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

Rotating Dynes

Meyer's last generator has no moving parts at all. It was built, it was photographed, and nobody wrote down what went inside the tube. This film tries to fill that in, and finds that the answer he chose is the one answer that cannot work.

The archive's own copy, 15 September 2026 56 MB

7 views

What was shown

Most of these films read a document and check it. This one tries to build something.

Meyer's Magnetic Spin Generator is the last machine in the Electrical Particle Generator line, and the strangest. The EPG pumps a magnetised fluid round a loop past pickup coils. The Magnetic Spin Generator throws the pump away. The medium sits still; two orientation coils at an air gap flip the particles ninety degrees; the field they carry swings from across the gap to along the tube and back, and the pickup coils see it move. His own drawing labels it: power input moves gas only.

It was built. The archive holds twelve photographs of it on a bench, a ring of non-magnetic tube wound with dozens of pickup coils.

What nobody wrote down is what went inside the tube. This film works that out.

The specification with the blanks in it

Meyer gives the geometry and the switching. He gives two options for the medium — a liquid slurry of micro-size magnetised particles in liquid Teflon or light oil, or magnetised particles carried on argon or nitrogen — and the patent follows the gas.

The size, loading, drive rate and coil count are usually said to be "not stated. Anywhere." Two of those four are stated. The EPG specification gives the flow (EPG #1: "Specification for EPG was 50 inches/second of movement for slurry. 90 inches/second for gaseous medium", with the Little Giant B-500 pump rated 500 GPH), and section Q of the very chapter cited winds the pump coil: "Coil (46) is wound (400 turns 30 ga. insulated wire) on molded nylon bobbin (47)". What is genuinely missing is the particle size and the loading, and those two are what this film is about.

Everything after this point comes out of the archive's own matrix engine: the Magnetic relaxation time, Stokes settling, Chain formation onset and Medium remanence and coercivity pages, with their material constants for iron, nickel and cobalt and their fluid constants for argon and light mineral oil.

The window

Two demands pull against each other.

A grain must be big enough not to tumble away from where the coils left it before the next pulse arrives. That is Brownian rotation, and it slows down as the grain grows. And it must be small enough not to fall out of a tube that nothing is pumping. That is Stokes settling, and it speeds up as the grain grows.

There is also a ceiling over both: above 60 nm a cobalt grain is many domains rather than one, they cancel, and its remanence collapses by a factor of three hundred. So take the best grain available, one right at that ceiling, and ask how long it keeps the direction the coils left it in:

holds its direction for
in argon at one atmosphere 2.3 µs
in light mineral oil 3.1 ms

Meyer's State of Quiescent — the magnetised gas sitting in the tube with the power off, holding its field across the air gap — lasts about two microseconds.

Settling is a second and separate failure. In argon, cobalt needs to be at least 483 nm to hold its direction for one half-cycle at 500 Hz, and at most 179 nm to stay up for an hour standing. Those two do not overlap. There is no particle size that works.

In light mineral oil, the lower edge falls to 40 nm and the upper edge rises to about seven micrometres. Cobalt goes multi-domain at 60 nm, so the working window is 40 to 60 nanometres of cobalt in light oil — which is a ferrofluid.

And the archive says that is what he actually ran. The easy conclusion is "he chose the gas, and the gas forgets everything in 2.3 microseconds", left there as the verdict. The physics stands; the claim about the man does not. EPG Speculations records a bottle of ferrofluid found at the estate sale — "Type is EFH1" — and identifies the green residue on the EPG's copper tubing as copper oleate, the residue of the oleic acid a ferrofluid is stabilised with. Jim Miner, who holds the photographs, told Bremen in 2019: "We have pretty good evidence that the EPG was originally thought to be using a ferrofluid... Stan actually picked the thinnest oil-based ferrofluid that had the most magnetic susceptibility of all the series that were available at the time."

He wrote the liquid down first and then bought the liquid. But this cuts both ways for a builder: Ferrotec EFH-1 is ~10 nm magnetite, far below the remanence window, so it has no permanent magnetisation at all. A machine filled with it is not a permanent magnet being carried round a loop; it is a transformer with a liquid core of relative permeability 2 to 3. That is an ordinary, measurable device, and it is not the one the drawing describes.

Which metal

He writes iron, nickel or cobalt, in that order, and that is usually called backwards, with iron "three thousand times worse". That comparison was made at cobalt's size, and it is withdrawn. Every metal has its own single-domain window: cobalt around 50 nm, nickel around 45, iron much lower and narrower at roughly 8 to 15 nm because it is magnetically soft. Inside its own window iron has the highest saturation of the three and gives 17,100 A/m at two per cent, above cobalt's 14,000. Metal-particle recording tape was iron for that reason, and spark ablation — the method his own figure 10 draws — makes primaries at 5 to 20 nm, which is iron's window and not cobalt's. His order is defensible; the line he never wrote is which size goes with which metal.

At 40 nm, which is the wrong size for iron and the right one for cobalt, the engine gives:

remanence the medium keeps
iron 4.8 A/m
nickel 4,800 A/m
cobalt 14,000 A/m

Iron is a thousand times worse than the metal he names last, because at 40 nm it is still many domains and it is magnetically soft. Cobalt's anisotropy is what buys the remanence.

A note on the filing

Everything here sits in this archive under Water Fuel Cell, and the device has nothing to do with water. That is not a cataloguing error. In one of the twelve photographs the words water fuel cell are printed on the base plate the ring is standing on. Meyer branded everything he built, and the filing followed the label.

He knew, and he tried to time it out

The best page in the file is section M of the Electrical Particle Generator chapter, headed EMF Counter Measures: Minimizing Power Input. He balances the pickup array's own field against a second ring, and switches the drive off the moment the reaction appears. Then he states the belief that makes the scheme look possible:

In the field of electronics it is known that an EMF field is "NOT" formed until after a magnetic field passes through a pickup coil.

That is the deepest mistake in the archive. The induced voltage is not a consequence that arrives afterwards; it is proportional to the rate of change, at the instant of the change. There is no after to switch away into, and going quicker makes the opposing force larger, not smaller.

The claim that can be tested

Meyer says the same thing three times across three documents: because the particles are permanently magnetised, transformer coupling between the orientation coils and the pickup coils cannot occur. The device's whole case rests on it.

Using the engine's own coil figures — a 0.338 mH drive coil switching five amps in a hundred microseconds, a 30 mH pickup — the breakthrough voltage at a coupling coefficient of 0.01, which is almost no coupling at all, is 1.6 V against a signal of 17.3 V. At 0.1 it is 15.9 V, and the two are the same size. The breakthrough also arrives at exactly the same instant as the signal, because both are caused by the coils switching, so no amount of gating separates them.

There is one experiment that settles it, and it takes an afternoon: run the machine with the tube empty. If the output does not change, the medium was never doing anything.

What he got right

The spark-gap generator of Electrical Generator Utilizing Magnetized Particles — vaporise the tips of two magnetisable electrodes, carry the particles up past a field generator to magnetise them — is spark ablation, and it is how magnetic nanoparticles are made commercially today. His count of iron's M shell at fourteen electrons is correct, and so is argon's outer eight. He diagnosed Lenz's law without ever naming it, called it the opposing magnetic field problem, and spent fifteen years trying to route around it.

The route does not exist. Twisting a magnetised medium inside a loaded pickup coil costs exactly what turning a rotor costs, because the induced current opposes the twist. Lenz moves from the shaft to the gas; he does not leave.

magnetic spin generator electrical particle generator magnetised gas ferrofluid WFC 418 EP0098897 matrix engine synthetic voice Evil Stan