Distinti Robert: Quality Engineering (Step 4b)

Date: 2026-10-03. Inputs checked: _text/manifest.json, all 15 digests in digest/, facts.json, claims.md, physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md, example.py. I spot-checked the extracted text in _text/. There is no deep/ folder: Step 1 ran before digests existed, and no deep reads were done.

Verdict: MINOR. The arbiter's headline verdicts are well argued and recomputed where it matters. I expect none of them to flip. But three long documents carry the experimental and derivation material behind claims 2, 3 and 6, and all three were only sampled: neThesis.pdf, nia1.pdf and nm.pdf. As a result, a whole experimental chapter (nia1 section 15), the thesis's data-correction procedure, and several measured or predicted results in nm.pdf and anomolies.pdf never reached facts.json, the reviews, the arbiter or the primer. These gaps can be fixed with deep reads and updates to facts.json, the arbiter and the primer. Steps 2-4 do not need to be redone.

---

1. Coverage

• All 10 manifest documents are digested. No document was skipped, and nothing with text went unread at the digest level.
• facts.json problems:
• (a) The how_read strings repeat the same boilerplate ("abstract, contents, key derivations and conclusion read; not every page") and then contradict it in a parenthesis. One example is paradox2: "sampled ... (nearly all 22 pages read)". These strings should say plainly what was read.
• (b) A spurious entry with path "(all files)" and how_read "filename only" describes the folder, not a document. It should become a note, or be marked type "other" with a clear title.
• (c) The nia1 entry says "Quantitative agreement tables were seen only in the companion thesis, not sampled here". This is wrong. nia1 pp.45-51 and 73-91 hold measured-versus-computed tables [read: nia1-8f67b3.txt lines 2174-2384, 3462, 3600-3860].

2. Shortcuts

3. Missed material (from the digests, confirmed in _text/)

3a. The thesis "measured" inductances are corrected per inductor (neThesis pp.67-70). - The method: - For each inductor, the author defines the true value as the reading where the no-plane and ground-plane curves intersect at low frequency. - He calls the difference between that reading and the 2 MHz reading the "stray effects". - He adds this offset as a "correction factor" to every 2 MHz reading of that inductor, with and without planes [read: neThesis-9619da.txt lines 2433-2517]. - Why it matters: - The 1.9% and 4.1% errors are therefore comparisons against processed data, not raw meter readings. - The no-plane "measurement" is taken at a frequency the author chose. - None of the reviews or the primer mention this. It bears on how much weight claim 6's "Plausible, untested independently" can carry. - A related point: the author explains why the ground plane has no effect below about 1 kHz by saying "a certain M-field intensity required to knock electrons loose" [read: line 2382]. The classical explanation is skin depth, which is far larger than 1.35-mil copper at 1 kHz [memory high]. That makes this a teachable misinterpretation the primer could use. - The abstract says "at least an order of magnitude faster", but the body says "nearly an order of magnitude" [read: lines 58, 2816]. facts.json's "10x" glosses over this.

3b. nia1 chapter 15 experiments (pp.73-91), entirely missed. The author built an "inductance template" PCB and measured it at 6 MHz with a transistor current source: - Direct coupling: 123 computed vs 125 nH measured. - Single ground plane: 40.2 vs 42.25 / 41.52 nH. - Parasitic loop: 74.62 vs 73.45 nH. - Trace inductance with one plane: 165 vs 169 nH. - Two planes: 27.5 vs 27 nH.

Notes on these results: - The two-plane "computed" value is not a computed sum. The 15-pair table sums to 23.5 nH and does not converge. The author then picks 27.5 nH by visual "Guestimation" in Excel [read: nia1-8f67b3.txt lines 3600-3790]. - I checked this: averaging the last two partial sums of his own table gives 27.9 nH [calc]. That is a legitimate way to sum an alternating series (Cesaro or Euler averaging), so the number is defensible. But the primer and arbiter should say how it was obtained. - The 97% "inductive reflectivity" of copper is a fitted constant used throughout. Nobody examined it, and it is presumably frequency dependent. - Earlier, the rectangular-loop test (nia1 pp.45-51, paper-and-glue loops) predicted 1.578e-7 H. It measured 1.49e-7 H (5.7% off) with one generator and 1.56e-7 H (1.3%) with another [read: lines 2174, 2340-2384]. - None of these numbers appear in any review file. They are more of the author's own supporting data for claim 6. Like the thesis, they test Neumann-type integrals and images, not NE as such, so the verdict should hold. But the evidence table should list them.

3c. A measured "anomaly" that classical theory already explains (anomolies.pdf p.7). - Distinti's table: 48-inch-perimeter loops of 26 AWG and 22 AWG wire measure 2253/2055 nH (circle) and 2144/1950 nH (square). He concludes that "the classical model for intrinsic inductance is incorrect" because wire thickness matters [read: anomolies-d00cc0.txt lines 268-275]. - Classically, wire radius b enters the external self-inductance through ln(8a/b), not through the mu/8pi internal term. - I computed L = mu0 a [ln(8a/b) - 7/4] for the circle and got 1754 nH (26 AWG) and 1641 nH (22 AWG) [calc]. The prediction has the same sign as the measured drop and about 60% of its size. The remaining offsets are plausibly from leads or the fixture, but the paper gives no information on them. - This is a misattribution, not an anomaly, and a good worked example for the reader. None of the reviews address it (zero hits for "AWG").

3d. nm.pdf content the reviews did not engage: - Absolute reference frame. nm pp.35-37 states outright that charge velocity is measured "relative to space"/ether, using an electron-gun-on-a-workbench thought experiment. The arbiter only infers absolute velocities. This is the author's explicit statement and bears directly on Lorentz invariance (claim 2). - "Stationary current" and "Gaussian current" (pp.29-42). This is how NM is made to give Galilean-invariant fields for neutral wires, with the ratio I_G/I_NG = V eps/(rho_F l^2). It is a concrete, checkable piece of the theory, and nobody mentions it. - Charged-disk prediction (p.35). A rotating charged insulating disk with sinusoidal angular acceleration "will" induce an emf in a coil. This is stated as a prediction, with no data. Classically there is also an induced emf, so the case does not discriminate unless amplitudes are compared. - Edge-current magnet model and MagnaView film observations (pp.55-67). These are qualitative magnet-attraction demonstrations, which the contrarian calls "magnet-attraction modelling without data in the sampled pages". The digests show photographs and observations exist. They are qualitative only. - Electron-beam coherency formula F = (Q^2/d^2)(K_E - K_M V^2) (p.53-54). The arbiter mentions the qvxb version. This should be compared explicitly with the exact relativistic result (repulsion falls as 1/gamma^2).

3e. Smaller items not carried forward: - The $50/hour challenge to solve the loop self-inductance problem, which the author says prompted NE (anomolies pp.5-6). - Guala-Valverde and Tramaglia's homopolar back-torque work, cited in anomolies p.26 and in qvxb (Am. J. Phys. 70, 1052, 2002). It is relevant to the contrarian's homopolar discussion and should be web-checked. - Paradox 2's admission that its screen captures are labelled 100 mV/div while the real scale is 100 uV/div [read: paradox2-08732d.txt line 439]. - ne_hydrogen's appendix table claims particleadventure.org gives r_e ~1e-18 m and r_p ~1e-15 m, the reverse of NE. The author's own source therefore contradicts his radii. - The ng.pdf suggestions that ether annihilation drives radioactive decay and Earth's molten core, and that light is faster at altitude. These are in facts.json but not in the review of claim 5. They are low priority.

4. Accuracy spot checks (against _text/)

All five load-bearing facts check out. I found no fabricated numbers.

5. Process

• Source tags. Present in every file, with tag counts as follows:
• physicist 23 [read] / 6 [web] / 19 [memory]
• mathematician 18 / 7 / 26
• contrarian 8 / 4 / 18
• arbiter 7 / 16 / 11
• primer 20 / 9 / 8

The tags look honest: [memory] items carry confidence ratings, and unchecked items are flagged. The contrarian is the weakest on [read] and carried a misreading (the square root on ng p.14) that the arbiter caught and withdrew. Good. - Web checks. Partly done (see section 2). The arbiter correctly declines to claim any replication or suppression. - Verdict scale. Applied consistently, with split verdicts where needed. "Plausible, untested" for the BMP Vt = c balance, while also saying the premises are inconsistent, is a slightly awkward label but it is explained. - Primer. - It follows the 9-part STEP 4 order plus answers. - It defines symbols and gives prerequisites with textbook chapters. - The worked example (BMP numbers) is concrete. - It is written for the reader profile, and the source gap is repeated. - example.py. Runs cleanly with the venv Python. It reproduces the electron-mass identity, the proton radius ratio of 547.8, hydrogen 1.00344 vs 1.00783 g/mol (0.435% low), and the zero residual between NI+NM and the quasi-static Lorentz force. - Earlier errors corrected by the arbiter. The proton factor (550, not 5.5e5), the corner formula (V_S, not V_S^2), and the square root on ng p.14. The physicist's and contrarian's files still contain the old errors, but the arbiter's "Corrections to earlier steps" section records them. That is acceptable.

6. Remediation actions

Deep reads (most important first): 1. neThesis-9619da, pp.43-79, Opus. Describe the measurement and data-processing chain exactly: - intersection choice, correction factor, frequencies used; - Tables 8-4/8-5/8-6, raw vs corrected values; - the seed-size sensitivity and the origin of the 97% reflectivity; - whether any accuracy (not speed) comparison with PI or a field solver exists.

Would the reported agreement survive using raw readings? 2. nia1-8f67b3, pp.28-41, Opus. Check the dipole derivation: - whether retarded time is applied consistently; - whether the longitudinal result (VK = 1/2 of transverse) follows from the wire-form integral; - what the ARRL -9 dB figure actually measures (height, ground, angle); - whether the "low frequencies travel farther" claim follows. 3. nia1-8f67b3, pp.42-51 and 73-91, Sonnet. Extract all measured-versus-computed couplings, the methods, instruments and frequencies, the role of the 97% coefficient and of "Guestimation", and any uncertainties stated. 4. nm-3bcad1, pp.19-42, Opus. Cover: - the author's derivation of the NM third term from BMP time dilation in all directions and from New Induction; - stationary and Gaussian currents; - the absolute (ether) reference frame; - the charged-disk prediction.

Is the derivation sound? Does it confirm or refute the mathematician's "added by hand" finding? 5. nm-3bcad1, pp.53-73, Sonnet. Cover electron-beam coherency, the edge-current magnet model, the MagnaView film observations and the homopolar derivation. Which observations were actually made, and which are only asserted? Compare the beam formula with the relativistic 1/gamma^2 result. 6. anomolies-d00cc0, pp.4-16, Sonnet. Cover: - the self- and intrinsic-inductance sections and the AWG table (any details on leads or fixture?); - reciprocity; - the flux anomaly; - Maxwell vs Kirchhoff.

Do any contain data, as opposed to argument?

File changes: - facts.json - Rewrite every how_read string so it states what was actually read, and add the deep reads. - Replace or relabel the "(all files)" entry. - Fix the nia1 summary: it has experimental tables. - Add the chapter 15 results, the thesis correction procedure, nm's absolute frame, stationary and Gaussian currents, the charged-disk prediction, and the AWG data. - claims.md - Update the Source gap note: the three long documents have now been deep-read. - Under claim 6, list nia1 ch. 15 as further author-only evidence. - Under claim 2, state the absolute-velocity premise explicitly. - arbiter.md: add a "QE revision" section that: - weighs the correction-factor procedure in the claim-6 verdict; - adds the nia1 ch. 15 results to "Tested and supported (author, narrow)"; - adds the AWG item to claim 3 as "Refuted by calculation" (the classical ln(8a/b) dependence; misattributed to intrinsic inductance); - records nm's explicit absolute frame under claim 2 and settles the "added by hand" question from deep read 4; - web-checks Guala-Valverde et al. (Am. J. Phys. 70, 1052, 2002) and, if possible, Michelson-Morley/Kennedy-Thorndike limits. - primer.md - Add to the evidence table: nia1 rectangular-loop and template results, the AWG "anomaly" (with the 1754/1641 nH classical numbers as a short second worked example), and the thesis correction factor. - In section 7, add the skin-depth versus "knock electrons loose" point as an instructive mistake, and the "Guestimation" series summing as a fair numerical technique. - example.py (optional): add the AWG loop calculation and the two-plane partial-sum averaging.

---

Recheck

Date: 2026-10-03. Checked: deep/ (4 files), facts.json (valid JSON), claims.md, arbiter.md ("QE revision"), primer.md, example.py (runs, exit 0).

Defects (section 1-3)

Remediation actions (section 6)

Remaining issues: (1) nia1 pp.28-41 dipole derivation not deep-read; (2) web checks outstanding.

Final: OPEN (2 issues)

Recheck 2

Date: 2026-10-03. Checked: arbiter.md (QE revision and Round 2), deep/nia1-8f67b3.md, facts.json, claims.md, primer.md.

Final: PASS