All journeys

[William A Washburn]

Thanks Doug,

I know it's got to be rough on you schedule to photo, video interview and writeup these little pieces
but for those of us wet behind the ears it is great information for us to really use to see what goes on.

Thanks, Bill

[Lee Hall]

We're having some problems (I think) at the moment. We got some vacuum fittings to secure our 1/4" copper tubing by pass loop into our 1" bellow tubing. We got the pressure back down to our previous level of 1 mTorr, but for some reason when I close the main stop valve and pull air just through the bypass loop, the pressure slowly rises (about .1 mTorr every second or so). This is a HUGE concern and I'm not sure how to trouble shoot. I already checked the flanges on the fusor for small leaks using alcohol and I didn't find any. The connections of the copper tubing are as perfect as we can make them (they look perfect to me). I'm really not sure where to go from here...

Perhaps more troubling is the video we just took. I brought the fusor down to 1.3 mTorr and at around 17 kV I leaked some air into the system. The pressure got to around 2.5 mTorr and we got a BRIEF glow that was really bright and then it dimmed a bit and, in less than a frame on the camcorder, it went out. The power supply is reading an overcurrent, but it is set at 25 mA for that. I put the pictures from the start of the glow to the end where it cut out below. If nothing else, they look different than what we've gotten before.

Beginning

Bright, middle

dimming, end

[Doug Coulter]

This kind of thing isn't uncommon to see when you have a leak and things are drifting. The reason the pressure goes up on bypass-only is that your leakage is big, and cutting down the pumping speed makes things equilibrate at a higher pressure. In any vacuum system, the net pressure is some equilibrium between leak/outgassing and pump rate -- even my really clean/tight system.

The pump says "we go below e-11 mbar" but heck, I've never seen that (and probably never will). In my case, I'm lucky to get within 2 orders of magnitude of that number -- and that's with a 500 liter/second pump! I can see permeation through the small amount of viton sealing I have, as well as outgassing on my mass spec, and pretty much a little of everything that has ever been in the tank since the last laborious detail-cleaning. When I pump down again after opening the tank, I can see my fingerprints as they evaporate on the mass spectrometer (takes a couple days sometimes).

When I was having such problems, I found that using a deliberately injected gas that has a different glow color than air (say, He or Ne) helps to see just how bad it is, and can sometimes also be used to find the doggone leak in the first place, by watching for a color change in the glow (just set current limit low, say 5 ma, no need for big power at this point, just need a glow to go by).

You have a large leak (by vacuum standards). That's got to be priority one on your project -- find and fix that bad guy. At this scale, the "large leak" may not even be enough to see light through (but you should try that -- because the human eye has a lot of dynamic range once dark-adapted), but it can still be big by vacuum standards. Think how much air it takes at STP to create a millibar in your tank -- in this example, it only takes 1/1000 of the tank volume as a leak. And even a drip can fill an ocean eventually. With a slow pump, it doesn't take as much leak to make trouble, either.

[Lee Hall]

Alright, so after a long day of troubleshooting and making some new connections, we've helped the situation out some. So we found a leak between a stop valve and a needle valve we have in our bypass loop, so I think its fixed. As far as all the connections, flanges, and valves, I used an alcohol-dampened paper towel and looked for pressure spikes. I saw none, while before the valve connection was fixed, I saw a pressure spike when i put the paper towel between them. I also shut both valves off and saw the pressure shoot down. We're getting to low pressures again (1.2 mTorr or so), but the pressure would go up again when only the bypass loop was used. This went up fast the first couple of times, the pressure would go up fast. However, after letting the pump run for a while, the pressure went up pretty slow (about .1 mTorr every 5 s.). Hoping this trend continues, we're going to leave it on pulling just through the bypass loop and see where we're at in the morning. To give people a better idea of our system, I have attached an updated set of pictures below.

Gas Reservoirs (one for D2, one for any other desired gas), and fill pressure gauge

Needle valve on the fill side

Bypass valves: Needle valve left, stop valve right.

Main line and main stop valve

Entire setup. Note the glyptal painted in rings on the sphere.

[Joe Jarski]

Lee, if I understand your setup correctly, then your bypass loop is the small diameter tube that you are trying to pump through. If that's true then it may be contributing to some of your problems. Leaks are always a problem, especially when it comes to contamination, but it's really difficult to get a good base pressure through small tubing - the bigger tubing the better with vacuum. Shortening any tubing runs as much as possible will help too.

[Lee Hall]

Joe,

That definitely makes sense to me. I was lobbying for the use of 1" bellow tubing for the bypass loop before we built it, but that was mainly because its easy to move around and change. Originally, our bypass loop simply curled out straight from the two Tees in the picture, so each tube up to them were less than 8 inches long. The downside to that is our view port points right at the valve location, so our adviser said move it. So, we basically had to move it where we did for adjustments while we're operating the fusor. I would have liked to re-organize the entire cart, but we're scrapping that setup soon and moving to a vertically oriented reactor so everything will change. Would you think simply running the vacuum continuously for days will help get out the contaminants? We also have some Argon we're planning on running through the system to condition things, will that help?

[Jerry]

Airgas is the last people you want to deal with for the kinds of parts you need for this setup. Possibly someone in their spec gas area might be useful but Airgas as a whole will be useless. They are primarily a welding supplier. Thats what they know.

[Doug Coulter]

I'll second the other guys here. Matheson Tri-gas is a lot better, and know their stuff, as they specialize in process gases for things like this (and the semiconductor business). But they aren't real cheap either, depending on who is asked. I'm told I actually got a great deal on that two stage regulator for a mere $400 from them. It IS a good one.

For your bypass, the whole point is to have less pumping speed so it's easier to control, not much worry there, but the main connection to the pump is much too long and too much surface area (bellows add a lot) to outgas quickly. Your pump is really marginal to get to the vacuum level you need anyway, so adding another big gas load (it's surface area, not volume that kills you) isn't a good idea. You should plan to add a diff pump or turbo at some point, because you're going to want it for sure once you get a little closer and see what the benefit is going to be. Old surplus leak detectors are a gold mine of the kind of parts you'll be looking for, and there's always ebay.

The best pressure for fusion is just before the point where you can't get it to run at all. So if you can't get to vacuum that just won't draw current at your full voltage, you're not there yet -- that point is only a few percent lower pressure than the place you want to run it. A factor of 10-1000 better than that is nice if you want to run in pure fuel rather than a few percent D, the rest contamination.

Argon is good for conditioning, but watch that you don't do it endlessly, because it's also good to sputter metal off the grid (and onto things you don't want it on). Keep the current and time on the low side till you get familiar with your setup and argon. At glow with dark space pressures, its very good at removing all the little surface irregularities from a grid. But of course, that stuff winds up somewhere and you hope it's not your insulators.

[Lee Hall]

This is troubling to hear to say the least. Right now, we have 5 feet of that 1 inch bellow tubing in the system. The least amount we can use (I think is 2 or 3 feet). The bypass loop itself probably contains 6 feet of the 1/4 inch copper tubing. After leaving the pump running all night with only the bypass open, it stabilized at 40.6 mTorr this morning. Since then, I have fully opened all of the valves and left the thing running. What I'm really concerned about is how our system will change as we add borated poly shielding (they're saying 20 inches!!). This would mean the length of the main line and all other lines would just get longer. Would using a short (2 feet or less) bellow tubing that then is adapted to copper tubing be a better main line? I'm thinking something on the order of 3/8 inch copper tubing or so. Also, would shortening up the bypass loop help us out enough to worry about it? Regardless, I still think we'll run some D2 through the system next week.

We'll definitely look for a leak detector ASAP.

[Doug Coulter]

Well, the minimum reasonable length of bellows is just enough to handle vibration -- call it a few inches. You are off the charts as far as (lack of) conductance of your plumbing goes. At the pressure levels involved, you are no longer talking about pressure -- where existing gas molecule collisions "force" other molecules into the pump. You are getting into molecular flow, which means that you are just waiting for them to find their way in there at random, which takes a lot longer -- the gas molecules rarely "hit" one another at that point. So, no, skinny copper pipe will only help with surface area to outgas, but not so much pumping speed. This is why you see 6" and larger pipe on all high vacuum systems, and usually short enough that you can see through it with maybe 30 degrees off axis viewpoint. Sad, but true.

For the bypass system, this is fine (once you outgas it) -- you want slow pumping via that loop for when you're trying to balance a slow deliberate gas inflow with a slow (and about equal) pumping speed. For getting there in the first place, even a foot of 1" bellows might be too much, particularly since you're trying to get near the limits of the pump in question. I have for example a 4" long by 6" diameter pipe between my final pump and my system for the big system, and well under a foot of 4" pipe for the small one. And the small one gets to e-5 where the big guy gets to e-8 in about the same time...That's with no leaks at all, just outgasing. Once you've got your bypass line "clean" then just ensure it doesn't have to go to STP everytime you open the system -- a valve here and there helps with that, so you never stop pumping on things like that if you can help it. Filling the tank with dry welding gas when you have to let it up to atmosphere does help and argon is a good one for that -- it's dry and doesn't "stick" so well to the tank and plumbing walls as say, water does.

See how it does with the main pumping scheme operating overnight (not just the bypass), that will tell a lot. And maybe whether you have a leak too. One way to tell is with more sensors, you might find my cheapo pirani sensor project worthwhile. For example, if you put a sensor right on the pump, and one on the tank, and then close off the pump valve and the pump sensor goes way down -- you know the system has issues. If the sensor at the pump is always reading much lower than one on the main tank, then you also know that you don't have enough "conductance" to get the job done. These are cheap to make, so you can afford to have a few of them here and there on your system, and it's very educational. I can give you a few of the sensors (grain of wheat bulbs, hard to find, so we bought hundreds when we did).

All that junk in the path limits your effective pumping speed a lot. As in orders of magnitude. You can probably find speed vs pressure curves for your pump (or one just like it) at Leskers website too. And note the funny units they use to handle the case that at lower pressures, you're not pumping as much actual gas as at higher ones.

I tell everyone to sign up and get the free Kurt Lesker catAlog, the old print version of which has a better section on all this than the document below -- but this is a start. You have to get it in your head that this isn't just water flowing through a pipe, and that you can get it all in a couple minutes of pumping. As you get close to the area fusion happens in (long mean free paths) a long skinny pipe may have more or less zero flow -- not enough pressure drop between an assumed perfect vacuum at the pump end and the millitorr at your chamber exists to "push" gas to the pump, and in the case of molecular flow, it's not "pressure" at all -- just random chance a molecule will bounce it's way into the pump (and not just bounce back out).

kjlc_ed09_sec17_web200910.pdf

Kurt Lesker Tech notes.

(2.41 MiB) Downloaded 300 times

However, I suspect you still have a leak. Most outgassing should be over with overnight, at least using the main pumping line (short, fat, shiny).

You can try this:
Bake the system. You can use heat gun, IR light, heat tape, whatever. At first turnon, that should cause the pressure to rise, and plateau. Then, it should slowly drop back to and even below what it was when you started out -- this may take hours. Then turn off the heat, and you should see the pressure drop further. If not, you've got a leak still. Raising the pressure via heat sounds dumb at first, but it's not -- it's making it easier for the pump to remove the junk off the tank walls -- and sans leaks, once it's gone, it's gone, which is why after cooling, you wind up with less pressure than at the start.

If there's no viton, you can get the system pretty hot (a couple hundred C) without harm, if there's viton, keep it to 100 C and wait (much) longer. Don't skip heating anything in the vacuum, including all the hoses up to the tank (but be careful about overheating valves that have viton or plastic in them).