Farnsworth-PT: Science Contrarian review

Reviewer role: Science Contrarian (judge by experiment; separate "failed", "never tested", "dismissed"; apply the same rigour to the inventor and his critics).

Source status: one document, US Patent 2,172,152 (filed 21 Feb 1938, granted 5 Sep 1939), full OCR text layer read by me [read: pdftotext, all 3 text pages]. No source gap for the patent. Everything about Farnsworth's wider multipactor work, test history and commercial fate is [web] or [memory]; the folder contains none of it.

Corrections to Step 1

1. facts.json key_claim 6 and the summary say "about one fourth of electrons striking the screen pass through". The patent says (col. 2) that of the secondaries produced, "some will pass through ... the ratio being very nearly that of the apertured area to the total area, or one to four" [read]. It is a statement about secondaries, and the equality of transmission with open-area ratio is asserted, not derived. For electrons with tens to hundreds of eV, wire-edge scattering and field penetration make this only a rough approximation [memory, medium].
2. claims.md claim 2 says the repeated trips give "a greatly amplified output current". The patent only says that by proper potentials repeated trips can be obtained "and, as is now well known in the art, a greatly amplified output current secured" [read]. The patent itself asserts, not shows, the amplification, and defers to prior art (the cited 2,071,515 and 2,071,516, not in the folder).
3. claims.md claim 3 is stated as a result; in the patent it is one sentence ("The output current is linearly proportional to the electron supply at the beginning of each cycle, as stated with reference to the photosensitive cathode"), with no derivation. Also, the "stated with reference to" photosensitive passage actually says linear only "in its maximum value ... up to the point of space charge saturation" [read]. So linearity is claimed for the maximum current, not clearly for the instantaneous response.
4. Minor: facts.json "mainstream_conflict: none" is fair for the physics, but "plausible" gain statements are not just unquantified: see claim 2 below, where the external record gives an unflattering number. Also the facts.json cathode "heater cathode" and "oxidised caesium" are correct [read]; the "1958" filing year is OCR noise as Step 1 said (text says 1938).
5. The patent says (claim 6 only) that the accelerating anode and cathode are "energized in opposite phase by a carrier current", a detail Step 1 omitted. Claims 1-5 do not require it.

Web checks done (budget 6; 3 used)

• Multipactor history: Gutton observed it in 1924 and Farnsworth studied it from 1934 as an amplifier/oscillator; the name comes from "AC electron multiplier". [web] https://www.microwavejournal.com/articles/11116-multipaction-discharge-in-coaxial-components and search summary of https://lampes-et-tubes.info/sp/The_Farnsworth_Multipactor_Tube.pdf. This upgrades Step 1's medium-confidence [memory] attribution; I treat the "Gutton 1924" detail as single-source, not confirmed.
• Resonance condition: electron transit time between impacts must be an (odd) integer number of half RF periods and secondary yield must exceed 1. [web] https://arxiv.org/pdf/1608.03096 and https://lss.fnal.gov/archive/test-fn/1000/fermilab-fn-1134-td.pdf (via search snippets; not read in full).
• Contemporary performance report: Halloran, "The Farnsworth Multipactor Tube" (Borderland Sciences reprint of a 1930s article): "an undistorted output of 25 watts was obtained from an input of 25 watts", 100 W undistorted from a small tube before the cathodes were destroyed by heat, 200 kHz to 60 MHz, a 1934 San Francisco to Honolulu/New York test; limited by cathode heating; no independent verification [web] https://borderlandsciences.org/journal/vol/44/n05/Farnsworth_Multipactor_Tube.html. This is a trade-press/promotional-era source reproduced by an alternative-science site, so reliability is modest. Search-engine summaries quoting gains of 100,000 to 1,000,000 and 600 W oscillator / 1,000 W amplifier figures are unsourced snippets; I do not rely on them.

Claim 1: A tube with a permeable secondary-emitting screen cathode between a helical accelerating anode and a collecting anode can multiply an electron current when driven by RF

a. Mathematics: The patent has no equations at all [read]. Qualitatively the physics is consistent: electrons accelerated by the helix hit the screen, yield delta > 1 secondaries, and some pass out the far side to the collector. Net current multiplication per pass requires delta x (transmission) > ... but the patent never writes a gain budget. One weakness: electrons that pass straight through the open ~25 percent of the gauze (not striking wire) are not multiplied at all; the geometry limits per-pass efficiency. Nothing internally contradictory.

b. Experiment: Secondary emission with delta > 1 for oxidised/caesiated silver is textbook [memory, high]. Multipactor discharge is real and heavily studied in RF cavities, satellites and accelerators, where it is a nuisance, with resonance conditions matching the patent's description of phase-dependent returning electrons [web, arxiv 1608.03096]. No established result contradicts the device concept.

c. Tested: Farnsworth's company built and demonstrated multipactor tubes in the mid-1930s (Halloran article, [web], promotional-era, not independently verified). The specific permeable-cathode/helix geometry of this patent: I found no test record; the patent itself gives no data. Not replicated by independent labs that I could find; this is a gap in my search, not proof of absence.

d. If true: nothing. It is a conventional (if exotic) electron-multiplier amplifier. No energy claim is made.

Verdict: Plausible, untested as a documented device for this exact geometry; the underlying principle (multipactor/secondary-emission multiplication) is Established.

Claim 2: The repeated-trip action gives greatly amplified output current

a. Mathematics: Internally consistent as a phase-focusing picture (electrons leaving when the helix is most positive return through it). But the patent gives no gain formula. Physically, repeated trips only multiply if the product of yield and capture per trip exceeds 1 while transit time matches half an RF period; the patent offers "proper adjustment of the potentials" as the only condition. Not checkable as written.

b. Experiment: Multipactor growth is a documented avalanche process [web]. But multipactor saturates by space charge and by phase slip, and is notoriously not a linear, low-distortion amplification mechanism. Fair question: how linear an amplifier an avalanche device can be. The claim that a single tube gives "greatly amplified" output is not contradicted by physics but is also not quantified.

c. Tested: The only performance number I found is Halloran's 25 W out for 25 W in "undistorted" [web, modest reliability], which is a power gain near 1, not "greatly amplified", although the article describes it as notable for the power level. I cannot tell whether that is power gain of the same tube type as in this patent. Be careful: the patent's own "now well known in the art" is a pointer to the 2,071,515/516 patents, which I did not read. Tube was commercially licensed in the mid-1930s per the same article, but the tube did not enter mainstream use; the article offers no reason, and I have no source on why. Dismissed without testing? No evidence of that. Failed after testing? No source shows that either. The honest status is that practical limits (cathode heating) are the only documented problem.

d. If true: nothing; at most it would make multipactor a useful gain mechanism, contrary to the later engineering view of it as a hazard.

Verdict: Never properly tested (no neutral, controlled measurement on record for this configuration; only inventor-affiliated reports) with gain "greatly amplified" Not testable as stated until a gain figure is given. Split: principle of electron-multiplication gain Plausible, untested; the word "greatly" is Not testable as stated.

Claim 3: Output current is linear in electron supply per cycle, so a control-electrode signal modulates the carrier in the same tube

a. Mathematics: If each cycle's output is (starting charge) x (multiplication factor M) and M is independent of the starting charge, then output is linear in it; this is trivially true in the regime below space-charge limits. The patent asserts it and does not state M's dependence on amplitude. Note the tension I flagged in Correction 3: linearity is stated for the maximum current, which hints at saturation behaviour. Consistent but unproven.

b. Experiment: Grid/control-electrode modulation of a beam is standard in electron guns [memory, high]. No contrary result.

c. Tested: no linearity data in the patent. Halloran article claims "undistorted" output [web, modest]. No independent distortion measurement found.

d. If true: nothing.

Verdict: Plausible, untested (no distortion or linearity measurement on record).

Claim 4: Above ~10 MHz a starting source of electrons is needed; photo-initiated output is linear in light up to space-charge saturation

a. Mathematics: Consistent in outline. Avalanche growth takes a number of RF cycles from a seed; at higher frequency the same number of cycles is shorter in time, so a seed helps. Numerical check I did: an electron accelerated through 100 V moves at about 5.9e6 m/s [memory, high; v = 5.93e5 sqrt(V) m/s], so crossing 1 cm takes about 1.7 ns, which is half an RF period at about 300 MHz. At 10 MHz the half-period is 50 ns, so a 1 cm gap at 100 V would have electrons arriving well within a quarter-cycle, so geometry/potentials would need to be set for lower-order resonance or larger transit time. The patent gives no dimensions or voltages, so I cannot check the 10 MHz figure; it is presented as an empirical finding ("it is found necessary"). Photoemission current proportional to light is the standard result below saturation [memory, high].

b. Experiment: Photoelectric linearity and space-charge saturation (Child-Langmuir) are textbook. Seed electrons for multipactor start-up are a known issue (stray electrons, cosmic rays normally supply them) [memory, medium].

c. Tested: no data in the patent, no independent test found. Not replicated.

d. If true: nothing.

Verdict: Plausible, untested.

Claim 5: No focusing fields needed; simple combined amplification and modulation

a. Mathematics: No equations. Self-consistent: the cylindrical symmetry and resonant return through the helix are presented as replacing magnetic focusing. A genuine weakness: electrons not emitted at the right phase are lost to the collector, which suggests a phase-selected, not full-current, efficiency. Never quantified.

b. Experiment: Nothing contradictory. Many multipactor devices operate without magnetic fields (multipactor in cavities is itself field-free).

c. Tested: The patent includes no measured gain, bandwidth or noise (Step 1 correct). Halloran gives power figures only. No independent bandwidth or noise measurement found. The "simple" claim is a design opinion.

d. If true: nothing.

Verdict: Not testable as stated ("simple", "no focusing fields necessary" are design remarks); the narrower statement "works without magnetic focusing" is Never properly tested.

Contrarian assessment: the strongest honest case and its limits

Strongest case for: Farnsworth was a competent inventor with working electron tubes behind him (image dissector). A multipactor amplifier built on secondary emission is physically sound, and a company demonstrated tubes at power levels (25 to 100 W) in 1934 [web, modest reliability]. The vacuum-tube field largely abandoned the AC-multiplier approach, and I found no documentation of a neutral head-to-head test against conventional tubes, only the absence of adoption. There is no evidence of suppression or blocked funding in the sources I have; I will not assert any.

Where I apply the same rigour to the inventor: the patent has zero data, no dimensions, no voltages and no gain figure; "greatly amplified" is borrowed from prior art; linearity is asserted without measurement; the 25 W in/25 W out report, if representative, means little amplification and is in tension with the "greatly amplified" language. Multipactor is inherently an avalanche that saturates, so claims of high-fidelity linear amplification deserve a distortion measurement. Critics, equally, have no recorded negative test; "it was never adopted" is not "it was tested and failed".

Single experiment that would settle it: build the patent geometry and record gain, third-order distortion and noise vs. frequency and bias, with and without a seed source.

Summary table

No claim here conflicts with mainstream physics, so "what would change in mainstream physics" is nothing for every claim.