Fusor Grid making

Tricks of the trade

Fusor Grid making

Postby Doug Coulter » Sat Feb 22, 2014 5:21 pm

Well, I'd recently tried using Ti for the grid ends - one carved from 1" dia rod, one made of end-welded 1/8" wire.
It worked, but Ti is a lousy conductor of heat, and the ends got hotter than I'd like. Also it was evident I'd made that one too long.

So, despite some concerns about hydrocarbon contamination (which I think I have a fix for later), I decided to do it all once again. Since how I did this isn't totally obvious to anyone but a pro machinist, here's what I did, in what order.

Starting with 1" graphite rod stock chucked into the lathe, I made the HV stalk-end first. This involved turning down the last .4" to .375 - I'm going to put a Ti washer to cover most of the back (feedthrough facing) end to reduce carbon sputtering onto the boron nitride I'm now using as an HV insulator right up to the grid (which, BTW, solved all my FT problems - it's great, if expensive). I then cut it off with a cutoff tool, so it was 140 mils thick where the rods are going to go in.

After re-facing the graphite stock end, I ran a 3/4" Forstner bit into it so as to create a 1/8" cross section ring - in fact, I went a little deep, and I can get another ring out of this piece later on.

People who've been paying attention to my grid designs will notice that this time I made the HV end a female 10-32 thread instead of trying to make a threaded 10-32 graphite end. I might try that later, but this stuff is on the brittle side, and I can simply use a good heat conductor metal threaded rod to mount to the female 10-32 on the end of my HV stalk.
This just looked like a better way - This kind of thing is a one mistake and all your work is lost thing, so you get conservative in design.

Then I applied a degree wheel to the lathe chuck, using magnets (like auto guys use for setting cam timing) and have a pointer mounted on the lathe (just copper wire). I then mounted up my homebrew toolpost grinder, which accepts 1/8th inch stuff - diamond wheels for cutting glass and so on, or solid carbide tiny drill bits - they have a 1/8th shank, and in this case are then necked down to .020 inch. Using the degree wheel to get exact (well, a tiny fraction of a degree) 45 degree
increments, and the tool slide to control drill depth, I drilled 8 holes in the ring, 100 mils deep each, and then cut it off the main chunk of stock.

I then repeated the process, going a bit deeper - 125 mils - for the HV end piece. Since I had turned down part of it to 3/8", I had a nice way to chuck it up, and got the exact same drill circle, since no other part of the setup had to change at all.

I did a bit of bevelling and sanding along the way, some while in the lathe, and some by hand - still needs a little more, as I don't want sharp edges - but I also don't want any of the holes to be "all the way through" - so I'm cutting a fine line here.
Here's what it looks like as a test-fit.
GridTestFit.JPG
Just test fitting - I have some more sanding and fitting to do, as well as punch out the Ti washer for the back end.


Then I had to cut (at least) 8 rods for this design. I wanted them 1.75" long, which will give me about 1.525" of active length for the rods. You really want this to have lens-like accuracy, though I'm not even to good optical, much less ion-wavelenth precision here - but you never know - it might be perfect for some part of its length. Since I've not yet adapted my dremel drill press for the new tool (the old one literally let out the smoke), I made this jig for a newer dremel (thanks BillF) that's a lot nicer motor and in better shape anyway.
RodDremelJig.JPG
Pretty simple *after* you figure it out.

These rods come very straight - the far more expensive wire from Alpha-Aesar, say, don't - and these are pretty cheap at the welding supply store. I'd used the "E3" type in the other grid, but that stuff (doped with various rare earth oxides) was just too good at emitting electrons, especially after being hit by a bunch of hot D (reduces them). So this time we're going with tried and true pure tungsten, 20 mils diameter. It tends to get pretty hot in the middle - even with a cylinder shape, and so far, almost no matter the length as long as it's a bit longer than it is in diameter, the poisser concentrates in the center of the length anyway - and that's where the wires get hot. The ends get hot too - because they are hanging out there, big, and at full volts, so they attract ion hits. At least with carbon, it's not a great secondary electron emitter, but it seems I do get some hydrocarbons from the sputtered off carbon atoms interacting with the hot D ions. I may vacuum coat this with titanium or some other high work function metal - I now don't care if it gets orange or yellow hot, since the graphite is what's holding the forces anyway, and it's quite happy well past white hot (at which point I'm more worried about my copper feedthrough stalk...). With all titanium, it got orange hot - at which point Ti has the strength of butter...this may get as hot (I'd bet it doesn't, since if nothing else, carbon radiates heat better than any metal as well as being thermally very conductive), but it won't be at risk of slumping, which I've had happen to other grids.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Fusor Grid making

Postby Doug Coulter » Sat Feb 22, 2014 5:27 pm

Oh, since I can't stake in the rods, as I did with the all-Ti model (which didn't help accuracy), I plan to try the old John Strong formula for high temperature glue - sodium silicate mixed with talc. After a few hours, this sets rock-hard, and after a few days, it's also dry (the water has to evaporate separately, it's a chemical reacton that makes this stuff set). This one came out a nice tight fit anyway, but there are a few grams of E field force trying to take it apart in use - and a lot more if there's an arc, which there usually is during break in. In other graphite-end designs, I'd been wrapping wire around the circumference at both ends to put enough tension on the rods to hold it together - looked really ugly, and since I had to use Ta wire for the temps involved, not cheap either. So we'll try this glue this time, even though it's going to play hob with my vacuum system base pressure till it's really dry - which might take a week or two under vacuum *after* I bake it in a toaster oven for an hour in air. We'll know soon!

I'm hoping this design will take much of what the new 100kv power supply will dish out, it's in the process of being commisioned as a separate project (no point running it till I have all the data aq and safety stuff in place and tested).
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Fusor Grid making

Postby Doug Coulter » Mon Feb 24, 2014 2:02 pm

I slathered two microscope slides, one with pure sodium silicate, the other with the previously baked out talc from the grocery store, mixed thick. I let them dry, then left them by the wood heater overnight at around 120f, before baking them in the toaster oven just now. Pictures:
BeforeHeat.JPG
Before heating - you can barely see the pure stuff on the slide - almost the same index as the glass and very clear.

After475f.JPG
The pure sodium silicate turns white and expands over 3::1, while the talc mix is almost unchanged.


So, even though it's kind of crunchy/bubbly, the pure SS probably wouldn't do here - it might break the graphite. I in fact should have tried this on a graphite sample anyway, since I can heat that hotter, and "who knows" what other chemical reactions might have happened with graphite or tungsten? But this is making the talc mix look better - I think it expanded a little, but I'll have to check closer and try higher heat as well. The pure SS had done its thing by the time the toaster oven said 250f. I didn't notice a change in either sample after going to 475f for half an hour.

Now, here's the real test - orange heat with a mapp torch, enough to make the glass sag under its own weight. No damage noticed to the talc mix. I didn't try the ss pure one, no point, the bubbling (sorry I couldn't capture that in a pic) would have pushed the grid apart before it could melt or whatever it would do at "orange".
AfterOrangeHot.JPG
After as hot as I could get it on glass...the talc mix, no damage noticed.

Not too surprised about it, but it looks like John Strong was correct - again.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Fusor Grid making

Postby Doug Coulter » Mon Feb 24, 2014 4:43 pm

Just for fun, our oldest record-setter and the new grid next to one another. The old one has the fat wires, and they're at an angle, a machining "mistake". The "twisted" grid like to be a bit further out into the main tank, but does not throw a concentrated beam of "something that can make SS incandescent and glass crack" like the straight ones do. Perhaps I should consider making that mistake - or even sweeping that parameter space - again.
OldVsNew.JPG


Note - the old one was shiny-smooth when first put in. It's rough now - carbon sputters a little.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Fusor Grid making

Postby Doug Coulter » Thu Feb 27, 2014 1:50 pm

Yup, it works, a little better than expected in fact, easily broke and held over 2m neuts/second at reasonable input power, though it would go higher and take all I could put in (which is quite a bit, I stopped at 50.2kv@25ma). And that's despite a wrong placement. We've found that placing the grid-end more or less flush with where the sidearm meets the main tank is far better than the 1/2 or more inch recess I have here due to having put in a spacer for the older, longer grid. Here's some eye candy.
NewGridEnd.JPG
Seen from the end. Not at max anything, was running at a fairly high pressure to get enough light to take a pic in daytime.

And, from the top. Kind of hard to see here, but the poisser does tend to concentrate around the middle of the length. You can barely see the rays here, I've done better shots of this effect, but this is what I got today.
NewGridTop.JPG
Top view via mirror and glass on the top of the main tank.


And a tiny bit of data. The green counter is 980 cpm/million neuts second.
NewGrid2.jpg
A litte data on the second run. First one was just to make it hot so I could pump out contamination.


Now the system is in pumpdown for a few days most likely. I may re-space this to the place we know seems to be best, or not (it leaks a bit of shop air in when I let the tubing coupler move), or just try the next step (deliberate pops mode) as is, since we don't know what will work best for that. Satisfyingly, it went down to 1e-6 millibar almost instantly when I turned on the real pumping after these check-in runs.
But I like to see e-8 and below before I trust things are pure. That will probably take a couple of runs.

I really like the way this handled heat - even with well over a killowatt in, the base got to dull red at most, and frankly I couldn't see the rods glow at all. Very minimal break-in arcing, just a couple sparkles and it was done. Oh, and no glue used at all - it was such a tight fit when I test-fit it I was afraid to damage it pulling it back apart - I wanted precision, and from the focus on the beams I saw just now - I got it. No bent rods.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Fusor Grid making

Postby Doug Coulter » Sat Mar 08, 2014 11:47 am

Here's the grid the new one replaced. It was pretty good, but was too long (wastes power) and got too hot (Ti isn't the greatest heat conductor). It got orange hot all too easily, and I was afraid the end ring (butt welded Ti wire) would slump. This one used E3 rods, and it just emitted too many electrons as well. Looks purty, but...that's not all there is to this. Staking in the rods spoiled the accuracy a little as well.
TiGridE3Rods.JPG
Lots of work to make, but a fail by comparison.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA


Return to Shop tips and kinks

Who is online

Users browsing this forum: No registered users and 3 guests