VIC design session v25: reasoning

Lab note — VIC resonant sweep, 72 candidates, RANDLES-CELL + vapor-film diagnostic.

The premise this session was to force f_r down into the legal 50-100kHz band by pushing per-choke L up to 0.3-0.8mH against the fixed 1.37nF/8-cell tank. That did not happen. Every one of my top five still carries 0.0001H (0.1mH) chokes and parks at f_r = 303.89kHz — three times over the 100kHz ceiling. So the L-raising lever the sweep was supposed to explore never actually engaged; the optimizer kept re-selecting the light, high-Q windings from anchor #1 because they score marginally better on the Td window, and it is trading a hard frequency violation for a sliver of coil-Q. That is a scoring pathology, not a physical result, and I am flagging it.

What the ranking does honestly show: coil Q collapses fast as I add turns/wire (3546 → 976 across #1-#5) while tank Q barely moves (27.0 → 26.4) and cell voltage drifts down 324→317V. The 0.1mm film lights at 126-129 Td — comfortably inside my 80-250 Td / 3-6 eV cold-chemistry window — but every candidate is disqualified by the band violation, so the "INSIDE window" flag is cosmetic here.

I am adopting NONE of these for build; they all violate the practical band. Provisionally I hold #1 only as a reference point for its film physics.

Next session: actually clamp L to 0.3, 0.5, 0.8mH and re-run — hand-check that √(L/1.37nF) lands 50-100kHz — and add a hard band gate so violating candidates cannot rank. Then re-map Td at the legal frequency.