Runs after new HV FT - some interesting!

[Doug Coulter]

I recently tore down the fusor to repair the HV FT, and put in a new design twist on it. Seemed to have REALLY improved things...wish I'd thought of this change earlier. I didn't have all my test gear and logging going on for this, so no really good hard numbers - I thought I was just breaking it back in (HV conditioning) after the teardown. But I hit a fantastic sweet spot after a little breaking in. Here's the breaking in part:
ww.youtube.com/watch?v=_b0OL1SHAY0



Then, a little fooling around at much lower than normal pressure (perhaps possible due to some contamination, who knows?) I hit the mother lode - something like 10m neutrons per second (on my log scale meter, same detector we calibrated up at HEAS) and at only 41kv and a few ma input! Check this out! I'm going to be trying to replicate this, you betcha.
http://youtu.be/cPJIcAzV0Rs

Indicated pressure was the lowest I've ever gotten this to light off well at - 1.4e-2 mbar (gage reads factor two high on D, so the real pressure, sans impurities, was half that).
Zowie - this is almost as good as my flakey pulsed high Q mode, which only drew about 100 microamps (but only put out 1/5 the neutrons at best too). I wouldn't even need all the fans to run like this if I can do it every time!

Here's some more footage where I show a view from the mirror on top of the thing. This camera works better in this light than my own eyes - I couldn't see diddly, but it shows fine on camera. I like that! At the low pressures, and high outputs, the fusor isn't as photogenic as when I up the gas and the current to show the rays better. Here it's just a line at the focus.
http://youtu.be/kfgJR92nWUg




This one shows the synergy you get using one tiny grid in the larger space to take advantage of Paschen's law to light it off with lower gas and lower voltage, so the main grid can run in more standard conditions. Neat demo of the effect here:
http://youtu.be/Qm5kzipMrmI

See this link for more on the law. Paradoxically, electricity going through a gas does not necessarily take the shortest path! To get an avalance of ionization, you need pressure times distance, which means that in some conditions, it tends to take a longer path and if no such path is available, no current will flow. I'm taking advantage of this in two ways in this demo - using the small grid in the larger tank gives me distance to get gas ionized at a lower pressure and voltage than I can in the place where the main grid lives is one. The other is that there is a grounded copper pipe right near the main grid HV input which draws NO current because the path is too short for the voltage and pressure there - cool, huh?

[chrismb]

Doug, all good work, as your norm, but may I suggest also to run a control gas. I do not dispute neutrons but there did seem to be a counter/flash co-incidence. I would suggest the more 'scientific' protocol would be to start running with a control gas to burn off the thick of the sparklers, until your count rate drops down to background whilst running at the same drive levels you'll run after the fresh D goes in. Then you have the hard data to discount claims of RF interference.

[Doug Coulter]

I'll just run some silver or indium. I now have no personal doubt that all three of my neutron detectors (using three different technologies and detecting neutrons of two different speeds!) are detecting neutrons only (and cosmic rays, but that floor is in the count/minute range, negligible compared to this). It's been a very long time since they weren't backed up by that on every test where both were used. Of course there is a correlation with flashing, that's also when it's drawing peak current. Did you look at the rather extreme lengths I've taken to get rid of EMI at the sources? And at how the sensor electronics are themselves handled? The system is quite tight now, after a couple years of working on that.

I've seen pulsing cause much better performance before, this is not a surprise, and a fairly simple batch of theories could explain that - in fact, the embarrassment is that there's more than one that could explain it and I don't yet know which if any is the right one. But it's pretty obvious that the conditions right around the onset of the first pulse are very different from those in stasis - the soup of stasis (dynamic equilibrium) isn't there, for one of the larger considerations. There isn't already a bunch of junk inside the grid...and so forth.

The problem is, there is no gas that really acts the same as deuterium, hydrogen would be "close" but not all that close, having half the mass...everything else has other issues including different ionization energies and levels and you name it - they surely would not produce the same emi under the same conditions. And of course, there's no proof that hydrogen wouldn't make a few neutrons of its own along with the other materials in the tank. Probably not many, but there's many reactions where a proton knocks something out of some other atom.

What I'm more interested in is why I got this super hot/high Q mode and was even able to run at a far lower indicated pressure than I've been able to do in the past. I strongly suspect that this is due to whatever contamination got onto things when the door was open, but sadly, it's quite difficult for me to take a mass spectra that means a whole lot. I have to first pump down to where the mass spec will work (factor of a thousand), make some assumptions about the errors due to that -- pump preferentially pumps some things better than others, sometimes the source is still producing some gas - which may not reflect what was really there before I did the further pumping, and so forth. Errors in the resulting spectrum well over 50% wouldn't surprise me a bit. And for the nonce, I can't separately pump that thing like they normally do for this situation.

For reference, the lowest pressure I've been able to run stably is 1.9e-2 mbar, with it becoming very unstable at 1.6 or so (indicated, PKR 251 reads 2x high on H). The hot run was at 1-1.4e-2 mbar indicated, flickering into the next lower decade. Normally nothing will even light off at all at that pressure, though you do generate some low energy X rays and phosphor inside the tank lights up. So something changed, but then went away with subsequent pumping down and refilling, when things returned to more or less my normal baseline, maybe a little better. Hard to find any other logic to explain the data -- I suppose out at several sigma, there was some surface effect/change on the grid or tank walls that just happened to coincide with the moment I did that. Like I said, pretty low probability on that one. It pretty decisively changed when the only thing I did was turn it off, pump out most of the gas (down to about e-5) and refilled, then turned it back on.

But I'm fully confident in the detectors at this point. They all even showed the normal count ratio between them - EMI wouldn't do that, or it never did in the past. Heck, even the rock solid geiger counter tracked the usual ratio. But as I said in one of the videos, yeah, of course the thing to do is run silver again on the standard counter - either way - if I get my "normal mode" back or this one.

[Doug Coulter]

Matching action to words, I just went down to the lab and made another run. Looks like my vacuum gage got flakey - but a reset seems to have fixed that...
At any rate, here's the results:

Std counter output screen.

You can see where I killed fusor power at about 06:17 or so, then put the activated silver on the counter. I scrolled the text log to that point in time so you could see that.
I let the silver decay until about 500 seconds (or slightly after). I let the counter count background for 10-12 seconds, then put on the indium, which isn't activated a ton out of my lab background today, which you can see at the end - after 700 seconds or thereabouts - but it was definitely above background a little and decays slower than the silver. Extrapolating the straight line silver decay (on the log plot) back to the instant of fusor turnoff gives us something like 1200 cpm at that instant through the glory of data averaging -- even at 1k cpm, the data is a little noisy due to randomness of silver decay.

Before that, you can see a place in the run where the neutron counter was reading very high for a little. That was me pulling out gas trying to get that mode again, and may account for the slightly higher than my average silver activation, though this time (despite pulsing and presumed some emi) the geiger counts didn't rise with it. This is not what I saw the other day at all, at that time everything went together, and this time the view into the tank didn't look the same either - looks like I *didn't* replicate "the mode" in actuality. Now I will have to try again, perhaps after letting in some air...which had to be the main contaminant last time. This time, it'd been sitting at e-8 mbar for a couple of days, so it had to be pretty clean in there. That's my normal situation, I leave the pump running all the time, as with the forepump only coming on as needed, it only averages about 30 watts out of my mains - my solar system can handle that pretty well, and it seems worth it for repeatability (except perhaps in this case - it would explain why I hadn't seen that before, though).

The neutron detector used in this plot is the one we calibrated up at Richard Hull's, the Hornyak detector that sees high-energy neutrons only. I have it propped up in a fixed position on some of the side ports on that side of the tank - it's about 1.5" from the outside shell of the fusor part. It seemed to be working just fine, and to agree with the 3He and B10 tubes I had on the stereo monitor. Obviously we need a many-input counter hardware device...someday.

You can also tell on this plot that I had it go "out" a few times - that's shown by the geiger counts having red bars to zero. I was playing with the gas in and out and overshot a time or a few.
Nice to have this going to help operate the things, real good feedback - remember that since these plots autoscale in both X and Y in realtime, during the beginning of the run you see a lot more detail of what little data you have so far.

I did see some neutrons from the secondary grid this time (and last time). I'm guessing that with all the other noises pretty well taken care of, I can now see a little lower neutron count rate from that (it's still pretty near the floor, but definitely there).

So - I refute EMI thus -- see above. That's real neutrons most certainly, and this verifies that all the measuring gear I've so painstakingly built and refined over the last couple years works just fine - still, again, every time.

[JonathanH13]

Doug,

The silver activation numbers are proof enough for me. And I think I understand what you are saying about Paschen's law allowing you to bring the grounded copper pipe so close to the grid, but I'm trying to visualise the electric field in this area. Surely the field intensity must be higher? Or are you suggesting that the field 'arcs' out in loop, like a solar prominence?

I also cannot help but notice some extra capacitance here (a few pF?), created by the surface of the copper tube, with respect to the inner HV conductor...

I think I'm going to find me some 1" copper pipe

[Doug Coulter]

Yes, though the field is more intense there, paschen's law isn't satisfied for a discharge at the pressures we run. There's about an inch, maybe a little less from the end of the pipe to the sold back of the grid, which is also an inch diameter. I'd guess the field lines are kind of straight...hmmm, not really, as there's the uninsulated feedthrough rod inside the pipe before it even comes out of there. At any rate, there's no funny activity, warping of the normal rays, or anything else I notice different, except it all seems to draw a little less current for the same output - not dramatically less, but some.

All I did other than that was replace the quartz, nothing else. I saw a "bad spot" on the end of the pyrex where it had been arcing before, and did nothing about that. Looks like I should have though, as next run set (trying to replicate that magic mode) I got things hot again and trouble started there again. So down we go again to cut off a bit of that pyrex and try again. I'll report on results. I think that issue was left over from last time and that cutting off the bad pyrex will finally put it to bed for good (I'll turn the quartz end for end just in case). Hard to see how ions would bombard the glass/quartz the way it is now, you'd think they'd be attracted out toward the grid or the metal stalk - and repelled by the copper pipe if coming from outside. I didn't want to muck with taking out or replacing the pryex this time as I'd put HV dope on the joint with the metal in a leak search (that turned out to be something else, but taking that off would have required some extra hours of re-cleaning stuff).

I tried to replicate that mode by letting in some air, a few mbar worth, even ran in some power with it like that (drive air ions into things??) then took it to normal pressure (1.9e-2 mbar indicated), let in a little D, ran, adjusted pressure down to the limit, ran, let in some more D, ran again - repeat 6-7 times. By the last time I should have been in pretty pure D again, but at no time did I get the choppy high output mode. I could get it unstable at the lower pressures, but nothing special happened.

I did find out something interesting about the meter on that hornyak. Electrostatic fields can make it read wrong! I put screen wire over it. Oops.