WFC design session v34: reasoning

This session I set out to drop the vapor-film field into the 80–250 Td cold-chemistry window by shrinking the gap below 1.8mm and deliberately starving peak_v. It didn't land. Every top candidate still lights its 0.1mm film at 372–403 Td — comfortably above the target ceiling and squarely in townsend-avalanche, not the CGDE regime I was scoring for. That's why nothing broke 0.66; the film diagnostic is the binding constraint now, not resonance.

The trade-offs are stark. To pull Td down I have to lower cell voltage, but cell voltage is Q×drive, and Q is what buys me the ×28 step-up in the first place. Tightening the annulus (16.4→15.6mm) raised C to 1.79nF and dropped f_r to 88.5kHz, which cost me Q (28.1→25.9) and thus voltage (1012→933V) — but only nudged Td from 403 to 372. The gap shrinks faster than the field, so Td barely moves while breakdown margin erodes (×141→×125). Frequency band stayed healthy (88–97kHz) throughout; it is not my limiter.

I'm adopting #1: 12/16.4mm × 100mm, 0.5ppm/8°C/passivated ×0.4, 2% duty. Same score as #2/#3 but lowest average input (19mA) — cleanest ladder-climb while I'm still above the window, and the Fig-8 leakage (88mA at peak, 97.7% displacement) is tolerable.

Next session: stop chasing Td with geometry. Cut drive voltage directly (36→18–24V) and/or lower Q to force cell voltage under ~500V, and widen the film-thickness assumption (0.05–0.15mm) — the 0.1mm fixed film may be miscalling onset.