VIC design session v105: reasoning

This sweep didn't do what I aimed it at, and that's the first honest note to log. I set out to push f_r into 55–75 kHz and walk peak_v down so the vapor film would light inside the 80–250 Td cold-chemistry window. Every surviving candidate instead parked at 99.84 kHz, 788 V cell, and a film at 314 Td — Townsend avalanche, but ABOVE the window I was rewarding. The scorer clearly traded my band target away for tank quality, so my scoring weights are miscalibrated: I'm not penalizing an out-of-window film hard enough relative to Q.

The trade-off structure is clear. Q sits at only 7.9 (coil Q 67.5), because the 126.5Ω electrode branch and 8.4Ω choke resistance load the tank; more turns (#5, 850t) actually LOWERED tank Q to 7.7 and coil Q to 55.6 for negligible film benefit — wire loss beats magnification. The top four are degenerate in everything but duty (2/5/10/20%), which only scales average input current (14→140 mA) while circulating tank current stays 0.70 A. Faradaic leakage is the real problem: 171.6 mA at peak blows my <10 mA target by 17×. The τ_EDL of 291 ms means the double-layer never charges in 29,000 carrier cycles — displacement current is 89.9% but leakage still dominates the DC path.

I'm adopting #1: lowest duty (2%, 14 mA avg) minimizes integrated leakage energy while I still don't have the window.

Next session: reweight to gate on film Td (hard-reject >250), drop drive voltage below 100 V, and raise ppm/temperature to shrink the film Paschen voltage rather than chasing L.