Correa · Source catalogue

Patent for a cold-cathode vacuum tube with large plates (at least 16 cm2 preferred) and a large gap (at least 2 cm, preferably 3.5 cm), fed from a DC source with enough series impedance to prevent a sustained vacuum arc, which self-pulses in a pulsed abnormal glow discharge (PAGD) regime. Pages 2-15 are drawings (FIG. 1-21B); the specification starts p.16. Background (pp.16-18): normal glow, abnormal glow and vacuum arc regimes; the only prior abnormal-glow art is Manuel US 3,471,316 (1969), externally pulsed, for coating cans. Examples (pp.23-31): Argon pumpdown tests show PAGD moving to higher pressure with more applied current and to lower pressure with more voltage; with an oil diffusion pump PAGD persists to about 5x10^-7 Torr extinction, and a 128 cm2 tube at 2x10^-6 Torr and 2.2 kV runs PAGD at 30 pps; below 1e-6 Torr and at higher voltage PAGD gives way to cathode rays and weak x-rays (p.23). Example 4 (pp.24-25, Tables 2-5) compares 16, 64 and 128 cm2 plates: bigger plates break down at higher pressure (128 cm2 at 3 Torr, 8x higher PAGD rate than 16 cm2 at 24 times lower pressure) and at lower voltage, which the authors say Paschen's law cannot explain, and fit area factors k_a (17 for a 10x pressure drop, 21.5 for a 50% voltage drop). Examples 5-6 (pp.26-28): sealed glass tubes at 2x10^-6 Torr (seal-off) with 5 cm gap; 16 and 64 cm2 tubes unresponsive (64 cm2 only a faint glow at 3.3 kV), a 256 cm2 tube fires at a much lower breakdown voltage; Alzak plates give 5x higher DC current and 20-30x higher pps than hardened aluminum; capacitance of the supply in parallel changes the rate; tube static capacitances 1, 2.05 and 4.1 pF for 64, 128, 256 cm2. A closed high-vacuum 128 cm2 tube at 668 V, 500 mA mean current and 200 pps gives pulses above 300 V. Cathode study (pp.29-31): aluminum, zinc, nickel, soft iron and silver work best; copper, brass and steel are poor; tungsten works at 750 V. Later pages (32-36): emission-site current densities of 1-2.5x10^9 A/m2 after a 75x correction to videographic areas; crater volumes (Alzak 6.2e-7 cm3, H34 3.1e-8 cm3); breakdown fields 2,730-30,000 V/m against more than 2x10^9 V/m for field emission, so a field-enhancement factor of 10^6-10^7 is needed; input pulse energies 1 mJ to above 112 J; Table 14 reactor and output energies of under 3 to above 50 times the input; peak reactor currents above 150 A from 1.2-1.7 A input; the worked reaction-force example (86.4 J net, 3.1e-8 cm3, 143.6 dyn calculated, 245.2 dyn balance). Cathode lifetimes (H34 1.2x10^9 total pulses; the digest's 10^8 is an OCR exponent slip) and Table 17 specifications (gas 10^-6-10^-3 Torr, PAGD rate 0.01 to above 50,000 pps). Claims pp.37-38. QE round 2 deep-read corrections (from the compilation text): (i) Vacuum was gauged: the Edwards rig had thermocouple and Penning gauges (rotary and diffusion vacua), the tubes 'measured 10^-6 Torr at time of seal off' and Table 6 col. G is 'the vacuum measured when the respective housings were sealed'; what is not known is the pressure inside the sealed tube after seal-off or running, and its gas (residual argon, Silicone 705 oil vapour, desorbed gas; no getter or residual-gas analyser is described). (ii) 'Input energy per pulse' in Tables 13-14 is supply voltage x DC current x pulse conduction time (every Table 13 row reproduces); it omits the supply energy delivered between pulses. At the worked example's 560 V, 1.6 A, 0.3 PPS the per-period input would be about 2,987 J against 253.5 J booked output, if the current was continuous; the largest output/input ratios (28-60x) occur at 0.2 PPS, as that omission predicts. The integration window and integrand for the output energies are never stated. (iii) The 'reactor energy' and the 0.2-80 mF 'dynamic capacitance' are consistent with 1/2 C V^2 of the 17.5 mF external C3/C5 bank at Vp of about 90-205 V, not with a tube property. (iv) The worked example: 86.4 J cannot be traced to Table 14 (nearest independent figure 90 J/pulse from '0.25 kWh per 10,000 pulses'); the authors' own inputs give 4.2x10^5 eV per ion, not the printed 4.8x10^5 (which matches 100 J, not 86.4 J); the 143.6 dyn reproduces only if the mass is divided by the 50 ms end of '40 to 50 msec'; 1.86 g/cm3 is not aluminium's density (2.70); the crater volume comes from a different run. (v) Table 15 normalises force by the 1.6 A DC input current; the same text reports peak reactor currents above 150 A, and k falls to the vacuum-arc value (about 100) at a pulse current of only 12-16 A, so the '10-100x vacuum arcs' claim holds only if the pulse current stayed below that. The 245.2 dyn 'reaction balance' reading has no apparatus, calibration, error, trial count or controls, and peak versus time-average (a factor of about 70 at 0.3 PPS) is not stated. (vi) Plate-area data (Tables 2-6): the area dependence appears in two datasets (pumpdown in argon, n of about 0.11; sealed tubes at 2x10^-6 Torr, n of about 1.5-2.2) and contradicts plane-parallel Paschen scaling as the authors apply it, but 4 cm plates at a 5-5.5 cm gap (edge and wall fields), the known area effect in vacuum breakdown, and (at 2x10^-6 Torr, pd of about 10^-5 Torr cm, about 4x10^-5 ionisations per electron crossing) the fact that gas avalanche is not the route all remain unaddressed; the sealed series is one tube at 256 cm2 (Vb 287 V, Vx 284 V, no hysteresis) with different gaps and currents. Example 1's Fowler-Nordheim 'positive slope' is an assertion from one graph with no fit or field calibration; the mean field is 1.7x10^4 V/m, so beta of about 2x10^5 would be needed.

Key claims

• A cold-cathode tube with large plates and a gap of at least 2 cm, fed through enough impedance to avoid a sustained arc, spontaneously self-pulses (PAGD) with emission-site current densities 10-1000 times lower than Fowler-Nordheim theory requires.
• At fixed pressure, gap and current, PAGD breakdown voltage falls as plate area rises, and the effect contradicts Paschen's law and space-charge scaling.
• PAGD occurs in tubes sealed at 2x10^-6 Torr (and an example at 10^-6 Torr) at 128-256 cm2 plate area.
• The breakdown field is only 2,730-30,000 V/m, so the field-enhancement factor needed is 10^6-10^7.
• Per-pulse reactor energy is under 3 to above 50 times the input pulse energy, with peak currents above 150 A from 1.2-1.7 A input (Table 14).
• Mean net pulse output of 86.4 J ejecting 3.1e-8 cm3 of aluminum gives 4.8e5 eV per ion as printed (4.2e5 eV from the authors' own inputs; 86.4 J is not traceable to Table 14) and a reaction force of 143.6 dyn (245.2 dyn by balance).
• 'Input energy per pulse' (Tables 13-14) is V x I_DC x pulse conduction time and omits supply energy delivered between pulses; at 560 V, 1.6 A, 0.3 PPS the per-period input would be about 2,987 J against 253.5 J output.
• Reaction-force constant k is about 10^4 only when normalised by the 1.6 A DC input current; at the reported >150 A peak pulse current k is about 1, below the vacuum-arc value of about 100.

Relevance: Describes the reactor itself and carries the area-effect and high-vacuum emission claims. Holds the only data on tubes at 10^-6 Torr, which a gas-breakdown (Paschen) argument does not cover.

Mainstream conflict: Claimed failure of Fowler-Nordheim scaling and of Paschen's law. Mainstream view: field enhancement at microprotrusions and adsorbates, vacuum breakdown triggered by microparticles and desorbed gas, edge and wall fields of small plates at a gap wider than the plate, a known electrode-area effect in vacuum breakdown, and relaxation oscillation of a discharge with negative differential resistance. At 10^-6 Torr Paschen scaling does not apply at all. Pressure: the 10^-6 / 2x10^-6 Torr figures are Penning-gauge readings on the pumping system at seal-off (factor-of-two uncertainty); the pressure and gas inside the sealed tube after seal-off or running were not measured (QE round 2 deep read; this replaces the earlier 'not gauge-measured' wording).

Pages 1-14 are the front page and 13 drawing sheets; text starts p.15. The device is a PAGD tube in a double-ported circuit: a DC drive pack (DP, about 46 lead-acid 12 V cells, about 560 V) feeds the tube through diode D1, ballast resistor R1 and returns through D4; the output side is a capacitor/bridge network (C3, C5, bridge D2/D3/D5/D6, C7a/C7b, D7/D8) that charges a charge pack (CP, 29 cells) at E1-E2. Deep read of FIG. 8-11: the drive loop is one series loop and the CP is coupled to the tube only through C3 and C5 (or C4 in the motor variant, FIG. 11); there is no closed DC path between the packs, and the authors say the design is for DC isolation of the two packs. Four measurement methods are used: (1) open-circuit-voltage mapping (Table 5), (2) time-integrated resistive measurements (Table 6), (3) direct DC V and I means (Table 7), (4) battery discharge curves before and after (Table 8), plus an operational video check (FIG. 20, 485% +/-18%). The authors admit the methods differ ten-fold. Deep-read findings: Table 7 col. 8 'DP Watts' equals (X - Y) x I, i.e. the dissipation in the ballast and input diodes; the DP delivers X x I, so on terminal power 10 of 14 runs are at or below 100% (57-146%, median about 92%) against the printed 218-563%. Table 7 has misprints in rows 3, 9, 10 and 14 (e.g. row 14 CP W prints 257.7, V x A = 357.7). Table 8 is not credible as an energy balance: its drive-pack losses (5.6-89 W average) are below the 98 W floor set by the authors' own 0.2 A minimum current, and the headline 3,485% does not reproduce from its own columns (3,245% or 3,423%). Table 5 run 5's 6,750% is an artefact (loss rate mis-divided, 0.016 for 0.034 kWh/h, giving about 3,200%; the drive pack is nearly empty by the method's own scale, 3.2%, capping the booked loss). Table 5 runs 1 and 6 imply 2.8 kW into and 4.3 kW out of the packs, impossible through a 300-ohm ballast. Other points: the output battery sets the oscillation (29 to 31 cells cuts PAGD rate ten-fold, more extinguishes it); the AC rms of each pulse can be up to 7.5 times the applied DC current at about 500 mA; removing the C7 bridge suppresses PAGD; the authors state the output energy 'is not directly related' to that used to trigger the pulses; current meters were Beckman rms multimeters in an unstated mode.

Key claims

• A PAGD tube in a double-ported circuit captures at the output a substantial proportion of the energy released by collapsing emission pulses (abstract).
• Breakeven efficiency by method: Table 5 96-906% plus a 6,750% run, Table 6 504-776%, Table 7 218-563%, Table 8 406-3,485%, FIG. 20 485% +/-18%.
• Table 7 'DP Watts' is (X - Y) x I, the ballast and diode dissipation, not the drive pack's output; on X x I the ratio is 57-146% (median about 92%).
• Table 8 drive-pack losses are below the minimum implied by the stated 0.2-0.6 A current range, and its 3,485% does not reproduce.
• The PAGD frequency depends on the charge pack's cell count and state of charge, so the 'output' is part of the oscillator.
• A motor on the auxiliary arm runs at 2800-3000 rpm for 10 pps (Table 7 run 12, 'electrical flywheel'). This figure appears only in this patent; no US 5,416,391 example gives a matching number, and by the synchronous formula 10 PPS would drive a 2-pole motor at 600 rpm.

Relevance: Core patent for the over-unity claim and the source of Tables 5-9, the only energy-balance data. The deep read settles the Table 7 input-power question.

Mainstream conflict: Output above input conflicts with energy conservation unless an unmetered store is found; the deep read finds the printed input omits most of the drive pack's output (Table 7), the Table 8 losses are below physical floors, and Table 5-6 power levels are impossible, so mainstream accounting (battery relaxation, surface charge, rms-meter error on pulsed current) explains the data. Patent still cites Aspden's non-mainstream electrodynamics.

Printout of rexresearch.com/correa/correa.htm that bundles: press release (p.1), Axelrad essay (pp.2-3), excerpt from Labofex report S3-001 (pp.3-5), US 5,416,391 (pp.5-15), US 5,449,989 (pp.15-27) and US 5,502,354 (pp.28-43), then an order page (p.44). Each is described in its own entry. All measurements are by the inventors; no independent test appears in it. QE round 2 deep reads now cover the motor examples, the vacuum processing and area-effect tables, and the per-pulse energy and reaction-force tables; their findings are recorded in the three bundled entries and in the scan entry for US 5,502,354.

Key claims

• See the entries for the bundled documents.

Relevance: Only readable-text copy of all three patents together with the inventors' framing; carries Tables 5-9 and 12-17.

Mainstream conflict: See the bundled entries.

Announces the three US patents (5,416,391; 5,449,989; 5,502,354) and an energy conversion system using cold-cathode vacuum-discharge plasma reactors in PAGD that is said to produce electrical power in excess of input ('over-unity'). Says there is no fusion and no nuclear radiation. Proposes electric vehicles, stand-alone power, housing, pulsed lasers, inverters and motor circuits; licensing agreements are said to be under negotiation.

Key claims

• PAGD reactors produce electrical power exceeding input, directly as electricity.

Relevance: Shows how the claim was publicised.

Mainstream conflict: Net energy gain conflicts with energy conservation absent a source.

Axelrad, Correa's former graduate supervisor at the University of Toronto and a biomedical scientist (polycythemia vera research), says he is not trained in plasma physics. He describes seeing a demonstration of an 'Aether Motor' said to run from orgone accumulator boxes, insulated bodies or a ground pipe, says the Correas ran Labofex for 15 years without grant funding and were shutting the lab for lack of funds, and cites endorsements from H. Aspden and E. Mallove. The same pages note that many demonstrations to backers failed to produce funding.

Key claims

• An 'Aether Motor' delivers electrical power with no external input other than orgone accumulators, insulated bodies or ground.

Relevance: Only source for the Aether Motor claim; it is a lay testimonial, not data.

Mainstream conflict: Conflicts with energy conservation; orgone is not accepted physics.

Literature survey and theory. Covers Tanberg and Kobel (1930) anomalous cathode reaction forces, Reich's claimed 30 kc oscillation in high-vacuum diodes and orgone motor, Kapitza's plasma balls, Kimblin's cathode-spot measurements, Compton's (under 2% vapour-jet result), Aspden's 1969 Law of Electrodynamics F = (qq'/r^3)[(v'.r)v - (m'/m)(v.r)v' - (v.v')r], Graneau's water-arc explosions, Chernetskii's 1989 Novosti-announced reactor, Shoulders' EVs (10^8-10^11 electrons per event), T.H. Moray, Manuel's 1969 patent, and G. Spence's patent. States that PAGD use closed systems where energy deployed in a second arm consistently exceeds that lost in the first, attributes it to vacuum (zero-point) polarisation, and says Fowler-Nordheim cannot explain arc initiation at the observed fields. Mentions IVAD (interrupted vacuum arc) reactors and a second laboratory, ABRI (Aurora Biophysics Research Institute).

Key claims

• Closed PAGD systems deliver in the second arm substantially more energy than is spent in the first arm.
• Surplus energy comes from vacuum (zero-point) energy via Aspden's law.
• The anomalous cathode forces obey Aspden's law rather than the Lorentz force law.

Relevance: Gives the inventors' theoretical framing and historical lineage.

Mainstream conflict: Aspden's law and the vacuum-energy source conflict with the Lorentz force law, Newton's third law and conservation of energy; Reich and orgone claims are outside accepted physics.

A DC power transducer: a PAGD tube across a current source, capacitively coupled to an AC induction, synchronous or hysteresis motor, with speed set by pulse frequency (source current, parallel capacitance, auxiliary electrodes). Pp.11-13 repeat the discharge regimes, give five motor examples (the text states almost no rpm, torque or power numbers; every result is in FIGS. 7-13, which are absent from the printout) (a single-phase squirrel-cage motor with a 64 cm2 H34 tube at 2 Torr; a 1/10 hp hysteresis motor; torque measured by rope and pulley) and inverter circuits (FIG. 14-15), and tube construction (Pyrex housing, 16-256 cm2 plates, 3-20 cm gap). Examples 6-8 on pp.13-14 repeat the high-vacuum and argon pumpdown tests of US 5,502,354. Deep-read detail: all examples use the FIG. 3 circuit (capacitors 10 and 11 in series with the motor, triode tube). Ex. 1: squirrel-cage motor, 2 uF auxiliary, 64 cm2 plates, 5.5 cm, 2 Torr air, rpm by stroboscope against series capacitance. Ex. 2: 1/10 hp hysteresis motor, 2.4 uF. Ex. 3: 36.6 uF parallel capacitance, extinction voltage about 330 VDC, standstill (rotor-blocked) torque by rope-and-pulley meter at 11-45 PPS, so no mechanical power is measured (torque x speed = 0). Ex. 4: 1.75 Torr, 440 uF and 8 uF in series = 7.9 uF (checked), DC amps, volts, wattage input, AC rms current and voltage, true and apparent power and rpm all measured against pulse rate to about 40 PPS, but only in FIGS. 11-12. Ex. 5: V per PPS plateaus near 25 PPS. Words only: triode raises the maximum pulse rate from 30 to 43-45 PPS; rpm of the hysteresis motor is proportional to pulse rate; the patent concedes the pulses are poorly suited to transformers. The patent makes no efficiency or over-unity claim for the motor circuit, only controllability. The 2800-3000 rpm at 10 PPS figure is NOT in this patent; it is a US 5,449,989 Table 7 run 12 remark, and no example here supports or contradicts it (a 2-pole synchronous motor at 10 PPS would run at 600 rpm; about 5 cycles per pulse of ring-down would be needed, an unshown hypothesis). Stored energy 1/2 C V^2 of the 36.6 uF at 330 V is about 2 J (upper estimate for pulse energy; not in the patent).

Key claims

• A self-pulsing tube across a DC source can drive an AC induction or synchronous motor through capacitive coupling.
• Motor speed is controlled by pulse frequency.
• Standstill torque of the hysteresis motor is proportional to pulse frequency and to input voltage (Example 3, FIGS. 9-10, no values in the text); no running torque or power is reported.

Relevance: Shows the PAGD tube as a DC-to-AC inverter, the least controversial application.

Mainstream conflict: None for the inverter function itself (a relaxation oscillator driving a load).

Same patent as the scan, here with a clean text layer including Tables 1-9 (Table 5 pp.22-24, Table 6 pp.24-25, Table 7 pp.25-26, Table 8 pp.26-27, Table 9 p.27). Deep-read findings: Table 8 'does not survive' as an energy balance (drive-pack losses 5.6-89 W average, below the 98 W floor; 3,485% not reproducible; each Wh figure is an extrapolation to a 115 W (DP) or 90 W (CP) cutoff from discharges starting only 15-30 minutes after loading, and a 0.4 W shift in starting power equals a whole booked loss); Table 5 run 5 (6,750%, motor on, 1,200 rpm) is an arithmetic and method artefact (about 3,200% corrected, with a nearly empty drive pack); Table 5 run 6 books 4.33 kW out of the DP, 3.7x the 1.16 kW maximum through a 300-ohm ballast; Table 6 run 1 implies 3.46 kW into the CP (about 10 A) against Table 7 maximum 1.36 A; Table 9 (H200 plates, DP 46 / CP 23 cells) gives net 17.55-35.2 mWh/s; lifetime projections of 20 MWh per H34 cathode rest on 0.25 kWh per 10,000 pulses. Batteries: 46x12 V 6 Ah gel cells (DP) and 28-29 cells (CP); measured half-charge capacities 1.033 kWh (DP, 100% = 2.066 kWh) and 660 Wh (CP, 100% = 1.320 kWh), about 50-60% of the manufacturer's theoretical figure.

Key claims

• Breakeven efficiency ranges by method: Table 5 96-906% (and 6,750%), Table 6 504-776%, Table 7 218-563%, Table 8 406-3,485%.
• The authors accept ten-fold variation between methods but say all indicate an anomalous transduction.
• Table 7 and Table 8 inputs are understated (deep read).

Relevance: The only quantitative energy-balance evidence for the over-unity claim.

Mainstream conflict: As in the scan entry.

Same patent as the scan, with a clean text layer. Deep-read content (QE round 2). Vacuum processing (p.30): Edwards oil diffusion/rotary pumps (Silicone 705 oil), thermocouple and Penning gauges to 1e-7 Torr, UHP argon cycles, 25 kV glow cleaning, 400 C bakes, x-ray 'hard vacuum' criterion with a Geiger counter, Tesla-coil 'black vacuum' test, seal-off 'never above 10^-6 Torr'; no getter or residual-gas analyser. The authors themselves raise oil-vapour molecules as a possible gas in diffusion-pumped tubes (p.31). Example 1: device 1 (H34, 128 cm2, 5 cm, 10^-6 Torr seal-off) shows PAGD from about 10 to 150 mA RMS with Vb plateauing near 850 V and a 'positive slope' they call evidence of auto-electronic emission (mean field 1.7e4 V/m; beta of about 2e5 would be needed); negative polarity gives 2 A against 200 mA with a 10x supply. Example 2: Alzak gives about 9x the pulse count of H34 and pulse rate is limited by the supply's recovery time (relaxation oscillator). Example 4 (Tables 2-5): 16/64/128 cm2, area dependence of pressure (prf 0.52, 0.44, 0.24) and voltage (91.7, 85.9, 78.7% of the 16 cm2 value); k_A of 17 and 21.5 are linear extrapolations; the Table 4 sample-7 row has the opposite sign and several typos. Example 5 (sealed, 2x10^-6 Torr): 16 and 64 cm2 unresponsive (>3.5 kV); 128 cm2 at 774 V; 256 cm2 (one tube) Vb 287 V, Vx 284 V; the 'fell by 1/3' is a factor of 2.76. Example 6: tube capacitances 1, 2.05, 4.1 pF (eps0 A/d gives 1.13 and 2.27 pF for the first two), and device 1's Vb is 850 V or 668 V depending on supply and current. Pp.39-40: Table 13 input energy = V x I x t_pulse for every row (omits inter-pulse supply energy); Table 14 output/input 1.5-60.5x with the largest ratios at 0.2 PPS, reactor energy ND in 8 of 14 rows and consistent with 1/2 C V^2 of the 17.5 mF external bank at Vp about 90-205 V (so the 0.2-80 mF 'dynamic capacitance' is the bank); the integration window is never stated; 'Table 6' in the text means Table 14. Worked example: crater volume, mass, ion count and 143.6 dyn reproduce (the last only with a 50 ms divisor); the energy per ion is 4.2e5 eV, not 4.8e5; 86.4 J is not traceable to Table 14 (nearest 90 J from '0.25 kWh per 10,000 pulses'); 245.2 dyn from an undescribed balance, peak versus average unstated. Table 15 normalises with the 1.6 A DC input current; at >150 A peak k is about 1, and k equals the vacuum-arc value at 12-16 A. Table 12: row 5 prints 12,200 V/m for 11,200; Table 16's 'work function' is X converted by X = 0.44 phi - 0.15, not a measurement, and the text mixes up X and phi for Li, Be, Mg, Cs. Pp.36-39 cathode materials and conclusions and Table 17 were read as digest only. Unlike the 5,449,989 tables, these are not energy-balance data from batteries, but the per-pulse energies are an energy-balance claim and rest on the omissions above.

Key claims

• Plate-area effect contradicts Paschen scaling (Tables 2-6); the area dependence is in the authors' own data, but edge/wall geometry and the vacuum-breakdown area effect are not excluded, and at 2x10^-6 Torr Paschen scaling is outside its domain.
• PAGD at 10^-6 to 2x10^-6 Torr in sealed tubes (Penning-gauged on the pumping system at seal-off; pressure and gas inside the sealed tube not re-measured).
• Reaction force 143.6 dyn computed, 245.2 dyn from an undescribed reaction balance, with 86.4 J net pulse energy (output minus input, with input counted as V x I x t_pulse; not traceable to Table 14); ion energy 4.8e5 eV as printed, 4.2e5 eV from the authors' inputs.
• Per-pulse output of 1.5-60x the input (Table 14); deep read finds input omits inter-pulse supply energy and the 'reactor energy' matches the external capacitor bank.
• Reaction force is 10-100x vacuum arcs (k of 10^3-10^4 against 10^2); this holds only if the pulse current stayed below about 12-16 A, whereas >150 A peaks are reported.

Relevance: Carries the area-effect, high-vacuum, cathode-material and reaction-force data.

Mainstream conflict: Claimed failure of Fowler-Nordheim scaling and of Paschen's law. Mainstream view: field enhancement at microprotrusions and adsorbates, vacuum breakdown triggered by microparticles and desorbed gas, edge and wall fields of small plates at a gap wider than the plate, a known electrode-area effect in vacuum breakdown, and relaxation oscillation of a discharge with negative differential resistance. At 10^-6 Torr Paschen scaling does not apply at all. Pressure: the 10^-6 / 2x10^-6 Torr figures are Penning-gauge readings on the pumping system at seal-off (factor-of-two uncertainty); the pressure and gas inside the sealed tube after seal-off or running were not measured (QE round 2 deep read; this replaces the earlier 'not gauge-measured' wording).

Order page for a 'civilization kit'; no technical content. The reference list (about 108 references) cited in the earlier Step 1 was not located in the digests and is unverified.

Relevance: None.

Mainstream conflict: none