JLN Labs MEG Replication - Detailed Experiment Logs & Anomalous COP Benchmarking

I am trying to replicate Bearden's MEG based on Jean-Louis Naudin's v3.0 logs. I ordered a Vitroperm 500F nanocrystalline core. The hardest part is cutting the air gaps without fracturing the core. What tool is everyone using for precision cutting to replicate JLN's setup?

Do not use standard hacksaws or rotary tools, they will shatter the ribbon layers of Vitroperm. You must use a diamond-coated wire saw or wire EDM (Electrical Discharge Machining) under a steady cooling water drip. Secure the core tightly in a jig before cutting. Once cut, sand the edges smooth and coat them with resin to prevent oxidation. This is exactly how Naudin achieved the clean flux transitions.

I finished my replication. Using a Tektronix TDS3024B oscilloscope and a true RMS wattmeter to measure input and output. Naudin pointed out that standard digital meters get confused by the high-frequency switching spikes (35kHz). My initial test shows a COP of 1.45, with input power at 12.4W and output load resistor dissipation at 18.0W. I verified this calorimetrically in a water bath—the heat output is real!

Also remember that the primary coils need to act as fast flux switches. Keep their inductance as low as possible. You are not trying to push magnetic field through the core, you are using the coils to momentarily alter the path of the permanent magnet's field. Timing is critical; start with 35kHz and adjust from there.