VIC design session v19: reasoning

Lab note, 2026-07-10 — mid-L sweep (0.3-0.8mH band + 0.2-0.3mH edge)

All five top candidates converge on the same 0.0002H choke value regardless of AWG/turns fill — the geometric C_eff (1.19nF) at this cell geometry pins f_r to 230.96kHz no matter how I vary turns/wire gauge across the swept band. That's the real finding: turns and AWG move coil Q (150.8 down to 91.3) and wire loss, but they don't touch f_r, which is set entirely by L·C_eff. So this whole band is locked out-of-spec — every candidate violates the 1-100kHz practical drive window even though the vapor-film diagnostic likes them (82-87 Td, squarely in the 3-6 eV CGDE window at 202-218V across the film).

I'm adopting #1 provisionally as the reference point (highest Q 18.2, best breakdown margin ×548, cleanest Td at 87) but I'm not calling it a build candidate — it fails the frequency constraint outright, and Stan's documented driver topology (choke-coils 9/12 as step-charge current limiters) was never characterized above ~100kHz. Higher turns (180→260) buys breakdown margin at the cost of Q and cell voltage, a straight trade I'd rather spend on lowering f_r instead.

Next session: hold this cell geometry (9mm/13mm/100mm, C_eff 1.19nF) fixed and push L up an order of magnitude (2-5mH) to drag f_r back under 100kHz while watching whether the Td/CGDE window survives the lower cell voltage that comes with it — that's the real unknown, not turns count.