epg: Lab Note — EPG Faraday, epg-faraday-v1

Lab Note — EPG Faraday, epg-faraday-v1
9/18/2026

Bumped best from 0.8879 to 0.8947 (score 0.895 run). η now 65.1%, net 79.72 W.

What changed: pushed slurry loading to φ=28.89% against documented 28.89%... actually the real lever was mismatching the load — 73.4Ω against a 39.27Ω coil, well off matched-load. Modulation depth held at 0.900 via the spiral divider, pitch 20mm on a 200mm loop, bore 88.5mm. Flow stayed slow and laminar (0.19 m/s, Re 2138), so hydraulic input is trivial at 33 mW.

Why the model rewards it: gross electrical output (122.4 W) is dominated by Lenz reaction, not flow work — the mechanical input term is basically the electrical reaction term (122.42 W) plus a negligible hydraulic/eddy contribution. So the model is telling me the "engine" here is coil geometry and load mismatch, not pumping harder. Coil loss (42.66 W, 39.27Ω across 3758 m AWG15) eats a third of gross — that's the real ceiling on η.

Bench checks: verify actual coil resistance and wire length match calc (3758 m, 39.27Ω) — small drift here swings I²R loss a lot. Confirm ferrofluid remanence/M at 32%/41.2 kA/m holds at working temp — EFH-1 datasheet numbers are lab-cond, not hot-running. Scope the 117 V/9.5 Hz output into the actual 73.4Ω to confirm no clipping.

Next lever: sweep load impedance toward the 39.27Ω match point — model predicts eddy and coil loss trade off differently there.