Perreault-Bruce: Quality Engineering (Step 4b)

QE run 2026-10-04. Inputs read: _text/manifest.json, _textprep.log, both digests, both deep reads, facts.json, claims.md, sites.md, all 11 research/ notes (research/files/ is empty), physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md, example.py (run). Spot checks were made against the extracted text and against page images I rendered of US7800286.pdf pp.4-6 (scratch: /tmp/qe-perreault/).

Verdict: MINOR. Coverage is complete, both documents were deep-read in full with page-image checks, the central calculations are correct and reproduced by example.py, and the verdicts are consistent across the three reviews, the arbiter and the primer. The defects are: one factual misreading carried into the arbiter and primer (the "0.1 atm of radon" vessel), a gap in the arbiter's and primer's "what would settle it" test design (the patent's radon source is a long-lived emanating material, so the device would not fade with the 3.82-day radon half-life), and several pertinent items from the Investigator and the reviews that the arbiter or primer left out. None of these changes a headline verdict.

---

1. Coverage

No document was skipped. The textprep log matches the manifest (2 documents, 2 chunks). Coverage: no defect.

2. Shortcuts

• Both documents are short and both were deep-read in full. Equations, isotope numbers and the vacuum fraction were taken from page images, not the OCR. No defect.
• Digest errors were caught by the deep read. The US7800286 digest is wrong on several load-bearing numbers: it gives the target as Ge-74 → Se-74 (the patent prints Ge-72 → Se-75), lists Po-211/Po-209 (the patent prints Po-212/214/216/218/210), and gives the vacuum as "1/100 atmosphere" (the patent prints 1/10). The deep reader flagged and corrected all of these. facts.json and claims.md use the corrected values. I confirmed the corrected values on the page images (see §4). This shows the digest-plus-deep-read design working as intended. Note for later users: do not cite the US7800286 digest for numbers.
• The real shortcut is outside the manifest. The inventor's fuller statements found by the Investigator were read only in part or not at all: the 26-page "Radon Gas is the Ultimate Fuel" (pp.1-2 only), "HF-box-lab-tests.pdf" (possible test data), the 2002 book Harnessing Cosmic Energy, about 30 other radioionics.com PDFs, the Abbood & Ebrahiem 2020 paper (abstract only) and NASA TM-20220019348. Every review, the arbiter and the primer say this openly, and research/files/ is empty because downloads need the owner's approval. This is acceptable for a first Investigator run. It is the main open item for the next run (see §7).

3. Missed material (from the digests, deep reads and Investigator notes)

Items that facts.json, the reviews, the arbiter or the primer missed or under-weighted:

1. The patent's radon source is a long-lived emanating material, not a charge of radon gas. The patent places "radon gas emissive radioactive material 4 ... at the base inside vessel 1" and says "It is preferred that a natural alpha source be used" [read: US7800286 p.5 (PDF p.5-6), text lines 408-431]. Claim 1(c) recites radon gas that "expands and thereby fills the cavity". The source is therefore a radium-, uranium- or thorium-bearing material that keeps producing radon. Two consequences that no review states: - (a) The arbiter's and primer's proposed control test ("a real source decays with the 3.82-day radon half-life"; "fade at 3.82 d confirms the source"; primer §7 and Answer 3) holds only if the radon is sealed in with no parent. With the patent's emanating material at the base, radon activity is held at equilibrium with the parent (Ra-226 half-life about 1600 years [memory, high]), so a real radon-powered output would not fade in days. The proper control is to remove or replace the emanating material, or to compare activity-matched sources. The verdicts stand, but the test design is wrong for the device as patented. - (b) The 0.114 W/Ci ceiling still applies, but the relevant activity is that of the emanating source (or of the parent in equilibrium), not of a gas fill.
2. Thoron's half-life. The patent prefers Rn-220 because its parents (Th-228, Th-232) are abundant [read: p.5]. Rn-220 has a half-life of about 56 seconds [memory, high], so it barely leaves its source before decaying. The patent's own picture of radon gas filling the vessel and being drawn to the wire does not fit thoron well (and Rn-220's 6.29 MeV alpha is in any case below the 6.40 MeV threshold, which the reviews do note). This is a practical point no review raises.
3. The patent's Fig. 2 feedback loop. The block diagram shows lines from the high-voltage capacitors 9 and the low-voltage capacitors 11 back to the step-up converter 7 [read: deep/US7800286, Fig. 2 page image]. This is the drawing's version of "self-generating": output used to keep the priming supply running. The physicist and contrarian mention it. The arbiter and primer leave it out, though it is the clearest statement in the patent of a closed, self-powered loop and the reason the net-power test matters.
4. Helium "blocking". The patent says helium from the alphas makes the device "electrically blocked" because "helium gas is electrically non-conductive" [read: p.5]. The primer quotes this as one sign that the patent is inconsistent about self-generation. No reviewer checks the statement itself. Helium ionises readily in a discharge, and at curie-level activity the helium made is tiny: about 3.5×10^18 atoms, or about 0.1 mm³ at STP, per curie-year of Rn-222 chain alphas [calc, QE: 3.7e10 s⁻¹ × 3 alphas × 3.156e7 s]. This is a minor check. It deserves one sentence in the primer, marked as a QE calculation.
5. The leaflet's filing date and "continuation". The leaflet says "U.S. Patent Filed on May 07, 2007" and that its circuit "is a continuation of U.S. Patent No. 7,800,286" [read: leaflet pp.1-2]. The patent was filed 4 May 2007 [read: US7800286 p.1], and Google Patents shows no family members [web: https://patents.google.com/patent/US7800286B2/en, per Investigator]. So "continuation" is informal (no continuation application on record), and the date is three days off. facts.json shows both dates but nowhere says that no formal continuation exists. This is minor and belongs in facts.json and primer §9.
6. Betz drawing details. The Betz note's "Remember- 'It's the ions, not the electrons!' Moray", the "three types of oscillator tubes", starting "with a static discharge like from a spark coil", and the lead and silver bars [read: deep/Atomic-Decay §1] appear in the deep read only. The ions quotation is pertinent to claim 5: it fits the inventor's later shift to atmospheric ions (2018 plasma valve). It is not in facts.json, the arbiter or the primer.

4. Accuracy spot checks

example.py runs cleanly (exit 0, "All checks passed"). Its numbers match primer §5 line by line (0.114 W/Ci, 6.50 µg/Ci, 3.25 mA, 29× bias power, 4.22/4.82 keV, the 1001× BTU slip). Part 3 prints "100 cm3 of pure radon at 0.1 atm", which is an honest hypothetical in the script. The defect is only in how the arbiter and primer attribute it to the patent.

5. Retention

Pertinent material that the Investigator or reviewers recorded, but the arbiter or primer dropped or reduced. No item appears to have been dropped because a reviewer judged it negatively, but these are retention gaps under SOURCE HONESTY ("keep pertinent information"):

1. 1999 "star-mode" observation (teslatech tape page): Perreault and assistant Calvin Bahlmann say they "rediscover[ed] the Moray semiconductor material" and saw a star-shaped discharge between electrodes coated with it at certain frequencies. Photos are credited to International High Technology, and the account compares it with a Farnsworth fusor videotape by Richard Hull (Jan 1998) [web: https://www.teslatech.info/ttstore/conftapes/archives/1999jul/600214.htm]. This is a reported anomalous observation, a witness-type account by interested parties. It is in sites.md and the research note, but in no review, the arbiter or the primer. It belongs in the arbiter's "kept in the record" list and the primer's evidence notes under claim 5, as an interested-party account with no data.
2. Moray-family performance account relayed by Perreault: a prototype of about 42×26×22 in, under 60 lb, about 4 kW before witnesses [web: same tape page]. The contrarian keeps it. The arbiter and primer refer only to "accounts conflict". They should quote the figure with its source and stake.
3. Perreault's 2018 lineage list of "plasma field discharge" devices: Coler (British Government report), Hendershot, Hubbard, Moray, Plauson US 1,540,998, Prentice CA 253,765 (fifty 60-W lamps from 500 W input, as he reports it), Britten US 1,826,727, Chernetski, Correa, Gray US 3,890,548 / 4,595,975 / 4,661,747 [web: Wayback radioionics.com, per Investigator]. The physicist and contrarian keep it. The arbiter mentions only "2018 Plasma Valve and Battery paper", and the primer omits it. It is related work under Investigator priority 3 and should be listed in primer §9 as the inventor's own lineage.
4. 1998 priority and wording. The 1998 Exotic Research Report article is the earliest dated statement of the idea and says "alpha fission and thermoelectric effects" [web: https://www.teslatech.info/ttstore/report/articles/v2n2art/radiant.htm]. The same article claims that Moray's valve was germanium-based and that Moray's first cousin Fletcher worked at Bell Labs (the inventor's claim). The primer lists the article in §9 but gives neither the 1998 date of the idea nor the fission/fusion wording change. The arbiter keeps the article only for its Tesla-like passages.
5. The 1927 germanium dating question. The physicist raised that "very very pure germanium" in a 1927 affidavit is unusual for the date, because high-purity germanium semiconductors came in the 1940s [memory, medium], and that the quote comes from a later book edition, so its dating is uncertain. This is put as a question, not a refutation. It bears directly on the leaflet's Moray premise. The arbiter and primer drop it. It should be kept as an open question, stated both ways, and web-checked (see §6).
6. "Radon emissive material" offered in the 2002 book ("how to obtain author's radon emissive material") [web: https://www.teslatech.info/ttstore/books/480002.htm]. The contrarian keeps it. The arbiter and primer omit it. It bears on the historical record and on what the patent's material 4 was in practice (see §3 item 1).
7. Inventors Weekend and the Don Smith e-mails (2001-2005) are in the arbiter's list but not in the primer. They are peripheral, and their absence from the primer is acceptable if the arbiter keeps them.

6. Process

• Source tags: present and honest throughout. [calc] is used consistently for the reviewers' own computations. Interested-party material (the 1999 tape page, the inventor-hosted Jaynes calculation, the inventor's "millionth of a watt" description of prior art) is labelled with its stake. Absence of evidence (2014 lapse, Alpha Fusion Energy Corp, shift of mechanism, Lumen notices) is analysed with competing explanations in the arbiter and primer, as SOURCE HONESTY requires. OK.
• Web checks: the physicist used 2 of about 6 calls, the mathematician 1 (failed fetch) and the contrarian about 1-2. Medium-confidence facts that verdicts depend on were checked by the arbiter: fair-weather current, Se-75 decay, Ge(α,n) data range, the Abbood abstract and the Moray accounts. Rn-220's 6.288 MeV alpha was checked by the physicist. Left at [memory, medium]: Ge isotope abundances, the Ge work function (the arbiter rightly calls it not verdict-bearing), and the 1927 germanium anachronism (not checked because it was dropped). Under-use of the reviewer web budget is a minor process note. Verdict-bearing facts were covered by the arbiter.
• Correction carried by the arbiter: the arbiter rightly corrected the Investigator and reviewers that Abbood & Ebrahiem's abstract does not cover (α,n) and gives no conclusion on the patent. Good practice.
• Verdict scale: applied consistently. Split verdicts are stated part by part. The arbiter resolved the claim 2 wording difference ("Plausible, untested" for the low-power cell, "Never properly tested" for the comparison with prior art) and the claim 5 scope difference with reasons. OK.
• Primer vs STEP 4: all nine sections are present in order, with answers at the end. Symbols are defined (Q, u, MeV, λ, Bq/Ci, W, T_max, the Richardson terms). The worked example is numeric and reproduced by example.py. Textbook chapters are named for Krane only. Krane ch. 8 (alpha decay) and ch. 11 (nuclear reactions) are correct for the 1988 edition [memory, high]. The reader profile is respected: there are no tensors or QFT, and the Gamow factor is explained in plain words. OK, apart from the content defects in §3-5.
• example.py: runs, exit 0, numbers match the primer. OK.
• Investigator: used 6.5 of 10 minutes. Five sites were "not reached (time limit)" and correctly marked Recheck yes. Queries are recorded and negative results list their queries. No files were saved (owner approval needed). OK.

7. Remediation (one round; MINOR)

Deep reads: none required. Both manifest documents have been deep-read in full with page-image checks, and the facts needed for remediation are quoted above with line references. The material still unread is outside the manifest (Investigator leads), so it belongs to the next Investigator run, not to a deep read: - Investigator, next run (priority order): (1) full "Radon Gas is the Ultimate Fuel" pp.3-26 and "HF-box-lab-tests.pdf" (Wayback radioionics.com), looking for any measurement; (2) Abbood & Ebrahiem 2020 full text, for any (α,n) or (α,γ) cross sections near 6-9 MeV; (3) NASA TM-20220019348, for measured NTAC power; (4) USPTO Patent Center file wrapper 11/744,769 and NH QuickStart; (5) a web check of the 1927 date of the "pure germanium" affidavit and of when that wording first appears in The Sea of Energy editions. If (1) or (2) yield data, run Steps 2-4b again for claims 1-3.

facts.json - US7800286 entry: add that the radon source is "radon gas emissive radioactive material 4" at the base, with "a natural alpha source" preferred [read: p.5]. State that the 1/10 atmosphere figure is the evacuated air pressure. - Atomic-Decay entry: add that the leaflet's filing date (7 May 2007) differs from the patent (4 May 2007), and that "continuation" is informal (no family on record [web via Investigator]). Add the Betz note's "'It's the ions, not the electrons!' Moray", the three tube types and the spark-coil start.

claims.md - Claim 1: replace "above a 6.06 MeV threshold" with the corrected wording the arbiter adopted (6.06 MeV is |Q|; threshold 6.40 MeV), so that claims.md matches the arbiter's corrected claim list. - Claim 2: say the radon is supplied by an emanating radioactive material in the vessel.

arbiter.md: add a "QE revision" section that: 1. Corrects the "Vessel scale in claim 1(c)" row. The 0.1 atm is residual air pressure, not radon. The 14,000 Ci figure is a hypothetical upper scale, not the patent's design. The realistic inventory is set by the emanating material's activity. 2. Revises "What would settle it" items 2 and 3 (claims 2-3). The control is removal or replacement of the emanating material (or activity-matched sources), not a 3.82-day fade, unless the radon is sealed in with no parent. Note the 56-second thoron half-life against the patent's preference for Rn-220. 3. Adds the Fig. 2 feedback loop (capacitors 9 and 11 back to step-up converter 7) to claim 3 as the drawing's statement of a closed, self-powered loop. 4. Adds to "Investigator findings kept in the record": the 1999 star-mode account (Perreault and Bahlmann, interested parties, no data); the relayed Moray 4 kW / 42×26×22 in / under 60 lb account; the 2018 lineage list (named devices and patent numbers); the 1998 "alpha fission" wording as earliest dated statement; the 2002 book's offer of "radon emissive material". 5. Adds the physicist's 1927 germanium dating question to Open questions, stated both ways (the quote may reflect later wording in a later edition; or the affidavit may be as dated), with the web check result if done. 6. No verdict changes.

primer.md - §5 Part B Step 4: rewrite as a hypothetical ("if the vessel held pure radon at the 1/10 atm pressure the patent gives for its residual air ...") and add one sentence saying the patent's radon comes from emanating material 4 at the base. - §7 "genuinely open" bullet 2 and Answer 3: replace "fade with the 3.82-day half-life" with the corrected control (remove or replace the emanating source; fade at 3.82 d only if radon is sealed in with no parent). - §4 table or text: add the Fig. 2 feedback loop to the "self-generating" paragraph. Add one sentence on thoron's 56 s half-life to the Rn-220 row, and the helium check (§3 item 4) to the helium sentence, tagged [calc]. - §4 Moray paragraph: add the relayed 4 kW / 42×26×22 in account, the star-mode observation (interested parties, no data) and the 1927 germanium dating question, each with source and stake. - §9 inventor's works: note the 1998 article as the earliest dated statement (with "alpha fission" wording); list the 2018 lineage devices with patent numbers; note the leaflet's informal "continuation" and the 7 May vs 4 May date difference.

example.py: no change required. Optionally relabel the Part 3 line "100 cm3 of pure radon at 0.1 atm" as "(hypothetical: pure radon at the residual-air pressure)".

Recheck

Rechecked 2026-10-04 against the current claims.md, facts.json, arbiter.md, primer.md and example.py.

Remaining defect (introduced by the QE text itself, then copied):

1. Helium volume is wrong in facts.json and arbiter.md. My section 3 item 4 gave "about 0.1 mm3 at STP" for 3.5e18 helium atoms. The correct figure is 3.5e18 / 6.022e23 = 5.8e-6 mol, times 22,414 cm3/mol = about 0.13 cm3 (130 mm3). The primer line 98 has the correct 0.13 cm3. facts.json (US7800286 summary, "about 0.1 mm3 at STP") and arbiter.md line 132 repeat my error and must read "about 0.13 cm3". The atom count 3.5e18 is correct, and the conclusion (a tiny amount of gas) is unchanged. No verdict is affected.

Final: OPEN (1 issues)