Skip to content
Stan’s Legacy The Stanley Meyer Archive
Sign in

Amp Leakage

Meyer drew the right graph, in the right patent, with a number on every rung of it. What an impedance meter sees looking into his water capacitor, and what the interface at the electrode is actually doing.

The archive's own copy, 15 September 2026 75 MB

5 views

What was shown

Meyer's whole method rests on one sentence in Method For The Production Of A Fuel Gas: the water is processed as a dielectric medium in an electrical resonant circuit. Not an electrolyte. A dielectric. And so the object of the circuit, in his words, is to prevent electron flow through the circuit.

He drew the graph that governs that. It is Figure 8 of Gas Generator Voltage Control Circuit, the 1985 Gas Generator Voltage Control Circuit — not of the fuel gas patent, where it is usually credited. Applied voltage potential up the page, amp leakage along the bottom, and six levels, each one a circuit measure that let him hold a higher voltage before the current came.

The cell the patent specifies

Example I is not a sketch. It gives dimensions.
two concentric cylinders, long 4 inches
inner cylinder, outside diameter 0.5 inch
outer cylinder, outside diameter 0.75 inch
water gap, so the bore is 0.625 inch 0.0625 inch
capacitance, Coaxial cell capacitance 2,029 pF
resonance reached at 26 V to the toroid primary, 10 kHz

A bench note in the archive describes a shorter cell — 3 inches, with a 0.75 inch bore — which comes out at 837 pF. That is a different cell. Every number here is on the one in the patent.

The six levels, with the numbers on them

These rows are often reproduced as named, ruled, and left blank. They are not blank. The patent gives all of them in one sentence: “The same rise and fall in amplitude of the voltage versus current flow repeats at amplitude levels L-2, L-3, L-4, and L-5; again, in a given size apparatus, voltages of 4, 5.5, 7, and 8.5.”

level the arrangement volts
L-1 below the minimum, constant applied voltage 2.5
L-2 uni-polar pulsating DC 4.0
L-3 duty-cycle pulses on the pulse train 5.5
L-4 exciter plates connected serially 7.0
L-5 resistor 60, negative plate to ground 8.5
beyond L-5 the sandwich limiter of FIG. 7 the limit is removed

And “Infinity!” is not a meter floor. It is the patent's stated procedure for the FIG. 7 limiter: “A second resistor of the variable type is serially connected to the unique limiter for tuning … The variable is employed until the ammeter reads zero or close to zero as possible.” The sandwich is a laminated resistor in the earth lead, not an electrode material.

One number, doing two jobs

τ = ε/σ. The cell's dimensions cancel out of it entirely, so this is a property of the water and nothing else.

water τ the corner frequency
ultrapure, 0.055 µS/cm 129 µs 1.2 kHz
distilled, standing, 1 µS/cm 7.1 µs 22.4 kHz
the engine's default, 5 µS/cm 1.4 µs 112 kHz
tap, 500 µS/cm 14 ns 11.2 MHz

It is the cell's RC time constant and the time after a pulse edge before the field collapses off the gap and onto the two interfaces. The Debye length says where the field ends up; τ says when. His pulse at 10 kHz is 50 µs wide, so below about 0.4 µS/cm the whole gap is dielectric for the whole pulse — exactly as he describes it.

Can it ring?

A 0.125 H choke tunes this cell at his 10 kHz. The water damps the tank, so Q = R/ωL:

water Q at 10 kHz
ultrapure 8.1
distilled, standing 0.45
the engine's default 0.09

Below one there is no resonance to find. His insistence on distilled water, which reads like superstition, is the condition under which his own claim of resonance is arithmetically possible.

The staircase that does climb

The bulk cannot hold charge between pulses in any water you can keep. But the charge is not destroyed — it lands on the electrodes, and the electrodes are a 450 µF capacitor.
charge handed over per 100 V pulse 203 nC
pulses to put 1.23 V on the interface 2,730
at 10 kHz, that is 0.27 s

Which is what “from several millivolts to several hundred volts during each pulse train” looks like on a meter. He described it. It is happening three tenths of a nanometre from where he thought.

The fork
Faraday, mixed gas, per ampere 10.45 mL/min at STP
a blocking cell at 1,000 V, 10 kHz 10 W, and no gas
a conducting cell at 1,000 V 286 W, and Faraday's gas
Meyer's claimed rate 115 mL/min/A — Faraday × 11

Only gas measured against current can tell those apart, and nobody in this archive has posted both halves on the same line.

Six things usually got wrong about this page

  • Figure 8 is in Gas Generator Voltage Control Circuit, not Method For The Production Of A Fuel Gas.
  • The bench table has a number on every rung; the blank rows are a transcription, not a gap in his work.
  • “Infinity!” is the patent's stated tuning procedure, not a meter artefact.
  • The axes are the corrosion-science convention, and the curve is a load line rather than a Tafel plot.
  • The cell is Example I's — 4 inches with a 0.625 inch bore — not the bench note's.
  • He had words for two of the four Randles elements, and the fourth is negligible in his band.

electric double layer Helmholtz Randles circuit Nyquist Bode impedance water capacitor electron leakage matrix engine synthetic voice Evil Stan