Arbiter: HiIQ-Youtube ("How Partnered Output Coils Work Together", Chris Sykes / hyiq.org)

Step 3. I read facts.json, claims.md, physicist.md, mathematician.md, contrarian.md, sites.md and the research/ notes. There is no deep/ folder (one 143 s video, judged digest-sufficient). Tags: [read] from a project file, [calc] recomputed by me, [web] with URL, [memory] with confidence.

Source gap note

The project folder holds one file, videoplayback.mp4, and Step 1 extracted only its whisper transcript (about 1,070 characters). The Investigator closed most of the gap: on-screen text names hyiq.org, the author is Chris Sykes (forum name EMJunkie), the video dates from on or before 2015-01-16, and its written source is the 49-page "Guidelines to Bucking Coils" v2.5/v2.7 [read: research/2026-10-06-video-frames.md, research/2026-10-06-hyiq-org.md]. What remains open and must be said plainly: - No reviewer viewed the figures or scope traces in the Guidelines or the 2011 paper, or the inventor's other videos. Statements that "no power measurement exists" rest on the text layers, not on viewed figures. (Superseded by the QE revision: the figures in pp.13-16, 21-24, 34-48 of the Guidelines and the 2011 scope screens were later read by deep readers and hold no input/output power.) - Only page 1 of the 350-page overunity.com thread 15395 was read. The Cyril Smith paper, the gratisenergi.se page and the Energetic Forum thread were not read by anyone. - Several conclusions rest on [memory: high] textbook physics (ideal transformer, ampere-turn balance, Poynting's theorem). No verdict rests on a medium or low confidence [memory] fact (see Web checks). - Corrections to Step 1 that all three reviewers made and I accept: claim 4's "separation of the input coil" is a mis-transcription; the on-screen caption reads "Seperation of Charge Carriers and Excess Output Energy can be obtained!". "Jaker" is Jaycar (an Australian electronics retailer). facts.json author and year are no longer "unknown". The video's single primary with two secondaries is not the same circuit as the Guidelines' "one input coil per power coil, input coil over half the power coil" (mathematician), so the video under-specifies the device. (Retracted in the QE revision: the Guidelines' own p.1 and p.38 and the video agree on one input coil and a partnered pair.)

Corrected claim list (what the inventor claims, as the reviewers agreed to restate it)

1. Two "partnered" (bucking) output coils whose magnetic fields sum to a vector zero cancel, or greatly reduce, the load's Lenz-law reaction on the input coil. The Guidelines qualify this: total cancellation is "impossible", the field is confined to the coil space, "the Magnetic Field Component must cancel and the Electric Field Component must add" [read: Guidelines, per contrarian]. (QE revision: the circuit is now known. One input coil L1 is coupled through a shared core to the partnered pair L2, L3, which are wired in series through one 10 ohm load on the Guidelines p.38 schematic; the p.38 "Output Configuration" box also shows a parallel wiring for opposite-wound coils [read: deep/hyiq-guidelines-2016.md, page images pp.38-39]. In the video the input coil sits over only one of the two coils [read: deep/videoplayback-3c5c77.md]. See "QE revision" below.)
2. Each partnered output coil wound at 3 times the input turns gives a 1:3 voltage step-up.
3. Current is not stepped down as voltage is stepped up (Guidelines: "Current, as you may think, is not stepped down as a result of the Voltage being stepped up", placed directly after the correct statement 240 V x 1 A = 24 V x 10 A).
4. Feedback and feed forward in the partnered coils increase the "Separation of Charge Carriers", and "Excess Output Energy can be obtained". The Guidelines' mechanism: current in the partnered coils feeds back via confined fields, the input coil's impedance rises, input current and power fall; the author opens that section with "I believe I still don't quite fully understand this section" [read]. (QE revision: the same Guidelines also state "Not Free Energy" for his alternator feedback experiment on p.48, and in 2015 the inventor wrote on the forum "Energy can not be created or destroyed". Both qualify how far "excess output energy" goes as a claim of creation. See below.)
5. It is cheap to try (video: "around 20 dollars") and viewers should build and share. Not a physics claim. The Guidelines' own parts list (function generator 14.99, amplifier 12.67, coils 15.50 each, at least two) comes to 14.99 + 12.67 + 31.00 = 58.66 dollars [read: Guidelines pp.36-37; calc]; the video's figure is "around $20.00 if you already have some simple test equipment".

Recomputed numbers (where reviewers differed or quoted figures)

All [calc], rerun by me with NumPy. - Physicist's example: 12 V primary, two 1:3 outputs, 10 ohm each. Output 36 V, 3.6 A, 129.6 W each, 259.2 W total. Ideal input 259.2 W, so 21.6 A at 12 V. If the primary draws 1 A (12 W), output would be 21.6 times input. Confirmed. - Contrarian's example: 12 V at 1 A in; two 36 V outputs at 1 A each = 72 W; ratio 6; ampere-turn balance requires 3 x (1 + 1) = 6 A in. Confirmed. Both examples say the same thing: "current not stepped down" with a 1:3 ratio means 3x to 6x the input power. - Mathematician's three-coil AC model (1 mH primary, two 9 mH secondaries, 10 V peak, 1 kHz, 10 ohm loads): k = 0.99: primary 17.73 A peak, each secondary 2.93 A, input 85.77 W, output 85.77 W. Reversing one secondary: identical. k = 0.5/0.5/0.3 bucking: 2.55 A, 0.324 A, 1.048 W in and out. With 0.5 ohm wire resistance in each coil: 2.50 A, 0.317 A, input 2.615 W, output 1.005 W. All four rows confirmed to the digits quoted. Secondary current over primary current = 2.93/17.73 = 0.165, against the ideal 1/6 = 0.167. So with two 3N outputs each carries one sixth of the primary current. The physicist's "one third" is for the pair together; the reviewers do not conflict. - Bifilar charge-time test (see QE revision for the correction to the resistance): L/R = 0.231 H / 7.8 ohm = 29.6 ms; 75 ms is 2.5 time constants (90 percent in 68 ms): ordinary inductor behaviour. Non-inductive: 0.24 mH / 7.8 ohm = 30.8 microseconds (90 percent in 71 microseconds) against the measured 650 microseconds. The time ratio 75 ms / 0.65 ms = 115 is far below the inductance ratio of 962, so something other than the coil (source, instrument, wiring) set the fast case. All three reviewers agree; none shows the test bears on net energy.

Claim by claim

Claim 1. Pairing cancels the Lenz load

Where they agree. All three: the weak form (opposed coils cancel field in space, shift or reduce reflected load between coils, as in humbucking pickups, buck-boost transformers, bifilar non-inductive winding) is Established. All three: no neutral, controlled input/output power test of a partnered-coil device is on record; the claim was ignored by institutions rather than tested and rejected.

Where they differ. Physicist and mathematician call the strong form (cancelling the Lenz field lets a load be driven without matching input) Refuted by calculation. The contrarian calls it Never properly tested and declines to say "refuted" because the narration does not say whether the system is closed (a permanent magnet, battery or measurement error could account for apparent gain).

Stronger argument. The physicist and mathematician, for the device the narration describes: a coil set driven from one source, loaded by resistors, with no stated extra source. For any linear passive set of coils the inductance matrix is symmetric and positive definite, so real power in equals power dissipated in the resistances; winding sense changes only signs [calc: mathematician's table, reproduced above; the 20,000-case random check I did not rerun, but it illustrates a theorem, [memory: high]]. The ampere-turn balance the inventor wants to abolish is exactly how load power reaches the primary: the Guidelines call it Bearden's "wave/anti-wave" but state it for an ordinary loaded transformer. The mathematician's point that the text uses net-zero field both as the Lenz effect and as its cure is an internal tension the contrarian did not weigh. The physicist's reading, quoting Jensen via Sykes ("each secondary winding as the primary winding for the other"), is that if the two secondaries' ampere-turns truly cancel, one coil is being driven by the other; the primary stops seeing A's load because A's output is going into B. The contrarian is right on one point the others skip: the passivity theorem covers passive linear networks. It does not cover an active, nonlinear, saturating, resonant or magnet-assisted device, and the 2011 device the inventor cites contains a permanent magnet that his own paper calls "a second source of input energy" [read: Self-Assisted paper, per contrarian]. Conservation of energy covers every closed system, but whether a given device is closed is an experimental question.

Final verdict (split). - Weak form (opposed secondaries cancel field in space; reflected load shifts between the coils): Established. - Strong form, as a passive coil/transformer arrangement with no other energy source: Refuted by calculation (energy budget; passivity argument above). - Strong form in a nonlinear, resonant, self-oscillating or magnet-assisted device with a measured gain: Never properly tested (no instrumented input/output power on record that anyone could read; not dismissed by a controlled test, simply ignored by institutions while hobbyists built it).

What would settle it: same coil pair, sine drive, resistive loads, true input and output power (time-average of v(t) i(t) over whole cycles on a scope or power analyser, simultaneously), partnered versus conventional winding as control, same drive and load. For the inventor's rotor "no Lenz" claim: constant-speed rotor, shaft torque or motor input power with coils open versus loaded, partnered versus plain.

Claim 2. 1:3 voltage in each output

Agreement across all three. Established as the ideal, tight-coupling relation V2/V1 = N2/N1. Caveats worth keeping (physicist, mathematician): with loose coupling or a gapped core, which the Guidelines say works "better", the ratio is about k N2/N1 and falls further under load; two secondaries wound oppositely on a common flux have equal-magnitude, opposite-sign terminal voltages, so wiring them in series cancels. No disagreement.

Final verdict: Established (ideal/tight coupling; smaller under load and loose coupling).

Claim 3. Current is not stepped down

Agreement. Everyone separates a trivial reading from a strong one. Trivial: the output current is fixed by the load, I = V/R; true of every transformer. Strong: output current equals input current at 3 times the voltage; that is 3x to 6x power gain (see recomputed numbers). All three note the claim contradicts the Guidelines' own 240 V x 1 A = 24 V x 10 A paragraph.

Disagreement. Physicist: Refuted by calculation (energy budget). Mathematician: Refuted by calculation and Internally inconsistent. Contrarian: strong form Never properly tested; Internally inconsistent only if both sentences are unqualified, and not "refuted" without knowing closure.

Stronger argument. The author qualifies the sentence ("these devices use some Transformer Principals, but not others") without an equation, so strict internal inconsistency is only established for the unqualified reading; the contrarian is right about that. But as a statement about a passive 1:3 transformer pair, the calculation is clean and needs no experiment: I_in = 3 (I_out,A + I_out,B). The physicist's and contrarian's measurement-pitfall lists agree and should be kept for the reader: apparent versus real power (a 10 V, 1 A circuit at power factor 0.1 is 10 VA but 1 W [calc: mathematician's example, checked as 10 x 1 x 0.1]), non-sinusoidal waveforms, DC-supply measurement with pulsed draw (overunity.com reply #3589 warns of this [web per Investigator, page not read]). The inventor's 2011 paper claims "No to minimum (0-5%)" phase angle for a coil, a statement unsupported in the text [read: per contrarian]. The "1 to 100" input:output figure quoted on a Prezi page from an overunity.com post is a lead whose context nobody read; it can be neither used nor dismissed.

Final verdict (split). - Output current set by the load: Established (trivial). - As a statement that a 1:3 passive step-up delivers its current without the primary supplying it: Internally inconsistent (with the Guidelines' own transformer paragraph) and Refuted by calculation. - That some real device shows unmeasured gain of this kind: Never properly tested (no power figure for a partnered-coil device appears in the documents read).

What would settle it: the same bench test as claim 1, on a plain resistive load without rectifier, with calorimetry of the load as an independent check.

Claim 4. Feedback and feed forward; charge-carrier separation; excess output energy

Agreement. "Separation of charge carriers" is the inventor's name for induction through the Lorentz force; it is a legitimate picture for motional EMF (Cohn 1949) but is not defined as a measurable quantity, and for a stationary transformer coil the standard source is the induced electric field (curl E = -dB/dt). "Feedback and feed forward" has no stated coil, phase or gain. "Excess output energy" has no energy accounting and no measurement. The bifilar time-constant test does not support "the magnetic field slows charge flow" beyond L di/dt.

Disagreement on one piece. The impedance mechanism: mathematician says "Plausible, untested as stated"; contrarian says Established; physicist folds it into Not testable. The stronger argument is the contrarian's on principle (a returning source on the secondary side raises the primary's input impedance, and lowers input current and power, in a conventional loaded transformer, matching Z_reflected = Z_load (N1/N2)^2 [memory: high]) but the mathematician's on scope: the Guidelines give no magnitudes and no output measurement, so only the qualitative direction is Established. The contrarian's reading that the Guidelines describe power entering from a second source on the secondary side, not a gain, is a fair and important one; it is also the inventor's wording ("a load that just so happens to be another Power Source") [read, per contrarian].

Honest note on the inventor's own history, kept in the record: in 2011 he wrote twice that his first device "is below unity" and "not an Over-Unity Device", with 15.2 W DC input and a 3.6 W lamp load [read: Self-Assisted paper]. By 2015 the video caption says excess output energy "can be obtained". No measurement bridging the two statements is in the documents read. He also writes "For every one device that works, expect 10 devices that don't" and "Sweet's conductor would carry 10 times more current at the same temperature": the latter contradicts Joule heating P = I^2 R at fixed resistance (the heat in the same wire would be 100 times larger) [physicist; [memory: high] physics, [read] quote]. Both stay in the record.

Final verdict (split). - Output-side feedback changes the input coil's impedance and so its input current: Established in direction (conventional transformer behaviour); magnitude and power balance Plausible, untested as stated. - "Increases the separation of charge carriers": Not testable as stated. - "Excess output energy can be obtained": Never properly tested (no measured input and output on record); by energy conservation, refuted for any closed passive system, as in claims 1 and 3.

What would settle it: a stated operational definition of "separation of charge carriers" (a measurable quantity), and the same input/output power bench test with the feedback path in and out.

Claim 5. Cheap to try, build and share

Not a physics claim. Agreement. The idea is cheap to test and the inventor published sources and invited replication ("I have given everything freely"); a controlled test needs a scope or power meter with phase, not much more. The real cost is above "20 dollars" by the Guidelines' own list.

Final verdict: Established as a feasibility and cost statement. The implied outcome (that a builder will see excess energy): Never properly tested by anyone whose results could be read. 305,360 reads and 350 pages by Sept 2015 show the invitation was widely taken up [web: Wayback capture of overunity.com thread 15395, per Investigator]; that is evidence of interest, not of replication.

Evidence that stays in the record whatever the verdict

• Third-party results, each documented in one source and none yet read in full: gratisenergi.se, "I made a Bucking Coils Transformer and was dissapointed when I found no Free Energy" (Google snippet; the page returns 404); Cyril Smith, "Bucking Coils produce Energy Gain" (2015; first line "There are many claims of overunity for systems that employ bucking coils"; the title suggests a gain, the opening line a survey; unread); Energetic Forum "Bucking Coil Inverter" (6 W LED lit through 30 AWG partnered coils; input power not seen); aboveunity.com threads from 2025 showing builders still working on it. The inventor's forum is an interested party: "above unity" labels there are his and his community's characterization.
• Prior art the inventor names (his attributions, mostly unverified): Sweet's VTA, Jensen UDT, Koenig, Markov, Melnichenko, Frolov, Cook 1871, Bearden. Sweet's 1:5 step-up and 5 kW claims come via second-hand sources and interested promoters; no independent measurement of the VTA was found [web: scienceforums.net and KeelyNet, per physicist].
• Cohn 1949 cuts both ways: the inventor recommends it for separating motional from transformer induction; its page 1 says confusion of the two laws produced "innumerable d-c generators without moving contacts which have been invented (and do not work)" [read: page image p.1, per Investigator].
• The inventor's own candour and concessions are listed above; the claim hardened from 2011 to 2015 without data in the documents read.

Web checks (budget 8; six checks documented below)

Verdicts depend on high-confidence physics only. Checks made of Investigator and reviewer facts: 1. Koenig US 4,806,834 [web: https://patents.google.com/patent/US4806834A/en]: confirmed. "Electrical circuit for inductance conductors, transformers and motors", Earl Koenig, filed 1987-04-16, granted 1989-02-21, expired for lapsed fees in 2001. A pair of coils on a common axis in mirror-image winding, connected in parallel. Interested party: this is the inventor's own patent, reporting that a transformer stepping 110 V down to an expected 3.09 V produced 5.77 V, "efficiency of almost 100 percent can be achieved", and 25 percent less amperage for a motor. The fetch summary says it makes no explicit over-unity claim. This is a different circuit (parallel mirror coils, not a bucking pair) from the video's, and its numbers are unverified, so I note it as related prior art, not support. 2. Cyril Smith paper [web: https://www.scribd.com/document/572304933/Bucking-Coils-Produce-Energy-Gain2]: check failed to read the text. The fetch metadata names the document's author as "Konstantinos Tomazis" (probably the uploader) and its length as 6 pages. The content remains unknown; I cannot say whether it reports a measured gain. 3. hyiq.org and aboveunity.com [web: https://www.hyiq.org/, https://www.aboveunity.com/]: from this environment both refused the connection (ECONNREFUSED at 45.74.14.60:443, one shared address) on 2026-10-06. That agrees with the Investigator's finding that both sites do not respond, and shows both share one server. 4. Wayback CDX recheck for 2025-2026 captures of hyiq.org and aboveunity.com: check failed (Claude Code cannot fetch web.archive.org). 5. Cohn 1949 [web: search]: no independent index entry found by search; the quote rests on the Investigator's page image [read]. Check failed to corroborate; not evidence either way. 6. A search for independent measurements of partnered output coils by the inventor or others returned nothing relevant. Negative for "found".

Absence of evidence: the record goes dark

What is silent. (a) hyiq.org (the inventor's own site) and aboveunity.com (his forum) do not respond; overunity.com, which hosted the 350-page thread where the claim was tested publicly, is now a domain for sale; the last good Wayback capture of the Guidelines is 2024-10-27, a 2025-05-14 capture has no status [web: per Investigator]. (b) Nowhere in the 2015-2025 material is a published, instrumented input/output power measurement for a partnered-coil device; the Cyril Smith paper and the thread's later pages are the places where one would be expected.

Competing explanations for (a) and what would decide: - Hosting lapse or server fault. For: a shared server address refusing connections is consistent with an outage; Google still indexes pages and aboveunity threads are dated Dec 2025 [web per Investigator]; the Wayback copy of the Guidelines is stable to 2024. Against: the failure persisted across several tools and sites on one day, and a 2025 capture already has no status. - Platform move by the owner. For: the inventor has moved platforms before (overunityresearch.com, then overunity.com, then aboveunity.com). Against: no new address found; hyiq.org's menu points to aboveunity.com, which is also down. - Withdrawal by the owner. For: nothing documented. Against: his stated policy is to give everything away freely. The inventor says "PayPal De-Platformed me" and solicits donations on aboveunity.com [read: Wayback 2022, per Investigator]; that is his account of a funding or payment dispute, which no record here confirms or tests. It bears on funding, not necessarily on hosting. - Domain or payment lapse, or suppression by a third party. For: the "de-platformed" statement above. Against: no document of any takedown or complaint. Evidence so far favours a transient or hosting fault over withdrawal; one day's failed visits cannot decide. Records that would decide: a repeat visit in a few days, Wayback CDX captures of aboveunity.com and hyiq.org in 2026, the domain registrar (WHOIS) and hosting history, and an announcement on the inventor's YouTube channel (@SweetSQM).

Competing explanations for (b), the missing measurement: the test was never done to a standard that could be published (hobbyist culture, no instruments or no phase-corrected power measurement); it was done and showed no gain, so nobody posted a result (the gratisenergi.se report is one example, from one snippet); data were posted in forum pages nobody here has read (most likely, since only page 1 of 350 was read); or data were lost with the forum. Evidence for the last two is thin and absence alone does not favour either. The pages most likely to hold numbers are overunity.com thread 15395 replies #3589, msg469812 and msg471869, the Energetic Forum thread and the Cyril Smith paper. The Wayback captures of the thread are what would decide.

What would settle it (consolidated)

1. The bench test in claim 1: true input and output power on the same bench, partnered versus conventional, resistive load, whole-cycle v x i, with calorimetry as a check. Output over input at or below 1 with a calibrated meter settles the strong forms against the claim for that build; above 1 with a calibrated meter would be a first anomalous datum, to be repeated by an independent party.
2. A calculation the reader can do now: the mathematician's three-coil model, solving Z I = V with Z = j omega L + R for any coupling matrix; output power never exceeds input power.
3. Records: later pages of overunity.com thread 15395, the Cyril Smith paper, the gratisenergi.se page, the Guidelines' figures and scope traces, videos iJsVSMQqCOM, ZhQgch4L5XY, iUy_9jDKDvo, a 2026 Wayback check of both sites.

Open questions

• ~~Which circuit does the video mean~~ Resolved in the QE revision: one input coil, a partnered pair on one load (series on the main schematic, parallel as an option), the input over one of the two coils in the video.
• Is the "net-zero field" a cure for the Lenz load or a description of it? The text says both.
• How does the claim go from "below unity" (2011) to "excess output energy" (2015) without measurement in the documents read?
• Do the later forum pages contain any calibrated measurement, and what did the 2025 aboveunity.com builders measure?
• Could a nonlinear or resonant partnered-coil device show net power-factor or measurement artefacts that produce apparent gain? That is a measurement question, not a refutation of conservation.
• Why are both of the inventor's sites down on 2026-10-06, and will they return?

Summary of final verdicts

QE revision

Written after qe.md (verdict MINOR) and five deep reads in deep/: the Guidelines (page images pp.13-16, 21-24, 34-48), the 2011 Self-Assisted paper, overunity.com thread 15395 page 1, the video (frames at 1 per second), and Cohn 1949. Tags: [read] from a deep-read file or project file, [calc] recomputed by me, [memory] with confidence. I did not re-open the page images myself; I rely on the deep reads, and say where a statement is the deep reader's own analysis. I made no new web calls in this revision.

Verdict effect in one line

No verdict moves. Three wordings change (claims 4 and 5, and the bifilar numbers); one open question is resolved; one new piece of evidence (COP 1.7) and several pieces of the inventor's own material are added to the record. The reason no verdict moves: the strong-form verdicts rest on the passivity theorem and energy conservation, which hold for any coupling, winding sense and series or parallel connection, and the deep reads found no measured input and output power for any partnered-coil device that could replace them.

1. Open question 1 resolved: which circuit

• The Guidelines p.38 main schematic: function generator, then power amplifier, then one input coil L1 ("Input Coil") on a shared core; partnered coils L2 and L3 on the other side; L2 and L3 in series with one "Resistive Load 10" (10 ohm, "this can be decreased or increased as needed"). No diode, no capacitor, no galvanic connection to the input; no polarity dots [read: deep/hyiq-guidelines-2016.md section 3].
• The "Output Configuration guide" on the same page has two wirings. "CW/CW 1 Flipped Coils" is a series loop (L2, L3, load). "CW/CCW Coils" has the two coils in parallel across one load (deep reader's description; the load value is not given in either box) [read: same, section 3]. QE's statement that the circuit is "the series-connected pair" is therefore right for the main schematic and for one box, and incomplete: the Guidelines show a parallel option too.
• The video: one input coil (green) sits over one of the two output coils (gold), which are joined end to end by one wire with a lead at each end; the two field arrows point head to head. The deep reader marks the arrow orientation on the input-bearing coil as low confidence at 640x360 [read: deep/videoplayback-3c5c77.md section 2].
• Forum reply #12: "On the input, really anything that works! I have always used a single coil" [read: deep/overunity-15395-p1.md].
• Result: the mathematician's "the video's single primary is not the Guidelines' circuit" is not supported (QE check 10 agrees). I retract the sentence in my own source-gap note that accepted it. Open question 1 is closed: single input coil, partnered pair wired to one load, series on the main schematic, parallel as an option. The physicist's and contrarian's worked examples (each output on its own load) are energy bookkeeping, not the inventor's circuit. The mathematician's asymmetric-coupling cases and example.py case B4 are the closer model [QE; I did not rerun example.py].
• One physics point from the deep reader that I keep as the reader's analysis, not as the inventor's statement: with two coils on a common flux wired in series so their fields oppose, the two induced voltages subtract, so output exists only to the extent the flux through the two coils differs. That fits the video, where the input sits over only one coil, and it fits the inventor's own remark that total cancellation is impossible. It also means "fields sum to zero" and "an output appears" pull in opposite directions, which is the internal tension already noted under claim 1. It does not change the verdict: whatever flux difference exists, the coils are passive and linear unless a magnet or nonlinear element is added.

2. Measurements the figures do and do not contain (DR1, DR3)

[read: deep/hyiq-guidelines-2016.md] In the Guidelines pages read, no figure, scope screen, meter photo or table gives an input or output voltage, current or power for a partnered-coil device, the rotor test (p.45, prose only: "No or next to no Lenz's Law and at the same time, Electricity is Generated"; no speed, load, drag or output figure) or the alternator test (p.48, prose only). The only instrument readings are two OWON oscilloscope screens of the bifilar test (pp.14-15). The only power figures are second-hand and unsourced: Sweet's Generation Six VTA "in the order of 5000 Watts" (p.13) and Rosenthal's 25 W minimum load (p.46). This turns my earlier "no power measurement appears in the text layers" into "none appears in the text layers or the figures of the pages read". Pages outside pp.13-16, 21-24 and 34-48 were not given a page-image read.

[read: deep/hyiq-self-assisted-2011.md] The 2011 paper has three scope screens at about 200 Hz, voltage only except Fig 4. Fig 2: input coil 15.78 V RMS, A-Vector coils 8.557 V RMS. Fig 3: output coil 1 at 3.659 V RMS, A-Vector coils 8.570 V RMS. Fig 4 (3.6 W automotive globe load): voltage 1.977 V RMS, "current" channel 82.41 mV RMS with no shunt value given. The Specifications section (p.9) has no input voltage, current, power or phase. The 15.2 W DC appears only in prose (pp.5-6) with no stated measurement point. So output power cannot be computed from the paper.

My arithmetic on the deep read's numbers [calc]: - The globe is rated 12 V, 300 mA = 3.6 W. At the measured 1.977 V, scaling by V squared gives (1.977/12)^2 x 3.6 = 0.098 W, ignoring the change of filament resistance; so the 3.6 W is a nameplate, not a measured output. - If the shunt were 1 ohm, output would be at most 1.977 x 0.0824 = 0.163 W, about 1.1 percent of 15.2 W; if 0.1 ohm, about 1.63 W. The unknown shunt sets the answer by a factor of ten. Both are below the 15.2 W DC figure, which fits the inventor's own "below unity" statement and does not test it, because the DC input may include amplifier losses. - The paper's voltages do not reconcile: 8.56 V in Figs 2-3, 1.977 V in Fig 4, and "10V RMS" in the text (p.5). Whether the figures are one operating state is not stated. - Internal tension in the paper: p.2 says the input coil has no mutual inductance with the A-Vector coils (opposed windings); p.5 finding 1 says a load on Output Coil 1 raises the input current "through Mutual inductance". These concern different coil pairs, so it is not a contradiction, but the paper does not make the distinction plain.

Effect on verdicts: claims 1, 3 and 4 stay "Never properly tested" for the active or magnet-assisted device. The wording can be sharper than before: the inventor's documents contain no instrumented input and output power for a partnered-coil device; the only on-record input figure (15.2 W DC) has no matching output measurement.

3. Correction to my bifilar numbers

[read: deep/hyiq-guidelines-2016.md section 2] The circuit puts a 1 ohm current-sense resistor in series with the coil, so the series resistance in the time constant is 7.8 + 1 = 8.8 ohm, not 7.8 ohm. I omitted it. [calc]: - Inductive case: tau = 0.231 / 8.8 = 26.3 ms (not 29.6 ms); three tau is about 79 ms, still consistent with the inventor's "75 ms". - Non-inductive case: tau = 0.24 mH / 8.8 = 27.3 microseconds (not 30.8); the deep reader reads the observed rise as about 400 microseconds from the graticule against the inventor's "650 microseconds", still some 15 to 25 times slower than L/R predicts, so the earlier conclusion stands: something other than the coil (MOSFET gate drive, generator edge, probe) limits the fast case. - The two runs also differ in supply (15.4 V against 7.0 V) and drive frequency (about 6 Hz against about 647 Hz), as well as in connection; the inventor's "same resistance" comparison holds only for the coil. This is the deep reader's observation, and I agree with it. - The stated inductances imply a mutual inductance M = (0.231 - 2 x 0.0556)/2 = 0.0599 H and a coupling k = M/L = 1.08 [calc], above the physical limit of 1; read as near-perfect coupling within measurement error. - Neither the inventor nor any reviewer shows the test bears on net energy. The verdict on this sub-claim is unchanged: ordinary inductor behaviour.

4. Evidence added to the record

All stay in the record whatever the verdict, with source and stake.

1. "COP 1.7". The critic MileHigh (reply #8, 2015-01-16, self-described "hostile witness") writes of "your alleged COP 1.7 device" and "how are we even supposed to know that your COP 1.7 claim is even credible? I looked through your pdf doc and I didn't see anything like that". COP means coefficient of performance, output power over input power. Sykes quotes the post in reply #11 and answers: "I have decided that the only proof is in the eyes of the end user". The deep reader's wording is that reply #11 "does not confirm, correct or dispute the figure"; QE's "did not dispute" is the same fact. So: a gain figure attributed to the inventor's device by a critic, in one forum post; no measurement, circuit or origin given; not confirmed or denied by the inventor on page 1; the origin may lie in the earlier thread (14974, "from about page 76 on") or the "Net Zero Magnetic Field.png" attachment, neither read. It cannot be used as a measurement and cannot be dismissed. It replaces my earlier statement that no performance figure exists, apart from the Prezi "1 to 100" lead [read: deep/overunity-15395-p1.md]. Both figures are of unknown origin, and 1.7 and 100 differ by a factor of about 60.
2. Named testers and the critic's protocol. MileHigh names "TK or Itsu ... with some help from MarkE and/or Picowatt" as the replication team. The page shows no test by any of them; their results, if any, lie in later pages that nobody has read. I say neither that they replicated nor that they did not. MileHigh's 10-point protocol (measure the load resistor, state both test points, give error tolerances, do not trust bench-supply current displays) fits my bench test below.
3. Sykes in his own forum words: "Energy can not be created or destroyed. You and I agree with Science on this one!" (#7); "Lenz's Law is ONLY apparent when Current is drawn" (#4); "Loose Coupling is a requirement!" and cost "no more than $20.00" (#11); "I have found 3 configurations to work! I have not included one because it over complicates the situation for beginners" (#12), no description of any of them and no definition of "work". The inventor does not claim to break conservation; he names an outside source (the Guidelines p.47: "there was no new form of Energy"; the Dirac sea and "the Aether" in the Guidelines, per QE). Whether the energy exists outside the circuit is the physical question; a measurement would settle it.
4. Alternator experiment, Guidelines p.48 [read: deep]: an automotive alternator driven by a motor, field coil shorted or in series with the output coils, gives "enough Feed Back ... to get an Electrical Output". "Not Free Energy, but this shows that Feed Back Techniques can excite the Coils", and "Lenz's Law is Only a Result of the Electrical Output, NOT the Cause!!!" This is generator self-excitation, a standard effect [memory: high], and the inventor presents it as not a gain.
5. Inventor material that was only in physicist.md. (a) CRT-magnet experiment (p.23): a magnet on a CRT screen and a second one pushed in with poles opposed; "Result: The Magnetic Field will be squeezed Nearly to Zero - NO MAGNETIC FIELD!!!", no numbers. The physicist's judgement, which I accept: a gauss-meter or Hall probe would show the field redistributed in the gap, and zero field at a point is not zero field everywhere. (b) "Mechanical Energy is NOT the Source" (p.8). (c) "A is the Electric Component of the Magnetic Field". I keep (b) and (c) as the inventor's statements; the physicist judges them against textbook definitions of the vector potential and the Lorentz force, and I have not re-argued them here.
6. Further experimental claims in the Guidelines [read: deep]: the rotor test (p.45) with no numbers; "I seem to get a better result if the Input is resonated, for example, LC Tank" (p.24); "Magnetic Gating" with a diode (pp.40-42), marked "work in progress" and "only appears to be valid under some configurations"; "expect 1 in 10 devices to work" (pp.34, 45); "many hundreds of hours, possibly ... many thousands" with "I still do not have all the answers" (p.34). The first device dates from 2010-11 (p.24); the Guidelines PDF is dated 2016-02-25 [read: PDF metadata].
7. Cohn 1949 [read: deep/cohn-1949.md]: the deep reader finds that beyond p.1 Cohn does not treat coupled windings, bucking coils, load current, Lenz back-reaction or power transfer. His p.1 line about "innumerable d-c generators without moving contacts which have been invented (and do not work)" is about those machines, where a transformer term was left out; it is not about bucking coils. My earlier "cuts both ways" wording stands but is narrower than I implied: Cohn neither supports nor refutes partnered coils. The step from his equations to a bucking result is the deep reader's inference [memory: high].
8. A note to the sceptical reader: the inventor's earliest document here states "below unity" (2011), and the Guidelines carry the p.48 "Not Free Energy" and the 1 in 10 failure rate. Stated beside them: the 2015 video caption ("Excess Output Energy can be obtained!") and "COP 1.7" in the forum. The record is thus not a straight line from "below unity" to "gain".

5. Changes to my earlier statements

• "The claim hardened from 2011 to 2015 without data": qualified. It is true that the 2015 caption is stronger than the 2011 wording and that no measurement bridges them. But the inventor wrote "Not Free Energy" for his feedback experiment, says energy is conserved, and the 2011 paper already says the device might lead to "above unity" ("It's the door to achieving above unity Devices"). The change is one of emphasis between documents, not a demonstrated reversal.
• Claim 5: "20 dollars" against 58.66 dollars is computed above, not just "well over".
• Web-check count: the heading said eight calls made; six are listed and documented in the section above. I cannot document the other two, so the heading now says six.
• Verdict wording: "Established in direction" (claim 4) is replaced by "Established (direction only); magnitude Plausible, untested". "Established as a feasibility statement" (claim 5) is kept as a feasibility, not a physics verdict, and the Curator should read it as such.
• Source-gap note: the Guidelines figures pp.13-16, 21-24, 34-48 and the 2011 scope screens have now been read by a deep reader and show no input/output power. Still unread by anyone: later forum pages (2-350), topic 14974, the Cyril Smith paper, the gratisenergi.se page, videos ZhQgch4L5XY, iUy_9jDKDvo and zSadMqEnO7I, and the figures on Guidelines pages outside the ranges above.

6. What would settle it (update)

Unchanged in substance. Add: (a) read the later pages of thread 15395 and topic 14974 for the origin of "COP 1.7" and for any posted result by TK, Itsu, MarkE or Picowatt; (b) apply MileHigh's protocol to the bench test, on the circuit of p.38 (one input coil, partnered pair in series, 10 ohm), then repeat with the parallel wiring and with the input over one coil, as in the video.