John, you may have missed that we no longer use the maggie ion source, as the separate ion grid out in the big space does a better job (And lots more ionization/watt at our operating points), though it won't go to quite as low a pressure as the electron cyclotron would. For grins, I had built a 1/4 wave 2.4 gHz antenna in there to see if injecting some RF into the tank proper did anything at the power levels one can get reasonably (about 8kW peak) and nope, no good - no measurable difference whatever, things are going too slow for that. If I'd added some 980 Gauss magnets and made the electron cyclotron out in enough room to hit more atoms, well, who knows? Never got that far, that's a lot of work just to support the weight, and things in there tend to get too hot for the curie temperature of the magnets one can get. For example, the solder joints on my pinhole camera, which was well away from "the action" kept melting and I took it out of there before the lead faceplate wound up melting and falling into the turbo(!). I just use that antenna as a Faraday probe now (it was the blue trace on the scope in the video, Far_B, for "back of the tank"). You can see it jump a couple scope divs down - that's our net negative charge being shown. A few hundred volts at that spot (100 meg ohm load).
I'm doing a lot better than a relaxation oscillator now (I was using a series L vs an R, FWIW). Now I have the possibility and initial test of something much nicer here - an active device (the fusor itself!) that has electrical power gain...and with the right other stuff can be made to oscillate at a range of frequencies, even more than one at a time (ask anyone who's made a high power linear ham transmitter amp about parasitics, or just study a super-regen detector...The latter is close.). These can be partially timed or synced to charge-disturbance flight times between the active parts - something John helped measure when he was here last. (And thanks for that one, John, the extra set of skilled eyes really made a difference - come back any time!)
I've learned a lot more about what goes on in there, and how dead wrong (or
not even wrong as W. Pauli would have said) everyone is about how things act in a fusor. As John has pointed out (fusor.net somewhere IIRC), nope, we are not dealing with widely separated arranged + and - charges - that would take serious amounts of energy to even create (perhaps enough to vaporize the entire lab) and people who
do work with well-separated charges, like CERN in their beams, in any decent current, have issues with the ion charge field just yanking electrons right out of the tank walls (field emission)! Which gives them issues for what they are attempting as the electrons eat some of the proton energy, and then bash the tank walls at the focusing magnets - a real mess, making X rays and heat at your superconducting magnets. John deserves some serious credit for getting me thinking along the right lines as to what to measure, at least.
What we actually have is a slightly net-negative plasma - the extra electrons are from our power supply (or pump) - since that's the only thing that will get kicked out of any conductor by impact, or flow through one. My wires just don't pass deuterium (rats?). You can't accelerate a plasma with an E field or perhaps at best with a slightly imbalanced plasma - only a little (and here it'd be in the wrong direction!). All you can do, is like a bar magnet in say the earth's field - align the polarization, maybe increase it a little. It takes a
gradient to accelerate one. And your slightly polarized plasma affects that gradient itself, by being polarized and where it is. There of course is no re-circulation in such a system - it's at best a resonant circuit with losses (scattering) so that idea was just all wrong. I measured for that, and there's nada. That whole theory is bunk. Sorry, Philo and friends. FWIW, since the electrons weigh so much less, that polarization can itself oscillate from side to side of our cylindrical collapsing plasma sheet. At some other frequency which varies with density and where it is in the overall machine..just to make it simpler(!).
Our idea of space-charge limited focus was
also wrong, because we don't have a big charge separation for there to even be what would be called space charge in say, a vacuum electron tube. It's messier than that. The parameter space is huge in fact, and makes the math for quadrupole mass specs look simple as well as a static solution to a dynamic problem that would have to add terms and be iterated like a fractal to even apply.
Past my speed, and my new friends at CERN, too. When I try to frame this problem in these terms every math guy just throws up his hands.
It turns out that in this case, the universe is its own best model and runs in realtime, though with pauses as the experimenter has to shut down to change some things.
But like the Mathieu equations, there are predictably interacting things - higher voltages (or gradients) mean higher speeds (very loosely) and its intersections in a number of functions is where the fun is.
This pdf has a nice little picture of
one of the most used intersections in a mass spectrometer, the largest one near the origin of the synthetic variables plotted (which is mislabeled as THE intersection, but on a larger plot there are many). The full plot looks like a couple spiders having kinky sex, and all the overlaps work at least somewhat, though the ones further out are more picky. I think we found
a spot, but don't know if it's
the good one yet. Unlikely I got that lucky.
Now, we have two e/m ratios, a plasma that's polarized but can't be completely independent re each charge vs the other, so add all that to the math and start trying to work it out. Good luck! (A good thing you don't need and can't patent math for this) The amount and direction of the polarization itself is dynamic here! Although it appears that once we map the space better, no existing standard model laws need to be violated, finding this by math synthesis is well,,,,heat death of the universe kind of hard. But here we're not going forward from a theory we're not sure anyone can produce. We HAVE an existence proof of at least one such good spot, though we have all too little data on those parameters past what I could observe, as last time it happened all the data aq failed instantly (and I was sick for awhile and too busy not-dying to write it all down).
I therefore think we'll get this one, since we know it's there (or at least I trust myself,
I completely understand others' skepticism, I'd have it too if I wasn't a first person witness - that's how good science works.). People who say "no way Jose" are like the American generals who looked at how hard the Manhattan project was and projected it'd be that hard for the Russians, and were WAY off with that. Why? The Russians weren't risking it all - they had an existence-proof (Japan), as do I. We didn't know if it would work, or even which of the possibilities might work. They knew from watching us! And they weren't dumb, as our egoists wanted to believe.
So, I can't predict timing here, and like with the Mathieu stuff, it's a little slippery and more than one thing can work. I could simply create a gonzo arb waveform generator (actually, I have two already) and a huge linear amplifier (would have to build that, and zowie, without knowing what is the required bandwidth in advance) and find out what the waveform parameters were by brute force/try everything. Heat death of the universe turf. Nope! My own expiration date is closer in, and I want to beat that one!
What I do hope to do is find that or another of the sweet spots, do a little diddling and gradient following, and find the big one if I haven't already.
That should in theory be a lot easier. And I have a built-in power amplifier with power gain > 100 (electrical), "interesting" transit time characteristics which are related to the very thing I want to manipulate (maybe good?), and wow - pretty much as great a power handling as I could ever want. I don't feel like that is going to be the limiting factor, though I may have to move some electrode mechanically to get the transit time sweep into the right range - but I already know
one of the good spots doesn't even require that, hence my confidence if I can hang on to my health long enough.
Note, I am leaving out one very crucial insight here, but heck, might as well share what I can right now. You would think even what I'm saying here would be "obvious to a practitioner of the art" but the evidence is otherwise. What I see in most other workers is deliberate lack of willingness to work out all the ramifications of their assumptions - and as my email sig says "Why guess when you can know? Measure!".
Like the out and out crackpots I often have to deal with, they refuse to acknowledge that if their silly theory was correct, other things we observe simply wouldn't happen. Or things we never observe would be common. At least I'm staying fully standard model (that's my story and I'm sticking with it!). I doubt any new extensions under consideration to that model by the far-out theory guys make any difference to mere nuclear physics involved here (FWIW, a couple of them agree on that).
I guess I just relaxed the rules about talking theory/interpretation here. At least it isn't on the front page of the news...If it shows up there we'll just have to call it fake news, right? (for now)
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.