Vacuum Chamber – Design and Build Advice

[Michael Gilligan]

On Michael Gilligan Said:

I have evidently lost the plot again, Dave

… I thought you were seeking alternatives to your PVC pipe.

MichaelG.

No, not yet!  If PVC works and is safe I’ll stick with it.

I’ve done some sums and a risk assessment suggesting my 3.7mm thick PVC pipe is safe, but it’s not confirmed:

• are my sums correct?  Will it blow up, and if it does, how much will it hurt?
• do others know more about using a PVC soil pipe to hold a vacuum?

Naturally enough, the discussion has wandered into alternatives, which are also valuable.

So far, confirmed there is cause for concern, but no-one has confirmed PVC pipe is a non-starter.  If they do, then I now have a list of alternatives, with different pros and cons.  Taking ABS as an example, it’s a lot more expensive than PVC and unlike PVC I haven’t found an easy way of sealing the open end.  It’s about 5× stronger than PVC, hurrah, but much more trouble.  So I’m in no rush to switch to ABS just because it’s stronger.   As PVC ticks a lot of other boxes, abandoning is only justified if PVC is proven unsuitable.   And it might be.  We’re working the problem, not jumping to conclusions.

Have to admit this project is complicated.  I’m also having trouble following all the moving parts. The design overview is 36 pages long, and more is on the way.

Dave

 

[duncan webster 1Participant @duncanwebster1]

It’s not strength that matters in avoiding implosion, it’s stiffness and thickness. I don’t know how ABS compares on stiffness, all plastics seem to be a lot lower than metals. However, sums suggest implosion isn’t going to be a problem, mainly because the tube is quite short, but outgassing could be. The large vessel I designed was aluminium. Feeling guilty now as I think it was me suggested PVC in the first place

[cedric 1Participant @cedric]

Would a piece of common a garden 4″ galvanised water pipe do the job? Threaded both ends you could attach a standard blanking cap to the top and a standard threaded flange to the bottom, which could then be bolted to a thick steel base plate.

Substitute a threaded reducer for the cap on the top and you could run a half inch pipe to your vacuum pump.

I wonder how a workshop air compressor would go as a vacuum pump if you hooked that pipe up to its intake? Might pull a bit of oil up past the piston rings but it could still pull a good partial vacuum?

[duncan webster 1Participant @duncanwebster1]

Zinc gasses off according to the font of all knowledge (wikepedia)

[cedric 1Participant @cedric]

LOL. Yes fount of all wisdom these days. Maybe just plain black steel pipe then? Or stainless? (Could be getting outside of hobby level prices though!)

[Michael GilliganParticipant @michaelgilligan61133]

https://highvacdepot.com/2025/08/08/how-outgassing-of-polymers-affects-vacuum-systems/

MichaelG.

[cedric 1]

On cedric 1 Said:

Would a piece of common a garden 4″ galvanised water pipe do the job? Threaded both ends you could attach a standard blanking cap to the top and a standard threaded flange to the bottom, which could then be bolted to a thick steel base plate.

Substitute a threaded reducer for the cap on the top and you could run a half inch pipe to your vacuum pump.

It would, other than the diameter being a bit too small.   The pendulum swings on a support tower inside the pipe, and, because the tower needs to be rigid, the pillars are spaced to use all the available space:

Apart from being too small 4″ pipe is a good answer.  If only it were 5″!

Pendulum clocks are usually made with 1 second pendula, needing a 1 metre rod.  At least 150mm (6″) is needed inside the tube at the bottom because the bob swings over a 5° arc inside the tube at the bottom.  The tower inside the pipe takes up more space too, and insufficient space results in a tall spindly erection that wobbles.   I’ve compromised with a shorter pendulum on a relatively squat tower – the best I can do with a cheap available 4½” pipe, which happens to fit on the cast-iron block in my scrap box.

I wonder how a workshop air compressor would go as a vacuum pump if you hooked that pipe up to its intake? Might pull a bit of oil up past the piston rings but it could still pull a good partial vacuum?

Reasonable unless a deep vacuum is needed.  Oil isn’t a problem because the vacuum is pulled on the intake before the air gets to the pistons.  Unfortunately my workshop compressor has no easy way of attaching a pipe to the intake.  And alhough they’ll pull a partial vacuum, they’re optimised to compress air, and the vacuum may be weak.  In theory my vacuum pump should do better, except it may not be in good condition.

Dave

[Michael Gilligan]

On Michael Gilligan Said:

https://highvacdepot.com/2025/08/08/how-outgassing-of-polymers-affects-vacuum-systems/

MichaelG.

Michael’s google-fu is strong!  Best overview I’ve seen. thanks.  I notice no mention of PVC, which might be because it’s beneath contempt!

Didn’t know until this morning that Zinc gasses!

Dave

[cedric 1Participant @cedric]

You can get 5″ and even 6″ threaded pipe from industrial suppliers. Not as common as 4″ but it’s out there.

[noel shelleyParticipant @noelshelley55608]

If you are going down the route of metal tube I have both stainless and ali About 6″ dia. The ali is about 10mm wall, the stainless 3 or 4mm. and both about 15″ long. Noel.

[parovozParticipant @parovoz]

Being experienced in vacuum testing, one thing to verify is that all your connectors and feed throughs are vacuum rated, many are not, an a pin head leak in such a small chamber will kill the vacuum. ALSO, make sure you have no wires passing through to the chamber. Even if they are fully sealed ‘potted up’ around the cable, air can leak down between the conductors. If we get a customer with lots of cables going into a chamber, we have to dissect them to get the likes of ‘silca set’ into them to prevent any leakage.

All the best.

 

[John HaineParticipant @johnhaine32865]

If you used metal pipe then would it not be best to also use it as the support as it must be more rigid than your “tower”?  Obviously gets complicated assembling the beast but technically better.

[MacolmParticipant @macolm]

Just find a large enough Thermos/Dewar vessel. It would clearly be strong enough to hold the vacuum.

[Wade BeattyParticipant @wadebeatty78296]

You might keep your eye out for used refrigeration vacuum service pumps, the are common and can evacuate to quite low pressures. I still think that an alloy tube sealed with a cap with o-rings is a good alternative to all the above issues with various plastics. A steel pipe with an o-ring cap is also the same. A screwed on cap with pipe threads?…I can see the struggle with your sensitive mechanism and 2 pipe wrenches, please send a video, with audio!

[cedric 1]

On cedric 1 Said:

You can get 5″ and even 6″ threaded pipe from industrial suppliers. Not as common as 4″ but it’s out there.

Thanks for the hint – I haven’t thought of that.  My local metal emporia does RSJs and pipes, none of it on show, you have to ask, and no running away when they present the bill.

On noel shelley Said:

If you are going down the route of metal tube I have both stainless and ali About 6″ dia. The ali is about 10mm wall, the stainless 3 or 4mm. and both about 15″ long. Noel.

I’ll stick with PVC for the moment, partly because sealing the top is easy.  If it doesn’t work I’ll be in touch,  I think my lathe can just about do 6″ diameter.  Then all I have to do is put a cap on it!

Many thanks both.

Dave

[old martParticipant @oldmart]

Having worked on aircraft instruments I used vacuum chambers often and would feel safer with metal tubing for the clock. We used to go down to 5millibars (10mb is about 105000feet altitude, and 115.9mb is 50000 feet). The common test chambers we used were about 1 cubic foot with a tempered glass door about 12 x 10 inches and 5/8″ thick, sitting on a rubber ring which sat in a groove milled to the chamber opening, hinged downwards. The Scott (USA) chambers were also about 12″ square and used an acrylic lid which simply rested on the top of the chambersealed by a similar rubber ring. The acrylic see through top was about 1″ thick and at about 40000 feet equivalent height had a bow of about 1/8″ when a steel rule was placed diagonally across the top. The British made chambers were cast aluminium about 1/2″ thick, thicher where the seal sat and the Scott chambers were weldments of a similar thickness.

The aluminium with the 10mm wall that Noel has would be my first choice, and rubber vacuum pipe would be better than any plastic.

[Wade BeattyParticipant @wadebeatty78296]

Vacuum fittings, valves and hoses can be sourced from refrigeration supply houses.

 

[John Haine]

On John Haine Said:

If you used metal pipe then would it not be best to also use it as the support as it must be more rigid than your “tower”?  Obviously gets complicated assembling the beast but technically better.

It would, and it’s in the “Potential Improvements” Annex.  Latest version of the Design Overview here on Dropbox if anyone has insomnia!

I rejected PVC as a support because it’s too bendy.  A length of Noel’s 6″ diameter, 10mm thick Aluminium pipe would do a super job.

Dave

 

[parovoz]

On parovoz Said:

Being experienced in vacuum testing, one thing to verify is that all your connectors and feed throughs are vacuum rated, many are not, an a pin head leak in such a small chamber will kill the vacuum. ALSO, make sure you have no wires passing through to the chamber. Even if they are fully sealed ‘potted up’ around the cable, air can leak down between the conductors. If we get a customer with lots of cables going into a chamber, we have to dissect them to get the likes of ‘silca set’ into them to prevent any leakage.

All the best.

 

Your input will be most welcome!

There’s another question coming about what I’m calling the Bulkhead Penetrator (in Naval parlance a device for passing cables safely between compartments when one of them is accidentally filled with water.)

It will carry 6 wires and the exhaust pipe.  In the companion “Clock Measuring” thread, Robert explained that plastic cable insulation leaks, so initial thoughts are a brass screw-fitting, with o-ring on the outside, pipe central surrounded by six metal spikes, the whole lot potted in epoxy.   CAD model coming later.  Very grateful for a professional critique when the time comes, because I’m an amateur taking a walk on the wild side.

Might be 7 wires because John Haine suggested a d**n good reason for adding another one this morning!

As a taster of innocence abroad here’s the valve I made today.

It’s modified Presta bicycle valve, inverted, with an adaptor to fit a 4mm diameter PTFE pipe.   If it doesn’t hold the vacuum, Michael has identified a commercial source of vacuum valves, so they’re Plan B.  Unless someone has a better suggestion! (I’ve noted Wade’s advice, ta.)

Fans of unusual threads will notice  the valve. It’s a DIN7756 (Metric 6.0mm 0.8 pitch)  I’m confused: some sources say it’s compatible with M6 × 1.0 and has the usual 60° V form. I had to make an adaptor and thread it with a special tap…

Dave

 

[Wade BeattyParticipant @wadebeatty78296]

For through bulkhead I would recomend these ROV bulkhead connectors. These are rated to 1000M for smaller subs. Our bigger higher HP ROVs are rated to 3000M and stay submerged for days, sometimes weeks at a time.

Female Bulkhead Connector + Male Inline Cable Connector (Micro Circular series)

You are operating at 15 psi pressure differential

[MacolmParticipant @macolm]

Incidentally, I expect you are all well aware that silicone rubbers generally make excellent molecular sieves.

[cedric 1Participant @cedric]

What about the rubber inside the bicycle valve? Might it be better to make a metal to metal valve, like a model boiler safety valve but with a lighter spring?

But will the low pressure in the chamber be enough to push the valve open against even a light spring? Perhaps a springless check valve would work better once the vacuum is getting down a bit?

Or a very light reed valve from a workshop air compressor etc. You could then have a larger area for the very low pressure to push on and open the valve?

[SillyOldDufferModerator @sillyoldduffer]

To reduce cross-talk I’ve started a new topic “Vacuum Bulkhead Penetrator Design – Advice Please”.  Turns out I don’t understand O-Rings either.

Dave

[Macolm]

On Macolm Said:

Incidentally, I expect you are all well aware that silicone rubbers generally make excellent molecular sieves.

I do now, ta!  Silicone Rubber is now banned.

On cedric 1 Said:

What about the rubber inside the bicycle valve? Might it be better to make a metal to metal valve, like a model boiler safety valve but with a lighter spring?

I don’t know.  Rubber, being squishy should seal better.

But will the low pressure in the chamber be enough to push the valve open against even a light spring?

There is no spring in a Presta.  Fitted to a tyre, its held closed by pressure in the tyre.  I’ve reversed it so it’s held closed by atmospheric pressure.  I hope!  Untested.

 

Perhaps a springless check valve would work better once the vacuum is getting down a bit?

Or a very light reed valve from a workshop air compressor etc. You could then have a larger area for the very low pressure to push on and open the valve?

Wrong way round, but I know what you mean.  A larger valve has a bigger leak area which is counterbalanced by more force on the seal.  More sums!  Gut feel is the force on the seat area is much the same.  Have to draw it to be sure.  A problem with a small valve and exhaust pipe is it will take a long time to remove all the air through them.  The tiny orifice may not matter provided the pump gets there eventually and the vacuum holds.

Worth trying even though it might not work.  Not expensive in time or money.

Making a good vacuum valve is quite difficult I think, as is making a good vacuum pump.  The slightest leak reduces performance dramatically.

Dave

[noel shelleyParticipant @noelshelley55608]

You want the ali ? I will measure it in the morning. Send me a PM I will tell you more. Noel.

[Author]

Posts