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).
Paper on the design and failure analysis of valve springs.
| Identifier | ExFiles\Box 56\2\ Scan059 | |
| Date | 1st January 1929 | |
| January, 1929 X290 11 Notes On Valve-Spring Design Why not design valve springs with frequencies so high that they will not vibrate until they have reached forced camshaft speeds? By W. T. Donkin Cleveland Wire Spring Company* WHEN automotive valve-springs fail after comparatively short service it is common practice to investigate the failure as being due to one or more of three causes: (a) defective wire, (b) improper heat-treatment or (c) faulty design. Unfortunately, the trouble cannot always be traced back to one of these causes, hence solution of the problem is not always simple. Investigations of failures often reveal that the wire was free from defects, that the heat-treatment was correct, and that the stress and stress range existing in the spring, as calculated from the conventional spring formulas, were of a low order. Analysis of Failures These failures, so difficult to classify as to cause, present several peculiar features. Life of the springs has been found in general to be somewhat proportional inversely to the camshaft speed. This proposition should be accepted with reservations, however, for reasons which will be explained later. Another fact, which no doubt has been noted commonly, is that, from a given number of springs made from the same material under conditions seemingly identical, some fail after a very short life, whereas others give good service for a long period. This phenomenal type of failure after comparatively short service nearly always occurs as a fracture in the region of 1 to 1 1/2 coils from the end of the spring. Microscopic examinations of these fractures disclose them to be fatigue failures. The fracture usually makes an angle of 30 deg. with the axis of the wire and, as a rule, numerous secondary or “incipient” fatigue cracks, making the same angle with the axis, appear in the vicinity of the fracture. ABNORMAL compression and opening of coils cause the existence of a stress and stress range much greater than those calculated by conventional formulas. Furthermore, the stress range is passed through in a time equivalent to that required for the completion of one wave of the natural period of the spring. The combination tends to accelerate fatigue and is a prolific cause of valve spring failure. It may be contended that these fatigue failures are due to the fact that stress of a higher order than that calculated through the use of static formulas existed in the neighborhood of the fractures. If such a contention is permitted, this stress concentration is definitely linked with the rate of forced vibration of the spring, the natural frequency of the spring, and such factors as the action of inertia forces due to cam acceleration and the weight of reciprocating valve-gear parts. In the following notes the parts that these variables may play in causing this stress concentration are discussed in some detail. Laws of Spring Design Valve-springs, like other springs, obey certain laws and can be designed accordingly. The first item to be considered is the rate of the spring in pounds per inch, in order to find the loads with the valve closed and open. The load in pounds per inch of deflection can be obtained from the formula. Rr = Gd⁴/8D³N where Rr = rate, in pounds per inch of deflection G = torsional modulus of elasticity, 12,000,000 lb. per sq. in. D = mean diameter of spring, in inches N = number of active coils After determining the rate, it would be well to check up the stress by the formula S = 8 PD / πd³ where S = stress, in pounds per inch P = load, in pounds D = mean diameter of spring, in inches d = wire diameter, in inches Experience has set a value of about 60,000 lb. per sq. in. for the maximum stress and 25,000 lb. per sq. in. as the maximum stress-range. There is a tendency today to treat the stress-range as being unimportant, yet tests and experience tend to show that stress-range is really more important than the maximum stress, up to the elastic limit. Diagnosis of the reason for this shows that, with a high stress-range, the material is put through a large angle of twist which tends to fatigue it quickly and open up any discontinuity in the metal, causing early failure. Therefore, high maximum-stress with a low stress-range is much less destructive, as the material is worked only slightly with each successive compression. Materials and Heat-Treatments Compared Material and heat-treatment also must be considered in valve-spring design. It used to be *An abstract from two papers presented at meetings of the Society of Automotive Engineers by Mr. Donkin and Mr. H.{Arthur M. Hanbury - Head Complaints} H.{Arthur M. Hanbury - Head Complaints} Clark and reprinted from the Society's Journal. | ||