Tech-Spring Report 5A THE STRESS PROFILE IN LARGER WIRES
Introduction Three batches of compression springs were supplied to IST to investigate the effect of the stress profile in cold-coiled springs made from relatively large wire. The springs supplied and information from their respective wire test certificate were as follows: Batch VS VBlue VL Wire 5.09mm 7.49mm 7.98mm Material Oil hardened SiCr Oil hardened SiCr Induction hardened SiCr T.S. 1843MPa 1730MPa 1874MPa
IST fatigue tested these springs in order to compare their relative performance, and when we found significant differences, the springs were sent for evaluation of their residual stress profiles. This interim report is supplied only to the supplier of the springs, but will be sent to all project partners once the residual stress data is available.
Fatigue Testing Each batch of springs was accurately load tested so as to establish the relationship between load and stress. They were then fatigue tested to establish the stress conditions necessary to cause failure. They were tested on mechanically driven machines at 380rpm, a speed that would not result in dynamic stresses being additional to those calculated. The results were: Batch VS Load Range / N Stress Range / MPa 111 - 1109 100 - 1000 55 - 1109 50 - 1000 Life to Failure 2 survived one million Survived one million
This batch could not be failed by fatigue. The solid stress was 1,170MPa, and testing closer to solid than 1000MPa would have given unreliable results because of the risk of coil contact. Batch VBlue Load Range / N Stress Range / MPa 253 - 2393 100 - 945 253 - 2532 126 - 2532 100 - 1000 50 - 1000 Life to Failure 4 survived one million 3 survived one million & 1 failed at 645k 4 failed at 477k - 944k
The fatigue performance of this batch, like the VS batch was good - ISTs software would predict a risk of failure after about 200k when tested between 100 - 1000MPa.
1
L:\PROJECTS\810CollectiveResearchProg\90331 No 5A
Batch VL
Load Range / N Stress Range / MPa 235 - 2246 100 - 950
Life to Failure 1 survived one million, 3 failed at 327k, 391k and 561k.
The fatigue performance of this batch was relatively poor - ISTs software would predict a risk of fatigue failure after about 400k cycles when tested between 100 - 950MPa. Hence the fatigue testing has shown that the fatigue performance can be ranked VS better than VBlue better than VL.
Optical Metallography Transverse and longitudinal sections were prepared from a tested spring of each type. This was done to check whether there were metallurgical differences that could account for the fatigue life differences. The results were: Batch VS Microstructure Tempered martensite Defects / Decarb No defects, slight partial decarburisation, usual effect of shot peening No defects, slight partial decarburisation, shot peening very light. No defects, slight partial decarburisation, shot peening o.k., longitudinal oxidised crack (Figure 1) Hv 560 / 579
VBlue
Tempered martensite
525 / 525
VL
Tempered martensite
554 / 592
Figure 1
x 430
It would usually be expected that a higher hardness would bring better fatigue performance, hence VS better than VBlue. Shot peening is vital to give the higher fatigue performance
2
L:\PROJECTS\810CollectiveResearchProg\90331 No 5A
observed here and the VBlue springs appear to have been peeened with very small shot compared with the other two. Nonetheless, the VBlue springs have given good performance. The larger longitudinal and oxidised crack in the VL spring maybe a full explanation for its relatively poor fatigue performance, but IST have not observed cracks of this type previously and are in the process of examining untested springs for evidence of similar cracks. With this type of crack the fatigue performance would have been expected to be very much worse than that observed, if the crack has been present before the start of the test. It is not clear why this spring should have developed a longitudinal crack during fatigue test, the crack and fatigue origin was at the outside surface of the spring which is unusual springs of this type should have fatigue initiation at the inside surface, as was observed in the VBlue springs.
Residual Stress Results
Stress profile under max. load (calculated values)
Residual stress profile after shot-pening without load (mean values of three measured springs)
3
L:\PROJECTS\810CollectiveResearchProg\90331 No 5A
Resulting stress profile under max. load (calculated values)
4
L:\PROJECTS\810CollectiveResearchProg\90331 No 5A
Additional batch of Stainless Steel Barrel shaped springs These springs had been manufactured from EN 10270-3 1.4301 wire and had been shot peened, but the appearance of the shot peened finish, shown as figures 2 and 3 was variable, which curtailed this test programme. Innotech photographed the shot peened finish shown in figure 2 on the scanning electron microscope and expressed doubt that it was satisfactory, but the fatigue performance was good and so this topic was left for future exploration. The spring design enabled IST to illustrate the printout from their non-standard compression spring program for the first time in this project, as figure 4. The fatigue test results on the well peened springs figure 2 52 804 MPa stress range Two springs survived 10 million cycles The fatigue test results for the springs figure3. 52 804 MPa stress range Two springs failed at 82,820 and 292,370 cycles
Figure 2 100% peening coverage
x 11.5 Figure 3 < 50% peening coverage
x 11.5
These photographs show that peening has not obliterated the coiling scratches on the inside surface of these springs.
5
L:\PROJECTS\810CollectiveResearchProg\90331 No 5A
Figure 4 Barrel shaped spring design
6
L:\PROJECTS\810CollectiveResearchProg\90331 No 5A
