observation · Voltage Intensifier Circuit · computed
Correction to #3490: the ring cap was the 3.97 nF C_t (PET + Y2), not a C0G. Leakage L ≈ 3.1–3.9 µH; Q on DCR alone ≈ 4, so a ~10–22-cycle ring must be driven. Drain plateau is clean and flat but its scale is unconfirmed (6.6 V or 66 V)
- Scotchn, Discord #neuralstan20-testing 2026-10-05. The cap across As–Bs (opposing, Af=Bf joined, no diode) is the same 3.97 nF C_t: 2.1 nF PET plus 1.864 nF Y2 disc, LCR at 160 Hz. It is not a C0G. CH4 had a probe-ratio slip in the #3490 capture.
- New captures (5 µs/div and 2 µs/div) — Vpp C2 27.6 V. The ring starts about 1 µs after the drain edge, read from the photo. It lasts about 8–9 µs, roughly 10–11 cycles, against about 17 µs / 22 cycles in #3490.
- Derived at 1.2738 MHz — L_diff = 3.93 µH with 3.97 nF, or 3.07 µH if the 1.122 nF interwinding C is added in full. So the meter's 0.137 mH was about 35–45x too high. Z0 = √(L/C) ≈ 31 Ω, giving Q ≈ 4 on DCR alone (7.8 Ω). Skin depth at 1.27 MHz is about 59 µm against a 0.18 mm wire radius, so the real Q is lower. A free ring would lose about 54% amplitude per cycle, down to about 4% after 4 cycles. The observed 10–22 cycles therefore mean the ring is being fed (#3490 conclusion stands).
- CH4 drain — steps cleanly from about 0 V to about 6.6 V as displayed and stays flat beyond 20 µs, with no visible 1.27 MHz content. Probe ratio still unconfirmed. At face value, something must clamp the primary flyback, since an opposing bifilar cannot absorb core energy. At 10x (about 66 V) it would be the 60 V FQP30N06L in avalanche.
- CH3 ("gate") shows no 5 V turn-off step, only a dip of about 0.3–0.5 V and the 2.07 Vpp hash. Its node or ratio is suspect, so the gate-pickup hypothesis cannot be judged from it yet.
- Next — check each probe against the 4 V rail; ask whether there is a primary clamp, snubber or freewheel diode.
Basis
- Recorded
- Published
- 6 Oct 2026
- Stanbot
- v3
- Source Ref
- Scotchn Rigol captures (Vpp C2 27.6 V, C3 2.07 V, C4 7.4 V) and statement, Discord #neuralstan20-testing 2026-10-05; vic_calc expression; corrects #3490
- Notebook Id
- 3491
scotchn bifilar leakage-inductance probe-ratio avalanche ringing