How to Test an Over-Unity Energy Claim

1. Write the claim in units

A testable claim is not "it powers itself." It is a statement such as: with all inputs measured at the boundary, the device delivered 620 watt-hours of electrical work to a resistive load during a six-hour run while consuming 410 watt-hours from the supply and ending with no decrease in stored energy.

Use joules or watt-hours for energy and watts for power. Report voltage, current, power factor, frequency, waveform shape, temperature, mass flow, pressure, and chemical products as relevant. The meaning of free energy becomes less ambiguous when the claim is written in units.

2. Draw the system boundary before the test

A boundary diagram should show every crossing: electrical leads, shafts, belts, gas lines, water lines, RF coupling, optical input, acoustic input, thermal contact, mechanical supports, and control electronics. It should also identify stored energy inside the boundary before and after the run.

If the device has a battery, capacitor bank, flywheel, magnet assembly, hot reservoir, chemical reagent, pressure vessel, or electrolyzer, measure its state before and after. A device can look over-unity for minutes or hours while silently spending stored energy.

3. Use instruments matched to the waveform

Many demonstrations fail because they use meters outside their valid conditions. Pulsed loads, non-sinusoidal waveforms, high-frequency switching, reactive power, ground loops, and probe placement can all produce false readings. The instrument must be calibrated for the signal being measured.

For electrical output, prefer direct calorimetry into a known resistive load or a power analyzer that reports true power under the actual waveform. For thermal output, measure mass, temperature rise, heat loss, phase change, and calibration runs. For chemical output, account for all reactants and products by mass and energy content.

4. Publish the uncertainty budget

A claimed 3 percent excess is not meaningful if current, voltage, temperature, and timing uncertainty combine to 8 percent. State each measurement uncertainty, how it was estimated, and how it propagates to the final energy balance.

The uncertainty budget should be written before the result is interpreted. If the claimed surplus is not larger than the uncertainty interval, the honest conclusion is inconclusive rather than positive.

5. Require independent replication

Independent replication is not an insult to inventors. It is how a claim leaves the demonstration stage. Replication should use a separately built apparatus, separately calibrated instruments, and raw data available for outside review.

The replication standard should be higher when the claim is broader. A new solar cell efficiency record needs careful lab validation. A device that would overturn thermodynamics needs multiple independent labs and a complete accounting of environmental coupling.

• Publish a boundary diagram and complete parts list.
• Record raw instrument logs, not only edited video.
• Use loads that cannot feed energy back into the device unless that path is measured.
• Run long enough to exhaust plausible stored-energy explanations.
• Let a skeptical lab choose and place the instruments.