VIC design session v26: reasoning

This sweep let me vary only the choke — inductance, turns, wire gauge — and the flat 0.274 plateau across the top four tells the real story: with L pinned at 100µH each side, I've fixed the resonant point at 303.9kHz regardless of how I wind the coil. That's the headline violation. Every candidate lights its vapor film inside the 80-250 Td cold-chemistry window (127-129 Td, ~4 eV) and holds 1000+ breakdown margin, so the CGDE physics is happy — but I'm resonating at 3× the top of my practical 1-100kHz drive band. That's not buildable with my current switching.

The turns/gauge sweep only trades coil Q, not tank behavior: 20 turns AWG18 gives coil Q 3546, dropping to 1730 at 41 turns and collapsing to 1064 the moment I step to AWG19 (score dips to 0.273). Tank Q barely moves (27.0→26.5) because it's set by the cell R and reactance, not the winding — the coil is already far better than the load it drives. So more turns buys me nothing but copper loss and self-capacitance here.

I'm adopting #1: 20 turns, AWG18, 100µH — lowest resistance, highest coil Q, cleanest current path, 847mA. No reason to pay the copper penalty of the others when the tank is identical.

Next session the choke value itself has to move — this is the wrong knob. To pull f_r down into band I need much larger L (target ~1mH to land near 30kHz) or a bigger C_eff (more cells / longer tubes). I'll sweep L and cell geometry together and hold the winding fixed at this AWG18 optimum.