Building the new beam device

[Doug Coulter]

Well, the beam device project is coming along, just at the mockup stage now. Still a few more things to do before it sees a vacuum station, but it's coming along. I finished the remote positioner (or I think I did, but tests will tell). Here's what it looks like:

Moveable target holder

For the moment, I have a piece of Ti sheet I punched out in there. Should be good enough for intial tests, but I of course have fancier things in mind for later. The first test will probably be with a phosphor screen to deal with focus and aiming issues which are expected to be kinda marginal at first -- I'm not an owner of Simion and so on, just going by the seat of the pants here.

This will go into the front hole of the cross depicted here. The top gets a window, the back gets an 8 pin lv feedthrough for diagnostices, and the tubing couplers will accept whatever beam tubes we wind up making (probably a few revs there -- so it's nice not to have to buy gaskets for every test).

Beam device target area

This setup should be versatile enough to do a few things and get some learning in. I can put a beam tube into one of the couplers, and if I like, a HV feedthrough into the other one so as to use split supplies for more volts in beam->target mode. Or I can use two beam pipes for beam on beam stuff. The movable target holder can hold whatever target, with whatever holes and has an extra spot to connect to the lv feedthrough for diagnostics -- for example, I can sweep a target with a hole in it across the beam and see both the stuff that doesn't hit the hole and the stuff that gets through on another target behind this on the other coupler.

Assuming the Phillips book is correct, we should be able to see on the order of a million (to ten million) neutrons/second per microamp with such a setup, using DT, or about 1/100 of that (still pretty respectable) with just D. Should be a hot rod. We're bulking up on shielding materials just in case!

Next is modding a gas inlet valve as described elsewhere here...I'll put that in the bottom adapter that goes to the pump station. This not being my first, I'm willing to take it at the right speed and get all the parts pretty close before worrying about getting to vacuum on it, I'll save a few gaskets that way. But it won't be long now anyway.

[Doug Coulter]

More progress to report on the beam device that's about to get real here. As with a lot of projects, most of it is getting ready, not doing the simple and fun parts right off.
So, I drug the little Pfeiffer pump station out from under the fusor-2 table, and took off the old tank, substituting the new stuff. This system didn't yet have a valve in the foreline, so I adapted one from a refrigeration unit to work with the 1/4" ID plastic hose this one uses for foreline - it's only a diaphragm pump, not a great forepump. To do this I turned out some little hose nipples on the lathe, from brass, that were a solder in fit to the .325" ID of the valve copper pipe. The valve is sitting on the station on the left front, yet to be plumbed and wired in. I'll put a little 24v transformer in there across the power to the forepump, and some sort of switch/pushbutton arrangement to allow me to have the forepump on, but the valve off - for pressure adjustments. Or, of course, just have the valve open whenever the pump is running. I suspect that having that mode will cut down how often the forepump has to run - seems like it's the thing that has any back-leak at all.

Basic Mockup

The long sloping pipe towards the camera is what I use to pump down tanks for gas transfers, glass blowing, making detector tubes and so forth - useful accessory. Has it's own pirani sensor, inlet valve, and there's a gas manifold (with a gage) down there on the bench. You can pull the manifold to zero, then let in this many mbars of this or that gas to get a mix you want, then let the desired amount into that sidearm, for filling gas tubes and detectors with a specific mix.

The lower cross has ports for that, the main pressure gage, and the mass spectrometer for this system. The upper 6 way cross gets a fixed target and feedthrough on the side away from the camera, the back is the push-pull-twist toy for the movable target, the top gets an 8 pin feedthrough for low voltages (faraday probes and target current monitor) , the front gets a window (which I've got to make a mica sacrificial shield for) and the side with the Al plug will accept the beam tube when I make that - the easy and fun part probably. But none of it works till it all works.

I figured I could use a target with a bunch of thermal mass, that I could either measure current on, or put voltage on and also measure current. Eventually, the idea here would be to evaporate some Zr or Ti onto this, then impregnate it with one of the hydrogen isotopes I have available. At the moment, it's just a big chunka copper.

Final target/beam stop

I cut the main feedthrough stalk to position this right flush with where that sidearm joins the cross.

Now, the movable target is, well, movable yet still wants to have a connection to the outside world. Insulated. Well, stranded teflon didn't seem like a good idea, fish beads not so great either, as the way this is laid out, the target connection and the feedthrough are very close when the target would be in the beam, but it would still need a long wire to allow the target to be pushed or pulled out of the beam - what's the extra wire going to do then, get in the way somehow? And it's required to be able to be "pushed" out of the beam, so I can get it hooked up in the first place - that needs around 3", but in position it's more like half an inch. After a lot of head scratching and searching, I finally found a real weak and stretchy spring. Not in my vast explicit spring collection either - but on a gasoline engine, the little spring they use to take up lost motion in the governor linkage. So, I had a kit to make into what I want:

Parts for the target feedthrough and connection

I took an 8 pin 5kv FT and shortened the inside vacuum wires to make the 8 pin ceramic octal tube socket be flush in the cross, soldered that on (yeah, that's a no no, but I hope it won't get hot enough to give me a significant lead vapor pressure), and bent the pins on the other side to fit into another ceramic 8p tube socket. I needed one wire for the target connection, and from experience, I knew I'll need some sort of shield for the insulators in this - there's always at least a little sputtering. So, making a virtue of necessity, I decided to fab a copper sputter shield on its own wire, to be a kind of offside faraday probe in the bargain. I just needed a hole for the spring wire/ceramic tube insulator to pass, and a rod welded on the back of the copper disc to hold it and connect it to the outside world. It could I suppose also be used to deflect a beam some - but also from experience, I know that the 5kv rating is pretty much the limit, and a 125kv beam isn't going to be moved much by 5kv over an inch or so...but I do have 6 more connections I could do something with later if I'm willing to take apart this ship-in-bottle to do that.

Assembled

Now its time for lunch and a beer...after which I'll try to get this somewhat assembled and hooked up - going to take some finesse to get that spring end under a screw while it's stretched and all tightened up. Then I can see if this will work OK and not short to something when the target is in position. Once I put in the foreline valve, and wire up a supply for it, I can at least test for leaks - I have a plug to put where the beam tube is going to go for that one. Then, ion source and focus tests, then the real thing. Baby steps, but those are sometimes the fastest way to get somewhere, as if you get them all right, there's no need to go back and check/fix each part - just move forward. Eyes on the prize!

[Doug Coulter]

That "final" in the title is probably optimistic...anyway, all I have to do now is tighten up the bolts, replace the fake beam tube with a plug for leak testing, and well, leak test. Got that ship in the bottle, with a mod to the spring, had to cut it shorter and then rotate the top flange to get it to wiggle right in there and not short to ground when the target was out in the beam area. The rest of the time, who cares...

View from where the window will be.

This took two beers to get together, and some very steady hand work with a couple little bent-wire tools I made to do it.

But first, you don't put a window near a big particle interaction zone and not protect it. So I cut some mica up to fit under the copper gasket, just inside the knife-edge. McMaster sells this as "india glass". I measured, used a compass, and fingernail (curved) scissors to cut it. This will get ruined most likely - better this cheap piece of mica than the nice quartz window...

Mica to protect the nice window I swiped from Bill.

For the picture, I put in a piece of pyrex to represent the beam tube (which will probably be made of that or quartz). I'll just test ion sources on this coupling first, probably that nice wire one Carl W posted about (and I got the papers for here), just to get going. Also I need to learn a bit about practical focusing, beam control, all that junk at hopefully nearly space-charge limits.
Nice that the coupler is easy to break down and put things on and off the system with. For leak testing, I turned a nice plug out of 6061 on the lathe to replace the beam tube. Baby steps again. I know this plug doesn't leak.

Mockup.

Now to get out the 12 point wrenches and sockets, and a beer, and start pumping. Once I get this leak tested (and presumably leak-free, whatever that takes), I'll add the foreline valve, power and control for that.

[Joe Jarski]

Doug, nice work! The beam system is looking good. BTW, I've had decent success with sputtering Ti lately - after getting the parameters right. If you need some targets made with a Ti thin film I can probably throw some together for you.

[Doug Coulter]

Of course. I've got some numbers, it wants to be fairly thick (half a mil or so) for my early tests. There's another scheme by which I might be able to load up that bit of sheet, by heating it in a D atmosphere and then letting it cool - might work, dunno. The books also mention Zr, which I've got and should be relatively easy to sputter. In the Ti case, the issue is prep of the Ti and doing it in an atmosphere really free of O and N and COx - so you'd have to ship it to me in argon or something, else it will refuse to accept any D (or that other isotope).

Right now, I'm fighting the usual leaks. I was much too lax shoving the thing together and evidently got a CF fitting gasket cocked - now that's not an easy mistake to make, but it seems I did; of course I'll know after I've fixed it for certain. Also, one of my pirani sensors leaked bigtime - the bulb wasn't sealed to the base well, but that one was easily fixed with a dab of Hysol 1-c.

Some vid of the leak hunting demo - the mark I eyeball is a good piece of test gear sometimes:
http://youtu.be/jIVyFwLku_I


I used a couple tricks with the bag, which I left on for the video here to show the "inverse trick" which I should explain. Since by the time I got the glow change, the bag was full of He, and removing the gas reverted to air color almost instantly, I know the leak *wasn't* in the stuff bagged near the top, anyway. Normally you'd just use the bag as part of divide and conquer to see if anything in it had the leak - He rises real nicely to keep a bag full of it and displaces the air out of the bag real quick. If I get 'round 'tuit I'll write this up better for the vacuum tech section, this set of tricks really works well. You could use a cheapo student optical spectrometer, or do things with a discharge tube in the foreline for real tiny leaks, but this is cheap, easy, and quick.

A huge key to the workability of this technique is limiting the pumping rate such that you can get a glow at all - this 60l/s turbo will, as JohnF says, suck this thing inside out, and will easily reach a nonconducting vacuum if set at full speed. I'd done that adjusting the night before for this demo. You could use a valve or whatever in your own system if you can't do what I did here.
I don't normally have that backing pump running all the time, but I had to set the turbo speed so slow for this, the cycling of that forepump made too much pressure excursion for this test.

[Doug Coulter]

Well, the divide and conquer strategy led me to just work on the pump station - looked like it was having trouble even without my new stuff. And it was, and now it's fixed. Turns out that a tiny bit of glass had been ingested by the turbo, with no damage to it, but the resulting powder played hob with the diaphragm/flap vale forepump. So, I pulled that (gheesh, they made that really hard, I think they built the whole station around it), cleaned up its innards, re positioned a gasket that had evidently been a little crooked the entire time, tightened up some hose fittings, fixed a leak in the gas ballast valve and now that's all better than it was when new. The pump is almost a zero clearance piston, shaped like the head, kind of cool, two stages. So, that was my issue, and it just happened to happen between taking off the old chamber and putting on the new - not what you would have guessed, but having a good troubleshooting plan means less guessing.
So, here it is, headless.

Headless station - lower cross prechecked, used for gages

While I had it that far apart, it seemed time to put in the foreline solenoid valve I know I'll want - a smaller copy of the bigger system downstairs. I got a referigerant valve from Burdin's Surplus center for about 7 bucks, rated 24v AC but I've also tested these on 6v dc - work fine either way. For simplicity, I just used AC this time, but I may change back as it operates kind of "too positive" with a big clack. The AC accomplishes this by drawing far more current when the slug isn't in the solenoid, so it has less inductance and draws more current till it has pulled in. I guess this will break it in fast. This valve is fully sealed re the outside, all brazed shut, so I know it's not going to leak to atmosphere. I made a couple tubing adapters to solder in on my lathe. It draws about 5W actuated, and can be run continuously. I had an old radio crap 24v .4a transformer laying around to run it. Basically, the idea is to run the transformer off the forepump AC power, so it gets switched when that does by the pump station logic, which allows for intermittent operation based on what the turbo pump is drawing. Very handy feature if you're saving your solar power, as well as runtime on that pump. So I pulled out the control panel and inserted the valve, the transformer, a switch and a pushbutton. The idea with the pushbutton is that you can run the forepump continuously, but then only take out gas in pulses. This allows fairly fine control of the system pressure. This one works a little better than the big one, actually, I'd guess because there's just a lot less volume in the foreline, so each press of the button removes less gas. In this mode, the turbo just acts like a fixed compression ratio, which you can separately set via setting it's rotation rate on the controller. For "high" pressures, like fusors, that's usually about 1/3 the nameplate, or about 1/9th the normal pumping compression and speed. It's a nice thing to be able to adjust to be sure.

Foreline valve hardware

After a quick test of this with the new stuff still not there, I put the new stuff back on and it seems I'd already fixed any leaks in that. But it's taking forever to outgas. The mass spec (the old Dycor) seems to be saying all I've got is water and H, but....who knows. I'll look in on it in the morning. Right now it's stuck at 1e-5, which is about where the headless one was for the first few hours too. By the morning, it was down to 7e-7 or so - a lot better. This has not much volume, but a heck of a lot of surface, so it might take awhile.

So, now I have a switch that will basically disable the flow in the foreline, and a pushbutton to override that switch when I'm in gas-control mode, and I probably have a basic setup with no leaks, which I'll know more about tomorrow. I ran in some He and ran a glow on it to heat the upper part up - I don't have a heater on that (and may not put one on there) and it seems the feedthroughs are all OK electrically.

Finished, and pumping back down

So, while it would have been nice to be here last week, I'm here now. Onward!