Effect of speed on fatigue life? 1

[meenogre]

Spring: about 5 in dia x 9 in tall barrel style coil spring using about .5 in wire. Fully compresed height is roughly 1.5 in.
Fatigue test: cycle spring from almost compressed height to about 4 in height at some rate until it breaks.
Question: How does that rate affect the life of the spring?

 

[israelkk]

Can you give more info? Are all the coils have the same diameter? If so then if the springs is squared and ground there is only one active coil?

 

[TVP]

Speed is only significant when the spring is near its resonant frequency. Engine valve toss is one example. Have you calculated the resonant frequency of your design?

 

[meenogre]

The 3" OD end coils are squared off so the thing stands up straight on the bench, then the coils expand out to ~5" OD middle coils.
I know that it doesn't actually operate near the resonant frequency, but I was thinking about cycle testing as fast as 5 or 6 Hz. What does resonating actually do to the spring? I'm guessing inertial forces would add to the spring force pushing on whatever the spring is hooked up to. But how does that affect fatigue life of the spring?
OK, are there some numbers to crunch to figure out the resonant frequency? Or should I hit it with a hammer and time the wobbles?

 

[TVP]

efunda has a nice section on springs, including natural frequency:

http://www.efunda.com/DesignStandards/springs/spring_frequency.cfm

 

[meenogre]

These springs have gone through some type of stress-relief exercise like shot peening during manufacture. (No, I don't understand all this stuff - just enough to be ;-D dangerous.) I also know that brand new springs demonstrate a force output decline over the first X hours of their life. The rate of decline decreases to almost nil after several days. Since I'm talking about testing brand new springs, how does this new-spring force decay affect the life cycle testing? I wondered if the two phenomena (force decay & cycle fatigue) are related. OK, not if, but HOW?

 

[meenogre]

OK, I ran the numbers on frequency & came up with a resonance of around 90 Hz. I'm nowhere near that, so the only weird thing I can think of will be the total time of compression during the life cycle test.
Let me put it another way.
If four identical springs are compared before & after a life cycle test:
# action result
1 sits relaxed no change
2 sits compressed loses force
3 cycles quickly loses some force, fatigues
4 cycles slowy loses more force, fatigues differently?

 

[israelkk]

The new-spring force decay does affect the life cycle test but it will actually prolong the life because the force reduce cause a stress reduce. Therefore the spring will have longer life. However, can you tolerate the force reduce?

The force reduce can be a result result of a relaxation or simply because the stress was (higher than the yield shear stress where the spring actually gone into the plastic zone and yielded. The result is a minor preset (scrag). If this is the case you do not have a sufficient safety factor.

You should calculate the stress at the desired force and compare it to the material yield shear stress. It should not be more than 35% to 50% of the ultimate tensile stress depends on the spring material.

You described the spring as barreled spring, is the gap (pitch) between the coils constant? or it is variable? Is the 5" coil bottoms first or all the coils bottom (touch) each other at the same time when spring is pressed to solid?

What is the type of material the spring is made off?

 

[meenogre]

Wire is chromium silicon SAE 9254.
There are 4½ coils total, including the squared-off ends of about ½ coil at the small diameter.
I can live with losing some force.
My goal is to do cycle testing faster without compromising the results. Slower cycling will help the spring last longer? OK. But that means faster cycling will kill it sooner, invalidating my test?

 

[TVP]

As I mentioned previously, the speed at which the spring is cycled during testing is irrelevant below the natural frequency. israelkk is talking about force loss due to compression set, which is a totally different phenomenon.