VIC design session v88: reasoning

This sweep is honest about a tension I can't dodge: every one of the top-five scores 0.271 because the vapor film lights beautifully — 118 Td, dead-center in my 3-6 eV cold-chemistry window at ~295V across the 0.1mm film — yet all five VIOLATE the practical band at f_r = 277kHz. The scorer rewarded the CGDE regime and ignored that I can't cleanly drive 277kHz with a 12V bench rig and real switching losses. So the raw ranking is misleading; I'm treating the band violation as disqualifying.

The physics is clean. With C_eff fixed at 1.37nF, f_r scales as 1/√L, so to drag 277kHz under 100kHz I need L up ~7.7×, i.e. from 0.12mH toward ~0.92-1mH — exactly the anchor I was bridging to. The good news the fine grid confirms: as f_r crosses under 100kHz and series current drops from ~707mA toward a few hundred mA, the 0.1mm film STAYS LIT. The window is set by cell voltage (still ~×24 tank gain), not by current, so restricting current with a bigger choke doesn't extinguish the discharge — it just trades tank Q (24.6 here) for compliance.

I'm therefore NOT adopting #1-5 as-drawn. I adopt candidate #1's cell/water geometry (18mm rod / 34mm tube, 8-parallel, 2ppm) but push each choke to ~0.9-1mH so f_r lands near 90-100kHz. The turns/AWG check shows AWG18 coil Q stays >1000 across 20-60 turns, so wire loss isn't the constraint.

Next session: build the 1mH choke explicitly, re-sweep L on a fine grid from 0.7-1.2mH, and confirm Td holds 80-250 as current falls.