I like it when experiments turn out more or less as I expect - hopefully, I thought it through, and got a little improvement, or at least learned something new. People who know me know that me being at a loss for words is quite rare.
Here's the really short couple vids - holy crap, batman!
http://youtu.be/7JVfR6wZ1pU
This might be close to that. While I'm sure some of what I just saw was equipment error/artifact, some was certainly not, and it's crazy too-good-to-be-true.
I was running a new setup. I have a 4+::1 turns ratio transformer, with the high turns winding in series with our main grid, and the secondary coupled to our "ion source" grid. Since I'd earlier discovered that we have a triode here with a lot of power and voltage gain, I thought it might be elegant to see what would happen just making it oscillate on purpose rather than the usual parasitics, which we'd seen in the past with nothing at all special to help it happen (Hams will know what happens if you bias a big power tube in its active range with even untuned circuits in the grid and plate - same here). It was a wild shot. Electron transit times etc would make things oscillate way too fast - which we'd seen by accident. Ion times a little too slow, as they'd probably hit the walls and be lost, so the idea was, make this broadband transformer (it's really good from around 20khz up to around 28 mhz - anyone who knows RF knows that means I did quite a job there). No deliberate tuning, just strays.
We are feeding the main grid through a 50k ohm 225w resistor, then the transformer primary. We are feeding the second "ion source" grid through another 50k resistor, but with the transformer secondary hooked straight to the grid, and with the other side of the secondary connected through a .01uf/40kv cap to a 6.8k "arc protect" resistor on the ground side of the cap.
The results were, I think, astonishing. But there was so much "RF in the shack" that my data aq blew up - actually, the host computer for it refused to respond instantly with even just the ion grid power on, and it free-run oscillating.
What is VERY STRANGE is that all my neutron detectors were showing counts in the range for 8-9 million neuts/second off just that - 36kv or so at 12 ma (current limit of the little spellman that we use for that). If I assume square law for distance, for detectors calibrated for their distance from the main grid, but further from this one - it works out right around 4 times that! Holy cow!
Obviously I need to get to beavering away on shielding and get my normal data aq back going again. And oh yes, when just foolin around (all I could do without some decent data aq other than eyes and ears and handheld counters) with both grids turned on - it was FAR better than usual. The ear isn't a computer/counter, but it's not utterly stupid either, and we connect our detectors through a multichannel audio rig for tuning all the time as a "backup" which in this case was about all I had.
So, Pix and such. Here's one of it all going at once:
For once it really does look like a star in a jar - not the usual recombining D spectra.
First shot with new 2nd grid wiring:
Here's another with just the 2nd grid going - you can see where I turned off its DC supply - the X rays and neutrons (from a far away detector!) go down then.
Sadly, once I kicked on the main supply there was no data take-able. Even my multimeter with the HV probe on it went blink and all the digit bars on....must have been a LOT of RF. In fact, I put on the scope and picked some up from across the room. I couldn't even standard-count my silver, but next time I'll have a backup at hand...could be tomorrow. If that shows the numbers I expect, well then....this was the right trick, and did I say I haven't even hooked up a tuning cap yet? Nor tried the other polarity of the transformer (it's in the polarity now that ought to be negative feedback if not for transit times).
But the nice new scope was safely 15 feet away, and caught some of what was going on with nothing but a probe tip floating in midair a long way from the fusor.
Here's a little of what it saw.
I'm combining a couple of really short movie clips on another computer and I will add them to this post so you can see/hear it in action. Either we're onto something here, or I just had every single piece of electronics in my lab tell me lies, or go dead.
Seems unlikely. It's probably not as good as it looks - that would be incredible luck, but it's certainly better than before - seems to have been a good guess on the new lashup which BTW, still has the 3/4 scale main grid in it. I don't open that door any more than I absolutely have to - it takes days to get back to really good purity in the vacuum each time.