Grid size

[bthoma]

I have a rather large fusor made from a 10" x 8" cross. Basically, it is a cylindrical chamber. I have been using a spherical grid made from tungsten. The grid I did first fusion on was about 1" and dwarfed by the chamber. Considering the size of a nuclei, a long distance, nonetheless. Advice from some of the more experienced members have suggested a cylindrical grid, which makes sense for a balanced geometry. I am still thinking about a design. I pulled the feedthrough for cleaning today, and dropped the grid, cracking one of the gridwires. I made another spherical grid, but this time at about a 2" diameter. It works well, runs cooler, and I am up to 170k n/s peak. Thanks again to all for the continued advice. This forum really is the best.

[Chris Bradley]

I seem to recall Frank S. posting that in his experiments he found a cathode size of around 70% of the anode size to be optimum. Can't find the post right now to double check that.

There also seems to have been reports recently from folks who have run 'small' (3") type chambers of good neutron numbers.

I guess there must be some optimum size of chamber for maximum reaction rate, because the 'surface area' to 'volume' ratio changes and one might presume there is an optimum that balances out, consequent to the effects of both of those values on rate/efficiency.

I've speculated that ions useful for fusion may be created by secondary electrons from the shell, because that is the location from which they can get the biggest acceleration, so it may be those ions that might make a particular contribution to the overall reaction rate. If that were so, then perhaps secondary electrons might then play a role in a relatively larger fraction of the volume of a small chamber than a large one, if we assume the actual area of the shell from which secondary electron emissions happen at the end of the beams is independent of chamber size (because the beam intersects the shell at a point, which is therefore independent of size).

Just a few thoughts...

[Doug Coulter]

Here's what we've found here, so far. It's not definitive by any means.

It seems ideal grid size is about 1/6 the shell size, maybe slightly larger. The inch to 1.25 inch range in a 6 inch tank consistently gives better results here than either 3/4" or 2" grids. We have messed with dual grid systems, so we have also made some up to 50% of the tank size and since it was then easy, tried those alone - not good at all.

Though to a certain mindset, it would seem a vane grid might be good, tried those too. Not so good with any conditions we've yet run. You just get a very blurry version of what you get with a wire grid, and the kernel doesn't collapse lengthwise as it does in the better performing grids.

At another extreme, we've been using an about 1/2" grid out in the big (14" ID by 24" tall) tank, as an ion source for the fusor in the side arm. While this works very well as an ion source, even at pressures below where the main fusor will light off alone, it makes very few neutrons even when run at 40kv and 10 ma. There is more that could be done in testing there, but I haven't yet because for now, it's doing the job I want it to. A fun demo is to set up fusor conditions with with pressure on the low side so my main grid won't light off at our 50kv max input. Then flip on a current limited 40kv supply to that second tiny grid out in the big space -- and even though it will current limit down to 7-8 kv, you can turn the main grid on and off with it, vary current the main grid draws and control neutron production. This continues working better and better as pressure is reduced to the point where the second grid itself won't light. The action in that case is reminiscent of a control grid in a vacuum tube, though of course the mechanism is quite different.

Not tried (at least, not with success) is to run the main fusor in the main larger tank, to find the ideal grid size which might generate some sort of scaling rule. Data from one point doesn't really tell us enough to extrapolate. Even two is only two...But with various things so far tried, the big tank isn't better *except* that due to Paschen's law operation, lets you get light-off with lower pressures. Even accounting for the increased distance to the detectors and square law, it's just not given us any neutron counts to brag on like the smaller tank/grid has. Yet.

[Richard Hull]

I have always mused that the buried deterium molecules over the entire anode under constant bombardment by electrons, coupled with the secondary emission electrons ionizing near the shell is probably the source for the bulk of the fusion ions that will ultimately undergo fusion. This is the largest volume and surface area in the fusor. Volumetrically there is a lot of gas ready to be ionized and a lot of secondary electrons. Area wise, the shell is a nice large surface to bury fast or slow neutrals, ready to be popped back out by the torrent of electrons streaming in.

It is important to remember that hydrogen loves to nestle into the surface structure of metals. I often wondered if an ultra heavy electroplating of Pd on the inner shell, would assist in a simple boosting of our reactions due to it ability to absorb hydrogen. Pure nickel or, at a lower level, titanium might also see an increase.

"Palladium black" which is often applied as a suspension, much like an aquadag, might also serve.

Richard Hull

[Dustin]

[Chris]
I seem to recall Frank S. posting that in his experiments he found a cathode size of around 70% of the anode size to be optimum. Can't find the post right now to double check that.

viewtopic.php?f=6&t=2936&hilit=fusor+efficiency#p18103

Steve.

[Chris Bradley]

[Well linked, Steve.]

So if Doug's work suggests a 1.25" grid works best, and Frank's analysis says the anode should be 141% of the grid size, and we have had a recent spate of 3" fusors performing good, then I think that adds up to a 'virtual anecdote'!!

[Frank Sanns]

I am a little late to this but Steven H. has listed one link. Here is another of my earlier posts on the matter.

viewtopic.php?f=6&t=2939#p12449


Frank Sanns

[Doug Coulter]

You know, Richard, that one's worth a try, I think. While my model is a little different (but despite more data than most have taken is only weakly backed up by measurements), you make a good point here.

You'd have to hope that the ions created in the surface layer would come loose and not be held in the metal (with a very short life before being neutralized). However, that IS the ideal place to start an ion, rather than somewhere partway down the field gradient, and you should see a definite difference in the neutron threshold voltage - loosely described as that point where your detector is definitely counting above background decently.

We here see, and others have reported this threshold in the 16kv region (very roughly), where most theory would indicate you should see it around 5kv if (and only if) the ions see the full field.

I subscribe to the Feynman idea that a theory that doesn't predict anything you can test won't buy a cup of coffee, so there's the test.

I would bet Ti, which works well for beam on target, would work here, and would be easy to try by evaporating some onto the tank shell. Or even Pd, since you'd need very little for that. In fact, I'd think too thick a layer would simply give you gas pressure control issues, as it acted like a reservoir and released gas on heating...

I guess you're suggesting Pd as it holds the gas so loosely, whereas Ti is superior in beam->target devices for clinging much more tightly at higher temperatures?

I have noticed in experiments here that when you have high enough gas pressure to keep a fusor lit on its own, it's not as good as an ion-gunned fusor that goes out when you turn off the ion gun. This would seem to allow you to run in that low pressure mode as well.

I'll put this on my TODO list. I need to make a few more instrument/fusor calibration runs with things as is, which will give me some more confidence in my baseline so I can see if there's any significant difference with a wall coating designed to hold D in place. AT present, my lab is busy producing a batch of calibrated neutron and geiger counters to enhance our ability to compare things between labs, and that's important enough to finish before starting something new.

This is what our boards are really here for -- good idea interchange.

[Frank Sanns]

Cold start up of a fusor that has been open to the air is a long process. It is more than outgassing the metal chamber interior. Somewhere in the posts, I have reported that on start up, I pump overnight at full vacuum. When the chamber is down to 10E-5 torr the next day, I shut the vacuum off and bleed back with deuterium up to around 1 torr. This seems like a waste of gas but letting the chamber sit at that pressure with the deuterium fill sure puts the fusor along in the neutron producing curve. Glow clean as the chamber is slowly pumped down for a fusion run and you are guaranteed a better neutron count. Still not as good as just running for days but when you are short on time, it definitely helps. Just more evidence that conditioning the fusor has much to do with what is being embedded in the top atomic layers of the fusor shell.

Frank Sanns

[Doug Coulter]

Thanks for the data point, Frank. I think I might have noticed something like this myself, but not enough to be dead sure. These days I just keep them at base pressure between runs, and get pretty good results by just letting in gas and going right away. I don't see a short term increase in performance in the first 20 min or so of running that I can't attribute to just tweaking the conditions better, however -- if anything they work best right at the start. This could be because most materials don't hold D well when hot? My outer shell for sure gets hot, enough to char paper if run hard and long, to the point where I've been trying to think of better cooling than the little fans I have on there.

For awhile there, I was getting really good results with Ti grids (but not as good as the current, more accurate W/C composite one). Could one suppose that this might have been due to Ti getting sputtered onto the shell walls and holding D more tightly there, even at the higher temperatures?

Seems like the thing to do would be to deliberately put on a good layer of that and see if the difference can be measured repeatedly & repeatably?

Edit:
Slightly OT, but this very thing is the reason I've been putting on a push to record in a database just about any bit of data I can collect during a fusor run -- power supply, pressure, neutron counters, geiger counters and "operator's impression notes" for starters - perhaps realtime audio and scope traces as well. Then, if a topic like this came up, one could simply go back through the data looking for things one didn't notice before, and have an instant "confirm or deny" sort of indication of whether there was something there or not. I had gotten that to a fairly unstable rev 1 - and am working now on rev 2, which as any software guy knows, is always the first rev to work right. In fact, that is nearly all the effort being put forth here along with producing some "standard counters". I really believe a database of past runs to have significant value when working out what to do next. I was taught long ago to always save the raw data, not just the current interpretation of it, for this very reason.