calculation · Voltage Intensifier Circuit · computed
Scotchn rewind topology: series-connected radial bifilar pancakes (Tesla-style); self-C model and an SRF test to replace #3388/#3426 capacitance figures
- TOPOLOGY (Scotchn, Discord 2026-09-25) — each "pair" is two wires wound together radially, centre outward. End of A (outer) goes to start of B (centre); end of B goes to start of A of the next pair. It is all ONE series conductor inside one choke. So there is NO strand-to-strand capacitance across the two legs (the C_strand/4-across-cell concern does not apply). Within a pancake it is a Tesla-style series bifilar. Unresolved: "3 pairs per window, 5 face-to-face" implies 6 layers, which does not fit 3 bifilar pancakes. Confirm.
- MODEL (equal volts per turn, core-linked turns) — A turn k sits at k·v_t and B turn k at (n+k)·v_t. Every in-plane A–B gap therefore sees ≈V_p/2, where V_p is the voltage across one pancake. Energy ½·C_AB·(V_p/2)² gives each pancake ≈C_AB/4 across its own terminals. Adjacent pancakes wound the same way (outer→centre crossover) are offset by a uniform V_p across the face gap, so each gap is worth ≈C_ff. Whole choke with N_p pancakes: C_self ≈ [N_p·C_AB/4 + N_gaps·C_ff]/N_p². Example (C_AB = 50 pF, 18 pancakes, 12 in-window gaps, C_ff assumed 50 pF): 0.69 + 1.85 ≈ 2.5 pF. So a few pF, not nF. Summing per-pair C_AB in parallel (35 × 50 pF = 1.75 nF, #3426) is NOT the terminal self-capacitance for this topology.
- TEST — measure SRF with loose coupling (the scope probe's ~10 pF swamps a few pF). Illustrative with the old 28.61 mH choke: 2 pF → ~665 kHz; 1.75 nF → ~22.5 kHz.
Basis
- Method
- unversioned-legacy
- Recorded
- Published
- 25 Sep 2026
- Stanbot
- v3
- Source Ref
- Scotchn, Discord #neuralstan20-testing 2026-09-25 (winding topology); derived energy model; supersedes capacitance conclusions of #3388, #3426; extends #3428
- Notebook Id
- 3429
bifilar self-capacitance pancake tesla-bifilar srf-test