Vacuum Chamber – Design and Build Advice

[Mark EasingwoodParticipant @markeasingwood33578]

Not really following this thread, so this may be irrelavant, but, one of of the neighbouring businesses to me sends a 100 lengths of 16mm PVC tube for me to cut into 1300 shorter pieces. The tube they supply me with is made in Italy by Lareter, it is marked as being WRAS approved, and rated at 20 Bar. I think they buy it HERE

It appears to be available in larger sizes at rated at 10Bar. Maybe all PVC pipe is not equal!

The Gas and Water mains seem to be made from HDPE/MDPE, no idea what pressure rating it is, but would that be better than PVC, if you could find a suitable offcut.

Mark.

 

[Michael GilliganParticipant @michaelgilligan61133]

I don’t know how I/We missed this page:

https://en.wikipedia.org/wiki/Materials_for_use_in_vacuum

MichaelG.

[Michael Gilligan]

On Michael Gilligan Said:

I don’t know how I/We missed this page:

https://en.wikipedia.org/wiki/Materials_for_use_in_vacuum

MichaelG.

Another glorious find!  I’m summarising the information in my design overview at the moment.  Quite surprising to find that a few PSI influence materials so much.   I would never have guessed that Zinc sublimates in a deep vacuum to the extent that Brass might dezincify!

I occasionally notice posts go missing from the forum.  When I read Michaels comment last night it wasn’t the last one in this topic.  It is now.   I expect the missing ones will reappear later!   Makes it appear people aren’t reading the thread properly, and are repeating comments made shortly before by someone else.  They’re innocent!

Dave

[SillyOldDufferModerator @sillyoldduffer]

Quick update:

Death sentence passed on PVC tube!  Except I’m sticking with it pro-tem for testing.   Replacing with metal, thanks Noel, means I have to make a new cast-iron base, which is now in the Mark 3 TO DO list.   It’s a major change, and there are a bunch of other problems in hand that can be fixed and tested without a vacuum.

Other Mk3 improvements have resulted from Michael’s find – Wikipedia’s summary of materials in a vacuum. Surprised to find that otherwise trustworthy materials behave badly when pressure drops below 14psi!  Examples:

• Mild-steel. Free-cutting Mild-steel being extra bad.
• Any alloy containing Zinc, Cadmium or Sulphur, so some common Brasses, and Brazing Rods are out
• Most plastics, tapes, glues, paints, rubbers, fibreglass, greases and lubricants.
• Tin, Lead and Antimony, so soft solder is dodgy
• Aluminium is good, but not if it’s anodised.
• Copper, unless it’s de-oxygenated
• Some stainless steels.

Others can be used with care:

• PTFE, ok but not as a barrier
• Nitrile O-Rings.  Viton better.
• Polythene and Nylon stabilise after thorough de-gassing.  i.e leave them in a deep vacuum for a long time before applying them.
• Mild-steel can be tamed by Nickel-plating.

The list of approved materials includes many expensive exotics:

• Beryllium
• Indium
• Gold
• Platinum
• Zirconium
• Tungsten
• Molybdenum
• Tantalum

Fortunately, Aluminium performs well, as do most stainless steels.

I have two advanced textbooks on Engineering Materials.  No mention in either of vacuum issues.

How badly materials misbehave increases as the pressure drops. What’s good enough at 800mb, might become problematic at 500mb.  The real trouble starts below about 100mb.   Zinc sublimates out of Brass and then rains or condenses on something else, causing damage and reducing the vacuum.  My design has a soft-soldered circuit board and Tin grows whiskers that cause short-circuits.

Probably helps that my vacuum pump is unlikely to stress the materials.  It can’t pull a very low vacuum.

Currently working on software deficiencies.  One nasty solved yesterday.  One nearly fixed, needs refinement.  And I’m baffled by the third!

Dave

 

[Grindstone CowboyParticipant @grindstonecowboy]

Ddn’t the Victorians use glass bell-jars? I wonder what sort of vacuum they can go down to? Although I don’t think I’d fancy the effects of an implosion much!

Rob

[blowlampParticipant @blowlamp]

Does it even matter that much if you can monitor and adjust the vacuum easily?

Tyres can often remain at pressure for months, so it might not turn out to be much of a problem for you.

 

Martin.

[Grindstone Cowboy]

On Grindstone Cowboy Said:

Ddn’t the Victorians use glass bell-jars? I wonder what sort of vacuum they can go down to? Although I don’t think I’d fancy the effects of an implosion much!

Rob

An Experiment on a Bird in an Air Pump (Exhibited 1768)

MichaelG.

[SillyOldDuffer]

On SillyOldDuffer Said:

Quick update:

Death sentence passed on PVC tube!  Except I’m sticking with it pro-tem for testing.   Replacing with metal, thanks Noel, means I have to make a new cast-iron base, which is now in the Mark 3 TO DO list.   It’s a major change, and there are a bunch of other problems in hand that can be fixed and tested without a vacuum.

 

–

–

–

 

Dave

 

Dave

What is the benefit of using cast iron for the base? (other than it was something you already had)

Also, on your existing clock you have installed the electronics on the underside of the base so its completely inaccessible, wouldn’t it be best to have the absolute minimum components within the chamber and just bring out the connections to the sensors and electromagnet?

The Vacuum chamber that Robert offered literally looks like it was designed specifically for this clock project, especially as it has a most suitable pass-through and anodising is removable.

I’ve no calculations to offer but from (underwater) experience I would think that for aluminium tube up to say, 6″diameter, a wall thickness of 3mm would be OK. I’ve just had a quick look at a supplier I use which stock ali tube up to 14″ OD and they cut to length (squarely) I recently bought several ‘slices’ of 9″OD 1″ wall tube 20mm thick)

Probably not relevant to the level of vacuum that your existing pump will attain, but one metal supplier told me that for high vacuum purposes his customer would only use ali plate/sheet (not bar) for end plates or large flanges in vacuum chambers. The reason he was given was that ali round bar can potentially contain microscopic longitudinal holes.

Ian P

 

 

 

 

[Ian P]

On Ian P Said:

On SillyOldDuffer Said:

Quick update:

Death sentence passed on PVC tube!  Except I’m sticking with it pro-tem for testing.   Replacing with metal, thanks Noel, means I have to make a new cast-iron base, which is now in the Mark 3 TO DO list.   It’s a major change, and there are a bunch of other problems in hand that can be fixed and tested without a vacuum.

 

–

–

–

 

Dave

 

Dave

Excellent questions!

 

 

What is the benefit of using cast iron for the base? (other than it was something you already had)

It’s heavy and has a dead property that absorbs vibration – one reason machine tools are made of it.  It doesn’t ring like a bell and, properly cast, is low porosity.  I’ve got some big lumps of it in my junkbox, and was lucky that it’s not on the ‘avoid’ list.

Also, on your existing clock you have installed the electronics on the underside of the base so its completely inaccessible, wouldn’t it be best to have the absolute minimum components within the chamber and just bring out the connections to the sensors and electromagnet?

Ah, not completely inaccessible.  A 3D-printed tray is bolted to front panel and is easily removed.  The wires are long enough so the clock still works, and they can be unplugged if need be.

The Vacuum chamber that Robert offered literally looks like it was designed specifically for this clock project, especially as it has a most suitable pass-through and anodising is removable.

Agreed, except a kind friend has found me a 150mm diameter x 10mm thick CO2 cylinder, which I can fit to a bigger cast-iron base.  Robert’s box being Aluminium will have a higher centre of gravity and a smaller base, making it more likely to vibrate.  I’m rather hoping Robert might be inspired to build his own, and that unit would be a good start.

I’ve no calculations to offer but from (underwater) experience I would think that for aluminium tube up to say, 6″diameter, a wall thickness of 3mm would be OK. I’ve just had a quick look at a supplier I use which stock ali tube up to 14″ OD and they cut to length (squarely) I recently bought several ‘slices’ of 9″OD 1″ wall tube 20mm thick)

Agreed.  The sums say 3.7mm PVC (which I have) will take 0mb with a Safety Factor of nearly 2x. Aluminium is about 5x stronger than PVC, so 3mm is very safe.  Four advantages to a repurposed CO2 cylinder though:

• No need to make an end-cap.
• The extra mass compared with a 3mm tube absorbs vibration.
• A thick tube can support the pendulum more rigidly.
• Increasing up to 150mm diameter creates enough space, I think, for a MIL-STD connector

Probably not relevant to the level of vacuum that your existing pump will attain, but one metal supplier told me that for high vacuum purposes his customer would only use ali plate/sheet (not bar) for end plates or large flanges in vacuum chambers. The reason he was given was that ali round bar can potentially contain microscopic longitudinal holes.

Now, thanks to the forum, I know more about it, I’m not surprised.  What happens inside a deep vacuum is something else.  In all my other projects Free-cutting mild-steel, Brass, plastics, and goo can all be taken for granted.  Not on this one!

There’s a Requirement I’m trying not to break:

‘Build requiring no more than a lathe, mill, and basic electronic methods – strip-board and/or wire-wrap construction. Off-the-shelf components, preferably affordable so that other people can build it.”   

I’d be pleased if anyone else attempted similar and hate projects that depend on unobtainium parts and materials.  Repurposing CO2 cylinders and Robert’s box is close to the knuckle.  I’m not a purist, but, so far, everything else in the clock is available.  Steel could be substituted for the cast-iron base, or Aluminium.  Steel would be good for a frame that bolts more rigidly to a solid wall than my free-standing arrangement.

Dave

 

 

[Robert Atkinson 2Participant @robertatkinson2]

I don’t think I’ll be building a pendulum clock Dave. I did briefly think about it but I have too many projects. Even on clocks I’ve got a Rubidium based clock / frequency standard to finish off. Just the last few bits to do. The clock part is old school, MSI logic chips and nice HP dot matrix displays in yellow with built in decoders. Not a microprocessor in sight. The front panel was one of the first jobs I did on the the mill.

[Ian PParticipant @ianp]

Four advantages to a repurposed CO2 cylinder though:

• No need to make an end-cap.
• The extra mass compared with a 3mm tube absorbs vibration.
• A thick tube can support the pendulum more rigidly.
• Increasing up to 150mm diameter creates enough space, I think, for a MIL-STD connector

There’s a Requirement I’m trying not to break:

‘Build requiring no more than a lathe, mill, and basic electronic methods – strip-board and/or wire-wrap construction. Off-the-shelf components, preferably affordable so that other people can build it.”   

I’d be pleased if anyone else attempted similar and hate projects that depend on unobtainium parts and materials.  Repurposing CO2 cylinders and Robert’s box is close to the knuckle.  I’m not a purist, but, so far, everything else in the clock is available.  Steel could be substituted for the cast-iron base, or Aluminium.  Steel would be good for a frame that bolts more rigidly to a solid wall than my free-standing arrangement.

 

Dave

The CO2 cylinder is definitely a good find and will add a lot of mass, yes possibly top heavy on an ali base, but a block of iron could be bolted to it

Not sure how ‘thick tube can support the pendulum more rigidly’ (if that is the CO2 cylinder?) as the pendulum is, or was, hung from a structure inside and floating from the tube.

Are you able to swing the CO2 cylinder in your lathe to create whatever end feature needed to seal it to the base? (just askin in case I might be able to help)

Ian P

 

 

How does the 6″ tube give more space for the mil-std connector, is it ?

[duncan webster 1Participant @duncanwebster1]

If the top closure was bolted on last, and the tube were bolted to the base, you could have jacking screws twixt top plate and tube. Thicker steel tube is obviously more stiff, but 3mm ally would be hugely stiffer than the present 4 columns, suspect diminishing returns.

I reckon SOD is too far down the road to start again. Bit like the Irish man who was asked for directions, I wouldn’t start from here.

[Grindstone Cowboy]

On Grindstone Cowboy Said:

Ddn’t the Victorians use glass bell-jars? I wonder what sort of vacuum they can go down to? Although I don’t think I’d fancy the effects of an implosion much!

Rob

They did use bell-jars, and got down to about 0.0013mb by filling a tube with Mercury and inverting it.  Takes forever to pump out a significant volume of air with a Torricellian vacuum though and these days Mercury is difficult to source.  Expensive too, I’d need at least a kilogram, about £200 wholesale.

Glass meets most vacuum requirements; it resists pressure extremely well, provided the pressure is applied slowly.  Unfortunately, glass is no good when a requirement insists the container shall not shatter!   Nonetheless it is used, for example by Prof ET Hall’s Littlemore Clock.

Professor Hall is interesting – exposing the Piltdown Man fraud, dating the Turin Shroud and other significant achievements.  Wikipedia says “He built the Littlemore Clock in the 1990s, which is the most accurate pendulum clock ever built.”  Not so, he got close, but didn’t quite make it.  Problem may have been location, the clock bouncing on shaky ground near a railway line despite being mounted on a buried 6 ton concrete pillar.

Dave

 

 

 

 

[blowlamp]

On blowlamp Said:

Does it even matter that much if you can monitor and adjust the vacuum easily?

The vacuum is monitored, and the clock already compensates for barometric pressure change, and could be extended to cover a leaking vacuum chamber.  But there’s a limit, and the accuracy of the pendulum will degrade due to increasing turbulence as the vacuum slowly fails.

I can’t adjust the vacuum easily.  Problem is vibration, not the mechanicals.  when measuring time accurately, Pendula are extremely sensitive to vibration.  The Mark 1 version of this clock detected the Moroccan earthquake and lorries driving past the house.   The clock goes to pot whenever the vacuum chamber is pumped, so best avoided.  I’m going to try a micro-vacuum pump idc.  It will be less intrusive than my big Edwards Rotary Pump, but I’m fairly certain it will still upset the timing.

 

Tyres can often remain at pressure for months, so it might not turn out to be much of a problem for you.

Exactty what I thought at first, and not so.  There’s so much air compressed inside a tyre that tiny leaks go unnoticed.  The same tiny leak in a vacuum is an unacceptable  flood.  Same problem with PVC: my soil pipe is good for 10bar and does not leak water, so should be OK right?   Nope, below about 0.8bar, PVC gases and disintegrates.  It’s not impermeable to air either.

I’m learning the hard way that experience of atmospheric and higher pressures is misleading in a vacuum.   I’m still struggling with the news that free-cutting mild-steel is unsuitable for what I’m doing.   Thank goodness Aluminium is OK!

Dave

[Ian P]

On Ian P Said:

Four advantages to a repurposed CO2 cylinder though:

• …

 

Dave

The CO2 cylinder is definitely a good find and will add a lot of mass, yes possibly top heavy on an ali base, but a block of iron could be bolted to it

I could, but there’s no advantage given I happen to have a suitable cast-iron block.  Someone who didn’t have a big lump cast-iron might well go that way.  They both work.

Not sure how ‘thick tube can support the pendulum more rigidly’ (if that is the CO2 cylinder?) as the pendulum is, or was, hung from a structure inside and floating from the tube.

In the existing design the pendulum hangs from a structure because PVC isn’t strong or rigid enough to support it.  A 10mm thick Aluminium tube is, so I can get rid of the internal tower,   Duncan did the sums on the tower and showed it will sway as the pendulum moves, and it’s worth bracing.  Bracing isn’t easy due to lack of space inside and the tower is less stiff than a hefty round pillar.

It’s not a simple upgrade.  Hanging the pendulum from a tower means the pipe is only a cover that can be taken off so the pendulum assembly can be worked on.  And the clock still runs.   Attaching the pendulum from the pipe introduces a shower of new problems, because its delicate suspension is inaccessible – the pipe is in the way!  How best to solve it may be another forum question.

Are you able to swing the CO2 cylinder in your lathe to create whatever end feature needed to seal it to the base? (just askin in case I might be able to help)

I hope my lathe is big enough, but it may not be.  Did 4½” PVC OK.  6″ will be a tight-fit faceplate job, with the tool-post on the end-stops.   Your kind offer might well be accepted!

Ian P

 

 

How does the 6″ tube give more space for the mil-std connector, is it ?

Here’s the floor plan.  PVC interior is the blue circle,  Aluminium in purple. Current tower shown in black:

The tower is a tight fit because a large base is more stable than a narrow.   However, that leaves less space for feedthroughs.  A 6″ tube increases the floor plan, but there’s still not much room for a MIL-STD socket.  Makes it easier though.

Dave

 

[AlanParticipant @alan14594]

Dave said..

It’s not a simple upgrade. Hanging the pendulum from a tower means the pipe is only a cover that can be taken off so the pendulum assembly can be worked on. And the clock still runs. Attaching the pendulum from the pipe introduces a shower of new problems, because its delicate suspension is inaccessible – the pipe is in the way! How best to solve it may be another forum question.

 

If you look at photos of the Shortt pendulum, the actual pendulum is in a tube, with the top part, including the suspension, on a frame mounted on the top ring of the tube.

The suspension is covered by a glass bell jar, so that it can be seen, and the whole can be evacuted…

Alan

[duncan webster 1]

On duncan webster 1 Said:

…

I reckon SOD is too far down the road to start again. Bit like the Irish man who was asked for directions, I wouldn’t start from here.

Very true.  Not made it clear, but this project is in it’s third iteration.

• A prototype explored the art of the possible, and eliminated several dead-ends.  It emphasised the need to maximise rigidity.
• Mk1 was much more solid construction.  Had an eye on vacuum containment, but focussed on the clock’s digital features and how to measure and analyse it’s performance statistically.   Result promising.  Many problems and improvements identified.
• Mk2 builds on Mk1 and addresses most of it’s shortcomings, but not all of it’s design compromises.  Mk2 includes vacuum containment, which is proving more difficult than expected, particularly if I insist on a deep vacuum.  But PVC is acceptable down to 800mb, maybe lower, not a complete waste of time.  It’s entirely acceptable as a draught and light excluder,  both nasty problems.  The design is not wrong, it’s not good enough when pushed to the limit.
• Disaster!  Illness struck 18 months ago after I’d dismantled Mk1 to do a major rebuild.  I put the notes and components in a box, which, when opened last month was empty!   Mk2 became a recovery programme based on faulty memory rather than a well-prepared advance.   I’ve had to type up a Design Overview that’s up to 40 pages and rising.  Desperately needs an edit!
• Mk2, still on the slipway, is proving to be a challenge.  Not unexpectedly so because every improvement to a precision clock gets progressively harder.

Another issue is I’m not building a better clock by copying the best of earlier clocks.  It’s experimental.  For example, it’s accepted that the Gravity escapement is the best possible, therefore that’s what I should use.  Absolutely not.  I’m investigating the possibility that I can do better with a precision timed electromagnet.

Several reasons why an electromagnetic escapement might not pan-out, but I want to prove it by measurement.  And even if the electromagnet only achieves equivalent performance, they are much easier to make!

Forgive me if I ignore advice.  It’s because the obvious answer may not be appropriate because I’m trying something new, or some other constraint applies. Otherwise they’re highly valuable, so keep ’em coming.

Dave

 

 

[SillyOldDuffer]

On SillyOldDuffer Said:

<p class=”bsp-quote-title”>… these days Mercury is difficult to source.  Expensive too, I’d need at least a kilogram, about £200 wholesale.</p>
 

Shame, a few years ago, I could have got you a plastic container full – at those prices, about £400 worth! – found when clearing a friend’s workshop. His relatives had no end of trouble getting rid of it, I successfully avoided getting involved.

Rob

[John HaineParticipant @johnhaine32865]

IIRC the Shortt has a ring that is attached firmly to the wall, the pendulum support being built on the ring.  The tube fits on the ring from below and there’s a glass “bell jar” sitting on the upper surface.

[SillyOldDufferModerator @sillyoldduffer]

Progress report:  I hooked my Edwardes vacuum pump up to a 300mm length of ⌀50mm plastic rainwater pipe, wall 1.9mm thick.  Not ABS as I thought – it’s polypropylene, about half as strong as PVC.  The tube fits into grooved discs: one end has a spike into the heavy rubber air-hose connected to the pump.  The other end is fitted with a bourdon vacuum gauge.

I put Plumbers-Mait into the groove as gasket material.  Far from ideal – soft and oily, so very likely to gas and pull out of the joint.   But I had some!

The pump worked better as a compressor as it warmed up, so I let her run for 10 minutes before testing the vacuum.

Pressures:

• Less than 1 minute to 600mb absolute.
• 11 more minutes to roughly 500mb
• 15 more minutes to 370mb
• letting the pump run for another 20 minutes got down to 350mb but then the needle varied ±10mb around 360mb
• I could hear a hiss at the gauge end – a leak.

The pump appears to be close to specification – I believe they were only good for 300mb when new.

360 millibars absolute is -9.5PSI in old money, so this is a rough partial vacuum.   Good news – it suits my  electronic sensor, and reduces the risk of gassing etc, and surpasses a previous test that only managed 500mb. Bad news, it reduces the pendulum’s time-performance.

Despite being made of thin weedy polypropylene the pipe showed no sign of stress. Sorry to disappoint those expecting an implosion – I’m not posting from A&E or Hades.

Dave

 

 

[SillyOldDuffer]

On SillyOldDuffer Said:

Despite being made of thin weedy polypropylene the pipe showed no sign of stress. Sorry to disappoint those expecting an implosion – I’m not posting from A&E or Hades.

Dave

 

 

Things might be different once you scale up the pipe size. 😉

 

Martin.

[Ian PParticipant @ianp]

Plumbers Mait into grooved discs and it got down to 350mB, that quite impressive, oh and there was a leak as well!

I’ve tried to follow this fascinating thread but, have I missed any numbers on drag reduction at different vacuum levels?

Is Helium, being much less dense than air worth considering?

Ian P

[Ian P]

On Ian P Said:

Plumbers Mait into grooved discs and it got down to 350mB, that quite impressive, oh and there was a leak as well!

I was pleased too!

I’ve tried to follow this fascinating thread but, have I missed any numbers on drag reduction at different vacuum levels?

Mentioned in passing.  For completeness:

• At atmospheric and higher pressures drag is proportional to pressure squared and the volume of the bob.  Bad news!
• In a rough vacuum, say between 10 and 700mb,  drag becomes linearly proportional to pressure and the volume of the bob.  Much better.
• In a deep vacuum, drag becomes tiny.  The main source of friction is the suspension and the metalwork bending.
• Zero drag  In a perfect vacuum..

Is Helium, being much less dense than air worth considering?

Yes, Duncan suggested it a while ago. Well worth trying, and not difficult to do.  It’s in the Potential Improvements Annex.    Compared with air with Helium reduces drag by 85–87%,  Hydrogen even better,  92–94%.

And, I could flush all the air out of the chamber with Helium and get a further improvement by vacuuming that.  Ditto Hydrogen, though I’d need to assess the risk of the pump or a spark igniting the exhaust.   3.4 litres of Hydrogen exploding in my garage would ruin my day!

Ian P

So much to do, so little time!

🙂

Dave

[blowlamp]

On blowlamp Said:

On SillyOldDuffer Said:

Despite being made of thin weedy polypropylene the pipe showed no sign of stress. Sorry to disappoint those expecting an implosion – I’m not posting from A&E or Hades.

Dave

 

 

Things might be different once you scale up the pipe size. 😉

 

Martin.

They are.  No need to guess, do the sums.

The maths predicted my 300x50x1.7mm test pipe would collapse at about 200mb, which I was confident the pump could not do.  Kept a close eye on the gauge though!

A 300x50x1.7mm drainpipe is predicted to collapse at about 820mb.  No good at all!

Many ifs and buts – in particular sitting the tube in a grooved disc is weaker than spherical ends, and a tube that’s not round or is damaged in any way is much weaker than a perfect one.

What I did was low risk in a short test, but the safety factor is too low for permanent use.  So beware chaps on the internet who imply “It’s safe because I did it a few times in my shed and nothing bad happened.”  All too often that means “Guessed, didn’t test it thoroughly, and got lucky.”  Ignorance is bliss, until it goes wrong!

Dave

 

[cedric 1Participant @cedric]

With atmospheric pressure of circa 14.7psi as the driving force you are unlikely to blow the roof off your workshop with such a small installation.  But I’d definitely be wearing safety glasses!

[Author]

Posts