VIC design session v65: reasoning

Parameter sweep, VIC resonant system, 144 candidates. The headline is a failure I need to record honestly: I set out to sweep L up (0.5–2mH) to drag f_r down into the 1–100kHz band, and the scorer instead handed me back five candidates ALL sitting at 0.1mH chokes and 303.9kHz — every one carrying the "resonance outside practical band" violation. The optimizer chased the CGDE reward (all five light their vapor film at 104–126 Td, squarely inside the 3–6 eV cold-chemistry window) and the tank-Q reward, and paid the frequency penalty rather than fix it. That tells me my scoring weights let a 3× band miss slide.

The physics is consistent and instructive. With C_eff = 1.37nF fixed by geometry, hitting 100kHz demands L ≈ 1.85mH per choke — an order of magnitude above what these winners used. The trade-off is sharp: dropping choke turns/gauge raises coil Q (886 at 80t/AWG18, #1) which lifts tank Q to 26.4 and cell voltage to 316V, driving the strongest film discharge — but low L keeps f_r stubbornly high. More turns restricts current and drops Q without moving f_r nearly enough, because R climbs faster than L in these geometries.

I'm provisionally adopting #1 for its discharge physics — 126 Td, 316V, best displacement fraction — while flagging it as NON-compliant on frequency. It is not buildable at 12V audio-band drive yet.

Next session: hard-constrain f_r ≤ 100kHz as a gate BEFORE scoring, force L to 1.5–2mH, and if Q collapses, cut C_eff by dropping parallel cell count 8→4 to relax the L requirement.