WFC design session v12: reasoning

Lab note, 7/3/26.

This sweep confirms what I suspected: at fixed geometry (21.8mm rod/26.8mm tube, 60mm) and 0.1ppm Na+ water at 5°C, the Randles double-layer parameters barely move the needle. Candidates #1-4 are identical electrically — same C_eff (1.39nF), same bulk R, same tank Q (91.2), same 915.8mA current, same 1094V output. Surface_factor sweeping from ×0.5 to ×4 only changes C_dl (3.8µF→30.1µF) without touching resonance, because at 96kHz the double-layer impedance is so low relative to the bulk/geometric branch that Cdl just isn't in the current path — displacement current stays pinned at 99.0%. So surface treatment, at least in this ppm/temp regime, is not a lever for Q or current headroom. That's useful negative information — it tells me the ionic branch is already dominated by bulk R, not electrode kinetics.

Comparing #1 to #5: going to 2 parallel cells drops Q (91.2→70.6) and f_r (96→74kHz) while current rises toward the 1A ceiling (923.6mA) despite better breakdown margin. Confirms the Q-vs-current-vs-cell-count trade I've logged before — cell count is a bigger lever than surface finish.

Adopting #1: best score, lowest current (915.8mA, more headroom under 1A cap), highest breakdown margin at single-cell simplicity, and surface ×0.5 means less electrode prep — cheapest build for the same performance as #2-4.

Next session: sweep ppm and temperature independently (not just surface_factor) to see if bulk R — the actual dominant term — can be pushed lower without wrecking breakdown margin.