Spring stress calculations - unique compression spring loading 1

[mmacconney]

I have a compression spring that is failing in application. The spring meets industry standard design criteria for axial loading. However, in application this spring also has a moment load generated as the opposite ends rotate relative to each other (see attachment). The questions I have are:
1. How do I calculate the increase in stress throughout the range due to the "pushing" effects on the coil?
2. How do I calculate the load change if the spring helix were in the opposite direction. The load requirements on the spring are important for the design performance.

 

[israelkk]

This is a combined case of compression spring and torsion spring. The disadvantage is that the coil is loaded in the "coil diameter increase" mode. Torsion spring creates bending stresses (tension and compression stresses) and the compression spring created shear stresses. From my experience there is no public information and practical rules regarding such springs but you can try to use Von Misses criteria based on the shear and tensile stresses. Alternatively you can use FEA to analyse the Misses stresses in the spring. The problem will be to what spring wire strength data to compare the Misses stresses.

Spring formulations are based on analytical and long time empirical experience but in your case you can not use that. You need to treat your spring as any mechanical products but you can not rely on a any spring specific knowledge.

If you can provide detailed information of the spring dimensions, loads, deflections, etc. and mode of failure may be it will be possible to analyse the spring independently as compression and torsion spring and see which mode is more detrimental to the failure.

 

[unclesyd]

israelkk,

We had a similar problem on some of our textile type machines where we were experiencing the same type of spring failure as the OP. This problem was eliminated by initially installing a thrust bearing and if our recall correctly it ended up as a bronze thrust washer.

Could this type arraignment possibly help the OP?

http://www.qbcbearings.com/BuyRFQ/ThrustB_LoadRatings.htm

 

[mmacconney]

I appreciate the responses. I am testing now with a roller thrust washer and two flat thrust washers top and bottom. This is to validate the twist as the root cause. So far, we have improved cycle life by 60% and still increasing. This solution is not practical in real life due to cost and the performance affect. I also neglected to mention the spring twist is 44 degrees. I have included the spring specs below for anyone who wants to help estimate the stresses. Our tier two supplier is making springs with a LH helix vs. the current RH for trial. I am interested in a method to calculate the life improvement and the load change throughout the stroke if I use the LH instead of the RH to see if we are wasting our time with samples. The spring is used in a mechanical CVT where the load balance from the primary to the secondary is critical for vehicle performance.

Spring Specifications
Ends: Closed ground
Grade: Commercial
Wire: Chrome Silicon 280 ksi min
Wire dia: .192 in
Rate: 50 lb/in
Spring Index: 13.0625
Mean coil diameter: 2.508 in
Coil OD: 2.7 in
Active coils: 2.477
Total coils: 4.477
Pitch: 1.5408
Pitch angle: 11.0649
Free Length: 4.2 in
Load point 1 at installed length: 105 lbs at 2.1 in (stress 105.1 ksi considering compression only)
Load point 2 at full travel length: 162.5 lbs at .95 in (stress 162.6 ksi considering compression only)

 

[israelkk]

I think I misunderstood your problem. In the drawing you posted the spring is not closed (squared) and ground but just closed. I thought you intentionally applied the torsion load added to the compression load on the edge of the wire. Now I understand that the "torsion load" is actually a parasite load.

Theory says that if the two spring edges are fixed and can not rotate the spring rate will be ~25% higher than a case where at least one end can freely rotate. In your case the spring can rotate but a friction torque doesn't allow it to freely rotate. Therefore the rate will be higher than the calculated rate but below the 25% extra.

My quick calculations show that this spring should be presetted/scragged and supposes to give ~12000 cycles if repeatedly loaded between the two points of loads. This is a conservative estimation but it indicates that this spring has "limited" life cycle. Any scratch, dent, tool mark or surface imperfection will cause a large shift in the life cycle of this spring. The calculations assumed only compression without any torsion load. Any additional load due to friction will seriously decrease the calculaetd life cycle.

You should note that with fatigue calculation you need to assume a factor of 4 to 10 to be sure that all batches of spring will survive the designed life. Therefore, you can only guaranty ~1200 to ~3000 cycles.

You may need to test every batch of production to see if it match the design goal.

 

[israelkk]

It may be that you can get a better spring using a custom designed spring but you need to figure out the maximum space (spring maximum and minimum working lengths and maximum outside diameter or/and minimum inside diameter) that you can use.

 

[tbuelna]

mmacconney,

The torsional moment on the spring due to friction is about 20 in-lb at max compression (assuming a Mu=.10), and that would equate to a bending stress of about 29ksi in your coils. And that friction torque alone should produce a "twist" in your spring of about 23deg. At least according to my rough calculations.......

Terry

 

[unclesyd]

This is primarily a paper on rectangle cross section spring, but it does consider circular cross section

http://scholar.lib.vt.edu/theses/available/etd-05062002-004428/unrestricted/Etd.pdf

 

[mmacconney]

Thanks everyone for your responses. They have been helpful.
I ran a test this weekend with a needle bearing under the spring. So far, the spring has completed 150,000 cycles with the max load matched. I had to reduce the stroke 7mm due to the bearing/washer thickness. This is more than sufficient for this application. So this leads me back to my original questions.
- Will the LH helix reduce the stress/increase the cycle life? How much?
- What will be the load difference between the RH, LH given the same deflection with the 45 twist?
We are working on samples to test, but I would like to have a good idea of what to expect before the samples arrive.

 

[israelkk]

If you completed 150000 cycles then your actual spring wire probably doesn't have the minimum tensile properties that the spec allows. You should check the actual tensile strength of the spring wire and calculate the life cycle to see if this will give at least the 150000 cycles. Then you should calculate the life cycle for minimum tensile properties to see if it complies with your need.

LH should not change anything.

 

[tbuelna]

mmacconney,

With reversing sliding frictions, like you would have at your helical spring wire's end, there is a difference in frictions between the "approach" and "recess" actions. The friction created by the approach action (ie. where the sliding is in the direction of the wire end) will be higher than the recess action (ie. where the sliding is away from the wire end).

Other than that, the direction of helix should not matter, Since the loads are reversing.

Terry

 

[unclesyd]

I'm sitting here looking a caged ball and spring valve rotator off a big marine engine. Could this be used to take the rotation seen by OP spring? In other words could the rotator be made to take the rotation of the spring instead of rotating the valve?

 

[Cockroach]

Fixed end, does the other slide during compression movements? I can't see the issue here unless both ends are fixed and not allowed to slide. Then you have the added wire stretch load which I think is doing you in.

What is the end condition of the far end (lower)? Is this free to slide against the member?

Also, pass the numbers and geometry of the Spring. Maybe something a little more freaky (internal to your geometry).

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada