Mathematician review: HiIQ-Youtube ("Partnered Output Coils", Chris Sykes / hyiq.org)

Role: PhD mathematician (Step 2). I checked derivations, notation, units and hidden assumptions. Source tags: [read] = I read it; [calc] = I computed it (script in the session scratchpad, NumPy; nothing was published); [memory] = recalled, with confidence; [web] = taken from the Investigator's notes, not re-fetched by me. No web calls were made. No verdict below depends on a medium or low confidence [memory] fact; the only recalled facts are the ideal-transformer relations and the passivity theorem, both high confidence.

Source status. The project folder holds one 143 s video. The Investigator identified it as Chris Sykes's (EMJunkie) "How Partnered Output Coils Work Together", and downloaded its written source, "Guidelines to Bucking Coils" (49 pp., v2.5 2015, v2.7 2016). I read the v2.7 text layer myself (requirements list, "Step Up/Down", the net-zero-field section, the bifilar experiments, the Lenz/Bearden sections). I did not view the figures or the scope shots, and I did not read overunity.com beyond what the Investigator saved. No input or output power figure for a partnered-coil device appears in the text I read. Conclusions about measurements therefore rest on the absence of numbers in the text layer, not on viewed figures.

What the video and Guidelines actually say [read]

Video narration (transcript): voltage on a primary gives a field; a loaded secondary gives an opposing field by Lenz's law; a "partnered" second secondary is added and "the vector of each magnetic field sum to zero"; each partnered output coil has 3x the primary's turns; 1:3 ratio; "Current is not stepped down as a result". On-screen caption (Investigator, from frames): "Feed Back and Feed Forward in the Partnered Output Coils can Increase the Seperation of Charge Carriers and Excess Output Energy can be obtained!"

Guidelines, the passage the video copies [read: v2.7, "Step Up/Down"]: the author first states the ideal transformer correctly ("240 Volts - 240 Turns ... 24 Volts - 24 Turns ... Input Power = Output Power = (240 Volts * 1 Amp) = (24 Volts * 10 Amps)"), then says: "these devices use some Transformer Principals, but not others. 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."

Guidelines on the field sum [read: v2.7, section on Lenz's law]: "Important: Here we can see that when the Transformer is loaded, there is a NET ZERO Magnetic Field. This means the Ampere Turns on the primary are equal but opposite to the Ampere Turns on the secondary, or close to it. This is Tom Bearden's Whittaker Wave/Anti Wave. Where the Vector of each Magnetic Field sum's to ZERO!!!" He then says that for his device it is the secondary pair that is to be cancelled, and that "this process can ONLY work, if Current is flowing in the Partnered Output Coils."

Notation and a symbol key (reader-facing)

• N1, N2: turns on primary and secondary. V, I: voltage and current (RMS, root-mean-square, unless stated). P: real (average) power, P = V I cos(phi), where phi is the phase angle between V and I.
• L: self-inductance; M: mutual inductance; k = M / sqrt(L1 L2): coupling coefficient, 0 to 1.
• MMF: magnetomotive force, N I (ampere-turns), the "magnetic push" a coil exerts on a core.
• Ideal transformer: V2/V1 = N2/N1 and I2/I1 = N1/N2, so V2 I2 = V1 I1.

Corrections to Step 1

1. Claim 4 as written in claims.md is wrong in its wording. "Separation of the input coil" is a transcription error; the on-screen text is "Separation of Charge Carriers and Excess Output Energy can be obtained". The Investigator is right (sites.md), and I accept it. The claim now has two parts: (i) feedback/feed forward "increases separation of charge carriers" (a term defined only by the author's "Induction is the Separation of Charge Carriers"); (ii) "excess output energy can be obtained". Part (ii) is an energy claim and is the more consequential one; Step 1 missed it.
2. facts.json says the opposing-field arrangement "is not in itself in conflict with standard electromagnetism". That is true but incomplete, and it hides the key point: by the author's own text the field sum already equals zero in an ordinary loaded transformer (ampere-turn balance, N1 I1 = N2 I2 apart from the small magnetising current). "The vectors sum to zero" is therefore not a feature that distinguishes the partnered arrangement from an ordinary one. See claim 1.
3. claims.md claim 1 describes "a primary coil with two partnered secondary coils". The video animation shows one primary and two secondaries, but the Guidelines require "one input or Excitation coil per Power Coil" (requirement 5) and the input coil to cover only half of the power coil (requirement 6). The arrangement in the text is two input coils (one per power coil), not one shared primary. The two descriptions are not the same circuit and the video does not say which it means.
4. facts.json: "I_out/I_in = N_in/N_out, so a 1:3 voltage step-up normally steps current down by about 3" is correct for one secondary. With two secondaries each at 3x the turns, on a common core, each carries one sixth of the primary current and the pair together one third (see the worked example). Step 1 did not distinguish. Minor.
5. The "Source gap" in claims.md is largely closed by the Investigator's identification; it remains for figures, scope traces and the 2015 threads (only page 1 of 350 read).
6. Step 1 did not note that the author states the ideal-transformer rule correctly in the very paragraph that contradicts it. That bears on how to read the claim (see claim 3).

Worked example: the energy balance with real numbers [calc]

I modelled the circuit as three magnetically coupled coils: primary L1 = 1 mH (N = 1 unit), two secondaries of 3 turns per primary turn, so L2 = L3 = 9 L1 (inductance scales as turns squared). Drive: 10 V peak at 1 kHz on the primary. Each secondary is closed on its own 10 ohm load. Coupling k = 0.99 (common core), solved as a linear AC circuit (Z I = V with Z = j omega L + R).

Reading the table: - Ratio of each secondary current to the primary current is 2.93/17.73 = 0.165, close to the ideal 1/6 = 0.167 (two outputs of 3N: ampere-turns N1 I1 = 3N1 (I2 + I3), so I2 = I3 = I1/6). Current is stepped down, voltage up, power equal. Step-up never leaves the current the same. - Reversing the winding sense of one independently loaded secondary changes nothing in the table: it only flips the sign of that coil's current. Direction of winding matters only when the outputs are joined (series or shared core), and then it changes which EMFs add, not how much power the source supplies. - For "current not stepped down" to hold with a 1:3 voltage ratio, I2 = I1 and V2 = 3 V1 per output. Output power per coil is then 3x input per coil; with two outputs, 6 x the share of input each pair accounts for. A device doing this would deliver 3x to 6x the power it takes in (COP, coefficient of performance, of 3 to 6) with the missing energy from an unstated source.

General check [calc, 20,000 random cases; memory: high for the theorem]: for any set of linear passive coils the inductance matrix is symmetric and positive definite (energy stored is positive). The complex power into the network is S = I^H (R + j omega L) I, whose real part is I^H R I >= 0: the real power in equals the power dissipated in the resistances (loads and wire). The inductive part contributes only reactive power. I generated 20,000 random positive-definite coupling sets (including negative couplings, i.e. any winding sense) with random loads: output power exceeded input power in none. This is a proof for the linear passive class, not a sample: no placement of coils, no field cancellation and no winding direction can change it. The result holds for any coil geometry because it uses only symmetry and positivity of L.

Scale check on the author's one quantitative experiment [calc]

The Guidelines' bifilar test [read: v2.7]: series (inductive) connection L = 0.231 H, R = 7.8 ohm; non-inductive connection residual L = 0.24 mH, same R. Charge times he measured: 75 ms and 650 microseconds. He concludes "This proves that the Magnetic Field Slows the Charge Flow Rates dramatically." - Time constant tau = L/R. Inductive: 0.231 / 7.8 = 29.6 ms; 2.3 tau = 68 ms (time to reach 90 percent), matching his 75 ms. - Non-inductive: 0.24e-3 / 7.8 = 30.8 microseconds; 2.3 tau = 71 microseconds, against his 650 microseconds. The ratio of times (75 ms / 0.65 ms = 115) is far smaller than the ratio of inductances (962), so something other than the coil (supply, instrument, wiring) set the fast case. - The slow case is the textbook inductor response (back-EMF from Faraday's law, EMF = -L dI/dt). It is not evidence that a magnetic field "slows charge", and it says nothing about Lenz cancellation or output power. It is consistent with established experiment and with no new physics.

Claim-by-claim review

Claim 1. Two partnered secondaries whose field vectors sum to zero make a transformer whose Lenz opposition is cancelled

a. Internal consistency. Split. - Weak form (fields from the primary and loaded secondary sum to zero at the core): true and standard. The Guidelines themselves say so for an ordinary loaded transformer (ampere-turn balance, quoted above). Mathematically N1 I1 + N2 I2 = (core reluctance) x (core flux), and the right-hand side is the small magnetising MMF. - Strong form (so the primary no longer "feels" the load): contradicts the model. The reflected load in a transformer is exactly the primary current that makes N1 I1 balance N2 I2; that balance is the cancellation. It is the reason the primary draws more current when loaded. The author writes that the reflected power "must be compensated" by increased input current (his words, Lenz section). Cancelling the Lenz field in the core by having a second coil's field oppose it leaves the second coil's own current opposed by its own back-reaction on the source (my table: input power equals output power in every row). The text is also internally inconsistent: the same page says the net-zero field is the ordinary loaded-transformer condition and that it is the Lenz effect, and then proposes the same net-zero as the cure. - The text concedes total cancellation "is also impossible" because the two wires cannot occupy the same space [read, per Investigator note; I did not locate that exact sentence]. - Notation: "the vector of each magnetic field sums to zero" mixes the field B (a vector at each point) with ampere-turns (a scalar per coil along a path). The sum rule holds for MMF around a closed path, not for B at every point; in a bucking pair B cancels only between the coils, not inside each. b. Consistent with experiment? The weak form is the well-tested transformer law. The strong form is contradicted by every measured transformer efficiency (large units are 98-99 percent efficient, a number the author himself quotes, as input minus loss, not as gain). Humbucking pickups and bucking (buck-boost) transformers are real, in routine use (the author cites them), and cancel noise or subtract voltage; none delivers more power than it takes in [memory: high]. c. Tested? By the author informally. Rotor test: "No or next to no Lenz's Law" with no numbers [read, per Investigator]. A third party (gratisenergi.se, snippet only) built a bucking-coil transformer and "found no Free Energy". Cyril Smith's 2015 "Bucking Coils produce Energy Gain" is title-only for me; I cannot weigh it. overunity.com thread 15395 (350 pages) is the public test record; I did not read beyond page 1. No controlled, neutral replication that I can read is on record. d. If true: Poynting's theorem and the symmetry of the inductance matrix would fail for a passive linear network, which would mean energy non-conservation, or a source of energy not in the circuit. Only an active/nonlinear/parametric mechanism (not shown) could do it. Verdict: split. Weak form (net zero MMF in a loaded transformer): Established. Strong form (the pairing removes the load's reflection and gives power gain in a passive linear coil system): Refuted by calculation (energy budget and the positive-definite inductance argument above). Whether a nonlinear/saturating or self-oscillating core could behave differently: Never properly tested (no measured input and output power on record that I could read); not dismissed without testing, since builders tried it.

Claim 2. Each partnered output coil at 3x the primary's turns steps the voltage up 1 to 3

a. Consistent: V2 = V1 N2/N1 for the open-circuit or lightly loaded ideal coupling. One caveat that the video does not state: with partial coupling (k < 1) and a load, V2/V1 falls below 3 (my loose-coupling case); and the Guidelines prefer the input coil over only half the power coil, which reduces coupling. b. Consistent with every transformer measurement. c. Tested for centuries (standard). d. Nothing changes. Verdict: Established (as the open-circuit turns ratio; for a loaded, loosely coupled coil the ratio is smaller).

Claim 3. Current is not stepped down as voltage is stepped up

a. Internally inconsistent with the author's own preceding paragraph, which states Input Power = Output Power (240 V x 1 A = 24 V x 10 A). With V2 = 3 V1 and I2 = I1 (unchanged), P2 = 3 P1 per output coil. His qualifier, "these devices use some Transformer Principals, but not others", is not given any equation. Requirement 4 in the document (P = V I, I = sqrt(P/R)) says "the Resistance of the power Coil must also be taken into account, according to normal theory. Here, not so" with no derivation. One reading that is mathematically fine: the output current is not stepped down relative to what the load would draw by Ohm's law, i.e. I2 = V2/R_load, since the secondary does not "know" the ratio. That is also standard transformer behaviour; but then the primary must supply the extra power, which the author asserts is not the case. b. Contradicts the general energy budget verified above for passive linear networks. His 2011 paper says of his first device: "The device I have shown is below unity" and "not an Over-Unity Device" [read, per Investigator]; the 2015 video and Guidelines say "excess output energy can be obtained". No measurement bridging those statements is in the text I read. c. Tested? Claim of gain: not with a controlled power measurement that I can read. Measurement pitfalls I can name from first principles [calc]: apparent power versus real power. A coil circuit with current lagging 84 degrees (power factor 0.1) drawing 1 A at 10 V RMS has apparent power 10 VA but real power 1 W, so multiplying RMS volts by RMS amps misreads the input by 10x in either direction depending on which side is measured; DC-supply measurements with ripple or switching have the same problem (the Investigator's lead, reply #3589 on "mistakes in measuring DC input power", is about this, but I did not read the reply). The Prezi quote of input:output "1 to 100" has no read context and cannot be assessed. d. If true: conservation of energy would need a hidden source (the author invokes the Dirac sea, Whittaker's scalar potentials and Sweet's VTA; none is quantified in the text), or Lenz's law (a statement of energy conservation for induction) would fail. Verdict: split. As a passive-linear-transformer statement (V up 3, I unchanged, no extra input): Refuted by calculation and Internally inconsistent (with the author's own 240 V/24 V paragraph). As the broader assertion that an unmeasured gain exists in some nonlinear or resonant mode: Never properly tested (no input/output measurement on record that I could read; third-party builds reported none; not shown to have been dismissed without testing, since it was built and discussed in public).

Claim 4. Feedback and feed forward in the partnered coils increase "separation of charge carriers", and excess output energy can be obtained

a. Not mathematically defined. "Separation of charge carriers" is explained by the Lorentz force F = q v x B (that is, the Lorentz force law, a standard physical law, [read] "Induction is the Separation of Charge Carriers"). That is a true description of motional EMF (Cohn 1949, which the author cites, derives EMF = B v l for a moving conductor). For a stationary transformer coil the EMF comes from the time-varying field via the induced electric field (Faraday's law, curl E = -dB/dt), not from a v x B term on moving conductors: Cohn's own paper separates the two cases, and the author applies the motional picture to a transformer without bridging them. The Guidelines' account of feedback: current in the partnered coils "can feed back via confined magnetic fields ... the Input Coil's Impedance is increased dramatically. This brings back the input current and as a result the input power is dramatically lowered" [read, per Investigator]. Higher input impedance gives lower input current: standard (reflected load impedance raises with a lighter secondary load, Z_reflected = Z_load (N1/N2)^2). The author states: "I believe I still don't quite fully understand this section." If input power falls while output is unchanged, that is a power-gain claim; no output measurement accompanies it. A resonant LC input seems to him "better" [read]; resonance raises circulating current and voltage without raising real power [memory: high]. b. Resonant step-up and impedance reflection match known experiment. "Excess output energy": no experiment I can read. c. Tested? Not under controlled conditions that I can read. The Guidelines warn "For every one device that works, expect 10 devices that don't" and "I may be wrong". d. If true: same as claim 3. Verdict: split. "Feedback raises the input impedance and so lowers input current": Plausible, untested as stated (consistent with transformer theory; magnitude and power balance not given). "Increases separation of charge carriers": Not testable as stated (no operational definition, no measurable difference from ordinary induction). "Excess output energy can be obtained": Never properly tested (no measured input/output power on record).

Claim 5. Cheap to try (about 20 dollars) and the viewers are urged to build and share

Not a physics claim. Nothing mathematically at issue; a coil, a signal source and two meters is a realistic cost. The call to measure is itself sound; the shortcoming is that the video gives no measurement protocol (true power, phase, waveform, equal-conditions control). This helps: it is testable by anyone. Verdict: Established as a feasibility statement (not a physics claim); testing of the underlying energy claim: see claim 3.

Hidden assumptions I found

1. "Lenz's law is the result of drawing current from the already separated charge, not the cause" [read]. Lenz's law in a transformer is the sign in Faraday's law and is how energy conservation shows up; treating it as a removable by-product gives a free-energy conclusion by definition rather than by derivation.
2. That cancelling B in a region cancels energy flow. Energy flows by the Poynting vector S = E x B / mu0; opposed fields in the gap between two coils reduce stored field energy there but the coils' currents still do work against their EMFs. Cancelling the sum of two vectors does not cancel what each vector does to its own coil.
3. That noise cancellation is an analogue. Active noise cancellation removes sound at one point by spending energy in a speaker; total sound energy is not destroyed, it is moved or dissipated elsewhere [memory: high]. The analogue supports, not undermines, conservation.
4. The sign of the winding is treated as physically meaningful independent of connection; as the table shows, it is a relabelling for independently loaded coils.
5. Real versus apparent power is not distinguished anywhere in the narration.

Investigator's findings that bear on my claims (kept whatever the verdict)

• The identification of the video (hyiq.org, Chris Sykes, EMJunkie; 2015-01-16 thread; Z-V1z2TdQJA) and the correction of claim 4 [web/read, Investigator]. Accepted.
• The 2011 paper "Self-Assisted A Vector Potential Oscillation": the author states the device is below unity and not over-unity, and calls the effect "the door" to above-unity devices [read per Investigator]. This is the author's own evidence, in his own words, and cuts against a claim of gain for that device.
• The 2015/2016 Guidelines: candour passages (10 failures per success; "I may be wrong"; admits not fully understanding the feedback section). Kept.
• Third-party results: gratisenergi.se ("found no Free Energy", snippet only, page 404); overunity.com thread (Wayback page 1; later pages unread); Prezi quote of "1 to 100" (unread context); Energetic Forum inverter lighting a 6 W LED (power source not stated; a 6 W lamp does not by itself show gain); aboveunity.com 2025 threads, which show continued building. Cyril Smith's 2015 paper: title supports gain, content unread. All stay in the record. Single-source or snippet-only items are described as such, not dismissed.
• Prior art named by the author (Sweet VTA 1:5, Jensen UDT, Koenig US 4,806,834, Markov, Cook US 119,825, Bearden/Hooper): attributed to him; I did not verify any of them. Sweet's 5 kW "Generation Six VTA" claim is the author's report of a second-hand record, with no measurements in the text I read.
• Cohn 1949: a mainstream source the author recommends; it says confusion between motional and transformer induction produced "innumerable d-c generators without moving contacts which have been invented (and do not work)" [read per Investigator]. Cuts both ways, as the Investigator noted.
• Historical record: hyiq.org and aboveunity.com stopped responding on 2026-10-06; the Guidelines' last good Wayback capture is 2024-10-27; overunity.com is for sale. Competing explanations (hosting lapse, platform move, withdrawal) are the Investigator's, with no evidence yet to choose between them; a site going dark is not itself a mathematical datum, but the loss of the thread removes the best public test record.

What would settle it (calculation or measurement)

1. One controlled bench test with true-power instrumentation: measure input power as the time-average of v(t) i(t) on a scope (or a power analyser), simultaneously on the output, for a 1:3 build with and without the second partnered coil, same load and drive. If P_out/P_in is at or below 1 in all cases, the strong claims fail; if above 1 with a calibrated measurement, that is the first anomalous datum.
2. Use a plain resistive load, with no rectifier, so waveform distortion does not hide errors.
3. Read the later pages of overunity.com thread 15395, the Cyril Smith paper and the gratisenergi.se page for actual numbers.

Summary table