Electropolishing springs

[dgallup]

I have some prototype springs made from ASTM A313 type 316 round wire that about 20% have broken tips after ~200 million cycles. 316 was chosen initially for low permeability, the spring is in the middle of a solenoid. These are quite small springs, 2.46mm OD x 8.8mm long, 0.457 mm wire. The corrected stress at maximum working load is less than 25% of UTS and the movement when they cycle is less than 0.2 mm. These springs have closed and ground end coils but no dead coils. The breaks are clearly fatigue failures. I want to pursue higher strength material (17-7, Inconel, Elgiloy, etc.) and adding a dead coil to better support the ends. Others have suggested electropolish but I have always been leery of it. It seems to me that to remove enough material to really smooth out the ground end would remove way too material and significantly alter the spring properties. I also question how consistent and repeatable the process is. Any chemical process like this seems to vary dramatically as the baths age and the processors always seem to wait too long before replacing them. So you get a couple of good lots of parts & then it all goes down hill.

Does anyone have any really good spring electropolishing experience to change my mind?

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[israelkk]

Can you post pictures of the broken tips and a picture of non broken spring?

 

[MikeHalloran]

Some of my friends tried electrolytic grinding of axle shaft bearing seats. The metal removal rate was terrific, and the shafts looked really good. ... but the fatigue life was severely decreased, because the process removed all the compressive stress trapped near the surface.

I'd expect electropolishing to similarly decrease the fatigue life of a spring, by removing the beneficial compressive stresses trapped near the surface.

I'd also expect it to affect the shape of the wire, because of the impossibility of building the complementary electrode in such a way as to produce a uniform current density over the circumference and length of the wire in a formed spring.







Mike Halloran
Pembroke Pines, FL, USA

 

[israelkk]

25% of UTS means 43.3% of ultimate shear (0.25/0.577=0.433). Unless spring wire surface is absolutely perfect (unreal) the local stress at any imperfection on the spring wire surface due to wire manufacturing quality, coiling tool markings, etc. the stresses probably are beyond the maximum shear stress that in theory guaranty infinite fatigue life. Looking at the spring wire surface using at least X10 magnification glass may reveal such defects.

 

[israelkk]

The fact that the springs last 200 millions cycles is probably due to:
1. The amplitude of alternating shear stress due to the ~0.2 mm deflection is small compared to the 43.3% of Ultimate Shear Stress
2. The wire actual UTS is larger than the minimum UTS allowed by the wire spec.

Can you give more details of the spring, i.e. total number of coils, number of active coils, load and spring length at first load point when installed in the solenoid, wire spec (ASTM, MIL-SPEC, AMS, etc)?

 

[dgallup]

Here is a picture of the fracture surface.

Total coils = 12.6
Active coils = 10.6
Length 1 = 7 mm (this is the max the spring is ever compressed, typically a good bit less as the load is varied to adjust the solenoid response)
Load 1 = 7.9 N

The wire spec and other info is in the first post.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[israelkk]

Thank you for the info.

I was expecting a picture that can show if the wire surface contains defects such as tooling marks, dents, kicks, etc.

I checked the spring and found that according to standard design procedures such a spring is not legal. The stresses at maximum deflection are too high.

Did the spring gone special secondary operations? If not I an sure the spring free length after compression is shorter than the free length before load.

 

[dgallup]

It never gets compressed below 7 mm, it does not take a set. I don't care what the solid stress is, it's not a factor.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[israelkk]

What will happen if one will pinch it between his fingers? If the spring force was 500N I may understand your reasoning. But, this spring force is very small. You can not guaranty that the solenoid assembler or any other person will unintentionally play with the spring and compress the spring to solid.

Just for the record, due to numerous accidents, any spring designed for the US navy must be designed such that even if compressed to solid it will not set.

According to my calculations the 7.9 N should occur at a nominal compressed length of 6.74 mm and the nominal free height should be 8.53 mm, rate = 4.42 N/mm. Maximum force after extra deflection of 0.2 mm is 8.78 N.

Taking all this into account still gives an illegal spring according to common standard spring design norms. The calculated stresses are too high for 316 SS wire at minimum UTS allowed by the spec.

 

[dgallup]

israelkk - Nobody plays with the springs, the assembly is automated. Even if the spring was compressed to solid and took a set prior to installation, I don't believe that would be detrimental to fatigue life.

I've looked at the springs at magnification up to 80x, I inspected every single one at 20x prior to build. The spring wire is very good quality, I do not see any defects visible at that magnification.

Can we get back to the question, would electro polish improve spring life?

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[israelkk]

I do not believe that electro polish will help on the contrary. It will remove material which no one to my experience can guaranty a constant amount and affect the spring rate and force because the spring rate is a function of the wire diameter powered by four. Add to this the risk of hydrogen embrittlement.

You still did not commented to the fact that the calculated stresses are beyond the recommended by any spring design reference especially the Spring Design handbook by SPEC. which is considered the Bible for spring design and is referenced even by the SMI.

 

[dgallup]

Thank you. Stress is why I plan on redesigning the spring in 17-7 or Elgiloy. I can add a dead coil and reduce the stress as a % of minimum tensile strength.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[dgallup]

Some update on this, the 17-7 springs have gone 616 million cycles without a single failure and continue to run. We will keep testing to ~1 billion cycles. We also have some elgiloy springs running but they only have 150 million cycles so far as the wire took quite awhile to procure and the spring vendor had a very hard time winding it.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[TVP]

Thanks for the update. Very interesting. Was the 17-7 just processed normally: spring coiling following by grinding, cleaning/passivation, and aging (477-488 C for ~ 1 hour)?

 

[dgallup]

Yes and done by the same supplier that made the 316 springs.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[dgallup]

We completed over 1 billion cycles on the 17-7 springs without a single failure.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.