This all brings up the idea of a complex "error surface" loud and clear. Optimizing along a gradient from wherever you happen to be gets you to a local minima in that surface for sure -- it just many not be the global one -- to find a global in a complex surface requires another strategy, like simulated annealing or a genetic algorithm (and you, Chris, are our wildcard there). One thing I seem to be finding out is that it's effective density at the interaction zone that seems to matter most, and that at least in a local sense, going to lower gas pressure, higher voltage helps with that (along with other things), even though the basic media is less dense -- this is taken from years of work with electrons, same issues in various electron-based devices -- same issues, just a different scaling factor involved there (and with ions, it tends to be worse re defocusing). It only needs to have high luminosity in the collision zone, having it elsewhere just makes it want to "Coloumb explode", attract neutralizing electrons, and be hard to manage. Hence the various approaches by others that seem to think that having a bunch of electrons in the mix will cancel some of that out. I personally am having troubles making that work philosophically, due to some experience with what the electrons actually do -- where they wind up in the mix at least in the rigs I've tried so far.
In fact, in my estimation, the electrons are the very main source of losses in a normal fusor, from visible light, to X rays to heat to neutralizing ions you've invested significant energy in and taking them out of the game. Not too surprisingly, a Faraday probe "out in the tank" always reads very negative (in non pulsed running) -- there are far more electrons in play than ions in my particular lashup, which has encouraged me to try grid materials with less secondary emission, and propose a new design of shielded grid (more on that in its own thread, soon). Having run the numbers and thought about it some, I think Jon and I might be on the exact same page re a bit of magnetic shielding. For example, in a cylinder fusor, one might consider a couple of donut magnets with the H field aligned down the length of the cylinder. "In theory" this should inhibit the electrons released from the grid due to ion impacts from having a straight path to the tank walls. It's worth trying, I've got the stuff, but just haven't gotten to it. A minute worth of math says that it should not take a lot of field to have quite an effect on electrons while barely perturbing any fast ions -- you might even get some good out of increasing their path lengths, due to the fact that they will help ionize more neutrals, like in a magnetron ion gage. Of course, I know from experience that any permanent magnet will quickly be fried in there, but speaker type ceramic donut magnets are cheap enough to be worth roasting a couple in the attempt -- if it works, then one could solve the other problems involved. Probably best to try a duct tape class solution first to see if it's worth going down that road, in an otherwise well characterized situation. When I have that (not quite, yet), then will be the time to give it a go. At this instant, cracking the door to put them in changes so many other things that it wouldn't be a conclusive test, and making that level of field from outside the tank is onerous to do for the "just try it" attitude. Probably need about a k gauss for that.
As Jon says -- reducing input power is the way to go for a home situation. If you got to breakeven, or even factor 100 better than now, at these input powers you'd be life threatening danger at hundreds of watts input. Our little finding of higher Q's in a pulsed mode still made millions of fusions/second -- but at 5 watts input. That's about the limit of what I want to have in my house!
TD fusion is known to be about 100x better than DD in most setups, but T is a little hard to get hold of, and is a controlled substance in most countries, so sadly, we can't talk about it all that much here without risking some issues we don't want to have with our governments. It's available in tiny quantities (glow in the dark things like gun sights and exit signs), however, so at some point I suppose we try it. Those super high energy neutrons from that reaction are going to be an issue, as ITER is just figuring out, re damage to everything else. A possible issue in a fusor is the differeing charge/mass ratio, which would affect focus if one believes the classic math on that one. All other known fusion reactions take more input energy, and are harder to measure the output from till you get a lot closer to breakeven. I'd love to be working with say, P->Li for example, but how to see the hot alphas in a sea of hot protons? Further, if working with Li metal, it would get all over everything in the tank, especially insulators, and wreck them. LiF is a possible candidate, but would release some F in the process, also not very good for the gear.
I second the idea of skipping the cryo. Where I live it's just not realistic to tool up for. The nearest place one can have it delivered is an hour drive from here and that is huge added expense. I'd have to see a really good justification to want to go there, and the only one I have now is that it'd be nice to have a really good gamma spectrometer (Ge), which needs that. For now, I'd settle for a mediocre one instead.
As many in the big-science fusion field point out, you don't even really need breakeven to be useful -- you can use the neutrons to breed things and drive non-chain reactions that have gain when driven as well, I don't like this much, but it's there.
But Jon and I agree in (more than) one thing -- going big before we go at all seems not that smart. If we could, in the limit, fire just two ions at one another and get enough precision and so on to get a very good interaction probability -- we'll have done something no one else has yet (including the big science boys), and at that scale (or much nearer that scale than we now are) and without the complications of a bunch of electrons, we have a better chance than just pouring the coals to it and hoping it all works out. After all, that's been done since Farnsworth, and while we are close to duplicating his most optimistic results, even exceeding them in some cases, it's not by much at all -- that road is well trodden, and I'm not so vain as to think that for sure I'll find a secret there that no one else has noticed. It's happened to me in other endeavors, but it's an outlier, statistically. I much more favor ideas that have worked out well in cases where few ions have been precisely controlled and marshaled to go where you want, when you want, simply refined from what's been done and with raised energies, as
I've proposed on another thread here. A dense group of charged particles is about the most difficult thing there is to control, so why not skip the density except for just where you need it -- at the interaction zone, and just when you need it -- the time they are all there, and let things spread out the rest of the time?
Seems logical to me, at any rate, not to try the extremely difficult before trying the obviously quite possible. The limit there seems to be -- you need a short focal length to get the time they "see" one another and repel short in both space and time, but to have low enough density to control outside the focal area, you need room to let them expand when they are not actively headed to the interaction point.
It's just taking awhile to build and get characterized -- if my back of envelope math is close, that's going to need one heck of a good video amplifier, the likes of which really don't exist off the shelf. Luckily, it looks do-able with very similar kit I already mostly have. I have some other plans to engage before that one -- I want to try my shielded grid idea, and get some more probes in there to measure what exactly is going on with what we have already for reference first, as a logical next step before I take the big leap to the new design, as I think what I'll learn (and share) from that will be valuable regardless, and it's a relatively easy thing to do right now. As we used to say in the troubleshooting business -- always check the easy stuff first.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.