epg: Lab Notebook — EPG, epg-faraday-v1

Lab Notebook — EPG, epg-faraday-v1
9/17

Pushed the Red Pump config from 0.6546 to 0.8804. Big jump, worth writing down carefully before I trust it.

What changed: bumped ferrofluid loading to φ=30.21%, nearly 4x what Ferrotec lists for EFH-1 (30.21/7.9 baseline). That drags remanence to 29% and M up to 38.4 kA/m, giving B=363 G at the coil face. Kept the 9x979-turn AWG14 array (28.45Ω) and conductive tube wall unchanged. Flow stayed slow — 0.25 m/s, Re 3730, so we're transitional not turbulent, ΔP only 5 Pa.

Why the model rewards it: net output is dominated by the Lenz/electrical term (407.6 W gross, 259.8 W net after 147.8 W I²R loss), while hydraulic input is trivial at 0.139 W. The model is basically converting magnetic loading into voltage almost for free — 178.83 V at 12.4 Hz into a 50Ω load that's not even matched to source Z (28.45Ω). Eddy loss in the tube wall is negligible (5.8e-8 W), so nothing in the model currently taxes the fluid concentration.

Bench check: verify EFH-1 can physically sustain 30%+ loading without settling/clumping — that's outside the datasheet range, so treat M and B as unverified until measured directly (Hall probe at coil bore). Also confirm 12.4 Hz output frequency against actual pump RPM, and load-match to 28.45Ω to see if real efficiency tracks the modeled 63.7%.

Next lever: try VIC impedance matching (drop load to ~28Ω) before pushing φ further — cheaper test than chasing exotic ferrofluid concentrations.