Session close 2026-09-30: pin headers tested good (the clips were too hefty for them). No A–B fault found anywhere; the #3445–#3455 anomalies remain unexplained and retracted (#3459) until Scotchn's repeat next week

Scotchn, Discord #neuralstan20-testing 2026-09-30 — he tested the pin headers on either side of the black plastic separator and they are good. His LCR clips were too hefty for the header pins, which explains the earlier false 3.6 Ω. The header-leakage candidate (#3461) is cleared at DMM level.
STATE AT CLOSE — no winding short, no header path, no core DC path (varnish + Kapton). Voltage-dependent leakage at operating voltage is untested. Still unexplained: start leads pinned together (u ≈ 0), a diode/cap split that ignored C_t, and a ms-scale shunt tail. All of it stays retracted (#3459) until repeated.
PLAN FOR THE REPEAT (early next week) — (1) DMM A↔B with everything lifted; expect OL. (2) The As→Bs differential capture first, same drive (4 V, 20 kHz 50 %, 1 kΩ, C_t 3.97 nF). (3) If u is still pinned, repeat the 2 nF vs 3.97 nF split and the diode shunt with the jumper test. Checklist: Math sources, channel Invert, probe lead directions on the drawing, every capture labelled.

Basis

scotchn bifilar session-close open-question