Meyl-Konstantin: Quality engineering (STEP 4b)

QE agent (Opus), 2026-10-08. Inputs read: _text/manifest.json, all 56 digests in digest/ (read in full), the 19 files in deep/ (headings and the parts that bear on the checks below), facts.json, claims.md, sites.md, research/ (notes list and leads), physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md, example.py (run). Spot checks were made against the extracted text in _text/. Tags: [read] = project text or review files; [calc] = my own arithmetic (.venv Python, 2026-10-08); [memory] = recalled, with confidence.

Verdict: MINOR

The central verdicts on claims 1-6 hold up. I spot-checked five load-bearing numbers against the source text and all five match. The review is careful, well tagged and honest about its source gap. The defects can be fixed by updating facts, the arbiter and the primer. Two kinds of defect remain: 1. A large part of the material is still digest-only. This covers the textbook's device, own-experiment and eclipse chapters (PDF pp.183-241 and 295-455; the earlier deep reads of these pages were lost) and Neutrino Power pp.1-60 and 133-302. These pages hold pertinent material that no later file uses: Meyl's own negative results, his own informal experiments, and his dated eclipse and earthquake forecasts, which can be checked. 2. Some pertinent items are missing from the arbiter and the primer. These include Meyl's 1990 capacitor-Q experiment, his failed reproduction of a magnetic converter, his kit manual's own "unplug the cable" result, the Bonn rouleaux test, and several Investigator leads. The claim 7 summary ("no measurement data or trials in the documents") is not accurate as written.

None of this is likely to change a verdict on claims 1-6. Claim 7 and the eclipse part of claim 5 need new or reworded sub-verdicts.

1. Coverage

• Manifest: 24 documents, 56 chunks, nothing skipped. Every chunk has a digest (56 of 56, checked by script) [read: manifest.json, digest/]. Five pages were OCR'd across the set. One page could not be recovered: Barkhausen notes p.9, which is garbled and belongs to a document by another author (Robert Shannon, 1996). This is acceptable.
• facts.json has 24 entries, one per document, and each how_read matches what was done [read: facts.json]. The textbook is correctly marked text_available: partial, because equations are missing from its text layer.
• No defect in coverage.

2. Shortcuts

Acceptable: - Digest-only coverage is acceptable for the Barkhausen notes (another author), the Elektrosmog article, the FGF newsletter, and the complementary-medicine lecture (Kaiserslautern 2001), all of which are peripheral. - Objektivitätstheorie 1 and 2 were deep-read. They carry the same argument as textbook ch.6, so ch.6 (PDF pp.106-146) resting on its digest is a small gap, not a defect.

Defects: 1. Textbook PDF pp.183-241 and 295-455 (ch.8.2-10 and 14-20). The deep reads were lost in a concurrent write and never redone [read: claims.md coverage note; arbiter.md source gap]. No trace of them exists in deep/. These chapters hold experimental claims with dates, people and numbers that no deep read has handled: - ch.19: the charge amplifier (1989); the Kirlian "phantom leaf"; the 1990 capacitor-Q test; the 1996 microwave-oven comparison; the 14.06.1997 detection test with an "electrosensitive" subject; the 12.10.1999 transmission line built by M. Andresen and M. Rehm; and book p.411. - ch.20: the eclipse "experiment" of 11.8.1999 and dated forecasts. - ch.14-18: third-party device claims and Meyl's own assessments of them.

The arbiter and primer say plainly that these pages are digest-level and give no verdict on the devices. That is honest, but the rule is clear: an experimental result resting on a digest alone is a defect. The only remedy is a new deep read. 2. Neutrino Power pp.1-60 and 133-302 are digest-only (only pp.61-132 were deep-read) [read: facts.json]. These pages hold the following: - the TU Clausthal "1000%" claim (p.18); - the claim that cylindrical coils drop the efficiency below 100% (p.20); - the Gallex/Kamiokande argument (pp.27-29); - the April 1990 Annalen der Physik rejection (p.214); - Meyl's own failed reproduction of a 3 kW magnetic converter (pp.239-241); - the railgun numbers (p.223); - the eclipse, supernova and pole-reversal claims (pp.280-293). 3. Claim 7 rests on digests for its two main sources: the Wiesbaden scalar-wave-medicine lecture (2001) and the Kaiserslautern lecture. Claim 7 is one of the seven central claims. The material is largely qualitative, so this is a smaller defect than 1-2. A deep read would still settle what data, if any, are offered, and which universities are named as having "measured" the kit ("Edinburgh to Vienna, Stanford, TU Peking") [read: digest Skalarwellenstrahlung p.3].

3. Missed material (from my reading of all digests, confirmed in _text/ where marked)

Each item below is pertinent, sits in the project documents, and does not appear in the arbiter or the primer (checked by grep across all review files).

1. Meyl's own manual shows the link needs the cable. In Experiment 2 the manual says: "If the cable connection is carefully unplugged, the light emitting diodes installed onto the transmitter side shine again. The LED's, installed on the receiver side, extinguish" [read: _text kit manual, PDF p.19]. The Fortschrittliche lecture adds that if the earth lead is disconnected "the oscillation stops" [read: digest p.15]. The Wiesbaden lecture says the same [read: digest p.4]. This is the most direct evidence in Meyl's own words that the wire carries the transfer. It belongs in arbiter claim 4 and primer 4.4, beside the patent's "Rückleitung erforderlich".
2. Meyl's over-unity figures disagree with each other. The 2003 book's own kit advertisement says "proof of free energy (1 to 3 times over-unity)" [read: _text textbook PDF p.649]. Other figures:

claims.md and the primer give only "500-1000%". The spread is pertinent to how the figure was produced. 3. Meyl's own negative results and scepticism. - Neutrino Power p.240 reports a magnetic converter that its inventor said gave "drei Kilowatt Ausgangsleistung im Dauerbetrieb". Meyl adds: "die Effekte ließen sich in meinem Labor bislang nicht reproduzieren ... einen Over-Unity-Effekt konnten wir noch nicht erreichen" [read: _text Neutrino Power p.240]. He offers the inventor's absence as one possible cause. - Book p.346: "I judge these measurements rather sceptical, after I already in three cases had to break [it to inventors that theirs was an energy] destroying machine and they merely had measured wrong". He asks for a closed-loop demonstration [read: _text textbook PDF p.346]. - Book p.238 / Neutrino Power p.238: his own Möbius-winding rebuilds did not work [read: digest]. - Book p.412: the 1990 capacitor-Q measurements were "silver coated Q = 48, polished off Q = 28, acid-stripped Q = 9". Repeats "drowned in the mean variation", and the colleague who had proposed joint publication withdrew [read: digest pp.410-412]. claims.md records this, but the arbiter's claim 2 says the only evidence is "one qualitative plot". That is incomplete.

These are honest negative self-results. They bear on claims 2 and 4 and on how much weight Meyl's positive reports carry. 4. Dated eclipse and earthquake claims that can now be checked (book ch.20, PDF pp.424-449; Neutrino Power pp.287-292). None of the reviews checks them. Meyl's after-the-fact claims are: - the 17.8.1999 Izmit earthquake and its aftershocks, "in a rhythm of 6 to 7 days"; - the 26.2.1998 Montserrat eruption; - a "proven temporary acceleration of the rotation of the earth" and Foucault-pendulum anomalies on 11.8.1999; - a Salt Lake City tornado.

Forecasts written in the 2003 book [read: _text textbook PDF p.448]: - "After the mentioned eclipse of the sun of 29.03.2006 there once more exists danger of earthquakes (approx. 4.4.2006) for the Island of Crete and Asia Minor"; - the 29.3.2006 eclipse is predicted to slow the Earth's rotation (p.446); - "no good" from 13.11.2012; - 21.8.2017: "The thankless role of Turkey ... this time takes over South Carolina".

These are falsifiable, like the proton-radius ratio. They can be checked against the USGS earthquake catalogue and IERS length-of-day series. My recollection is that no major Crete or Columbia, SC earthquake followed those dates, but that is [memory, medium] and not a substitute for the check.

Meyl also cites Shu-wen Zhou, 21st Century Science & Technology 12(3), 1999. That paper reports caesium-clock and spectral anomalies during eclipses (Allais-type effects). It is a third-party anomaly report that the contrarian should weigh; at present it appears only in facts.json. 5. The eclipse "lens" refutation rests on the wrong mechanism. The physicist, and the arbiter after him, refute the 27-68x focusing with the Moon's gravitational lens (focal length 1.4e16 m) [read: physicist.md claim 5; arbiter.md claim 5]. Meyl's mechanism is different: he says slow neutrinos are refracted toward denser matter "like a convex lens" [read: digest pp.424-426]. The right check is the neutrino refractive index in rock from coherent forward scattering: n - 1 = sqrt(2) G_F N_e / E, which is about 1.3e-19 at 1 MeV and 4.9e-19 at 0.26 MeV in lunar rock (N_e about 1.0e24 per cm^3) [calc; G_F and the formula memory, high]. That still rules out any focusing, but the calculation should answer Meyl's model, not a different one. Meyl's "slow neutrino" premise is itself untestable as stated, and the record should say so. 6. How the claim was received in 1990. Meyl submitted "Potentialwirbel in der Elektrodynamik" to Annalen der Physik in April 1990 and it was rejected. The referee's reported argument was that Maxwell's theory covers all phenomena and that cases it does not cover only need better approximation methods [read: _text Neutrino Power p.214]. This is single-source (Meyl's own account in an interview book). It belongs in the contrarian's "how was it rejected" analysis and the arbiter's absence-of-evidence section, stated as Meyl's account. 7. Meyl's informal biological test (book p.418, 14.06.1997). An "electrosensitive" woman was asked whether the grounding was clamped or not, every 5 s. She could see neither the audience nor the oscilloscope. "After a training round her proportion of hits was lying at 100 percent!" A second subject with a pendulum also answered correctly. Meyl adds: "I already have repeated this experiment several times with different success" [read: _text textbook PDF p.418]. The experimenter was in the room and knew the condition, so the test was not blinded on his side. The arbiter's claim 7 ("no measurement data or trials in the documents") and primer 4.7 need to say instead that one informal, unblinded on/off test with mixed later repeats is reported, single-source, by Meyl. 8. Rouleaux (blood "money-roll") test. The FGF newsletter carries a study by Gollnick and Conrad (Physiologisches Institut, Bonn). It found rouleaux in many samples whatever the test conditions, and no effect of a phone call, and it judged the public demonstration method non-standardised [read: digest Newsletter pp.5-7]. Meyl cites the "blood roll phenomenon" with mobile-phone fields [read: digest About Vortex p.6-7]. facts.json has this, but neither the arbiter nor the primer uses it under claim 7. It should be stated as a test by an industry-funded association (FGF members include network operators and manufacturers) [read: digest Newsletter pp.13-14], that is, an interested party. 9. The Kelvin-Tesla 1897 story. Every Meyl document uses it as historical support ("Both are right, Tesla as well as Hertz!"). Its only source is Tesla's own 1919 Electrical Experimenter account [read: digests Elektrosmog p.2-3, textbook p.202, p.460]. The PIERS 2013 and 2012 papers date the visit 1887, before the 1893 Helmholtz meeting it follows. The arbiter lists the 1919 article as "not obtained" but does not weigh the story. It is web-checkable, and it is Tesla's own characterisation, so he is an interested party. 10. Meyl's two readings of Tesla's patent. The textbook (p.214) gives a second reading of US 787,412: 0.08484 s, "about 6 Hz", and 1.6 c using a polar radius of 6363 km [read: digest textbook p.214]. Elsewhere he uses 12 Hz against 7.8 Hz. The two readings are inconsistent. Arbiter correction 1 should mention both. 11. Further numerical inputs that no review checks: - Kamiokande "3 billion" against Gallex "66 billion" neutrinos per cm^2 s is read as "4% positrons, 96% electrons" [read: digest Erdwachstum p.4; JGG p.2]. The two detectors have different energy thresholds and see different parts of the solar spectrum [memory, high], so the ratio is not a charge split. - Aristarchus's year of 365.25062 days gives "0.22 s/a" [read: digest JGG p.3-4]. - The free-neutron lifetime is given as 918 s (book p.168); the measured value is about 880 s [memory, high].

These belong in facts.json and claim 5 as further misread inputs. 12. Stakes not recorded: - the SWG-A "scalar wave device" at 3950 EUR, sold for placing substances with "biological information" in the coil field [read: digest kit manual p.34]. This is a commercial stake in claim 7. - Meyl says telecom firms bought his books but declined funding, and that his work runs on donations and volunteers [read: digest Teslastrahlung p.15]. Neutrino Power p.295 says his staff are paid from a private donation fund. - Meyl is president of a "Society for the Advantage of Physics" [read: digest textbook p.640]. - He sits on the egm "Wirbelwandler" advisory board [read: sites.md]. 13. Unnamed confirmations stated more specifically than the arbiter records: - "over 50 universities worldwide (Edinburgh to Vienna, Stanford, TU Peking)" [read: digest Skalarwellenstrahlung p.3]; - "an EMC laboratory checked and confirmed this [the cage test] ... a conformity declaration and CE mark are available" [read: digest Teslastrahlung p.14]; - "hundreds of measurements at more serious institutes"; "one German university bought three kits" [read: deep/Kreuzfeuer]; - press items Meyl lists: Spiegel "Schwarzwälder Kopernikus" (21.5.2001); Südkurier "Gebäude der Physik wankt" (23.4.2003); J. Gruber, "Großer experimenteller Erfolg von Prof. Meyl" (DVR 4.4.2003) [read: digest textbook p.653].

These are leads for the Investigator, especially the named universities and the 2003 "report".

4. Accuracy spot checks (against _text/)

There is one digest error, with no effect downstream. The JSE digest (p.7) says "Pulling the receiver earthing: receiver LED still lights". The source says the opposite: "If I take down the receiver by pulling out the earthing, then the lighting up of the LED signals the mentioned effect back on the transmitter" [read: _text JSE p.7]. The LED that lights is the transmitter's. facts.json does not repeat the error.

example.py runs without error (exit 0), and its numbers match the primer [calc: run 2026-10-08].

5. Retention check

The question here is whether pertinent material from the sources, or from the Investigator, was left out of the arbiter or primer because a reviewer judged it negatively.

Retained well: - Gao Peng's torsion-balance test. - Runckel's opinion. - The proton-radius prediction, credited as the programme's only falsifiable number. - Hilgenberg's expanding Earth as a respectable minority position. - Weaknesses of the IGF report (eye-matched LED brightness, the wavelength inconsistency, no cage data, no wire-removed control). - Bruhn and the Furtwangen dean, marked as interested parties. - The retractions, with their duplicate-publication cause stated. - Moray kept as a lead. - Absence-of-evidence analysis with competing explanations.

These are good practice.

Defects (all MINOR): 1. Claim 2: Meyl's 1990 capacitor-Q experiment (Q 48 / 28 / 9, later lost in the scatter) is in claims.md but missing from the arbiter's claim 2 and from primer 4.2. It is pertinent whichever way it is read. 2. Claim 4: the manual's own "unplug the cable, receiver goes out" result is missing from the arbiter and primer (missed item 1). 3. Claims 4 and 5, devices: Meyl's failed reproduction of the 3 kW converter and his three "mis-measured" inventors (missed item 3) are absent. The arbiter gives no device verdicts, which is right, but these are Meyl's own results, not device claims, and they belong under "Pertinent material kept". 4. Claim 7: the Bonn rouleaux test and Meyl's 1997 on/off test are missing (missed items 7 and 8). The statement "no measurement data or trials in the documents" (arbiter claim 7 agreement; primer 4.7) understates what the documents report. 5. Investigator leads in sites.md that the arbiter and primer drop (grep shows they appear only in physicist.md or contrarian.md, if anywhere): - DE 10 2014 109 660 A1 (2014, withdrawn): information transfer with organic fluids. This is Meyl's own patent-form of claim 7. - DE 10 2021 127 321 A1 (late work). - The 2005 Hamburg industry-hosted lecture and the 2008 Südwest-Messe show (interest, not tests). - The Bearden MEG / Evans AIAS circle cited in support at DGEIM 2003. - egm "Wirbelwandler" (Meyl on the advisory board; a stake). - Kühlke's critique (Furtwangen physicist). - H. J. Herrmann's statement that his Stuttgart institute never worked on scalar waves. - The plant ATP and plant-growth papers (2015-2016) and the PIERS Prague 2015 proton paper.

No sign that any was dropped because it was judged negatively; they look dropped for length. Each should get at least one line in the arbiter's "Pertinent material kept" and in primer section 9 or 7.

6. Process

• Source tags are present and used honestly throughout. Counts of [read] / [web] / [memory] tags:

The arbiter retags [memory] items to [web] where it checked them, and it lists what remains [memory]. - Web checks: physicist 6, arbiter 7 of 8; the mathematician and contrarian have [web] items with URLs. The verdict-bearing medium-confidence items were checked: US 787,412, geodesy, PRad/CODATA, SN 1987A, Hafele-Keating, spin ice, day-night. Acceptable. - Verdict scale: - Claim 6b reads "Refuted by calculation / not controlled tests". "Not controlled tests" is not a scale verdict. Give the anecdotes (Moon, Hubble, Tamarack) their own scale verdict, for example "Tested, not supported" for Tamarack, or "Not testable as stated" if no record can be obtained. - Claim 3b, "Refuted by calculation within Maxwell's equations (Gauss's law)", uses the theory under challenge to refute the challenge. The arbiter already notes that the experimental statement is narrower. The arbiter should consider whether "Internally inconsistent" (the derivation) plus "Tested, not supported" (kit scale) says it better, and say why either way. - Primer against STEP 4: all nine sections are present, in order, with answers at the end, and symbols are defined. It is written well for the reader profile, and the worked example (Earth growth) is excellent. Small errors: - Section 2 cites "Kleppner and Kleppner"; this should be Kleppner and Kolenkow, An Introduction to Mechanics [memory, high]. - Section 5, step 2: "the n = 5 case would give 66 m/yr" is the circumference figure (the radius is 10.5 m/yr) [calc]. Say "66 m/yr of circumference", as example.py does. - example.py runs, and its comments explain each step. Pass.

Corrections to earlier documents

1. digest/Meyl-Scalar-Waves-Theory-And-Experiments-2001-f744b9.01.md, line "p.7: ... Pulling the receiver earthing: receiver LED still lights". This is wrong. The source says pulling the receiver earthing takes the receiver down and the transmitter LED lights up [read: _text JSE p.7].
2. arbiter.md claim 7, "no measurement data or trials in the documents", and primer.md 4.7. These are incomplete. The textbook reports an informal on/off detection test (14.06.1997; 100% after a training round; later repeats "with different success") [read: _text textbook PDF p.418], and the FGF newsletter reports a Bonn laboratory test of the rouleaux claim with a negative result [read: digest Newsletter pp.5-7].
3. arbiter.md claim 2, "one qualitative plot from a thesis". Add the 1990 Q measurements (48 / 28 / 9) and their failed repeats [read: digest textbook pp.410-412].
4. arbiter.md claim 5 and physicist.md claim 5, eclipse lens. The gravitational-lens calculation answers a mechanism Meyl does not propose. Replace it with the refractive estimate n - 1 of about 1e-19 (section 3, item 5) [calc], and add the dated 2006 and 2017 forecasts as testable items.
5. claims.md claim 4, "over-unity of 500-1000%". Add the range across Meyl's own texts: 1-3x (2003 kit advertisement), about 10x, up to 5x, 4.9x (gross 21%).
6. arbiter.md correction 1 (Tesla's 12 Hz). Add that the textbook (p.214) gives a second, different reading: 0.08484 s, about 6 Hz, 1.6 c with a polar radius of 6363 km.
7. primer.md section 2: "Kleppner and Kleppner" should read "Kleppner and Kolenkow".

Remediation

Deep reads (most important first; at most 8)

1. Textbook PDF pp.401-455 (ch.19-20), Opus. List every experiment Meyl reports, with date, people, instruments, numbers and controls. List every eclipse or earthquake claim and every dated forecast with its page. Say which can be checked against public records (USGS, IERS length of day) and what each record would have to show.
2. Textbook PDF pp.295-400 (ch.14-18), Sonnet. Record each third-party device claim (railgun, Testatika, RQM, Bedini, Ecklin, Seike, Coler, Patterson, Meyer, Searl, Podkletnov) with the numbers and sources Meyl cites. Record separately every statement of Meyl's own assessment, failed rebuild or scepticism.
3. Neutrino Power pp.223-302, Sonnet. Railgun figures; the converter visit and Meyl's failed lab reproduction; the eclipse, supernova and pole-reversal claims; and every number given.
4. Textbook PDF pp.183-241 (ch.8.2-10), Sonnet. The historical claims about Tesla (Kelvin 1897, Colorado Springs 200 lamps at 42 km, 0.08484 s and 6 Hz and 1.6 c, the single-wire patent) and their cited sources, plus the "oscillating interaction" argument for galaxy rotation.
5. Neutrino Power pp.1-60, Sonnet. The TU Clausthal claim and the flat- versus cylindrical-coil efficiency claim; the Gallex and Kamiokande figures; the Tesla, Kelvin and Morgan history as told.
6. Skalarwellenstrahlung (Wiesbaden 2001) pp.1-14, Sonnet. The claim 7 mechanism (nerve, tinnitus, cancer); any data offered; which universities are named as having measured the kit; efficiency figures.
7. Textbook PDF pp.514-545 (ch.24-26), Sonnet. Vibration medicine, nerve conduction and the antenna-loss argument: what is claimed, what evidence is cited, and Meyl's own caveats.
8. Newsletter pp.4-7 (rouleaux article), Sonnet. The Bonn method, samples and results, and the newsletter's funding and membership (stake).

Changes to later files (MINOR route)

• facts.json:
• Textbook: update how_read after the new deep reads.
• Neutrino Power: add the failed converter reproduction, the Annalen 1990 rejection and the Clausthal p.18 wording.
• Kit manual: add Exp. 2 "unplug cable", the SWG-A price and purpose, and the 2003 "1 to 3 times" advertisement (textbook entry).
• JGG and Erdwachstum: add the Kamiokande/Gallex 4%/96% and Aristarchus 0.22 s/a inputs.
• claims.md: claim 4 over-unity range (correction 5); claim 5, add the eclipse focusing mechanism and the dated forecasts; claim 7, add the 1997 test and the rouleaux test.
• arbiter.md, new "QE revision" section:
• (a) claim 2: add the 1990 Q test;
• (b) claim 4: add the manual's own unplug result and Meyl's failed reproductions under "pertinent material";
• (c) claim 5: redo the eclipse-lens argument with the refractive estimate; check the 2006, 2012 and 2017 forecasts and the 1999 "rotation acceleration" against USGS and IERS on the web and give them a scale verdict (expected "Tested, not supported" if the records show nothing, but the check decides);
• (d) claim 6b: give it a proper scale verdict;
• (e) claim 3b: reconsider the label;
• (f) claim 7: reword the "no data" sentence, add the 1997 test and the Bonn rouleaux test (with the FGF stake), and check whether any split verdict changes;
• (g) add the dropped Investigator leads (section 5, item 5), the Annalen 1990 rejection (Meyl's account), the Kelvin-Tesla 1897 story with Tesla as its interested source, Zhou's eclipse-clock report as a third-party anomaly, and the stakes (SWG-A, donation funding, egm board, society presidency).
• primer.md:
• 4.2: add the 1990 Q test.
• 4.4: add the unplug result and the spread of over-unity figures.
• 4.5: add the eclipse claims with the refractive calculation as a short worked number, and the dated forecasts as a "check it yourself" item.
• 4.7: add the 1997 test (unblinded, single-source) and the rouleaux test.
• Fix "Kleppner and Kolenkow" and the 66 m circumference wording.
• Section 9: add Meyl's 2014 patent DE 10 2014 109 660 A1 and the Bonn rouleaux paper.
• sites.md (next Investigator run):
• Tesla, Electrical Experimenter, Feb. 1919 (Kelvin visit);
• the 2003 report behind "Gebäude der Physik wankt" (Südkurier 23.4.2003) and Gruber's DVR note;
• the named universities (Edinburgh, Stanford, TU Peking) and the "EMC laboratory" with its CE declaration;
• USGS and IERS records for the forecast dates;
• Shu-wen Zhou 1999.

Output file: review/Meyl-Konstantin/qe.md

Recheck

QE recheck (STEP 4b), 2026-10-08. Checked against the current arbiter.md, primer.md, claims.md, facts.json, deep/ and digest/ by grep and spot reading. [read] throughout.

Section 2, shortcuts

1. Textbook pp.183-241 and 295-455: FIXED. A deep read now covers them (deep/ textbook file, 418 KB; facts.json lists pp.183-241, 295-400, 401-455, 514-545). The arbiter has an "QE revision" section built on it.
2. Neutrino Power pp.1-60 and 223-302: FIXED (deep parts 1 and 2 in deep/Meyl-Neutrino-Power-2000-5d7650.md). Pp.133-222 stay digest-only, as facts.json states. REMAINING (disclosed gap, includes the p.214 Annalen rejection, which I read directly in _text).
3. Claim 7 sources (Wiesbaden, FGF newsletter): FIXED (deep reads exist; arbiter claim 7 cites them).

Section 3, missed material

1. Manual "unplug cable" result: FIXED (arbiter claim 4; primer).
2. Spread of over-unity figures: FIXED (arbiter claim 4 and QE table, section 4).
3. Meyl's negative results (3 kW converter, three mis-measured inventors, Q 48/28/9): FIXED.
4. Dated eclipse and earthquake claims: FIXED (checked against USGS/IERS in arbiter QE section 3; Zhou added).
5. Eclipse lens mechanism: FIXED (refractive estimate replaces gravitational lens).
6. Annalen 1990 rejection: FIXED (stated as Meyl's account).
7. 1997 on/off session: FIXED (arbiter claim 7; primer 4.7 and table row 7d).
8. Bonn rouleaux test: FIXED (with FGF stake).
9. Kelvin-Tesla story: FIXED (Tesla 1919 article read on the web, February 1919).
10. Second reading of Tesla's patent (0.08484 s, about 6 Hz, 6363 km): FIXED.
11. Kamiokande/Gallex, Aristarchus, neutron lifetime inputs: FIXED (in facts.json and primer).
12. Stakes (SWG-A, donations, egm, society): FIXED.
13. Unnamed confirmations: FIXED.

Section 5, retention

1-4. Claims 2, 4, 7 and device results: FIXED. 5. Dropped Investigator leads (DE 10 2014 109 660 A1 etc.): FIXED (restored in arbiter).

Section 6, process

• Claim 6b and 3b verdict labels: FIXED (6b anecdotes get individual scale verdicts; 3b relabelled Tested, not supported).
• Primer "Kleppner and Kolenkow": FIXED. "66 m/yr of circumference": FIXED.

Corrections to earlier documents

1. digest/Meyl-Scalar-Waves-Theory-And-Experiments-2001-f744b9.01.md line 31 still says "Pulling the receiver earthing: receiver LED still lights". REMAINING. The arbiter acknowledges it and facts.json does not repeat it, but the source digest itself is unchanged and still wrong. 2-7. FIXED (arbiter claim 7 and 2, claim 5, correction 1 addendum, claims.md, primer 2).

Minor note, not counted: arbiter.md line 7 still carries the old source-gap sentence, but line 9 explicitly supersedes it.

Remaining defects: (1) digest error in f744b9.01.md line 31; (2) Neutrino Power pp.133-222 digest-only.

Final: OPEN (2 issues)

Orchestrator fix (2026-10-08)

• digest/Meyl-Scalar-Waves-Theory-And-Experiments-2001-f744b9.01.md:31: "receiver LED still lights" -> "the transmitter LED lights" (with note quoting JSE p.7).

Orchestrator fix (2026-10-08), deep read

Deep reader then arbiter (orchestrator fix) for the second remaining defect of the Recheck. [read] throughout unless marked.

What was done: - Neutrino Power pp.133-222 deep-read in full from _text/Meyl-Neutrino-Power-2000-5d7650.txt. The text layer is clean prose. Page images were rendered from the original PDF (manifest) for the figure and formula pages p.142, 155, 163, 171, 186 and 191. The result is appended to deep/Meyl-Neutrino-Power-2000-5d7650.md as "Part 3, pp.133-222 (2026-10-08)", tagged [read: p.N]. It covers experiments (there are none in these pages), the particle model, the dual field approach with the book's own formula misprints, the ether and c, Meyl's statement that the length law cannot be tested in a laboratory, the Apollo "36", gravitation, the objectivity theory, and the Annalen account with its context. It ends with registers of people, dates, places and numbers. - arbiter.md: new section "QE revision 2 (orchestrator fix, 2026-10-08)", claim by claim. No existing verdict changed. New sub-items: - Claim 5: neutrino-induced beta decay and the unipolar-generator "rumours", both Not testable as stated. - Claim 6, Not testable as stated: the length law undetectable from inside, in Meyl's own words; no strong interaction; orbits, day length and light bending as field effects. - Claim 6, Tested, not supported: proton moment "opposite its spin", neutron = proton + electron, anti-suns as common as suns, electrons larger than nucleons. - In-place corrections: the claim 6 Moon line and correction 14; the summary table row 6; the claim 6 verdict list; section 8's Annalen bullet; the source-gap note; the Annalen record added to the absence-of-evidence records. - facts.json (Neutrino Power entry only, checked by diff): how_read now reads deep read in full (pp.1-302); the summary has a pp.133-222 paragraph; the Annalen wording is refined; six key claims added; sources updated. The JSON loads. - claims.md: coverage note (Neutrino Power deep-read in full), claim 6 additions (claims only, no verdicts), Moon-36 correction, Annalen wording. - primer.md: source gap; 4.6 (Moon 36 in Neutrino Power, star-brightness check with the magnitude scale defined, Meyl's "cannot be measured from inside" quote); evidence-table rows 6g', 6h, 6i; the Annalen sentence in section 7. - physicist.md: two stale statements that "36 appears only in Objektivitaetstheorie 2" corrected, with a pointer to arbiter QE revision 2 (correction 32).

Corrections to earlier documents, stated plainly: 1. The arbiter, physicist and claims.md said the Moon factor 36 occurs only in Objektivitaetstheorie 2. That was wrong. Neutrino Power p.191 gives "auf das 36-fache" [read: page image p.191], and the Objektivitaetstheorie 1 deep read records 36 too. Only the 2003 book uses 2.457. Fixed in all four files. 2. The Annalen passage was summarised as "the referee arguing that Maxwell covers all phenomena". At source, the "covered by Maxwell" ground is in the passive voice. The sentence attributed to "ein Gutachter" is the one about approximation methods. The journal is named by the interviewer, and Meyl does not correct it [read: p.213-214]. The substance is unchanged; the wording is fixed in facts.json, claims.md, primer.md and arbiter section 8.

QE check of these additions: - Accuracy: 46 key German phrases quoted in the deep read were checked by script against the extracted text (whitespace and hyphenation normalised). All are present; one spans the p.217/218 page break. The formulas quoted (p.163, 171, 186, 191) were read on the page images. Arithmetic re-run [calc]: sqrt 6 = 2.449; 2.5 log10(1296) = 7.78 and 2.5 log10(6) = 1.95 magnitudes; Sirius +6.32 and +0.49; equinox day excess 6.7 min at the equator; 1e-3 cm / 1.41e-13 cm = 7.1e9. - Source honesty: - Every new fact is tagged. The new Tested, not supported items rest on [memory, high] measured values (proton moment sign, neutron spin, antimatter-domain limits, electron size). They were not web-checked, and arbiter QE revision 2 says so; the web-check rule covers medium and low confidence items. The one [memory, medium] item (the Annalen's 1990 publisher) bears on no verdict. - The Annalen account is described, not labelled: it is documented in one published source (the inventor's interview book, 2000), the record that would settle it is named, and both readings of the referee's ground are kept. - Interested parties are marked: the interviewer and publisher co-promote the book. - Retention: the pertinent material from these pages is kept next to the verdicts, including what cuts against them and what cuts for Meyl. Meyl's own concessions are kept (the length law is unmeasurable from inside; vacuum energy is "not in sight"; making energy from waves would cost more than it gives). So are the points that are fair to him: refractive light bending is a legitimate equivalent picture; source-independent light speed agrees with experiment; the "almost three times" proton moment is right. - Scale: every new verdict uses the STEP 2 scale. The electron-size item carries its alternative ("Not testable as stated if a field extent is meant"). - Reader profile: the primer additions define the magnitude scale. They use no tensors or field theory and give a worked number.

Remaining defects from the Recheck: 1. digest f744b9.01.md line 31: CLEARED (fixed by the concurrent orchestrator agent; see the entry above). 2. Neutrino Power pp.133-222 digest-only: CLEARED (deep-read in full; folded into arbiter, facts.json, claims.md and primer).

Not counted as defects: the digest of Neutrino Power chunk .05 has small page drifts and marks two formulas "[OCR garbled]" that read cleanly as images. Digest slips are expected and were caught here, and nothing downstream repeats them. aaa-progress.md (line 109) still records Meyl-Konstantin as OPEN; it is the orchestrator's status log, to be updated when the site is next rendered.

Final: PASS