Machining and welding custom high vacuum chamber

I have a need for a vacuum chamber that is approximately 8" tall and 24" in diameter. Operating temperature is room temp.
Vacuum at the range of x10^-6 torr, so just getting in to molecular flow, but nothing nasty that takes days to pump down. The vacuum is really just to rid the environment of oxygen.

I've worked with high pressure a reasonable amount, but never at the other end of the spectrum. High pressure, you just tighten it a bit more. Vacuum...well, not quite the same.

My biggest design motivator right now is honestly cost, I need to do it on the cheap. That doesn't mean free, just cheap. I am using off the shelf surplus parts for a majority of it, but the chamber is just not available. I have a good roughing pump and a diffusion pump, all mounted in a nice rack with ion gauge.

This won't be a dive in and hope you can swim project...but it sort of is at the same time. I don't have the budget to have it custom made by a shop that specializes in vacuum.

BUT...I know there are some sub-atmospheric pressure guys here.

Right now, my thoughts are simply a sandwich of ~1" thick 6061 aluminum plates that have an o ring groove milled (high speed, carbide, air jet coolant, ramp down not dive, polish any lines that aren't concentric)....but what to use for that spacer to get the depth?

I can have a piece of stainless rolled, then get a couple of flanges made up, then weld it on (stitch welded on outside for strength, fusion weld on inside for seal). But, instead of flanges, would it be foolish to try to roll stainless angle, weld it up, then machine it flat and polish it?

I have also considered just getting some "thick" rings of aluminum water jetted and stacking them, of course with appropriate seals and such.

My understanding is that at x10^-6, I'm not yet in to the range where I'll see wicked virtual leaks from off gassing of non-electropolished material...but I'm also inexperienced.

Any ideas?...

Its way more involved than it seems... check this out for some inspiration. If you spend some time there is lots of little snippets of info you can pick up from this guy.

Ryan Hussey (@bagofwelds) • Instagram photos and videos

I would acquire a decommissioned centrifuge and use the chamber. High speed centrifuges have a 2" thick safety barrier so they are quite heavy. However they already have an "O" ring seal top and bottom and are made for high vacuum usage. You could probably pick up an older high speed for scrap price and just use the barrier and toss the rest. Of course you could use the vacuum system that is already there. They usually have an Edwards or Sargent-Welsh roughing pump and a silicon oil diffusion pump.

kustomizingkid said:

Its way more involved than it seems... check this out for some inspiration. If you spend some time there is lots of little snippets of info you can pick up from this guy.

Ryan Hussey (@bagofwelds) • Instagram photos and videos

Click to expand...

I understand that it's an art form...but x10^-6 is just entering in to high vacuum territory, most of what I see him posting is ultra high vacuum. I'm not yet to the point where I'll be chasing a finger print for days.

I've done a decent amount of reading on it, so I get that it's pretty involved...but realistically, I could do with just a rotary vane. I just need it to seal up so I can pump it down fast, back purge with argon...then do that two or three times and I should be good for my purposes.

I'm not in the realm of knife edge...still solidly in static o ring territory.

crossthread said:

I would acquire a decommissioned centrifuge and use the chamber. High speed centrifuges have a 2" thick safety barrier so they are quite heavy. However they already have an "O" ring seal top and bottom and are made for high vacuum usage. You could probably pick up an older high speed for scrap price and just use the barrier and toss the rest. Of course you could use the vacuum system that is already there. They usually have an Edwards or Sargent-Welsh roughing pump and a silicon oil diffusion pump.

Click to expand...

What are these centrifuges used for? As in, where do I look to find one?

They are research instruments used to separate biological (usually) components that are very light such as dna. Beckman-Coulter comes to mind for one major manufacturer. Sorval is another. The barrier is used to contain the titanium or sometimes AL rotor when it explodes as I have seen many times. These are considered "high speed" analytical centrifuges as opposed to the "low speed" units that are typically used to separate blood components which don't require the same forces. High speeds run in the 80,000 to over 100,000 rpm range. The chamber is approximately two feet in diameter by 1 foot deep. I'm sure there are a couple on ebay as just about every lab in the country has a least one centrifuge. I bought one from UVA at scrap price. I have taken many to the dump as trade ins. Look up Beckman L-60, L-70 or L80. These are older machines that are scrapped quite a bit. The chamber comes right out with four bolts.

Ok, wasn't aware that they were ran under vacuum for biological purposes. I'll check it out.

I should have mentioned that what you want is an "ultra-centrifuge" not a "high speed". Other even older models that would work are L-2, L5-65 etc. Any ultra has the same barrier/chamber system.

Rotor explosion

kustomizingkid said:

Its way more involved than it seems... check this out for some inspiration. If you spend some time there is lots of little snippets of info you can pick up from this guy.

Ryan Hussey (@bagofwelds) • Instagram photos and videos

Click to expand...

Guy does some really nice work, but I wonder if his shielding gasses are setup correctly - seeing a little more oxide buildup than I think is ideal.

Tangentially related to vacuum, make sure you have adequate stiffness (ideally convex rounding) on your 24" endplates. Atmospheric pressure on flat endplates is going to make a tough job of sealing tougher, due to deflection.

10E-6 is a good place to start, not too easy, not too difficult. You'll find that aluminum plate can be porous and leak forever. There are processes to seal it, but IMO way too much trouble. You have to vacuum impregnate it with a suitable sealant. Large surface area makes things more difficult than you might think. +1 on the idea of salvaging a chamber from something.

Conrad Hoffman said:

10E-6 is a good place to start, not too easy, not too difficult. You'll find that aluminum plate can be porous and leak forever. There are processes to seal it, but IMO way too much trouble. You have to vacuum impregnate it with a suitable sealant. Large surface area makes things more difficult than you might think. +1 on the idea of salvaging a chamber from something.

Click to expand...

Even when using 6061? I know from another post that cast tooling plate will be pretty porous, but I thought 6061 was supposed to be better.

I can probably find some stainless plate if that would be better. But it's going to have standard mill scale.

I'll go check out an ultracentrifuge this week, but I don't think the diameter is large enough.

Made one from a beer keg. Micro polished SS to be food grade. Tig welded con flats to it. Held 10^-7 no problem. Had to touch up the welds with torr seal. Be sure to have argon in it when welding.

10^-6 is going to be a struggle with primitive equipment. Clean Dry and Empty, it will still take HOURS to pump out with a dif pump. But the mean free path at 10^-6 is quite long.
Mean free path Perhaps you don't need to go so far. Viscous and molecular flow is not usually a meaningful metric for many vacuum applications. "Getting the oxygen out" is not difficult, unless there are leaks ;-)

I worked for years producing PVD optical films at 10^-5.

Try here for a cast off chamber / bell jar
Used Semiconductor Equipment | Used Semiconductor Manufacturing Equipment

work on the price if they have something.

A 1 inch base plate is "self supporting" out to say 30 inch diameter with an o-ring or boot seal.

Aluminum is NOT a good choice, you will constantly be fighting scratches, porosity etc.

I have a very nice 18 inch box coater (small) in the downstairs lab. Turbo pumped, and fully automated. I would sell it, but it's not cheap ;-)

ETA

IT's always the WATER that takes it's time moving out of any vacuum system. Hard stuff to get moving and keep moving. It wants to cling to everything it contacts.

Milland said:

Guy does some really nice work, but I wonder if his shielding gasses are setup correctly - seeing a little more oxide buildup than I think is ideal.

Click to expand...

I think he does it for show since EVERYTHING is electropolished after welding.

CalG said:

Perhaps you don't need to go so far. Viscous and molecular flow is not usually a meaningful metric for many vacuum applications. "Getting the oxygen out" is not difficult, unless there are leaks ;-)

Click to expand...

I'm honestly not too concerned about it actually leaking, much more concerned about virtual leaks.

And realistically, for most things I won't need to go that far down. This is for melting platinum group metals which just love to adsorb anything they can get their greedy little molecules on.

kustomizingkid said:

I think he does it for show since EVERYTHING is electropolished after welding.

Click to expand...

True enough - it is prettier than plain SS.

snowman said:

I'm honestly not too concerned about it actually leaking, much more concerned about virtual leaks.

And realistically, for most things I won't need to go that far down. This is for melting platinum group metals which just love to adsorb anything they can get their greedy little molecules on.

Click to expand...

24 inch diameter base and top plate seals are going to permeate enough to make virtual leaks trivial.

Again, it's the WATER (H2O) that takes the time, and those Os add up fast. By the time the water has released it's self from the interior walls and fixtures, the virtual leaks will have been pumped. Unless you are completely disregarding good venting practice. Heat helps release the H2O.

One could consider a Getter with an alu evap. pre-process if you are in a hurry.

10-6 is not going to be quickly achieved just pumping. The Argon flush is good. Evolution from the walls will persist, and then it all repeats when the chamber is opened.

But hey, Try stuff, see how it goes, we learn by doing.

ETA For an 8 inch height, a feed through collar would be an ideal "chamber wall".

There are lots of them sitting around unused... somewhere......;-)

Your best bet here is probably purchasing a used bell-jar system. My suggestion is to avoid diffusion
pumps in favor of a small turbopump, backed by an oil vane pump. You will need a simple gage to
determine what the vacuum really is, this can be a small ion gage or a small discharge gage.

With a turbopump you do not need the valve components that a diffusion pump requires, nor do you
need any kind of liquid nitrogen trap to keep the diff pump oil out of the system.

To put this in perspective, a bell jar system with a diameter of about 12 inches and about 18 inches tall
has a rubber gasket on the bottom, which seals to a flat plate. The turbopump need not be more than
three inches in diameter and the backing pump can be pretty modest. With a system like this
you can easily achive low 6s or even mid 7s after a while if you are careful about what goes into it.

Again, no valving needed. You rough right through the turbo, when the pressure drops below a torr or
so the turbo spools up and pumps.

Look at surplus items and you might be suprised.

CalG said:

But hey, Try stuff, see how it goes, we learn by doing.

ETA For an 8 inch height, a feed through collar would be an ideal "chamber wall".

There are lots of them sitting around unused... somewhere......;-)

Click to expand...

Lol...I like learning from other people's mistakes if I can.

I've been looking...just haven't found em. ISO-K 250 and below are easy.

Probably a lot easier to just open the check book.