Meyers-Stan (Stanley A. Meyer, "Water Fuel Cell"): QE (Step 4b)

QE, 2026-10-09. Inputs read: prompt.txt; _text/manifest.json and _textprep.log; facts.json, claims.md (in full); arbiter.md and primer.md (in full); physicist.md, mathematician.md and contrarian.md (by section); sites.md (Leads for review in full); the deep/ reads where a check needed them; the extracted text and rendered page images listed under "Accuracy spot checks". example.py was run. Tags: [read] extracted text; [read: page image p.N] a page I rendered (to /tmp/qe-meyer); [calc] my arithmetic (.venv python); [memory] with confidence.

Verdict: MINOR

The existing verdicts hold up. Every load-bearing number I checked against the source is right, and example.py runs and reproduces the primer. Three gaps keep this from a PASS: 1. A central claim is missing from the claim list. In both 1989 talks Meyer says outright that the system is self-sustaining: it "maintains itself as long as there is water in the fuel cell", making electricity "without consuming the fuel source". This claim is in the Einsiedeln deep read (section 5.11) and in the New Zealand transcript. It is not in claims.md, any of the three reviews, the arbiter or the primer. 2. Step 0 was re-run after the fact-finder. The re-run produced a 153 KB transcript of the 2 h 38 min New Zealand video and an OCR of the photos. Every later file still says the video was "never transcribed", and nobody has read the transcript. 3. Meyer's own 1983 experimental graphs were never read. EP 0 111 573 A1 has graphs of gas against voltage, current, plate spacing and water type, with a 12 V, 1 A operating point. The patent was only front-page skimmed. Its graphs may hold the only quantitative bench data Meyer published in a patent.

None of these is likely to reverse a verdict. The self-sustaining claim will almost certainly get "Refuted by calculation" from the energy accounting already in primer section 3.4. The 1983 graphs, if they show gas rising with current and the water heating, support the existing reading. So the fix is the MINOR route: deep reads, then updates to facts.json and claims.md, a "QE revision" section in arbiter.md that gives the new claim a verdict, and a primer update. The orchestrator may escalate to MAJOR if it wants the three reviewers to weigh the self-sustaining claim themselves.

1. Coverage

2. Shortcuts

• Claim 1 (water capacitor). The verdicts rest on the deep reads of WO 92/07861, US 5,149,407 / WO 89/12704 and US 4,798,661, the fact-finder's full read of US 4,936,961, and Birth pp.1-56. That is adequate for the 1990s form of the claim. The 1983 generator patent (EP 0 111 573 A1) is the shortcut: it is the earliest document with Meyer's own operating point and graphs, and nobody read it.
• Claim 2 (fracturing). Adequate. Memo 424 was read in full by three reviewers.
• Claim 3 (vehicles). The verdict rests on Appendix A (read), WO 92/22679 and WO 92/08046 (deep-read). The 1982-83 injector family was not read. Probably no engine data there (my grep found no numeric operating points in US 4,389,981 or CA 1,231,872), but it is unconfirmed.
• Claim 4 (EPG). claims.md and the arbiter say "no numbers, no measurement". The New Zealand talk gives Meyer's own numbers and mechanism: a 10 W laser "struck a thousand times" compounds the field; 1,000 turns per coil; a 20-mile loop to Auckland; a home unit "half size of a hot water tank, about 18 inches in diameter" with a garbled "2 2300 amp" capacity; a magnetized gas of iron ions and argon at 2,000 to 90,000 V [read: NZ transcript, folded lines 264-298 in my scratch copy]. Section 5.11 of the Einsiedeln deep read also has a "gas battery" scale figure (4 x 10^9 electrons = 6.4 x 10^-10 C). The verdict "Not testable as stated" should be revisited against these statements. The laser-compounding claim is an energy-accumulation claim, and a calculation can check it.

3. Missed material (claims, experiments, people, events, numbers)

There are no digests, so I searched the unread source text directly (NZ transcript; Birth pp.126-221; photos OCR; family patents).

A. The self-sustaining (closed-loop) claim. Missing everywhere downstream of the deep read. - Einsiedeln 1989: "so the process is a self-sustaining oscillation system that once you excited from an external source it maintains itself as long as there is water in the fuel cell"; and of the gas-pressure turbine, "I'm now producing electrical energy without consuming the fuel source" [read: deep/Video-...Einsiedeln...md, section 5.11, transcript L.557-581]. - New Zealand 1989: the "electron extraction circuit" draws electrons from the water to make electricity ("100 volts times 100 amps ... 10 kilowatts of electrical energy"), which is fed "back through the voltage intensifier circuit ... so I now have a self-sustained oscillation system ... it goes until ... there's no more water" [read: NZ transcript]. - In the same passage Meyer reports that at the Pentagon "General Abramson" (whisper's spelling) said "man, it's a perpetual motion device". Meyer replied that it is not perpetual motion because "I'm using water now as the energy source just like ... gasoline" [read: NZ transcript]. The identity of the general is unverified. A Lt. Gen. James Abrahamson headed the Strategic Defense Initiative Organization in 1984-89 [memory, medium], but nothing in the folder links him to this meeting; a web check is needed before any name is given. - Why it matters: this is Meyer's own over-unity claim in its plainest form, and it is the reason the "complies with thermodynamics" statement and the barrel figures conflict. It should be claim 7 (or a sub-claim of claim 2), with its own verdict. The calculation is short: water is the product of hydrogen combustion, so any loop that splits water and burns or recombines the gas returns at most ΔH = 285.8 kJ/mol against at least ΔG = 237.1 kJ/mol to split it. Every conversion loss then makes the loop run down.

B. Meyer's own 1983 measurements (EP 0 111 573 A1, Hydrogen Generator System, 38 pp.). - "for the embodiment shown in Figures 2 and 10, the source 30 supplied 12 volts at 1 ampere" [read: p.12]. Faraday's law gives 0.685 L/h of mixed gas for 1 A [calc]. - "the voltage current versus gas generated is illustrated graphically. The gas generated is a linear function to magnitude of voltage applied ... As the voltage/current is increased the temperature of the water increases and eventually reaches a condition where steam is generated" [read: pp.12-13]. "The current versus gas generated is also a function of the type of water utilized" (Fig. 9). Efficiency against plate spacing is "a linear decrease with spacing" (Figs. 8 and 15, "for each of the waters tested") [read: pp.13-16 by grep]. - The patent says prior-art electrolysis "can produce hydrogen with an efficiency of only 37%" (OCR "377") [read]. - Why it matters: gas linear in current, heating and dependence on water type are the signatures of ordinary electrolysis, written by Meyer himself in 1983. claims.md's "No gas rate, current, input power or energy balance is given anywhere" is already corrected by the arbiter for the 1995 binder. It should also cite this patent. The graphs may give numbers, so read them as page images.

C. Meyer's own water analysis (Birth p.163, Fig. 7-15, Memo 426). - An "Analysis of Water" chart credited to "R & D Laboratory, A Certified Biological Testing Laboratory, Columbus, Ohio 43204, C. Thomas Bell, Ph.D., Lab Director, Test Date: January 10, 1983, Test Method: Spectrophotometer Analysis, Contaminates: Sodium (Na) and Potassium (K)". Values: ocean 26.1K ppm, river 45, drilled 24, lake 20.2, rain 16.1, tap 11, distilled under 2 ppm [read: page image p.163]. The same chart labels the dielectric value of water "78.54Ω" (the unit error the reviewers found on pp.142-143). - Newsletters 10 and 11a reprint the chart; their deep reads noted it and read the date as "1963" [read: deep/...No-10...md L.38; deep/...No-11a...md L.44]. The page image p.163 of Birth prints 1983. - Why it matters: the primer's Q_cell argument uses a generic tap-water range (100-500 μS/cm). Meyer's own data gives a source-based figure. Read as NaCl, 11 ppm Na is about 60 μS/cm (σ ≈ 0.006 S/m, τ ≈ 115 ns, Q_cell at 5 kHz ≈ 0.004). Read as KCl, it is about 42 μS/cm (Q_cell ≈ 0.005). Meyer's "distilled, under 2 ppm" is about 11 μS/cm (Q_cell ≈ 0.02) [calc, molar conductivities Na+ + Cl- 126.4 and K+ + Cl- 149.9 S cm^2/mol, memory high]. Hardness ions (Ca, Mg) were not measured, so real tap conductivity is higher. The conclusion (the cell is a resistor at kHz) holds even on Meyer's own water figures. This is a stronger, source-based way to make the primer's central point. - p.190 (Memo 429) cites a "WFC Patent Validation Report dated January 14, 1983 as per WFC Test-Results ... to demonstrate operability" under 35 USC 101 [read]. This is a named test document not in the folder, and it belongs in the arbiter's records list. The 1995 binder transcribed at stanslegacy may contain it; the Investigator should check.

D. New Zealand 1989 house meeting: other content not in the record [read: NZ transcript; whisper speaker attribution is not given, so who said each line is my inference from context] - Context: an investor/manufacturing meeting. Meyer proposes decentralised manufacture in New Zealand, "small amounts of money ... to set up the manufacturing procedures in New Zealand", and a production unit "in 18 to 24 months". One participant (apparently not Meyer) says a demonstration model is not important: "it doesn't need to be demonstrated it just needs to work". This bears on the finance/stake section and on the "no demonstration" silence (arbiter section 10A). - Operating ranges stated: drive "from zero up to 10 kilohertz ... 20 kilohertz"; "zero to 2,000 volts"; fracturing gas at "2,000 volts to 10,000 volts ... 20,000 volts up to 90,000 volts". Memo 425 likewise says voltage rises "up to and beyond 90,000 volts range within a millisecond or less" [read: Birth, Memo 425]. The arbiter's open question 2 asks for pulse lengths. "Within a millisecond" is the only time scale I found, and it is a rise time, not a pulse length. - Aircraft: "we can literally fly a tactical aircraft like an F-18, an F-16 ... off the atomic power of water"; "retrofitting ... a Learjet like the T-39". This adds to the Jet Commander and F-15/F-16 statements already in facts.json. - Military contacts: "General Bachard" (spelling as transcribed) and staff at Wright-Patterson Air Force Base; the Pentagon meeting above. Identities unverified. - Mechanism analogies: muon-catalysed fusion ("the muon process") offered as precedent for "hydrogen fracturing"; "Unaculum's law" (whisper, apparently Coulomb's) and 6.25 x 10^18 electrons per ampere-second. Meyer again says the "2.5 times gasoline" burn figure comes from the "National Bureau of Standards"; Appendix A says the same: "water is 2.5 times more powerful (gtnt) than gasoline (U.S. National Bureau of Standards)" [read: page image p.221]. The arbiter discusses the 2.5x figure but not Meyer's attribution of it to NBS.

E. Birth Memos 425 and 427-430: claims not in claims.md [read: Birth pp.126-219, headings and scan] - Memo 428 "Exhaust Air Reclaimer": the electrical polarization process "liberating chemical-oxides oxygen O2 molecules once again for reuse", with atmospheric oxygen "less than 20% now as compared to 30%". The NZ talk gives "38 percent or 28 percent" historically. Both are checkable statements [memory, high: about 20.9% today]. - Memo 429 "Optical Thermal Lens: A Miniature Controllable Sun", which asks "How is it mass is being created in our universe?" Memo 430 "WFC Steam Resonator". Memo 427 "Energy Vectoring" by cavity shape. - Memo 425: fuel-injector voltages "up to and beyond 90,000 volts ... within a millisecond"; capacitor film "endure ... -40 to 155 C"; a "lab tested less than .0001/year decomposition rate". - These are mostly qualitative, but the atmospheric-oxygen figures and the steam resonator are claims a reader would want listed (claims.md "Meyer's claims" or a peripheral-claims note).

F. photos/ (OCR) [read: _text/photos-d68f0b.txt] - Most images carry the watermark "Gries Petty Research, RWGresearch.com / AlexPetty.com". The photographs therefore come from a later third party, and facts.json should give that provenance; their relationship to Meyer or his estate is not in the folder. - One placard reads "FlameLess Central Water Heating System Using 'Water as Fuel' ... WFC Heat Resonator ... Proposed Packaging Concept. Now, under WFC Industrial Contract: Europe. Copyrighted 1997 By Stanley A. Meyer". That is a 1997 claim of a European industrial contract, which bears on the Belgian-investor story and on the post-1998 silence (arbiter section 10C). Others show a "Voltage Intensifier Circuit Multi-Coil Spool Assembly" listing US patents 4,826,581, 5,293,857, 4,936,961, 4,389,981 and 4,613,304, and a "GMS Unit" panel with "Hydrogen Fracturing Process / Resonant Cavity Fuel Cell". - Surviving hardware is a physical record. The probate file shows the "83 DuneBuggy" transferred 5 May 2000 (sites.md). Neither the buggy nor the photographed hardware appears in the arbiter's "what would settle it" list.

4. Accuracy spot checks (against extracted text and page images)

5. Retention (Investigator's findings and source material left out of arbiter or primer)

Most leads are kept, and kept fairly: Nagypál with his caveats, Perkins, Hindley/Laughton/Griffin, Nieper, the NIST and DTI reports, the court, the coroner, the Belgian investors, and the patent lapse. The primer's "Kept in the record whatever the verdict" paragraph is good practice. Items dropped between sites.md and the arbiter/primer:

1. Lammertink's scrap-steel alloy hypothesis for why Meyer's first cell "worked then stopped", linked to Meyer's 1982 steel-analysis affidavits (Robert L. Queen), and the Pollack exclusion-zone interpretation (sites.md, Closely related work). Neither appears in any review except a passing Lammertink citation in the contrarian (about the ABC footage). These are hypotheses outside mainstream science that bear directly on claim 1. Under SOURCE HONESTY they stay in the record with their source, whatever the reviewers think of them. Defect.
2. Griffin's Warsash lecture numbers (Sep 1993: 10 W in, gas at 10 psi within 10 s, vessel cold). The physicist and mathematician weigh them. The arbiter omits them, and the primer names the lecture with no numbers. They are a quantified witness observation and should be in arbiter section 9, with what is missing (vessel volume).
3. Dublin Institute of Technology explanation in the 1995 binder: dropped.
4. Contemporary press and broadcast record: Channel 4 "It Runs on Water" (17 Dec 1995); Ship & Boat International / RINA (Jun 1991); Asian Defence Journal (1992); IECEC 1993 (Bailey and Grotz); Fortean Times (1997); Air & Space Smithsonian (Oct/Nov 1991). Only Spy and Air & Space reach any review. Add these to arbiter section 12's records and to the primer's further reading under "Later accounts and media".
5. Other recorded talks (1992 Global Sciences Congress "Voltage Disassociation of Water", 1993 interview, Internet Archive lectures): not in the arbiter's records list.
6. Earlier investor litigation: Franklin County 79CV004853 (Schuler/Leukart v. Meyer, 1979-81) and 93CV005477 (Schneider, Brooks and Gilvesy, settled 1994; Brooks invested over $300k). The arbiter mentions 93CV005477 only to separate it from the 1996 case. A 15-year pattern of investor suits is pertinent to the stake and finance section, and the primer has none of it.
7. Surviving hardware (the dune buggy transferred in probate on 5 May 2000; the hardware in the Gries/Petty photographs) is not listed as a record that could settle claim 3b.
8. The primer's further reading omits Kucherov US 5,632,870, Graneau et al. (Phys. Lett. A 1986), Gary L. Johnson (Kansas State) and the Shimizu pulsed-electrolysis papers. All four are related work the reviewers discussed. Meyer himself cites the Northeastern "Powerful Water-Plasma Explosion" and the Kansas State "Electrically Induced Explosion in Water" reports in Appendix A [read: page image p.221].

None of these omissions changes a verdict. Items 1 and 4 are the ones the reader would most miss.

6. Process

• Source tags: present and honest throughout. The primer carefully marks the stanslegacy material as "machine transcriptions ... unverified" and the interested parties' stakes (Meyer, Nagypál, Perkins, Griffin). Good.
• Web checks: done and logged (physicist and arbiter in full; contrarian 2 of about 6, with a reason). One medium-confidence [memory] fact still supports a verdict: the bond-scale field of 10^9-10^10 V/m behind "Refuted by calculation" for bulk-field bond breaking. The arbiter argues, reasonably, that the web-checked Wien-effect threshold (10^8 V/m) already exceeds the applied field. The check is still owed; one search would close it.
• Verdict scale: applied consistently. Split verdicts are clearly separated. "Historically superseded within Meyer's own work, but not by a test" (claim 5) is worded as a note, not a scale verdict. Acceptable.
• Primer vs STEP 4 and the reader profile: all nine sections are present and in order. The worked example is strong and reproducible, symbols are defined, and the answers are at the end. Small items: "STP" is used in answer 5 without definition; "Debye relaxation" (mathematician) does not reach the primer, which is fine. The primer's "distilled, 1 μS/cm" row is optimistic against Meyer's own "under 2 ppm" (about 11 μS/cm); see 3C.
• example.py: runs (exit 0) with .venv/bin/python. Every printed figure matches primer section 5 (821 pF, 1.23 H, Q_cell 0.00044 to 3.97, peak gain 4.0 at 4.9 kHz, 69/83 kW, 56,317 A, 28.6 vs 27.4 L/h, 2.54 x 10^8).
• Stale process statements: "Step 0 left no manifest.json" (claims.md, arbiter) and "New Zealand video not transcribed" (claims.md, facts.json, all three reviews, arbiter, primer) are now false. facts.json how_read for newsletters No. 4 and No. 8 and photos/ says "NOT extracted by Step 0"; they are now in the manifest, with 0 text chunks for the newsletters and an OCR file for the photos.

Corrections to earlier documents (for the orchestrator)

1. arbiter.md section 5 and Correction 4; mathematician.md: "he writes '740 ml/sec', which should be 0.74 ml/s" is wrong. Birth p.221 prints ".740 ml /sec" [read: page image p.221]; the decimal was lost in OCR. Delete that slip and keep "3.3 amp/hr" (real).
2. arbiter.md Correction 9; contrarian.md Corrections item 6: the "1989" date for licence #490606 is on p.1 of the viXra 233-page copy (research/files/vixra-0811.0001), not on p.1 of the folder copy (a cover). The folder copy p.18 prints 1988 [read: page image p.18]. Record it as a discrepancy between two Meyer editions. sites.md Leads ("490606, 15 Nov 1989 ... cited on Birth p.1") should name the viXra edition.
3. claims.md, facts.json, arbiter.md, primer.md, all three reviews: "the New Zealand video was not transcribed" is now wrong. A whisper transcript exists (_text/Video-Stanley-Meyer-1989-House-Meeting-in-New-Zealand-4ed44c.txt, Step 0 re-run 2026-10-09 02:20). "Step 0 left no manifest.json" is also wrong now.
4. deep reads of newsletters 10 and 11a: the water-analysis chart's test date is "January 10, 1983" in Birth p.163 [read: page image]. Check the newsletter scans before keeping "1963".
5. claims.md claim 4 (EPG) "No numbers": the 1989 talks give Meyer's own figures (section 2 above). The statement should say "no measurement".
6. claims.md claim 1 "No gas rate, current, input power ... anywhere": besides the arbiter's 1995-binder correction, EP 0 111 573 A1 (1983) gives 12 V at 1 A and graphs of gas against voltage, current, spacing and water type.

Remediation (one round)

Deep reads, most important first (slug from manifest.json):

Process action (orchestrator): run Step 1a (Haiku) on all 84 chunks so a digest record exists for the recheck.

What each later file must change: - facts.json: add the NZ transcript entry (how_read, summary, key claims). Update how_read for newsletters 4 and 8 and photos/ (now in the manifest; photos OCR'd; Gries/Petty provenance). Update EP 0 111 573 A1 and the Birth entry after deep reads #2 and #3. Add C. Thomas Bell's 1983 water analysis and the 14 Jan 1983 "Patent Validation Report" as a cited, not-obtained record. - claims.md: add claim 7, the self-sustaining loop ("once excited ... maintains itself as long as there is water"; electricity "without consuming the fuel source"; electron extraction feeding the VIC) [Einsiedeln deep read 5.11; NZ transcript]. Amend claim 4 ("no measurement", not "no numbers") and claim 1 (1983 operating point and graphs). Update the source-gap note (manifest exists; NZ transcribed; Memo 424 closed). List peripheral claims (exhaust air reclaimer and atmospheric oxygen, steam resonator, optical thermal lens) under Meyer's claims. - arbiter.md, "QE revision" section: give claim 7 a verdict, with the calculation (expected: Refuted by calculation). Revisit claim 4 against the EPG figures and the laser-compounding statement. Add Meyer's own 1983 data (deep read #2) and the Bell water analysis to the claim 1 discussion, and the "within a millisecond" rise time to open question 2. Apply Corrections 1-2 and 4 above. Restore the retention items in section 5 (Lammertink/Queen alloy and Pollack hypotheses; Griffin's Warsash numbers; the Dublin IT explanation; the press and broadcast record; other recorded talks; the 1979 and 1993 investor suits; the surviving buggy and hardware). Add the 14 Jan 1983 Patent Validation Report and the Channel 4 1995 programme to section 12. Web-check the bond-scale field (10^9-10^10 V/m). - primer.md: add the self-sustaining claim to section 1, to section 4 (a row: "loop splits water, burns the gas, feeds the electricity back" against ΔG ≥ 237 kJ/mol in and ΔH ≤ 286 kJ/mol out) and to the section 6 evidence table with the arbiter's verdict. Use Meyer's own water analysis (tap 11 ppm Na+K, about 40-60 μS/cm; distilled under 2 ppm, about 11 μS/cm) as a source-based row in steps 1, 3 and 5; the conclusion is unchanged. Drop "never transcribed" for the NZ video. Add the press/broadcast items, the investor suits and the related-work items (Kucherov, Graneau, Johnson, Shimizu) to further reading or the "Kept in the record" paragraph. Define STP in answer 5. Re-run example.py if new rows are added. - physicist/mathematician/contrarian.md: no rewrite required under MINOR. Apply Corrections 1-3 as notes.

Recheck

Recheck, 2026-10-09. Read: claims.md, arbiter.md (sections 5, 12, 14), primer.md, facts.json, sites.md and the three review files, searched against each QE defect; the new deep/ files exist (NZ transcript, EP 0 111 573 A1, Birth pp.126-222, photos, newsletters No. 4 and No. 8, US 4,389,981, Puharich book). I did not re-render page images; checks are of what the files now say.

No verdict changed; the new claim 7 verdict follows the energy accounting in primer section 3.4. Items 11-13 do not affect any verdict.

Final: OPEN (3 issues)

Orchestrator fix (2026-10-09)

Item 11 (reviewer files): inline notes "[corrected 2026-10-09 after QE: was ...]" added in contrarian.md (New Zealand video "not transcribed"; licence #490606 date attributed to p.1 of the folder copy, in two places) and mathematician.md ("740 ml/sec" in two places; "not transcribed" in the source-gap paragraph). arbiter.md: "section 11" for the 1995 binder energy figures corrected (no section covers them by itself; they are at binder pp.59-60, cited in sections 2-3 and in physicist.md and contrarian.md) in the top-of-file note and in section 14 item 5. primer.md line 161 "arbiter.md section 11" was left untouched (it refers to the verdict list, which is correct). Item 12 (Bell chart year): I rendered Newsletter No. 10 PDF p.6 and No. 11a PDF p.17 at 200 dpi (the same chart on a page printed "Page Six"). Both print "Test Date: January 10, 1983" (C. Thomas Bell, Ph.D., R & D Laboratory, Columbus, Ohio 43204; copyright 1994, "UCC 1983"). "1963" was an OCR error. No. 10's page also prints tap 11 ppm, rain 16.1 ppm, ocean 26.1K ppm; the No. 10 deep read's OCR values (tap 19, rain 15.1, ocean 28.1, 43206) and the No. 11a read's (rain 36.17, ocean 2.1K) are OCR errors, noted in the No. 10 deep read; the No. 11a deep read's other figures on that line were not individually re-edited. Fixed in both deep reads, claims.md (item 13) and arbiter.md (section 14, two places). Item 13 (no Step 1a digests): left to the orchestrator as a process item.

Item 13 (no Step 1a digests) is left to the orchestrator as a process item.

Final: PASS