I mean, it's all laid out....but, I will try and reiterate...Stan did several different things, the demo cell (Dave Lawton replication) is the worst. The next iteration Stan developed right before the miniaturized plug version, was by his own words "tweaked" to produce more gas. That is the autotransformer and diode positive feedback loop. The same circuit drives the last iteration (water fuel plugs). The conversation I initially had with chat gpt is correct, the second theory I had about instantaneous voltage distribution across an inductor (how a variac can be ran in step-up or step-down) is ALSO correct...I initially didn't think both could happen...but. after running the ignition coil/diode chain feedback loop at 12v with a 1/2 watt limiting 220ohm resistor in series with the primary....I swept every harmonic up until it drops off at around 3mhz. The first resonance was found at right around 150hz. In resonance at such low power, the phenomena I showed when dumping a capacitor straight into the primary (disappearing forward induction, straight to flyback polarity) did not happen. Instead using two channels, I set channel one to trigger using the cmos signal on the isolated side of my opto. Then I started by probing the negative of the primary. I found a much higher voltage mirrored signal, with the same phase (lowest input current, and voltage rise on primary in resonance). Probing the primary LV+ before the diodes (hooked to LV+ but, unhooked from the secondary HV lead) I saw almost nothing. Moving to the end of the free-floating diode chain I saw a curious sight, but it would not trigger, i attached the second channel ground to the primary negative while second channel probe was attached to end of diode chain, and shorted the common ground in my oscilloscope. I detached the triggering ground of channel 1, and checked again (now grounded to LV+) ...on the other side of the diode chain, free-floating in air...was a HV wave of opposite polarity but in phase. The diode chain was blocking it from causing issues and the standing resonant tension across the primary kept the two inverted ground potentials separated across the primary despite the HV lead conducting through a neon load to primary ground. Two opposing currents standing still across the primary. The same thing happening across the secondary...these are the opposing counter currents of self/mutual induction....not opposing.....instead, moved out of the way, just waiting all politely lol.
When allowed to conduct, this lower powered manifestation does increase the overall power seen in the load neon....but not like before (not as intense, and the wrong polarity).....the neon shows that the polarity is NOT flyback polarity...primary ground is positive biased to the diode chain/HV lead. So, without the current density, the magnetic flux isn't strong enough to produce enough opposing self/mutual induction counter-flow to overcome the normal flow and pull it all into the backside of the secondary (postive feedback loop).....because, the higher the magnetism, the longer it takes the forward current to get across the primary coil and the higher the counter-emf voltage climbs , yet the primary voltage travels across the coil instantly...and here's the key, this autotransformer voltage distribution (instantly across the primary......and the much higher impedance secondary, as if one long variac in step-up mode) results in the LV+ overextending all the way to the top of the secondary and becoming HV+. This results in a much stronger/faster pull on the primary current lagging behind 90deg. This in turn creates an even larger counter self/mutual induction than normal, the assymetry of the primary + secondary inductance as one coil, means the voltage of the mutually induced counter current in the secondary overcomes the autotransformer voltage (HV+) faster than the primary surge-current can get through the primary (the increased voltage=greater magnetism=greater opposed self/mutual induction=greater initial current lag....quite counter-intuitively) and it makes the diode chain start to conduct the opposing mutual-induction from the secondary, and the opposing self-induction from the primary to practically zero ohm backside of the secondary. This not only removes the opposition out of the way of the primary current making its way all the way to the tip of the secondary, these two flows, now create their own opposing self/mutual induction flows that pile up on the primary current/voltage, as such, the primary current moves faster than it would normally cross the primary and now wants to go up to the LV+ turned HV+ at the top of the secondary, the opposing flows head to same place in an opposite direction through the diode chain...this compounding positive feedback loop continues until the primary current empties from the capacitor. The moment the capacitor empties, the magnetic field collapses and both currents reverse polarity and all of the currents are in the secondary (either in the coil, or on it's backside in the feedback loop)...the 100uf capacitor is not enough to make if all the way across the primary (if it makes it at all), the self/mutual induced currents flip to HV+ , the primary HV+ at the top of the secondary flips to negative and through the low resistance HV lead, keeps the diodes conducting. The HV+ counter-flows are more opposed to ground, and when I sprayed water in a confined glass bottle containing the triple electrode setup I showed before, it avoided the middle diode chain electrode and skin effect conducted on the wet glass around it to the ground electrode....but, if the path to the middle electrode is the lower resistance than the ground electrode, the HV+ will strike lower HV+ of the primary current coming out of the diode chain, and together they will then pull even more current from the ground electrode.
There is some wave destruction as the discharge voltage is lower then normal flyback (normal flyback spark can span the entire distance, and this phenomena has to be positioned to span two 1/2 distances)....but ALL the current (minus loses) is applied to HV....the primary current is not destroyed as normally happens.....hopefully it clicks now.