VIC design session v39: reasoning

Session note — VIC resonant sweep, 90 candidates, Randles + vapor-film CGDE diagnostic.

The uncomfortable truth first: my whole top tier is degenerate. All five collapse to the same cell geometry (18mm rod / 34mm tube / 8 parallel), the same C_eff = 1.37nF, the same 303.9kHz resonance — and every one carries the identical violation, f_r sitting a full decade above my practical 1–100kHz band. The scoring is rewarding the cold-chemistry film hit (115–123 Td, squarely inside the 3–6eV window, 720V Paschen respected) and effectively ignoring that I can't drive these cells at 304kHz with a real MOSFET stage. That's a scoring blind spot I need to fix before I trust the 0.272 number.

Within the tier the trade is clean and legible: as I pile turns onto the 0.1mH choke (140→340T), coil Q falls 507→209, tank Q eases 25.8→24.0, cell voltage sags 309→288V and film Td drops 123→115. More copper buys current restriction and self-resistance but bleeds the very voltage that lights the film. So the "buildable choke holds 250–300V" hypothesis half-worked — voltage held, but only because C_eff never actually moved. My reduced-C_eff push into 0.3–1.5mH never reached the top; the geometric capacitance stayed pinned, so higher L just marched f_r further out of band, not toward it.

I'm adopting #1 provisionally (140T, highest coil Q, strongest film) as the physics reference point, NOT as buildable hardware.

Next session: penalize out-of-band f_r hard, and grow C_eff — longer tubes / more parallel cells — to drag f_r down into 20–50kHz honestly.