VIC design session v43: reasoning

This sweep was supposed to test mid-to-high chokes (2.5–6mH) to pull resonance honestly into the 1–100kHz band, but every candidate the scorer surfaced snapped back to the 0.1mH reference anchor and lands at f_r = 303.9kHz — three times over my practical ceiling. That's the headline failure, and I won't paper over it: the objective function is rewarding the CGDE film window (all five light at 111–124 Td, squarely in the 3–6 eV cold-chemistry band) and coil-Q, but it's not penalizing the frequency violation hard enough. So the "top" candidates are physically the ones I said I wanted to move away from.

What the ranking does teach me is the Q-versus-loss trade at fixed L and C_eff. Going #1→#4, adding choke turns (120→340, AWG18) drops coil Q from 591 to 209 and shaves cell voltage 312→288V and current 816→755mA — more copper, more AC resistance, less magnification. Dropping to AWG20 (#5) collapses coil Q to 163. So thin/many turns buys nothing here; fat/few wins on Q but can't fix frequency.

I'm provisionally adopting #1 only as a Q/current reference point, NOT for build — it violates band. The real lesson: with C_eff pinned at 1.37nF (8 cells), no realistic choke lands me in-band; f_r = 1/(2π√LC) demands either much larger L or much larger C.

Next session: force the sweep to actually hold 2.5–6mH (fix the anchor leak), and simultaneously raise C_eff — fewer cells is wrong; I need MORE capacitance or longer tubes — targeting f_r near 40–70kHz while keeping the film in the Td window.