Rankings Primer Arbiter Physicist Mathematician Contrarian QE Claims Sources

Gray-EV: Physicist review (PhD physicist, mainstream view)

Source gap (repeated from Step 1). The four patents are scanned images with no text layer. I read the image of US 3,890,548 p.9 (cols 1-2, background) and p.17 (claims 1-10), and GB 2,030,801 p.5 (specification and claims) myself. Description pages of US 3,890,548 and nearly all of US 4,595,975 / US 4,661,747 were not read. There is no independent test report, no waveform, and no instrument list anywhere in the folder. Several conclusions below rest on [memory] alone or on the promoter's own letter; these are flagged.

Web check: the session's web-search budget was exhausted, so no [web] verification was possible. Every [memory] item is marked "web check failed" and I did not let any verdict depend on one.

Corrections to Step 1

I rendered crosby_letter.pdf and read the page image [read: page image p.1]. The Step 1 OCR reconstruction was wrong in several places:

The letter does NOT say "exceeds 100%". It says "The measured overall system efficiency exceeds 99%". Step 1 (claims.md item 2 and the crosby_letter summary) said it exceeded 100 percent. The ">100%" over-unity reading comes from the tabloid press, not from this letter.
The consumption figure is 26.8 W per hour, not 26.9. The letter reads "was only 9.75 watts; this equals 26.8 watts per hour". (The TATTLER reprint gives 26.9.) Step 1's 26.9 is the press figure.
Signer: "J. A. Maize, Director of Engineering" (Step 1 had "A. Maize"). The letter carries a notary seal (Carol L. Dye, 9/28/73). A notary only attests the signature, not the measurements.
The letter states test dates May 1-10 1973 and a confirmation on Sep 25 1973, both by Pan World Enterprises and "EvGray engineers", that is, the inventor's own side. It also says "Tests reports supporting the above information is available at EvGray", so the underlying data are not in the letter.
Step 1 said the 5,454 Wh was an OCR garble of "51454". Confirmed: it reads 5,454 watts for one hour (5,454 Wh).
The letter says "exceeds 99%" while also giving numbers that imply an efficiency of ~27,800%. Step 1 did not note that the letter contradicts itself (see Claim 2).
Patent US 3,890,548 text (read): the background says the engine is "believed to operate on the principle of conservation of energy". It also says the new engine "produces torque and stores the excess energy for subsequent use". That is ambiguous wording, not an over-unity claim. Step 1's summary ("does not claim over-unity") stands.

Minor: US 3,890,548 claim 1 (read) requires the spark-gap capacitor discharge and a charging source that "continues charging said capacitor means following each discharge". It is a conventional pulsed-power circuit claim.

Primer physics used below

Energy stored in a capacitor: E = (1/2) C V^2. Energy in an inductor: E = (1/2) L I^2.
Mean power = (energy per pulse) x (pulse rate). A meter that reads only the average of a DC level, or a meter whose bandwidth misses fast spikes, can read far too low if the input current is a short, high-current pulse train.
Mechanical power: P = torque x angular speed.

Claim 1. Pulsed capacitor-discharge motor (US 3,890,548) drawing current only at each discharge

a. Mathematics consistent? Yes, as far as it goes. The patent gives no equations. Check of the dynamometer figure from the letter: 48 ft-lb = 65.1 N m; 1,100 rpm = 115.2 rad/s; P = 65.1 x 115.2 = 7,497 W = 10.05 hp, consistent with "10 hp = 7,460 W" [read: page image p.1 of the letter, with my own calculation]. For the patent mechanism: a capacitor C at V stores (1/2)CV^2; discharge through a coil gives an impulse on the rotor. Nothing in the described circuit is mathematically inconsistent. The statement "current is needed only at the instant of a capacitor discharge" [read: US 3,890,548 col.1] is true of the circuit, but the energy per pulse still has to be supplied by the source, then multiplied by the rate of pulses.

b. Consistent with experiment? Pulsed-capacitor drive of coils is conventional (capacitor-discharge solenoids, spark-gap pulse power, CD ignition, pulsed electromagnets). Same-polarity repelling cores are conventional too. A working low-power version is physically plausible. [memory, medium: standard pulse-power practice; web check failed.] The patent does not claim a measured efficiency, so nothing in it conflicts with known measurements.

c. Tested? The 1973 Crosby letter (promoter-side) and a TRW scientist's remark in the AP item (Cronin saw a ~10 s run that "wasn't convincing" [read: Step 1 summary of various_articles.pdf, garbled OCR; not independently re-read by me]). No independent, instrumented test of the original spark-gap motor on record in the folder. No replication known to me; I cannot cite any [web check failed].

d. What would change? Nothing, for the mechanism as a plain motor. Mainstream physics already includes it.

Verdict: Plausible, untested for the claim as a working motor. Established (as pulse-power practice) that such a circuit can produce torque; Never properly tested for any performance figure (power, efficiency, sustained operation).

Claim 2. 10 hp output at 1,100 rpm for about 26.9 W (letter: 26.8 W) from batteries; "efficiency exceeds 100%"; 203 hours

Step 1 corrected: the letter says "exceeds 99%", not 100%. The press (TATTLER) gives the 26.9 W figure.

a. Math internally consistent? No. The letter contradicts itself.

Claimed output 7,460 W. Claimed input 26.8 W. Ratio 7,460 / 26.8 = 278, i.e. 27,800% "efficiency", which is incompatible with the letter's own statement "exceeds 99%" (a 99% efficiency would mean about 7,540 W input).
9.75 W (the letter says "watts"; it can only be watt-hours) consumed in a 21 min run: 9.75 Wh / (21/60 h) = 27.9 W average, not 26.8 W (a 4% mismatch; the letter's "26.8 watts per hour" is also a unit error; W per hour is not a power).
5,454 Wh / 26.8 W = 203.5 h, matching "203 hours".
But the same 5,454 Wh at the claimed 7,460 W output lasts 5,454 / 7,460 = 0.73 h = 44 min if there were no gain. So "operate continuously for 203 hours at 10 hp" requires an energy output of 7.46 kW x 203 h = 1,514 kWh from a 5.45 kWh battery pack, a factor of 278.
A 21-minute run does not demonstrate a 203-hour run; the 203 h is an extrapolation.

b. Consistent with experiments? No. Conservation of energy has been tested to extreme precision in electromechanical systems. A device with 278 times more mechanical energy out than electrochemical energy in, with no external generator, would contradict every measured energy balance of motors and batteries. The "electrostrictive power" the letter invokes is not an energy source: electrostriction is the (quadratic) strain of a dielectric in an electric field, which is a mechanism of energy conversion, not generation. [memory, high; web check failed, not needed for the verdict.]

c. Tested? The only record is the promoter-side letter and its press reprints. No instrument list, no waveforms, no description of how pulsed input power was measured, and "test reports ... available at EvGray" were not provided. An independent look by TRW's Cronin (AP, 1975) saw about 10 s of running, not convincing [Step 1 summary of garbled OCR]. No replication known. I cannot verify that any other lab ran a controlled test [web check failed].

d. What would change? Conservation of energy (first law of thermodynamics) and the electrochemistry of batteries would need to fail for this circuit class; the Noether link between time-translation symmetry and energy conservation would be broken. Prior probability of that is far below that of a measurement artefact.

Likely measurement error (mundane explanation, [memory], medium; web check failed): for a pulsed supply, input power is the time average of v(t) i(t). A DC ammeter or averaging voltmeter that integrates poorly, or measurement of battery current at a point where the capacitor bank supplies the spikes, would read low. Also, the battery's recovered "back EMF" charge may be counted as free. Also, the efficiency was perhaps computed from a short run in which capacitor energy stored beforehand was spent. A proper test needs a battery energy balance (before and after, weighed in watt-hours with a constant-load discharge reference) over hours.

Verdict: the over-unity claim as stated (278x or ">100%") is Refuted by calculation (energy budget: 5.45 kWh cannot deliver 1,514 kWh) AND the letter is Internally inconsistent (">99%" vs 27,800%; W per hour unit error; 27.9 vs 26.8 W). A measurement of the true performance of the device itself has Never been properly tested by a neutral party.

Claim 3. Back-EMF of collapsing coil fields can be captured and returned to a battery or capacitor, recycling most of the energy (GB 2,030,801; US 4,595,975; US 4,661,747)

a. Math consistent? I read GB 2,030,801 p.5 [read: page image]. Phase B: capacitor discharges through the inductive load into a second battery; at zero capacitor voltage the load current would overshoot (reverse-charge); a diode lets the current continue in a closed loop through load and recovery battery. This is circuit-consistent: the energy left in the inductor, (1/2) L I^2, goes into the battery (or into the capacitor in the later patents). The claim itself is only "substantially as hereinbefore described", so no quantitative claim is made. The patent says "according to the preferred embodiment ... this is an energy conserving system", and the inventor "believes" it is "a more efficient use of energy". Not over-unity.

Worked example: L = 10 mH, I = 20 A. Energy left in the coil = 0.5 x 0.01 x 400 = 2 J. If a capacitor charged to 2000 V had C = 10 uF, its stored energy is 0.5 x 1e-5 x 4e6 = 20 J. So only a fraction (here 10%) is left in the magnetic field at the moment recovery starts; the rest was dissipated in the load and its resistance or became mechanical work. What can be recovered is bounded by the energy left in the field, and then recovery efficiency is limited by diode and battery losses (battery charging efficiency ~70-90%) [numbers illustrative, my own choices, not from the source].

b. Consistent with experiment? Yes. Flyback and freewheeling diodes, regenerative braking, and energy-recovery snubbers are standard power electronics, and are used in switched-reluctance drives. [memory, high; web check failed.] The recovery is real but returns only the residual stored energy, never more than was put in.

c. Tested? Standard technique, extensively tested in industry. Gray's particular circuits: no test report in the folder. "Most of the energy is recycled" has no measurement behind it (the patents give no data).

d. What would change? Nothing for ordinary recovery. If "recycled" is read as a net surplus, the first law would fail (see Claim 2).

Verdict (split): weak form (a diode path can return the inductor's residual energy to a source) Established. Strong form (most of the total energy recycled so the motor needs much less than conventional drive; net gain) Never properly tested for actual numbers; if it implies net gain, Refuted by calculation (energy budget).

Claim 4. "Cold", partly static electricity (and possibly atmospheric static) rather than ordinary battery current, so the motor runs cool and produces more than it consumes

Source: promoter and press only (ems_electronic_power, various_articles); not in the patent text [read per Step 1; I did not re-read these articles in detail].

a. Math consistent? No definition or equation exists. "Cold electricity" is not a quantity. Static electricity (charge on conductors, voltage without sustained current) is real, but power requires current times voltage; sustained power from "static" needs a continuing charge source. Not defined, so no checkable mathematics.

b. Consistent with experiment? The motor running cool is expected for any low average-power pulsed system (Joule heating is I^2 R averaged); the demonstration where a magnet flew off and was caught in the hand shows a short impulse, not free energy. Atmospheric electricity is a real but tiny power source: the fair-weather field is ~100 V/m with a current density of order 1e-12 A/m^2 [memory, medium; web check failed], so the harvestable power from a car-sized collector is far below watts.

c. Tested? No test of "cold electricity" as a distinct phenomenon; no measurement of any different kind of electrical energy is on record in the folder. Not replicated.

d. What would change? One would need a new energy source or a new form of electric energy not described by Maxwell's equations and the first law.

Verdict: Not testable as stated ("cold electricity" has no measurable definition). The atmospheric-static-power part is Refuted by calculation for the power needed (order-of-magnitude, from the fair-weather field and current density above; [memory, medium]).

Claim 5. The 1974-76 LA District Attorney's raid and prosecution were meant to suppress the invention rather than being a legitimate investor-fraud case

a. Math consistent? Not a mathematical claim (not applicable).

b. Consistent with evidence? Per the folder's articles (Step 1 summary of garbled OCR; I did not re-read them), the AP item records a deputy DA saying the device "would violate all the laws of physics" if it worked, and the charges (eight grand theft counts and others) concern investors; the outcome was a plea to two minor securities counts (1976, per the magazine feature). Gray raised about $2 million from 788 stockholders per the same feature [read per Step 1]. That pattern (investors' complaints, securities counts) is consistent with ordinary securities enforcement. The suppression motive is an allegation by a tabloid and the inventor; I found nothing in the folder that documents suppression intent (no internal DA documents, no testimony). [memory, low: general history; web check failed.]

c. Tested? There was no neutral inquiry into motive that I can cite. The DA also apparently did not have the device independently tested; whether the raid affected the technology cannot be determined from the folder. Court/DA records (named in the Source gap) are needed.

d. What would change? Nothing in physics. Historiography only.

Verdict: Not testable as stated from the sources held (motive claim; the folder has only allegations and a plea). The weaker factual part (raid, charges, plea) is attested only by press in the folder; it is Never properly tested in the sense that the suppression motive has no documentary support here. It does not bear on whether the motor works: a prosecution of the promoter says nothing either way about the physics.

Overall physicist summary

A pulsed capacitor-discharge repulsion motor with back-EMF recovery is ordinary, plausible engineering that gives efficiency improvements at most modest ones.
The headline "10 hp for 26.8 W" is excluded by an energy budget, is contradicted within its own letter, and is most plausibly a measurement artefact of pulsed power, though no raw data exist to prove which.
The single most decisive test: run the motor on a calibrated load for several hours while measuring battery energy in and out (calibrated, fast power analyser on the pulsed input, battery watt-hour discharge before/after).