Comparison of standard grid vs pointed electrodes

[Bob Reite]

I figured that I was wasting too much energy heating the gird, so I came up with an idea for throwing away the grid and using two pointed electrodes to generate a plasma discharge between them. As it turns out, someone before me already tried this. Frank Sanns patent # 7,550,741. I did discover that it was not necessary to have a potential difference between the two electrodes. I got the same tying both to the 40 KV supply.

By the time I got around to trying this, my bubble detector had worn out. I bought a REM ball from Ebay. The previous user had already determined the voltage setting and had noted it on a label attached to the REM ball, but I had to use the calibration factor in the manual. So measurements made with the REM ball may not agree with my earlier runs with the bubble detector, but I could still compare the standard grid vs pointed electrodes by using the REM ball for both measurements. The center of the REM ball is 26.67 cm from the center axis of the chamber.

Best run with "standard grid" using REM ball.

Code: Select all

Eg 40 KV
Ig 7 mA
Pressure 15.0 microns
Flow rate 4.0 sccm
4.8 mREM/h
calculated isotropic neutron emission 334,000 n/s

I have no idea why I could not make 500,000 n/s unless the calibration of the REM ball is off. But in any case, I can still compare.

Picture of pointed electrodes in operation.
Best run with pointed electrodes.

Code: Select all

E in 40 KV
I in 5 ma (the pointed electrodes did not want to draw as much current as the grid)
Pressure 20 microns
Flow Rate 5.0 sccm
5 mREM/h
calculated isotropic neutron emission 348,000 n/s isotropic.

As can be seen there was no real improvement, although one could argue that I'm getting the same amount of neutrons for less input power, but in any case I will try something else.

[Doug Coulter]

Some regular-light pix of your gear when it's not running would make this a lot easier to understand.

We've run with pointed-end wires a bit when a grid end-cap blew off. It seemed to work well enough, and we seemed to be able to go a little lower in pressure, as field emission from the points in our peculiar tank layout seems to help with ionizing the D. That might not apply to this if I understand what you're saying, though.

[Bob Reite]

Here is a visible light picture of the chamber interior taken through the viewport with the system idle. The only camera I have that gives a wide enough angle to see anything is the CCD camera used to monitor the system while operating, so it's NTSC TV, but good enough. I added black arrows to highlight the two pointed electrodes. The bottom one melted, actually the copper connection melted. If I want to pursue this approach further, I would need to get another feedthrough like the top feedthrough who's center conductor is stainless steel. The top electrode seemed to hold up well enough. Indeed, the field emission from the sharp points does help to ionize the gas and establish a vertical channel of hot plasma between the two points.

[Doug Coulter]

For that approach you might consider using tungsten TIG rods - cheap. They make a version without the the thorium for use in aluminum welding, which is recommended (less secondary emission and you don't get radioactive dust in your lungs grinding a tip). I've done mounting by just drilling a hole in the end of whatever and poking a rod in - you can hold it with a setscrew or just pinch the holder so it's tight in there.

The advantage is that tungsten not only won't melt - you can get a sharper tip, and it is a lousy conductor of heat, so your mount won't melt unless it's heating on it's own from ion bombardment.
40 mil rods at any welding supply store, but ask them if they can special order 20 mil ones. That's what I'm using for grid rods. About 2x the cost of the cheap guys, but they come really straight (well, most of them).

A neat trick for making super sharp (maybe even one atom) tips.
Make up a lye solution (strong but not critical) in a beaker. Use a strip of lead or something like it as the cathode electrode (iron would probably be fine but I haven't tried it). Suspend your rod so the tip is straight down and just a little below the surface.
Put on around 5v with the tungsten as and anode. Very quickly, it'll eat that end off, and due to capillary action, leave a very sharp tip as it winds up above the surface of the etching solution.
People have made STM tips this way - it really works and it's pretty fast - 5-10 minutes depending on current etc.

What's the chamber geometry? Pic of the outside?

[Bob Reite]

So you all don't have to run over to fusor.net and hunt for it, here is a picture of the outside, with the lead shield removed for the photo. It's an 8" cylinder, with four ports. The top and bottom have the freedthroughs. Yes, I know I should put an anti corona ball on the top feedthrough but I have not gotten around to it.

[Doug Coulter]

Really nice! I use rounded (or at least radiused) ends on my feed throughs. It's not a big deal till around 50kv, and there and above everything becomes a big deal. When I'm not actively working on it, my HV stuff is put inside of a piece of PVC pipe that tightly fits the outside of the flange, which PVC is also coated outside with copper screening and grounded. When you get a real arc, the parasitic capacity of stuff - if not the deliberate C in the power supply, has really high Q and you get crazy peak currents. Which in turn, easily radiate - fast rise times work well with a short antenna, and I've lost a few PC's nearby - not even connected, as the power or mouse cabling etc are enough pickup to let the magic smoke out of the mobo.

If it sparks inside my HV "coax", it's not such a big deal. I shield the small signal stuff too - try and get this problem at both ends. Until it happened to me, I had no idea a ground loop could be that potent!