Leak Troubles

Every fusor and fusion system seems to need a vacuum. This area is for detailed discussion of vacuum systems, materials, gauging, etc. related to fusor or fusion research.
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

Confident it's still leaking, much more slowly, but still a leak. About 1 micron every 8 seconds. Up from 20 microns to 2 Torr over the course of 3 hours. This compares to 3 microns a second before greasing.

I'll reapply grease properly as per Dennis' directions (didn't wipe the layer the first time) and see how things go. I actually have some new o-rings unopened, so will see how they work out too.

The pressure rise seems linear based on my observations, so I'm leaning away from outgassing. Following the gut Steven!

Tom
User avatar
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: Leak Troubles

Post by Dennis P Brown »

Your leak rate is high for such a tiny system but may be out gassing. Your pump isn't back flowing gas? You have a valve between the pump and gauge? Remember, without a anti-backflow device, pump oil will out-gas into your system and you may be seeing that.

You changed the oil but often if a pump has a lot of bad stuff, one needs to open the pump and clean it out manually using solvents and lint free rags (I've done that to pumps before that needed it; such a job isn't easy and one can also easily damage o-rings/seals so it isn't a job done unless one is ready to deal with those issues and is careful. Of course such seals can be very bad and will fall apart when the pump case is opened; in which case they needed to be replaced.

In any case, I would carefully check the surfaces of all flanges and look for tiny scratches - remember, grease is like a sponge to water and other things that can out gas; that is why one wants as little as possible.
Jerry Biehler
Posts: 975
Joined: Tue Nov 24, 2009 3:08 am
Real name:
Location: Beaverton, OR

Re: Leak Troubles

Post by Jerry Biehler »

You are not trying to reuse conflats, right?
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

Nope, not reusing any conflats... As it stands I actually have 0 conflat connections for testing. Chamber is all conflat but I don't have it hooked up, just trying to get these KF fittings under control.

Dennis, I do have a valve on the pump inlet, so I don't think there should be any oil outgassing. I test the leak rate with the pump off and valve closed, so in theory the gauge is completely isolated from the pump. Hopefully the pump doesn't have badness inside... It's producing much less vapour now, no fog in the shed air after running for a while (I know, I'm getting a longer exhaust hose to run out the window), though it's a little noisy. The oil remains clear after ~6 hours operation. 3 x 2 hour runs, a little more maybe.

I'm pretty sure it's leaks in the KF fittings rather than the pump. I'll keep at them for a few days and see if things get better.

Tom
User avatar
Richard Hull
Moderator
Posts: 14991
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Leak Troubles

Post by Richard Hull »

KFs with smooth faces rarely leak or outgass. The key factor is no fine sleeks should be visible on the faces of the mating flat surfaces. When properly tightened, the elastomer ring really spreads out over the flat, but micro sleeks and ultra fine scratches of some length or depth can defeat the nice broad area of contact.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

Fairly certain I'm only dealing with out-gassing now, and possibly vanishingly small leaks. Pressure rose from 20 microns to 1.5 torr in about 3 hours, and 12 hours later it's only at 2 torr, so the rate is definitely sloping off. It's still a fairly large 'leak' - 1micron/5 seconds at the start, but I hope that'll reduce with some pump downs.

To fix, I followed Steven and Jim's advice - Sanded each KF face and polished to a mirror finish. Cleaned all centering rings and KF faces with Acetone, and greased all o-rings. I'll post some pictures of them later. Peter Schmelcher's post on polishing up KF faces was helpful: viewtopic.php?f=10&t=4175. For the actual process of sanding + polishing I first asked a man I know who works in a metal shop to sand + polish the faces on one piece for me and then I worked from them, trying to copy his work on everything else.

I started with 120 grit, then went 240, 320, 600, 1200, 1500, all manually, followed by Peek polish with a rag. I did everything manually, placing the sandpaper on a cloth on the bench surface and rubbing the piece against that. Each face took about 30 minutes to do from start to finish. I followed Dennis' advice for greasing the o-rings.

Now, I haven't checked all the parts to see if they're leaking, but the gauge, KF nipple and valve, which accounted for most of the leak (if not all) seem leak free.
If I can get the out-gassing dealt with I should be good to get the turbo pump up and running.
User avatar
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: Leak Troubles

Post by Dennis P Brown »

If your "leak" rate is just one micron in five seconds (if that includes your chamber volume, that isn't bad at all), why aren't you running your turbo!? That "leak" rate should easily allow your system to get in the very low 10^-5 torr range and possibly the 10^-6 torr range.

Remember, once air is allowed in, water vapor will cover all vacuum surfaces and easily create the leak rate you are measuring. That is why I built an air dryer system for my vent air source. Improves things a great deal; however, a dryer is not needed for a fusor - I use the air dryer only for my accelerator.

More to the point of what you are doing, if your system gets to the mid/low 10^-5 torr range with the turbo, then it is fusion grade! That means your system, running in the micron range with deuterium being leaked in (against the turbo that is being gated) will be more than sufficient to allow fusion. But again, if your system reaches the mid/low 10^-5 torr than it is doing very well and you don't need to pursue leak methods; even 10^-4 torr will mean the system can do fusion; however, with the gate valve (or what ever valve system) fully open, 10^-4 torr means there is real leaking (not just out gassing.) At that level of vacuum, using 'dust off' or alcohol with your high vacuum gauge on will locate such leaks easily!

My large fusor had to have its chamber run with a plasma before my out gassing drops from a micron every few seconds to one micron every minute (water vapor.) A plasma is a great way to clean a fusor! I always pre-clean my fusor in that manner before making a fusion run.

Aside: any HV power supply can be used to "plasma ash" your fusor; that is, a simple NST (AC is fine) will 'strike' a plasma that will clean your system. That is a simple way to improve the chamber and reduce out gassing.
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

Today I ran the turbo pump - thanks for the prompt Dennis.

The RV3 pumps the chamber down to 40 microns in 10-15 minutes and in another 10 minutes down to 20 microns. I find isolating the system from the RV3, letting the pressure rise a little (to 60-80 microns) with outgassing and then open up the pump again, it gets the pressure a bit lower. The RV3 is already running through the Leybold Oerlikon Turbovac 150, so once the system was down to 20 microns, I started the turbo pump. Pressure in the fore line quickly rose to 80 microns, a graph in the pump manual shows a rise in forepressure, so this is to be expected. The Pirani gauge is located between the RV3 and turbo pump, not on the chamber. I have a Penning gauge on the chamber, part of a LEISK VC15 unit, good for 10^-5 to 10^-10 mbar. The readout for this with the turbo running (after brief acceleration phase) was a steady 2 x 10^-7 mbar // 1.5 x 10^-7 torr, this was after 5 minutes max. I find this suspect, it seems quite low for a Fusor system with no bake-out or rigorous cleaning, though maybe I'm lucky. It just seems a little too good to be true, though comparing to Andrew Seltzman's post (http://fusor.net/board/viewtopic.php?f=10&t=11373) he's getting a small volume with a small pump down to 10^-5 quickly. Does 10^-7 sound sane here? Pumping speed is 115 l/sec with a 14l chamber.

However, the Penning gauge operates based on ionisation currents and I read that above 1 micron, the response is not reliable. I.e: Instead of reading at the top of the 10^-5 - 10^-10 scale if it were at 20 microns, the gauge could be giving some wonky reading anywhere in the range with no useful relation to what the actual pressure inside is. I'm suspicious the turbo didn't actually get really pumping and the

Anyways, after 10-15 minutes with the turbo on normal operation, the failure light lit up on the controller and the turbo spun down. The controller attributes this to too low a rotational speed (below 80%) in the turbo. The turbo pump manual advises checking that fore vacuum pressure is not too high, or else gas volume is too great (leak) for the pump. I'm going to try running again tomorrow to see if the turbo doesn't shutdown and check if the Penning gauge pumps down through the 10^-5 and 10^-6 ranges before getting to 10^-7, as opposed to jumping straight there with a wonky reading. I can start identifying problems then.

Maybe I have a tight system that's performing exceptionally well. Here's hoping.

Also, I finally got an extended exhaust hose for the RV3, no more smelly oil mists in the shed.
User avatar
Richard Hull
Moderator
Posts: 14991
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Leak Troubles

Post by Richard Hull »

Puzzling report. The turbo shutdown due to speed fall out sounds like gas drag due to high foreline pressure like the manual said.

Was your slow leak rate stated in the past for the foreline only or the foreline+ shut off turbo and the fusor chamber?

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

The leak rate for the foreline + shut off turbo + chamber averages at 1 micron / 10 seconds from 20 microns up to 80 microns. This is a little faster at the start and slower with higher pressure (as expected, outgassing), but it takes 10 minutes to rise from 20 microns to 80 microns.

I did another run today and the turbo ran fine for 15 minutes without any failure, I switched it off then. I think yesterday's failure may have been a once-off I'll run again tomorrow or tonight and see if anything goes wrong again.

Regarding the Penning gauge - When I started the turbo, the gauge moved down the range through the e-4, e-5 and e-6 readings and sat at 2 x 10^-6 after 3 minutes with the turbo running. A few minutes later, it was down at 3 x 10 ^-10, which seems absurd for a system like this - no bakeout, no rigorous cleaning. I'll see if this happens again with next run, the 2x10^-6 reading seems plausible though.
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: Leak Troubles

Post by John Futter »

Tom
your penning gauge is dirty if it stops like that and reads an impossibly good vacuum.
I looked at your system and could not see the gauge head unless it is a philips Granville unit
the Edwards ones we regulary clean @ work every 3-5 years on a clean system or twice a year on a sputter system or ion implanter.
cleaning consists of of dissassembly and all parts polished with increasing grades of carborundum paper using a lathe to quicken the process. A wash
of the parts with acetone then ethanol and reassembly. takes about half a hour per gauge head
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

John, pictures attached of head. It's a LEISK unit.

Looks like there's another, smaller conflat fitting on the head, halfway along the length, behind black part. The port is a CF 2.75" fitting, I guess the inner one would be a 1 1/3" CF, though I can't get any manual so I'm not sure. I have a second unused one, just have to find it, it's lying around somewhere. Otherwise I'd rather hold off disassembling as I'd have to replace internal gasket and I don't have any gaskets less than 2.75" at the minute. If I can confirm size I'll order gaskets and disassemble.
CF 2.75 Port
CF 2.75 Port
Full head
Full head
Cable plug
Cable plug
User avatar
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: Leak Troubles

Post by Dennis P Brown »

I am both surprised and concerned by your turbo indicating the pressure of your system is too high or gas load too high. Either your gauge is not accurate for the leak rate you indicated or else the turbo pump has issues (remember, high load can also be caused by bearing's having issues ... ); I see you are addressing the gauge issue. Nothing can be done if one's vacuum gauge is in extreme error; that 10^-7 torr is in error for a unbaked system with that leak rate.
Ten minutes to go from 20 microns to 80 microns and your turbo indicates stalling after running!? That does not follow. My fusor far exceeds that leak rate after exposure to air and my DP pumps it into the low 10^-5 torr in under ten minutes; my turbo system has a similar leak rate and bottom outs 10^-5 torr in under five minutes (it is both an old and very small turbo.) So, definitely clean or check your gauge.
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

Wary that there's a small CF gasket inside the heads, I was reluctant to dismantle until I got said gaskets for replacing, but I took apart the gauge head today. Looks like the internal CF gasket is a DN25/ 2 1/8" fitting. OD of gasket is 30mm , ID 20mm, DN25 gaskets are the closest I could find anyways. Though I don't have them here yet.

I was surprised on dismantling to find the whole thing quite clean to my eyes. No discolouring coatings or dirt buildups. The sensor itself is almost immaculate on the outside. I haven't tried to get at the inside, and I'm not sure if I should, seems delicate and better left alone. The dirtiest part is a streak along the inside of the CF connect between port and sensor.

Does this look clean to you guys? I've to wait for a set of DN25 gasket replacements before trying it out again, and I can compare to the other gauge head, which I found yesterday and dismantled too. The unused head is actually a lot dirtier, with discolouration, so I'll have to give it a clean.

John, you mentioned sanding the parts with high grit paper, I'm not sure I should go near the sensor though. What do you think, just give it a good wipe on the outside with acetone and ethanol, or should I even try and get inside the sensor? Doesn't seem like a good idea to me as I'd have to break and resolder parts.
Penning head disassembled
Penning head disassembled
Notice browny streak 3/4 of way along inside of tube
Notice browny streak 3/4 of way along inside of tube
Side 1 of Sensor
Side 1 of Sensor
Side 2 of sensor
Side 2 of sensor
Roberto Ferrari
Posts: 364
Joined: Tue Mar 02, 2004 12:21 pm
Real name: Roberto Ferrari
Location: Argentina
Contact:

Re: Leak Troubles

Post by Roberto Ferrari »

Hi Tom
This kind of gauges rely on measurement of very low currents.
So you need to pay attention to cleaniness on insulators. Look inside the conflat.
Roberto
User avatar
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: Leak Troubles

Post by Dennis P Brown »

No knowledge of that specific gauge head but have cleaned an Edwards. That had a filament that one could not clean since it had a coating (the rest could be cleaned and that fixed the gauges issues nicely.) Bottom-line: call the company and confirm that cleaning is allowed and what solvent is acceptable before trying to clean it. Destroying a gauge isn't fun.

Is there a way to cross reference the gauge head's performance? I used someone else's gauge system to cross calibrate mine when I was unsure if it was any good. Also, cheap TC gauges are available on ebay that can partly serve that purpose; I later bought a cheap ionization gauge/head to backup and cross reference my existing High Vac gauge.
Peter Schmelcher
Posts: 228
Joined: Sat Nov 13, 2010 1:56 am
Real name: Peter Schmelcher

Re: Leak Troubles

Post by Peter Schmelcher »

Tom have you tested your penning gauge electrical feedthroughs as they can leak, mine did and on a nearby part I had a cracked weld.

Wrap nearby or all fittings with plastic wrap and/or tape. This helps reduce dust off overspray when hunting for multiple leaks and if possible bag your gauge in sections.

Unfortunately some leaks can have time constants measured in hours.

The joys of using recycled parts.

Good hunting
-Peter
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

Thanks for the help guys.

I'm a bit stuck at the minute on any testing, as I can't get a gasket to fit the internal CF fitting in the gauge. It doesn't seem to fit any of the standardised fittings. The gaskets outside diameter is 29.4mm (1.157") and this is an exact fit for the port, I tried DN25 gaskets, 32.79mm (1.291") OD and they won't fit, too big. ID is 21.7mm (0.853") and thickness is 3.8mm (0.15"). Does anybody have any idea of what sort of size fitting this is? The port itself is a bit strange too, compared to normal conflat. It only has a single recess, so the edge of the fitting itself acts as the knife edge, as far as I can tell - instead of having two recesses, with the middle one being the knife edge. See attached picture.

I tried running with the old gasket, with some hope that it might hold, but no luck. When I find replacements I can try and follow the advice given. I'm also in a hard spot regarding technical support on the gauge - LEISK closed down a number of years ago, and they've left absolutely no information behind. Whoops.

Tom
Attachments
See single recess
See single recess
gauge port.JPG (27.26 KiB) Viewed 6186 times
User avatar
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: Leak Troubles

Post by Dennis P Brown »

That normally uses a copper gasket and that is often best (consider getting some); however, special "o-rings" can be purchased that fit those types of flanges and you need to get the correct sized one from the company's that sell those types of fittings - that is, as you have found out, those are not a standard size because they are not o-rings but square (flat sided) gaskets that look very much like O-rings. Failing that, cutting your own flat "washer" shape gasket is not exactly difficult and will work until you can buy the correct gasket.

A semi-safe cleaning fluid is absolute ethanol alcohol. Can't say it wouldn't harm any coatings but it is the best compromise choice. A dip in a warm bath of the Alcohol might do the trick - do not use compressed air to dry since that might damage filaments.
Peter Schmelcher
Posts: 228
Joined: Sat Nov 13, 2010 1:56 am
Real name: Peter Schmelcher

Re: Leak Troubles

Post by Peter Schmelcher »

If your gauge works fine at higher pressures and then stops as the pressure drops you might have a magnetic field strength problem. If you adjust the magnet position and center it over the cell it might operate to a lower pressure. It is also possible heat has demagnetized the magnet.

The voltage between your electrodes could also be low. You might read a few other penning manuals for typical voltages.
Penning manual TINA14E1_Web.pdf
(752.87 KiB) Downloaded 861 times
BTW there is nothing magical about conflat seals, they worked well enough, others used them and it became one of the survivors. Wheeler also designed a wire sealed flange, just a squished copper wire between two flat surfaces.

If you can’t buy a replacement seal, repair the old one with a little solder in the groove or make your own by filing a CF gasket to make it fit.



I meet a retired engineer enjoying life as a part time docent in the Santa Fe technology museum. A piece of trivia is vacuum chambers for the Manhattan project were made of plywood with big diffusion pumps. I asked how thick the metal liner was and he responded just painted on the outside.

The physics of some high vacuum gauges
Ionazation CERN p145.pdf
(1.3 MiB) Downloaded 1740 times
-Peter
Jerry Biehler
Posts: 975
Joined: Tue Nov 24, 2009 3:08 am
Real name:
Location: Beaverton, OR

Re: Leak Troubles

Post by Jerry Biehler »

Look up copper crush gaskets, they are used on all sorts of things like hydraulics. Aluminum ones too. Either will work and there is a huge selection in size.

Heck, you could put a nitrile o-ring in there and it would be fine. You would be limited on bake out temps but I would not worry about it for what you are doing.
Tom McCarthy
Posts: 405
Joined: Wed May 08, 2013 3:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy »

Whoa, thanks for the tip Jerry, copper crush turned up a treat, and affordable too. Ordered a pack of 1.5mm thick ones, they should do, old ones were 2mm. Arriving Monday.

Thanks for the gauge info Peter, the chamber actually seems to be holding up at 20 microns with the old gasket in the gauge, I've yet to check readings with the turbopump running though. Reading through those docs.

The more you get into it, the less mystifying the nuts and bolts of science hardware seems to get. From the outset it seems incomprehensible, but things simplify as you work on them. Point being it seemed to me at the start as if conflat etc had to be rigorously justified by in depth models and so on. In reality, the only model that looks like it matters is experience.
User avatar
Richard Hull
Moderator
Posts: 14991
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Leak Troubles

Post by Richard Hull »

That is what the hands-on imperative is all about.....Do and learn. It is why "labs" are used in college. Labs in college are only for exposure. Living of life must forever be a constant "lab" experience with tremendous and continuous hands-on events, going far beyond mere exposure to border on or achieve expertise.

When no working with the hands you should reading with the eyes.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Post Reply

Return to “Vacuum Technology (& FAQs)”