Follow-up Q17: Multipaction as the mechanism of Moray's cold-cathode valve

Date: 2026-10-08. Investigator: STEP F follow-up investigator (Opus). Web time: about 3 minutes by date +%s (start 1791506985, end 1791507137; Chrome, one tab, closed at the end). Stance: the Investigator's (no debunking or substantiating; leads, with tags).

The question (owner's words, lightly polished)

Multipaction as the mechanism of Moray's cold-cathode valve: a nonlinear secondary-electron avalanche that RF engineers try to eliminate. Could it explain conduction in unheated tubes (the Patent Office said the cathodes could not emit), radioactive coatings as seed electrons, a voltage and frequency window that acts as the "spill over" threshold, one-way behaviour from unequal secondary yields, and the "surges"? Is there any measured case of multipaction giving net energy rather than absorbing RF power? [The folder's aa-new multipactor papers (Kishek flyer, "Multipactor discharge on a dielectric", "Temporal evolution of multipactor discharge") are not yet weighed. Compare Farnsworth-PT (multipactor tube) and Dollard-Eric.]

Short answer

Multipaction explains several of the features in principle: cold cathodes carrying large currents, seed electrons from a radioactive source, a lower and an upper voltage threshold, a DC component when the two surfaces have unequal yields, and transient current surges. It has never been measured to give net energy. In every measured case it takes its energy from the RF field or from a DC anode supply. Farnsworth's cold-cathode multipactor oscillators of 1934-36 converted battery power to RF, with claims of up to 90% efficiency. Two points weigh against multipaction as the mechanism of Moray's valve. Moray himself calls his tubes "ionic" and says that "ionization occurs in the gases" of the tube, which describes a gas discharge. Multipaction also needs a hard vacuum and a gap-frequency product of at least about 0.4 GHz mm, so a gap of tens of centimetres at about 1 MHz.

What was read in the project

• aa-new/Kishek_flyer_v01.pdf (2017 University of Michigan colloquium flyer, Rami Kishek) [read]. Multipactor is "an electron avalanche occurring when stray electrons are trapped and accelerated by radiofrequency (rf) waves in a vacuum, hitting walls and ejecting other electrons by secondary electron emission (SEE)". It "creates noise, heat and possibly damage". Design rules use Hatch-Williams (1950s) "susceptibility diagrams" with "arbitrary margins". Kishek's chaos-theory method maps all modes, including non-resonant ones, and identifies "parameter regions that are multipactor free".
• aa-new/MULTIPACTOR DISCHARGE ON A DIELECTRIC.pdf (Kishek and Lau, 1998 IEEE, pp.3198-3200) [read]. This is single-surface multipactor on an RF window. An RF field runs parallel to the surface, and a DC field normal to it comes from the surface's own charging, measured "as high as 4 kV/cm". The DC field "does not impart any energy to the electrons"; the electron "gains energy only from the rf electric field". Eventually "a large amount of energy gained from the rf electric field will be deposited on the surface, possibly leading to surface damage or breakdown". The yield delta exceeds 1 only between two crossover energies E1 and E2. This gives a lower and an upper RF-field boundary, Erf,min = 2E1/(1 - cos(4 Eom/EDC)) and Erf,max = 2E2/(1 - cos(4 Eom/EDC)) in normalised units. Table 1: alumina delta_max 1.5-9 (Emax 350-1300 eV), Pyrex 2.3, quartz 2.4. The remedies are to change the RF level, add a slightly conductive coating, or coat with low-yield TiN.
• aa-new/Temporal Evolution of Multipactor Discharge.pdf (Kishek, Lau and Gilgenbach, 1996 IEEE, pp.1602-1604) [read]. The model is a two-plate gap driven by an RF cavity of quality factor Q. "The multipactor current saturates primarily by its loading of the cavity". At saturation "the secondary emission coefficient is unity, corresponding to the 'first cross-over point'". In a high-Q cavity "the multipactor current may reach a very high level in a transient manner before it settles down to steady state". The stored energy "may relax via a multipactor discharge, albeit transiently in time, and locally in space ... as much as 20% of the energy stored in a high Q cavity may be discharged in a single overshoot". Footnote 8 explains why E1 is stable: more secondaries drain more energy from the cavity, the gap voltage falls, and the impact energy returns toward E1. Example in dimensional units: 1 GHz, a 0.22 cm gap, delta_max 1.2 at 400 eV.
• Moray's own descriptions of the tube (Sea of Energy 4th ed.) [read]:
• p.107 (text line 4593): "The special tubes which appear to be the key to the success of this device are Ionic cold cathode tubes which require no external power sources."
• p.45 (text line ~2111): "The R. E. tubes received this energy in surges which may last only a few micro-seconds but the pressure and the current in those surges are so large that sufficient energy is delivered to the equipment in resonance" (OCR partly garbled).
• pp.55-56 (quoted in Q16): "when a certain voltage is reached, ionization occurs in the gases of the discharged tube", followed by the "bucket brigade" passage.
• pp.43-45 contain a long list of "Methods of electronic excitation" in gases and at surfaces. Item B.1 is "Bombardment of metal by fast electrons, alpha particles, positive ions or recoil atoms giving secondary atoms [sic]", which is secondary emission. Item A.1 is "electrons and beta particles from radioactive changes". It is a textbook-style list, not a description of his tube.
• p.42: "Special provision is provided to stop R. E. tubes from becoming blocked in their disipation of the charges ... the oscillatory capacity back rush effect common to capacitors and herein applied in vacuum tubes."
• Patent Office, 1932 action, as quoted by Moray's side [read: 4th ed. p.155]: "Claims rejected as being drawn to an apparently inoperative device. It is thought it will not produce a sensible current because the cathodes are not heated to a point where they will emit an appreciable number of electrons" (the book adds "such an objection is now obsolete"). The 5th ed. p.161 version reads: "Because no means was provided for causing the cathode to emit an appreciable number of electrons, the current produced in the cathode by the antenna will not heat the cathode". The arbiter's reading of the surviving action adds that "radioactive ionisation gives no space current without plate or grid potential" [read: arbiter.md, Claim 6, from the gov-Commerce image].
• Comparison projects [read: review/Farnsworth-PT/arbiter.md; review/Dollard-Eric/arbiter.md]. The Farnsworth-PT arbiter rates the mechanism of US 2,172,152 "Established" with no quantified gain, and records the 1934 Halloran article ("25 watts" out from "25 watts" in, an ambiguous figure). The Dollard-Eric arbiter records Dollard's view of the multipactor as "probably the only real free energy device that anyone ever demonstrated which can be reproduced", which it rates "Not supported (externally powered negative-resistance oscillator; no measurements given)".

Web research (Chrome; pages saved to research/pages/ and grepped)

What was found

1. Cold cathodes can carry large currents by secondary emission, but only with a supply. - Halloran, RADIO, Oct 1934 [web, Borderland reprint]: "Think of a vacuum tube without filament or grid, thus requiring neither an A nor a C battery, which generates high-frequency oscillations!" - Construction: the cathodes are "coated with caesium silver oxide" in "an evacuated glass tube". - Range: "200 kilocycles to 60 megacycles"; "an undistorted output of 25 watts was obtained from an input of 25 watts"; one small tube gave "100 watts of undistorted energy before the cathodes were destroyed by heat". - The energy source is named: "Aside from the energy which is required for the magnetic focusing field ... the only external source of energy is that which maintains a positive potential on the anode." - Wireless World, 3 Apr 1936 [web, r-type.org], on Farnsworth's IRE demonstration of 4 March 1936: - Tubes of 200, 1,000 and 4,000 W were shown: "kilowatts of power in a cold-cathode tube, at frequencies as high as 300 MHz", "with efficiency as high as 90%". - The oscillator "takes energy from the battery in the anode circuit and converts it into radio-frequency energy in a tuned circuit". - Range: "Frequencies from as low as 100 kHz to as high as 300 MHZ". - Secondary yield: each impact "liberates from three to ten 'secondary' electrons". - These efficiency figures are the company's, reported in the trade press. They are not a neutral measurement. - US 2,184,910, "Cold cathode electron discharge tube" (Farnsworth, filed 4 Nov 1936) [web, Google Patents]: the object is "an electronic oscillator tube capable of converting direct current energy at extremely high frequencies". Facing cathodes are "treated so that they will emit secondary electrons at a ratio greater than unity". It works "simply by connecting the two cathodes by a resonant circuit and supplying the anode with an accelerating potential". - Bearing on the Patent Office ground: by 1934-36, a cold-cathode tube with no heater carrying amperes was public (Farnsworth). The examiner's wording "cathodes are not heated" is therefore too narrow as physics, as the arbiter already found on other grounds. The examiner's second point, no space current without plate potential, describes Farnsworth's tube exactly: it ran only with a positive anode supply.

2. Radioactive seed sources are standard practice in multipactor testing. - ECSS-E-ST-20-01C, clause 6.5.5d [web]. "At least one of the following seeding sources shall be used". The list is: UV light; an electron gun; a charged wire probe; and "Radioactive β source, which produces high-energy electrons that, after impacting with metal surfaces or propagation though metallic or dielectric walls, yield a supply of secondary low energy seed" electrons. - Clause 6.5.2a: "An electron seed source shall be used for CW Multipactor test." - Google snippets name the isotopes: "Radioactive source (Typically Cesium-137 or Strontium-90)" (Aerospace TOR-2014-02198); "Sr-90 radioactive sources to provide additional electron seeding capability" (OSTI 2022) [web, snippets only]. - Langellotti 2022, footnote 2 [web]: "The source of the electron is not important. Many natural processes can generate this initial seed electron, including cosmic radiation or spontaneous field emission." - Halloran 1934 [web]: "There are always some free electrons present in the inter-electrode space, if only those due to photoelectric emission from the cathodes". - Wireless World 1936 [web]: the start comes from "the shock to the tube when the HT battery is suddenly connected". - The seed sets when the avalanche starts, not how much energy it has: growth and saturation are set by the RF field and the yield curve (Kishek 1996) [read].

3. A voltage and frequency window, with a self-limiting threshold. - Kishek and Lau 1998 [read] give both a lower boundary (E1) and an upper boundary (E2). Kishek, Lau and Gilgenbach 1996 [read] show that the steady state locks to delta = 1 at E1 by loading the cavity, which clamps the gap voltage. - ECSS Table 5-1 [web], lowest breakdown voltage against the product f x d (frequency times gap): - aluminium: 33.8 V at 0.43 GHz mm, falling to a minimum of 23.8 V at 0.70 GHz mm; - copper: 36.6 V at 0.47 GHz mm; - silver: 38.3 V at 0.50 GHz mm; - gold: 41.0 V at 0.49 GHz mm. - Below about 0.43-0.50 GHz mm there is no first-order breakdown. - SEY (secondary electron yield) parameters for aluminium: E1 = 17 eV, SEY_max 2.92 at 276 eV (Table 9-1). - ECSS scope [web]: "high vacuum conditions (pressure lower than 10-5 hPa)"; "does not include breakdown processes caused by collisional processes, such as plasma formation". Clause 6.2c: testing "shall be performed at pressures below 1,0 x 10-5 hPa". - Rasch (Chalmers) [web]: "The transition from gas breakdown to multipactor appears in the pressure region where the mean free path of electrons, l, is of the same order of magnitude as the system dimensions, L."

4. Net energy: no measured case found. - Every source read treats multipactor as a load on the RF (Kishek 1996: saturation "by its loading of the cavity"; Kishek and Lau 1998: energy from the RF "deposited on the surface"). The ECSS detection methods (close-to-carrier noise, phase noise, harmonics, phase nulling, pulse envelope, optical, electron probe; clause 7.2) all look for the discharge's disturbance of the RF signal. - The one device built as a source, Farnsworth's oscillator, names its source: the anode battery (Halloran 1934; Wireless World 1936; US 2,184,910). - Borderland Sciences' 1989 editor's note [web] (an interested party: an alternative-energy foundation testing a tube lent by Farnsworth's widow) says the multipactor "holds promise in the 'free energy' field as a truly workable electronic device that may produce over 100% efficiency". It adds that "25 watts out of 25 watts in ... does not take into account radiated heat and possible residual losses", that "Preliminary tests show the tube to be good" and that "further tests are being planned". No published result was found in this search (query 7). That is a gap in what was found, not a negative result. - US 11,588,421 B1 (Robert M. Lyden, 2023, "Receiver device of energy from the earth and its atmosphere") [web]. Lyden, a patentee with a stake, writes: "In U.S. Pat. No. 2,139,813, 'Secondary Emission Electrode,' Farnsworth discloses a multipactor vacuum tube that can operate cold cathode also over unity". Farnsworth's own text in US 2,139,813 [web] says the electrode is "capable of liberating secondary electrons at a ratio greater than unity upon electron bombardment". That "unity" is the secondary-emission ratio delta, not an energy ratio. In the same patent the device runs from "the battery being connected to the anode". Lyden's "over unity" appears to read the yield ratio as an energy gain. This is recorded as his characterization, beside Farnsworth's wording.

Worked example (for the reader)

Symbols: f is the frequency, d the gap, delta the secondary yield (electrons out per electron in), l the electron mean free path, p the gas pressure.

1. Gap needed. First-order multipactor in ECSS Table 5-1 needs f x d of at least about 0.43 GHz mm (aluminium) [web]. - At 10 MHz (0.01 GHz): d >= 0.43/0.01 = 43 mm. - At 1 MHz: 430 mm. - At 100 kHz: 4.3 m [calc]. - Moray's tank frequencies are not known; the review uses 100 kHz-1 MHz as working values (arbiter, lamp-current argument). A two-plate multipactor at those frequencies would need a gap of 0.4-4 m. Farnsworth reached 100 kHz with an anode DC field and magnetic focusing, which change the transit time, so the rule is for the simple two-plate case only.
2. Vacuum needed. The electron mean free path in air is about 0.3 mm at 1 Torr and scales as 1/p [memory, medium]. For l to reach a gap L, p <= 0.3 mm Torr / L. - L = 43 mm: p <= 7 x 10^-3 Torr. - L = 430 mm: p <= 7 x 10^-4 Torr [calc]. - A gas-filled "ionic" tube at glow-discharge pressures (about 0.1-10 Torr [memory, medium]) is collisional, so it would break down as a gas (Townsend or glow) discharge, not as a multipactor (Rasch; ECSS scope) [web].
3. Unequal yields give a DC component, not a one-way block. Take two plates with yields delta1 = 2 and delta2 = 0.5, so delta1 x delta2 = 1, the steady state. - A sheet of charge sigma leaves plate 1 and arrives at plate 2, which releases 0.5 sigma back. Plate 1 then releases sigma again. - Per RF cycle, plate 1 receives 0.5 sigma and emits sigma, so it loses 0.5 sigma of electrons. Plate 2 receives sigma and emits 0.5 sigma, so it gains 0.5 sigma. - A DC electron current of 0.5 sigma per cycle therefore flows from plate 1 to plate 2 inside the tube, and the external circuit must return it [calc, from the sheet model of Kishek 1996]. - This is a rectified current driven by the RF. Electrons still cross in both directions every cycle, so the gap does not block reverse flow the way Q16's quench gap does.
4. Surges. In Kishek 1996 a high-Q cavity loses "as much as 20%" of its stored energy in one multipactor overshoot [read]. Take a 10 nF tank at 1,000 V, which stores 0.5 x 10^-8 x 10^6 = 5 mJ. One overshoot then dumps up to 1 mJ in microseconds [calc]. That is a surge of stored energy, not new energy.

Reasoning: how the evidence bears on each part of the question

• Conduction in unheated tubes. Yes, in principle. Farnsworth's public cold-cathode tubes (1934-36) carried watts to kilowatts by secondary emission. The Patent Office's "not heated" wording is narrow. Its "no space current without plate potential" point is what Farnsworth's tubes obey: they need an anode supply or an RF drive. Moray's tubes, he says, "require no external power sources". In his cascade the earlier tank's RF could serve as the drive, so the energy question passes to the tanks (Claim 3), not to the valve.
• Radioactive coatings as seeds. Yes. This is current practice (ECSS β sources; Sr-90 and Cs-137 named in snippets). The same role exists in gas tubes, where radioactive pre-ionisation shortens the start delay [memory, high: radioactive primers such as Ni-63 or Kr-85 in spark gaps and gas-tube surge arresters]. So the seed role fits both the multipactor reading and the gas-tube reading. It does not decide between them.
• The spill-over threshold. Partly. Multipactor has a lower threshold (E1), an upper cut-off (E2) and a gap-frequency window, and above threshold it clamps the gap voltage by loading the circuit. That is a threshold load. The question is whether it also passes energy onward. It can do that only if a collector (Farnsworth's anode) takes the current into the next circuit. Moray's own words ("ionization occurs in the gases"; "Ionic cold cathode tubes") describe the gas-gap threshold of Q16. Multipactor as a pure vacuum effect would need about 10^-5 hPa and gaps of tens of centimetres at his likely frequencies.
• One-way behaviour from unequal yields. Only as a DC component of an RF-driven current (worked example 3). Single-surface multipactor (Kishek and Lau 1998) is asymmetric too, but it dumps energy on the surface. Neither form blocks reverse energy flow the way Moray's "blocked circuit of no return" describes.
• Surges. Kishek 1996's transient overshoot is a real multipactor surge: a burst far above the steady current, fed by stored cavity energy. It matches the microsecond "surges" in form. In every model it releases energy already stored.
• Net energy. None measured. Multipactor is measured as an RF load. The one oscillator built from it converts its anode supply's power, with the efficiency reported as up to 90%. The two "over unity" readings found are a promise by an interested party with no published result (Borderland 1989) and a patentee's reading of "secondary emission ratio greater than unity" as an energy gain (Lyden 2023).

Competing readings of Moray's valve (the record allows more than one)

1. A hard-vacuum multipactor or secondary-emission tube. - For: "cold cathode"; secondary emission (B.1) is in his own list; the radioactive seed role; the p.42 "vacuum tubes" wording; the threshold and surge behaviour; and Farnsworth was building such tubes in the same years. - Against: Moray's "ionic" and "ionization occurs in the gases" wording; the vacuum and gap needed at his frequencies; and no hard-vacuum pumping, getter or secondary-emission surface (caesium-silver oxide, barium) appears in the review files.
2. A gas-discharge threshold gap (Q16 reading). - For: Moray's own pp.55-56 and p.107 wording; the bucket-brigade firing and extinction; radioactive pre-ionisation known in gas tubes. - Against: gas tubes do not need a radioactive "detector" stone, and the stone's role remains open.
3. A crystal or contact rectifier (arbiter Claim 4 (i)). It fits the receiver use (p.139) but not the cascade.

A mixed reading is also possible: a low-pressure gas tube can show multipactor-like secondary-emission effects at the boundary pressure (l about equal to L). The review files hold no pressure, gap or frequency for any Moray tube, so this cannot be decided from the record.

Open: what would settle it

• A surviving tube (Perreault's, or one in the House of Moray archive): measure the gas fill and pressure (a spark spectrum or residual-gas analysis), the electrode gap and the surface composition (caesium, barium, magnesium oxide, radioactive coating). The arbiter's gamma and alpha spectroscopy test can be run on the same tube.
• Moray's tank frequencies and valve voltages: the notebooks, the p.115 cascade schematic as a page image, and the REA test reports.
• Whether Borderland published a result of its 1989 multipactor test (Journal of Borderland Research 1989-1991 issues).
• A bench test: drive a two-plate copper gap in vacuum at f x d of about 0.6 GHz mm. Compare the forward, reflected and transmitted RF power, the anode DC power and the calorimetric heat, with and without a Sr-90 seed. Energy conservation predicts input = output + heat. The seed should shorten the start time only.

Sources

• [read] aa-new/Kishek_flyer_v01.pdf; aa-new/MULTIPACTOR DISCHARGE ON A DIELECTRIC.pdf (Kishek and Lau 1998); aa-new/Temporal Evolution of Multipactor Discharge.pdf (Kishek, Lau and Gilgenbach 1996).
• [read] Moray, The Sea of Energy in Which the Earth Floats, 4th ed., pp.42-45, 55-56, 107, 155; 5th ed. p.161; arbiter.md (Claim 4, Claim 6); review/Farnsworth-PT/arbiter.md; review/Dollard-Eric/arbiter.md.
• [web] ECSS-E-ST-20-01C (2020): https://ecss.nl/wp-content/uploads/2020/07/ECSS-E-ST-20-01C(15June2020).pdf
• [web] Halloran, RADIO, Oct 1934 (Borderland reprint, 1989 note): https://borderlandsciences.org/journal/vol/44/n05/Farnsworth_Multipactor_Tube.html
• [web] Wireless World, 3 Apr 1936, "New Electron Multipliers": https://www.r-type.org/articles/art-480.htm
• [web] Wikipedia, Multipactor effect: https://en.wikipedia.org/wiki/Multipactor_effect
• [web] Langellotti, PhD thesis, U. Michigan 2022: https://plasmabay.engin.umich.edu/wp-content/uploads/sites/281/2023/01/2022-Stephen-Langellotti.pdf
• [web] Rasch, Chalmers: https://publications.lib.chalmers.se/records/fulltext/164741/164741.pdf
• [web] US 2,139,813: https://patents.google.com/patent/US2139813A/en ; US 2,184,910: https://patents.google.com/patent/US2184910A/en ; US 11,588,421 B1: https://patents.google.com/patent/US11588421B1/en
• [web, snippets only] Aerospace Corp TOR-2014-02198; OSTI 2022 (Luhmann); UAM doctoral thesis on anti-multipactor coatings.
• Saved pages (not published): research/pages/ecss-e-st-20-01c.pdf, ecss.txt, wiki-multipactor.txt, farnsworth-halloran-1934., rtype-art480., langellotti-2022., rasch-chalmers., US2139813A., US2184910A., US11588421B1.*.

Corrections to earlier documents

None found wrong. One sharpening for the orchestrator to consider: the arbiter (Claim 6, "Patent grounds") calls the examiner's cathode ground "partly weak" because cold-cathode tubes were known [memory, high]. That point can now carry a [web] source: Farnsworth's cold-cathode multipactor tubes were public in 1934 (Halloran, RADIO, Oct 1934) and 1936 (Wireless World, 3 Apr 1936). Both dates are after the Dec 1932 action, so they show the physics, not what the examiner could have known.