Radiant_One_Youtube: Primer

"Ground Energy Extraction" (GndNrgEx): a one-transistor radio-frequency oscillator that lights 52 white LEDs only when one end of the LED array is wired to an earth rod. This primer explains the AC circuit physics behind it and shows how to audit where the LED power comes from.

Source tags: [read] = from the project folder (here, one image); [web] = Internet source, URL given (most come through the Investigator's notes in research/2026-10-04-youtube.md, marked "via Investigator"); [memory] = recalled from training, with a confidence level. The verdicts come from arbiter.md. I recomputed the numbers in section 5 myself, and example.py reproduces them.

Source gap (repeated from claims.md and arbiter.md). The project folder holds one file: a schematic screenshot titled "Ground Energy Extraction (GndNrgEx) Circuit 1.1 (Corrected)", "YouTube Channel: radiant_one" [read: page image p.1]. It has no video, transcript, measurements or author name. 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, via Investigator]. No reviewer re-watched the videos, and neither did I. To go further you would need the videos themselves (especially the missing GndNrgEx_4, _6, _7 and _8), the original LTspice file, and the 2015 overunity.com thread the author answered.

---

1. The idea in one paragraph

An anonymous experimenter posted twelve YouTube videos between June 2015 and March 2018 on the channel radiant_one (469 subscribers) [web: https://www.youtube.com/@radiant_one7857, via Investigator]. In eight of them he builds a small radio-frequency (RF) oscillator (battery-powered in the 2016 build), which he describes as "pretty much" Dr Ronald Stiffler's SEC-15 exciter. He feeds its output through a coil into two strings of 26 white LEDs each, and wires the far end of the LEDs to a metal rod driven into the ground. With the rod connected the 52 LEDs glow brightly. With it disconnected they stay dark. After he tunes the circuit, the supply current is the same, about 30 mA, in both states (a 2015 video; the type of supply used then is not stated). He reads this as the circuit drawing "ground energy". He cites T. Henry Moray, who in the 1920s-30s claimed a "radiant energy" device that took power from the earth. In most of his video descriptions he also writes "No overunity performance is suggested or implied." The mainstream reading is that the earth rod closes an AC circuit: high-frequency current can flow through capacitance (a capacitor passes alternating current but not direct current), so the rod completes a loop through the ground and the stray capacitance between the earth and the battery-powered circuit. The LED power then comes from the supply. The record contains no measured LED output power: the author's "4 mW per LED drawn from a battery" (2016) most naturally reads as the battery draw per LED, not light or electrical power delivered to the LEDs, and the 30 mA comes from a different video a year earlier. So the author's figures neither show nor exclude a surplus (arbiter, QE revision). Nobody has run the control experiment that would separate the two readings, and that experiment is cheap.

2. What you need to know first

Power and energy in a DC circuit. Power P (watts, W) is voltage V (volts) times current I (amperes, A): P = VI. Energy (joules, J) is power times time. A battery's capacity in ampere-hours (Ah) times its voltage gives watt-hours (Wh). A 12 V, 42 Ah battery stores about 504 Wh. Read: Halliday, Resnick and Walker, Fundamentals of Physics, the chapters on current and DC circuits.

Capacitors, inductors and reactance. A capacitor (capacitance C, in farads, F; 1 pF = 10⁻¹² F) stores charge: Q = CV, and the current into it is i = C dv/dt. An inductor (inductance L, in henries, H; 1 µH = 10⁻⁶ H) resists changes in current: v = L di/dt. For a sinusoidal signal at frequency f (hertz) and angular frequency ω = 2πf (radians per second), each behaves like a frequency-dependent resistance called reactance X (ohms): X_L = ωL for an inductor and X_C = 1/(ωC) for a capacitor. So a capacitor blocks DC (X_C is infinite at ω = 0) but passes high-frequency AC easily. Read: Horowitz and Hill, The Art of Electronics, chapter 1 (reactance, impedance, LC circuits).

Resonance and the quality factor Q. In a loop with an inductor and a capacitor in series, the two reactances cancel at one frequency, where ωL = 1/(ωC). Solving gives the resonant frequency f₀ = 1/(2π√(LC)). At resonance the current is limited only by the loop's resistance R. The voltage across the inductor (or the capacitor) is then Q times the driving voltage, where Q = X/R is the quality factor (a dimensionless number). So a resonant loop can turn 12 V into 75 V across one element without a transformer. The energy is not created. It sloshes back and forth between the inductor's magnetic field and the capacitor's electric field, and each cycle the source tops up what the resistance loses. Read: Horowitz and Hill chapter 1; any intro text on RLC circuits.

Displacement current. Maxwell added a term to Ampère's law: a changing electric field acts like a current, with density ε₀ ∂E/∂t, where ε₀ = 8.854 × 10⁻¹² F/m is the permittivity of free space and E is the electric field. That is how "current flows through a capacitor" even though no charge crosses the gap. It is fully real for circuit purposes: a loop can close through air, through a person's body capacitance, or through the capacitance between a bench and the ground. Read: Griffiths, Introduction to Electrodynamics, chapter 7 (Maxwell's displacement current).

Self-capacitance and "ground" in RF work. Even an isolated object has capacitance to its surroundings. For a sphere of radius r far from everything else, C = 4πε₀r, which is about 11 pF for r = 10 cm. A battery-powered circuit on a bench therefore has tens of picofarads to the room, the floor and the earth. Radio engineers make deliberate use of this. An "earth" or "counterpoise" (a wire or plate acting as the other half of an antenna or circuit) gives RF current a return path. Read: the ARRL Handbook or ARRL Antenna Book, sections on grounds and counterpoises [memory, high that these books cover the topic; chapter numbers not given].

Diodes and LEDs as rectifiers. A diode conducts in one direction only, once the forward voltage across it exceeds a threshold (the forward voltage V_f): about 0.7 V for a silicon signal diode such as the 1N4148, and about 2.8-3.6 V for a white LED [web: https://circflow.com/blog/led-forward-voltage-chart/, via arbiter]. Twenty-six white LEDs in series need roughly 26 × 2.9 ≈ 75 V before any current flows. Driven by AC, a diode passes only one half of each cycle (rectification). Read: Horowitz and Hill, chapter 1 (diodes).

Thermal resistance. A transistor's datasheet gives θ_JA (junction-to-ambient thermal resistance, in kelvin per watt): the steady temperature rise per watt of heat dissipated. For the MPSA06 transistor used here θ_JA = 200 K/W [web: https://www.mccsemi.com/pdf/Products/MPSA06(TO-92).pdf, via arbiter]. So a rise of ΔT kelvin means a power of roughly ΔT/θ_JA watts.

Atmospheric electricity. In fair weather the air near the ground carries a vertical electric field of about 100 V/m and a tiny downward current of about 2 pA/m² (picoamperes per square metre) [web: https://space.fmi.fi/~makelaa/fairw.html; NOAA/AFRL handbook ch. 20, https://www.ngdc.noaa.gov/stp/space-weather/online-publications/miscellaneous/afrl_publications/handbook_1985/Chptr20.pdf; via arbiter]. The earth's crust also carries natural "telluric" currents, with voltage gradients of order volts per kilometre [memory, medium; not web-checked]. These are the known natural sources an earth rod could tap. Read: Feynman, Lectures on Physics Vol. II, chapter 9, "Electricity in the Atmosphere" [memory, high].

Measuring power at RF. Ordinary multimeters assume DC or 50-60 Hz signals. At several megahertz their leads act as antennas, and their internal diodes rectify whatever they pick up, so they can read nonsense or be damaged. The author warns of this on the schematic [read: p.1]. Measuring RF power correctly takes a shunt resistor with bypass capacitors, a battery-powered meter placed away from the field, or calorimetry (measuring heat) or photometry (measuring light) against a calibrated reference.

3. The textbook view

What the circuit is, as an engineer reads it [read: page image p.1; lineage via Investigator]:

• Supply. V1 = 11-13 V. The battery is documented for the 2016 build: in GndNrgEx_9 (2016) it was a 12 V, 42 Ah sealed lead-acid battery, and the author says the circuit "should not be powered by a mains power supply" [web: https://www.youtube.com/watch?v=K0YO3o4Tu9o, via Investigator]. The 30 mA reading is from the 2015 build (GndNrgEx_1 and _3), where the descriptions say only "the power supply" and do not state its type [web: https://www.youtube.com/watch?v=JAxUJODVMfY and https://www.youtube.com/watch?v=6XTDBlYBqfI, via Investigator]. This matters. If the 2015 runs used a mains bench supply, the earth-rod loop could close conductively through the supply's protective earth or its filter capacitors, not through tens of picofarads of stray capacitance (arbiter, QE revision R2). The later mains warning may be the author's own recognition of this; that is a possibility, not shown.
• Oscillator. Transistor Q1 (MPSA06, an NPN transistor). R1 = 1 MΩ biases its base from the supply. A feedback branch of C1 = 470 pF in series with Lb = 1.4 µH runs from base to ground. L1 = 8.6 µH sits in the collector lead to the supply. Two 1N4148 diodes in series run from base to ground, and their midpoint is brought out to an "Alligator_Clip" lead. Both have the anode toward ground and the cathode toward the base, the usual clamp that stops the base being driven far below ground in a Slayer exciter [read: page image p.1, per QE; the arbiter records this as the QE's reading]. The author says this left half is "pretty much" a SEC-15 exciter, a descendant of the hobbyist "Slayer exciter" [web: GufTw6ECoz4 transcript, via Investigator]. Lb is about 20 turns of 24-gauge wire on a hot-glue stick, L1 is 18-gauge wire on an acrylic spool, and all coils are air-cored [web, same].
• Output. From the collector node, L2 = 26.7 µH runs in series to an LED group: two parallel strings of "26 White LED's in series". Two 1N4148 diodes are drawn across the whole group, one above and one below. Both point the same way, anode toward L2 and cathode toward the earth rod, so they act as one clamp across the group. The far end of the group goes to "Earth Rod" [read: page image p.1, as corrected by the arbiter]. The LED orientation is not drawn. The LEDs can light only if they conduct in the direction opposite to the diodes (earth rod toward L2); otherwise the diodes would clamp the group at about 0.7 V.
• Notes printed on the schematic. "Warning: This circuit produces RF. Multimeters and oscilloscope can be damaged or read incorrectly!" and "Tuning: Unearthed current draw to match earthed current draw. Tune by adjusting the physical proximity of alligator clip to L2" [read: p.1]. The two orange vertical bars over the middle of the image are probably a video-editing overlay. They cut through the title text and have no electrical meaning [read: p.1; arbiter correction 2].

How the mainstream explains the LEDs lighting. The oscillator makes the collector voltage swing at some megahertz frequency. That swing drives a loop: collector → L2 → LED group → earth rod → the ground → the stray capacitance C_s between the earth and the battery-powered circuit (with the room around it) → back to the circuit's negative rail. (If a mains-earthed bench supply was used, as may have been the case in 2015, the loop could instead close through the supply's earth wire; see Supply above.) Each piece of this loop is ordinary physics:

1. The loop can close without a wire, through displacement current in C_s.
2. It can be resonant. L2 and the capacitances in the loop resonate when f = 1/(2π√(L2·C)). Near resonance the voltage across the LED group can ring up to several times the 12 V supply, which is enough to reach the ~75 V a 26-LED string needs.
3. Rectification. On one half-cycle the LED strings conduct (if the voltage is high enough). On the other half the 1N4148 clamp conducts. A capacitor cannot pass a net DC current, so the charge each half moves must balance over a cycle. The diodes provide the return half. Remove them, or put a single diode in series with the rod, and the loop has no way to balance its charge, so the LEDs go dark. This is the mainstream reading of the author's report that "ground energy appears to stop when it encounters a IN4148 diode" [web: GndNrgEx_10 description, https://www.youtube.com/watch?v=Bhj8N7_stZY, via Investigator; reading from arbiter].
4. Unearthed, the same loop is open. The oscillator still runs and dissipates power in the transistor, the coils and the alligator-clip network. The tuning procedure moves the clip until that unearthed power draw equals the earthed draw.

In this picture the earth is a counterpoise, a return path, not a source. Stiffler's own documentation reportedly says the same about his exciters: "earth ground or a long wire (antenna) but in reality a counterpoise" [web: Google snippet of the Scribd document "Wireless", via Investigator; full document not read]. The earth's role as an RF return is Established in general. Whether it is the whole story for this circuit has not been measured (section 6).

Energy conservation. Over any closed boundary drawn around the battery, the circuit and the earth rod: energy out of the battery = LED power + heat in the transistor, coils, diodes and wiring + power radiated as radio waves + whatever, if anything, enters through the rod. The only known inputs through the rod are atmospheric and telluric currents, and section 5 shows they are far too small.

4. Where the inventor departs

The exact point of departure is the first row. Both readings agree on what is seen: the LEDs light only with the rod connected. They differ on whether the rod is a path or a source. The author's own descriptions point both ways. He disclaims overunity, and he also describes a "much larger response" from the ground. The arbiter records this tension and does not treat the loose wording as a formal inconsistency: "a much larger response" may describe the visible LED response, not an energy ratio. The earlier argument that "his measured numbers fall on the no-overunity side" is withdrawn (arbiter, QE revision R1), because no LED output power was measured; his figures neither support nor contradict either statement.

The author deserves credit for good practice here, which the contrarian review stresses. He published component values, coil construction and a tuning method. He warned that his own meters were unreliable. He answered forum objections with a heat test. He used a battery rather than a mains supply (documented for the 2016 build). He tried a metal-enclosure control against the radio objection. And he repeatedly disclaimed overunity.

5. Worked example: where could the LED power come from?

This is an energy audit with real numbers, reproduced by example.py. Run it with: .venv/bin/python review/Radiant_One_Youtube/example.py

Step 1. What frequency does it run at? Resonance is where X_L = X_C, so ωL = 1/(ωC), which gives ω₀ = 1/√(LC) and f₀ = 1/(2π√(LC)). - Lb with C1: f = 1/(2π√(1.4×10⁻⁶ × 470×10⁻¹²)) = 1/(2π × 2.565×10⁻⁸) = 6.20 MHz. - L1 with C1: 2.50 MHz. Lb + L1 with C1: 2.32 MHz.

The schematic does not say which inductance sets the frequency, and the frequency was never measured in anything seen. So treat 2.3-6.2 MHz as the plausible band. The wavelengths (λ = c/f, with c = 3×10⁸ m/s) are 48-130 m, far larger than the circuit. The circuit is therefore a poor antenna and radiates little.

Step 2. What voltage do the LEDs need? One string: 26 × 2.9 V ≈ 75 V. From a 12 V supply that is a step-up of about 6.3. The inductance ratio alone gives only √(L2/L1) = √(26.7/8.6) = 1.76, so resonant ring-up with Q of order 6-10 is needed. That is a modest Q for air-core coils.

Step 3. What capacitance would resonate L2? C = 1/((2πf)² L2). - At 6.2 MHz: C = 1/((3.90×10⁷)² × 26.7×10⁻⁶) = 24.6 pF, and L2's reactance is ωL2 ≈ 1,040 Ω. - At 2.5 MHz: C ≈ 151 pF, and the reactance is about 420 Ω.

For scale, an isolated sphere of radius 25 cm has C = 4πε₀r = 4π × 8.854×10⁻¹² × 0.25 = 27.8 pF. So about 25 pF is just the self-capacitance of a bench-sized rig. About 150 pF is six times more, and harder to reach with a rod and LED packages alone. Measuring the frequency would therefore test the mechanism (arbiter, "What would settle it", item 2).

Step 4. Can that stray capacitance carry the LED current? This step is an illustration: if about 200 mW reached the LEDs (52 × 4 mW = 208 mW, taking the author's figure as LED power, which is only one of two readings; see Step 5), could the capacitive path carry it? At 75 V that is a mean current of 208/75 = 2.8 mA in total, or 1.4 mA per string (plausible for small LEDs). If it flows in half-sine pulses, the mean is (peak)/π, so the peak is about π × 2.8 ≈ 8.7 mA. The voltage needed across 25 pF at 6.2 MHz to pass 8.7 mA is V = I/(ωC) = 0.0087 × 1,040 ≈ 9 V, which is easily available. So the ordinary capacitive return path can carry the current the LEDs need. Nothing in the numbers requires a new source.

Step 5. What the author's two numbers can and cannot tell you. There are only two power-related figures in the record, and they come from different videos a year apart: - 30 mA supply current, earthed and unearthed (GndNrgEx_1 and _3, 2015; supply type not stated) [web, via Investigator]. - "Each led is only drawing 4mW from a battery despite being brightly lit!!" (GndNrgEx_9, 2016) [web: https://www.youtube.com/watch?v=K0YO3o4Tu9o, via Investigator].

The second sentence can be read two ways, and the arbiter keeps both.

Reading (a): 4 mW is the battery draw per LED (input). "Drawing ... from a battery" describes what is taken from the supply, and "despite being brightly lit" fits an author stressing a low draw. Then the total draw is 52 × 4 mW = 208 mW, and the implied battery current is

So the 2016 build drew about 17 mA, not 30 mA, which means the builds or their settings differ. Under this reading there is no LED output measurement in the record at all.

Reading (b): 4 mW is measured LED power (output). Then a ratio can be formed, but only by assuming the 2015 input of 30 mA also applied to the 2016 build, which is not shown:

Earlier versions of the reviews, the arbiter and this primer treated reading (b) as fact and concluded "58%, so no surplus". The QE found that this mixes two videos and probably misreads the sentence, and the arbiter withdrew it. The present position is: the record contains no measured LED output power, so the author's figures neither show nor exclude a surplus. Two further cautions still apply. (i) The "4 mW per LED" has no stated method, and the author himself warns that meters misread in this field. (ii) Even a measured ratio below 100% would not prove the ground contributes nothing. It would show only that no contribution is needed to explain the numbers (the contrarian's point, adopted by the arbiter).

The lesson for a learner: before dividing one number by another, check that both describe the same quantity (input or output), the same build and the same date.

Step 6. The heat test, in milliwatts. The author ran the circuit for 5 minutes, switched it off, and read the transistor case with an infrared (IR) thermometer. He measured it after switch-off because the thermometer read about -50 while the RF field was on. Results: +1.2 K above the 22.5 °C ambient when unearthed and +0.7 K when earthed [web: https://www.youtube.com/watch?v=6XTDBlYBqfI, via Investigator]. Divide by θ_JA = 200 K/W: - unearthed 1.2/200 = 6 mW; earthed 0.7/200 = 3.5 mW; difference 2.5 mW. - If about 200 mW (the 208 mW of reading (b)) moved out of the transistor and into the LEDs, the temperature change would be 0.208 × 200 ≈ 42 K, not 0.5 K. This is a counterfactual, since no LED power was measured.

So the transistor accounts for only about 6 mW of the 360 mW drawn when unearthed (360 mW = 12 V × 30 mA; the heat test and the 30 mA are both from the 2015 videos). Allowing for the case being cooler than the junction, and for cooling between switch-off and the reading, it might be a few times that, which is still far below 360 mW. The rest (about 350 mW) is unmeasured. It could be heat in the coils (at MHz, current crowds to the wire surface, the "skin effect", which raises resistance), in the diodes and wiring, or it could be meter error. Radiation is unlikely to be large for a coil much smaller than 48-130 m [memory, medium]. The heat test has the right sign for diverting power to the LEDs, but if about 200 mW had moved to the LEDs, the 2.5 mW change would be too small by about a factor of 80 (208/2.5 = 83). It cannot confirm the energy balance in either direction.

Step 7. Could natural earth current supply the LEDs? The fair-weather air-earth current density is about 2 pA/m² [web]. Even collected over 100 m² that is 2×10⁻¹⁰ A, against the 2.8×10⁻³ A the LEDs would need on the Step 4 illustration (a mean of 208 mW at 75 V; even 1% of that would leave the shortfall above 10⁵): short by a factor of about 14 million. A rod collects far less than 100 m². Atmospheric electricity cannot be the source at this scale. (Telluric currents in the ground are a separate quantity. No one in this review web-checked their size, and that figure stays [memory, medium].)

Step 8. The cheap long-run test. 12 V × 42 Ah = 504 Wh. At 0.36 W (the 2015 figure of 30 mA) the battery lasts about 1,400 h (about 58 days); at 0.208 W (reading (a) of the 2016 figure) about 2,400 h. Running two identical batteries, one circuit earthed and one unearthed, and logging their voltages would directly show whether the earth changes the battery drain (the contrarian's proposal). Because the draw is set equal by tuning, this test mainly checks the tuning and the meters. The cleaner test is the return-path control in section 6.

What the example shows. Every number in the observed effect is consistent with an ordinary resonant RF loop closing through about 25 pF of stray capacitance (or, if the 2015 supply was mains-earthed, through the supply's earth), powered by the supply. The weak points are in the data, not the physics. The frequency, the LED output power and the losses were never measured with RF-safe methods, and the only two power figures come from different videos and may both describe input.

6. Evidence table

Verdicts are from arbiter.md. "Replicated?" means by anyone other than the author.

What would settle it (arbiter, in brief): 1. Replace the earth rod with (a) a floating metal plate of measured capacitance, (b) a capacitor to the battery negative, and (c) the rod through a resistor. If brightness tracks capacitance, the earth is a counterpoise. If the rod beats a floating plate of equal capacitance, something else is happening. 2. Measure the oscillation frequency with a pickup loop and a scope kept away from the circuit. About 6 MHz fits ~25 pF; about 2.5 MHz needs ~150 pF. 3. Measure true input power (shunt resistor with bypass capacitor, battery-powered recording meter away from the field) and true LED power (differential probe, or photometry or calorimetry against a DC-driven reference LED). After the QE revision this is the most important item: it would be the first LED output measurement in the record, not a refinement of an existing one. 4. Run a long-term battery discharge with and without the rod. 5. Confirm the LED orientation from the LTspice file or the video. 6. Establish what supply gave the 2015 30 mA reading (bench supply or battery; mains-earthed or not), and rerun the earthed/unearthed comparison on an isolated battery with a series shunt. The frequency test in item 2 settles the stray-capacitance question only for a battery-isolated build.

The record that goes dark. GndNrgEx_4, _6, _7 and _8 are missing from the channel's public list (the numbering runs 1, 2, 3, 5, 9, 10). GndNrgEx_4 was the promised measurement of light output, exactly the data an energy balance needs [web, via Investigator]. The arbiter sets out competing explanations: - Never made, or renumbered. For: the series labels are loose (GrdNrgEx_2, GrnNrgEx_5). Against: _4 was explicitly promised. - Deleted or made private by the author. For: he does reissue videos, as with the "Corrected" diagram. - Removed by the platform or under pressure. No evidence found. - Hidden because they showed an anomaly. No evidence; he kept publishing similar material afterwards.

Wayback Machine captures of the channel and playlist PLgevW9RSTBFwr5FD3a4V5CY20BYCP18Go, the upload order, and the author's comments could decide between them. The arbiter judges a mundane cause most likely but does not treat the gap as settled either way. The wider silence (no independent replication, an anonymous author, the channel stopping in March 2018) is the expected record for a small hobby channel. It is also partly a search limit: the Investigator's run used 7.6 of its 10 minutes and did not reach overunity.com's own search.

Other material kept in the record [web, via Investigator; all single-source and the sources' own characterizations]: - Three 2015 high-voltage videos with "electro-radiant" language: a flyback driving a 25 W 240 V bulb "via water capacitor, double arc and ground wire"; a "radiant green arc" with a felt "shockwave"; and copper foil attracted to "electro-radiant emissions". - A Joule thief replication. - Related YouTube grounding demonstrations (Okki Moeljadi, "Electro Radiant Energy - Grounding Effect"; George Chaniotakis, "Radiant Ground Energy!"), not examined. - A 2018 commenter's story attributing a quartz-crystal device to "Henry Murry", with a lab fire and a body in a ditch. This is the commenter's claim, unchecked, and it does not match the Moray record as known [memory, low]. - Another commenter: "You didn't show the circuit working." - The GndNrgEx_9 thumbnail shows two meters reading "12.00" and "0.000"; what the second meter measures is not stated [web: https://www.youtube.com/watch?v=K0YO3o4Tu9o, via Investigator]. If it is a current meter on an amps range, 0.000 does not fit a draw of 17-30 mA, which suggests RF upset, a different quantity, or an off state. Unresolved. - One Wire Energy (xKaYttqzHcQ): "The power consumption in the driver does not alter with the earth connection" [web: https://www.youtube.com/watch?v=xKaYttqzHcQ, via Investigator]. This is the same "no change in input" pattern as claim 3, in a second setup. It is equally consistent with an oscillator whose draw stays constant while its output varies, so on its own it does not point to an outside source. - GrnNrgEx_5 "Path to Ground" (7LkbYBLM4Dc): the description "suggests a real ground connection is important if higher output levels are to be obtained", and the thumbnail says "First insert the diode into the connection to ground" [web: https://www.youtube.com/watch?v=7LkbYBLM4Dc, via Investigator; transcript not read]. "A real ground matters" is a testable statement, and it is exactly what the rod-versus-floating-plate control (item 1 above) would test. - Energetic Forum, same circuit lineage: kcarring reports "12V / 75 mA - 15 Watt 18" Fluoro", i.e. about 0.9 W of input lighting a fluorescent tube rated 15 W; jonnydavro (2011) reports tuning his own SEC variant by placing a small magnet on its L2 choke [web: Energetic Forum thread above, via Investigator]. These are third-party forum posts with no measurement shown; a tube's rated power is not the power actually delivered to it. - Google legal removals: the Investigator's search on Stiffler showed "3 results removed on a legal request (LumenDatabase)" [web, research/2026-10-04-energeticforum.md]. This is a record gap, so the arbiter sets out competing explanations: a copyright notice on kit documents or manuals (judged most likely, since Stiffler-related documents are products); a takedown unrelated to the claims; something else. There is no evidence for any of them yet. The Lumen Database notice itself, which names the requester and the URLs, would decide. It is not treated as evidence of suppression. - Interested parties: the author has a stake in his own interpretation. A Google result shows an "SEC Exciter 15/20 Kit Assembly Guide" (Scribd) exists [web: Google snippet, via Investigator], but that Stiffler is the seller is not confirmed, so the "commercial stake" label for Stiffler is provisional. His "counterpoise" wording is his own characterization either way.

7. What a learner should take away

Solid. - Stray capacitance and displacement current are real and can close an RF circuit through the earth. - Series resonance can raise 12 V to the 75 V a long LED string needs. - Diodes in such a loop must provide a return half-cycle, because a capacitor passes no net DC. - The oscillator design is a well-known hobby circuit, and the LEDs do light. - RF really does fool ordinary meters.

Wrong, and why the mistake is instructive. - Equal current in two states, set by tuning, is not evidence about an extra source. Ask of any "no change in input" claim: was the equality measured, or arranged? - Battery current alone does not tell you where the power goes. The heat test was a good instinct, but at 200 K/W a 1.2 K rise is about 6 mW. Always convert a temperature rise into watts before drawing a conclusion. - Naming a path as a source. "Ground energy stops at a 1N4148" is exactly what a capacitive return path predicts. An observation that both hypotheses predict cannot pick between them. - Undefined terms ("ground electrons", "magnetic-dielectric interaction") cannot be tested until they carry units and a predicted number.

Genuinely open. - What frequency does the circuit actually run at? - Where do the ~350 mW go when it is unearthed? - Is "4 mW per LED" input or output, and how was it measured? - Was the 2015 supply mains-earthed? (If so, the earth rod may have closed the loop through the supply's own earth.) - What do the two meters in the GndNrgEx_9 thumbnail ("12.00", "0.000") measure? - What was behind the three Google results removed on a legal request? - Does a floating plate of equal capacitance work as well as the earth rod? If it does, the counterpoise reading is confirmed. If it does not, there is a real anomaly to chase. - What did the missing GndNrgEx_4 show? - What objections did the 2015 overunity.com thread raise?

All of these are bench experiments within a hobbyist's reach.

8. Check yourself

1. The author tunes the alligator clip until the unearthed supply current equals the earthed current, both 30 mA. Why can this equality not, by itself, show that the earth supplies energy?
2. Compute the capacitance that resonates a 26.7 µH coil at 4 MHz. Is it closer to a bench rig's self-capacitance or not?
3. A transistor with θ_JA = 200 K/W warms 3 K above ambient. About how much power is it dissipating? What would a 100 mW change in its dissipation do to its temperature?
4. Why does a single 1N4148 diode in series with the earth lead stop the LEDs lighting, on the mainstream reading?
5. Design the one control experiment that would most directly distinguish "earth as counterpoise" from "earth as energy source", and state what result would favour each.

9. Further reading

Mainstream - Horowitz and Hill, The Art of Electronics (3rd ed.), chapter 1: reactance, resonance, diodes; and the oscillator chapter. - Griffiths, Introduction to Electrodynamics, chapter 7: displacement current and Maxwell's equations. - Feynman, Lectures on Physics Vol. II, chapter 9, "Electricity in the Atmosphere". - ARRL Handbook / ARRL Antenna Book: grounds, counterpoises, RF measurement. - MPSA06 datasheet: https://www.mccsemi.com/pdf/Products/MPSA06(TO-92).pdf - Fair-weather atmospheric electricity: https://space.fmi.fi/~makelaa/fairw.html

The inventor and lineage [web, via Investigator] - Channel radiant_one: https://www.youtube.com/@radiant_one7857. Key videos: GndNrgEx_1 (JAxUJODVMfY), GrdNrgEx_2 Circuit & Tuning (GufTw6ECoz4, source of the schematic), GndNrgEx_3 Analysis (6XTDBlYBqfI, heat test), GrnNrgEx_5 Path to Ground (7LkbYBLM4Dc), GndNrgEx_9 (K0YO3o4Tu9o), GndNrgEx_10 Asymmetrical Pathways (Bhj8N7_stZY), One Wire Energy (xKaYttqzHcQ, kQK7yOgGKbI). - Dr Ronald Stiffler, "Spatial Energy Coherence" (SEC) exciter documents (Scribd, studylib; not read in this review). - Energetic Forum thread "Joulethief SEC exciter and variants" (link above). - T. Henry Moray, The Sea of Energy in Which the Earth Floats [memory, high]. See also this review's Moray project, ../Morray-TH/primer.md.

---

Answers

1. The tuning procedure makes the two currents equal: the clip is moved until the unearthed draw rises to match the earthed one. Equal input power in two states says nothing about whether a further source feeds the LEDs. In the unearthed state the same power simply goes into other losses. You would need to measure the output powers and the losses, or run a control with the rod replaced.
2. C = 1/((2π × 4×10⁶)² × 26.7×10⁻⁶) = 1/(6.32×10¹⁴ × 26.7×10⁻⁶) ≈ 59 pF. That is about the self-capacitance of a 0.5 m sphere (55.6 pF), so it is within reach of a rig plus a person and some wiring, but larger than the ~25 pF needed at 6.2 MHz.
3. P ≈ 3/200 = 15 mW. A 100 mW change would shift its temperature by 0.1 × 200 = 20 K.
4. Through a capacitive return path the current must alternate: a capacitor cannot pass a net DC current. A series diode lets charge through in one direction only, so the stray capacitance charges up to the peak voltage and current stops. The LEDs then have no continuing current. Nothing about an energy source is needed to explain it.
5. Replace the earth rod with a floating metal plate (not touching earth) whose capacitance to the circuit's surroundings is measured, and compare LED brightness (photometer) and battery draw against the rod, with the frequency measured too. If brightness tracks the measured capacitance and a plate matched to the rod's effective capacitance performs as well, the earth is a counterpoise. If the rod clearly beats a plate of equal capacitance, at the same measured input power, the earth contributes something not yet explained, and the true RF power balance (arbiter item 3) becomes the next experiment.