Embrittled 17-7

[swall]

One of our plants supplies a Belleville made from 17-7 bar, 3" in diameter, with thickness of .25" thick. It is heat treated to the TH1050 condition. So far, so good. But, for some yet to be determined reason, the plant gives these a supplemental age of 1060F for 3 hours. A high percentage of the parts crack as they are loaded up in the assembly. My question: is this overage the plant is performing embrittling the 17-7?

 

[metengr]

swall;
I will ask the obvious question, have you or someone performed a proper metallurgical analyis of the failed parts?

 

[metengr]

You might be interested to review this publication

http://oai.dtic.mil/oai/oai?verb=getRecord&metadataPrefix=html&identifier=AD0287893

 

[swall]

I have not performed one of my own, but have reviewed SEM and metallographic work performed on 3 of the cracked washers. The washers have cracked in the radial direction, so being machined from bar, the fractures are in the transverse direction and as expected, appear "woody". The morphology is a mix of dimple rupture, quasi cleavage and isolated IG. The fractures on 2 partially cracked springs show initiation on the I.D. which of course in a Belleville is the area of highest stress. The micros show banding of elongated delta ferrite stringers. There are also possible fine carbides in the grain boundaries, although this may merely be an artifact of an overly aggressive electrolytic etch. This is a classic case of the metallurgist being brought in after the problem has already been going on for a year.

 

[metengr]

It may be that the washers in the overaged condition were too low in strength and would fail in the manner you described upon assembly. Was chemical analysis and micro- hardness testing performed and compared with the standard TH1050 heat treatment?

 

[swall]

Yes, chemistry was found to conform to the limits stated in the Armco publication for 17-7, ditto for hardness. Armco states Rc 43 typ, the incoming parts were Rc 41-43. The 3 additional hours at 1060F dropped the hardness to Rc 39-41. I am still trying to get clarification as to what the plant hoped to accomplish with the extra thermal processing, although sometimes spring people will do wierd things like that to try to change the spring rate.

 

[btrueblood]

Or maybe they did it because the parts were cracking with the original treatment too, and they thought the higher toughness of overaging would help?

 

[swall]

Although this is very limited sample size, six of the parts w/o the extra 3 hours @ 1060F were pressed to solid with no breakage. Three more are on hand to experiment with and will be pressed to solid, examined, and if not cracked, will be exposed to 1060F for 3 hours and pressed again. It takes 30,000 pounds to go to solid; bad parts can crack at a load of 18,000 pounds. I am not convinced this test will be conclusive, as the initial press to solid will change the residual stress pattern compared to as-received material.

 

[swall]

This experiment was performed and the parts did not crack either in the as received condition or after re-ageing at 1060F for 3 hours. Yes, I know, small sample size, not conclusive. The plant also told me that the extra 3 hours came about due to hardness of the original material not meeting customer spec. I have advised them that standard practice for 17-7 is to spec min UTS and TYS with hardness for reference only. But, I doubt the customer will change his requirement.

 

[israelkk]

Are these disks machined or grinded?
If they are grinded the grinding process may create micro cracks. See MIL-STD-866 GRINDING of Chrome Plated Steel and Steel Parts Heat Treated to 180,000 PSI or Over at

http://assist.daps.dla.mil/docimage...3CAA214B100C4&jsessionid=063067a6c48a3eb2f95b

The way to convince the customer to drop the hardness requirement is to show him that standard specs for the bars require only tensile properties. To my best knowledge hardness can vary up to ~10RC for same H1050 treatment with no change in tensile properties.

 

[EdStainless]

17-7 isn't like 17-4, over aging does not give a big increase in toughness.
I would look very carefully at any cleaning operations.
The parts need to be very clean before aging.
Also look to make sure that the ID bore and edges are smooth.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[mfgenggear]

it would be helpfull to know how parts are being manufactured. This type of material can be work harden.
maybe the post heat treat was a stress relieve?

it may be as Israelkk states there may be cracks in the parts. to remove any posibilities recomend to NDT flo pen.

if any parts are cracked before assembled then do a section & micrograph.

is it a material issue or a workmanship problem.
if these parts are process like double disk grind
or just turned with high feed rate inducing work hardening or stress. this can cause the above.

where parts ever pickled?

 

[swall]

The parts are cnc machined from 17-7 bar stock, condition A and are then given the TH 1050 age. That is the condition the plant receives them in. The parts are then tumble cleaned in an alkaline soap that also contains sulfonates. The parts were rinsed and were given the supplemental 3 hours at 1060F in an attempt to bring down the hardness. As the order has been completed, we will have to wait until next time before NDT of the incoming parts is undertaken.

 

[mcguire]

Any cracking under elastic loading of a martensitic material should always be suspect as hydrogen induced. Is any of the heat treat done in a atmosphere which has a hydrogen component? This is a very tough material. The fracture you describe sounds brittle. I don't know of any way to induce brittle fracture at room temperature at non-cyclic stresses below the yield strength, except by hydrogen.

Michael McGuire

http://stainlesssteelforengineers.blogspot.com/

 

[mcguire]

I should add that martensitic 17-7, after annealing in a hydrogen atmosphere, cannot be deformed without delayed fracture due to residual stress inducing hydrogen embrittlement. It takes a lengthy outgassing at room temperature before it is safe to form.

If you baked your parts at 500F for a few minutes in air to remove any hydrogen and could then assemble them without failure, you would know if hydrogen was the culprit.

Michael McGuire

http://stainlesssteelforengineers.blogspot.com/

 

[swall]

McGuire--After the standard TH1050 age, which is done in air, the parts were heated at 1060F for 3 hours--I am not sure if that is in air or atmosphere--thought I had that detail in my notes, but I just posed the question to the plant. I should hear from them by the end of the day.

 

[swall]

The 3 hours at 1060F is also performed in air.

 

[israelkk]

Why the disks are machined in the A condition. From my experience machining in the heat treated condition is easier and give better surface finish. Then chemical cleaning is not needed before passivation.

Generally, belleville spring are made sheet or strip of 1t-7 treated the CH900 condition. Therefore, the TH1050 treatment seems to be too low in strength.

According to NASA specifications only the CH900 (and in your case I assume HT900) condition has a high resistance to stress corrosion. See MSFC-STD-3029 GUIDELINES FOR THE SELECTION OF METALLIC MATERIALS FOR STRESS CORROSION CRACKING RESISTANCE IN SODIUM CHLORIDE ENVIRONMENTS st

http://www.goes-r.gov/procurement/flight_documents/MSFC-STD-3029_RevA.pdf

If the parts are not grinded I will suspect the base material may be bad and may include inclusions and internal defects. Did you checked the bars? Are these a premium quality bars?

What are the calculated tensile stresses in the disks? Usually belleville spring are governed by compressive stresses but in this case it seems the tensile stresses are the problem. Can you supply the full dimensional details of the disks and applied deflections?

 

[swall]

The section size is heavier than what is available in condition C, so CH900 is not an option. The processing information I have from Armco states that normal processing sequence is to obtain condition A material, fabricate, then age to the TH 1050 condition.

 

[mfgenggear]

swall

can you pull parts from stock & do a flo pen?

turning & heat treat should have not caused this problem.
it is possible it was a bad batch of material.

this material is susceptible to IGA & corrosion.

general at this heat treat TH1050 the material is easily machined. however it is easier to machine from full bar stock for this type of part.