Arbiter: Radiant_One_Youtube (Step 3)

Inputs weighed: facts.json, claims.md, physicist.md, mathematician.md, contrarian.md, sites.md and research/ notes (no deep/ reads exist; research/files/ is empty). I also viewed the PNG myself and zoomed on the diodes [read: page image p.1].

Source gap (repeated from claims.md)

The project holds one image: a schematic frame, "Ground Energy Extraction (GndNrgEx) Circuit 1.1 (Corrected)", channel radiant_one [read: page image p.1]. There is no video, transcript, measurement or author identity in the folder. Every measurement quoted below (30 mA, 4 mW per LED, transistor temperatures) comes from the Investigator's reading of the author's YouTube descriptions and auto-generated captions [web: research/2026-10-04-youtube.md]. None of the three reviewers or I re-watched the videos. Conclusions beyond circuit analysis therefore rest on that second-hand web record, plus [memory] where tagged.

Corrections to the reviews and to Step 1

1. Diode orientation (all three reviewers and facts.json are wrong or imprecise). Zoomed on the PNG, both 1N4148 symbols across the LED group point the same way: anode on the L2 side, cathode on the earth-rod side [read: page image p.1, crops of both diodes]. The physicist and mathematician said the two diodes are in opposite orientations. They are not: they are two diodes in parallel, one drawn above and one below the group, both conducting current from the L2 node toward the earth rod. Together they act as one clamp spanning the whole two-string group, not one per string, and not an anti-parallel pair of each other. "Anti-parallel" is true only relative to the LEDs, whose orientation is not drawn: the LED strings are shown as labelled lines ("26 White LED's in series"). The circuit can only light the LEDs if they are oriented to conduct from the earth rod toward L2, opposite to the diodes. If they were drawn the other way the diodes would clamp the array at about 0.7 V and nothing would light. This matters for the mechanism (below) and the reader should know the schematic leaves the key polarity implicit.
2. Orange bars. Two orange vertical blocks above and below the L2 line, cutting the title text ("Ex[]action") [read: page image p.1]. The contrarian's description is right; facts.json's "coupling between L1 and L2" is unsupported. They are probably a video overlay; no electrical meaning is assigned. L2 is drawn in series between the collector node and the LED group; there is no coupling marking to L1.
3. Claim 3 wording. Equal current is the tuning target (the clip is moved until unearthed current equals earthed current), not an independent finding. All three agree [read: tuning note on the PNG; web: GufTw6ECoz4 transcript via Investigator].
4. Claim 1 wording. The author says in most descriptions "No overunity performance is suggested or implied" and also frames the work as Moray-style ground-energy extraction and says "A small energy input from the driver produces a much larger response from the ground electrons" (GndNrgEx_10) [web, via Investigator]. Both statements are kept; they are in tension (the mathematician calls this internal inconsistency; see claim 1).
5. Reviewer disagreements on what the measurements are. No reviewer's verdict depends on a number the author did not state.

Numbers recomputed (Python, my own run)

The physicist did not draw out the heat-test magnitude; the mathematician did. I confirm it. A case rise of 1.2 K implies only about 6 mW in the transistor (up to a few times that allowing for case-to-ambient being lower than junction-to-ambient). The supply delivers 360 mW. So either the transistor is about 98 % efficient (unusual for a single-transistor oscillator [memory, medium]), or the 30 mA reading, the IR reading, or both are unreliable (the author's own warning and the IR thermometer's -50 reading in the field), or the loss is in parts he did not measure (coils, diodes, LEDs, wiring). The heat test therefore cannot confirm an energy balance in either direction. The 0.5 K difference between states is the right sign for diverting power to the LEDs but corresponds to about 2.5 mW, not 208 mW.

Web checks of [memory] facts the verdicts depend on

• Fair-weather atmospheric field about 100 V/m and air-earth current density about 2 pA/m^2 (2.3 over continents) [web: https://space.fmi.fi/~makelaa/fairw.html and search summaries of NOAA/AFRL handbook chapter 20, https://www.ngdc.noaa.gov/stp/space-weather/online-publications/miscellaneous/afrl_publications/handbook_1985/Chptr20.pdf]. Upgrades the physicist's and contrarian's [memory] to [web]. Consequence: collecting even a 100 m^2 area at 2 pA/m^2 gives 0.2 nA; a rod gives far less, many orders of magnitude below 208 mW. (Telluric currents are a separate quantity; I did not web-check them, so the "volts per km" figure in the mathematician's review stays [memory, medium].)
• MPSA06: Vceo 80 V, 200 C/W junction-to-ambient, 625 mW [web: https://www.mccsemi.com/pdf/Products/MPSA06(TO-92).pdf and other datasheets]. Confirmed.
• White LED forward voltage 2.8-3.6 V, commonly 3.0-3.4 V at 20 mA [web: https://circflow.com/blog/led-forward-voltage-chart/]. Confirmed in range; the 2.9 V used is slightly low-typical, which does not affect any verdict.
• Not web-checked (no verdict depends on them): single-transistor oscillator efficiency 60-85 % [memory, medium]; radiation resistance of a small loop or coil at MHz being tiny [memory, medium]. Both are used only to say where the unmeasured power most plausibly goes. I flag them as [memory, medium]; the verdicts below do not rest on them.
• Moray's record [memory, high that he claimed a "radiant energy" device and wrote "The Sea of Energy"] was not examined by anyone here; not weighed.

Corrected claim list and verdicts

Claim 1. A "Ground Energy Extraction" circuit can draw usable electrical energy from the earth through an earth rod.

Where reviewers agree: the author's own numbers (0.208 W out, 0.36 W in) show no surplus; no neutral controlled test exists; no independent replication was found; the earth rod can serve as a return path; natural earth-field sources are far too small. Where they differ: (a) the physicist and mathematician say the strong form is "Never properly tested" and give the weak form as Established (physicist) or Plausible, untested (mathematician, contrarian). (b) The mathematician adds that the author's "much larger response" is Internally inconsistent with his "no overunity" statement.

Weighing. The weak-form dispute is about whether to call it Established. No one has measured it on this circuit: no floating-plate control, no capacitance or frequency measurement. The mechanism (series resonance with L2 and stray capacitance, rectification by the LEDs and the 1N4148 pair, displacement current closing through the rod) is standard physics, but its application here is an inference, and the frequency check above shows the 25 pF figure depends on an unmeasured 6.2 MHz. I side with Plausible, untested for the weak form on this specific circuit; the general principle (earth/counterpoise as an RF return path) is Established.

On the inconsistency: "a much larger response" than the driver input is not necessarily an energy statement; it could describe the visible LED response relative to a small change in driver behaviour. But taken literally it contradicts "no overunity". I do not assign "Internally inconsistent" as a verdict because the author's measured numbers (58 %) sit on the no-overunity side and the phrase is a loose description, not a derivation. The tension is recorded.

Dismissed or untested? Not dismissed by anyone; not tested by anyone independent. It sits on a small channel (469 subscribers, 14,645 views on the schematic video) [web via Investigator]. The one question the claim raises has a cheap, clean test (below).

Final verdict (split). - Strong form (net usable power from the ground beyond supply input): Never properly tested. Not "Refuted by calculation": the true RF input and output powers were never measured, and the author's 4 mW per LED and 30 mA are RF-contaminated, method-unstated, single-source numbers from different videos. [Revised, see QE revision: the original "at face value they show 58 %" is withdrawn; the record holds no measured LED output power, so the author's figures neither show nor exclude a surplus.] No known ambient source approaches the size needed (about 2 pA/m^2 [web]). - Weak form (earth rod completes an RF return path so the LEDs light with no rise in supply draw): Plausible, untested on this circuit; mechanism Established in general.

Claim 2. MPSA06 oscillator (L1 8.6 uH, Lb 1.4 uH, C1 470 pF, 1 Meg, 11-13 V) plus L2 26.7 uH can drive an earth-rod-connected load of two 26-LED strings with a 1N4148 clamp.

All three agree: Plausible, untested (physicist and mathematician say the base design is widely built; contrarian same). The lit LEDs are shown on video by the author and are not in dispute; no independent build of this exact LED array was found. Correction: the clamp is two parallel same-direction diodes across the group, and the LEDs must be oriented against them (see Corrections). The base circuit lineage (Stiffler SEC-15, Slayer exciter; 12 V, 30-75 mA builds 2011-12) is documented on the Energetic Forum [web: https://www.energeticforum.com/forum/energetic-forum-discussion/renewable-energy/5174-joulethief-sec-exciter-and-variants/page139.html], but Stiffler's own documents were not read.

Final verdict: Plausible, untested.

Claim 3. Supply current with the earth rod connected can be tuned (alligator clip near L2) to equal the current without it.

Disagreement: the physicist says Established as a tuning result; the mathematician and contrarian say Plausible, untested. Reasoning: nobody independent has reproduced it, and the only evidence is the author's report of 30 mA in both states (single source, with the author's own warning that meters can misread). That it can be done in principle is trivial (continuous adjustment). Because the physicist's verdict rests on theory, not on a test, and the observed equality is one source, I use Plausible, untested as to the observation. As evidence for ground energy it is Not testable as stated: equal input power in two states says nothing about whether a further source feeds the LEDs; the equality is set by the procedure. Where does the 0.36 W go when unearthed? Unmeasured; the heat test accounts for about 6 mW at the transistor.

Final verdict (split): observation Plausible, untested (reported by one source, not contradicted); interpretation as evidence of ground energy Not testable as stated.

Claim 4. The circuit produces enough RF that meters and oscilloscopes can be damaged or read incorrectly.

All three: Established as a general property of RF circuits. The severity here is the author's report (IR thermometer read -50; personal meter damage) [web via Investigator]. Consequence they all draw: his own 30 mA, 4 mW and temperature figures carry unquantified error.

Final verdict: Established (general); the specific severity is the author's report.

Added item (not in claims.md): the author's mechanism statements

"Ground energy stops at a 1N4148", "asymmetric pathways", "ground electrons", "magnetic-dielectric interaction" of the fibreglass table (GndNrgEx_9, _10) [web via Investigator]. The 1N4148 observation is reported by one source and has a mainstream reading (a series diode in a capacitive return path leaves no way to close the net current). The "ground electrons" and "magnetic-dielectric" wording is Not testable as stated (no definition, units or numbers). The "One Wire Energy" and "Full Metal Jacket" videos (earthed receiver brighter; driver in a metal can still lights LEDs through one wire) are single-source and compatible with ordinary displacement-current return; they do not separate "radiant energy" or longitudinal waves from near-field coupling. Untested by any proper return measurement.

Where the reviewers' arguments were stronger

• The physicist gave the clearest mainstream mechanism and the useful floating-plate control. The mathematician added the heat-test magnitude check, the unit and notation points, and the 4 mW method gap. The contrarian gave the fairest account of the author's good practice (battery supply, published values and coil data, heat test attempt, repeated no-overunity disclaimer, metal-can control) and correctly said that the experiment as run cannot distinguish "ground supplied some LED power" from "the supply supplied all of it". That point is the most important epistemic one: lack of a surplus on the numbers is not proof that no ground contribution exists, only that none is shown.
• The one overstatement I corrected: the physicist's "Established" for two items that rest on theory rather than a documented test of this build, and the diode-orientation error shared by two reviewers.

What would settle it

1. Return-path control (settles claim 1 weak form and the mechanism). Replace the earth rod with (a) a floating metal plate of measured capacitance, (b) a capacitor to the battery negative, (c) the rod through a resistor, battery-powered, rod isolated from mains earth. If LED brightness tracks capacitance, the earth is a counterpoise. If a rod outperforms a floating plate of equal capacitance, something else is happening.
2. Frequency and waveform measurement (new). Measure the oscillation frequency with a pickup loop and a scope far from the circuit (RF-safe). If it is about 6 MHz, 25 pF closes the resonance; if 2.5 MHz, about 150 pF is needed and the mechanism is tested.
3. True input and output power (settles claim 1 strong form). Input: battery plus a shunt resistor with a decoupling capacitor, read by a battery-powered recording meter away from the field. Output: LED voltage and current waveforms with a differential probe and RF-safe shunt, or calibrated photometry or calorimetry against a DC-driven reference LED. Verdict rule: output above input minus all other measured losses would show an unaccounted source; below it shows an inefficient RF converter.
4. Long-run battery discharge (42 Ah at 0.36 W, about 1400 h) with and without the rod, as the contrarian proposes.
5. Verify LED orientation from the original LTspice file or video (the schematic does not show it).

Open questions

• Which inductance sets the oscillation frequency, and what is the frequency?
• Where does the 0.36 W go when unearthed (the transistor accounts for only about 6 mW)?
• How was "4 mW per LED" measured?
• What did GndNrgEx_4 (promised light-output measurement) show, and why are _4, _6, _7 and _8 missing?
• Which overunity.com thread of 2015 prompted GndNrgEx_3, and what objections were raised?
• What do Stiffler's SEC documents actually say about earth/counterpoise (only a Google snippet seen)?
• Does the 1N4148-blocks-ground-energy observation hold under a controlled test, and does a series capacitor of equal reactance do the same?

Absence-of-evidence analysis: the missing videos and the silent record

Where the record goes dark: (i) GndNrgEx_4, _6, _7, _8 are absent from the public list (numbering 1, 2, 3, 5, 9, 10), and _4 was the promised light-output measurement, the very data an energy balance needs; (ii) no independent discussion, replication or citation of radiant_one was found; (iii) the author is anonymous with no patent, paper or company; (iv) the channel stopped in March 2018.

Competing explanations for the missing videos: - Never made or renumbered. For: numbering by topic in a hobby series can skip (the series has odd names GrdNrgEx_2, GrnNrgEx_5, showing loose labelling). Against: _4 was explicitly promised. Decided by: playlist order and video upload dates; Wayback Machine of the channel and playlist PLgevW9RSTBFwr5FD3a4V5CY20BYCP18Go. - Deleted or made private by the author (for example because the result was unimpressive, or measurement errors were found, or to reissue as corrected versions as was done with the "Corrected" schematic). For: GrdNrgEx_2 was explicitly reissued with a corrected diagram, showing the author does revise and replace. Against: nothing seen. Decided by: Wayback captures, the author's later descriptions or comments. - Removed under pressure or by platform action. For: no evidence found. Against: the channel kept its other overunity-disclaiming videos and a Moray-framed ones public; nothing on record suggests a takedown. Decided by: platform records, which are not obtainable; the author's own statement. - Hidden because they showed an anomalous result. For: none in the record; this is speculation. Against: the author continued publishing similar claims afterwards. Decided by: recovery of the videos. The most probable is a mundane one (not made, or removed by the author), but the record cannot yet separate them. I do not treat the gap as evidence of suppression or of a hidden result.

On the wider silence: a 469-subscriber channel, with the longest video 4:30, is the expected record for a hobby experimenter and the lack of independent coverage is not contradiction. But the Investigator's run was 7.6 of 10 minutes and did not reach the overunity.com forum's own search, Stiffler's documents on Scribd/studylib, the remaining transcripts, or the Wayback Machine. So the silence found is partial and reflects search limits as much as the world. Records that would decide: overunity.com's own 2015 thread (a contemporaneous independent audience with objections the author answered), Wayback captures, YouTube comment threads, the author's replies.

Investigator findings retained in the record (whatever the verdict)

Twelve videos 2015-06 to 2018-03 (eight on ground energy; others on single-wire energy and HV "electro-radiant" arcs); construction details (Lb about 20 turns of 24 gauge on a hot-glue stick; L1 about 18 gauge on an acrylic spool; all air-core; 470 pF yellow capacitor); 30 mA both states; transistor +1.2 C unearthed versus +0.7 C earthed over 22.5 C ambient after 5 min; 12 V 42 Ah sealed lead-acid battery; "should not be powered by a mains power supply"; the author's Moray framing; the Stiffler SEC-15 and Slayer-exciter lineage; related YouTube grounding demos (Okki Moeljadi, George Chaniotakis; unexamined); a 2018 commenter's unchecked story about "Henry Murry", a fire and a body in a ditch (the commenter's claim, not checked, does not match the Moray record as known [memory, low]); another commenter's "You didn't show the circuit working" (2018). All are single-source and are the sources' own characterizations. The author is a party with an interest in his claim; Stiffler is a seller of exciter kits (commercial stake) [memory, low; not verified], so his "counterpoise" language is his characterization too.

Verdict summary

QE revision

Source: qe.md (verdict MINOR). No deep reads exist (deep/ absent; the manifest holds one image, and the open questions need the videos, which go to the Investigator). I re-checked the QE's points against research/2026-10-04-youtube.md and the PNG notes already in this file. The QE's verdict-level conclusions are accepted; the corrections are to reasoning and numbers.

R1. The "58 %" figure is withdrawn as a finding (changes reasoning, not the verdict)

The only source for "4 mW" is the GndNrgEx_9 description (2016-09-05): "each led is only drawing 4mW from a battery despite being brightly lit!!" [web: https://www.youtube.com/watch?v=K0YO3o4Tu9o, via Investigator]. The 30 mA figure is from GndNrgEx_1/_3 (2015) [web, via Investigator]. I and all three reviewers divided one by the other. Two readings are possible: - (a) Input per LED: "drawing ... from a battery" describes what each LED takes from the supply. Total 52 x 4 mW = 208 mW = about 17 mA at 12 V (16-19 mA over 11-13 V), in the 2016 build. Then the record contains no LED-output measurement at all. "Despite being brightly lit" fits this reading: the author is stressing a low draw. It also means the 2016 build drew about 17 mA, not 30 mA, so the builds or settings differ. - (b) Measured LED power: then 58 % output/input only if the 2015 input (30 mA) also applied to the 2016 build, which is not shown. Both readings stay in the record. Under (a) nothing is known about output. Under (b) the figure is cross-video. I do not choose between them; the QE reads (a) as more natural, which I find reasonable but unproven. Replacement statement: the record contains no measured LED output power; the author's figures neither show nor exclude a surplus. This strengthens the contrarian's point that the experiment as run cannot separate "ground supplied some LED power" from "the supply supplied it all". Every earlier line here that says "own numbers show no surplus" or "58 %" is superseded by this paragraph. The heat-test magnitude check (about 6 mW at the transistor against 360 mW supply) is unaffected as a statement about the transistor, but the "factor of 80" and "42 K" counterfactuals assume about 200 mW reaching the LEDs and are conditional.

R2. Supply type for the 30 mA reading is unknown (new open question)

The battery (12 V 42 Ah sealed lead-acid) and the warning "should not be powered by a mains power supply" are from 2016 (GndNrgEx_9). GndNrgEx_1 (2015) says "The power supply registers no change in the current" [web, via Investigator]; the 2015 supply type is not stated. The physicist, the primer and my own text joined the 2016 battery to the 2015 30 mA. If the 2015 runs used a mains bench supply, the earth-rod loop could close conductively through the supply's protective earth or filter capacitors, not through about 25 pF of stray capacitance. That would change the capacitance/frequency test (item 2 under "What would settle it"): a frequency check settles the resonance question only for a battery-isolated build. The later mains warning may reflect the author's own recognition of this; that is a possibility, not shown.

R3. Diode orientation, extended

QE confirms my diode correction (two parallel same-direction 1N4148s, anode at L2, cathode at the earth rod). It also finds the two clip diodes in series with anode to ground and cathode toward the base (the usual base-reverse clamp of a Slayer exciter) [read: page image p.1, per QE]. I have not independently re-zoomed these two; I record them as the QE's reading.

R4. Material added to the retained record (QE section 3, points 2-6)

• GndNrgEx_9 thumbnail shows two meters reading "12.00" and "0.000"; the second meter's quantity is not stated. If it is a current meter on an amps range, 0.000 is not consistent with 17-30 mA, suggesting RF upset, another quantity, or an off-state. Unresolved [web, via Investigator].
• One Wire Energy (xKaYttqzHcQ): "The power consumption in the driver does not alter with the earth connection". Same "no change in input" pattern as claim 3 in a second setup; single source; also equally consistent with a constant-draw oscillator whose output varies, so it does not by itself indicate an external source.
• GrnNrgEx_5 (7LkbYBLM4Dc): "suggests a real ground connection is important if higher output levels are to be obtained"; thumbnail "First insert the diode into the connection to ground" [web, via Investigator; transcript not read]. Bears on the rod versus floating-plate control: the author claims a real ground matters, a testable statement.
• Google legal removals: the Investigator's Stiffler query showed "3 results removed on a legal request (LumenDatabase)" [research/2026-10-04-energeticforum.md]. Competing explanations: a copyright notice on kit documents or manuals (most likely, since Stiffler-related documents are products); a takedown unrelated to the claims; something else. For none is there evidence. Decided by: the Lumen Database notice itself, which names the requester and URLs. Not treated as evidence of suppression.
• Energetic Forum, kcarring: "12V / 75 mA - 15 Watt 18" Fluoro" (about 0.9 W input for a tube rated 15 W) and jonnydavro's tuning with a magnet on L2: third-party reports from the same circuit lineage, forum posts, rated power is not the same as delivered power; no measurement shown. Kept as quoted items.
• Stiffler "sells kits": retagged. The Investigator's Google result "SEC Exciter 15/20 Kit Assembly Guide" (Scribd) shows a kit guide exists [web: Google snippet, via Investigator]; that Stiffler is the seller is not confirmed. The "commercial stake" label for Stiffler is therefore provisional.

R5. Settle-it list and open questions additions

• What would settle it, new item 6: establish the supply type used for the 30 mA in 2015 (bench supply or battery, mains earth or not), and rerun the earthed/unearthed comparison on an isolated battery with a series shunt.
• What would settle it, item 3 is now more important: it is the first LED output measurement in the record, not a refinement.
• Open questions added: Is "4 mW per LED" input or output, and how was it measured? What do the two meters in the GndNrgEx_9 thumbnail read? Was the 2015 supply mains-earthed? What was the 3-result Lumen removal?
• Investigator tasks (from QE, not done by me): watch GndNrgEx_9; supply type in _1 and _3; transcripts of _1, _10 and GrnNrgEx_5; Wayback, overunity.com 2015 thread, Stiffler documents, Lumen notice.

Verdict changes