First Hands-free fusor runs

[Doug Coulter]

We have now made our first successful hands-free fusor runs. The data on these will be a little sparse compared to the standard I wish to set, but it's what I have.
I will edit and add data from the database soon.

Basic run conditions:
Main 1" x 2.5" grid, 8 .040" rods, in 6" sidearm, recessed.
Second grid in main tank, about 1" by ~3/8" four elements (two loops of W/Re wire, .010"), pear shaped
Pressure 1.6 e-2 mbar, indicated (probably high)
Gas flow rate during runs -- zero.

Power:
Main grid, Spellman Sl2KW, set at 51 kv, current limit at 12 ma
Second grid: Spellman PTV-200 10-20kv, current limited
uWave ion source -- on, low output in the .5 mA region.

Main power current usually 4-7 ma, variable.

Measurements:
3He tube counter on audio, not recorded these runs (software still coming along)
B10 tubes tested on one run, approx 9k cps during tests.
Silver and Indium in activation oven, silver measured at end of runs.
Faraday probe, far from either grid -250v at 100meg load, bouncing
PKR-251 pirani/ion gas gage
A peak at mass 6 appeared at about twice the noise floor on the mass spec appeared after the first run.


Results:
Friday night Aug 27 1560 cpm silver on the first 10 sec interval
Saturday morning, Aug 28 786 cpm silver, same measurement protocol
Delay to count -- logged but not yet reported (have to go pull it out of the DB) -- 20-30 seconds. Counter allowed to count one full 10 sec period at background
between shutting off HV and starting silver count.

Comments:
The second grid is used as an ion source for the main grid, primarily. It rarely reached voltages that make it produce neutrons at any rate well out of the noise.
Turning this off also shut down the main grid current draw. In other runs, varying this would vary main grid current over wide ranges.
Main grid was "blinking" on and off for both runs. No particular sync with the 3He audio was noted this time. Power seemed to be fluctuating with the blinking,
but only using LCD meters, hard to tell about sync.
The geiger counter counts in the 1100cpm region during runs, though it is behind a 2" thick lead brick, near the front of the fusor tank.
Backscatter? It comes right back to background of about 42 cpm when the fusor power is shut off, we then count silver for a few periods, then the indium if
the silver was pretty hot. We generally see on the order (not close order) of about 10x less activity in the indium doing this, which usually isn't well enough
above the background to really suit me.

The mass spectrometer has shown me peaks at 6 after runs before, usually too near the noise floor to get too excited about. One reason is that I have to
remove gas to make it usable, and usually go too far to see the best signal. I will have to tweak the post run pump technique to get into a higher
pressure range, nearer the top of the mass spec working range, to really see this well. I always have peaks at 1,2,3,4 after a run, from various H isotopes and molecules
like HD. The rest of the spectrum stays about the same before and after a run -- water and pals, some air, some CO2, and so on.

What enabled this to run hands free was a new gas flow system. We have added a small solenoid valve around the main manual valve in the forepump line, which
allows taking out small increments of gas via slapping a microswitch. Gas is initially let in to a higher pressure than desired (around a millibar), partly to flood the
tank and help the turbo spin down to the standby speed more quickly, then gas is removed in small pulses until a desired pressure is reached.
Good clean vacuum system maintenance allows this pressure to be pretty stable over hours-long periods. Gas inlet is via a tiny needle valve from McMaster
limited by a capillary tube. At some point, I twisted off the needle between the threads and the actual needle/o ring part, and it's actually much better that
way. Now the knob only controls whether the needle is pushed into the seat, or floating ever so slightly above it. It takes several minutes of this full open
with 10psig feeding in to reach the ~millibar tank pressure region. During these runs, the valve was closed off all the way, including for some time before
the runs, to allow the gas in the plumbing to reach the same pressure as the tank, so adjusting the final tank pressure by removing gas resulted in a stable
pressure in the tank.

After the Saturday run, we wanted to test some old B10 type tubes on Chad's new Ludlum electronics unit (we had to make some adapters and change the strange front
panel connector do to this). I had pre-drilled a 22" long 6.5" dia piece of HDPE to use as a moderator for them. We tested the thresholds using the supply voltage
recommended by Tyler (650v) and the threshold set at the Ludlum manual suggestions, which cut the counting rate on torbermite ore sample (about the hottest
and broadest band everything-emitter in my shop) to nil. We then placed the moderator/tube near the fusor, approximately 18" from center of grid to center
of tube, and slapped the HV back on -- instant neutrons, around 9k counts/second (not timed, just eyeballed) on all three tubes we tested.
This was about an hour or two after the hands free run had been made, and the gas pressure had stayed the same on the gage the entire time.

[Doug Coulter]

We are getting ready to tear the system down now to install what we hope are some improvements, so we did another couple of runs yesterday to make sure we had a baseline for comparison.
Sadly, some system upgrade seems to have broken the data aq software, which just locks up when you hit go at this point, so our activation counts can't be timed vs end of run, but we got there fairly quick with the silver. Run conditions were as close to the same as those above as I could make them (which is now fairly easy) and as the system had been on the pumps for a long time, we started from a base pressure of 1e-8 millibar (one of the improvements we're going to put on this go around is to remove some of the viton gaskets and a small leak from the system in a spark-plug feedthrough we don't use anyway).

On strictly timed 5 min runs (no cheating here) we got into the 1100-1200 count range on silver both times, both neutron detectors sang happy songs throughout on our stereo audio monitor, and it just worked -- I'm beginning to really like that.

Process:

1. Set turbo to standby.
2. Close main valve to forepump, but turn forepump from auto to always on.
3. Let in gas till we see perhaps 2-4 e-2 mbar in there -- this helps the turbo slow down quicker to the standby setting.
4. Turn on 2nd grid power and let it current limit while waiting, the theory being we are making some UV and blowing some junk off the tank walls by so doing. At that point, it's limiting to about 5kv or so.
5. When the turbo stabilizes at the lower speed (by which time the gas pressure has risen due to the lower compression ratio), use the solenoid valve to the forepump in pulses (we do this manually) to get into the range of 1.6 to 1.9 e-2 millibar. After that we can turn the forepump back off.
6. Set up the main power supply to 52kv and 12 ma current limit.
7. Start the run and the timer.
8. After 5 minutes, stop the power, get the silver and put on the geiger, write the number down.

Between runs, simply pump down at full speed, while waiting for the silver to get back to background levels.

We actually got the higher of our two activations at the lower pressure -- 1.6 e-2 mbar, and about 9 ma current on the main supply, with the 2nd grid current limiting at about 25 kv, not terribly stable. For whatever reason, with the second grid alone doing that, and current limiting at about 12 ma, our neutron counters don't say much about it. The grid is a pretty lousy one, just a couple loops of type C thermocouple wire, small, to act as a good ion source out in the larger tank where Paschen's law allows it to draw current at lower voltages.

In fact, one the the things that amazed the new lab assistant (because it's a little counter-intuitive) is that the little grid will light off and draw current just fine at 20kv or so out in the big tank, where the main grid in the 6" sidearm, alone, won't light off even with 53 kv -- you can turn the main grid on and off via the second one, it's a pretty good switch, actually.

We did have the microwave ion source on too, but one of the reasons for the teardown it that it has quit making ions into the tank, I suspect the wire to the extractor electrode has broken, it lights up fine still. No way to find out without going in there. I'm going to modify the gas inlet system that used to flow the inlet gas through it -- don't need that anymore and it makes troubles as the gas in the supply tube can light off as it gets down to system pressure in there -- as in the previous post, during the run, all the valves are shut, so the supply line eventually reaches system pressure, and even a foot of pyrex tubing in there isn't enough -- nice "neon sign" with D, but that drags down the push electrode voltage.

So, once we finish making our "kit", we're going to open up the thing, move quite a few flanges around for new work, replace both grids with new designs, move the wiggle stick to a better place re distorting the E field where it is, and some other things. We are getting ready to try this with much lower pressures and higher main voltages, so the big Spellman will now run the second grid, and the really big Wallis-parts supply I'm now finalizing will be the main grid power. We plan to go 80-100kv for starters, but we could have up to 180kv available.
That's going to take new feedthroughs for the HV and some other big changes, and I have some good type N RF feedhthoughs to weld into the main tank to get ready for some RF work in there.
Whether the scheme mentioned here will get to be tried on this rev or not, I don't know, but I do know that without some of this, I can't try it, so, baby steps, but steps.

Although some of the discussion should be on another thread, I'd have to start it I guess, and no one else is reporting results yet anyway, so here goes with some general observations.
Grid accuracy matters. When I had 1/8" errors, I had a 1/8" focus. Now that the errors are .03", I have a focus that size. Neutron output seems to go up faster than linear, and we are going to shoot for .001" errors next time. Whether exact grid size vs tank size matters a lot, or number of elements and their spacing is crucial, we just don't know yet, but small changes (25%) in those don't seem to make much difference we can see. I think there are two reasons we don't see much out of our second ionizer grid -- one, it's really bad as a lens, and two, it's a longer distance from all the detectors, and running at much lower voltages. The visual focus really stinks on that one -- that wasn't what I was going for at the time.

So far, as we go down in gas pressure and up in voltage (limited by how well our secondary ion sources can work at low pressures) neutron output and Q is going up much faster than linear, and going up parabolic, no slowing or peak in sight at the levels we've been able to run till this point. I can guess some of the "why" but I'll leave that for another day. I will just say, it's true and certain that this is happening here.