12_1_2013, trying for Q and getting it.

[Doug Coulter]

With the new grid, I thought I'd check for the high Q mode on the edge of instability, and here are some results of that. It looks like we can make this oscillate - a PNP tube with power gain - at a number of different frequencies and in a number of different modalities, probably due to things like the differing transit times of electrons and ions between the grids, and "I don't know" what other factors. But we still see large Q at onsets of fusion, very large (400x), even though here we are very limited by the data aq time resolution - most of the bursts were much shorter than a second, its resolving time slot. I notice (and I'm guessing this is a good thing) that the Spellman main power supply is reporting actual current, not limited by the peak current setting, or what it delivers before its control loop can cut it back to the set current - same thing. You can see the current control hitting it's set limit around 400 seconds, and sitting there - a moment of stability. I've fixed the math to account correctly for my 50k ballast resistor - used to be set up for 100k and that depressed the voltage reported at high currents.
All correct now. I took a lot of video clips of all this, which I'll link in a few hours - when they finish uploading to youtube.
Here's the very raw data screenshot of the entire data aq.

Raw data aq

In the meantime, here's a screen shot showing raw (but in real units) data, and as my version of Q (which is neutrons/min counted divided by watts input, not a real Q or we'd be there already by a fat factor!).
Warning, this is one big screenshot...4 hd monitors worth. Oops, it wouldn't even fit here, I'll do some slicing and dicing so it will.
Here's the plots themselves. The one on the left is raw neutrons, the one on the right is "Q". The rest I cut out, as it partly is dupes or just the UI of the plot program.

Raw and Q plots

Long movie, some for just the sound of my stereo N detectors and commentary, oops, I could have had the camera pointed better in a few places.
http://youtu.be/7XgeFFYsXWo


As a result of the limitations of the data aq, I'm fairly sure we are actually understating the instantaneous Q we get in this mode (intermittently) but I can't be sure. I have to replace a rather expensive o-scope that let out the magic smoke when 50kv hit one of the probes...darn, it's well over a month's income...Once I do that we can tease apart cause and effect, and get real precise time-lines for what happened when during one of those little bursts. And see if it's oscillating at some odd RF frequency as it seems to do sometimes (and which seems unrelated to the rest, other than proving that yes, we do have a thing with power gain here).

One bad thing about the scatter plots. They don't show time at all. Now, I can change the code so that for instance, it draws lines between the little dots with arrows on the ends, which will tell something about the time, but makes the plots "hard to see" with all the extra pixels from those lines, and the tiny arrowheads. I will probably try that at some point, and if I like it (for this type work) I'll add it as an option in plotdat (the software) which is posted elsewhere up on the forums. At the moment, I still remember when I was doing what - for example, those red dots - that was startup with too much gas and a 5 ma current limit. Going up that line towards the full supply voltage was me taking gas out till we went under 5ma, or just went out. Then I started fooling with the ion source. At one point, I let in more gas again (the pressure was dropping due to "cleanup" in the tank) and pulled some back out - but it's hard to see on this, since it would overlay more or less the same line. Near the end of the run (as you can see on the time plot screen shots) I DID raise the current limit for a few seconds, just to see what that was like. No surprises there - the neutrons went up as Q went down, as usual.



Note, this one's taking awhile, check back later if you see this. The movie was still "processing" at youtube when I typed this - and the first time I played it here, their sound was garbled - I might have to try again, or a different encoding, because in this one, the sound is really important.

[Doug Coulter]

Here's a version of the above vid with better sound. One reason I was pointing the camera down sometimes is to pick up the stereo neutron detectors in well, stereo.
http://youtu.be/KjsSoTdpHxM


I feel like there might be a huge advantage to "driving" this oscillation, rather than just letting it "motorboat" at some frequency largely determined by the power supply caps and ballast resistors...it seemed pretty certain it was also able to oscillate at RF type frequencies, but IMO, those are going to be too fast for the transit times I've measured, which are on the order of 10usec (for what combo of particle types, dunno)...stay tuned, I'm working on that one. 10 microseconds is a LONG time for the speeds that are implied by the power supply voltages, but other indicators do seem to be telling us that we don't (usually) get a very big fraction of that onto our ions or electrons - the way things are now, that is. Maybe a little organization of where the space charge is, and when - make a large difference. I'd bet it does, and will be putting my time/effort into that.