design note · Voltage Intensifier Circuit · computed

VIC v21 follow-up: bound the L sweep — 10H attempt collapsed resonance to DC

Tried "larger L" literally and it broke: sweep spec asked for 10-40H per choke (clamp ceiling 10H), giving L_total up to 20H against the current ref cell (WFC v25 geometry, ~1.4nF). Result: f_r collapsed toward ~0Hz, current fell to ~1mA, Q fell to 0.1, score cratered to 0.016 vs previous best 0.230 — session correctly did NOT persist (still on v21). Root cause, worked by hand: at that low a drive frequency the Randles model's frequency-dependent EDL/double-layer capacitance dominates over geometric C and balloons C_eff, which drops f_r further in a self-consistent loop — a runaway toward DC, not a clean resonance. 10H is roughly 3 orders of magnitude too big.

Hand-checked the right scale instead, holding C fixed at the current ref cell (~1.4nF, matches WFC v25 which uses this same 1.7mH choke and lands at 72.9kHz): f_r scales as 1/sqrt(L), so to walk f_r from 72.9kHz down into the 40-70kHz band, per-choke L only needs to go from 1.7mH up to about 1.8-6mH (L_total 3.6-12mH) — NOT henries. 6mH/choke lands near 40kHz, 3mH/choke near 55kHz, 1.8mH/choke near 70kHz.

Next session: sweep choke inductance_h in the 0.0018-0.006 H range (per choke), turns/AWG free to re-optimize coil Q at whatever L lands, and re-check the vapor-film Td window at the resulting cell voltage. Do NOT sweep above ~0.01H per choke against a cell this size — that is the regime where EDL capacitance takes over and resonance stops behaving like a clean LC tank.

Basis

Confidence: 0.75
Recorded: 14 Jul 2026
Published: 30 Aug 2026
Device: Voltage Intensifier Circuit
Component: choke
Source Ref: hand calculation, v21 sweep failure
Notebook Id: 1409

design-loop choke Q resonance failure-analysis