The Biefeld-Brown Effect as a Misread Corona Wind: Martin Tajmar's Sealed-Box Pendulum Test and Its 10 µN Limit

"The Biefeld-Brown Effect: Misinterpretation of Corona Wind Phenomena" is a paper by Martin Tajmar, then at ARC Seibersdorf research in Austria and a lecturer at the Vienna University of Technology, published in the AIAA Journal in 2004. It does two things. It reviews what Thomas Townsend Brown and others claimed about thrust from high-voltage capacitors, and it reports one experiment designed to catch any force that is left over once the ordinary electric wind has been cancelled out. The paper's summary of that experiment is blunt: "No thrust was detected within the accuracy of the experimental set-up."

The claim under test

Tajmar traces the effect to Paul Biefeld and Brown in the early 1920s: a capacitor charged to a high DC potential is said to push itself from its negative towards its positive terminal, more strongly when the electrodes are unequal, such as a wire facing a plate or disc. According to the paper, Brown's pendulum trials led him to say the effect grows with closer plates, higher dielectric strength, larger conductor area, higher voltage and a heavier dielectric. The last of these, Tajmar writes, is what prompted Brown to suspect a link between gravitation and electromagnetism.

The crucial assertion is about vacuum. The paper cites Brown's US Patent 3,187,206 for the claim that the thrust persisted at pressures down to 10^-6 Torr, which would rule out an electric wind, since a wind needs gas to push. Tajmar's objection is about the record: almost all of Brown's work sits in patents and a popular magazine piece, with no thrust figures and no account of how the effect varies with voltage, current or pressure. One of the patents on his list is the one described in US 2,949,550: Thomas Townsend Brown's Electrokinetic Apparatus Patent (Wire-and-Plate Thrusters, Discs and a Torpedo Body); why a patent cannot settle such a question is set out in Patent as Evidence.

What earlier studies had found

The review lines up four pieces of prior work.

• De Seversky's "Ionocraft" patent (US 3,130,945, 1964) gives figures Brown did not: 0.1 m² of electrode area, a 5 cm gap, 20 kV at 0.5 mA, and a force of 50 mN. De Seversky also said efficiency changed with humidity and air pressure, which Tajmar notes is what a corona wind would do.
• Christensen and Møller (AIAA Journal, 1967) measured thrust from a similar arrangement in ambient air and compared it with electric-wind theory. Tajmar describes the agreement as very good. He also cites theoretical work by Cheng (1962).
• Talley, in two reports for US Air Force laboratories (1988 and 1991), tested Brown-type electrodes in vacuum up to 10^-6 Torr. The paper says he found no thrust under a static DC potential, but noticed an anomalous force during electrical breakdowns, when current was flowing.
• Williams's five-dimensional theory, which couples gravity and electromagnetism, would allow a diverging current to induce a mass flow if its coupling constant is not zero. Tajmar treats this as a possible explanation for the claimed effect, and his sealed-box test is designed to probe it.

The experiment

The test follows from that last point. A corona discharge between unequal electrodes carries a diverging current, so if the five-dimensional coupling existed, the discharge should push the whole apparatus. An ordinary corona wind sealed inside a closed box cannot do that; it can only make the box wobble. A steady displacement of the box would therefore be the anomaly.

The electrodes were an aluminium cylinder and ring, held 6 cm apart by four dielectric rods. Tajmar chose this shape, after Christensen and Møller, over Brown's wire and plate. They sat on a wooden bar in the middle of a wooden box 50 cm on each side with 5 mm walls, wrapped outside in grounded aluminium foil to act as a Faraday cage. Box and cables hung from fixed rings so the assembly could swing as a pendulum. The supply was a Heinzinger high-voltage unit; displacement was read by a Keyence laser displacement meter placed 105 cm from the centre of the box to avoid electrostatic influence.

With a suspended mass of 7.499 kg, a pendulum length of 70 cm and a position sensitivity of ± 0.1 µm, the paper gives a force resolution of 10.5 µN. The acknowledgments place the measurements at ARC Seibersdorf in 1999.

What was measured

The voltage was raised by hand from zero to about 38 kV.

• Before the run, the box was already swinging with an amplitude the paper puts at about 50 µN.
• By 20 kV the corona had ignited, drawing roughly 0.01 mA. The swinging did not change.
• At 38 kV the current reached 0.6 mA and the swinging grew to 100 µN, which Tajmar attributes to air circulating inside the box.
• Throughout, the box swung about the same mean position.

From the unchanged mean position Tajmar concludes that any steady thrust in this configuration must be below 10 µN. At the peak input of 22.8 W that means at least 2280 W for each millinewton, against the 20-70 W/mN he quotes for Hall, ion and FEEP thrusters. Even a real effect at that level, he argues, would lose to existing electric propulsion.

For comparison he computes the corona wind the same power should drive in dry air, using Christensen's measured relation with an ion mobility for air of 2.15 × 10^-4 m² V^-1 s^-1. The result is 163 mN, which he calls similar to de Seversky's claim. That relation has the same form as the ion-drift estimate, current times gap divided by mobility, that came within a factor of 3 of a lifter's weight in Force on an Asymmetric Capacitor: The 2003 Army Research Laboratory Lifter Tests and Why Their Ion-Wind Estimate Fell Short (Bahder & Fazi); the Army report had found a collision-free ion-wind model far too weak.

What the paper does and does not show

The result is an upper limit for one geometry, in air, at up to 38 kV. It shows that enclosing a corona discharge removes the net force to within the stated resolution, which is what a wind explanation predicts.

Three limits deserve notice. First, the paper contains no vacuum measurement of its own; for vacuum it relies on Talley, and its account of Brown's vacuum claims, a wind fed by gas still escaping from the electrodes, is offered as most likely, not tested. Second, the 10 µN limit is drawn from the mean of a swing whose amplitude was five to ten times larger, with the data shown in a single figure. Third, the printed numbers do not all agree: the paper equates 50 µN with 0.95 µm, while its own formula and parameters give about 100 µN for that displacement, and 163 mN divided by 10 µN is roughly 16,000, a little over four orders of magnitude where the text says at least five.

Tajmar does not dismiss corona wind itself. He points to wind-tunnel studies of drag reduction by corona discharge and concludes that the phenomenon may be useful for propulsion, "but not along its original claims".

Further reading

A May 2004 NASA survey by Marc Millis cites this paper, an earlier Tajmar conference paper and Talley's 1991 report as the tests behind its judgement that Biefeld-Brown devices are not candidates for new propulsion physics: Prospects for Breakthrough Propulsion From Physics: Marc Millis's 2004 NASA Survey and Its Verdict on Biefeld-Brown Devices, Lifters and Asymmetrical Capacitors.

A 2009 NASA Glenn study credits this sealed-box test with showing that blocking the wind removes the thrust, then adds balance and thrust-plate measurements showing the two are equal in size, and tests whether corona thrust could propel aircraft: An Investigation of Ionic Wind Propulsion: NASA Glenn's 2009 Balance Tests Showing Lifter Thrust Equals the Ion Wind, and Why It Stays Too Weak for Aircraft (Wilson, Perkins & Thompson).