Here's some of my nice 4d plots. The plot on the left is raw neutrons, roughly calibrated such that each 1000 on the vertical scale is about 1 million neutrons/second (it's actually about 980 cpm ~~ 1m n/s, so I'm being conservative here). The plot on the right is the same data, but the neutron counts are divided by power input, and both are corrected for actual grid voltage post the ballast resistor. That super high Q single point might be an artifact of the response time of the data AQ, and I should probably find it and edit it out of the log so the auto-scaling doesn't depress the rest and we could see a little more, but even with that fault...well, look at the picture (click as usual for a larger version).
- Raw data on the left, normalized by power on the right.
Locally, I have the option of twirling these around the main 3 axes, of course, but that's kind of hard to show here without taking a movie, which doesn't have the resolution.
I may add one later, that's the nice thing about acquired data - I can fool with it to my heart's content, as far as mining it goes.
This indicates to me, despite whatever outliers or artifacts, that Q can be increased on the order of hundreds in a standard fusor. Other data seem to indicate that it's the onsets from zero that have the high Q, but I don't have a saved high time rez trace at the moment (my cool GW instek scope fell to EMI, and I can't right now afford another 4ch 2.5 ghz scope). I've seen this whole thing many times, even managed to get it to happen about 50% of the time with a series inductor in the ballast and the thing kind of self-oscillating - maybe reproducible about half the time. There are most likely other more practical ways to "switch" the fusor, since with our separate ion source grid, we can make a "triode" that has very significant power gain...more on that elsewhere.
