Dunning-Davies and Norman: Quality Engineering report (Step 4b)

Reviewer: QE (Opus). Date: 2026-10-03. Inputs read: _text/manifest.json, all 8 digests in digest/, facts.json, claims.md, physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md, example.py. Spot checks made against _text/*.txt. There is no deep/ folder: Steps 1-4 ran before digests and full-text extraction existed.

Verdict: MINOR. The central mathematical verdicts (claims 1 and 2) rest on a full reading of the 2020 preprint, and I confirmed them against the extracted text. example.py runs and reproduces them. Other verdicts are unlikely to change. However, the review was built on partial reading of three documents that now have full extracted text: the 2024 Hadronic Journal paper (OCR of all 29 pages), the Gravity v.2 appendices (pp. 28-69), and the end of the 2018 aether paper. Several statements in facts.json, arbiter.md and primer.md are now stale or wrong. One quoted-number finding (the Sweet-Bearden "inconsistency") comes from a misreading, and the arbiter's statement that nothing in the folder reports a measurement made by the authors needs qualifying. All of this can be fixed by updating facts.json, claims.md, arbiter.md and primer.md.

---

1. Coverage

Every manifest document was digested and none was skipped. No digest is missing.

2. Shortcuts

1. Gravity v.2 appendices (pp. 28-69) were read by heading only [facts.json; arbiter "Source gap"]. These pages carry the evidence for claim 6 (Santilli's hadronic fuels and waste neutralisation, with specific numbers: 1.294 MeV resonant photons, Zn(70,30) and Mo(100,42) fuels [read: Gravity text lines 1551, 1649, 1663]) and for claim 5e (neutroid and neutron synthesis, including the authors' own co-authored experimental paper, see section 3). They also hold the Carstoiu/Dunning-Davies gravito-Maxwell note (pp. 60-69), which the arbiter uses for the "Internally inconsistent across documents" point on claim 4a. Its only basis is the mathematician's [read], and the mathematician's own Source gap paragraph says the appendices were "only skimmed". A central claim (claim 6) rests on [memory] while its primary text sits unread in the folder. Defect.
2. 2024 Hadronic Journal paper: about a third read at 70 dpi. It is the latest and most complete statement of the project. Its OCR text is now available. The arbiter's open question 2 ("What does the 2024 paper actually show beyond facts.json?") can now be answered. Defect.
3. Brown patent (Gravity pp. 51-59): OCR is badly garbled [digest .02, .03]. It is peripheral (historical, and covered by the Brown verdict), so digest-only coverage is acceptable. The "million to one" efficiency claim is stated legibly in the main text [read: Gravity lines 603, 611] and reaches facts.json.
4. Enigma of Light, CIA memo, 2020 preprint: fully read. Acceptable.

3. Missed material (found in the digests and checked in the text)

Important (affects a claim, a verdict, or a statement in the arbiter or primer):

• The authors co-signed an experimental paper. Gravity v.2 cites "Experimental Confirmation of the Synthesis of Neutrons and Neutroids from a Hydrogen Gas", American Journal of Modern Physics 6(4-1), 85-104 (2017), by Norman, Bhalekar, Beghella Bartoli, Buckley, Dunning-Davies, Rak and Santilli. It lists links to detector readings (Ludlum model 375, Berkeley Nucleonics SAM 940, Polimaster PM170) on thunder-energies.com [read: Gravity text lines 1486-1503]. The arbiter's statement "Nothing in the folder reports a measurement made by the authors" is true of the folder's own documents. It should still say that the authors put their names to an experimental claim cited from the folder, and it should name the conflict of interest: Norman is "Scientific Advisor, Thunder Energies Corporation" [read: Gravity p.2; Final consequences p.1], and the detector links are on that company's site. This bears on claim 5e ("Never properly tested") and on the contrarian's "tested and failed vs never tested vs dismissed" split.
• Neutroid hypothesis and earlier experiments. The Poynting-vortex appendix cites delayed neutron detections from Sternglass, Borghi, Santilli and Trounson as evidence of a spin-zero "neutroid" intermediate (A=1, Z=0, J=0) [read: Gravity lines 1215-1220; digest .02 p.34-36]. No review file mentions the neutroid, Borghi or Sternglass. These are named experiments that a claim-5e verdict should address. Did anyone independently replicate them? This needs a [web] check.
• "No neutrino has ever been directly observed" [read: Gravity p.34]. It is an explicit, testable statement. Claim 5d covers it only by [memory: high, not web-checked]. Give it a web citation (Cowan-Reines 1956).
• B-2 "perpetual propulsion". Gravity v.2 pp.15-17 says the B-2 bomber can fly "indefinitely ... with essentially zero fuel consumption", powered by a self-generated gravitational gradient, quoting the Valone-edited Electrogravitics Systems [15] [read: Gravity lines 520-568]. It is a concrete, checkable claim about a real aircraft, and it is absent from every file. It belongs under claim 4 (over-unity) and claim 7 (suppression).
• Hadronic-fuel numbers. Waste neutralisation by 1.294 MeV photons; fuels Zn-70 and Mo-100 with the photon "not lost" [read: Gravity pp.44-48; 2024 paper pp.22-25]. These are specific enough for an energy-budget check against tabulated nuclear masses (both nuclides are double-beta-decay candidates). That check could move part of claim 6 from "Never properly tested" to "Refuted by calculation" or confirm the current verdict. No reviewer saw the numbers.
• Carstoiu note's own statements (pp.60-69): its gravitational fields propagate at c; the gravitational analogue of the Alfvén speed is imaginary; Dunning-Davies proposes D/Dt to make the equations Galilean-invariant [digest .03]. This supports the arbiter's 4a cross-document inconsistency, but it should be read first-hand, not taken second-hand.
• 2024 paper content not in facts.json [digest .01]: Jack's T field linked to heat via Bridgman and the Ettingshausen effect (p.6-7); "time itself is heat" (p.16); graviton as transient photon-antiphoton pairs (p.11); the neutron needing "a neutrino ... 0.782 MeV more positive energy" (p.12, confirmed [read: text line 449]), which settles the arbiter's "[from facts.json only]" tension; a seven-point list of proposed advances (p.12); Einstein's 1920 Leiden address on the ether (p.9).

Worth adding (teaching value, no verdict change):

• Einstein's 1920 Leiden "ether" address [read: 2024 paper p.9]. The authors use it to support a material aether. The primer should explain that Einstein's GR "ether" has no state of motion or rest frame, so it is not Thornhill's gas. This is a good lesson for the reader.
• Van Flandern's eclipse claim (maximum eclipse "about 40 seconds before" gravitational alignment) [read: Gravity lines 85-86]. Not mentioned anywhere. It is a nice worked example of light-time and aberration.
• Twin paradox, Sagnac effect, "absolute time" as interaction frequency in the aether (Gravity pp.24-25). Not mentioned anywhere.
• Brillouin's ε = −1/G = −1.5 × 10⁷ (cgs) [read: Gravity p.11]. The arithmetic is correct (1/6.674e-8 = 1.498e7). A units example for the primer.
• Wesley's electrodynamic gravity (Gravity pp.68-69); Santilli's nitrogen synthesis "explaining thunder" and Earth's core heat (Gravity p.50; 2024 p.27); Tesla's 1932 longitudinal-pulse quote.

4. Accuracy spot checks

Minor factual slips: facts.json says Gravity v.2 has "about 113k characters" (manifest: 143,724). Thornhill's papers are dated 1983 in the 2024 paper's reference list and 1984 elsewhere [read: 2024 text lines 44, 734, 1125]. facts.json should note both.

5. Process

• Source tags: present and generally honest throughout. The arbiter and primer clearly separate [read], [web], [memory] and [calc], and they say plainly when a conclusion rests on [memory]. Good.
• Stale source-gap notes: claims.md, arbiter.md and primer.md all say the 2024 paper has "no text layer". OCR text now exists. facts.json text_available: none and how_read: page images must be updated once it is read.
• Web checks: the arbiter made 7 calls and lists the failed checks honestly (search quota). Medium-confidence [memory] facts that a verdict depends on and that remain unchecked: lensing achromaticity (verdict 4b), the NATO AGARD 1990 report (the papers say it concluded a Laithwaite-type device "could counteract gravity"; this bears on 4c), Santilli/neutroid replication (5e, 6), and Michelson-Morley successors and neutrino detection (3, 5d; these are high confidence, so less urgent). Defect (minor): redo them.
• Verdict scale: applied consistently, and split verdicts are stated. One off-scale label ("Notation-only form: true and uncontested") is explained in the text. Acceptable.
• Primer: follows all nine STEP 4 sections in order. It defines symbols, introduces quaternions from scratch, explains gauge freedom, "elliptic" and "d'Alembertian", and gives numbers in the worked example, with answers at the end. It fits the reader profile. Its section 5 / §7 Sweet-Bearden lesson and Check-yourself Q5 depend on the misreading in spot check 3.
• example.py: runs cleanly with .venv/bin/python and reproduces every number quoted (c√3 = 5.193e8 m/s, c/√3 = 1.731e8 m/s, ratio −1/3, elliptic ω = ±ick, gain 5.714e6, 197-235 MeV, 0.782 MeV). Part E's "quoted range" comment carries the same misreading.

6. Verdict

MINOR. Claims 1-3 and 4a are well founded on fully read text and correct mathematics. Remediation affects claim 5e and claim 6 (possible split verdict after an energy-budget check), the reasoning behind 4e, the stale source-gap notes, and missing material. It does not affect any central verdict on the T field, scalar waves or superluminal gravity.

7. Remediation actions

Deep reads (write deep/{slug}.md, tagged [read: p.N]):

1. Gravityv-2-283abf, pp. 28-51 (Opus). Poynting-vortex and hadronic-energy appendices. Extract every experimental claim (neutroid; Sternglass/Borghi/Santilli/Trounson detections; the 2017 co-authored AJMP neutron-synthesis paper and its detector evidence) and every number (0.782 MeV, 1.294 MeV, Zn-70, Mo-100, energy output claims). Do an energy-budget check of the hadronic-fuel reactions against tabulated atomic masses ([web] for mass data): is net energy positive once the photon is accounted for? Note the conflict of interest (Thunder Energies).
2. NormanDunning-Daviescopy-70e992, pp. 1-29 (Sonnet). Read the full OCR text (render page images where equations are garbled) to replace the 10-page 70-dpi reading. List all claims: T field and heat (Bridgman, Ettingshausen), "time is heat", graviton as photon-antiphoton pairs, the seven proposed advances, the neutron/neutrino passage, Einstein's 1920 Leiden quote in context, waste-neutralisation and nitrogen-synthesis claims, aether particle mass, and the reference list.
3. Gravityv-2-283abf, pp. 60-69 (Sonnet). Dunning-Davies's note on the gravitational equations analogous to Maxwell's (Carstoiu): confirm first-hand that the fields propagate at c and are transverse, the imaginary gravitational Alfvén speed, the D/Dt Galilean-invariance proposal, and Wesley. Does this contradict the main text's superluminal gravity?
4. Gravityv-2-283abf, pp. 14-27 (Sonnet). Main-paper evidence section: B-2 "perpetual propulsion" quote and its source [15], Brown "million to one", Laithwaite and the NATO AGARD 1990 citation (exact wording), Schauberger/Coats, Sweet-Bearden, Twin paradox, Sagnac, absolute time. Also pp. 3-4 for Van Flandern's 40-second eclipse claim.
5. Finalconsequencesofanaetherrevised-6ecfe8, pp. 6-10 (Sonnet). The unread end: strong force as "spherically confined nuclear scale induction", Gamow window, the neutrino and quantum "logical inconsistencies" lists, aetherino, Santilli neutron. Do any equations or numbers appear?

Web checks (arbiter, about 6 calls): AGARD 1990 report content; Borghi/Santilli neutron-synthesis replication or critique; the 2017 AJMP paper's existence and venue; lensing chromaticity bound; Cowan-Reines neutrino detection; B-2 propulsion (public engineering record).

facts.json: update the 2024 paper (how_read → OCR full text + deep read; text_available → yes; add the missed claims). Correct the how_read for Final consequences (partial, then deep read) and for Gravity v.2 (front read + deep reads pp. 14-27, 28-51, 60-69). Add the 2017 co-authored experiment, the neutroid claim, the B-2 claim, the hadronic-fuel numbers and the Thunder Energies affiliation. Fix the character count and note both Thornhill dates.

claims.md: remove "no text layer" from the Source gap note. Add the B-2 / zero-fuel flight to claim 4. Add to claim 5 that the authors co-signed a 2017 experimental neutron-synthesis claim.

arbiter.md (add a "QE revision" section): - Withdraw the "Sweet-Bearden figures are internally inconsistent" finding (spot check 3). Say instead that the gain range and the 175 µW / 1 kW figures describe different experiments, and that the quote claims self-running battery-less units, which is a direct over-unity claim (strengthens 4f). - Qualify "Nothing in the folder reports a measurement made by the authors": the folder cites a 2017 experimental paper co-authored by Dunning-Davies and Norman, and Norman advises the company hosting the evidence. Reassess 5e accordingly (most likely still Never properly tested, now with the reason "claimant-run, no independent replication"). - Claim 6: add the energy-budget result from deep read 1, and give a split verdict if the arithmetic rules out part of it. - Claim 4: add a verdict for the B-2 zero-fuel claim (Refuted by calculation as perpetual motion, or Not testable as stated, with reasoning). - Close open questions 2 and 3, and update the Source gap paragraph.

primer.md: remove the Sweet-Bearden "figures disagree" lesson in §7 and rewrite Check-yourself Q5 and its answer (for example: what gain does 175 µW → 1 kW imply, and why is a "closed battery-less system with positive feedback" a direct conservation-of-energy question?). Update the Source gap paragraph and evidence-table rows 4e, 5e and 6. Optionally add in §3 or §7 a short note on Einstein's 1920 "ether" versus Thornhill's gas aether.

example.py: change the Part E comment that calls the gain "outside the quoted range" so that it does not claim an inconsistency. Otherwise unchanged.

Recheck

Reviewer: QE recheck (Sonnet). Date: 2026-10-03. Checked facts.json, claims.md, arbiter.md (including its "QE revision"), primer.md, example.py and deep/ against the defects and remediation actions above. I ran example.py with .venv/bin/python: it runs cleanly and prints the Part F fuel budget (Zn-70 0.872 MeV back per photon, Mo-100 1.804 MeV) and the corrected Sweet-Bearden gain line.

Defects (sections 1-5)

Remediation actions (section 7)

Outstanding

1. Gravity v.2 pp.14-27 not deep-read (deep read 4; Van Flandern, twin paradox, Sagnac, Laithwaite, AGARD wording rest on the digest).
2. Gravity v.2 pp.52-69 not deep-read (deep read 3; Wesley, D/Dt proposal rest on the digest).
3. Web checks for AGARD 1990, independent replication of Santilli neutron synthesis, lensing chromaticity, B-2 record and Cowan-Reines still undone (verdicts rest on [memory] for these and say so).

None of these changes a central verdict: the remaining gaps are acknowledged in claims.md, arbiter.md and facts.json, and the affected verdicts (4a, 4c, 4g, 5d, 5e) are tagged accordingly.

Final: OPEN (3 issues)

Recheck 2

Reviewer: QE recheck 2. Date: 2026-10-03. Re-read facts.json, claims.md and arbiter.md (including "Round 2"); checked deep/Gravityv-2-283abf.md pointers. Scope: the three items the first recheck marked as remaining.

1. Gravity v.2 pp.14-27 not deep-read: fixed. facts.json (how_read, "second-round deep read pp.3-5 and 11-27") and claims.md (Source gap) record a second-round deep read of pp.3-5 and 11-27 with all items [read]. The arbiter's Round 2 section uses it (B-2, Laithwaite/AGARD wording attributed to Cook, Sweet-Bearden, Van Flandern, twin paradox and Sagnac).
2. Gravity v.2 pp.52-69 not deep-read: fixed. facts.json records a second-round deep read of pp.52-69 (Brown patent GB 300,311, Carstoiu / Dunning-Davies note, D/Dt proposal, Wesley) with page images, and the arbiter corrected its earlier Carstoiu remark in Round 2.
3. Web checks (AGARD 1990, independent replication of Santilli neutron synthesis, lensing chromaticity, B-2 record, Cowan-Reines): remains. claims.md and arbiter.md (lines 49, 53, 59, 70, 108, 171, 205) still tag these [memory] and "not web-checked", and say so. No central verdict changes, but the checks are undone. The user paused remaining work on this item pending their own quality review.

Final: OPEN (1 issues)