Flynn-C: Physicist review (Step 2)

Reviewer role: PhD physicist, mainstream view.

Source gap (repeated from Step 1)

The only document is US 6,246,561 B1 (Flynn, filed 1998, granted 2001). It has no text layer. I re-read the description (pp.34-45) and claims (pp.46-49) by rendering pages at 200 dpi and OCR with tesseract [read: page image p.34-49, OCR]. The drawing sheets (pp.3-33) were not studied. The patent contains no measurements. Google Patents also notes no test data [web: https://patents.google.com/patent/US6246561B1/en]. No test reports, no motor performance data and no independent evaluations are in the folder. A web search for independent tests could not be run (search budget exhausted), so statements about the absence of independent tests rest on "none found in the folder or the patent page", not on a survey.

Corrections to Step 1

1. Drawing sheet count: facts.json says "32 drawing sheets" while claims.md says "31 drawing sheets (pp.3-33)". Pages 3-33 are 31 pages; the figures run Fig. 1 to Fig. 50 [read]. I treat it as 31 sheets, unverified.
2. Claim 3 wording in claims.md ("cuts the holding force on each armature to one quarter, not one half") is correct as a per-armature statement, but the "unexpected" framing hides that the TOTAL force over both armatures is one half (see Claim 3). Step 1 recorded the patent's claim without this caveat.
3. The "twice the coil energy" figure in claim 4 is the patent's own figure (p.39: "twice that delivered to coils 122, 124"). The Step 1 text calls it "coil energy"; the patent also states it as "one half the power" (p.37). It is a static comparison of coil input to hold a force, not an energy accounting over a motion cycle. Step 1 said this loosely; I make it explicit below.
4. Confirmed: the patent does not say output exceeds input [read]. Also new: current assignee QM Power Inc since 2008; patent expired [web: https://patents.google.com/patent/US6246561B1/en].
5. Everything else in facts.json matches what I read.

Background physics (for the reader)

A permanent magnet (PM) behaves like a magnetomotive force (MMF, the "magnetic voltage") in series with an internal reluctance (magnetic resistance). Flux Phi (webers) plays the role of current. A coil with N turns carrying current I adds MMF = N*I. Force pulling an iron armature onto a pole face is F = B^2 A / (2 mu0), where B is flux density (tesla), A is face area (m^2) and mu0 = 4 pi x 10^-7 H/m [read: p.34-36, the patent uses this formula].

Worked numbers: B = 0.50 T on A = 1 cm^2 = 1e-4 m^2: F = 0.25 * 1e-4 / (2 * 1.2566e-6) = 9.95 N per face. Halve B to 0.25 T: F = 2.49 N, one quarter. Quadruple force needs only double B.

Claim 1. Pulsed coils on two parallel paths switch essentially all the PM flux from one path to the other.

a. Mathematics: consistent. Treat the PM as MMF source Fm with reluctance Rm feeding two parallel gap paths of reluctance R. A coil in path 1 whose MMF cancels the magnet's MMF drop across that path forces path-1 flux to zero; the flux goes through path 2. If Rm is much larger than R (typical of real magnets with a small gap), total flux changes little, so path 2 carries about double its previous flux, as the patent says (p.37). If Rm is not dominant, the flux does not exactly double; the patent assumes constant magnet flux, a good first approximation only. b. Experiment: consistent. This is the working principle of polarised latching solenoids, hybrid stepper motors, "magnetic flux switching" holding devices and flux-switching PM machines; also the principle of switchable magnetic base chucks. [memory, high confidence on the general existence of these devices; no URL fetched]. c. Tested: the patent has no test data. Flux switching in general is routinely built and tested by many others. Whether Flynn's exact configurations were built and measured: unknown from the folder. d. If true: nothing changes. Established engineering. Verdict: Established (as a magnetic-circuit effect). The word "substantially all" needs the coil MMF to match the magnet's drop, which holds only for a well-designed circuit.

Claim 2. A small electromagnetic excitation controls a much larger PM flux.

a. Consistent but ambiguous. The patent's own figures give the coil MMF needed as equal to the magnet's MMF drop across the path, i.e. not "small": p.7 says it takes "one half the power it would require for a coil alone". The objects (p.34) say "a lesser level of electromagnetic flux". "Lesser" is true; "much smaller" is not supported by the patent's own numbers, which give a factor of two. b. Consistent with latching relays and hybrid actuators. In those, a small pulse can release or reverse a large holding force because the system is near a force balance (unstable equilibrium), not because the coil does the work. c. No test in the patent; generic devices tested widely. d. Nothing; no violation of energy conservation, since the work comes from the supply plus the stored energy changes in the gap. Verdict: Established in the weak form (a control pulse can release or redirect a large PM force near a balance point). The strong form ("much larger" with tiny excitation, in continuous operation) is Plausible, untested for the specific designs; the patent's own number is a factor of 2.

Claim 3. Splitting a given flux over two equal paths cuts the holding force on each armature to one quarter, not one half.

a. Arithmetic correct: F proportional to B^2, B halves, so F/4 [read: p.35-36]. But it assumes constant total flux, uniform B, and equal areas. The "unexpected" label is a framing trick: there are two armatures, so the total holding force over both is 2 * F/4 = F/2. A single armature does see one quarter. It is not unexpected to any magnetics engineer: force is quadratic in B. b. Consistent with the Maxwell stress F/A = B^2/(2 mu0) (well-tested, e.g. electromagnets lift-force formulas). Limits: real flux is not constant when the circuit changes, because the PM operating point moves. c. Tested: not in the patent, but the B^2 law is textbook and routinely checked. d. Nothing. Verdict: Established (per armature). The "unexpected" description is misleading, since total force is halved.

Claim 4. Switching PM flux gives more coupling force and work per unit of input electrical energy than an all-electromagnet equivalent (comparison device needs twice the coil energy).

a. The force comparison is internally consistent as a static copper-loss comparison: the PM supplies half of the MMF needed to reach the same B in the armature, so the coil needs half the MMF; since coil loss goes as I^2 R, the loss for the same coil could fall by up to 4x, not 2x, so "twice" is conservative or counts differently (p.39: not derived in the text). The step from "greater force per unit of input energy" to "more work per unit of input energy" is NOT justified. Force at one position is not work. Work is the integral of F dx over the stroke, and over a closed cycle of a conservative PM, net work out = electrical energy in minus losses. A PM contributes stored field energy that can shift but not net energy per cycle. I computed nothing beyond the above; the patent gives no cycle accounting. b. Consistent with experience that PM-assisted (hybrid) actuators use less coil power than pure electromagnets for a given holding force [memory, medium confidence; not web-verified here]. No result shows net work per cycle above input. c. No test in the patent. No independent test result in the folder. d. If read as "more work out than electrical energy in over a cycle", conservation of energy would have to fail. As the patent states it (a ratio to an electromagnet design, not to input), nothing changes. Verdict (split): Weak form (less coil power than an electromagnet-only design for the same force): Plausible, untested for these designs; physically reasonable. Strong form (net energy gain, which the patent does not assert): Refuted by calculation (energy conservation for a conservative magnet in a cycle). Statement "more work per unit input" as written: Not testable as stated (no cycle, stroke, or loss definition).

Claim 5. Motors built on this are more efficient and have better torque than many conventional motors.

a. No mathematics given; no efficiency number, torque curve, or loss budget. A comparison to "many currently used motors" is not a defined quantity. b. Consistent in principle with flux-switching/hybrid PM machines reaching high efficiency; also consistent with standard motor limits (copper loss, iron loss, friction). Nothing in the patent contradicts them. c. No tests in the patent [read]. No data in the folder. Independent tests not found here. d. Nothing, unless efficiency above 100% were claimed, which it is not. Verdict: Never properly tested (no neutral controlled test on record in the folder; not shown to be dismissed). As worded, partly Not testable as stated ("many currently used motors").

Claim 6. With fixed armatures and secondary coils, switching converts static magnet flux to time-varying flux, giving transformer or inverter-like conversion.

a. Consistent. Faraday's law: EMF = -N dPhi/dt. If switching changes the flux linking a secondary coil between Phi0 and 2*Phi0 (p.44: "varies by a factor of twice"), a voltage appears. This is a switched, magnet-biased transformer (like a fluxgate or magnetic-modulator principle). The patent's statement that static PM flux is "useless" until time-varied is true as far as induction goes. b. Consistent with Faraday induction and magnetic modulators. c. No test in the patent. No data. d. Nothing, provided the energy in the secondaries comes from the control-coil supply. The control coils are the "primary" (the patent says this on p.43); the PM only biases the core. The PM adds no power. A reader who sees "the flux that would normally be supplied by a primary winding is supplied by the static flux of the permanent magnet" (p.44) might infer a free primary; it is not. The secondary power must be supplied by the control circuit, less losses; the PM does no net work. Verdict (split): Principle (flux modulation induces EMF): Established. Implied claim that this is a more efficient converter or that the magnet supplies the power: Refuted by calculation (energy conservation; power out <= control power in, minus losses). Practical efficiency: Never properly tested.

Overall

Most of the patent is ordinary, correct magnetic-circuit engineering written up as new. Its weak points are rhetorical, not mathematical: (i) a per-armature statistic called "unexpected", (ii) a static force-per-energy ratio presented as "more work per unit of energy", (iii) uncontroversial "efficiency" claims with no data. Over-unity is not claimed in the patent as read [read]; any such use comes from outside it.

Source tags: [read: page image OCR p.34-49]; [web: https://patents.google.com/patent/US6246561B1/en] for status, assignee and absence of data; [memory: medium-high] for existence of latching/hybrid actuators and flux-switching machines (not web-verified; search budget exhausted).