RankingsPrimerArbiterPhysicistMathematicianContrarianQESites logClaimsSourcesResearch notes

Primer: Chris Sykes, "Guidelines to Bucking Coils"

Step 4 study document. Subject: Chris Sykes, Guidelines to Bucking Coils: Lenz's Law Free Power Extraction, version 2.7 (26.02.16), 49 pages, together with what the Investigator found about his later forum work and about Floyd Sweet's device. Verdicts come from arbiter.md; this primer adds no new ones.

Source tags: [read: p.N] means the Guidelines, page N. [read: research/...] means a file the Investigator saved. [web] gives a URL. [memory] is recalled textbook knowledge, with its confidence. [calc] is a calculation reproduced in example.py.

Source note: the Guidelines have a good text layer. Deep reads from page images now cover pp.4-19 and 21-42: the coil experiment (pp.14-19), plus pp.4-13, 21-24, 25-33 and 34-42, which were added after the quality check (QE). Pages 45-49 (the rotor experiment, the lamp "swipe" experiment and the alternator rig) have no page-image deep read yet. What this primer says about them rests on the text layer. The Guidelines contain no power-in/power-out measurement of any device. Every power figure in this primer comes from outside the document: Sykes's forum posts and website, a replicator's post, or Floyd Sweet's 1986 lab notes as transcribed by Sykes. Those were read through the Investigator's notes (forum threads: first page only). The arbiter also read the lab-notes transcription directly for the lines quoted below. Itsu's replication data, the scans of Sweet's notes and Sykes's device videos were not seen.

1. The idea in one paragraph

When you draw current from a generator or a transformer, the input has to work harder. In a generator the shaft gets stiffer; in a transformer the primary takes more current. That reaction is called Lenz's law. Chris Sykes, an independent experimenter who runs the hyiq.org and aboveunity.com sites, proposes a way round it. Put two output coils side by side ("partnered" or "bucking" coils) and wind them in opposite directions, so that their magnetic fields cancel each other while their voltages add. If the magnetic part cancels, he argues, the load's reaction cannot reach the input coil, so the load can be powered without the usual extra drag. He adds practical conditions: a loosely coupled (gapped) core, a step-up turns ratio, and enough load current. He also describes a diode and square-wave drive, but he calls it "a work in progress", and the circuit he gives beginners uses a sine-wave drive and no diode. He reads Floyd Sweet's 1980s "Vacuum Triode Amplifier" (VTA) and many other free-energy devices as working on this principle. He is also careful in places: he says this is "no new form of Energy", that one build in ten works, and that any claim of more output than input must come with measurements.

2. What you need to know first

Faraday's law of induction. A changing magnetic flux through a loop of wire drives a voltage around it. The flux Φ (capital phi, unit weber, Wb) is the magnetic field strength B (unit tesla, T) times the area it passes through, for a uniform field at right angles to the area. The induced voltage is called the emf (electromotive force), and emf = −N dΦ/dt for a coil of N turns. The minus sign is Lenz's law. Read: Griffiths, Introduction to Electrodynamics, ch.7 (Electrodynamics) [memory, high]; Feynman Lectures Vol. II ch.17 (The Laws of Induction) [memory, high].

Inductance and the RL circuit. A coil opposes changes in its own current, because a change of current changes its own flux and induces a back-voltage L di/dt. L is the self-inductance (unit henry, H). Switch a coil of inductance L and total resistance R onto a steady voltage V and the current rises as i(t) = (V/R)(1 − e^(−t/τ)), with time constant τ (tau) = L/R. The coil also stores energy W = ½ L i². Read: Griffiths ch.7 (inductance, energy in magnetic fields); any introductory circuits text on first-order RL circuits.

Mutual inductance and transformers. Two coils that share flux are coupled by a mutual inductance M: a changing current in coil 1 induces M di₁/dt in coil 2, and the same M couples coil 2 back to coil 1. The coupling coefficient k = M/√(L₁L₂) lies between −1 and +1. A transformer is two tightly coupled coils on an iron core; when the secondary is loaded, the primary current rises to supply the power. For an ideal (lossless, perfectly coupled) transformer with N₁ primary turns and N₂ secondary turns, V₂/V₁ = N₂/N₁ and I₂ = I₁N₁/N₂, so V₁I₁ = V₂I₂: stepping voltage up steps current down by the same factor, and power in equals power out. Sykes states this rule correctly on p.10 (240 V × 1 A in, 24 V × 10 A out) [read: p.10]. Read: Feynman Vol. II ch.17 (mutual inductance) and ch.16 (Induced Currents) [memory, high].

AC circuits with complex numbers (phasors). For a sinusoidal drive at angular frequency ω (omega, rad/s; ω = 2πf with f the frequency in hertz), each voltage and current is written as a complex number whose size is the amplitude and whose angle is the phase. A resistor has impedance R, an inductor jωL and a capacitor 1/(jωC), where j = √−1 (engineers write j, not i, to avoid confusion with current). Kirchhoff's laws then become linear algebra with complex numbers. The average power into an element is P = ½ Re(V I), where I is the complex conjugate of I and V, I are peak amplitudes. Only the real (resistive) part of an impedance absorbs average power. Read: any first circuits text, the chapter on sinusoidal steady-state analysis.

Power measurement: mean versus RMS. RMS (root mean square) is the square root of the time average of a squared signal. Average power is the time average of the product v(t)·i(t). It equals V_RMS × I_RMS only when voltage and current are in phase and the same shape (a pure resistor). For a coil, a pulse or a diode circuit, the product of RMS values can be far larger than the real power. This matters a great deal in this project. Read: a circuits text on "real, reactive and apparent power".

Potentials. The magnetic field can be written B = ∇×A, where A is the vector potential, and the electric field as E = −∇φ − ∂A/∂t, where φ (lowercase phi) is the scalar potential in volts. These come up because Sykes uses "A" and "Phi" in his own way. Read: Griffiths ch.10 (Potentials and Fields) [memory, high]; Feynman Vol. II ch.15 (The Vector Potential). The Guidelines quote Feynman on the vector potential on p.20 but give no chapter [read: p.20]; the "Vol. II ch.15" reference comes from Sykes's hyiq.org page "A Vector Potential" [read: research/2026-10-04-hyiq.org.md]. (The Guidelines call him "Sir Richard Feynman"; Feynman was not knighted [memory, high].)

Magnetic domains and the Bloch wall. Inside iron the magnetization is split into regions (domains) each magnetized one way. The thin boundary where the direction rotates from one domain to the next is a Bloch wall, typically tens to a few hundred nanometres thick [web: https://www.sciencedirect.com/topics/materials-science/domain-wall, as read by the mathematician]. Read: Feynman Vol. II ch.37 (Magnetic Materials) [memory, medium on the chapter number]; Kittel, Introduction to Solid State Physics, the ferromagnetism chapter.

The B-H curve, its "knee", and hysteresis loss. H is the magnetising field set by the coil current (ampere-turns per metre of core, unit A/m); B is the flux density that results in the core (tesla). Plot B against H for an iron or ferrite core and you get the B-H curve. At small H, B rises steeply, because the core multiplies the coil's field many times. Then the curve bends over and flattens, because the core is saturating: nearly all its domains already point along the field. The bend is called the "knee". Above it, extra current buys little extra flux, and the coil's inductance falls. Drive the core back and forth and B lags behind H, so up and down trace different paths. The closed path is the hysteresis loop. The area inside it, times the core volume, is the energy turned into heat in each cycle. That is why hysteresis loss grows with frequency. Sykes says his devices work "in and around the knee of the BH Curve" and do "not move through Hysteresis" [read: p.42]. Read: any electrical-machines or power-electronics text, the chapter on magnetic cores [memory, high]. Ferrites are ceramic magnetic materials with low eddy-current loss. N87 is a common manganese-zinc ferrite grade for power transformers, made by TDK (EPCOS) [memory, high].

COP (coefficient of performance). In this field COP means P_out / P_in: average output power divided by average input power, both measured. COP < 1 is an ordinary lossy device. COP > 1 would mean more power out than in, the "over-unity" claim. (In heat-pump engineering COP has a different meaning, heat moved per unit of work. That meaning does not apply here.) The Guidelines never use the word. It enters through the forums: "COP 1.7" was attributed to Sykes by a forum critic, and a replicator reported "about 0.82 COP" (section 6).

3. The textbook view

3.1 Induction and Lenz's law

For a coil of N turns, emf = −N dΦ/dt. For a straight conductor of length l moving at speed v through a field B at right angles, emf = B v l (l is a length in metres; then T × m/s × m = volts). When current i flows, the same field pushes on the conductor with force F = B i l, opposing the motion. The mechanical power needed to keep it moving is F v = B i l v = emf × i, which is exactly the electrical power delivered. So:

  • No current, no reaction force: the generator turns freely (only friction).
  • Current flows: the reaction force appears, and the shaft must supply emf × i.

For a rotating generator the same statement reads P = e i = τ_mech ω_shaft, where τ_mech is shaft torque (N·m) and ω_shaft is shaft angular speed (rad/s). Lenz's sign is energy conservation written as a direction.

3.2 Coupled coils, reciprocity and passivity

Two coils with self-inductances L₁, L₂ and mutual inductance M store energy

W = ½ L₁ i₁² + M i₁ i₂ + ½ L₂ i₂².

This must be zero or positive for every pair of currents (you cannot store negative energy in a magnetic field), which is why M² ≤ L₁L₂, i.e. |k| ≤ 1 [memory, high]. M is the same in both directions (M₁₂ = M₂₁): whatever geometry lets coil 1 induce voltage in coil 2 also lets coil 2's current push back on coil 1. This is called reciprocity.

For any circuit made only of resistors, inductors (coupled or not), capacitors and diodes, the power delivered by the sources equals the rate of change of stored energy plus the heat in all resistors. Over a full cycle of steady operation the stored energy returns to its starting value, so

average power from sources = average heat in all resistors (load included) ≥ average load power.

No winding direction, core gap or coupling arrangement changes this inequality. A diode or a saturating iron core can change waveforms and where energy is stored, but not the sign of the balance: a diode only dissipates, and an iron core's hysteresis loop also only dissipates [memory, high]. This is the result the reviewers call "passivity".

3.3 The loaded transformer

In a loaded transformer the primary voltage fixes the core flux. When the secondary draws current, its ampere-turns (N₂I₂, also called magnetomotive force, mmf) would change the flux, so the primary draws extra current until N₁I₁ ≈ N₂I₂ and the flux is restored. That extra primary current is how the load's power is paid for. Sykes describes exactly this on p.21: when the transformer is loaded, "the Ampere Turns on the primary are equal but opposite to the Ampere Turns on the secondary" [read: p.21].

In circuit language the load appears at the primary as a reflected impedance. With primary (R₁, L₁), a secondary loop of total impedance Z₂ = R₂ + R_L + jX₂ (R₂ the coil resistance, R_L the load, X₂ the loop reactance) and mutual inductance M_x between them, the input impedance is

Z_in = R₁ + jωL₁ + (ωM_x)² / Z₂,

whose added real part is

R_reflected = (ωM_x)² (R₂ + R_L) / ((R₂ + R_L)² + X₂²).

R_reflected is positive whenever M_x ≠ 0 and current can flow. That positive resistance is Lenz drag in circuit form. (Derived symbolically in example.py, Part 2.)

3.4 Bucking coils in ordinary engineering

Coils wound or connected in opposition are common and useful [memory, high]:

  • A humbucking guitar pickup uses two coils wound oppositely with opposite magnet polarity: the string signal adds and the mains hum (same in both coils) cancels.
  • Non-inductive wirewound resistors wind half the wire each way so the inductance nearly vanishes.
  • Common-mode chokes, buck-boost transformers and degaussing coils all rely on fields that add or cancel by winding direction.

The rule is simple: two coils in series add their emfs if the flux change through them has the same sign relative to their winding sense, and cancel if not. Whatever makes the emfs add also makes their reaction fields add on the core they share.

3.5 Generator efficiency

Large generators convert shaft power to electrical power with efficiencies of about 95-98% [web: https://eureka.patsnap.com/report-comparison-of-efficiency-between-synchronous-generator-and-asynchronous-generator, search summary read by the arbiter]. The missing 2-5% is loss (copper heat, iron loss, friction, windage). The 95-98% that is converted is the useful output.

4. Where the inventor departs

The real physics under Sykes's starting points is sound, and he says much of it correctly. The departures are at specific steps.

Topic Textbook Sykes Where it holds or fails
Lenz's law and current No current, no reaction (3.1). "Lenz's law is the Result of drawing Current from the already Separated Charge, not the cause!" [read: p.10]; with the secondary open "Lenz's Law is not in effect" [read: p.21]. Holds. This is the textbook statement in different words (arbiter: Established).
Can the reaction be avoided while current flows? No: the reaction force times speed is the delivered power (P = e i = τω). "If we were able to turn off the Magnetic, but at the same time keep the Electric components, there would be no Reflected Power on the primary" [read: p.21]; "Mechanical Energy is NOT the Source of the Energy Generation!" [read: p.8]. Fails. Sykes himself says "A Moving Charge constitutes a Magnetic Field" [read: p.10]. Remove the reaction and you remove the delivered power (arbiter: Refuted by calculation).
Bucking output coils Emfs add only where the reaction fields also add (3.4); reflected resistance (ωM_x)²Re(1/Z₂) > 0. Opposite-wound "partnered" coils cancel the magnetic component while "the electric field component adds" [read: pp.40-42], so no drag reaches the input. The weak form (a bucking pair cancels external field and can reduce coupling) is ordinary engineering. The strong form (load power with no reaction, or net gain) fails by reciprocity and passivity. See section 5.
Step-up ratio and current Ideal transformer: I₂ = I₁N₁/N₂. A 1:5 step-up with 1 V in gives 5 V out, and a 1 A load needs 5 A at the primary: 5 W in, 5 W out [calc; example.py Part 4]. States the ideal rule correctly on p.10. Then: "Ideally, these devices work with a Step Up configuration. 1:3 is a common Ratio ... Current, as you may think, is not stepped down as a result of the Voltage being stepped up"; "Floyd Sweet used a Step Up configuration of 1:5" [read: p.11]. Fails if read plainly: secondary current that is not reduced by the turns ratio means more power out than in (arbiter: Refuted by calculation; the wording is ambiguous and no measurement is attached). The "1:5" in the Sweet drawing is FB1, a 48-turn feedback winding, to P1, 240 turns, not drive to output. The drawing is a 120 V, 60 Hz VTA of unstated origin. Sweet's 1986 notes describe a different device: a 250-turn excitation coil and a 12-turn output coil, "90 Degrees Uncoupled" [read: deep read pp.10-13; lab-notes transcription].
Potentials A is a potential with B = ∇×A; φ is the scalar potential in volts; Φ is flux in webers. "A, the Electric Component of the Magnetic Field and, Phi, the Magnetic Component" [read: pp.4, 6]; "Phi" also used for flux [read: page image p.46]. One symbol is used for three things; "A is the electric component of B" is not a statement the equations can express (arbiter: Internally inconsistent as vocabulary).
emf = Bvl l is a length. "l = typically the Area of the Element" [read: page image p.9]. Unit error: B v × area has units T·m³/s, not volts. The same kind of slip occurs elsewhere. "Reflected power = −dΦ/dt" (p.21) has the units of a voltage, not a power. p.4 attributes field decay to an "I²R factor", which is a power loss, not a decay law. p.9 defines velocity as "Speed * Distance Travelled" [read: deep reads pp.4-9, 21].
Bloch wall Nanometre-scale boundary between domains inside a magnet. A macroscopic region "between the bucking coils" that "moves in time with the Field" [read: p.23]. A different object under the same name, with no equation (arbiter: Not testable as stated).
The coil experiment τ = L/R. Same coil, windings aiding vs opposing: "This is an expected behaviour ... when the Inductance Changes ... This proves that the Magnetic Field Slows the Charge Flow Rates dramatically" [read: p.16]. The first sentence is the textbook account. The second renames it; both views predict the same curve. See 5.1.
Input falls as output rises In a resonant loaded transformer, input current can fall when a load is connected, while P_out < P_in (section 5.3). The output coils "feed back", input impedance "is increased dramatically ... the input power is dramatically lowered" [read: p.46]. The current effect is real in a passive model. The power claim does not follow; current is not power. Sykes's own analogy (load = a second power source) shows why (5.4).
Generator efficiency 95-98% converted, 2-5% lost. "100% of Rotational Shaft Energy, with a 1-2% loss in friction/heating ... and the other 98-99% are the effects of Lenz Law creating Drag" [read: p.23]. The converted output is relabelled as drag. Removing it removes the output (arbiter: Internally inconsistent; Refuted by calculation).

On energy itself the document says both things. "There was no new form of Energy" [read: p.47] and the alternator test is "Not Free Energy" [read: p.48]; in a 2015 forum reply he writes "Energy can not be created or destroyed. You and I agree with Science on this one!" [web: Wayback copy of overunity.com topic 15395, via research/2026-10-04-overunity.com-wayback.md]. But also: "if Lenz's Law could be by passed, then every Generator on the planet could be a perpetual motion machine of the first kind" and Lenz's law is "a law that is not valid under certain conditions" [read: p.22], "Free Energy Future is here NOW" [read: p.49], and in a 2020 builders' recipe "Input can become Negative!" [web: research/2026-10-04-overunity.com-wayback.md]. His resolution is that electromagnetic systems are "Open Systems" connected to the "Aether" [read: pp.46-47], with no stated energy source or measurable prediction.

5. Worked example

Everything here is reproduced in example.py:

.venv/bin/python review/Sykes-Chris/example.py

Part 5.1 uses Sykes's own coil data. Parts 5.2-5.4 use a model of his three-coil arrangement with illustrative component values (the ones the mathematician and arbiter used), not values from any Sykes build.

5.1 The coil experiment (pp.14-16): what τ = L/R predicts

Sykes's data [read: pp.15-16; page images pp.14-15 via the deep read]: two windings of 55.6 mH each, 3.8 Ω each; in series aiding 0.231 H and 7.8 Ω; in series opposing ("BiFilar Wound Non Inductive Mode") 0.24 mH and 7.8 Ω; current measured across a 1 Ω sense resistor. Experiment 1 ran at 15.4 V and about 6 Hz; Experiment 2 at 7.0 V and about 650 Hz.

Step 1. Total resistance R = 7.8 + 1 = 8.8 Ω (switch and supply resistance neglected).

Step 2. Experiment 1: τ = 0.231 / 8.8 = 0.02625 s = 26.3 ms. Three time constants (95% of final current) = 78.8 ms. Final current V/R = 15.4 / 8.8 = 1.75 A. The trace shows τ ≈ 22-27 ms and a plateau ≈ 1.6 A [read: deep read, page image p.14]. Sykes's "75 ms" is about 3τ. This is a clean textbook RL curve.

Step 3. Experiment 2: τ = 0.24×10⁻³ / 8.8 = 27 µs; 3τ = 82 µs; V/R = 0.80 A. Sykes quotes "650 µs". That is 24 times the L/R time and close to the pulse width the generator imposed at 650 Hz (deep read: ≈ 640 µs).

Step 4. Ratio test. If both quoted times were L/R times, their ratio would equal the inductance ratio. 0.231 / 0.00024 = 962; 75 / 0.65 = 115. They are not both L/R times.

Step 5. What limited Experiment 2? The trace rises roughly linearly to about 0.7 A in about 0.4 ms. For a linear rise, L_eff ≈ V/(di/dt) = 7.0 / (0.7 / 0.0004) = 4.0 mH, which is 16.7 times the quoted 0.24 mH. Something other than the quoted inductance set that curve. Candidates are the drive circuit, core behaviour at 0.7 A, and the pulse width; the document does not say which. The drive circuit is two MOSFETs (metal-oxide-semiconductor field-effect transistors, used here as electronic on/off switches). As drawn, the upper one is wired in a way that would limit how hard it can switch on. The measured plateau current in Experiment 1 suggests the bench wiring differed from the drawing [read: deep read, page image p.14].

Step 6. Coupling check. For two identical windings, series aiding gives L_s = 2L(1 + k) and series opposing 2L(1 − k). Opposing: 0.24 mH = 111.2 mH × (1 − k) gives k = 0.998 (very tight coupling). Aiding: 0.231 H = 111.2 mH × (1 + k) gives k = 1.077, which is impossible (k ≤ 1). The ceiling is 4 × 55.6 = 222.4 mH; 231 mH is 4% above it, which suggests the inductance readings were taken under different or uncontrolled conditions (instrument and test frequency not stated).

What the example teaches: "the magnetic field slows the current" is a fair plain-English description of inductance. It is not evidence of an extra "friction on charge", because the textbook predicts the same curve. The steady currents (V/R with the same 7.8 Ω) show no change of resistance, so Floyd Sweet's quoted claim that the same wire "would carry 10 times more current at the same temperature" [read: p.17] is not tested by these runs.

5.2 The three-coil circuit: does opposite winding remove the reflection?

Model: a primary L₁ = 1 mH, R₁ = 0.5 Ω, driven by a 1 V peak sine wave at 400 Hz (ω = 2π × 400 = 2513 rad/s). Two output coils, 10 mH each, with mutual inductance −0.5 mH between them (they buck), total coil resistance 0.4 Ω, in series with a load R_L. The loop inductance is L_loop = 10 + 10 + 2(−0.5) = 19 mH. M_x is the net mutual inductance from the primary to the series pair: the signed sum of the two individual mutuals.

Case A: both output coils see the same flux change and are wound oppositely. Their emfs cancel, so M_x = 0. Then the reflected impedance (ωM_x)²/Z₂ = 0, the loop current is zero and the load gets nothing. The primary draws |I₁| = 1 / |0.5 + j2.513| = 1/2.562 = 0.390 A and takes P_in = ½ × 0.390² × 0.5 = 0.038 W, the same as with no output coils. This is the humbucking case: perfect cancellation, and also no output.

Case B: the output coils sit where the flux changes in opposite directions (the geometry Sykes describes, "half the conductors" seeing the opposite change [read: p.8]), wound oppositely, so their emfs add: M_x = 4 mH (illustrative). Take R_L = 10 Ω.

Step 1. ωM_x = 2513 × 0.004 = 10.05 Ω; (ωM_x)² = 101.1 Ω².

Step 2. Loop impedance Z₂ = (0.4 + 10) + j(2513 × 0.019) = 10.4 + j47.75 Ω; |Z₂|² = 108.2 + 2280 = 2388 Ω².

Step 3. Reflected impedance = 101.1 / (10.4 + j47.75) = 101.1 × (10.4 − j47.75) / 2388 = 0.440 − j2.021 Ω. Its real part, 0.440 Ω, is positive: this is the load showing up at the primary.

Step 4. Z_in = 0.5 + j2.513 + 0.440 − j2.021 = 0.940 + j0.492 Ω; |Z_in| = 1.061 Ω; |I₁| = 1/1.061 = 0.942 A (up from 0.390 A with no load).

Step 5. P_in = ½ × 0.942² × 0.940 = 0.417 W. Loop current |I₂| = ωM_x |I₁| / |Z₂| = 10.05 × 0.942 / 48.87 = 0.194 A. P_load = ½ × 0.194² × 10 = 0.188 W.

Step 6. Check the books: primary heat ½ × 0.942² × 0.5 = 0.222 W; output-coil heat ½ × 0.194² × 0.4 = 0.008 W; 0.222 + 0.008 + 0.188 = 0.417 W = P_in. The script asserts this balance for every case.

Load sweep (no capacitor) [calc]:

| R_L (Ω) | |I₁| (A) | P_in (W) | P_load (W) | P_load / P_in | |---|---|---|---|---| | open | 0.390 | 0.038 | 0 | – | | 0.5 | 1.491 | 0.601 | 0.025 | 0.04 | | 2 | 1.375 | 0.573 | 0.084 | 0.15 | | 10 | 0.942 | 0.417 | 0.188 | 0.45 | | 50 | 0.461 | 0.165 | 0.111 | 0.67 |

The bucking arrangement did not remove the reflection. Whenever the opposite windings give output (Case B), the load reflects into the primary; when they cancel the reflection (Case A), they also cancel the output. That is reciprocity at work.

5.3 Resonating the output loop: input current falls, power does not "come free"

Now add a capacitor in the output loop to cancel its reactance at 400 Hz: C = 1/(ω² L_loop) = 1/(2513² × 0.019) = 8.33 µF. With X₂ = 0 the reflected impedance becomes purely resistive, (ωM_x)²/(R₂ + R_L).

For R_L = 10 Ω: reflected = 101.1 / 10.4 = 9.72 Ω; Z_in = 10.22 + j2.513 Ω; |Z_in| = 10.52 Ω; |I₁| = 0.095 A; P_in = ½ × 0.095² × 10.22 = 0.046 W; |I₂| = 10.05 × 0.095 / 10.4 = 0.092 A; P_load = ½ × 0.092² × 10 = 0.042 W.

| R_L (Ω) | |I₁| (A) | |I₂| (A) | P_in (W) | P_load (W) | P_load / P_in | |---|---|---|---|---|---| | open | 0.390 | 0 | 0.038 | 0 | – | | 50 | 0.282 | 0.056 | 0.0995 | 0.079 | 0.79 | | 10 | 0.095 | 0.092 | 0.0461 | 0.042 | 0.91 | | 2 | 0.023 | 0.098 | 0.0117 | 0.0096 | 0.82 |

Best ratio over R_L from 0.01 to 1000 Ω: 0.915, at R_L ≈ 9 Ω [calc].

Look at the R_L = 10 Ω row. Connecting the load made the input current fall from 0.390 A to 0.095 A, a factor of four, exactly the kind of effect Sykes reports ("This brings back the input current" [read: p.46]). Yet the input power rose from 0.038 W to 0.046 W. The reason: with no load the primary is mostly an inductor, its current is large but nearly 90° out of phase with the voltage, so little real power flows; the loaded, resonant loop makes the primary look resistive, so a smaller current carries more real power. An ammeter on the input, or V_RMS × I_RMS, would mislead you in both directions. Only the time average of v(t)·i(t) tells you the power.

Sykes also resonates the input. He writes: "I seem to get a better result if the Input is resonated for example, LC Tank ... There seems to be a time constant in the Magnetic Fields of the Partnered Secondary Coils that appear to play a factor in the Operating Frequency" [read: p.24]. He gives no circuit and no numbers. An LC tank is an inductor and a capacitor together, here at the input. Energy then swings back and forth between L and C, and the current drawn from the source can fall a long way. As in the table above, that changes the current, not the power balance. The arbiter rates "a resonated input gives a better result" Plausible, untested as stated; net gain stays Refuted by calculation.

5.4 Sykes's own analogy: a power source as the load

On p.46 Sykes explains the falling input this way: imagine the load is "another Power Source ... adjustable Power and Frequency so we can get a phase Lock on the input. Then we turn up the secondary and at a point, the input power is reduced simply because we have a source of power returning back to our Transformer" [read: p.46].

Put a second sinusoidal source V₂ in the resonant loop of 5.3 (R_L = 10 Ω), in phase with the loop current, and turn it up [calc]:

| |V₂| (V) | Power from source 1 (W) | Power from source 2 (W) | P_load (W) | Source 1 + source 2 (W) | |---|---|---|---|---|---| | 0 | 0.0461 | 0 | 0.0422 | 0.0461 | | 0.2 | 0.0380 | 0.0094 | 0.0441 | 0.0474 | | 0.5 | 0.0258 | 0.0242 | 0.0474 | 0.0500 | | 1.0 | 0.0055 | 0.0509 | 0.0541 | 0.0563 | | 1.5 | −0.0149 | 0.0801 | 0.0620 | 0.0652 |

Sykes's description is correct as far as it goes: the primary source's power falls, and at 1.5 V it even goes negative (power flows back into source 1). But the second source pays for all of it, and the total supplied always covers the load plus the coil heat. His analogy needs a real second source; in the bucking-coil device the "source" in the output loop is the output coils themselves, which are driven only by the primary. The analogy explains why input power can fall in a two-source system. It does not supply the missing source in a one-source one.

5.5 Smaller calculations

Generator "drag" (claim 6). A 10 kW generator at 1000 rpm: ω_shaft = 1000 × 2π / 60 = 104.7 rad/s, so the torque needed is 10 000 / 104.7 = 95.5 N·m. At 98% efficiency the shaft supplies 10 000 / 0.98 = 10 204 W, of which 204 W is loss. The "98-99% drag" is the 10 kW itself: remove it and the 10 kW goes too [calc; arbiter].

Sweet's 1986 flux (claim 5). The transcribed notes give a 12-turn output coil of 1.5 in diameter delivering 24.2 V at about 400 Hz [read: research/files/hyiq-Lab-Notes-Transcribed-Floyd-Sweet.pdf, via the Investigator's notes]. For a sinusoid, V_RMS = 4.44 f N Φ_peak, so Φ_peak = 24.2 / (4.44 × 400 × 12) = 1.14 mWb. The coil's bore area is π × (0.01905 m)² = 1.14×10⁻³ m², so the flux density must swing by about ±1.0 T through the coil, if all the flux passes inside the bore. Ceramic (ferrite) magnets of the common grade Y30 have a remanence (the flux density a magnet keeps on its own) of about 0.37-0.40 T [web: https://www.stanfordmagnets.com/magnetic-properties-of-sintered-ferrite-magnets.html and https://www.first4magnets.com/us/ceramic-magnet-information/grades-of-ceramic-magnets, search summaries checked by the arbiter]. Walt Rosenthal, as quoted in the rexresearch.com file, gives Sweet's magnets as ferrite "grade 5 or grade 8" [read: research/2026-10-04-rexresearch.com.md]; those grades were not checked. The notes give no core geometry, so this is a question to ask, not a disproof [calc, example.py Part 4; arbiter]. Note also that this 1986 device has a single output coil mounted at 90° to the excitation coil "so that the output Currents do not couple with the excitation coil". It is not a bucking pair [read: lab-notes transcription, via the deep read pp.10-13].

Sweet's two stated inputs (claim 5) [read: lab-notes transcription; calc, example.py Part 4]. On 24 May 1986 the notes give 0.023 mW in and 19.1544 W out: a ratio of about 832,800, i.e. 83,280,000% (the notes write "83,260,000%", a rounding difference). On 4 June they give "143µa, 7v ... 1mw": 7 V × 143 µA = 1.001 mW, and 111.32 W / 1.001 mW ≈ 111,000. The two inputs differ by a factor of about 43. The implied drive-coil impedance, 7 V / 143 µA ≈ 49 kΩ, agrees only to within an order of magnitude with the notes' description of the drive coil as "hundreds of thousands of ohm".

6. Evidence table

Verdicts are the arbiter's. "Tested?" means a measurement exists on record, of any quality.

Claim Tested? Result Replicated? Verdict (arbiter)
1. Opposite-wound partnered output coils cancel B, "A adds", so no Lenz drag on the input By Sykes, informally (videos 2010-11 and later; 2014: "Input is 1 Watt (5V 200ma) ... Globe ... rated at 12V 300ma (3.6 Watt) ... slightly more Output" [web: hyiq.org, research/2026-10-04-hyiq.org.md]; a rated nameplate figure, no method stated). By forum replicators. Negative reports: dewetw (aboveunity.com, 30 Oct 2019): three-coil build on an N87 ferrite core (see section 2), sawtooth reproduced, "never saw any result where I am getting more than about 0.82 COP" (one forum post with photos; Sykes did not dispute the figure). patrick1: "successful of course, but thats because it isnt OU". Itsu (2020): called "failed" by Sykes; Itsu's data are behind a login and known only through Sykes's account. Positive reports, all from interested parties and none with power data seen: Sykes (aboveunity.com, 2025): "Marcel [FringeIdeas] has succeeded, but now has the refinement phase" and "I have shown you how to get an output, how to Generate Energy in excess of the Input" (in the same post: "we have not succeeded, in general"); the forum's "Our Above Unity Machines" list (Baerndorfer's Experiment, Captainloz's A-R-E, Aetherholic's COP>4.0 and others); the 2019 thread is tagged ABOVE-UNITY [read: research/2026-10-04-aboveunity.com.md]. No neutral, fully metered test on record. Neither the positive nor the negative informal reports include calibrated power-in/power-out data. Weak form (bucking pair and loose coupling reduce reflected load and external field): Plausible, untested as a quantified claim. Strong form (no reaction, or net gain): Refuted by calculation; the informal replications with numbers point the same way but are not a neutral test. A/Phi vocabulary: Internally inconsistent. "Current is not stepped down" by a step-up (p.11), read plainly: Refuted by calculation; wording ambiguous.
2. Lenz's law is the result of drawing current, not its cause Every generator ever measured (weak form). Sykes's p.45 rotor test reports "No or next to no Lenz's Law" with no numbers [read: p.45, text layer only; no page-image deep read of pp.45-49 yet]. Weak form confirmed universally. Strong form: no numbers. – Weak: Established. Strong (power delivered with no reaction): Refuted by calculation; bare wording alone Not testable as stated.
3. The magnetic field slows charge flow (75 ms vs 650 µs) Yes, by Sykes (pp.14-16). Exp. 1 matches τ = L/R; Exp. 2's 650 µs is about the pulse width, 24 times the L/R time; supply, drive and frequency all changed between runs. No. Observation: Established. The quoted times are not reproducible from the traces. Strong reading ("friction on charge", seen as heat, p.17): Never properly tested; this experiment cannot tell it apart from inductance (arbiter's QE revision, replacing "Not testable as stated"). Sweet's duty factor (10x current at same temperature): Never properly tested.
4. Conditions (loose coupling, Bloch wall, step-up ratio, diode and square wave at the BH knee, enough load); input falls as output rises; alternator test Sykes (2019 three-coil build, 12 V 400 mA input, output "some data off screen"); replicators as in row 1. Loose versus tight coupling: one informal before/after comparison by Sykes. He rebuilt a gapped device with the laminations "tightly coupled ... like a transformer" and "was not able to make the device work the same" [read: p.32]. No numbers, no definition of "work"; a video is referred to but not in the PDF. Alternator: self-excitation demonstrated, "Not Free Energy" [read: p.48]. Measurement basis fragile: on Sykes's own measurements thread, RMS-based power of 59.5 W vs a scope mean of 17.1 W on the same 0.25 Ω shunt [web: research/files/aboveunity-measurements.txt]. The beginner circuit Sykes prescribes (p.38) is a sine drive, an input coil and two bucking output coils in series with a 10 Ω load, with no diode. The diode and square-wave scheme is "a work in progress" that "only appears to be valid under some configurations" [read: deep read pp.38-42]. Partzman (2020): schematic diodes reversed relative to the scope traces; Sykes replied "My Diagram is correct"; which circuit Partzman meant is not established. Walt Rosenthal, quoted on p.46: "The VTA 'likes' to always see a minimum load of 25 watts" (a stated condition, untested). Untested prediction (pp.47-48): with a big coil, a magnet and a lamp, the lamp normally lights after a delay set by the coil's impedance, but "in this technology" the eye sees it light "nearly immediately"; judged by eye only, no numbers. Sykes resonates the input with an LC tank for "a better result" [read: p.24]. "COP 1.7" is attributed to Sykes by MileHigh (2015); Sykes's own statement of it was not found. Sykes: "For every one device that works, expect 10 devices that don't" [read: p.34]. No neutral test. Alternator: Established (with mechanical input). Input current falls under resonant load: Plausible, untested as stated (reproduced in 5.3 with P_out < P_in). Net gain: Refuted by calculation (passive metered circuit), never neutrally measured. Conditions (gap, diode, square wave, minimum load): Never properly tested. One informal loose/tight comparison by Sykes exists, with no numbers. The diode is not part of his beginner circuit. Swipe-lamp prediction and 25 W minimum load: Never properly tested. Bloch wall and knee: Not testable as stated.
5. Sweet's VTA and the devices of Moray, Meyer, Jensen, Koenig, Frolov, Melnichenko and others share this principle Sweet: 1986 lab notes (transcribed by Sykes from scans from "an anonymous source", pages missing), demonstrated 24 May 1986 with witnesses Mark Goldes and Darryl Roberts. The notes record: "4.6A x 24.2V = 111.32 Watts" into lamps at 388-402 Hz, "measured twice, observed by 3 persons", "It seems to be affected by the quality of the Sine Wave, negatively by harmonic distortion"; "19.1544w / 0.023mw = 83,260,000% Gains" (24 May); "Oscillator put 143µa, 7v ... approximately 400hz 1mw into the Excitation Coils" and "105,000 x - 111,000 x Gain" (4 June); a 10-hour run "with no noticeable heating"; Beckman and Keithley meters; "a Wavetek Signal Generator, with line power, producing a sine wave, vs. Tom's battery supplied square wave (20w?) oscillator" [read: lab-notes transcription]. Bearden's account (preserved on rexresearch.com) of a weight-reduction test: a load box of parallel 100 W lamps fed with "120 volt, negative AC sine-wave power at 60 Hz", weight recorded for each 100 W step. That is an interested party's account, and no data tables were seen [read: research/2026-10-04-rexresearch.com.md]. Bearden reports about 6 W from an earlier Sweet unit [read: p.19]; Sykes quotes 5000 W for "Generation Six" with no source [read: p.13]. Other devices: a drawing or photo and one remark each (pp.25-31); the only performance statements are the captions on Frolov's figure ("A Toroidal Over-Unity Generator") and Alek's ("EFFICIENCY >> 100%"), Moray's lamp lit after about 10 minutes of tuning, and a Meyer video remark [read: deep read pp.25-31]. Arithmetic is internally correct. Gaps: V × I at 400 Hz into a highly reactive drive coil is apparent, not real, power unless phase is measured; the notes themselves state two input figures (0.023 mW on 24 May, 1 mW on 4 June) about 43 times apart, without explanation (5.5); the generator was mains-powered and its draw (bench units of that class about 15-25 W [web: arbiter's checks]) was not counted; the flux needed (5.5) is large for ceramic magnets. The notes also record a part Sweet withheld ("there is a little part in the Generator that he has withheld from disclosure ... (he calls this unit a 'Vector Oscillator')"), so the notes alone do not specify the device. The 1986 device has one output coil at 90° to the excitation coil, not a bucking pair. Koenig US4806834 shows mirror-image coils, a grant, not a measured gain [web: https://patents.google.com/patent/US4806834A/en]. No independent laboratory test of any listed device is on record. Sweet VTA and each other device: Never properly tested (not dismissed without testing either). Weight-reduction account: Never properly tested. Unifying principle and Bearden's wave/anti-wave language: Not testable as stated, and weaker for the 1986 data, whose device is not a bucking pair.
6. 98-99% of generator energy goes to Lenz drag; remove it for efficiency; "open system connected to the Aether" Calculation only. The converted output is counted as loss (5.5). – Internally inconsistent; Refuted by calculation. Aether and "Free Energy Future": Not testable as stated, and contradicted by the author's own "no new form of Energy".

Kept in the record, whatever the verdicts (from arbiter.md and the Investigator):

  • Sykes's good practice: full coil data and photographs; he calls the coil result "expected behaviour" [read: p.16]; he set the rule "Any COP > 1 Claims must be accompanied by Measurements" and told users that the scope mean, not RMS, gives power; he published dewetw's 0.82 on his own forum; he says the work "Can not be Patented" and shares it openly; he located and posted Sweet's primary papers.
  • His own admissions: one can have "Asymmetrical Regauging and still be Below-Unity"; "we have not succeeded, in general" (aboveunity.com) [web: research/2026-10-04-aboveunity.com.md].
  • MileHigh (16 Jan 2015) asked for scope shots, load resistance and error bounds for the "alleged COP 1.7 device"; no data reply was found in the part of the thread read (page 1 of a thread that ran to 514+ pages and 1.14 million reads by January 2017).
  • Tinman, quoted second-hand by Sykes: "a bucking coil arrangement dose indeed add energy to a system" (mechanical devices; original not read).
  • Sweet's own theory text "Nothing is Something" (24 Jun 1988) contains E = (B × V) + (−B × −V) = 2(B × V). The algebra is correct and trivial: flipping both signs in a cross product changes nothing.
  • Sykes rejects Bearden's account of Sweet's "magnet conditioning" ("simply wrong" [read: p.17]; "nonsense", 2025), so the two main promoters of the VTA disagree about how it worked.
  • Positive informal reports (interested parties, no power data seen): Sykes's 2025 "Marcel [FringeIdeas] has succeeded" and "I have shown you how to get an output, how to Generate Energy in excess of the Input"; the aboveunity.com "Our Above Unity Machines" list (Aetherholic's COP>4.0, Captainloz's A-R-E, Baerndorfer's Experiment); the ABOVE-UNITY tag on the 2019 thread [read: research/2026-10-04-aboveunity.com.md]. They stand beside the negative reports in row 1 of the table.
  • Sweet's claimed origin of the effect. Sykes quotes Jeane Manning: "On rare occasions, Sweet saw this effect, called self-oscillation, occur in electric transformers". He also quotes Bearden: "Sweet was also a transformer designer and expert, and he remarked that he had also observed specialized self-oscillation in certain transformers" [read: p.12]. Both are second-hand accounts by promoters, with no measurement.
  • The withheld part. Sweet's notes (transcription p.46) record: "Checks still not in – 13 days late – Sparky explains that there is a little part in the Generator that he has withheld from disclosure – A $1.50 part from Radio Shack ... that is not a capacitor ... that puts a small µa current into the Power Coil as well as the Excitation Coil ... (he calls this unit a 'Vector Oscillator')" [read: lab-notes transcription]. Two readings are open. For a rebuild, a copy built "as the notes describe" may lack a part, so its failure would not refute Sweet. For a sceptic, an undisclosed part in a metered circuit is exactly where an unmetered power source could sit. "Checks still not in" also records payments owed to Sweet at the time; it belongs with the interested-party record (Sweet, Goldes, Roberts, Bearden).
  • Drive sources in the notes. The signal generator ran on mains power, and a separate battery square-wave oscillator of about 20 W ("20w?", a question mark in the notes) is mentioned [read: lab-notes transcription p.5]. Neither was metered.
  • Bearden's weight-reduction account (100 W lamp steps, "negative AC sine-wave power"), kept as a claimed anomalous result that a weighing test could check cheaply.
  • The full lineage Sykes cites. In the device survey (pp.25-33) and the credits (p.49): Clemente Figuera, T. Henry Moray, Lester Hendershot (who used a bell ringer, pp.43-44), Raymond Kromrey, Stanley Meyer, Paul Raymond Jensen, Floyd Sweet, Earl Koenig, Gennady Markov (bidirectional current transformer), Alexander Frolov, Andrey Melnichenko, Vladimir Utkin, sergdo, Alexander Mishin, Mishail1971 (NE555/IRF640 sketch dated 2011), Bill Alek (SmartPAK), Tom Bearden, John Bedini. Sykes also notes that Melnichenko's windings are "not reversed but results are still possible" (p.29). W. J. Hooper is cited through Bearden, who calls him "the only physicist who ever methodically investigated EM stresses ... in the laboratory" (p.19) [read]. No review followed Hooper up.
  • Related-work leads from the Investigator, not checked by the reviewers: the Wiegand-wire field-cooling analogue (US3820090) and other industrial analogues raised by the forum analyst Anor; LordVëber's US4210859A; George I. Cohn on induction (cited by Sykes on p.9); and the later lineage Sykes himself cites after 2016 (Kapanadze, Akula, Don Smith, Gunderson, and the Motionless Electromagnetic Generator (MEG) "asymmetrical regauging").
  • The Aharonov-Bohm effect shows the vector potential A is physical: it shifts the phase of electrons passing where B = 0 [web: https://www.hitachi.com/rd/research/materials/quantum/aharonov-bohm/index.html, cited by the contrarian]. That supports "A is not mere bookkeeping"; nobody has derived load power without reaction from it.

Where the record goes dark (arbiter's absence-of-evidence analysis, condensed):

  • overunity.com, home of Sykes's 514+ page thread (topic 15395), is now a domain-for-sale page; its mirror overunityarchives.com is "Now Open for Paid Subscribed Members Only"; overunityresearch.com (Itsu's build) needs a login. The evidence fits commercial or hosting decay and the general move of the forum scene to other sites; no evidence of deliberate removal was found. Deciding records: a Wayback CDX listing of the topics, and Itsu's own posts or videos.
  • Sweet's notes exist only as Sykes's transcription of scans of unknown provenance. Explanations: ordinary loss of a hobbyist's papers; selective release; alteration (no evidence of it found). The notes themselves give one documented reason why they cannot reproduce the device: the withheld "Vector Oscillator" part. Deciding records: the scans and the original lab books (Goldes's and Roberts's papers, Magnetic Power Inc. files).
  • Sweet never had the VTA independently certified; Bearden, an interested promoter, says Sweet feared his life would be "snuffed out" if tested. A search summary gives Sweet's death on 5 July 1995 of a heart attack at 83 [web: search summary, arbiter's check]. Explanations: lack of funds and secrecy (defenders), fear of exposure of an artefact (critics), suppression (Bearden). Evidence for suppression is Bearden's recollection alone.
  • aboveunity.com says "We have been DEPLATFORMED by everyone" and "PayPal De-Platformed me! They REFUSE to tell me why!" This is Sykes's own account (an interested party); no second source was found. The processor's notices would decide it.
  • The most telling absence: no power-in/power-out table for any Sykes device, in the Guidelines or anywhere the Investigator reached.

7. What a learner should take away

Solid. - Lenz's reaction appears only when current is drawn. Sykes states this correctly, and the transformer description on p.21 (primary and secondary ampere-turns balance under load) is textbook. - Bucking coils are real engineering: they cancel external fields and common-mode signals, and they can almost eliminate inductance (Experiment 2, k = 0.998). - The coil experiment is a clean demonstration of τ = L/R. The curve in Experiment 1 matches the textbook to within the accuracy of reading a scope screen. - Input current can fall when a load is connected, in a purely passive circuit (5.3). This is worth knowing: it explains why builders see a falling input ammeter and think something unusual is happening. - Sykes's measurement advice (mean of the product, not RMS; COP claims need measurements) is correct.

Wrong, and why it is instructive. - Reciprocity: the same mutual inductance that lets the primary drive the output lets the output push back. A geometry that cancels the push-back also cancels the drive (5.2, Case A vs Case B). This is the single most useful idea in the project. - Current is not power. The 5.3 table shows input current falling by a factor of four while input power rises. Many over-unity claims rest on this confusion, or on its cousin, V_RMS × I_RMS on a reactive or pulsed load (Sykes's own 59.5 W vs 17.1 W example). - Sykes's p.46 analogy (load = a second power source) is right about two-source circuits and silent about where the second source's power comes from (5.4). - Calling the converted 98% of a generator's output "drag" is a bookkeeping error: the drag is the conversion. - "Current is not stepped down when voltage is stepped up" (p.11) contradicts the ideal-transformer rule Sykes states correctly one page earlier. If secondary current is not reduced by the turns ratio, output power exceeds input power. Check any such claim with V₁I₁ = V₂I₂ (example.py Part 4). - A list of necessary conditions (gap, ratio, diode, frequency, minimum load, winding direction), together with "for every one device that works, expect 10 devices that don't" (p.34), lets any failure be put down to a missing condition. The claim then becomes hard to falsify. A claim gets testable when its conditions are written down in advance. - Renaming is not explaining. "The magnetic field slows charge flow" and "inductance" predict the same curve; to claim a new effect you need a measurement where they differ (for example, steady-state heating at equal current).

Genuinely open. - No Sykes build, and no Sweet VTA, has had a complete, calibrated, time-aligned power measurement with calorimetry. The arbiter's verdict on net gain rests on calculation, and the calculation is sound for a fully metered passive circuit; but the experimental record is informal on both sides. - What limited the current rise in Experiment 2 (4 mH effective vs 0.24 mH quoted), and why the aiding inductance exceeds its physical ceiling. - What Sweet's two stated input figures (0.023 mW and 1 mW) measured, what the drive coil's real power and power factor were, and where the original scans are. - What Itsu measured. - Which device the 1986 witnesses saw (a single output coil at 90°, as the notes describe, or a bucking pair), and where the p.11 schematic (2000-turn excitation coils, 240-turn bifilar P, 48-turn FB, 120 V 60 Hz) comes from. - What Sweet's withheld "Vector Oscillator" did, and whether it was metered. - What Sykes saw when the loosely coupled device stopped "working the same" after he coupled it tightly, and whether his video of it survives. - Pages 45-49 (the rotor, swipe-lamp and alternator experiments) still need a page-image deep read before the statements about them count as fully checked. - Whether the "diode, knee, loose coupling" recipe reduces reflected reactive power at a fixed load by a useful amount, even with efficiency below 1. That is the part of the project a builder could use without any free-energy claim.

What would settle the main question (arbiter): a battery-powered build of the three-coil circuit with a differential probe and time-aligned averaging of v(t)·i(t) at every port, plus calorimetry of the whole assembly over several minutes, with the circuit and procedure agreed in advance with Sykes.

8. Check yourself

  1. Experiment 1 used a 0.231 H coil with 7.8 Ω, a 1 Ω sense resistor and a 15.4 V supply. What time constant and final current does the textbook predict, and how many time constants is Sykes's "75 ms"?
  2. In the model of 5.2, explain in one or two sentences why making the two output coils' emfs cancel (Case A) also removes the reflection, and why making them add (Case B) brings it back.
  3. In 5.3 the input current fell from 0.390 A to 0.095 A when a 10 Ω load was connected, yet the input power rose. How is that possible?
  4. A builder measures 5 V RMS and 2 A RMS at the input of a pulsed coil circuit and reports 10 W in. What is wrong with this, and what should be measured instead?
  5. Sykes writes that in the best generators 98-99% of the shaft energy is "Lenz Law creating Drag". For a 10 kW, 98%-efficient generator, how much shaft power is lost, and what happens to the output if the "drag" is removed?

9. Further reading

Mainstream: - D. J. Griffiths, Introduction to Electrodynamics: ch.7 (Faraday's law, inductance, energy in magnetic fields), ch.10 (potentials) [memory, high]. - R. P. Feynman, R. B. Leighton, M. Sands, The Feynman Lectures on Physics, Vol. II: ch.15 (the vector potential; the chapter Sykes cites on his hyiq.org page, not in the Guidelines), ch.16 (induced currents), ch.17 (the laws of induction, mutual inductance) [memory, high]; ch.37 (magnetic materials, domains) [memory, medium]. Free online at feynmanlectures.caltech.edu. - Any introductory circuits text (for example Nilsson and Riedel, Electric Circuits): sinusoidal steady state, real and reactive power, magnetically coupled circuits and reflected impedance. - C. Kittel, Introduction to Solid State Physics: ferromagnetic domains and Bloch walls. - On power measurement of non-sinusoidal waveforms: any power-electronics text's chapter on power factor and harmonics.

The inventor and related primary sources: - Chris Sykes, Guidelines to Bucking Coils, v2.7 (26.02.16): https://www.hyiq.org/Downloads/Guidelines%20to%20Bucking%20Coils.pdf [web]. Earlier v2.5 on academia.edu (not read). - Sykes's forums: aboveunity.com ("Non-Inductive Coil Experiment", 2019-2026; the Measurements thread); overunity.com topic 15395 (2015 on, partly on the Wayback Machine). - Floyd Sweet: "Lab Notes Transcribed" (1986 notes by Darryl Roberts, signed by Sweet, transcribed by hyiq.org); "Letter to Mark from Sparky"; "Nothing is Something" (24 Jun 1988), all at https://www.hyiq.org/Downloads/ [web]. - F. Sweet and T. E. Bearden, "Utilizing Scalar Electromagnetics to Tap Vacuum Energy", IECEC 1991 (as cited by the Investigator; not read). - Rex Research Sweet file, including the Bearden "hearsay" exchange: https://www.rexresearch.com/sweet/1nothing.htm [web]. - E. Koenig, US Patent 4,806,834: https://patents.google.com/patent/US4806834A/en [web]. - George I. Cohn, "Electromagnetic Induction" (cited by Sykes on p.9; PDF on hyiq.org /Downloads/, not read). - Related work Sykes or his forum cite (not read by this review): W. J. Hooper (via Bearden, p.19); the Investigator's leads listed under "Kept in the record".

Answers

  1. τ = L/R = 0.231 / 8.8 = 26.3 ms. Final current = 15.4 / 8.8 = 1.75 A. 75 ms / 26.3 ms ≈ 2.9, so about three time constants (95% of the final current).

  2. The output loop's emf and its reaction on the primary are both set by the same net mutual inductance M_x. If the two output coils' contributions cancel (M_x = 0), no emf drives the loop and nothing reflects; if they add (M_x ≠ 0), the loop current flows and reflects a positive resistance (ωM_x)²(R₂ + R_L)/|Z₂|² into the primary. You cannot have the first without the second, because M is reciprocal.

  3. With no load the primary is mostly inductive: 0.390 A flows but nearly 90° out of phase with the voltage, so real power is only ½ × 0.390² × 0.5 = 0.038 W. The loaded resonant loop reflects 9.72 Ω of resistance, making the input nearly resistive; 0.095 A in phase with the voltage carries ½ × 0.095² × 10.22 = 0.046 W. Real power depends on current, voltage and the phase between them, not on current alone.

  4. V_RMS × I_RMS is apparent power. It equals real power only for in-phase sinusoids into a resistor. For a pulsed or reactive load it can be much larger (or, with misaligned probes, smaller) than the real power. Measure the time average of the instantaneous product v(t)·i(t) over whole cycles, with time-aligned probes, at every port, including the supply that powers the drive.

  5. Shaft power = 10 000 / 0.98 = 10 204 W, so 204 W is lost. The other 10 000 W is the electrical output: the "drag" is the force through which the shaft does that work. Remove it and the output falls to zero; the most that cutting real losses could gain is the 2%.