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).
On valve-spring design, discussing natural frequency, surge, stresses, and the effects of the number of coils.
| Identifier | ExFiles\Box 56\2\ Scan062 | |
| Date | 1st January 1929 | |
| -2- Notes on Valve-Spring Design (Continued from page 12) mass of the spring. However if the design is good, springs designed for the same frequency will all come in, or surge, within a few vibrations per minute of each other. Hence, the formula is valuable for comparative purposes. Design As Related To Natural Frequency From our hypothesis of the nature of valve-spring surge, the obvious inference is that it is good practice to so design valve-springs as to obtain the highest possible natural frequency. This has been found to be true, within limits. Other factors, such as load and stress, being equal, a high-frequency spring is desirable since its wave-length is so high, compared with the wave-length of the disturbing force, that the spring will not vibrate in itself at the lower camshaft-speeds. Another, and true, inference that can be drawn from the hypothesis is that surge, because of its nature, cannot be eliminated except by recourse to some sort of friction damper which might prove expensive and annoying. It must be expected that, in the present state of the spring art, a spring always will vibrate when the frequency of the disturbing force is an integral part of the frequency of the spring, provided the frequency of the disturbing force is high enough to cause the spring to vibrate of itself. This proviso suggests a possible solution of the problem. Why not so design valve-springs that their frequencies will be so high that they will not tend to vibrate until they have reached forced vibration or camshaft speeds in excess of those encountered in customary practice? This idea of improving valve-spring surge conditions by increasing the natural frequency of the spring can best be illustrated by the study of an actual problem. Data relative to the characteristics of two actual springs are given in Table 1. The spring whose characteristics are given in the column headed "Original" was found to have noisy vibrational periods when running in the engine, and, moreover, some cases of breakage in service were reported notwithstanding the low static-value of stress as shown in the table. It was proposed to supplant this spring by one whose characteristics are given in the last column of the table, and in which an endeavor was made to hold to the same approximate valve-loads but to increase the natural frequency of the spring. It may be of interest to mention briefly the way in which this higher frequency of the spring is obtained by design. First, in this case, the pitch-diameter of the spring is decreased considerably, which tends to increase the rate of the spring. But to obtain the correct valve-loading and, at the same time, a reasonable free length, it is important that the rate be not increased excessively, hence the diameter of the wire is reduced somewhat, but the net result is a considerable increase in the rate. Number of Coils Affects Rate It should be noted that the number of active coils also affects the rate and that this number is varied to vary the rate. The weight of the active mass in the spring is decreased by reducing the diameter of the wire, the pitch-diameter, and the number of active coils. The increase of rate together with the decrease of mass effects a substantial increase in the natural frequency of the spring in accordance with Ricardo's formula. The two springs mentioned were studied by means of the equipment designed for this purpose. Their surge characteristics were noted and a marked difference in favor of the redesigned spring was found. Surge periods in the original spring were noted at camshaft speeds as low as 560 r.p.m. From this speed upward periods were noted at intervals obtained by dividing the spring frequency by whole numbers. It was found, however, not only with this spring but also in other cases, that at some times a period of a certain camshaft-speed gave less surge and noise than the one at the next higher speed at which surge was noted. This is difficult to explain and does not harmonize with the hypothesis presented as to the cause of surge. Perhaps the reason for this seeming discrepancy lies in the fact that the period of greater surge occurs at a speed which may be critical for one or more parts of the valve-gear, and that it is the resonant vibration of these parts, in addition to the spring, that causes the increased noise. So far as the spring is concerned, the discrepancy does not invalidate the surge hypothesis, for, as pointed out, the observed periods corresponded closely, in a general way, with the calculated spring-frequency. The redesigned spring had no periods of appreciable magnitude until a camshaft speed of 1330 r.p.m. was reached, and at even this period the magnitude was so small that it was difficult to observe. How Surge Affects Spring Stresses Surge, if excessive, may cause the valve to float open and thus impede the operation of the engine or it may cause excessive hum and make an engine more noisy than it should be. In any case, it always causes, to a greater or less degree, a building up of stress which is deleterious to the fatigue-life of the spring. Upon this last premise we account for the phenomenal failures of valve-springs, as mentioned in the initial part of these notes. Motion pictures reveal that, as the spring is vibrating in its own period, certain coils may, at a given instant, be compressed solid, while others may be opened up beyond their normal | ||