RankingsPrimerArbiterPhysicistMathematicianContrarianQESites logClaimsSourcesResearch notes

Naudin-JL: central claims (Step 1b, fact-finder)

Source gap

Almost all of Naudin's own pages (JLN Labs, 1996-2013) are saved as HTML text, but the experimental data they point to are mostly not in the text layer: scope captures, graphs, photographs and videos were not retrieved (README, "Saving method"), and many sub-pages that hold the numbers were left out (CFR v1-v4 and the 33-run data page, MAHG runs 59-83 and the data table, Lifter v2/v3/Maximus/theory and the Faraday-cage, enclosed-box, 'T. T. Brown EKP device in MINERAL OIL', smoke, gamma and ion-wind pages, GEGENE tests 1-20, TEP, DLE tests 1-25, Orbo/2SGen scope images). Works whose key numbers are NOT readable here (text_available "partial"): the CFR results page (the 33 runs appear only second-hand in Kowalski item 73), the MAHG one-hour COP 21 page (figures only), the GEGENE and WFC-booster pages (menus only), the dstl slides (charts; now read as page images by the QE deep read, values are estimates), and the Losonc vacuum analysis (formulas and figures missing). Shawn Bishop's original MEG critique, the Infinite Energy articles on Naudin, and the Gravitec vacuum video itself were not obtained. Later steps should say plainly when a conclusion rests on a description of a result instead of the result. To obtain: the CFR data pages (jlnlabs.online.fr/cfr), MAHG data table, Lifter Faraday-cage / enclosed-box / mineral-oil / ion-wind pages, the GEGENE test pages gegene07en.htm (calorimetric), gegene08en.htm (closed loop, Romero UK) and gegene13en.htm (closed loop via grid-tie inverter), the CFR replication pages (nfrcnam/, aehodme/, lorio/report/, logbook/rh/, ppclkrs/, html/cfrmhmc.htm, html/cfrppcglv.htm), Tajmar, Kossling and Neunzig (Sci. Rep. 14, 19427, 2024), GEGENE tests, the Gravitec video, Talley's Air Force reports (AFAL-TR-88-031, PL-TR-91-3009), Stein's Purdue vacuum paper, and the Dstl Lifter report.

Provenance. The folder separates Naudin's own pages and patents ("primary", 38 entries) from third-party tests, critiques and theory ("secondary", Related-Research/, 17 entries). The README is a compiler's retrieval catalogue. No document is an earlier AI-written analysis. Many performance numbers on Naudin's pages are reports by interested parties (Naudin, Bearden's company, Bedini, Gravitec Inc., the patent applicants) and are marked as such in facts.json.

Central claims (what the documents claim; no verdicts)

  1. Lifter thrust is not simply ion wind. Naudin states that the Lifter's main thrust in air is an electrokinetic effect on the medium rather than a simple ionic wind, that a plate still flies toward a charged tungsten wire sealed in a vacuum tube with no corona (about 25 kV, 25 mW), and that asymmetrical capacitors still produce a weaker "electrogravitic" Biefeld-Brown thrust in high vacuum (Gravitec/NASA NSSTC video, 1.72e-6 Torr, 45 kV, as described on Naudin's page; Brown's own vacuum claims reach the record through the Appendix B letter in the ARL report). Bearing on it: the ARL (Bahder and Fazi), NASA/ISR (Canning et al.), Tajmar, Ianconescu et al. and Martins and Pinheiro explain air thrust by ion drift and momentum transfer to air; Talley (quoted in two reports) saw no static-DC vacuum force; Tajmar and the NASA/ISR report found none in sealed-box and low-pressure tests (with stated limits of sensitivity); a Dstl slide states calculated ionic thrust is more than an order of magnitude below the observed lift (the slide shows no calculation; the QE deep read of Dstl's own chart currents and wire heights with F = I d / mu, mu = 2e-4 m^2/(V s), gives 2 to 5 times the lifted weight, so the ion-drift bound is above, not below, the observed lift; Dstl also reports 'weight loss observed prior to lifting' on a balance, about 5-10% by 17-22 kV, which a gradually growing ion force would also produce); the Wikipedia copy cites Tajmar, Kossling and Neunzig (2024) only as a supporting reference, without describing its content, and says experiments 'both prove and disprove a force at very low pressure' without naming them; other items in the folder that bear on air dependence: Honda R&D oil-immersed capacitor weight loss (via Musha), Brown's air-versus-transformer-oil torque statement (ARL Appendix B) and a June 2001 Transdimensional Technologies test of the NASA thruster in the MSFC vacuum chamber (Naudin's one-sentence relay, no pressure, voltage or result); Losonc's calculation gives zero net electrostatic force in vacuum and concludes "electro-gravity" by elimination.

  2. Lifters are real, reproducible and scale. Naudin reports a 2.3 g Lifter1 hovering with 1 g payload at about 24 W (41.9 kV, 570 uA), a 32 g hexagonal Lifter v4 hovering at about 81 W and flying at about 133 W, a 194 g "Lifter Maximus" lifting 60 g, and 351 self-registered replications worldwide by October 2006. The ARL built three lifters and a wire that lifted at 30-37 kV; Ianconescu et al. matched four measured lift points (7.5 to 13 g) within about 6.5% by an ion-drift calculation.

  3. Naudin's MEG replication gives an "apparent" COP above 1, open loop. Naudin reproduced the waveforms of Bearden's Motionless Electromagnetic Generator (permanent magnet plus Metglas core, two actuator and two collector coils) and reports measured COP 1.75 (7 Nov 2000) and, for the 9 W lamp run, 6.76 W out against 3.25 W total input or 1.5 W after subtracting the 1.75 W control-board draw (Horky reads this as COP above 4.5; 6.76/3.25 is 2.08); Bearden's paper (inventors' own data) shows about 2.7 per coil and a claimed net 5.4. Both say the loop was never closed. Horky's paper (p.2, no source given) adds that Bearden's team developed 2.5 kW closed-loop automotive MEG modules (four joined for 10 kW, 'not a single piece' produced), that Naudin 'is now a member of Bearden's team' (contradicted in tone by Horky's own p.14, 'independently of Bearden's team', and not mentioned in Naudin's 2008 e-mail), and that Magnetic Energy Limited was founded; these are a student advocate's unsourced statements, kept in the record as such. Naudin records that the load resistor warmed less than the apparent power implies and that "a measurement artifact" remains possible. Bearing on it: a Tektronix support note (via Flickinger) says RMS of a math power waveform "has no value"; the paper's own scope readings are such "M RMS" values; a Czech student replication (Horky, COP 1.57-2.89, apparent power, undefined no-load subtraction, no heat measurement); and a Poynting-theorem argument (Jelinek and Pekarek) that mean output equals mean input, which the deep read finds sound in outline with flaws in detail. Smith (hosted by Naudin) replies that Bishop's critique used a wrong power formula.

  4. The Mizuno-Ohmori plasma-electrolysis "Cold Fusion Reactor" gives excess heat reproducibly. Naudin replicated the reactor on 7 May 2003 and reports 33 runs by 25 Jun 2003; Kowalski quotes run 28 (170 g of water evaporated, 1200 W in, 3093 W of heat) and a blank run; Naudin's index lists replications at the CNAM Paris (undated on the index; said to be presented at ICCF-12, Yokohama, Nov-Dec 2005, in a transmutation paper by Dattilo, Iorio and Cirillo), by Ronald Hohls in South Africa (19 Sep 2006), at Boulder, Colorado (15 Mar 2006), by Morgan H. and Marissa C., and by an unnamed French lab, all described as successful in Naudin's own words; the reports are not in the folder. Naudin thanks Prof. Pierre Clauzon (help in some test runs) and Prof. Jean-Francois Fauvarque and Gerard Lalleve of the CNAM Laboratoire d'Electrochimie Industrielle (scientific support), and wrote that 'some French physicists work closely with me'; none of them is quoted on the results. Only the summary is in the folder; the run data are not.

  5. Moller's Atomic Hydrogen Generator (MAHG) gives COP 5 to 21. Naudin's logbook lists COP 2.43 (v1.0), 5.26 (v2.0) and 21 (v2.1, 1 h, 5% duty cycle, 51 Hz); Kowalski cites 300% at 500 W input and 1153% on 23 Jun 2005, notes the net energy source is unexplained and does not accept or reject it. Figures and numbers per run are not in the saved text.

  6. Two Cornille-Naudin patents claim lower losses and "efficiency more than one". FR 2 789 818 A1 (1999, Advanced Electromagnetic Systems S.A.) claims a self-switching ferrite-transformer charger that charges a 1.3 F supercapacitor bank in 30 s instead of 83 s with over 90% charging efficiency (no definition or method given; no overunity claim). US 2005/0057116 A1 (priority 2001) claims that a plate capacitor with switched plasma-tube electrodes yields energy greater than that invested (abstract: "efficiency of more than one"), with a computed 338% for an assumed multiplier of 200, three different unreconciled accounts of the extra energy (work over Debye lengths, "surrounding magnetic ether", Earth's potential energy), and one measured trace (500 V to 1750 V at constant charge) with no input-energy measurement.

  7. Other devices Naudin hosts or replicates are presented as overunity or "regauging" machines, with mixed self-reports. These include Bedini's G-field generator ("120% or better", Bedini's statement), Mini-Romag (24 W self-sustaining, no data), Steorn Orbo (loading leaves the 400 mA input unchanged, input voltage and LED power not stated; Naudin's page says both that Larsko's loading test 'has demonstrated an overunity effect' with 'measured COP = 4.75' and, in his comment, that the device's COP 'is less than one' and 4.75 'is NOT A COP', without reconciling them), 2SGen (load powered from core demagnetisation), KAPAGEN (up to 2100 W of lamp light by tapping Earth electrons, input not given), GEGENE (induction cooker plus Tesla bifilar coil, "great efficiency"; the index page lists a calorimetric test #3, a closed-loop test #4 by 'Romero UK' and a grid-tie closed-loop 'Phase 2' test #9 by title only, with no results in the folder), Meyer-type water fuel cells, and Hendershot (Epps's "non-reciprocal" argument, which Epps says did not hold at higher voltage in a test with Naudin). Naudin's own 1999 closed-loop note says the Bedini Magnetic Gate shows no overunity.

What rests on the digest alone (no deep read)

All of Naudin's own HTML pages except the two patents are digest-only; their key numbers are therefore the digests' transcriptions of text, not page images. Central items resting on digest alone: the MEG v2.0/v2.1 pages (claim 3, Naudin's own COP data), the Lifter vacuum-tube pages (claim 1), the Lifter1 and v4 data (claim 2), the CFR and MAHG pages (claims 4 and 5), the Kowalski items, Smith's critique and the dstl slides (including slide figures pp.28-37 that are charts only). Done after QE (2026-10-09): deep reads of the dstl slides (page images), Horky pp.1-4, the CFR index, Kowalski item 73, the Orbo page, the ACT page, the Wikipedia copy and the GEGENE index. Still digest-only: the MEG v2.0/v2.1 pages (scope-screen figures if recovered) and the other Naudin pages.

Corrections to earlier documents

  • The step 1a digest of the Czech Horky paper (digest/Related-Research-2009-MEG-Horky-replication-...-2e29e5.01.md, p.18-19 and p.20 lines): the deep read shows (a) Horky's bullet "this is not a measurement error" reverses the emphasis of Naudin's own e-mail on p.29, which says "a possibility of a measurement artifact (not a measurement error)" and adds that the load resistor is "only a bit warm", both omitted by Horky; (b) the table row labelled "I1-I2" is really I1 - I0 (an undefined no-load current) so the printed "input power" is not measured supply power. facts.json describes it accordingly and cites p.29 for what Naudin said.
  • The ARL-TR-3005 report's copper-ion ionic-wind figure (0.002 g, eq. 6) recomputes to 0.023 g with the report's own inputs, so "3 orders of magnitude too small" is about 2 orders for that case (the electron case checks); the report's ion-drift estimate is the collisional (ordinary ion wind) form, so it should not be cited as showing that ion momentum transfer in air is too weak. Source: deep read of ARL-TR-3005.
  • Martins and Pinheiro (arXiv 1009.6111) contains no experiment, only simulation; any entry that describes it as reporting measured thrust is wrong.
  • Digest of the FR patent and the deep read agree; the patent itself gives 2.5 times and "divided by 3" for the same example (83 s vs 30 s is 2.77).
  • CFR index dates (facts.json, entry 2003-CFR-plasma-electrolysis-reactor-index; the digest and the first facts.json read the date lines in the wrong order): the HTML shows Ronald Hohls = 19 Sep 2006, Boulder = 15 Mar 2006, and the CNAM item carries no date on the index (ICCF-12 presentation Nov-Dec 2005). Earlier text that gives 19 Sep 2006 for the CNAM replication or 15 Mar 2006 for Hohls is wrong.
  • QE (qe.md section 3, item 2) suggested the CNAM replication may be the Fauvarque-Clauzon-Lallevé paper at ICCF-12 [memory, medium]. The deep read of Naudin's index finds that the ICCF-12 paper linked from the CNAM item is named as 'Transmutation of metal in confined plasma in water' by Datillo [sic], Iorio and Cirillo, not by the CNAM names; the link between that paper and Naudin's CNAM electrochemistry colleagues is not established by any document in the folder. Retrieve both before weighing.
  • The Dstl slide statement (p.37) that calculated ionic thrust is 'more than an order of magnitude less than that observed' was carried in the first claims.md and facts.json as a Dstl result; the deep read shows the slides give no calculation and that Dstl's own chart currents and wire heights with F = I d / mu give an ion-drift bound 2-5 times the lifted weight (d assumed equal to the wire height; chart values estimated). Anything citing the statement as an established result should add this.
  • Talley's vacuum measurement is dated 1988 in the Wikipedia sentence and the cited report 1990; do not write '1990' for the measurement.
  • Horky p.2 statements on Bearden's team, Magnetic Energy Limited, the 'National Matarials Science Laboratory' and the 2.5 kW / 10 kW closed-loop modules have no stated source; they must not be cited as documented facts. The Hungary location for the laboratory is the author's own guess ('zrejme v Madarsku - pozn. autora').
  • The digest of the Orbo page and earlier summaries that give either 'claimed COP 4.75' or 'rejected overunity' alone are incomplete; Naudin's page contains both statements (see claim 7).
  • The ACT page's TdT test (June 2001) is Naudin's relay with a photo link; earlier text that treats it as a documented vacuum result is overstated, and one that omits it misses a third Huntsville vacuum test alongside Gravitec/NSSTC 2003 and MSFC 2004.