Protect your glass!

[Doug Coulter]

Well, I knew this and always have, at least some. For example, protecting the main glass of my rather expensive hinged viewport, is a clipped in piece of 3/8" thick pyrex on the "fusion" side. This piece of sacrificial glass has turned brown, and now has circular tiny cracks all around - almost like a bullseye.
However, when running in a new, more precise grid this afternoon, I found another issue for which that wouldn't have been enough. It's pure luck I had added a stainless steel, fine-mesh screen to the very inside, as I was playing with RF injection and didn't want RF in the shack. So, in order from the "action", I have a grounded screen, then a sacrificial piece of glass, the viewport glass, then a 1/2" thick piece of leaded glass to look through, outside the tank, to protect the operator from the X rays.

Well, this new, more precise grid had another side effect - the beam out of the end (I do cylinders here) stayed focused and collimated all the way to the view-port, as you can see here:

http://www.youtube.com/watch?v=vkWyikizEbg

Or here, with more words and pictures: http://www.coultersmithing.com/forums/v ... f=25&t=678

Clearly, the inner pyrex would have shattered - it's not proof against anything, you know, just better than plain glass. That could have been one very expensive oops if the shards had gotten into the turbo. Had this been a bell jar - it would have gone bye-bye in an implosion pretty quick.

[Chris Bradley]

Doug Coulter wrote:
> Well, this new, more precise grid had another side effect - the beam out of the end
Why do you think this spot effect is because of the new grid, rather than because of the new grounded mesh screen?

That effect, where you get ivory coloured electron discharges across the glass, is very similar in appearance to effects I've seen in my plastic parts that have electrode structures embedded in them carrying HV. I think it is, 'simply', where the electrical potential of the charge across it is sufficiently high that it is no longer acting as an insulator.

There seems to be an electron path escaping in the '8 o'clock' position - has that reached your outer viewport, Doug?

[Doug Coulter]

The grounded mesh has been there for many runs and a couple other grids. This is the first time I've seen this, though. I suspect something similar was happening, just not as focused as this.

I've seen electrons hitting the glass (I assume that's what they are, due to the serious excess of electrons in the plasma, and their speed, and how it looks) before, lots of times, and using a magnet, made them twirl around easily - with only a weak (100 gauss) magnet that wouldn't have the same effect on heavier ions. There's a movie on my youtube channel of me doing that, actually, way in the past. Unlike X rays, which can also make the glass light up blue - these moved with the magnet, easily, and a long distance. The main odd thing this time is that they moved around quite a lot with me doing nothing on purpose at all - could just be break-in effects, we'll see if that behavior keeps up during "money runs" later.

Looks like I should put my pinhole camera back in there with some electromagnets to really determine what is going on here - electrons should move in opposite directions from positively charged ions, and more deflection, while X rays shouldn't move at all when the electromagnet is turned on.

So, I'm in that happy place for an experimentalist to be - more stuff to measure! Of course, we might also have some negatively charged ions here - other things I've tested indicate there are a fair amount of them, that make fusion when they hit D embedded in the walls.

I didn't notice anything very special at 8 o clock, but I'll look next run. The sacrificial glass is about half an inch smaller in diameter than the window/port, and has metal clips to hold it in at 12, 8, 4 o'clock.

I burned out a nice meter during this run. It was rated at 2kv, but I had it hooked across the feedthrough shield and it arced internally. I normally was getting only about 1400v with the other shield back there. In all cases, anything in there that looks like a faraday probe gets a large negative charge on it, sometimes even under a fairly stiff load to ground, which is why I assert we have a non-neutral plasma here. The proposed cause of that is secondary electrons from the grid wire itself, every ion that hits it knocks out quite a few, and they accelerate away from the grid. This is one reason I used carbon for the end cap - fewer secondaries there, and better heat radiation. Sadly, graphite sputters some in D, so I didn't want to use it at the feedthrough end, as it shortens the life of the glass when it gets covered with some conductor I don't control the potential of (and that's why I have a wire on that shield, to look at that, control it, maybe even drive some current into).

The real question is does this more-accurate grid give me higher Q, but with only the audible neutron detectors running, I couldn't say - that will have to wait till I do some computer repairs. Subjectively, it sounded really good, and it looked really good - but not hugely or obviously better than the twist grid.

[kcdodd]

If you could isolate the mesh and measure any current to or from it, that tell you what's hitting it. My first guess would be electrons. Ions hitting it doesn't make sense to me, since they should not have any energy there. The less likely possibility would be neutrals from charge exchange at the grid. But that would be a really cool result if it was.

[Doug Coulter]

Well, I could indeed isolate that screen, but all I'd find out is the same thing I find with the faraday probe next to it - the *net* of what's hitting it, not the individual species. The pinhole camera would be better as you can separate by charge polarity nearly all the possibilities. Though if it's heating up SS screen that bad - the camera will not last more than a couple seconds in that beam. I might have to make one much less sensitive, and much more energy resistant. Hmmm. I could put a wire on the wiggle stick and move it around in there.

Yup, there could be charge-exchange, and coating my tank with D loaded Ti proves that something at least hits the tank walls hard enough to generate fusion there. Charge exchange is really inefficient, at least in tandems (I believe), but a fusor is so bad that even a little help from there doubles my output (around 2 mill neuts/second at 40kv/20ma with the D in the walls, about 1 mil when things heat up and the D is gone from them). In other words, about the same as Hull's fusor without the D in the walls, twice that with it. (this was originally his idea, FWIW, but not for the reason he thought, I think).

I've proved (to myself) before that there is indeed quite a lot of something coming out of that hole in the end of my "cylinder" grid - I actually put a ceramic pipe on the end of one, and inches away, stuff is streaming out, almost like a fluid under pressure. This grid just just much more accurate...

Has it yet occurred to anyone that their "sphere" grids are actually *less* spherical than my "cylinder" grids? At least I have one end open...they don't, that's where all the wire loops cross. We're both stuck with a feed through on the other end. At least I have even spacing for a uniform focal point.

[kcdodd]

So does the Faraday probe say it's mostly positive stuff, or negative stuff?

[Doug Coulter]

Probes say hugely negative. I get around 1500v across 10 meg ohms on either of my probes. One is right beside the main window, the other is behind the grid, they both are about 3" long and give about the same results, other than the timing of fast events. It seems to take about 10us for some disturbance to get across the tank (15" or so) from the fusor area, FWIW.

They act like diodes. If you hook say, a neon sign xfrmr across either one, when it tries to drive it positive, it can't do it - more than a few hundred volts, and during that half cycle, fusion is suppressed.
During the negative half cycle on either probe - fusion takes place in little bursts with a rep rate of about 2 khz or thereabouts, and you see the probe swing positive (against the will of the NST) about 2kv a little after each burst of fusion (about 10us). I'm supposing this is + charges flying by and taking the probe up via induction. If you hook a capacitor in series with the NST feed, it charges up just like it would if there was a diode to ground on the faraday probe.

But all that is taking an average of what is more complex than that technique can show you - the emergent behavior of a "simple" system is actually quite complex when it's not at equilibrium or even when it is, as I've been measuring all along.

For example, doing nothing more than hooking a good Q inductor in series with my main ballast such that with the stray capacity it resonates at 1.8 mhz, is what gets me that unstable super Q mode - around 500 times the Q I get when things are all nice and stable. It's hard, but not impossible to reproduce - in that mode I get only about another double of neutron output, but with far less power input. But it is very hard on feedthroughs, as the inductor kick back can get to really high voltages before the fusor re-strikes - which is surely at least part of the higher Q I see there.

[kcdodd]

If there is lots of ionization right in the center, perhaps this could be used as a nice electron gun without a thermionic cathode? The ring shape on the open end would create beneficial focusing as electrons get pulled out. Ions get pulled back and recombine on the grid. I think you would get more current at lower volts with Argon, with its bigger ionization cross section, at similar pressure of course, if it really is making most of the electrons by ionization in the middle, and shooting them at your screen. If you get similar I-V, then they're from the grid surface no big woop. What do you think?