Can shot-peening change the free-length of a compression spring? 1

[prdave00]

Can shot-peening change the free-length of a compression spring? Keep in mind our springs are not conventional wound/coiled.

I'm seeing part-to-part & lot-to-lot variability in free-length of our finished springs. Our supplier insists that shot-peening is a likely culprit. I'm not convinced (I suspect it's loose process control during the spring mfg process).

Details: Spring 'machined' from MP35N bar stock (45-55% cold-reduced). Machining process is proprietary to supplier, but they have indicated material is aged prior to 'machining' coiled section. Last step is shot-peening with tumbling process using S070H shot at .007A intensity. Shot-peeing is outsourced to a 2nd supplier.

I have to assume all lots of springs are from same raw material lot until supplier can confirm.

 

[israelkk]

I assume the spring manufacturer is Heli-cal.

Shot peening induce compressive stresses to the surface therefore, uneven coverage may cause dimensional changes. How the dimensional changes affected your spring properties? Is it a compression or a torsion spring?

Although the manufacturing process is proprietary to supplier he can do it in three ways, milling, disc cutting. or wire EDM. When a small gap between the coils is needed only disc and wire EDM are possible. I designed and manufactured many machined spring using EDM and milling.

Two important points consider:

If it was machined by wire EDM there is a high risk of hydrogen embrittlement. You mentioned that the spring was machined after heat treatment which I assume is for high strength ~50 RC. I recommend you ask the manufacturer to be absolutely sure that the machining process of the coils process didn't produce hydrogen. If it is then you have a problem.

If you want to check the whole lot of springs you need to load them to the maximum expected load and hold them at that position for at least 200 hours. Then check for cracks (if they still in one piece).

The correct process is to wire EDM machine the coils before heat treatment and then use a fixture during heat treatment if dimensional change is absolutely required. There is always a chance that the spring properties will still be in the required range even with small dimensional changes.

If it was machined by a grinding disc then again there is a chance of micro cracks in the grounded area. Make sure the grinding/cutting was made according MIL-STD-866C Grinding of chrome plated steels and high strength steels heat treated to 180000psi or over. This spec requires grounded surface should be NDT checked to assure no cracks are present.

 

[israelkk]

There is one more machining option: diamond coated wire cutting.

 

[CoryPad]

I agree with israelkk that shot peening can lead to dimensional variation, but it usually is a small effect. How much variation do you see?

MP35N is a corrosion resistant alloy that has higher resistance to hydrogen that most alloys.

 

[israelkk]

We a an EDM wire machined spring made of Custom 455/465 H1000 that shutterd to pieces after one loading.

 

[israelkk]

Sorry for the typo error. It should be "We had an EDM ....".

 

[TheTick]

Given that most springs have a variance of ±15%, minor variations due to shot peening are likely to be insignificant in comparison.

 

[swall]

What is the section size of the spring? S70 shot is pretty small and how much effect it could have on distortion depends on how small your spring is.

 

[TVP]

I've been waiting to comment on this issue because I thought that the previous comments might elicit some additional information from the OP. Like the others, I am skeptical that shot peening is the cause of significant free length dimensional change, unless the parts have a small wall thickness/wire diameter.

 

[prdave00]

Sorry about the radio silence, but wanted to get my ducks in a row first.

Re. the amount of variation. First I want to clarify that this is more of a lot-to-lot variation issue. The part-to-part variation within a lot variation is pretty good. Magnitude of the range of overall spring length for Lot A = .004”, Lot B = .002”, & Lot C = .010”. The max/min length values for the first two lots spanned the upper half of the .010” tolerance band, while the median of length value for non-conforming Lot C was hovering just above the lower specification limit.

The raw material lot was different for Lots A/B versus Lot C. Yield and UTS of the raw material were within 2% of each other and %EL was within 10%. Although it would be convenient to blame the raw material, I can’t convince myself to do so given the meek difference in mechanical properties.

To address Swall's query, the cross-section of the ‘coil’ is .033 x .056”.

There is a question at

http://www.shotpeener.com

relating experience with the height of wound springs changing after shot-peening, but I’m not sure if the observation is applicable to my situation. See

http://www.shotpeener.com/ubb/ultimatebb.php?ubb=get_topic;f=1;t=000065#000000.

 

[swall]

prdave00--I suspect your problem has to do with variation in residual stress present in the bars from lot to lot. This would be best addressed by stress relieving the bars prior to machining. The shot peen issue discussed in The Shotpeener is well known--springs will take a set when cold pressed after shot peening unless they are stress relieved at 400-450F . After understanding your "wire" section size, I don't thing the peening with S70 shot has anything to do with it.

 

[prdave00]

According to Latrobe, there is no recommended, specific, effective stress relief treatment schedule for MP35N. But suggested the following: 'If stress relief is carried out at a temperature above 850F or so on non-aged, cold worked MP35N, the precipitation hardening (aging) reaction will occur. If a stress relief heat treatment is followed by another aging heat treatment at a higher temperature the effect of the prior stress relief heat treatment on the final hardness and strength properties will probably be minor, but it also may not be that effective in reducing residual stresses, unfortunately.'

There is some hope buried in Latrobe's response, but no empirical data to support it. I definitely don't want to do anything the would be detrimental to the strength of the material.

I'd like to hear from any other 'experts' on MP35N.