Ainslie-Rosemary-Ann · Source catalogue

Reprint on the author's blog of the 2002 Quantum experiment with an added abstract, intro, discussion and appendix. A 24 V battery (two 12 V 20 Ah in series) feeds a 10 ohm air-core wire-wound ceramic 'resistor' (150 mm x 32 mm, 48 turns, 8.64 microhenry) in series with an N-channel MOSFET (IRFPG50) switched by a 555 timer; a 1N4007 flyback diode is across the load and a 0.25 ohm shunt (four 1 ohm resistors in parallel) is in the negative lead. The 555 is said to run at 2.4 kHz, 3.7% duty, but with the gate resistor adjusted the circuit oscillates randomly. Battery power is computed from sampled oscilloscope shunt voltage (1 microsecond per division) as sum(V*I)/n, with the battery voltage taken as constant and positive and negative shunt readings treated as current out of and back into the battery. Load power is NOT measured electrically (authors say phase lag in the inductive load caused 'protracted discussion'); instead a platinum temperature probe on the load wall gives a temperature rise of 52 C above ambient, and a DC calibration run finds 13.32 V across the same resistor (V^2/R = 17.74 W) gives the same rise. Result: battery delivered 1.13 W average (67.6 kJ over 997 min), load dissipated about 17.5 W average (1.22 MJ stated), a ratio of roughly 16 to 18. Table 2 logs a 16.5 hour run in which battery voltage stayed about 24.3-25.1 V. The abstract claims a non-classical magnetic field model predicts energy dissipated can exceed supplied energy. Names Cape Metrology Laboratory (calibration), Spescom, ABB Electric Systems Technology Institute, BP South Africa, Sasol and Power Engineers as having attended or approved naming. Error margins: 10% on temperature, 10% and 1.5% on scope readings; claims a coefficient over 16 survives discounting negative samples. Conclusion: unity 'has been breached. It now remains to be disproved.' Appendix gives a verbal, non-mathematical account of current as 'imbalanced magnetic fields' that change 'justification'; it states no part of the model is developed mathematically.

Key claims

• A battery-driven switched inductive-resistive circuit dissipated about 1.22 MJ as heat in the load while the battery delivered only about 67.6 kJ over 997 minutes.
• Load heating measured by temperature rise matches a DC calibration of 17.74 W, while battery power computed from shunt-voltage samples is 1.13 W.
• The battery terminal voltage did not fall measurably over 16.5 hours (about 24.8 V to 24.3 V).
• The effect depends on counter EMF from a switched inductor and a MOSFET body diode giving a reverse-current path, and increases with switching frequency.
• Over-unity results are repeatable across many circuit variants including single and double wound inductors, AC supplies and recharging a second battery.
• Electric current is the movement of magnetic 'fields' that change justification, so the Second Law of Thermodynamics need not constrain electromagnetic interactions.

Relevance: The core experimental document of the project: states the measurement method and the headline result that every later review step must test. The method (battery power from a sampled shunt waveform treated as always at the nominal battery voltage; load power inferred from thermal calibration) is where an artefact would have to hide.

Mainstream conflict: Claims output energy exceeds input energy from the supply (violating energy conservation / the first and second laws) in a passive circuit with no stated extra energy source; claims Laws of Induction need revisiting. Mainstream expectation: apparent gain comes from measurement error with high-frequency switching waveforms, unmetered inputs, or non-steady-state heat balance.

The original published account (South African electronics trade journal 'Quantum', editor Pierre van Rhyn) of the experiment. Abstract invites readers to duplicate it. Circuit figure shows a 555 timer, IRFPG50 MOSFET, 10 ohm load, 0.25 ohm shunt, 24 V supply and a separate 12 V line. Text states the duty cycle is set to 3.7% on at 2.4 kHz, and that reducing the gate resistance makes the MOSFET oscillate 'between 143 kHz and 200 kHz' with duty defaulting to about 1.3%. Measurement apparatus: Fluke 199C 200 MHz scope, about 1200 samples, a 1.2 microsecond sample range. Energy dissipated in the load is judged by temperature rise (platinum probe, calibration against DC supply: 13.32 V gives 17.74 W for a 52 C rise). Battery power is from sampled shunt voltage, 1.13 W average; Table 1 repeats the 997-minute run (1.22 MJ at the load, 67.6 kJ from the battery). Conclusions page states error margins and that discounting negative samples by 10% still gives a coefficient above 16. Crucially it states the 555 switching circuit could not run from the 24 V battery, so a separate 12 V, 4 Ah dry-cell battery (measured 12.45 V before, 12.04 V after 16.5 h) powered the switching circuit, with a current of 0.039 A; the authors conclude the extra energy could not have come from that source. This separate supply is not included in the 1.13 W figure. Also lists calibration certificates and organisations (Cape Hope Metrology Laboratory, Spescom, ABB, BP, Sasol, J De Bruto of Power Engineers). Ends: 'Although a great deal of investigation of the phenomenon is still required to develop the mathematical constructs ... preliminary investigations ... may be consistent with claims.' Deep-read refinements. (1) The 0.039 A gate-supply current is not shown to be measured: the text gives only the open-terminal voltage before (12.45 V) and after (12.04 V) and says the drop was 'consistent with' 0.039 A; how the figure was obtained is not stated. Over 997 minutes it is about 2,333 C (about 0.65 Ah), roughly 28 kJ or about 42% on top of the 67.6 kJ battery figure; it is not in the energy balance (reviewer calculation). (2) 'A sample range was chosen spanning 1.2 micro seconds' (journal p.7); the oscilloscope stores 'about 1 200 simultaneous real time samples' (journal p.6), so the window and the memory length coincide but no sample count or number of stored waveforms is given. At 143-200 kHz the period is 5-7 microseconds, so the window is only about 17-24% of one cycle, and its position in the cycle is not stated. (3) The text says reduced gate current gives self-oscillation at 143-200 kHz with duty 'approximately 1.3%', yet also says results depend on 'the precise components used' and, later, that the circuit is 'easily repeatable'. (4) Error treatment: 10% on temperature, 10% (shunt) and 1.5% (battery) on scope readings; discounting negative samples by 10% 'still' gives a coefficient above 16. Using the arithmetic mean power 17.53 W the ratio is about 15.5, not above 16. (5) The named third parties (Cape Hope Metrology Laboratory, Spescom, ABB Electric Systems Technology Institute, BP Africa, Sasol, J De Bruto of Power Engineers) 'approved the inclusion of their name'; the article quotes no result from ABB or BP and does not state that any endorsed the over-unity result. Parts list also gives IRFPG50 (Farnell 355-744), NE555, 1N4007, 2 x 1N4148; calibration certificates dated Nov-Dec 2001.

Key claims

• Energy dissipated in a switched inductive load, measured by temperature, exceeded the energy delivered by the 24 V battery by a factor of 16 or more.
• Reducing the MOSFET gate drive makes the circuit self-oscillate at 143-200 kHz with about 1.3% duty cycle.
• A separate 12 V 4 Ah battery supplying the switching circuit drew only about 0.039 A over 16.5 h so could not account for the extra energy.
• The effect is easily repeatable by readers using the published parts list.

Relevance: Primary source. Contains the details the later blog reprint dropped or contradicts: the separate control battery and the 143-200 kHz oscillation versus the 2.4 kHz setting.

Mainstream conflict: Same as the 2009 reprint: apparent output/input ratio greater than 1 from a passive circuit.

Newspaper feature on Ainslie (age 54, 'former secretary and property agent') and her circuit that 'produces 17 times more power than it takes in'. MTN ScienCentre at Canal Walk, Cape Town put it on display. Quantum editor Pierre van Rhyn says 'It works ... But I cannot yet say why or how.' Sasol (John Marriott) offers a bursary to an academic institution to take it up. Ainslie says Stellenbosch University refused even to allow a demonstration; says the theory is a 'non-classical' magnetic field model in which all forces are underpinned by a magnetic force, and that academics refuse to evaluate because they are trained in classical thinking. ABB (North Carolina) is said to have conducted independent tests of instantaneous power and Cape Hope Metrology Laboratory to have certified the components; Prof Mike Bruton of MTN ScienCentre says it is taken seriously and there is a case for debate. The centre notes Ainslie 'couldn't give us the oscilloscope with which she displays the wave patterns and the computer on which she displays the models'. Sidebar: explains the proposal that current is transfer of discrete magnetic particles and that load energy comes from magnetic fields trapped in the resistor's structure ('atomic abodes') that 'peel off' as photons. Additional details from the digest: Ainslie, 54, says the discovery took five years; the prototype was built with engineers and technician Brian Buckley using two 12 V batteries; Fluke is said to have provided and certified test equipment, ABB (described as a NASA supplier) to have run tests in North Carolina; Prof Bruton says physics is not his expertise but the evidence merits debate. The 'refusal' by Stellenbosch is Ainslie's own statement; no response from the university is in the article. The paper's claims about ABB and Cape Hope are reported statements, not verified here.

Key claims

• The circuit gives an efficiency gain of about 17 times.
• Stellenbosch University refused to allow a demonstration of the device.
• Independent organisations (ABB, Cape metrology lab, Spescom/Fluke) tested or certified the apparatus.
• Energy dissipated in the load comes from magnetic fields stored in the resistor's structure.

Relevance: Documents the public reception and the suppression/refusal narrative (self-reported by Ainslie), and independent-test claims that need checking. A secondary, journalistic source.

Mainstream conflict: Headline: defies basic laws of physics; supporting sidebar posits resistor-bound magnetic fields as energy source.

Patent application bundle (43 pages, all OCR). Abstract: pass current through an inductor, repeatedly interrupt it (40 Hz or more) and harness the back EMF to supply an energy-receiving device; the description states comparative efficiency of 'unity' or 90% or more, and preferred duty cycles of 50-95% on (digest/Kelly). Three bench examples, read in the deep read; page numbers are PDF pages (printed page = PDF page - 2): Example 1 (pp.19-22, Fig 3A/3B): control is a 68 ohm 10 W resistor at 95 C drawing 0.0049 V / 0.025 ohm = 0.196 A at 12.28 V = 2.406 W. The test is a DIFFERENT 22 ohm resistor in parallel with an inductor, switched by a MOSFET at 5 kHz, 22.7% on, adjusted until the resistor reads 93 C. Mean sense voltage is read by eye from a sloped scope trace (0.01895 V peak-average x 8.5/37.5; the text prints 0.043 V, which should be 0.0043 V), giving 0.1718 A and 2.1097 W. Adding a 5% allowance gives 2.215 W and a stated 'efficiency 8.6 percent in excess of a factor of 1' (2.406/2.215 = 1.086; without the allowance 1.14). The ambient temperature is not stated. Reviewer arithmetic: the 8.6% is no larger than plausible reading error, and the temperature match is a single point with no calorimetry. Example 2 (pp.22-29, Fig 4): control is a 100 ohm resistor on two 6 V batteries in parallel (6.12 V), 0.0612 A = 0.374 W, 30 C against 22 C ambient. The test is a 47 ohm resistor with inductor, freewheel diode and MOSFET at 50:50 duty, load at the same 30 C. Three 0.05 ohm sense resistors: SR1 (load, peak 0.006 V; peak divided by 2 called 'rms', which is really the triangle's mean, giving 0.169 W), SR2 (battery 1: 0.0175 A on + 0.0075 A off = 0.025 A x 6.12 V = 0.153 W) and SR3 (battery 2: bipolar, peaks +-0.0055 V, about +-0.11 A). The SR3 trace is 'discounted' as 'too small to represent a meaningful measurement', although it is the largest current signal in the circuit. Two incompatible gains are given: about 10% (0.169/0.153) and '100 percent greater than unity' (0.0612 A vs 0.025 A, which is a ratio of 2.45). The text also says 'V x I' (battery voltage times the load's average current) is the right load power and that I squared R is 'erroneous'. Reviewer calculation: that is a category error for the load; a buck-converter estimate gives battery power about 0.184 W; and the three waveforms do not satisfy Kirchhoff's current law as described. The author admits 'calorific values ... were not accurately ascertained'. Example 3 (pp.30-32, Fig 5, Table 1): 12 V battery, 0.05 ohm sense resistor S1, 6R3 (6.3 ohm, 50 W) load, a 350 VA coupled-winding inductor with a diode in the second winding, MOSFET Q1; duty 90% to 50% (frequency not stated). Instruments: APPA 105 true-rms meter, Fluke Scopemeter 123, Fluke 79 II. Battery power = V1-2 x mean S1 current; load power = V3-4 rms x S1 rms current. This method is sounder than Example 2 but the table is internally inconsistent: V3-4 rms / I rms (which should equal the load resistance 6.3 ohm) is 4.89, 5.18, 5.24, 4.68 and 2.96 ohm. Using V squared / 6.3 the ratio falls to 0.91, 0.94, 0.86, 0.80, 0.46, i.e. below 1. The printed ratios are 1.176, 1.148, 1.033, 1.082, 0.976 at 90/80/70/60/50% duty. At 50% duty the net battery current is only 1.2 mV across S1 (0.024 A) against a 15.9 mV rms reading, so a 1 mV offset changes battery power by about 100%. The OCR misreads the 70% rms current (4.022); the page image shows 1.022. The three examples use different resistors, supplies, instruments and calculation methods; their claimed margins (8.6%, 10% or 145%, up to 17.6%) do not reinforce each other. The patent regime (50-99% duty, 40 Hz or more) is the opposite of the 3.7% / 1.3% duty of the 2002 heater circuit.

Key claims

• Interrupting current through an inductor and harnessing the back EMF gives energy-use efficiency of unity or more (the description says 90% or more) compared with a resistor temperature-versus-wattage calibration circuit.
• Interruption frequency should be at least 40 Hz, with duty cycle from 50% to 99% on, preferably 75-95% (digest; claims pages not deep-read).
• In Example 1 a 22 ohm inductive test circuit heated its resistor to 93 C on 8.6% less battery power than a 68 ohm control needed for 95 C (COP 1.086 with a 5% allowance).
• In Example 2 a 47 ohm inductive test circuit reached the same 30 C as a 100 ohm control while drawing 0.025 A against 0.0612 A (stated as 100% above unity; actual ratio 2.45).
• In Example 3 the load-to-battery power ratio exceeds 1 at duty cycles of 60% and above (0.976 at 50%, up to 1.176 at 90%).

Relevance: Primary source for Ainslie's earliest bench claims (priority January 1998), before the 2002 Quantum experiment. Examples 1-3 show the measurement methods and bookkeeping errors (different resistors in control and test, a discarded battery-2 waveform, mean called rms, an internally inconsistent Table 1) that the later 16-18x result must be judged against. Also shows the claimed regime (50-99% duty) differs from the 2002 circuit.

Mainstream conflict: Claims efficiency of at least unity, and in examples above unity, from a passive inductor switching circuit; energy conservation and the reviewer calculations above point to measurement or bookkeeping artefacts, not a violation.

Third-party chapter from Patrick Kelly's free-energy compilation. Opens with a polemic that university physics is 'deliberately incorrect' and that space is not empty; cites Harold Aspden's rotor energy anomaly (300 J to start, 30 J for restart within a minute), Bob Boyce's toroid electrolysis (claims over 1000% of Faraday efficiency), a South African child's toy car whose battery is recharged 30+ times by a Boyce-style device powered by the same battery, Don Smith's claim that extra energy is proportional to pulse frequency squared, and Bedini/Bearden battery chargers. Then describes the Ainslie circuit: NE555N timer, IRFPG50 FET, 10 ohm air-core load, 24 V. Kelly reports that with the 24 V battery connected and the gate resistor tuned, the 555 stops behaving as a 555 and runs at over 500 kHz with about 55% mark-space and paired non-square pulses; it radiates (interferes with a television) and only NE555N chips (not three other makes) do this. The tuning procedure: adjust the gate resistor until the battery voltage rises; Kelly states the battery voltage rise is 'caused by the inflow of external energy' and heating up to 17 times normal. Includes digest of WO 99/38247 with examples (COP 1.086; COP 2.45). Also says (digest pp.14-15) that Ainslie's back-EMF techniques give 'COP values from 4 to 17', reprints the patent's Example 3 table (COP 1.176 at 90% duty down to 0.976), and in pp.15-21 describes a mains-powered diode/transformer battery charger and a self-powering lamp/transformer/battery variant, with a stripboard build guide and a disclaimer. Third-party claims only: the 'four times input' prototype, 'COP 4 to 17' and self-charging battery are not backed by measurements in the document. Kelly's own commentary is advocacy and is not a measurement record.

Key claims

• With the gate resistor correctly adjusted the circuit self-oscillates at over 500 kHz with about 55% duty regardless of the 555 timing network.
• Only NE555N chips produce the runaway oscillation; three other 555 makes did not.
• Heating in the load is up to 17 times that expected, with part of the inflowing external energy returning to the battery so its voltage rises.
• Circuit gain comes from environmental energy inflow rather than from the battery alone.
• The patent method gives COP of 1.086 and 2.45 in the patent examples.
• Ainslie's back-EMF techniques achieve COP values from 4 to 17, and a battery-powered version can charge the battery that drives it (third-party claim, no data).

Relevance: Gives the community interpretation (zero-point or environmental energy inflow) and practical operating details that a replicator would need, including the observation that the oscillation is parasitic and frequency far above the 2.4 kHz setting.

Mainstream conflict: Attributes gain to external environmental energy; asserts universities teach 'deliberately incorrect' physics; cites Aspden effect and Boyce results that are unconfirmed by mainstream testing.

Verbal, non-mathematical cosmology. Abstract: the universe is structured from a single magnetic dipole moving at twice the speed of light; particles and field 'amalgams' make a ten-dimensional binary system; derived by symmetry and a 'principle of correspondence' (the whole field is the sum of its parts, so field parts are tiny magnetic dipoles). The author calls herself an amateur. The smallest part is the 'zipon' (from 'zero point energy'), a magnetic dipole tachyon with mass below that of a photon, superluminal speed, orbiting in closed strings. An electron is said to be three 'truants' and a proton three electrons; the mass ratio 1836 is claimed to follow from assigning zipon mass 0.5 and electron mass 1.5 and the proton three electrons plus extra truants of different sizes (see below). Claims the model addresses non-locality (quantum entanglement), dark matter, dark energy and gravity, and that fractal geometry may describe particle interactions; says it describes particles 'in defiance of the prescriptive use of mathematics'. Says the electromagnetic application was 'experimentally proven' in the appendix test (the heater experiment). Deep-read details (p.19). The author's mass-ratio sum: photon mass 1, zipon 0.5, electron 3 x 0.5 = 1.5; second-truant scale 1.5 x 4 = 6, third-truant scale 6 x 4 = 24; 3 second truants 6 x 6 x 3 = 108, 3 third truants 24 x 24 x 3 = 1728. 'This gives a mass of 1837.5, minus 1.5 for the quarks ... giving a total of 1836.' Reviewer check: 108 + 1728 is exactly 1836, so reaching 1837.5 needs an unlisted +1.5 that is immediately subtracted; the add-and-subtract is cosmetic. The factor 4 is never justified, other factors give other numbers (factor 3: 607.5), the model's own units give proton/electron = 1836/1.5 = 1224 rather than 1836, the real ratio is 1836.15267 [memory, high], and 'mass' is treated as length, area and mass in different lines. The proton is formed from three electrons, one of which is expelled to become the hydrogen electron, with no charge accounting (the text says the proton 'comprises three electrons'). The neutron is 'resolved' only as to mass, and the text itself says truant masses interchange continuously, which undercuts the fixed values used in the sum. The only experimental content in pp.17-22 is the qualitative bubble-chamber spiral and an unreferenced 'photograph' of an electron appearing and disappearing. Later pages (digest, pp.31-37) extend the model to gravity as Earth's mono-directional field acting on bi-directional trapped fields, a toroidal universe with anti-matter at the boundary, fire as unravelling zipon strings, black holes, and electric current as spin transfer; p.35 says the model 'has been tested and proven'.

Key claims

• The universe is built from magnetic dipoles (zipons) moving at about twice the speed of light.
• A proton is composed of three electrons, and the proton to electron mass ratio can be reconciled from zipon sizes.
• Magnetism is a primary force and electromagnetism is secondary.
• Entanglement shows superluminal communication, which the model's tachyons explain.
• Energy transfer via a break in magnetic field symmetry can exceed unity efficiency, as shown by the heater circuit.
• The proton/electron mass ratio is 3 x 6^2 + 3 x 24^2 = 1836 from the author's truant sizes (the arithmetic is an exact integer sum with a cosmetic add-and-subtract of 1.5; the real value is 1836.15267).
• Gravity arises from the interaction of trapped bi-directional magnetic fields with Earth's mono-directional field.

Relevance: The theoretical motivation for the experiment, by the author's own account. The paper states the experimental result does not prove the model; the model is not developed mathematically.

Mainstream conflict: Conflicts with relativity (superluminal particles), quantum field theory and the Standard Model (proton composed of quarks, not electrons; electron is an elementary particle), charge conservation, and the scattering evidence for quarks.