VIC design session v79: reasoning

Ran the sweep to bound f_r sensitivity outside the 4-7mH/55kHz window I'd been living in, and the result is a warning as much as a finding. Every top-scoring candidate the optimizer handed back clusters at a tiny realized choke — 0.0001H (100µH) — which drives f_r up to 303.9kHz, far outside my practical 1-100kHz band. The scorer is chasing the vapor-film CGDE window (all five land 97-124 Td, inside the 3-6 eV cold-chemistry regime) and the higher displacement current that comes with the smaller reactance, and it's willing to eat the frequency violation to get there. That's the core trade-off this session exposed: pushing cell voltage and film energy up the resonance ladder wants low X and high Q, but the honest turns/geometry realization of a switchable, buildable choke pulls me back toward mH-scale inductors and sub-100kHz operation.

Note the Q-vs-copper story is clean and real: #1 (AWG18) gives coil Q 545 and ×25.9 magnification for 311V; dropping to AWG20/22/24 bleeds coil Q to 150/122/93 and voltage to 276/266/252V purely on wire resistance. Current restriction stays healthy (700-810mA) across the set.

I'm not adopting any of these five — they violate the frequency band and the 100µH choke contradicts my own resonant-charging-choke truths (500µH-2mH). I'm provisionally tagging #1 as the "physics-ideal but unbuildable" reference point.

Next session: constrain the optimizer to 1-3mH realized L, re-run at 55kHz, and see whether I can still light the film inside 80-250 Td with buildable turns. If not, I need bigger C_eff (more cells) rather than smaller L.