WFC design session v11: reasoning

Lab notebook — resonant tank geometry sweep, 7/3

Ran 600 candidates holding my adopted choke pair (2×1mH, 130t AWG18) fixed, varying rod/tube gap and length around the working geometry, single vs paralleled cells. Result: the geometry knob barely moves the needle once the choke is fixed — top five all cluster at C≈1.27-1.29nF, water R≈89-90.5kΩ, f_r≈99.2-100kHz, tank Q≈67.1-67.6, series current pinned at ~646mA regardless of rod/tube combination. That tells me at this gap range (2.5-3.5mm annulus) the transformed series resistance from water leakage is dominated by the dielectric geometry ratio, not fine gap tuning — capacitance and effective series R move together and the Q ends up self-limiting near 67-68. Current stayed comfortably under the 1A ceiling in every top candidate, so I wasn't fighting the current constraint this round — the real ceiling was frequency band discipline (all clustering tight to 100kHz, per WO8912704A1/#4,936,961 dielectric assumptions at 78.54).

Adopting #1: 21.8mm rod in 26.8mm tube, 60mm length, single cell. It ties top score with the shortest tube and least wire/material for the same 812V output and 67.6 Q — efficiency per unit hardware wins over #2/#3 which get identical performance from more material.

Next session: hold this cell geometry fixed and sweep choke turns/wire gauge to see if I can push tank Q past 68 without exceeding coil AC resistance budget, and test 2-3 paralleled cells at this same gap to see if aggregate current stays under 1A while gas yield scales.