Help with Stainless Steel on how to keep it FLAT 1

[bsdme]

Hi Folks,

I need HELP.
We use flexure that is round, 17-7 PH A Stainless Steel and 3 mil thick, 3/8 in dia. weld onto another flat surface
that is 17-4 PH Stainless Steel. The problem here is how can we keep the flexure say 100% flat or close
to 100%. Another information is that we sandblast (micro) onto one side only, after sandblast, I observe
the flexure under microscope and there curve or dome on the opposite that is not sandblast. Our vendor receive
the material from it vendor as Annealed but the material later rolled down to 3 mil from unknown original thickness.
Our vendor then rolled out the material to get it to flat which I did check that after we receive from our vendor
and that they do look flat observing under our interferometer instrument.
Question: How can we keep the flexure flat after sandblast on one side?

Thanks Folks,

BSDME

 

[EdStainless]

This has been age hardened right?
If not then you don't stand any chance.
It needs to be aged, maybe two or three aging cycles to minimize residual stresses.
If you aren't worried about toughness I would go for maximum strength.
Though with the mixture of alloys it will be tricky.
You will need to re-anneal the fixture after welding,then austenite condition at 1400F for 1 hr, cool to room temp and hold 1 hr, then age 1050F for 90 min.
The 17-4 will end up in an overaged (softer) condition, but the 17-7 will be hard.

But if you are actually doing any work on it then it will distort.
Either blast the other side to correct the shape, or make the fixture stiffer.

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

 

[bsdme]

Hi Ed,

Thanks for you reply.

The flexure material itself did have anneal process as part of the requirement from
us to our vendor. Also, we did send out the flexure parts for heat treat (Harden)
when we receive the part from our vendor. After we weld the two part, we also
stress relief at 345C (653F) for 5 hours and we then sandblast the flexure on one
side after some process to the part. Most have good result but some flexure still
adjust to the stress in them. So now we have focus on how we can flatten it before
welding them.
The thing is that we can sandblast the other due to the space limited in the part
after welding and the design of the flexure is very space limit to access to sandblast
the other side.
Ed, if the material is harden, it should still be flat even after sandblast apply to
one side?
I'll take your suggestion into account while testing/researching on this project.

Thanks,

BSDME

 

[EdStainless]

It may be distorting due to residual stress, or maybe because of the stress that you are creating with the blasting.
If this has been hardened correctly it should be RC 40.

The anneal and age should have been done after the welding.
If you can't then you need to use a stress relief temp that is 50F below the aging temperature.
Uniform heating and cooling is critical, and multiple cycles may help a lot.

But since you welded one end of a tube you have created serious non-uniform stresses.
So was the part precision ground after it was hardened?
Can you use a less aggressive blasting?

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

 

[tbuelna]

I would agree with EdStainless. The sandblasting is likely causing the distortion. The sandblasting will result in a residual compressive surface strain in the flexure. And since it is thin it will tend to curl in the direction of the surface that was blasted. It's an imbalance between the tension/compression strains in the opposing surfaces.

 

[bsdme]

Hi Ed Tbuelna,

Thanks for your reply.

Before sandblast I did check the thin flexure and it is
pretty flat, after sandblast, I did there is the dome curve.
I know that the stress is created by the sandblast on one side
of the flexure, I just need to find ways to hopefully flatten
it out after this sandblast process.
We actually did 3 cycle of heat 150c then -80C plus another 48 hours
of burn-in at 250F, it help most unit but still some stress is in there.
Maybe more cycle perhap? Well, I'm still testing this idea.
Ed, could you please explain this " So was the part precision ground after it was hardened?"
What do you mean ground? Sorry since I am still learning.
Thanks for the help guys.

BSDME

 

[EdStainless]

How did you get it flat in the first place? Are you sure that that process didn't introduce stresses?

150C or 250F is not a stress relief temp for this alloy. You need to use 450C.
Often with multiple stress relief cycles people will use a lower temp each time by about 50C or so.
You must slow cool. Any non-uniformity in cooling will create significant residual stress

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

 

[bsdme]

Hi Ed,

We do burn-in at 250F after we put some other part to it, so we can
not go to 450C because other part will not be able to handle at that temperature.

The flexure that we receive from our vendor and I check it under our interfermeter
and it does look flat, but the question that we have was how much stress did the
manufacture introduce into the metal when they roll the metal sheet to 3 mil (.003 in.)
and did they do any anneal after the roll process. Our vendor did indicate that
the metal sheet was anneal when they received that product.
The reason we sandblast one side of the flexure is because we put another very small
part on it and the surface need to be clean before we glue it to the surface of the
flexure, without sandblast the epoxy does not glue very well.

Thanks folks,

BSDME

 

[EdStainless]

You should consider doing a higher temp stress relief before you assemble any of it. Then lower temp after you mount other parts.
Annealed, or annealed and aged material is not going to be free from residual stress.

And you should either try to blast less, or look at acid etching the location where you are bonding.


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

 

[tbuelna]

BSDME,

The depth of the surface layer having residual compressive strain will depend upon the properties of the blast media (size, shape, mass, hardness, etc), the blast intensity level, and the blast coverage. This compressive surface strain can be fully relieved using a conventional thermal process, or it can also be relieved by removing the affected surface layer. However, it is unlikely either approach will result in a consistently flat surface over a large number of parts.

Since the part in question is a "flexure" that will likely be subject to cyclic loading, if continued flatness of the flexure in service is important then it is critical to fully relieve the part. If there is any residual strains in the flexure when it is put into service cyclic loading will gradually relieve them, and this can affect the performance of the flexure.

In your post you also mentioned that the flexure is grit blasted to produce a rough surface texture intended to improve adhesive bonding. Firstly, I would mention that grit blasting is not a preferred method for preparing a metal substrate for adhesive bonding, since the grit blasting usually drives contaminants deeper into the surface rather than removing them completely. Chemical etching processes work far better. Secondly, if what you really want is a rougher surface texture for improved adhesive bonding, you might consider sanding the flexure surface prior to welding. Sanding will be far less likely to distort the flexure.

Hope that helps.
Terry

 

[bsdme]

Hi Ed n TBuelna,

Thanks you for sharing your share of knowledge there, they're great! and good tip.

Let me introduce a bit of our process
1) we receive the flexure from our vendor,
2) we send out the part for heat treat at ~1066C ( I think it for 1 hour heat treat)
3) we weld the flexure (Stainless Steel 17-7) onto another surface (stainless steel 17-4)
4) we stress relief the assemble parts at 345C for 5 hours
5) we done some process to the parts but does not distort the flexure, it is fully protected in housing.
6) we sandblast one side of the flexure to remove the oxide surface and reveal rough surface(can not see with naked eyes)
to allow epoxy to have better grip of the surface and for bonding our strain gage. (surface clean with solvent b4 use)
7) we then do some low temperature testing for stability for overall of the whole assembly.

Ed: anneal the flexure metal sheet would not remove all the internal stress that was created from rolling the metal sheet
from original thickness down to 3 mil?
Assume that the anneal process remove all the internal stress, I then sandblast one side of the flexure, the
stress will be created due to uneven sandblast?
I will definitely look into the sandblast process, maybe that is the key (I hope).
TBuelna, I will definitely look into your idea and with the grit since I don't know exactly the type of powder
is being use.

Thanks you for everything guys,

BSDME