Tightest realistic tolerance of a 200mm ID cylinder 1

[Poorpaulus]

Hi everyone,
I am working on a project that requires a stainless cylinder with an ID of 200mm to have a redicoulsly tight tolerance of +/-0.001mm, over an axial distance of 120mm (10mm wall thickness). That's not a typo on the decimal places - needless to say, this is challenging.

Does anyone, with appropriate experience, have any idea of the feasibility of generating a part with a tolerance this tight?

P.

 

[3DDave]

That's 5. ppm tolerance; steel has a coefficient of thermal expansion of 11-13 ppm/ºC, so you have only about 0.5ºC temp change to go from nominal to being out of tolerance.

Even verifying this is going to be a problem.

 

[Poorpaulus]

Agreed. It'll be used in a temperature controlled environment. Someone with a massive brain has done some very long calculations, which stated that it requires the tolerance for the concept to work.

 

[3DDave]

It will take days to stabilize to the hundredths of a degree required. This is the sort of temp control NIST uses for making it's reference items; like turning out all the lights as the heat radiated from the lights causes the parts to distort.

 

[EdStainless]

Is it steel?
Has it been heat treated?
Given multiple temper/stress relief cycles?
Was the HT done a few weeks ago, at least?
I hope that your grinder in a temperature controlled space.
Fixture the part, start running the grinder (and circulating coolant).
Make sure that all of your measuring devices are in the same space.
Let everything run for a few days.
Then you can start the finish grind.
With that tol and such a thin wall that will take a while.
Watch the temp on the coolant, a change of 1F would be enough to require you to stop and let things re-stabilize.

 

[dgeesaman]

Agreed. It'll be used in a temperature controlled environment. Someone with a massive brain has done some very long calculations, which stated that it requires the tolerance for the concept to work.

Not every massive brain idea that can be supported by calculations is a useful idea. The true genius is making math and manufacturability work together. The fact that the designer hasn't considered if this is manufacturable yet suggests you don't have that luxury.

But to EdStainless' point, you will need to throw every imaginable variable in your favor and possibly still include an assumed scrap rate.

I would also work very hard on locking down the acceptance criteria. Just physically moving the part seems like a very high risk for egging the piece so I would not be interested in having final acceptance occur off-site. Way too many variables added.

Might be worth a call now to a high end ring gauge manufacturer to see if they will produce a ring gauge to that accuracy. If they won't do it, can anyone?

 

[Poorpaulus]

Thanks everyone.

You're on the same mindset as me. I will go back to the "Big Brain" and tell him that he needs to take another look at his calcs.

 

[btrueblood]

Not sure why the cylinder needs to be stainless steel, but if it's for corrosion resistance more than strength, there are alternatives. An alternative corrosion resistant material would be glass tubing. It's formed to very tight (.001 mm) axial i.d. tolerances (shrink fit from the melt over centerless-ground graphite mandrels).

https://aceglass.com/html/3dissues/GB1002%20Trubore%20Precision%20Glass%20Tubing/GB1002%20Trubore%20Precision%20Glass%20Tubing.pdf

Precision Glass Tube Used with Pistons and Precision Instruments

Precision glass tubes with accurate dimension and 0.001 mm tolerance are made of soda-lime, borosilicate or quartz, used in laboratory, hospital, factory.

www.glass-product.com

You can probably find others.

 

[Poorpaulus]

Thanks!
I'll look in to that. It was assumed that it needs to be a non-corrosive metal.
I can't give exact details, but it needs to be 180mm diameter, corrosion isn't really an issue, but it will be in contact with atmosphere and it will be in contact in a non-corrosive liquid (not water).
I appreciate that there's a whole world of materials out there (and I'm no expert), so I feel that this journey could be very long, or very short.

 

[IRstuff]

For thermal stability, particularly thermal expansion, one might consider quartz, which has a small fraction of the thermal expansion of glass, albeit, it will be more expensive. Integrated circuit photolithographic masters tend to be quartz based to minimize thermal changes to the critical dimensions of the masks printed on the masters.

 

[3DDave]

Quartz is 0.55 ppm/ºC so that does improve it to an almost +/-10 ºC swing if made to the nominal size.

Per https://www.heliosquartz.com/prodotti/proprieta-del-quarzo/?lang=en#

 

[GregLocock]

If you put a kevlar wrap over it you could get zero expansion with temperature since kevlar has a negative COTE