VIC design session v74: reasoning

Lab note — VIC resonant sweep, 90 candidates.

The headline result is a failure I need to own: the session's stated goal was to drag f_r down from ~300kHz into the 40-70kHz band by raising choke L ~20-25x (0.0001→0.0025H). Every surviving top candidate still sits at L=0.0001H and f_r=303.9kHz, tripping the same VIOLATION — resonance outside the practical 1-100kHz band. Either the higher-L candidates got scored out, or the L knob never actually swept. I don't trust that the frequency lever moved this run; that's the first thing to fix.

What the sweep does show cleanly is the Q-vs-turns trade at fixed L and C. More turns buys coil Q at the cost of AC resistance, but tank Q falls (25.9→22.2) because winding resistance loads the tank, and cell voltage tracks tank Q down (311V→266V). Series current likewise slides 813→698mA. So within a fixed geometry, fewer turns / heavier gauge (AWG18, 130T) wins on both voltage and the CGDE film metric — #1 lights the vapor film at 124 Td, deepest in the 3-6 eV cold-chemistry window.

I'm adopting #1 provisionally: best tank Q, highest cell voltage, best Td placement. But I'm flagging it as band-illegal — 304kHz is not buildable with honest switching, and displacement current at 96.5% means we're capacitively coupling, not dissociating.

Next session: actually enforce the L sweep to 0.0025H and confirm f_r drops toward 60kHz; then re-run turns/AWG for Q at the legal frequency. Also vary C_eff (tube gap, cell count) as a second path to lower f_r before spending more on choke iron.