Science Contrarian review: Znidarsic-Frank

Reviewer: Science Contrarian (Step 2). Judges by experimental evidence, not consensus; builds the strongest honest case; applies the same rigour to the inventor as to the critics. Tags: [read] from the documents (page images where noted), [calc] my own calculation (project venv, CODATA constants), [web] with URL, [memory] with confidence. Inputs used: facts.json, claims.md, deep/ (both), sites.md "Leads for review", research/ notes, research/files/ (NEN-1999-03.pdf opened by me). I rendered and read these pages myself: Control pp.4, 9, 10, 11; Flow p.3. I read Control pp.9-13 as text (Step 1 had only digested them). Source folders untouched.

No source gap: both papers have full text layers. Note for the reader: the project's evidence base is two short papers (Control of the Natural Forces, rev. 2/2011, 15 pp.; The Flow of Energy, Physics Procedia 20, 2011, 8 pp.) plus the Investigator's finds. The inventor's earlier papers (JNE 1996 and 2000, Infinite Energy 1998, the 2005/2006 GSJ papers) were not obtained. Some points below rest on the inventor's later book and a 1997 letter, not on those papers.

0. The bottom line, up front

1. The one number everything hangs on, Vt = 1.094e6 m/s, equals alpha*c/2 (half the Bohr-orbit speed) to 5 figures [calc]. Znidarsic's "derivations" of h, the Bohr radius and alpha from Vt are rearrangements of how Vt was built (his tabulated Compton frequency already contains h). I confirm the Step 1 finding and extend it (section "Corrections to Step 1" and claim 3).
2. The experimental anchors are thin, and the inventor himself documents that. His own bench record, as found: one marginal unrepeated anomaly (1997), one explicit null (1999). The Podkletnov effect has no successful independent replication on record.
3. The strongest honest case is narrow: Znidarsic reported his weak and null results candidly (a credit, and rare); the numerology about ~1e6 m/s may point to something real (atomic-scale velocity ~alpha*c is the electron's natural speed in atoms, so "atomic speeds meet nuclear-scale springs" is a coincidence worth one calculation, section Claim 2); and LENR and the Podkletnov effect have not been dismissed on the basis of clean, large, controlled negative tests of the specific RF + rotation configuration he proposes. None of that rescues the central derivation claims.

1. How the claims were handled (rejection analysis, applied to this project)

Separating "tested and failed", "never properly tested" and "dismissed without testing":

Anomaly about the record going quiet (SOURCE HONESTY, absence of evidence): the claimed 1997 application, the 1998 NASA contact, the unfinished MSFC RF runs, and the "Zero Point Technologies" activity (1999) all vanish from the record after about 2000, with next appearance in 2005-2012 self-published or low-barrier venues. Competing explanations: (a) the null results ended the experimental programme (supported by his own 1997 and 1999 reports); (b) a spare-time hobbyist programme with no funding (supported: P.E. day job in steel, mining and utilities per NPA wiki); (c) abandoned patent (consistent; unproven); (d) suppression (no evidence found; the inventor says the patent office denied cold-fusion patents, his characterisation). What would decide: USPTO file wrapper; MSFC correspondence under FOIA; the omitted earlier papers. I take (a)+(b) as best supported by what is on record, while saying plainly that nothing excludes (c).

2. Claim by claim

Verdict scale from prompt.txt. For each claim: a (math consistent?), b (consistent with established experiments?), c (tested?), d (what would change in mainstream physics?).

Claim 1: "Transitional velocity" Vt ~ 1.094e6 m/s, observed in cold fusion and Podkletnov-type experiments, is the speed of sound in the nucleus.

Split into parts.

1A. Vt is a measured velocity seen in LENR and Podkletnov experiments. - a. The numbers do not carry the claim. In the Control paper Eq. 1, Vt = (2 pi fc/n)(n rp) = 2 pi fc rp; n cancels; the four-figure value comes from the tabulated Compton frequency fc and rp = re/2, not from any measurement [read: Control p.2 and deep read; [calc] 2 pi (1.236e20)(1.409e-15) = 1.0942e6]. The text calls it "empirical" but supplies only order-of-magnitude products: 50 nm x a thermal frequency in LENR (no number given in Control; Flow gives omega_N ~ 1e13-1e14, whose product with 5e-8 m spans 5e5-5e6 m/s, uncited [read: Flow p.2 per deep read]), and 3 MHz x 1/3 m for Podkletnov = 1.0e6 [calc]. These use different definitions (angular vs plain frequency; radius vs diameter). A factor 2 pi (6.3) ambiguity is larger than the "agreement" claimed. In Flow Sec. 3 the superconductor product 1e14 x 1e-9 = 1e5 m/s, ten times short of the 1e6 stated [calc; read: Flow p.3 page image]. - b. Not contradicted by data because no data are given. The only dated empirical origin is the 1997 letter: rule f(MHz) = 37/L(inch), about 9.4e5 m/s, from one anomaly at 11 MHz on a 3.5-inch disk "just above the noise", afterwards "We cannot now repeat this" [web, Investigator, 1997 letter]. One data point fixes a one-parameter rule; it cannot test it. The inventor's later book tells a different order of discovery (Podkletnov's disk and Poole's textbook remark first) [read: Investigator, research/files book, only partly read by me]. Both are his accounts; they differ. - c. Tested by the inventor in 1997-99: null (see section 1). By others: Podkletnov effect not reproduced by the groups listed above (none tested the 1e6 m/s hypothesis as such; the Hathaway RF + rotating-disk test is the closest and found nothing). LENR: no experiment has been done that looks for a 1e6 m/s product in the 50 nm domain. - d. If a universal ~1e6 m/s transition speed controlled LENR and superconductor anomalies, it would require a mechanism coupling lattice dynamics to nuclei at 1e-8 m scales, which condensed-matter and nuclear theory does not contain; no one has even proposed an equation of motion for it. Not a small change. - Verdict 1A: Never properly tested, with the inventor's own weak runs on record as "tested, not supported" for his RF-superconductor rig. As a measured constant it is not testable as stated, because no measurement procedure is given.

1B. Vt is the speed of sound in the nucleus. - a. Eq. 6 (Control; Eq. 5 in Flow): Vt = (1/2 pi) sqrt((Fmax/2rn)/Mn) x 2rn [read: Control p.4 page image]. Arithmetic is right: with rn = 1.36 fm and nucleon mass it gives 1.0932e6 to 1.0948e6 m/s depending on which mass (neutron, 1.67e-27) [calc: 1.0948e6 with 1.67e-27 kg; 1.0932e6 with neutron mass; 1.0940e6 with proton mass]. But the "spring" uses Fmax = 29.05 N, the Coulomb force between two electron charges at the classical electron radius. The paper's own justification is the sentence that the electrical spring constant "equals the elastic constant of the strong nuclear force at points where the expansive electromagnetic force balances the compressive strong nuclear force" [read: Control p.3]. This is asserted, with no calculation, and it mixes an electromagnetic force at 2.8 fm with a strong-force spring and then says "The electrical force is expelled from the nucleus, does not act between nucleons, and was factored into the calculation" [read: Control p.3]. That is self-contradictory as stated: the electric force is used to supply the spring constant and is said not to act between nucleons. - Sensitivity of the match [calc]: Eq. 6 reduces to Vt = (1/pi) sqrt(Fmax rn/(2 Mn)), so Vt scales as sqrt(rn). rn needed for exactly alphac/2 is 1.3597 fm with the proton mass. Values and results: 1.0 fm -> 0.94e6; 1.2 -> 1.03e6; 1.25 -> 1.05e6; 1.36 -> 1.094e6; 1.5 -> 1.15e6 m/s. So any nuclear length of 1.2-1.5 fm lands within 6% of 1e6; the 0.1% agreement is obtained because rn was quoted to three figures as 1.36. That is one tuned parameter reproducing one number, not a prediction. Origin of 1.36: cited only to Modarres, Europhys. Lett. 3, 1083 (1987), no calculation shown [read: Control p.15]. Standard nuclear-matter physics has a Fermi momentum k_F ~ 1.33 fm^-1 at saturation density [web: https://arxiv.org/html/nucl-th/0312020 summary: "k_f ~ 1.33 fm^-1"], an inverse length. The label "nuclear Fermi momentum spacing" with the value 1.36 suggests but does not prove a reuse of k_F (in fm^-1) as a length (in fm). I have not seen Modarres 1987. Mark: suspected units confusion, unverified (Step 1 flagged this as [memory medium-high]; I upgrade it to [web] for the k_F value only). - b. Actual nuclear sound speed: for nuclear matter, first-sound speed c1 = sqrt(K/(9m)) [web: https://arxiv.org/pdf/2207.01884, as summarized: "sound velocity c = c1 = sqrt(K/9m)"], with incompressibility K_inf ~ 220-235 MeV from giant-monopole-resonance data on 208Pb [web: same search, https://arxiv.org/pdf/2207.01884 and https://arxiv.org/html/nucl-th/0309002v3]. [calc] sqrt(230/(9 x 938.9)) = 0.165, i.e. c1 ~ 0.165c ~ 4.9e7 m/s. Vt = 1.09e6 m/s is 0.022 of that. So the identification "Vt = speed of sound in the nucleus" is refuted by calculation against measured nuclear compressibility, unless "speed of sound" is redefined as a property of the electron-force spring Znidarsic invented. The authors do not compare with any measured nuclear compressibility. - c. No test; not formulated as a test. - d. Would require nucleons to behave as oscillators on an electromagnetic spring constant; contradicts the strong-force-dominated structure measured in nuclear scattering. - Verdict 1B: Refuted by calculation for the plain meaning (speed of sound in nuclear matter, measured via the giant monopole resonance). In its redefined meaning (a spring built from Fmax) it is Internally inconsistent* (electric force used and excluded in the same paragraph) and rests on an unsourced rn.

Claim 2: Quantum transitions occur where the speed of light in the electron shell equals the nuclear sound speed Vt ("impedance match").

• a. Not a mathematical statement in the papers. Eqs. 21-22 (Control) are the standard transmission-line formulas (v = 1/sqrt(LC), impedance sqrt(L/C)); no L or C is given, nothing is computed [read: Control p.8 per deep read]. "Speed of light within the electronic structure of the atom" is never defined; the only electron speeds that appear are alphac/n (Bohr) and Vt itself. The statement that the speed of light in the shell "equals" Vt is a restatement that Vt = alphac/2 is, up to the factor 2, the Bohr speed [calc]. Hence the "match" at n = 1 is: the electron's orbital speed equals about twice the nuclear "sound speed" Vt, where Vt was built from the electron's own constants.
• Eqs. 24-32 (Control, which Step 1 had not read in depth; I read pp.9-11 images): Eq. 24 gives Vt = n omega rt, Eq. 25 sets n omega rt = n sqrt(K/M) rp, so the same omega gives rt = rp, whereas rt is later the Bohr-scale radius. Eq. 29 is Vt^2 = n^2 Fmax rp^2/(rt M), the same as Eq. 19, i.e. Bohr orbit radius again. Eq. 31 bracket = 4 pi Fmax rp^2/Vt, which [calc] equals e^2/(4 eps0 Vt) = h, so Eq. 32: rt^2 = n h/(8 pi^2 M fe) is declared "the known formulation for the amplitude of electronic harmonic motion squared". For a harmonic oscillator of energy n h f the amplitude squared is n h/(2 pi^2 M f) [memory, high; E = 2 pi^2 M f^2 A^2], four times larger than printed. The paper does not define rt as a half-amplitude, so Eq. 32 is off by 4 unless rt is redefined. The paper then asserts that the "probability of transition is proportionate to the amplitude squared" with no calculation of any transition rate, selection rule or Einstein coefficient.
• b. Established atomic physics (spectra, selection rules, transition rates from time-dependent perturbation theory, Lamb shift, g-2) agrees with QED to many digits; Znidarsic's model reproduces only the Bohr model output (energies/radii in hydrogen), which Bohr and Schrodinger already give. No fine-structure splitting, no He, no multi-electron atom, no line intensities computed. The "Z" extension, Eq. 19B, is the hydrogen-like formula rx = n^2 a0/Z [calc], plus a statement about "fractional values of frequency" that is undeveloped [read: Control p.7]. The muonic atom claim has no numbers.
• c. Never tested; no observable differing from QM is predicted. There is one testable-sounding claim: that transitions are gated by a velocity match at Vt. I can see no way to measure "velocity of light in the electron shell" independent of the Bohr speed.
• d. The Copenhagen picture of the quantum state would have to be replaced by a classical-causal one; Bell-test results (nonlocal correlations, loophole-free tests) constrain any classical local reconstruction. The papers do not address Bell tests. (Step 1 did not note this.) [memory, high]
• Verdict: Not testable as stated, with the specific mathematical content (Eqs. 24-32) Internally inconsistent at the points above (rt/rp, factor 4 in Eq. 32). Strongest honest case: the idea that atomic transitions involve a velocity matching condition between different degrees of freedom is a recognised flavour of phase-matching and resonance conditions; but nothing here is predictive.

Claim 3: Planck's constant, hydrogen radii, the fine-structure constant, the Compton frequency and the de Broglie wavelength "follow classically" from Vt.

• a. Confirmed from the deep reads and re-checked by me: Vt = e^2/(4 eps0 h) = alpha*c/2 [calc: 1.093846e6 m/s, both forms agree to all digits]. Fmax = 29.0535 N [calc] equals e^2/(4 pi eps0 re^2); rp = re/2. Therefore:
• Eq. 15 (E = h ft): bracket Q^2/(4 eps0 Vt) returns h identically because Vt contains h through fc. The capacitor model needs three unmotivated choices: area = lambda^2, spacing = lambda (Control) or lambda/2 (Flow Eq. 7; the two papers disagree on this and on the factor-of-two bookkeeping), and a factor 2 for "two degrees of freedom"; and wavelength = Vt/ft rather than c/f. Control Eqs. 9-10 give C = 2 eps0 lambda while Flow Eqs. 7-8 give C = eps0 lambda^2/(lambda/2) = 2 eps0 lambda, same end result via different geometry [read: Control p.5-6, Flow p.5 per deep reads]. Different choices, identical outcome: evidence the geometry was fitted to the target.
• Eq. 20 (rx = n^2 rh): equivalent to the Bohr orbit with v = alpha c/n [calc]; h enters via Vt^2.
• Eq. 23 (alpha = 2 Vt/c): definition of Vt restated. The fine-structure argument is in words: "difference between the longer fine line and shorter fine line divided by the longer line yields the fine structure constant" [read: Control p.8]. Real fine-structure splitting is of order alpha^2 of the gross transition: for H-alpha the splitting of the n=2 levels (~10.9 GHz, [memory, medium]) over 4.57e14 Hz is 2.4e-5, compared with alpha^2 = 5.3e-5 and alpha = 7.3e-3 [calc]. The claim as worded does not fit spectroscopy. (A "fine line" ratio could mean something else in his picture; he does not define it.)
• de Broglie (Eqs. 35-41 Control; 19-23 Flow): the phase-matching algebra gives ct = h/(2Mv) [calc; Control Eq. 40 as printed], then Control sets the wavelength equal to ct (Eq. 41), producing h/Mv without the factor 2 that Eq. 40 has; Flow instead sets lambda = 2ct. The two papers do the factor-2 step differently and neither justifies it. Both start with the Compton frequency mc^2/h, so h is again an input. [calc; read: Control pp.12-13 text]
• Eq. 3 vs Eq. 2 (from Step 1, verified: Fmax rp^2 = e^2/(16 pi eps0), a quarter of the Coulomb constant). It hides a factor 4 that cancels against Vt^2 = (alpha c)^2/4 in Eq. 19. This is a pairing of two errors that happens to give a0.
• Eq. 27/Flow junction spacing rJ = 8 pi a0 = 1.33 nm [calc]. With the stated Cooper-pair charge e = 3.2e-19 C the formula gives 3.3e-10 m, not 1.33 nm; the quoted number comes from e = 1.6e-19 C [calc; deep read]. This is a disagreement between the text and the number.
• b. Consistent with established values because the inputs were established values. This is the opposite of a test: the output matches CODATA because CODATA constants went in.
• c. Not applicable as a prediction. Possible independent content: Eq. 6 (nucleon spring) matching to 0.1%; see Claim 1B. Anything beyond: no new numerical prediction (e.g., a shift in a measured energy level, a value of a constant not already used).
• d. Nothing in physics would change; there is no new constraint. What would be required for the claimed "classical origin" of quantisation: reproduce a result not already used, such as the helium ground state energy (-79.0 eV) or the Lamb shift, from Vt.
• Verdict: Internally inconsistent in the sense of circular (outputs are inputs) for the h, a0 and alpha derivations; the de Broglie step carries an unexplained factor 2 that differs between papers. Split: the algebra is correct line by line (no arithmetic errors found except Eq. 3 factor 4, Eq. 32 factor 4, Eq. 27 e vs 2e), but the physical claim "derives h" is Refuted by calculation (it is an identity Vt = e^2/(4 eps0 h)). The one non-circular item (Eq. 6) is Never properly tested as an independent coincidence, with a source gap for rn.

Claim 4: The Josephson-junction scale (~1.33 nm) and the LENR domain share the ~1e6 m/s transition speed, so LENR releases are discrete bursts like electron-pair transitions.

• a. rJ = 2(4 pi hbar/e)^2 eps0/me [read: Flow Eq. 27 per deep read] is 8 pi a0 = 1.3300e-9 m [calc], an atomic length times 8 pi. It is not a derivation of the junction: any 1e6 m/s = fe/2 x rJ relation gives a nanometre scale for any ~1e15 Hz frequency (fe here = Vt/(4 pi a0) = 1.645e15 Hz, half the Rydberg frequency [calc]). The paper says fe is "of the range of the LENR" but its own LENR frequency is 1.6e12-1.6e13 Hz, 100-1000 times lower; and "about twice that of the gravitational anomaly experiments" though no such frequency is given and Robertson's 1e14 figure is 100 times lower [read: Flow p.7 per deep read]. The superconductor product 1e14 x 1e-9 gives 1e5 m/s, not 1e6 [calc].
• b. Typical Josephson tunnel barriers are about 1-3 nm and cuprate coherence lengths 1-2 nm [memory, medium; not rechecked], so 1.33 nm is "in range". But it is in range of every nanometre-scale structure. The step to "LENR energy releases are discrete bursts like electron-pair transitions" has no data: the paper reports none. Experimental LENR claims rest on calorimetry, helium-4 and tritium measurements and CR-39 pits [memory, medium]; the paper takes one of them (ref 6) and says "very little, if no, radiation", whereas that paper reported pits that its authors attributed to charged particles [memory, medium; per deep read]. Also ref 5 (Lawson 1957) is a light-ion thermonuclear criterion, cited for heavy-element transmutation [memory, high; per deep read]: the citation does not support its sentence.
• c. No test. Robertson's Josephson-junction mass-change mechanism (NASA MSFC 1999, NTRS 19990046249) is a separate, earlier, theoretical proposal tied to Woodward-style transient mass shifts; it makes a quantitative prediction and proposes an experiment [read: Investigator; I did not re-read]. I did not find a test of it. The Robertson chapter appears in Modanese and Robertson (eds.), Gravity-Superconductors Interactions, Bentham 2012 and is cited in Gallerati et al., Frontiers in Physics 2022 [web: https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2022.941858/xml]. That a mainstream-journal review cites the Robertson line of work keeps this part in the record. It does not support Znidarsic's Vt identification. Robertson is a co-author and the Flow paper's peer-review statement is by his institute, which is the same entity: an interested party for that statement [read: Flow p.1].
• d. A nuclear (LENR) energy source with no commensurate radiation would conflict with measured nuclear branching and gamma-ray emission; the proposed "long-range convergence of forces" has no equation behind it.
• Verdict: Never properly tested (no experiment reported or on record that targets the linkage), with the numerical case Not testable as stated until a measured LENR timing or length is given. The 1.33 nm figure is Internally inconsistent with its own stated charge.

Claim 5: At transition speeds at or above ~1e6 m/s new phenomena occur, and "free" (vacuum or dark) energy can be pulled from ~1e-9 m interactions.

• a. No equation. Hedged by the authors ("would seem to imply", "appear to place", "may lead to") [read: Flow pp.3-4, 7-8]. The statement "a minimum velocity ... from which free energy can be pulled" is not quantified: no energy, power, or balance is given.
• b. Vacuum energy extraction from a nanometre-scale system would be a violation of the second law as normally stated if cyclic and net-positive; the paper does not claim a device, so there is nothing to compare against measurement. If instead "free energy" means ordinary stored energy released in LENR, it is a different claim. Nothing in the Casimir/zero-point literature gives a 1e6 m/s threshold [memory, medium].
• c. The inventor's tests (Section 1) looked for conversion of reactive power to real power ("absorbs vars and produces watts") with a 3.5-inch disk: one marginal anomaly, then none [web: 1997 letter, via Investigator]. The 1999 spinning-disk run: "No anomalous energy was observed" [read: NEN-1999-03.pdf, p.2]. Detail from my reading of the 1999 notice: the measured quantity was the RF decay time of the spark-excited circuit (rung 400 times/s, dying in ~50 microseconds); no gravity measurement was made; the disk's material and size are not stated; rpm was at most ~1,000, below the 10,000 rpm the editor mentions for reported effects; and the frequency range is printed as "100,000 Khz to 40 Mhz", which cannot be right (100,000 kHz is 100 MHz, above 40 MHz), probably 100 kHz. The editor thanked him and noted the "Aspden Effect" at 3,500 rpm [read]. This is a single-bench, low-sensitivity test, so "tested, not supported" applies only to that rig. It does not test his later (2011) proposal to excite a Bose condensate at a stated MHz-metre product.
• d. A net energy gain from the vacuum would require rewriting thermodynamics and energy conservation in the form given by Noether's theorem; the burden is on a demonstration, which the authors do not offer.
• Verdict: Never properly tested as a physical claim; as a written sentence, Not testable as stated. The only energy balance anywhere in the record is the null of 1997/1999.

Claim 6: Forcing a macroscopic body into a single transitional quantum state by Bose-condensate vibration at ~1.094 MHz-metres would induce strong gravitational and long-range nuclear forces (propulsion, nuclear-waste reduction, new energy).

• a. One paragraph, no equations ("New technologies", Control p.13 [read: text]). Link to Eqs.: "the constants of the motion tend toward the electromagnetic in a Bose condensate that is stimulated at a dimensional frequency of 1.094 megahertz-meters" [read: Control p.11]. Note "megahertz-metres" = 1.094e6 m/s only when the "frequency" is plain Hz x length (consistent with Podkletnov's 3 MHz x 1/3 m) and not with Eq. 1 (angular frequency x radius). The claim that the electric force between two electrons (29.05 N) "is of the magnitude to induce the gravitational field of the electron" [read: Control p.11] has no equation.
• b. The inventor's scaling rule: 37 MHz-inch (1997, 9.4e5 m/s) versus 1.094 MHz-metre = 27.8 MHz-inch (2011) versus Podkletnov's 3 MHz x 1/3 m = 1 MHz-metre = 25.4 MHz-inch [calc]. The 1997 number is 35% above the later ones. The NPA wiki already asks whether the constant was ever established by a range of measurements or fixed by one point [web, Investigator; the wiki is a dissident-science society's page: stake noted]. Established gravity physics (equivalence principle, torsion-balance tests, gravimeters) sets limits at the 1e-8 to 1e-13 level on anomalous weight changes near test objects [memory, medium]. Noever and Koczor limit < 2 parts in 1e8 of g in their configuration [read: Investigator, NTRS 19990023209].
• c. The designated experiment (a rotating/vibrating Bose condensate driven at the MHz-metre product) is the Podkletnov configuration; see section 1: null in the replications that came closest (Hathaway et al. 2003), exact rig never rebuilt. The inventor says the RF runs at NASA were never completed because of budget cuts and quotes BAE (Greenglow, Ronald Evans) saying the RF method was not attempted for funding reasons [web: Investigator; the inventor's own account, no second source found yet]. I take this as "partly untested" for the RF part of the configuration. But Hathaway et al. did combine RF and rotation (their title) with no effect at 0.001%, which narrows the untested part.
• d. Gravity from a laboratory condensate at observable strength would require coupling of the gravitational field to superconductor quantum phase far beyond known weak couplings (general relativity's coupling is ~1e-39 of electromagnetism for protons [memory, high]); Modanese's Frontiers review discusses theoretical suggestions but also the null results [web, same Frontiers link; not read in full by me].
• Verdict: Never properly tested for the exact RF-driven Bose-condensate rig (with the near-neighbour test of Hathaway et al. and the inventor's own tests Tested, not supported); the predicted strong-force and "nuclear waste reduction" effects are Not testable as stated (no magnitude, frequency spec, or readout).

3. Honest case for the inventor, and where I turn it against him

For the claims (the "strongest honest case"): 1. A numerology that selects alphac/2 can be a legitimate observation if tied to data. A small, honest observation: three scales (atomic velocity alphac, nuclear-scale spring with rn ~ 1.36 fm, and ~1e6 m/s lab products) coincide to an order of magnitude. At least one genuine, non-circular match (Eq. 6, 0.1%) exists as written. Whether it is a coincidence depends only on where rn comes from; this is checkable by reading Modarres (1987). 2. The inventor reports his own null and marginal results, including "We cannot now repeat this", and states his patent was filed before seeking feedback. That behaviour is a mark of honest practice rather than salesmanship; the NPA wiki says the same [web, the wiki's characterisation]. 3. The Podkletnov question was not closed by a clean, exact replication. The mainstream tests differ from the original in disk size, temperature, drive; NASA MSFC could not complete its rig [web: Frontiers/Millis]. Mainstream summaries say "no replication", which is true, but it is different from "tested and falsified in the original configuration". 4. LENR is a field where the 2004 DOE panel was split; claims about LENR should not be dismissed from authority. But Znidarsic's papers add no LENR data.

Against the inventor (same rigour): 1. The defining equations are circular (section Claim 3). "Planck's constant emerges" is false: h is put in through fc. 2. Unit and notation faults: omega given in Hz; "10^-24" ratio not reproducible (I get 2.8e-8 from the paper's own inputs [calc], while the paper's image shows 10^-24 [read: Flow p.3 page image]); fe "twice" gravity-experiment frequency with no such frequency given; two papers differ on geometry and factors; 3.2e-19 C vs 1.6e-19 C; Eq. 32 off by 4. 3. The empirical origin is unstable: three different statements of how Vt arose (1997 own disk; Podkletnov 3 MHz x 1/3 m; Poole's textbook remark and 20 THz x 50 nm Pd), with constants 9.4e5, 1.0e6, 1.094e6. 4. Unfalsifiable language: "causative classical explanation", "convergence of the motion constants", "impedance match" never given as quantities. 5. Poole: the book cited for "1e6 m/s is the speed of light in a superconductor" was not checked by me; it may refer to the Fermi velocity of electrons in metals (~1e6 m/s) [memory, low]. If so, the premise is a misreading; if not, a textbook remark doesn't establish a gravitational link. Unverified either way. I did not search for it. 6. Peer review of the Flow paper was under the co-author's institute (self-referential) [read: Flow p.1]; Elsevier published the proceedings, but its statement is by the institute. This is a limit on weight, not proof of error.

4. Investigator leads kept in the record (with my weighing)

5. Corrections to Step 1

1. claims.md claim 1 says Vt was "observed". The documents give no measurement of Vt; the number is built from constants (Eq. 1) and the "observations" are order-of-magnitude products with different definitions. Better wording: "Vt is asserted to be inferred from".
2. facts/claims state Eq. 6 "matches to about 0.1%" as an independent numerical match. True, but it is a single-parameter fit: Vt scales as sqrt(rn), so values of rn from 1.2 to 1.5 fm give 1.03e6-1.15e6 m/s. Step 1 should say the match depends on rn quoted to three figures.
3. Deep read says the Control digest renders Eq. 19 with a factor 2; the image has none. I confirm the deep read (the Eq. 19 image is in Control p.7; I checked the algebra by [calc]).
4. Claims.md says "Pp.9-15 of the Control paper were read from the digest only". I read pp.9-13 and found content that Step 1 omitted: Eqs. 24-32 (an "amplitude of harmonic motion" derivation with a factor-4 mismatch, Eq. 32 vs the oscillator result), the "convergence of motion constants" passage, and the Control vs Flow factor-2 difference in the de Broglie step. These belong under claims 2 and 3 (reflected above).
5. Step 1 treats "speed of sound in the nucleus" as a claim to be judged on circularity. It is also testable against measured nuclear compressibility: c1 ~ 0.165c ~ 4.9e7 m/s versus Vt ~ 1.09e6 m/s [calc; [web] K_inf ~ 230 MeV]. This is a separate refutation of the plain-language claim.
6. The Flow deep read states that the 10^-24 ratio "I could not reproduce". I reproduce the same: 2.8e-8 [calc]. Agreement noted; I add that the printed exponent is -24 (page image Flow p.3, confirmed), not an OCR slip.
7. The NEN 1999 experiment was recorded by the Investigator as "RF up to 40 MHz". The printed range is "100,000 Khz to 40 Mhz" [read], which is inconsistent as printed. The measured quantity was RF ring-down time, not weight or gravity.
8. claims.md item 4 attributes "LENR energy releases are discrete bursts" to Znidarsic and Robertson as a claim; the paper's wording is hedged ("would seem to imply", [read: Flow p.3]). Step 1 recorded the hedge for claim 5 but not claim 4.
9. Step 1 and the Investigator give the NPA wiki assessment as independent corroboration of the identity. It is consistent with my calculation, but the wiki is a single dissident-society page and was not independently sourced; my own calculation is the stronger evidence. (Hathaway et al. as cited by the wiki is verified by my [web] check above.)
10. Not corrected but flagged: I did not verify Modarres (1987), Poole (2007) p.4, or the full text of the Frontiers 2022 review; these remain open.

6. What would settle things

• Claim 1B / 3 (Eq. 6): read Modarres (1987) and check whether 1.36 is in fm or fm^-1; if it is k_F, the one non-circular match is a unit slip.
• Claim 3: compute the helium ground-state energy or any line intensity from Vt alone. Failing that, no new content.
• Claim 1A / 6: a run of the exact Podkletnov rig (large YBCO disk, rotation, RF at 3 MHz, sensitive gravimeter) at a lab that is not an interested party. The MSFC rig was the one candidate; not completed.
• Claim 4: a measured time or length scale for LENR events at ~50 nm with a thermal frequency (a number cited and sourced), plus a Josephson junction measurement showing the 1e6 m/s product.
• Historical record: USPTO file wrapper of the 19 Oct 1997 application; MSFC FOIA for the 1998 visit.

Verdict summary