epg: Lab note — EPG-faraday-v1, 9/27

Lab note — EPG-faraday-v1, 9/27

Best score moved 0.9416 → 0.9483 (η 59.4%, 119.91 W net to a 50Ω load, 130.36 V at 12.4 Hz). Change was pushing the slurry loading to φ=32.88%, which is 4.16× what Ferrotec lists for EFH-1 — that's the part I don't trust yet, flagging it clearly as modeled, not measured. At that loading the model gives M 46.9 kA/m and B 443 G, which raises modulation depth to 0.900 and that's what's driving the voltage up. Coil I²R stays fixed at 81.97 W off 34.18Ω of AWG15, so all the gain is coming from field strength, not from squeezing more current through the coils.

Physics reward is straightforward: eddy loss in the conductive tube wall is negligible (3.5e-8 W), hydraulic input is tiny (0.067 W), so the model is basically crediting the whole gain to the Lenz reaction term (201.9 W) scaling with B². That's a red flag for a bench check — real ferrofluids saturate well below 4× nominal loading, so I expect actual M to plateau far short of 46.9 kA/m. Bench priority: measure B directly at the coil bore with a Hall probe at the claimed loading before trusting the voltage number.

Next lever: hold loading at documented spec (φ~7.9%) and instead try tightening spiral pitch to 15mm to recover modulation depth without leaning on the unverified ferrofluid assumption.