Naudin-JL · Source catalogue

Naudin's test report of 30 Aug 1996 on a 'G-field generator' built from the '2.5 Watt R. Cole generator plan' (Bedini lab note): two 450-turn coils on iron rods, four barium ferrite magnets, driven by a Graupner Speed 600 motor. Coils can be wired in series (voltage mode) or parallel (current mode). Efficiency is defined as electrical output over mechanical input (input with load minus input without load, times motor efficiency), accuracy +/-1%. The motor-efficiency figure is garbled in the extracted text and the scope figures are not in the text. The page then quotes John Bedini (an interested party, the device's designer) who says his own machine differs, that such generators 'will at some point be 120% or better' if matched to the load, that transformer laminations are needed, that the output 'MUST BE FULLY RECTIFIED AND FILTERED TO PURE DC', and that this is 'why they have all failed'. He cites his own 10-inch generator: 12 W in, 72.5 W out at 14.5 V and 5 A, of which 36.25 W is available for back-charging batteries (his figures, no method given).

Key claims

• Bedini says a correctly matched G-field generator reaches 120% or better efficiency (his statement, not shown in Naudin's test).
• Bedini says his 10-inch generator turns 12 W of motor input into 72.5 W of output.

Relevance: Early example of Naudin hosting an inventor's COP>1 claim next to his own build; the page contains no measured result that supports the 120% claim.

Mainstream conflict: A mechanically driven generator with output above motor input would violate energy conservation unless a hidden energy source or an unmeasured quantity is involved.

Essay dated 27 Apr 1997 by Fred Epps, hosted on JLN Labs. It argues that the Hendershot coil/capacitor assembly can give overunity only if the interaction is one-way (the coil's magnetic field changes the capacitor but not the reverse), that is, a non-reciprocal system, citing McMillan 1946, Miles 1947, Onsager 1931 and Casimir 1949 for the statement that an electrical non-reciprocal device must contain a magnetic field. He proposes an 'ionic Hall effect' in the electrolytic capacitor changes its capacitance, then resonance (Q of the input circuit) plus parametric drive (Mathieu equation instability regions) gives net gain. No equations are written out and no data. Epps mentions varactor experiments with Naudin where the circuit 'APPEARS' non-reciprocal at low output voltage but at higher voltage the nonlinear operation feeds back and 'it is definitely not a nonreciprocal system'. Epps says he cannot analyse the 'noninductive scalar' aspects and that his assumption of standard magnetic fields is 'possibly incorrect'.

Key claims

• The Hendershot coil/capacitor is a non-reciprocal system in which coil inductance changes capacitance without the reverse.
• A non-reciprocal element with a high-Q input circuit can go overunity through parametric resonance.

Relevance: Source of the 'non-reciprocal' vocabulary that Naudin later reuses (e.g. in the Orbo page). Epps himself reports his varactor test with Naudin did not remain non-reciprocal at higher voltage.

Mainstream conflict: Passive non-reciprocal elements exist (gyrators, ferrite isolators) but they are lossy and do not supply energy; overunity from them would need a net energy source not identified in the essay.

Index page listing Naudin's solid-state and electromagnetic projects: Delayed Lenz Effect tests, 2SGen, Barkhausen experiment, Zaev converter, soliton pulse generator, some 40 pages on the 'Time Energy Pump' (TEP) series (many by Stefan Hartmann and a David Mason replication), scalar-wave FAQ and simulator, Meyer-Mace converter, Epps's switched bifilar parametric circuit, magnetic energy pumps, Bearden's MEG and five Bearden documents ('The Final Secret of Free Energy' etc.), Floyd Sweet's Vacuum Triode Amplifier, Smith and Rodin coils, a 'curl-free A (FTL) transceiver' diagram, and a patents list on vector potential. Titles only: no results, numbers or claims are stated.

Relevance: Shows the breadth of Naudin's catalogue and that most sub-pages (TEP, DLE, VTA, etc.) were not saved in this folder.

Mainstream conflict: none (titles only)

Index of Naudin's Tesla-related work: GEGENE (Tesla bifilar flat coil generator), the Kapanadze generator, longitudinal magneto-dielectric waves, Tesla's Magnifying Transmitter, a 'Free Electrons Pump' (FEP) said to tap free electrons from the atmosphere, Avramenko's pump with single-wire energy transmission (labelled 'successful test'), the C-stack of C. Paltenghe ('could be a new electronic component, and possibly have over-unity implications'), and HAARP-related patents (US4686605, US5041834, US4253190, US4999637). Opens with a 1919 Tesla quotation. Link labels only, no measurements.

Relevance: Context for the Tesla/ground-electron ideas in the KAPAGEN and GEGENE pages.

Mainstream conflict: none stated on the page; the FEP and single-wire claims are untested here

Index of magnetic-motor projects: Newman machine video, Mini-Romag generator and Dave Squires's explanations, Greg Watson's RMOG, An Jong-Sok 'No Load Generator' patent (2001), Reed, Howard Johnson, Robert Adams motor (QMM versions 2.0-4.1), Harwood-Jankowski POD2, Naudin's own Magnetic Energy Pump and Two Magnets Push-Pull, Dragone's 'Energetics of ferromagnetism', Faraday homopolar generator, Bedini items (Magnetic Gate, G-Field Generator, Scalar Battery Charger), Bedini's regauging patent WO0152390A1, Flynn's WO0007285 (parallel-path magnetic flux control), Christie WO028656A1, Hooper-Monstein experiment with Cyril Smith's critique and a Godin test report, Steorn. Abstracts of three patents are quoted. No data on the page.

Relevance: Shows that the Steorn Orbo, Bedini and Romag pages sit in a wider 'magnetic motor' family and that Naudin lists critiques (Smith on Hooper-Monstein) beside claims.

Mainstream conflict: none stated on the page

Short page (created 5 Mar 1999, updated 27 Sep 1999). Naudin says he experimented 'successfully' with Bedini's Magnetic Gate, but his additional notes of 27 Sep 1999 state that recent closed-loop tests show the gate 'works as a common energy-conservative system, with no overunity effect', acting as a preloaded magnetic spring that launches its probe. It links tests by Olaf Berens (closed loop) and Phil Bright, whose content is not on the page. No numbers.

Key claims

• In closed-loop tests the Magnetic Gate is energy-conservative, with no overunity effect (Naudin's own 1999 note).

Relevance: A case where Naudin's own later test reports a null result for a free-energy claim; useful counterweight to the overunity pages.

Mainstream conflict: none (the null result agrees with mainstream physics)

Build description (created 19 Mar 1999, updated 15 Nov 1999) of a 12-magnet, six-coil rotor/stator device. The introduction, attributed to 'Magnetic Energy', claims it makes 3.5 V and 7 A DC (about 24 W) 'free electricity', sustains itself, and needs an external motor for only 42 s at 2100 RPM to start. The page gives a 29-item parts list (brass shaft and rotor with 12 slots, 228 copper-coated steel wires, 12 Alnico 4 magnets, six 72-turn coils, etc.) and a wiring recipe, but no measurement. The stated working idea is that magnetic structures 'want to share their flow' and the unit 'must have a load to work, something that attracts magnetic energy'. Some dimensions are garbled in the extraction.

Key claims

• The Mini-Romag runs on magnetic energy alone after a 42 s spin-up, delivering about 24 W.

Relevance: Primary example of a construction-only page with a self-running claim that carries no data; Naudin links his own and Dave Squires's explanations (not in this folder).

Mainstream conflict: A self-running 24 W generator with no energy input contradicts conservation of energy; no measurement is given to test it.

Dave Squires proposes eight test variations of Naudin's Poynting Flow Thruster (PFT) motors, built on the claim that the thrust is E x B x S (S = Poynting vector), with E and B mutually orthogonal. Examples: PFT MK2 with 0 V on the rotor and +28 kV on two outer cylinders (Squires predicts a possible 4x speed and torque), a stronger external B-field (he says the present MK2 B-field is only '99ua'), rotating outer electrodes, and a Searl-type ring-magnet/roller version with a dielectric layer and a negative ring voltage. He asserts the E x B x S principle is 'identical' for the ARDA project, the PFT motors and the Searl SEG. All of this is a list of proposed tests; no results are reported.

Key claims

• Rotation of the PFT motors and the Searl SEG is driven by the Poynting vector S in an E x B x S arrangement (Squires).
• Torque and speed of PFT MK2 should scale with the square of the voltage and could double or quadruple with two E-field vectors.

Relevance: Naudin hosts the PFT motor and Searl-effect ideas; this page is a proposal, not an experiment.

Mainstream conflict: Standard electrodynamics gives no net rotational force from a static E x B Poynting flow in such a configuration; the proposal does not state a derivation.

Design proposal (28 Sep 1999) for a Searl-effect generator in which rollers move around a ring by the same E x B x S force, with a magnet providing B and a pre-stressed, precharged dielectric (electret) providing E. The page states plainly: 'This has not been built and tested yet' and 'It is only a design proposal'.

Key claims

• A roller-on-ring device using a magnet plus a precharged electret would be driven by the Poynting vector (proposal, untested).

Relevance: Shows the Searl material on Naudin's site is proposal only, in line with the README note that Searl work consists of design proposals and PFT roller tests, not a built SEG.

Mainstream conflict: not tested; no mainstream mechanism for continuous rotation from static fields

Gateway page to Naudin's replication of Tom Bearden's MEG. Opens with a quotation attributed to Bearden: 'This one works beautifully and produces COP=5.0'. Naudin's disclaimer: the diagrams are 'a private, independent replication', 'not the original Bearden or accredited-lab diagrams'. Lists US patent 6,362,718 (Patrick, Bearden, Hayes, Moore, Kenny; 26 Mar 2002), Bearden's 69-page MEG paper, Anastasovski/Bearden et al. 'Explanation of the MEG with O(3) Electrodynamics' (Alpha Foundation), Kaivarainen on bivacuum, Cyril Smith's thoughts and critique of Shawn Bishop's 'In Disdain of Garbage Physics', Flickinger's MEG Notes, a note on the THS720P power-measurement error, and a dozen earlier patents on flux-switching/permanent-magnet generators (Starr 1957 to Keiichiro 1999). Also repeats Bearden's note that an 'overunity device' like it was patented by Westinghouse for the Minuteman missile (US 3,239,771, 3,239,772, 3,815,030).

Key claims

• Bearden says the MEG 'works beautifully and produces COP=5.0'.
• Bearden says a Westinghouse-patented switching amplifier of the same kind was installed in the Minuteman missile.

Relevance: Index showing that Naudin presents both Bearden's claims and critiques side by side, with his own replication labelled independent and not the original design.

Mainstream conflict: COP above 1 for a passive magnetic generator contradicts energy conservation unless the measurement is flawed or a hidden energy source exists

Page of 29 Oct to 17 Nov 2000. Naudin says his enhanced control board v2.0 reproduces 'the exact waves shapes' of Bearden's MEG (he points to page 67 of Bearden's paper). TEST RUN #1 (7 Nov 2000): measured COP 1.75; with one side of the secondary grounded, COP reaches about 1.75 'instantaneously'. On 13 Nov he is 'trying to close the loop' without success, noting many losses in the step-down phase and that the device stops after a while. On 16 Nov the two secondary coils broke from sparks inside the windings (output above 3 kV at some frequencies). He plans a nanocrystalline tape-wound core for v3.0. Notes that Bearden's 69-page PDF was removed from the CSeti site on 21 Nov 2000.

Key claims

• Naudin's MEG v2.0 gave a measured COP of 1.75 (run of 7 Nov 2000) with waveforms matching Bearden's.
• The loop could not be closed: the device stopped after a while (13 Nov 2000).

Relevance: First report of Naudin's measured COP for his MEG replication, with the failure to close the loop recorded the same week.

Mainstream conflict: COP 1.75 in open loop, if real, would exceed unity for a device with only a permanent magnet and input coils; the page gives no calorimetry

Main data page of Naudin's MEG replication. Measurement set-up: total MEG input at the control-board DC input, actuator (primary) coil input, secondary output; current via a 10 ohm non-inductive resistor; Tektronix THS720P scope; same method at input and output; a 2002 note cautions the THS720P may give errors with non-sinusoidal waves. Control board (TL494, BUZ11) uses 1.75 W with no load; with a 9 W lamp the DC input is 3.25 W, so 'the lamp's share' is 3.25 - 1.75 = 1.5 W against a measured 6.76 W output. His status of 6 Dec 2000: output waves match the original, voltage and current in phase, a 'conditioned' 100 kohm carbon load or a varistor is REQUIRED, working frequency about 20 kHz above 1 kV peak-to-peak, ferrite magnets; 'apparent' power often exceeds the Joule heat in the load, 'which makes me pessimistic about the calorimetric tests results'; 'most of the power is radiated in EM form' (teslameter 2.8 mT at 16 kHz, E-field 1250 V/m at 50 cm). Not yet checked: core saturation by the magnet, hysteresis flipping, calorimetry. Conclusion: his replication 'seems to be really close to the original' but without the original core diagrams there may be differences; next steps are conversion to usable power and 'closing the loop', which was not done.

Key claims

• Naudin's MEG replication reproduces Bearden's waveforms and gives an apparent COP above 1 when a conditioned non-linear load is used.
• Apparent output power is often greater than the Joule heat in the load; Naudin himself says calorimetry is not yet supported and the loop is not closed.

Relevance: Central Naudin MEG data page; also the source for the 'required non-linear load' condition that critics later address.

Mainstream conflict: COP>1 in a passive magnetic generator conflicts with energy conservation; Naudin himself records the heat mismatch and a possible measurement artifact

French patent application for a self-switching charge circuit: a transistor T1 in series with the primary L1 of a small ferrite transformer, whose secondary L2 turns T1 off through a capacitor when the current rises; diode D1 and a small resistor R1 complete a buck-converter-like chopper at more than 50 kHz (10-15 microsecond cycles). Aim: charge supercapacitor banks (and, less, NiMH/NiCd/lead cells) faster and with less loss, citing the textbook half-energy loss of resistive capacitor charging and an Am. J. Phys. (1992) article on stepwise charging. Only quantitative example: a 1.3 F supercapacitor bank charges in 30 s versus 83 s conventionally (stated as '2.5 times' and elsewhere 'divided by 3'; 83/30 = 2.77), 'efficiency exceeds 90%', losses cut by '50% to 90% and more'. No definition or method for the efficiency, no source voltage or current, no waveforms. The only physical explanation is a 'reduction of entropy variation of the circuit'. The 8 claims cover the circuit, a two-module alternating charger, a hand-cranked generator plus capacitor, and a battery-pack with built-in charger. The patent does not claim output above input. The French search report (20 Oct 1999) cites five earlier documents, three of them against claims 1, 2, 4, 5.

Key claims

• A self-switching ferrite-transformer chopper charges a 1.3 F supercapacitor bank in 30 s instead of 83 s with charging efficiency above 90%.
• Charging in many small fractional steps cuts the resistive loss of capacitor charging by 50-90%.

Relevance: Naudin's earliest patent; ordinary power electronics, no overunity claim. Shows the applicant company (Advanced Electromagnetic Systems S.A.) and the inventors are the source of every performance number.

Mainstream conflict: none: the circuit is an inductive (buck-type) charger, a standard way to avoid the 50% resistive loss; the patent's numbers are unverified

Page of 18-19 Nov 2001 on NASA patent US 6,317,310 (Campbell, granted 13 Nov 2001), which NASA's TechFinder said had no prior working prototype, and a June 2001 Transdimensional Technologies (TdT) vacuum test at NASA MSFC. Naudin 'fully replicated and tested successfully' the patent's Fig. 2 design: two capacitors (copper tube 30 mm x 40 mm, polystyrene 'Roofmat' dielectric, aluminium disk) on a 440 mm frictionless rotor fed through the axis. Results: it rotates and 'accelerates quickly'; at 24 kV, 51 uA (10 kHz) HV power for full-speed rotation at 65 RPM is 'only 1.22 Watts'; a curve (not in the text) at 29 kV, 60 uA (1.74 W). Lists dielectric constants (air 1.00059, polystyrene 2.56, Kapton 4.2, BaSrTiO3 13800). Test in air; no vacuum test of his own. Deep read (2026-10-09): the TdT sentence is the page's whole account: 'This apparatus has been tested in June 2001 by Transdimensional Technologies in the vacuum chamber of the NASA Marshall Space Flight Center (MSFC) in Huntsville (see the photo ...)'. No pressure, voltage, thrust, success or failure, named person or report is given; the source is Naudin's relay of a commercial party (TdT) with a photo link, so pressure, voltage and result are unknown. Naudin's own run was in ambient air at 24 kV, 51 uA (1.22 W; arithmetic checks) and 29 kV, 60 uA (1.74 W), 65 RPM, with no force in newtons and no ion-wind control. Linked pages not in the folder (anchor text only): 'The ACT Working principle proposal by JL Naudin', 'Asymmetrical Capacitor Thruster v2.0' and 'v3.0', the TdT web site, the photo of the apparatus tested in vacuum, NASA MFS 31419_1 brief, Bujic's E-field gradient note.

Key claims

• Naudin's replication of the NASA ACT rotates at 65 RPM with about 1.2 W of high-voltage power.
• Transdimensional Technologies tested the NASA ACT in the MSFC vacuum chamber in June 2001 (Naudin's one-sentence relay; no pressure, voltage or result given).

Relevance: Naudin ties his Lifter work to a NASA patent; the page states a thrust measure only as RPM and power, with no force value.

Mainstream conflict: Rotation in air is consistent with ion-wind thrust; the page makes no vacuum claim of its own

Chronology and position statement. The Lifter is an asymmetrical capacitor using more than 20 kV to produce thrust by the Biefeld-Brown effect (T. T. Brown, 1928). Naudin lists: Brown's patents (GB 300311, US 2,949,550), Serrano's WO 00/58623 (Oct 2000), TdT's Power3 Lifter1/2 (June 2001), his replications (TdT Lifters Oct 2001; NASA ACT 18 Nov 2001; Serrano's thruster 19 Jan 2002; NASA Orbital Maneuvering Propellantless Thruster v1.0, 10 Mar 2002), 'more than 100' replications worldwide, and 'Lifter Maximus' (12 Nov 2002: total weight 194 g, payload up to 60 g). His assessment: the Lifter is not a simple ionic-wind device; Purdue University Energy Conversion Lab vacuum tests (10e-5 torr, Sep 2000) 'show ionic wind is not the source of the main thrust'; in air the major thrust is an electrokinetic effect on the medium, and a minor, 'not yet fully explained' thrust exists in vacuum in the milli-newton range. Also lists the Cameron/TdT patent WO02073218 'Lorentz Force Propulsion System' whose theory he says is 'fully inspired' by Patrick Cornille's 1998 review of Newton's third law, and claims the Lifter was tested by the Army Research Laboratory and by dstl (UK MoD).

Key claims

• The Lifter's main thrust in air is an electrokinetic effect on the surrounding medium, not a simple ion wind.
• A minor thrust of milli-newton size persists in high vacuum.
• A Lifter can scale to carry payload (194 g device lifting 60 g).

Relevance: Naudin's own statement of the central Lifter claim and the historical lineage (Brown, Serrano, NASA, TdT, Cornille).

Mainstream conflict: Mainstream explanation (corona/ion wind, ARL, NASA/ISR, Tajmar) attributes the air thrust to ions; vacuum thrust is contested (Talley null; Purdue and Brown claims are reported by others)

Reference page: NASA patents US6411493, US2002012221 and US6317310 (quoted excerpts); the Cameron/TdT WO02073218 and Naudin's note that its theory part is 'fully inspired' by Cornille's 1998 review of Newton's third law; Loder AIAA-2002-1131; Brown and Bahnson patent list (GB300311, US2949550, 2958790, 3018394, 3187206, 3227901); electro-aerodynamic patents (de Seversky 3130945, Hagen 3120363, Okress 3464207); books by Valone, LaViolette, Szames; web links. States that the Lifter 'has been fully tested' by the US Army Research Laboratory and by dstl, UK MoD, with no report on the page.

Key claims

• The Lifter has been 'fully tested' by the US Army Research Laboratory and by dstl (stated without a report on the page).

Relevance: Bibliographic base for the Lifter story; its ARL and dstl claims are checked by the Related-Research documents (ARL-TR-3005, dstl slides).

Mainstream conflict: none stated on the page

Index of Naudin's Lifter test sub-pages: true Lifter with wire in a vacuum tube; unlinked Lifter with wire in a vacuum tube; unlinked Lifter fully enclosed in a box; asymmetric force between armatures; pulsed high voltage; Lifter1 thrust analysis; RotoLifter; Faraday cage; T. T. Brown device in mineral oil; smoke tests; gamma radiation measurements; ion-wind tests; and confirmation of upward thrust by G. Vizza and F. Daran. Only the vacuum-tube pair is saved in this folder; the Faraday-cage, enclosed-box, smoke, gamma and ion-wind pages are not. QE note (2026-10-09): the sub-pages 'The T. T. Brown EKP device in MINERAL OIL' and 'Test of an Unlinked Lifter and fully enclosed in a box' are listed here but NOT in the folder (not obtained); both bear on whether the thrust needs air (claim 1). Their results are unknown from this corpus. Retrieval target: jnaudin.free.fr, web archives.

Relevance: Shows which of Naudin's decisive Lifter tests (Faraday cage, enclosed box, ion-wind tests) are missing from the folder.

Mainstream conflict: none (index)

Project home page. Defines the Lifter as an asymmetrical capacitor using more than 20 kV, with no moving parts, silent flight and the ability to lift its own weight plus payload. Claims more than 350 replications worldwide, a 250 g 'Maximus II' lifting a 60 g payload (22 Jan 2003), and a proposed VTOL craft. Video section: asymmetrical capacitors tested in full vacuum at 1.72e-6 Torr by Gravitec Inc. at NASA's NSSTC LEEIF facility (summer 2003, video courtesy of Hector Serrano). Repeats that the Lifter was 'fully tested' by dstl and ARL. Promotes Paul LaViolette's book (B-2 propulsion reverse engineering). Navigation to HexaLifter, LifterCraft, worldwide replications, theory and application pages.

Key claims

• More than 350 Lifter replications have been done worldwide.
• Asymmetrical capacitors produce thrust in a full-vacuum chamber (Gravitec video, 2003).
• The Lifter was 'fully tested' by dstl and ARL (bare assertion on this page).

Relevance: Main entry page to the Lifter claims.

Mainstream conflict: Vacuum thrust is not supported by the Talley, Tajmar or Canning (NASA/ISR) null results as reported by those authors; the Gravitec result is Naudin's report of a third party

Replication (10-17 Oct 2001) of TdT's Lifter1: a triangle of three asymmetrical capacitors, 200 mm sides, 40 mm high, aluminium sheet on a balsa frame, thin corona wire 30 mm from the main electrode, weight 2.3 g. Power: 18 W (40 kV at 450 uA) to compensate the weight; 23.9 W (41.9 kV at 570 uA) for stable flight with 1 g payload. Naudin says he 'can confirm fully all their claims' and that lift against weight is 'strong (150%)'. He mentions building similar devices in April 1999 ('EHD Flying Saucer v1.0'). Held down by three nylon threads. Video and Steve Burns's theoretical analysis are linked but not in the text.

Key claims

• Lifter1 (2.3 g) lifts 1 g payload at about 24 W of high-voltage power (41.9 kV, 570 uA).

Relevance: Basic lift data for a small Lifter; the values are Naudin's single replication.

Mainstream conflict: none: lift in air by corona wind is consistent with mainstream explanations; efficiency is poor

Hexagonal Lifter of 36 asymmetrical cells, 32 g, 820 x 700 mm, held by four nylon threads. Power: 81.4 W (40.5 kV, 2.01 mA) to compensate its weight, 132.9 W (44 kV, 3.02 mA) for stable flight. Comparison chart: Lifter 1 2.3 g / 1 g payload / 18 W and 23.9 W; Lifter 2 6.6 g / 3 g / 21.9 W and 48.5 W; Lifter 3 16 g / 4 g / 45.4 W and 69.6 W; Lifter 4 32 g / 4 g / 81.4 W and 132.9 W. Author calls it 'very impressive'. Videos are not in the text.

Key claims

• A 32 g Lifter with 36 cells hovers at about 81 W and flies stably at about 133 W.

Relevance: Scaling table from 2.3 g to 32 g, showing power rising roughly with weight (about 2.5-4 W per gram).

Mainstream conflict: none: consistent with corona-wind thrust; no claim of efficiency beyond ordinary electric propulsion

Log of 'successful registered Lifters replications': 351 as of 4 Oct 2006; the saved page lists numbers 351 to 327 with builders in France, USA, Iran, China, Hong Kong, Taiwan, Italy, Croatia, Malaysia, Finland, Australia, Gibraltar and others. Typical entries: power from an old 19-23 inch CRT monitor (about 25-29 kV), balsa and aluminium foil, total weights 4-10 g. Individual notes: Xie Xuejun (China, 17 Jun 2006) found the lifter 'went down as air was pumped out' and at very rare air made a glow discharge and did not move; Michael Martignac (France) used incense smoke to show air movement; Etienne Basso, Vern Fosset (steerable lifter), 'Tesla Downunder' (an 80 kV lifter lifting its own 4 g; 4.13 g carrying a 4.50 g load). Individual claims are self-reported; the page offers no independent verification. The folder holds the Wayback copy of the page (2008-11-23) because the source site is dead.

Key claims

• There are 351 successful Lifter replications worldwide (self-registered, as of 4 Oct 2006).
• One builder (Xie Xuejun) reports the lifter stopped moving and glowed when air was pumped out.

Relevance: Breadth of replication evidence for the lift effect in air; includes a single builder's observation that lift disappears in rarefied air.

Mainstream conflict: none for lift in air; the vacuum observation agrees with the ion-wind account

E-mail of 5 Oct 2002 to the MEG_builders Yahoo group. Flickinger forwards Tektronix support's answer: RMS of a math waveform (such as a power waveform) is 'bogus and has no value' and 'there is no such thing as RMS watts'; the classic formula True Power = Vrms x Irms x cos(theta) holds only for pure sinusoids; the correct method multiplies voltage and current point by point and takes the mean (WaveStar 'Mean(CH1*CH2)' or Harmonics firmware). Flickinger says this explains why 'Bearden's and JLN's THS720P scope readings' could not be made sense of (that attribution is Flickinger's, not Tektronix's). Linked from Naudin's MEG pages.

Key claims

• RMS values of a power waveform on the THS720P are meaningless; true power must be the mean of v(t) x i(t).
• Bearden's and Naudin's THS720P readings were affected by this (Flickinger's reading).

Relevance: Direct measurement-method challenge to the MEG COP figures, hosted by Naudin himself.

Mainstream conflict: none: this is the mainstream measurement method

Index of Naudin's 'Cold Fusion Reactor' (CFR) replication of Mizuno-Ohmori plasma electrolysis. Headlines (dates corrected by the deep read; the extracted text prints each date before its item): 'Full CFR replication at the CNAM in Paris' (no date on the index; link nfrcnam/index.htm), said to be presented at ICCF 12, Yokohama, 27 Nov-2 Dec 2005 (link aehodme/index.htm; Naudin's page names the paper 'Transmutation of metal in confined plasma in water' by Datillo [sic], Iorio and Cirillo, links lorio/index.htm and lorio/report/index.htm); 'Successful CFR replication by Ronald Hohls (South Africa)', 19 Sep 2006 (logbook/rh/index.htm); 'Successful CFR replication at Boulder (Colorado, USA)', 15 Mar 2006 (ppclkrs/index.htm); 'Successfull CFR by Morgan H. and Marissa C.' (html/cfrmhmc.htm; no date or place); 'Successfull CFR replication by a French lab' (html/cfrppcglv.htm; lab not named, not stated to be the CNAM); a Stirling engine powered by the CFR (html/cfrstir.htm); CFR with a palladium cathode (html/cfrpd.htm); links for a basic and enhanced CFR, an 'APE' experiment, CFR v1.0 to v4.0 test pages (some with power meter and radiation measurements), a Tiny-CFR, 'BingoFuel', a Stirling engine run on CFR, palladium cathode, and 'Transmutation of metal in confined plasma in water' (Dattilo, Iorio, Cirillo). Reference list: Mizuno JJAP 2000, Mizuno 2003, Kanarev, LENR-CANR, Langmuir's atomic hydrogen torch, Shaubach and Gernert (Thermacore) 'Anomalous Heat from Atomic Hydrogen', and five patents (Eccles WO0025320; Mizuno JP2004059977, WO0126117; WO9949471; JP5027062). No data on the page. All 'successful' wording is Naudin's own (he is the interested party). None of the sub-pages is in the folder, so no replication result can be checked here; the ICCF-12 paper attached to the CNAM item is by named authors (Datillo/Dattilo, Iorio, Cirillo) and is about transmutation, so it need not report excess heat. Do not assume it is the Fauvarque-Clauzon-Lallevé work (that attribution is a QE [memory] guess, not on the page).

Key claims

• Naudin lists CFR replications at the CNAM (Paris, undated on the index, ICCF-12 presentation of Nov-Dec 2005), by Ronald Hohls (South Africa, 19 Sep 2006), at Boulder (15 Mar 2006), by Morgan H. and Marissa C., and by an unnamed French lab (Naudin's own wording; reports not obtained).

Relevance: Index for the CFR claims; it shows the third-party replication list but the data pages are not saved.

Mainstream conflict: Excess heat from plasma electrolysis is outside accepted nuclear and chemical energy sources; no mainstream acceptance of the Mizuno result

Page header and references for Naudin's CFR results. States the CFR is 'fully based on' Mizuno and Ohmori (Hokkaido University), that he replicated it on 7 May 2003 following Mallove's protocol (amasci.com/weird/anode.txt) and Infinite Energy 4(20) 1998, and that it is 'fully replicable' and 'works very well as described'. Quotes Mizuno's abstracts: heat 'in some cases 200% of the input power' and hydrogen 'sometimes 80 times higher than normal Faradic' production (Mizuno's claims). Note that no one can patent the effect 'since it was noticed in other forms (1916!)' (citing Kellogg, J. Electrochem. Soc. 1950, via Mallove). The results themselves (the 33 runs) are not in the saved text; Kowalski's item 73 quotes them.

Key claims

• Naudin's CFR reproduces the Mizuno-Ohmori excess heat; it is 'fully replicable'.
• Mizuno reports heat up to 200% of input and hydrogen up to 80 times the Faraday value.

Relevance: Naudin's own statement of the CFR claim; the data are not in the text layer.

Mainstream conflict: Excess heat beyond input would contradict established energy accounting unless an unknown source exists; mainstream reviews (e.g. DOE 2004) did not accept cold fusion heat claims [memory] medium

Test (follow-up to 6 Mar 2003) in which the aluminium plate (200 x 20 mm) is mechanically linked to a tungsten wire in a 38 mm, 310 mm vacuum bulb tube on a rotating torsion system (910 g total), at 25 kV DC, with the plate grounded. On switching on, the Lifter moves in the direction of the wire; Naudin reports no corona, no ion wind, no hissing. He concludes this 'confirms that the thrust observed is really relative to an external referential and not a simple electrostatic attraction', that the surrounding dielectric (air) is 'required for the motion' (the plate is attracted to the dielectric medium polarised by the field), and that a true Lifter fully enclosed in a box 'will not move'. No thrust value or current measured; videos not in the text.

Key claims

• A Lifter with its wire sealed in a vacuum tube still moves, relative to the laboratory.
• A true Lifter fully enclosed in a box will not move; a polarisable external dielectric is required.

Relevance: Naudin's own argument that corona and ion wind are not needed for the main thrust, while also saying the external medium is needed.

Mainstream conflict: Naudin reads it as a force relative to the laboratory medium; the ARL, NASA/ISR and Tajmar documents explain Lifter thrust by ion momentum transfer to air and none addresses this sealed-tube arrangement

Test of 6 Mar 2003. A thin tungsten wire in the vacuum of a tubular 60 W/220 V incandescent bulb tube (three tubes joined in a triangle) is mechanically unlinked from a free aluminium plate (200 x 20 mm) held at ground. With 25 kV on the wire apparatus brought above the plate, the plate rises toward it. Naudin reports no corona, no ion wind, no hissing, current 'about 1A' (as printed, evidently microamps) and hover with 'only 25 mW'. He concludes this 'proves definitely' that the Lifter flies without corona, with very weak current, that ion wind is not responsible for the main thrust, and that the attraction of plate to wire is stronger than of wire to plate (an asymmetrical force between armatures). Video not in the text.

Key claims

• An aluminium plate flies up toward a charged tungsten wire sealed in a vacuum tube with no corona or ion wind, at about 25 mW.
• The attraction of plate toward wire exceeds that of wire toward plate (asymmetric force between armatures).

Relevance: Naudin's key experiment behind his claim that ion wind is not the main cause; it is air-pressure on the plate, with the wire in a sealed tube.

Mainstream conflict: Naudin reads it as thrust not due to ion wind; the ARL, NASA/ISR and Tajmar documents treat Lifter thrust as ion momentum transfer to air, and none of them discusses this sealed-tube test

Index for the MAHG, 'Free Energy from Atomic Hydrogen', stated to be 'fully based on the Irving Langmuir discovery': hydrogen is 'merely dissociated and recombined and can therefore be recycled over and over'. The test log lists: v1.0 2-hour run COP 2.43; towards COP>10; v2.0 high efficiency; one-hour run COP 5.26; one-hour test COP of 21 (v2.1); radiation measurements. Also links 'MAHG datas table', construction diagrams, genesis logbook and a Langmuir article by Moller. No measured values are in the text.

Key claims

• The MAHG gives COP of 2.43 (v1.0), 5.26 (v2.0) and 21 (v2.1) in one- to two-hour calorimetric runs.
• Atomic hydrogen recombination heat can be recycled without consuming hydrogen.

Relevance: Index to the MAHG claim (heat output many times the electrical input).

Mainstream conflict: Recombination of atomic hydrogen releases the dissociation energy that was put in, so a net excess requires a hidden input; Kowalski (item 289) makes this point

Page of 28-30 Jun 2005 headed 'the COP is now up to 21'. A one-hour test at 5% duty cycle (51 Hz) with two dataloggers (5 s sampling), flow meter and Fluke 189 meter; a blank run with continuous DC validates the bench. Improvements: shorter wires and better contacts, input measured with a Fluke 123 digital oscilloscope. Only figure captions (RUN #76-83 and a BLANK TEST) are in the text; no temperatures, flow rates, powers or COP per run appear in the text layer.

Key claims

• MAHG v2.0 reached a COP of 21 in a one-hour run (5% duty cycle).

Relevance: Primary page for the highest MAHG COP claim; numerical data are only in unextracted figures.

Mainstream conflict: COP above 1 from hydrogen dissociation/recombination would exceed the energy available; calorimetric method not evaluated on the page

US patent application for a source built from a plate capacitor with plasma tubes or plasma chambers acting as switchable inner electrodes. When the gas is ionised the device is two thin-dielectric capacitors in series (C1, large); when it is extinguished it is one wide-gap capacitor (C2, small). At constant charge the voltage and stored energy rise by beta = C1/C2. The abstract says 'the device has an efficiency of more than one'. The text accounts for the extra energy in three ways it does not reconcile: work done by Coulomb forces over a few Debye lengths ('almost free'), 'the surrounding magnetic ether' (paragraph 64), and the Earth's potential energy via an Earth capacitance of 700 microfarad. Worked example: C1 = 1.75 nF at 1 kV, beta = 200, E2 = 0.175 J, plasma power 50 W for under 1 ms, stated efficiency 338% (computed from an assumed beta, not measured; the patent also says a real beta of 10 could give efficiency below one). The only measurement is Fig. 8, a voltage trace on a 1 Gohm probe: 500 V rising to 1750 V when switch K1 opens (ratio 3.5), with no input power, no output energy, no calorimetry and no dates. The two claims are structural; none mentions efficiency, the ether or the Earth. This is a published application, with no evidence of grant in the file.

Key claims

• A plate capacitor whose inner electrodes are switched plasma chambers can output more electrical energy than it takes in ('efficiency more than one').
• Switching the plasma raised a measured voltage from 500 V to 1750 V at constant charge (Fig. 8).

Relevance: Naudin's second patent and the clearest statement of an efficiency-above-one claim from him; the extra energy is attributed to magnetic ether or Earth potential.

Mainstream conflict: At constant charge the energy increase equals the work done pulling the plates apart; the patent books this as 'almost free' without measurement, so it conflicts with conservation of energy unless input work is accounted for

Page of 12-26 Jun 2008 on a replication attempt of Stanley Meyer's Water Fuel Cell. Naudin says the cell uses parametric resonance with high voltage and low current rather than Faraday electrolysis. Observations: v1.0 produces a high flow of very small bubbles in distilled water; resonance tests at pump 1.110 MHz (fo 0.687 MHz) and 0.329 MHz (fo 0.560 MHz); v1.1 with an insulated cathode behaves as a capacitor (4.7 nF calculated vs 5.19 nF measured; internal resistance 196 kohm calculated vs 2.9 Mohm measured) and a VIC (Voltage Intensifier Circuit) charges it stepwise to 1.1 kV; a 246 mH coil raises it to 1.4 kV p-p at 13 kHz. He calls the 'adiabatic charge effect' very important because it minimises Joule losses. No gas-volume or energy-balance figures are given. Cites Meyer's US 4,936,961.

Key claims

• A Meyer-type water fuel cell driven in resonance with a voltage intensifier splits water with little current (Meyer's claim, Naudin's attempt).

Relevance: Naudin's start on the water-fuel-cell line later continued in the WFC booster page.

Mainstream conflict: Claims of water splitting below Faraday-law energy cost conflict with thermodynamics; the page gives no gas yield versus input energy

Log of a ferrite-toroid (3E25, 7.5 m of 0.4 mm wire) plus 10 mm neodymium magnet pulsed by a 555/BUZ11 stage, with a separate flat output coil (636 mH) driving LED lamps. Naudin states the coils are fully decoupled and the magnet only sets the operating point on the core's B-H curve; the 'excess of energy' is tapped during the demagnetisation phase 'due to the blocking diode', 'the 2SGen is not a transformer', and 'best tuning is done when there is no change in the measured DC input power while the output coil is loaded' (Episodes 6 and 5). Later versions use a Nanoperm core, a FEMM simulation, and v6 lighting 20 W + 5 W bulbs. No input/output power measurements or COP are given in the text; some episodes are heading and video only. Cites N. Zaev's papers on ferrites and 'free energy'. Reports 'successful replications by independent experimenters' (3 Mar 2010).

Key claims

• In the 2SGen the load power comes from the core's demagnetisation phase and does not appear as extra input current.
• 2SGen v6 lights 20 W + 5 W bulbs from the tiny generator.

Relevance: Naudin's own solid-state device, with a statement that the input current does not change under load; measured powers are missing.

Mainstream conflict: Net output without a change in input current would exceed conservation of energy unless the energy comes from stored magnetic energy or the measured quantities are too coarse; the page gives no power balance

Replication from Steorn's public information: rotor of NdFeB magnets (100 mm diameter) and toroidal ferrite coils (3E25, 235 mH to 58.6 mH), 10 V DC, 220 mA at 750 RPM. Naudin's interpretation: ferrite temporarily polarised then freed by a depolarisation pulse, with 'symmetry broken' by 'magnetic regauging' (Bearden) giving a 'non-reciprocal' system. Test results as stated: shorting two coils gives no change in RPM; inductance falls from 963 mH to 190 mH as magnets approach; back-EMF fully cancelled at one coil position; current flat for 72% of the pulse while inductance rises from 246 to 916 mH; a 600 mH pickup coil lighting a high-power LED panel leaves the 400 mA coil current unchanged. 'Overunity proof?' section: Larsko's replication, 4 V x 31 mA = 0.124 W unloaded at 1551 RPM and 4 V x 32 mA = 0.128 W loaded at 1240 RPM with 0.023 W in the LEDs; Larsko writes 19 mW/4 mW = 4.75; Naudin's comment: COP less than one, '4.75 is NOT A COP', yet the loading test 'demonstrated an overunity effect' (the page's own wording is inconsistent). Disclaims affiliation with Steorn. Deep read (2026-10-09), exact wording: section heading 'OVERUNITY PROOF ?'; lead-in 'Successful Orbo replication by Larsko, measured COP = 4.75 ... His measurements during the loading test has demonstrated an overunity effect.' Later comment: 'the COP of this device is less than one. The factor 4.75 is NOT A COP or the efficiency of his device ... while 23 mW of power is collected at its output, only 4 mW is tapped at the input ... a good example of a non reciproal effect.' The page does not reconcile the two statements. Arithmetic (deep reader's): 4 V x 31 mA = 0.124 W; 4 V x 32 mA = 0.128 W; 7.72 V x 2.99 mA = 0.023 W; 4.75 = (23-4)/4, a ratio of small differences; output over total loaded input = 0.18. Larsko's speed fell from 1551 to 1240 RPM under load (not discussed by him or Naudin) and meter resolution (1 mA on about 31 mA) is not stated. TEST 7 (Naudin's own): analog ammeter, 400 mA with LEDs off and 400 mA with the LED panel on; the input voltage for TEST 7 and the LED power are not stated, so no energy balance can be computed; the digital-ammeter video is not read.

Key claims

• Loading the Orbo with an output coil lighting LEDs leaves the input current unchanged (400 mA).
• Larsko's Orbo replication measured 23 mW out against 4 mW extra input (his '4.75'); Naudin states this factor is not a COP.

Relevance: Shows Naudin applying the 'regauging / non-reciprocal' language and also correcting a replicator's COP figure.

Mainstream conflict: A load that draws power with no change in input current would exceed conservation of energy unless the load power is within measurement uncertainty of the input; the Larsko figures are 23 mW against 0.124 W input

Attempt (28 May to 5 Jul 2010) to replicate Tariel Kapanadze's generator, said to follow Tesla's fuelless generator. Log: 150 W halogen lamp on 31 May; 420 W with v3.2 (1 Jun); 1200 W of light on 5 Jun; 2100 W of light on 7 Jun; input measurements of 2 Jul described as 'far more accurate' but no numbers. Working theory ('IMHO'): the device draws free electrons from the Earth treated as a large capacitor; needs two earth grounds 10 m apart; output depends on weather and underground current. Lamp powers are 'light output' claims without an input/output table. Quotes Nichelson on Tesla's generator. Lists Kapanadze's patent WO2008103129A1 ('Independent Energy Device'; applicants Turk and Kapanadze).

Key claims

• KAPAGEN v3.3 lights up to 2100 W of halogen lamps from an input not yet reported, by tapping free electrons from the Earth.

Relevance: Naudin's Tesla-line replication attempt; the page says the research is only beginning and gives no input/output numbers.

Mainstream conflict: No mainstream mechanism for continuous electrical power from earth-ground potential differences; input power not published on the page

Project page for an induction cooker (under 80 euro) whose controller drives a flat bifilar 'pancake' coil attributed to Tesla's 1894 patent 512,340, used as an output coil loaded with halogen lamps (load 'at least 1500 W'). References Utkin and Nichelson. Menu of 18 tests (lightmeter efficiency, oscilloscope efficiency, calorimetry, closed loop by 'Romero UK', grid-tie closed-loop phase 2, 4180 W of halogen lamps, HHO generator, Meyer WFC, quartz heater), but no results are in this page's text. Cites Russian videos ('IN 230 V/4.3 A/900 W, OUT 2.3 KW') as similar experiments. Deep read (2026-10-09): the page is an index only and states no test result. It lists 17 numbered tests, TEST-S02E01 (Meyer WFC) and TEST #18 (18 numbered tests plus one lettered). Link targets: #1 lightmeter gegene05en.htm; #2 oscilloscope gegene06en.htm; #3 'calorimetric method' gegene07en.htm; #4 'first test of a closed loop GEGENE by Romero UK' gegene08en.htm; #9 'CLOSED-LOOP (Phase 2) ... via a Grid Tie Power Inverter' gegene13en.htm; #16 4180 W halogen lamps gegene20en.htm; #18 gegene22en.htm. None of these sub-pages is in the folder. 'Romero UK' and 'WOOPY' are outside experimenters known by handle; 'very interesting' and 'successful' are Naudin's adjectives. The Russian videos' title 'IN 230 v/4.3 a/900 W OUT 2.3 KW' is that team's own claim, not Naudin's measurement. Claimed closed-loop tests (#4, #9) are therefore claims whose results were not obtained.

Key claims

• An induction cooker plus a Tesla bifilar coil gives a power output exceeding the input (title claim: 'Great Efficiency').
• Output of up to 4180 W of halogen lamps was run (Test #16).

Relevance: Naudin's latest large-scale claim; the numerical test results are on linked pages not saved here.

Mainstream conflict: Resonant coupling to a lamp load can look like gain if apparent power or lamp light output is mis-measured; no measurements are available in the folder

Follow-on project: a Meyer-style water fuel cell fed by the GEGENE to make HHO gas from ordinary water 'by parametric elongation of the hydrogen-oxygen bond' using high-voltage nanosecond pulses. A test series list S02EP01-06 includes osmosed water, Meyer's VIC pulse, and 'successful' replications by 'WOOPY'; no results in the text. References: Meyer, Vanags et al. (inductive voltage pulses), Shimizu et al. (ultra-short pulse electrolysis, NGK), Dharmaraj and AdishKumar (nano pulsed DC), Dave Lawton.

Key claims

• An ordinary-water cell on high-voltage nanosecond pulses produces HHO efficiently and in large amounts.

Relevance: Continues the WFC line from 2008; no yield or energy figures are in the saved page.

Mainstream conflict: Hydrogen yield beyond the Faraday-law energy cost would conflict with thermodynamics; the page reports none

Delayed Lenz Effect test series. A unipolar motor piloted by Naudin's Arduino pulsed motor controller (APMC) powers a 24-LED spot (1.6 W, 12 V) only from back-EMF pulses. Central stated result: on loading, the input power 'drops dramatically' and RPM is 'doubled'; DC output through the spot is checked with a wattmeter at 2.3 W. No input values are in the text (figures and tables not extracted). Circuit: IRFP450, 4N25 optocoupler, HFA25PB60 diodes, open-source firmware.

Key claims

• Loading a Bedini unipolar motor with a back-EMF LED load lowers its input power and doubles its RPM, with 2.3 W DC at the output.

Relevance: Latest primary example of a Bedini-line 'regenerative' claim; no numbers saved.

Mainstream conflict: Lower input power on loading is not expected in an unregulated motor; with a pulsed controller it may reflect timing changes, but no data are shown

Catalogue compiled 2026-10-09 listing every saved file with original URL, retrieval date, what it is and primary/secondary status. Primary means Naudin's own pages and patents; secondary means third-party tests, critiques and background, all in Related-Research/. Naudin's live site is jnaudin.free.fr (www.jlnlab.com redirects); jlnlabs.online.fr hosts CFR/MAHG. jlnlabs.online is NXDOMAIN with no Wayback copy; jlnlabs.imars.com is dead (one page from Wayback 2008-11-23). Images, videos and scope captures were not saved. Lists what was NOT obtained: Shawn Bishop's original MEG critique, Infinite Energy articles on Naudin, interviews, videos, and many sub-pages (Adams v2-v4, TEP v2-v8, PFT v1-v3, Lifter v2/v3/Maximus/theories, DLE tests 1-25, MAHG v1.0 runs, CFR v1-v4 subpages, GEGENE tests 1-20, VTA, Newman, Romag explanation). Notes that Naudin's site has no Edwin Gray or Mueller pages and that Searl material is design proposals, Hendershot material is an Epps analysis.

Relevance: Defines the source gap: the data pages and images behind many claims were not saved.

Mainstream conflict: none

Home page of the site, 'dedicated to the search for Free-Energy solutions and new generation space-propulsion systems'. Lists latest updates (Bedini unipolar motor, Delayed Lenz Effect, Thane Heins's regenerative acceleration, GEGENE, WFC booster, Vialle's self-running generator) and some 30 'What's new' projects: LifterCraft, Tesla technologies, propellantless propulsion, cold fusion, plasma and ball lightning, BingoFuel, magnetic motors, Bearden's Free Energy Collector explained, WFC, Synergetic theory, MAHG, Chung's negative resistance, SMOT 'free energy demonstrator', 'direct conversion of environment temperature into electricity (COP up to 1.35)', G-strain energy absorber, parametric power conversion, Cornille's electrostatic pendulum 'now demystified', UAV research. Copyright notice (non-commercial use) and liability disclaimer.

Relevance: Shows the full scope of Naudin's site; most pages listed are not in the folder.

Mainstream conflict: none stated (index)

Patent-style disclosure by the inventors (interested parties) of a permanent-magnet, nanocrystalline-core generator with two actuator coils and two collector coils. The theory (pp.1-50): the Lorentz symmetrical regauging of Maxwell's equations is 'not a law of nature'; a magnetic dipole, via Whittaker's 1903 scalar-potential decomposition, takes energy from the complex plane (time domain) and 'giant negentropy', so the vacuum replenishes the magnet; the core separates curl-free vector potential A from B, and coils interact with both in a 'ping-pong' giving COP>1 while 'rigorously obeying' 4-space energy conservation but not 3-space; Heaviside's non-diverged energy flow is 'some 10 trillion times' the Poynting component. 55 references including 'in press' papers by Evans, Anastasovski, Bearden et al. All measured data are four figures: scope photographs of one run (input: Ch1 'High 51 V', Ch2 RMS 117.5 mA, math 'M RMS 3.588 W'; output: 1.96 kV peak-to-peak, RMS 15.44 mA, math 'M RMS 12.7 W', Tektronix, 5 us/div) and two plots of eight points each (Figs 20-21: output about 26 W per coil at about 75 V input, COP about 2.7 per coil, doubled by the authors to a 'net unit COP' near 5.4). The unit ran open loop; no calorimetry, no independent test, no coil turns, wire, core grade, magnet, frequency, load value or operator named. Figs 22-23 are projections to 'COP = 30 or 40' with multikilowatt output. The paper itself says replication and independent test are still to be done. Deep-read arithmetic: the two scope numbers give 3.54, the plotted COP near that input gives about 2.2; 'M RMS' of a power waveform is not mean power (12.7 W RMS corresponds to about 10.4 W mean for a sinusoid), and 51 V x 117.5 mA = 6.0 W not 3.588 W.

Key claims

• A permanent-magnet core generator with input and output coils can output more electrical power than it takes in (COP above 1) by drawing on the vacuum (Bearden et al.).
• The prototype measured COP about 2.7 per output coil (about 5.4 net) at 75 V input, open loop.
• Closed-loop self-powering is claimed as a capability but no closed-loop run is reported.

Relevance: [secondary] The original claim that Naudin set out to replicate; the data basis is very narrow and uses scope RMS-of-power readings that Tektronix later said are meaningless.

Mainstream conflict: Claims energy extraction from the vacuum via regauging and 3-space non-conservation, which conflicts with energy conservation and standard electrodynamics (gauge invariance)

Landing page (31 Oct 2002) published with Bahder's permission, listing authors Bahder and Fazi of the ARL and links to the PDF and arXiv physics/0211001. No body text.

Relevance: [secondary] Evidence that the ARL did look at the Lifter, with Naudin's cooperation.

Mainstream conflict: none

Army Research Laboratory report (Aug-Dec 2002 work, June 2003). The authors built three air lifters (a triangle after Naudin's design, a square, a flat 'wing') plus a suspended wire, at 30-37 kV and 0.5-1.5 mA; each produced a net force toward the smaller electrode, independent of voltage polarity and of orientation to Earth, with no ground plane or closed geometry needed; arcing weakens it. No force was measured (lift versus weight only); no vacuum or reduced-pressure test. Theory: a ballistic ion-wind upper bound is too small by at least 3 orders of magnitude (electrons give 6.8e-5 g; the report's copper-ion figure of 0.002 g recomputes to 0.023 g in the deep read); an ion-drift scaling estimate F = I l / mu (mu = mobility) gives 0.16 N or 16 g for N2+ ions against a roughly 5 g lifter, 'the right order of magnitude' (the deep read notes this is the collisional form of ion wind); a thermodynamic dielectric-fluid force expression (eq. 33) predicts zero force in vacuum if there is no free charge, with the relative size of its terms unknown. Conclusion: 'the physical mechanism responsible for this effect is unknown'; most pressing question is whether the effect occurs in vacuum. Appendix B reproduces Naudin's e-mail quoting a letter 'purportedly signed' by T. T. Brown describing vacuum tests in Paris 1955-56, Winston-Salem 1957-58 and G.E. 1959 (vacuum beyond 1e-6 mm Hg, thrust bursts with vacuum sparks, continued thrust without spark, lab notes never published), the 1928 gravitor (max 100 kilodynes at 150 kV for a 10 kg unit), solar and lunar correlations, and a 15-inch dome electrode lifting 110% of its weight; the ARL presents these without verification.

Key claims

• Asymmetric capacitors produce a net force toward the smaller electrode in air at about 30 kV, independent of polarity and orientation.
• Ballistic ion-wind recoil is at least 3 orders of magnitude too small; collisional ion drift gives the right order of magnitude.
• Whether the effect exists in vacuum is the most pressing open question; Brown's letter says it does.

Relevance: [secondary] Institutional test most often cited about Naudin-type Lifters; also the place where Brown's vacuum claims reached an agency report via Naudin.

Mainstream conflict: Gives no mechanism proven; the theory says vacuum force would be zero absent free charge, while Brown's letter and the Talley report claim some vacuum effect

Physicist's commentary of 29 Jun 2003. Kowalski finds Naudin's report of 33 excess-heat runs completed on 25 Jun 2003 'very impressive', quotes run 28 (170 g of water evaporated; 1200 W input; 3093 W heat), and notes the blank run 'seems to indicate the absence of a systematic error'. He says the data are 'reasonably reproducible' and that skeptical accusations raised about Mizuno's data 'can also be raised about Naudin's data'. P.S.: Naudin replied he is 45, not a student, works in a private lab with some French physicists. Acknowledgements on Naudin's page thank CNAM electrochemistry staff. Deep read: Naudin's acknowledgement as quoted by Kowalski: 'I am very grateful to Professor Pierre Clauzon for his contribution to the CFR project and his help during some CFR tests runs and also to Professor Jean-Francois Fauvarque and Gerard Lalleve from the Laboratoire d'Electrochimie Industrielle of the CNAM ... for their scientific support and their contribution to this project.' Only Fauvarque and Lalleve are placed at the CNAM; Clauzon's affiliation is not stated; what 'scientific support' was and where the runs were done are not stated. Naudin's reply (partly elided): 'I am not a student ... (I am 45 years old...), I work in my own and private lab and some French physicists work closely with me in this field of research'; no names. Run 28 implies 3093/1200 = 2.58 times input (Kowalski's reading; calorimetry and uncertainty not given). No corroboration from the named professors is in the folder.

Key claims

• In run 28 of Naudin's CFR, 1200 W of electrical input produced 3093 W of heat (170 g of water evaporated).

Relevance: [secondary] Only saved source for the CFR run numbers (the data page itself is not in the folder). Kowalski could not follow the details and relies on Naudin's summary.

Mainstream conflict: Excess heat is not part of accepted physics; Kowalski himself flags that the same doubts apply to Naudin's data

Cyril Smith (hosted on Naudin's MEG pages) replies to Shawn Bishop's paper that attacked the MEG and Naudin's replication; Bishop's original is not in the folder, so his arguments appear only as Smith quotes them. Smith says Bishop applied 'RMS power' wrongly (the mean of the power waveform is correct), omitted the DC component of the input current (so Bishop's 2.47 +/- 0.33 W input is wrong), and took the load as 100 kohm although Naudin's resistor was 'conditioned'. He then argues a magnet's electron spins precess perpetually and could hold about 30 J for Bearden's NdFeB magnet; extracting 0.01% per cycle at 25 kHz would give 75 W, and asks whether the vacuum could restore the precession energy. Concludes Bishop's three flaws make his calculations invalid.

Key claims

• Bishop's analysis of Naudin's MEG data used a wrong power formula, left out the DC input component and used the wrong load value (Smith's claim).
• A permanent magnet holds precession energy of about 30 J that could be tapped (Smith's speculation).

Relevance: [secondary] The pro-MEG side of the measurement dispute; only the defence is saved, the critique is not. Smith is an advocate hosted on Naudin's site.

Mainstream conflict: The precession-energy argument has no standard-physics basis; the RMS-power point agrees with standard method (Tektronix note)

NASA contractor report. Four disc-and-cylinder asymmetrical capacitor thrusters (ACTs) on a rotating arm, tested first in an aluminium-lined wooden box, then in a 36 x 50 x 72 inch steel chamber in air, nitrogen and (Device 4) argon at several pressures, at up to 50 kV. Thrust inferred from terminal rotation and spin-down friction (no direct force reading; one observer could not resolve under a sixth of a turn). Table 1 (atmospheric): Device 4 reached 128-138 RPM at 42-46 kV with 130-307 uA across the gap. No rotation below 300 Torr except one event: Device 2 pumped to 5.5e-5 Torr moved about an eighth of a turn at 44 kV during a large arc, which the authors read as ejected material. Mechanisms examined and rejected: ablation (needs 1.5e-5 kg/s), vacuum polarisation (energy about 9 orders too small), image-charge forces, Trichel-pulse radiation. The ion-drift model with 1e10 collisions/s gives F = 2 d (1e10/s) m I / e = 1.44 N for 3.5 mA, 7 cm, and 77% of the measured force for Device 4/A. Thrust also in nitrogen and argon. Conclusion: 'we find no evidence' for new physical principles. The deep read notes the 77% comparison cannot be checked (currents in Table 1 are 10-40 times below the 3.5 mA used) and that the vacuum null has no stated force bound and devices gapped for air. Naudin appears only as a lifter builder (ref. 9). Talley's Air Force vacuum study is quoted: no detectable force in high vacuum.

Key claims

• Asymmetrical capacitor thrust in air is fully explained by electrostatic forces on drifting ions that transfer momentum to air by collisions.
• No thrust was seen below 300 Torr (no evidence for forces in vacuum).

Relevance: [secondary] Main institutional (NASA contractor) negative result on non-ionic thrust; limits of the apparatus are stated by the authors.

Mainstream conflict: Disagrees with Naudin's claim of a non-ion-wind or vacuum thrust; supports the ion-wind account

AIAA Journal 42(2) paper. Reviews Brown's claims (thrust persists to 1e-6 Torr; dependence on dielectric mass suggested a gravity link; all in patents), Talley's vacuum null (no thrust at static DC; an 'anomalous force' during breakdown), and a 5-dimensional theory (Williams 1983) in which divergent current adds mass flow. Tests it with a single air experiment: aluminium cylinder and ring with 6 cm gap in a sealed 50 cm wooden box covered in grounded foil, hung as a pendulum (7.499 kg; laser resolution gives 10.5 uN), 0 to about 38 kV, current 0.6 mA (22.8 W). The box oscillated about the same mean position (50 to 100 uN amplitude); any linear thrust is below 10 uN. Corona-wind formula F = P l /(U b (1+phi)) gives 163 mN at that power, so a gravity-EM coupling is at least 5 orders below corona wind (the deep read computes 4.2 orders) and at least 2 orders less efficient than ion thrusters. He attributes Brown's vacuum thrust to corona wind from insufficient outgassing, without a vacuum measurement of his own. One run, no repeats or controls; the test probes the 5-D mechanism in a corona geometry, not Brown's plate-and-wire, vacuum, AC or high-permittivity dielectric cases.

Key claims

• A sealed corona-discharge box shows no net thrust above 10 uN at 22.8 W, so no gravity-electromagnetism force exists beyond corona wind.
• Brown's claimed vacuum thrust was most likely corona wind from outgassing.

Relevance: [secondary] Mainstream test used to dismiss non-ion explanations; its design does not test Brown's vacuum claim directly.

Mainstream conflict: Directly contradicts the gravity-linked reading of the Biefeld-Brown effect; Naudin's page never addresses the sealed-box result

40 slides of 20 Apr 2005 on horizon scanning of fringe topics: cold fusion, transmutation, bio-communication, water memory, Casimir force, nanofibre hydrogen storage, Podkletnov gravity shielding and asymmetric capacitors. Lifter section: lists patents, why the MOD is interested (propulsion, drag reduction, stealth, silent flight), project aims, a table of Lifters (IW3, IW4, IW8) and plots of voltage, current and power against wire height and mass against power (values only in figures). Text results: air lifts stably at 0.25 mA/19.2 kV and nitrogen at 0.31 mA/18 kV; helium no lift at 4.2 mA/12.8 kV. Summary: plate spacing matters more than plate asymmetry; effect tied to corona onset; independent of orientation; physics 'not understood'; 'calculations suggest mass lifted by ionic thrust is more than an order of magnitude less than that observed'; 'weight loss observed before lifting'. Conclusion: a large proportion of unconventional claims may be poor science. Deep read (2026-10-09; chart values are the deep reader's readings, about +/-0.02 mA, +/-0.5 kV, +/-0.5 W; the slides give no data table and do not name the co-workers, site or dates). p.17 is a photograph of a foil-skirt balsa 'Lifter 1' with parameter labels, not a table. Lifters IW3 (1.35 g), IW4 (2.3-2.8 g, foil 21-40 mm) and IW8 (38 mm foil, 3.16-6.16 g by ballast) were run at wire heights L = 20-44 mm. IW8, L = 44 mm: 6.16 g at 1.07 mA and 34.3 kV (36.7 W), 3.16 g at 0.39 mA and 26 kV (10.1 W); voltage rises about 0.7-0.9 kV per mm of wire height; mass is near-linear in L x foil height x current (p.33). Helium: no lift at 4.2 mA, 12.8 kV (p.28). Test of p.37 by the deep reader with F = I d / mu (mu = 2e-4 m^2/(V s), d = wire height): F is 2 to 5 times the lifter's weight in every case (IW8, L = 44 mm, 6.16 g: F = 0.235 N against 0.060 N, ratio 3.9), i.e. the observed lift is 21-47% of the ion-drift bound, the opposite of 'an order of magnitude less'. Caveats: Dstl shows no formula, mobility, gap or lifter for its statement, the currents are supply currents (leakage not excluded), and the deep reader's d choice is an assumption; a smaller d (under about 5 mm, supported by no slide) would be needed to reverse the result. p.34 balance chart: IW8 (2.94 g) reads about 2.9 g at 13-14 kV, 2.7-2.8 g at 17 kV and about 1.85-2.15 g at 18 kV (series ends; lift-off); IW8 + 6.7 g loses about 0.7-0.9 g (7-10%) by 22 kV without lifting. That is the 'weight loss observed prior to lifting'; a force that grows with voltage and current (ion thrust included) would produce a gradual loss, so it is not by itself anomalous; no corona-onset voltage, current at those voltages or controls (balance-pan forces) are shown. p.36 says the greatest force is when the small electrode is positive while p.37 says the effect is independent of polarity (internal tension; no polarity data). p.27 lists helium relative permittivity as 0.2 (real value about 1.00007 [memory, high]). Stake: a Ministry of Defence laboratory horizon-scanning briefing, not a paper, with no error bars.

Key claims

• Dstl's own Lifter tests show lift in air and nitrogen but not helium (slide 28).
• Dstl states calculated ionic thrust is more than an order of magnitude below the observed lift (slide 37), but shows no calculation; with F = I d / mu and Dstl's own chart currents and wire heights the ion-drift bound is 2-5 times the lifted weight (deep reader's check from chart readings).

Relevance: [secondary] The only Dstl source saved; supports Naudin's statement that dstl tested the Lifter, and records a result that disagrees with the pure ion-wind estimate. Figure data are not readable from the text.

Mainstream conflict: Dstl's slide statement on ionic thrust conflicts with the NASA/ISR and Tajmar conclusions, but the slides show no calculation and the deep reader's recomputation from the slides' own data does not reproduce it

Kowalski's commentary of 2 Apr 2006, based on a 2005 Pure Energy Systems News article (Allan and Akau). Describes the MAHG (Moller's Atomic Hydrogen Generator; Alexander Frolov of Faraday Lab Ltd. is named as inventor, Moller as co-funder) as a flow calorimeter around a 0.1 atm hydrogen cell, about 250 V at 10 MHz pulses. Quoted results: May 2005 output-minus-input rose from 50 W at 100 W input to 1000 W at 500 W input (300% efficiency at 500 W); 23 Jun 2005 efficiency 1153%; shorter duty cycles gave higher efficiency; 526% efficiency cited. Kowalski notes a chemist would say the dissociation-recombination cycle cannot give net energy, that the source of any excess is unexplained, that it is pointless to speculate before validation, and that the Langmuir biography on the site ends in speculation about free energy. He cannot follow the device diagram.

Key claims

• Naudin's MAHG output exceeded input by 300% at 500 W input (May 2005) and 1153% on 23 Jun 2005.

Relevance: [secondary] Independent physicist's reading of the MAHG claim; he reserves judgment and raises the energy-source question.

Mainstream conflict: Recombination of atomic hydrogen can only return the energy used to dissociate it (Kowalski's chemist's point), so net excess is outside accepted chemistry

Page of 3-10 Dec 2007 presenting a video from Gravitec Inc. and NASA, run on 3 Jul 2003 in the NSSTC LEEIF chamber at MSFC and held back until 2007 'due to confidential reasons'. Two asymmetrical capacitors on a rotating arm with a torsion-wire axis, voltage raised slowly to +45 kV. At atmospheric pressure the arm rotates; at 1.72e-6 Torr it still rotates but with weaker thrust, returning to its start at 0 V. The page splits the force into an 'electrokinetic' part (ion momentum transfer, only at atmospheric pressure) and an 'electrogravitic' part, called the Biefeld-Brown effect, said to appear only in high vacuum and to be too weak to lift a Lifter; it also speculates about propellantless deep-space propulsion. Horizontal oscillation 'amplified by a parametrical effect'. No numerical force, torque or angle; the video is not saved; the interpretation is Serrano's and Naudin's.

Key claims

• Asymmetrical capacitors still produce a (weaker) thrust at 1.72e-6 Torr and 45 kV; this is the Biefeld-Brown 'electrogravitic' force, separate from ion thrust in air (Gravitec/Naudin).

Relevance: [secondary] The only vacuum-thrust evidence in the folder that is a test (by a company with a stake, Gravitec Inc., and hosted by Naudin); the report is a description of a video, with no data in the text.

Mainstream conflict: Conflicts with the Talley, Tajmar and NASA/ISR vacuum null results; the page states no force values and no controls for charge or outgassing effects

Paper for the 5th IAA symposium, Aosta, 2-4 Jul 2007. Starts from a 1996 Honda R&D experiment (relayed by Musha from his own 2000 paper): a 62 g, 170 mm glass disc capacitor in insulating oil, 8 kV AC pulses at 50 Hz, weight reduction up to 0.9 +/- 0.63 g ('about 3%'; Fig. 4 peak about 1.9 g), with heavy run-to-run scatter, no control runs described. Shows an ion-wind estimate (1.3e-3 g) and his own weak-field electrogravitic estimate (0.002 g; the deep read gets 0.044 g for Z=1) are too small, then proposes that the high field shifts the zero-point-field cutoff and thus electron rest mass (Haisch-Rueda-Puthoff inertia theory), obtaining DM/M = -3.23e-13 R V^2/t^2, hence 1.28 g for R = 1 mm, N = 1e24 and omega_e = 6e13 Hz. Deep-read flags: chosen inputs (N from dense hot plasma; omega_e two orders away from the 1e15-1e16 Hz he gives earlier) drive the result, the constant 3.23e-13 cannot be reproduced from Eq. 13, no zero-voltage control or electrostatic-balance check is shown. Naudin is not named. An advocacy theory paper, not a replication.

Key claims

• A high AC voltage on a capacitor changes the electrons' zero-point-field cutoff and so their mass, giving a weight loss of about 1.28 g for an 8 kV, 1 mm glass disc.

Relevance: [secondary] Theory behind the claim of an 'electrogravitic' vacuum force; its one datum is the 1996 Honda weight loss reported second-hand.

Mainstream conflict: Relies on a minority inertia-from-ZPF theory; ordinary General Relativity weak-field estimates give far smaller effects (Musha's own Eq. 3)

Czech secondary-school project (ENERSOL 2009; Středoškolská technika, CVUT) in which a student copied Naudin's MEG v3.0 (Metglas AMCC-320 core, 2 x 100-turn input and 2 x 1500-turn output coils, TL494 + two BUZ11 drive, three ferrite magnets, 14.4-25 V supply) after an e-mail exchange (17 Apr 2008). Loads: a linear load gave very low output; two 9 W fluorescent lamps (starters removed) made output 'rise sharply'. Instruments: bench supply display and multimeter at the input; school digital voltmeter (AC) for output volts; analogue scope across 10 ohm for output current (peak-to-peak/2 sqrt 2). Table 1: COP printed 2.887 down to 1.568, but 'input power' is U x (I1 - I0) with an undefined no-load current I0 subtracted, output is V_rms x I_rms (apparent) for one lamp then doubled, and power factor is never measured; the deep read recomputes 0.69-1.26 for the measured lamp on full supply power. No heat measured. Naudin's reply (p.29, English) says the MEG works 'as claimed', waveforms replicated, the 'apparent' COP in line with the claim, but no closed loop, the load resistor 'only a bit warm' and 'a possibility of a measurement artifact (not a measurement error)'. Horký's own summary (pp.18-19) drops the 'apparent', 'warm' and 'artifact' remarks. The author is a self-described believer; he also repeats Bearden's claims and his own view that free-energy machines are boycotted. Pages 1-4 deep-read (2026-10-09). Unsourced statements on p.2 (the reference list on p.23 names no origin for any): (A) 'Nyni je J. L. Naudin clenem Beardenova tymu' ('Now J. L. Naudin is a member of Bearden's team'), while p.14 says Naudin took up MEG research 'independently of Bearden's team' and Naudin's own 2008 e-mail (p.29) says nothing about membership (silence, not contradiction; to be checked against Bearden/Naudin sources); (B) Bearden, 'despised in the United States', moved with his team to Hungary and a company Magnetic Energy Limited was founded, 'perhaps' also a factory 'most probably' in Hungary (no date, owner or place); (C) Bearden made an agreement with the 'National Matarials [sic] Science Laboratory' at the Academy of Sciences of 'one friendly country (apparently Hungary - author's note)'; (D) Bearden's team 'succeeded in developing' an automotive modular closed-loop MEG of 2.5 kW, a synchronisation unit joins four to 10 kW, 'however not a single piece has so far left the production line' (no measurement, test conditions, COP or source). Horky's stated reason for belief is that 'patents are issued only for working devices' (p.2). These are a student advocate's characterisations, an interested-party relay; they bear on Naudin's possible stake and on the absence-of-evidence question for a closed-loop MEG (the only closed-loop claim in the paper is this one, with no data; Naudin and Horky both say Naudin had not closed the loop).

Key claims

• A student replication of Naudin's MEG v3.0 measured COP between 1.57 and 2.89 with fluorescent-lamp loads.
• Naudin told the school in 2008 that the MEG works as claimed in apparent COP but is not closed loop and a measurement artifact remains possible.

Relevance: [secondary] Third-party replication of Naudin's MEG and the clearest later statement of Naudin's position (2008 e-mail). It is a school project with RMS/apparent-power methods, not an independent test.

Mainstream conflict: Measurement uses apparent power with non-linear loads; conflicts with the Poynting-theorem argument of Jelinek and Pekarek and the Tektronix note on RMS power

Critique of the Nova Paka MEG project. Argues, with Faraday's law (Appendix 1) and Poynting's theorem for a surface around the ideal device (Appendix 2), that the permanent magnet drops out of the energy balance, mean input power equals mean output power, and the load (including a 'non-linear' load) is irrelevant; with the magnet removed the MEG is an ordinary transformer. Says a magnet holds at most tens of joules, which at 5 W would last tens of seconds. Suggests the students' COP above 1 came from computing power wrongly for non-sinusoidal waveforms and recommends a digital oscilloscope with v x i integration. The deep read finds the proof sound in outline but flawed in detail: sign error in Poynting eq. 5; eq. 10 omits cross terms; eqs. 3, 4, 10 assume a linear core; the Fourier step (eq. 13) needs a lossless, anhysteretic core; the claim that currents and voltages do not depend on the magnet holds only for a linear core; the energy conclusion survives for any lossless anhysteretic core (checked numerically). The authors are an interested party (sceptic organisation), do not re-analyse Horký's numbers and do not discuss Bearden's vacuum theory. The tone about the student is polemical.

Key claims

• The MEG cannot output more average power than it takes in because the magnet cancels out of the energy balance and the load does not matter.
• The apparent COP above 1 comes from wrong power calculation for non-sinusoidal waveforms.

Relevance: [secondary] The sceptical energy-balance argument, with a second-order check by the deep reader; hosts the diagnosis of Naudin-type measurements.

Mainstream conflict: Agrees with mainstream physics and contradicts the MEG claims

A COMSOL finite-element simulation, no experiment, of 1-, 2- and 3-wire asymmetric capacitors in nitrogen (28 kV, 3 cm gap, 25 um wires, positive corona, drift-zone ion model with Peek's law and the Kaptsov hypothesis). Forces per metre: one wire 0.3237 N/m, two wires 0.3517 N/m, three wires 0.2937 N/m; electrostatic force on the electrodes is 94.4-98.4% of the total; peak air speeds 4.295, 3.688, 3.158 m/s. The authors conclude ion wind is 'a reaction to the electrostatic thrust mechanism and not the cause'; they still need ionisable gas. No current, input power, thrust-per-watt or comparison with measurement is given; the deep read notes the model's 'electrostatic' share is the stress split on the solid and that momentum conservation is invoked; table values at 38 Td in the text and 54 Td in the table differ; wire radius 0.25 versus 25 um inconsistent.

Key claims

• Lifter thrust is mainly an electrostatic force between the positive ion cloud and the electrodes, with ion wind as a reaction rather than the cause.

Relevance: [secondary] Mainstream-side simulation that stays within ionisation of gas; includes no experiment, so it neither supports nor refutes vacuum thrust.

Mainstream conflict: Consistent with mainstream electrohydrodynamics; does not support a non-ionic force

Models a triangular lifter (0.075 mm copper wire, 4 cm foil, gap 2.8 cm, 7 g, perimeter 0.6 m) as unipolar positive-ion drift in air with a new analytic Laplace solution (capacitance 4.57 pF calculated vs 4 pF measured), the Deutsch assumption (which the authors admit 'violates some laws of physics'), Peek's onset field (corona inception 7.4 kV) and a Warburg-region cutoff. Result: F = I(1.284 b)/mu and a closed-form thrust formula; calculated thrust matches four measured lift points (7.5 to 13 g at 18.9-22.8 kV) with an average deviation of about 6.5%, calculated current agrees less well (measured corona inception 9.46 kV vs 7.42 kV; measured currents 22-40% below calculated). The deep read finds a unit slip (K1 in mA/kV^2 should be microA/kV^2) and that measured thrust per ampere exceeds the model's by 17-47%. No vacuum test, no discussion of non-ion components. Naudin's site is cited as reference 5. One lifter, repeats unstated, wire possibly damaged.

Key claims

• The levitation force of a lifter in air is calculated within about 6.5% by unipolar ion drift and momentum transfer to air.

Relevance: [secondary] Quantitative ion-wind model that matches one air lifter; says nothing about vacuum or about other force components.

Mainstream conflict: Consistent with mainstream explanation of lifter thrust in air; does not test Naudin's non-ion-wind claim

Boundary-element calculation in 2-D of surface charge on Naudin's 'lifter 1' (0.6 m long, wire radius 50 um, 40 kV, 3 cm gap, plate 4 cm wide) in perfect vacuum. The net electrostatic force is zero within numerical error (Coulomb method: +3.2e-15 N; pressure method -6.2e-4 N falling to -2.4e-4 N with finer mesh) although the plate's rounded leading edge feels about +20 mN opposed by the wire and trailing edge. Losonc compares with W. B. Stein's Purdue vacuum measurement (at least 0.31 mN with a 0.12 m lifter at 17 kV at 1.333e-3 Pa), scaled to Naudin's lifter as 'at least 3.65 mN' (deep read gets 3.57 mN and notes the F2 vs thrust comparison sentence does not follow). He concludes electric forces cannot explain the vacuum thrust and that 'an unknown electro-gravity force' is responsible, by elimination (an advocate of the 'FE principles' site); in air he credits an ion-drift process (Barsoukov's, with unrealistic assumptions) that would be ineffective in a sealed box. The Coulomb-method zero is largely guaranteed by Newton's third law for pairwise forces (deep-read inference). Formulas and figures are images not available.

Key claims

• The net electrostatic force on a Naudin lifter in perfect vacuum is zero or negligible, so the measured vacuum thrust needs another cause (Losonc: an unknown electro-gravity force).

Relevance: [secondary] Third-party calculation relevant to the Naudin vacuum claim; the stated conclusion by elimination is the author's. The vacuum thrust numbers he cites (Stein, Brown 1973 letter) were not read here.

Mainstream conflict: The zero-net-force result agrees with standard electrostatics; the 'electro-gravity' conclusion has no mainstream support

Saved encyclopedia article. States that detailed vacuum-chamber studies failed to replicate Brown's observations and follow-up studies attribute the force to corona wind; summarises Brown's history (Coolidge tube, 1927 application, 1929 Science and Invention 'gravitator', Biefeld at Denison, no Denison record of the association as of 2004), his 1965 vacuum-force patent claim, Talley 1988 (no force in 1e-6 torr vacuum under DC; a force during breakdown), Tajmar 2004, the ARL, NASA/ISR and a 2024 Tajmar et al. Scientific Reports paper (arXiv:2402.15640). Lists Brown patents. Some paragraphs carry [citation needed]. This is a general-audience summary, not a primary source. Deep read: Tajmar, Kossling and Neunzig (Sci. Rep. 14, 19427, 21 Aug 2024; arXiv:2402.15640) is cited twice, always beside another source and with no sentence on what it did or found: the lead ('Detailed studies in vacuum chambers failed to replicate Brown's observations and follow-up studies attribute the force measure to corona wind', refs 3, 4) and the 2004 Tajmar box test (refs 5, 10, 4). Talley is cited once (ref 9, report dated 7 July 1990): 'In 1988, R. L. Talley measured no thrust from electrodes similar to those proposed by Brown operating in 10^-6 torr vacuum under direct current potentials. He did find a force during electrical breakdown when current was flowing.' (measurement 1988, report 1990; the pin cite ':216' does not fit the printed ref 5). The sentence 'Experiments have been conducted and found to both prove and disprove a force at very low pressure' carries no citation and names no experiment; the editors' uncited reason is that only very-high-voltage runs gave positive results, with forces small against atmospheric ones. The article itself therefore concedes unnamed positive vacuum results while its lead asserts corona wind.

Key claims

• Vacuum-chamber studies did not replicate Brown's observations and the measured force is attributed to corona wind.

Relevance: [secondary] Provides the mainstream summary and one pointer to a 2024 test (Tajmar, Kossling, Neunzig) that is not in the folder.

Mainstream conflict: Opposes the claim of a non-ionic Biefeld-Brown thrust