift. No engine speed is high enough so that the valve lift can ever come into resonance with the spring frequency. Any periodic motion such as the valve-lift curve, however, can be resolved into a series of harmonic components, which are sine curves of increasing frequencies. The ordinates of these sine curves, added together, equal the ordinates of the original valve-lift curve. When the order number of one of these harmonics multiplied by the camshaft speed in revolutions per minute is equal to the frequency of the spring, in vibrations per minute, a state of resonance results and surging occurs. Fig. 13 shows the harmonics of a valve-lift curve from the eighth up to and including the thirtieth. The lower harmonics are as a rule of a much greater amplitude, but the speeds at which they would come into resonance with the spring are far above the usual camshaft speeds, therefore they were omitted to avoid obscuring the more significant curves. Table 1 shows the maximum amplitude and the sign of each harmonic of six different cam contours. The sign of these harmonics has been arbitrarily chosen as positive in the direction of valve lift and negative in the opposite direction. From Fig. 13 it is seen that these harmonics have widely different amplitudes. In general, the higher ones have a smaller amplitude than the lower ones, although this is not true ¹ See A First Course in Physics, by R.{Sir Henry Royce} A.{Mr Adams} Milliken and H.{Arthur M. Hanbury - Head Complaints} G.{Mr Griffiths - Chief Accountant / Mr Gnapp} Gale; Ginn & Co. ² See The Automobile Engineer, August, 1926, p. 290. FIG. 6—PHOTOGRAPH OF MACHINE WITH VALVE-LIFT INDICATOR FIG. 7—DIAGRAM OF SPRING-VIBRATION INDICATOR FIG. 8—PHOTOGRAPH OF SPRING-VIBRATION INDICATOR FIG. 9—COMPLETE MACHINE, WITH SPRING-VIBRATION INDICATOR IN PLACE | ||