Robert Ellefson, I'm glad it's all sorted out!
To give you folks with electrical engineering background (and not only) a bit more insight to my thoughts, I want to highlight a very important characteristic on the waveforms, which is further argument that the interpretation of the observed phenomena is plausible and it's recreation in simulation is also plausible:
The inductor current's magnitude dictates the voltage rate of change on the spark cap parasitic capacitance, when there is no conductive plasma channel in the spark gap. Important to note, the higher the current, the faster the voltage rises on a constant-ish capacitance. If the current is constant into a constant capacitance, the voltage change is linear.
Let's see this in action on the measured waveforms...
On my setup if I have mid-value current (100-200mA), the spark gap capacitance voltage slew rate is higher, around ~3.5kV/us:
If my current is lower (~80mA), the spark gap capacitance voltage slew rate is lower, just 500V/us:
The LTspice simulation shows the same effect (observe on previous post).
Now the interesting part: if we analyze a waveform I took in a configuration made to give similar waveforms to Egely's ICCF osciiloscope capture (higher load resistor value, to create only unipolar spark-gap voltage, don't let the LC resonator to swing to negative current and voltage):
Note the current being low at beginning, it evolves to a maximum in the middle, then it dampens to low values at the end, in a similar way as previous waveform with lower dampening. The voltage slew rates change accordingly...
Now brace for the impact:
And I hope I'm wrong...
My biggest fear is the oxide layer causing dielectric barrier discharge and and it's typical periodic filamentary arcs, meaning the burst of current spikes are not fusion events.
DOI: 10.5772/intechopen.80433
One down, more to go?