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: 404
Joined: Wed May 08, 2013 7:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Leak Troubles

Post by Tom McCarthy » Mon Feb 20, 2017 8:59 pm

Turning to the more enlightened vacuum folks of the forum for help. Can't seem to get my foreline working.

I had my roughing pump (Edwards RV3) hooked up to a valve, couple of inches of hose and to my Pirani gauge. After some work and problems with the gauge ( http://fusor.net/board/viewtopic.php?f=10&t=11196 ) the small system pumped down to ~5 microns and I connected it up to my full, blanked-off chamber. Took a short few minutes longer to pump down (as expected), but got to single digit microns again and after more than 36 hours, the chamber was up to just about half a Torr, so it chamber was pretty leak tight.

Following this, I went to set up my turbo, bolted it onto the chamber, attached forepump through turbo exhaust and checked how the forepump was working. No joy, pump could barely get past atmosphere - turbo wasn't bolted properly (could hear the leak), so I tightened turbo conflat bolts and gauge now got down to single digit torr.

Wanting to make sure an issue hadn't appeared in the foreline, I disconnected it from the turbo so I could check it by itself and promptly ran into confusing results. Major discrepancy is between setups 3 and 4 below.

---------

Setup 1: pump -- valve -- two hoses -- gauge.

Takes only a few seconds to get down to 30 microns, slowing down as it goes and inching past 20 microns. If I turn the pump off with valve open, pressure shoots up to 2 Torr or so. Turn pump off with the valve closed there's no huge shoot up, but a quick, smooth leak of 3 - 4 microns/second. This is a problem. To try and diagnose where the issue is I took the line apart and added parts piece by piece to see if anything contributed dramatically to the leak.

Closing the valve, turning off the pump and then opening the valve still gives a shoot up, but less dramatic. Goes up to 800 microns - 2 torr and a lot less quick, but still quick.. 2 seconds as opposed to instantaneous. This isn't as much of an issue as I can just valve off the pump and there's no problem, also usually there's a small raise in pressure from atm when the pump starts and then it starts pumping down. Persists across all the setups described, so I won't repeat it below.

-----------

Setup 2: Gauge deadheaded directly on to pump inlet (with KF 10 -25 adapt)

As above, takes a few seconds, gradually slowing down as it goes to about 30 microns, then much slower inching down past 20 microns. Couldn't isolate gauge from pump, so couldn't check gauge for leak by itself as pressure just shoots up to 2 torr when I turn off the pump. When deadheading previously, it would get down quicker to single digit microns, seems to have slowed down, leak developed in the gauge connection perhaps.

-----------

Setup 3: pump -- valve -- gauge

Same pumpdown as above, gets to 30 microns, slowly gets past that. Closing the valve lets me check the leak rate, which I took some measurements as follows to do some real diagnostics.

40 -> 60 microns: 7 seconds
60 -> 80 microns: 7 seconds
80 -> 100 microns: 7 seconds
100 - 200 microns: ~45 seconds

Looks like a leak.
--------

setup 4: pump -- valve -- hose -- gauge

Here's a problem.

40 - 60 microns: 30 seconds
60 - 80 microns: 30 seconds
80 - 100 microns: 35 seconds

Looks like a leak. But it's nearly 4 times slower than in setup 3. This goes against my intuition, as all the offending fittings from setup 3 are still present along with the hose and an extra KF connection. Can anybody explain this?

---------

One thing I did notice is some grime/oily dirt on the pump inlet (pictured). From this post ( http://fusor.net/board/viewtopic.php?f= ... 4&start=30 ), could oil contamination be causing a problem?

Also, I need to find the exact locations of the leaks, I tried painting the rings with acetone one by one, but there was no noticeable shoot up on the pirani gauge. Will dust-off work better? I find KF connections hard to get at with this sort of spray-on/paint method as the clamp around them is a big obstacle. To get around that, I just turned on the pump and one by one removed a clamp from a connection, painted the o-ring circumference with acetone and put the clamp back on. No luck with that. From tests above, it looks like the gauge connection (gauge + adaptor) is leaking. What's the best way for me to deal with this? And once I've found the leak, should I just put a new o-ring in, or lather it up with some sort of grease, how to actually stop the leak?

Once I sort the foreline out I may have issues with the chamber, but need to get over this hump first!

-- Picture previews are showing up sideways, click on them and you'll get a full size image the right way up.
IMG_0039.JPG
Setup 3
IMG_0042.JPG
Setup 4
IMG_0041.JPG
dirt on inside of pump inlet
IMG_0037.JPG
two hoses, not described in setups
Tom

User avatar
Richard Hull
Moderator
Posts: 11635
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Leak Troubles

Post by Richard Hull » Mon Feb 20, 2017 9:28 pm

It looks like your pump is belching crud into your system and gauge!! This garbage looks like it is coming from your pump for sure! your gauge may now be fouled. Is that a bead of water I see in the photo??!! A big, big, no-no.

Never ever run with a gauge on the line when flushing!!! Blank off the pump inlet!

I once bought a used pump and dumped the oil which was filthy beyond belief. I ran a charge of fresh flushing oil for a few minutes and dumped it. The oil was still kind off cloudy. I ran another load of flusihing oil, but this time for two hours. This second flush of oil was more filthy than the first!!! A third flush for two hours of run time finally yielded acceptably clean oil. A charge of good pump oil and a T/C gauge showed the pump went down to about 15 microns in under one minute. I filled the T/C gauge tube with acetone and shook it vigously. I, then, dumped it out. after an hour of air drying, it went down to about 10 microns. T/C gauges can, over time, get fouled.

Just putting new oil in a pump of unknown origin is a losing situation. Run a first flush on a used, blanked off pump for 1 hour, minimum. Continue until additional flush oil drains clean and rather clear. Only then, put in decent oil and hook you gauge up to see what you have.

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
Retired now...Doing only what I want and not what I should...every day is a saturday.

Tom McCarthy
Posts: 404
Joined: Wed May 08, 2013 7:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy » Wed Feb 22, 2017 4:19 pm

Thanks Richard. I had decontaminated the oil a few weeks ago when trying to get my gauge working (ran for a couple hours with gas ballast open), but didn't replace it.

Today I cleaned out the sight glass, inlet and filter. I also drained the old oil, which was clear, but a kind of yellowy colour, see the picture below and flushed some new oil through. I then put a full fresh charge of oil in, and am currently waiting on the pump now, running it for 2 hours or so to see how the oil turns out. Note that in the picture of oil below, the solid particles you see (vast majority of them anyway) were already in the bucket, and didn't come out from the oil.

Hope this fixes things up, if not I'll give the gauge an acetone clean. It's an Edwards PR10K, Pirani gauge. Looks like I'll need to dismantle the gauge, as there seems to be a danger of damaging a filament. Is this true, or can I just immerse the thing in acetone as Richard's described and shake it around and leave to dry?

http://mfc.engr.arizona.edu/documents/E ... 157880.pdf

Page 11:
The interior of the gauge tube must not be cleaned as damage to the
filament may occur. If contamination is suspected the gauge should
be returned to BOC Edwards for exchange/replacement.
Will update ye when I've checked how the RV3 is performing with new oil change. May not have to go near the gauge at all, but from Richard's advice it looks like it will need to be done.

UPDATE: emptied the new oil out after > 2 hours running and about an hour to cool, it came out clear as water. Filled up with new oil and attached the gauge. Bit quicker on the pumpdown now, but not much difference. I guess the gauge needs cleaning. There's also definitely a leak in the KF fittings (KF10 - 15 centring ring, KF15 - 25 adapter) attached to the gauge, as isolating it leads to a rapid rise in pressure, as described in the original post.

Tom
IMG_0044.JPG
Pump oil, first change

User avatar
Jim Kovalchick
Posts: 344
Joined: Thu Apr 14, 2011 12:00 am
Real name:

Re: Leak Troubles

Post by Jim Kovalchick » Wed Feb 22, 2017 7:49 pm

Tom,
Your KF seals are leaking. In setup 3 with the valve closed you should see very little rise in pressure. Did you try a little grease on each viton seal ring?
My son and I have an RV-3 in our lab too. It's pretty reliable, but we did have a vane failure once that locked the shaft and shattered the motor coupling fingers. A rebuild kit fixed it up. I don't ever rely on that spring loaded back flow valve on the pump inlet. It will always leak back when the pump is off. This can lead to streaming of oil vapor into your turbo.

Good luck

User avatar
Steven Sesselmann
Posts: 2111
Joined: Thu Aug 11, 2005 1:50 am
Real name: Steven Sesselmann
Location: Sydney - Australia
Contact:

Re: Leak Troubles

Post by Steven Sesselmann » Thu Feb 23, 2017 2:03 am

Tom,

Trust your gut feelings, you probably know where the leak is but hope it isn't because its hard to get to :)

Pull it apart, clean every part with acetone, check every KF fitting with a loupe, and if you see even a tiny scratch, wrap a cotton rag on a clean sanding block and polish the fitting with a bit of chrome polishing paste, until it's so shiny you can see yourself. Then clean the part and the O'ring with acetone, and apply a smidgeon of silicone vacuum grease to the O ring before you put it back together.

Rebuilding a vacuum system takes less time than messing around trying to find a tiny leak.

For comparison my unorthodox chamber now drops only 3 micron in 1 hour, and I have several KF fittings and a chamber made of 20 discs with 40 surfaces glued together with Latex. I bet some chambers made with Conflat copper seals aren't even that good.

Steven
http://www.gammaspectacular.com - Gamma Spectrometry Systems
https://www.researchgate.net/profile/Steven_Sesselmann - Various papers and patents on RG

Tom McCarthy
Posts: 404
Joined: Wed May 08, 2013 7:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy » Thu Feb 23, 2017 3:52 pm

Jim, thanks for the confirmation - I suspected leaks, but hoped it was something with the pump instead.

I've ordered a tube of Dow Corning vacuum grease and picked up some chrome paste as you advised Steven, cleaning and inspecting all parts today. As mentioned I did have leaks in mind and was avoiding them, so you're right about that gut feeling.

One question - I've read docs advising against using acetone, spirits, cleaning solutions etc with o-rings. Is it ok to wipe them down with some acetone, or should I use something else for the o-rings?

Will report back when the work is done.

Tom

User avatar
Jim Kovalchick
Posts: 344
Joined: Thu Apr 14, 2011 12:00 am
Real name:

Re: Leak Troubles

Post by Jim Kovalchick » Thu Feb 23, 2017 4:47 pm

Don't use acetone except on metal. If for some reason your o-rings are dirty use a little alcohol.

Tom McCarthy
Posts: 404
Joined: Wed May 08, 2013 7:36 pm
Real name: Tom McCarthy
Location: Ireland
Contact:

Re: Leak Troubles

Post by Tom McCarthy » Thu Mar 09, 2017 12:08 am

Some progress:

Dow Corning grease arrived yesterday. Had already cleaned KF fittings with acetone and wiped down o rings a few days ago (played around with setting up MFC in the meantime). There's no detailed guide on applying vacuum grease online more than 'put a tiny bit on the o ring, less is more.'

I only got it tested with the valve and Pirani gauge fittings connected to the pump, but I've gone from 3 microns/s leak to 1micron / 3 seconds leak. Better, but still a leak. That may improve with runtime, as on first run with the newly cleaned and greased fittings, they leaked up much quicker, but have slowed down with more time in vacuum. Residues gassing off I guess?

I'll do some more runs and see how things are, if a noticeable leak persists I'll have to go at the system again.

Meanwhile, any advice on applying vacuum grease, in case I'm going wrong? What I'm doing at the moment is oozing a tiny bit out of the tube (like a quarter of a fat drop of water on a tap) wiping that on my glove finger and just massaging all around the o ring with it. There's a slight sheen of the o-ring, but no noticeable build up of residue.

Thanks,
Tom

User avatar
Steven Sesselmann
Posts: 2111
Joined: Thu Aug 11, 2005 1:50 am
Real name: Steven Sesselmann
Location: Sydney - Australia
Contact:

Re: Leak Troubles

Post by Steven Sesselmann » Thu Mar 09, 2017 5:24 am

Tom,

If you keep pumping down the chamber for a week or so it will improve significantly as water comes out and if there is a leak you will see that pressure falls linearly.

If it is outgassing only you will usually see the chamber stabilise between 500 and 1000 micron and stop falling, but if it leaks it just keeps going.

Once again trust to your gut feeling and fix the leak.

Good flanges and new O'rings - Unlikely
Professional welds - Unlikely
Tapered threads with teflon tape - Very likely
Conflat flanges fitted properly - Unlikely
Shut off valves bought on eBay - Likely

If you are still having problems take a few pictures and post them here, maybe we can spot the weak point.

Steven
http://www.gammaspectacular.com - Gamma Spectrometry Systems
https://www.researchgate.net/profile/Steven_Sesselmann - Various papers and patents on RG

User avatar
Dennis P Brown
Posts: 1716
Joined: Sun May 20, 2012 2:46 pm
Real name: Dennis P Brown
Location: Glen Arm, MD

Re: Leak Troubles

Post by Dennis P Brown » Thu Mar 09, 2017 7:18 pm

Using any vacuum grease to seal a leak means that either the o-ring or its mating surface have a minor defect. To use grease, follow the following: lightly apply a tiny amount of grease using proper gloves (to avoid lint and contamination of the grease) and spread as thin a layer as possible over the entire o-ring. Then using a lint free wipe, remove all the grease you can (a film will remain.) The grease should cover all the o-ring but should not be visible to the eye or get another wipe and wipe again. Then install the o-ring and test. If it still leaks, then either the leak is somewhere else or the 0-ring/mating surface is too damaged. Either case grease is not the answer.

Out gassing can easily go above 1 torr (1000 microns) but will never bring the system to atmosphere and will definitely slow as pressure increases.

Post Reply