Aether Cosmology · terrestrial tests · built 2026-11-04 08:00 UTC · companion to the VLBI analysis and its explainer

Terrestrial tests for the ether wind: ring laser, GNSS, fibre

The VLBI test samples the free ether far above the ground. This page collects tests with instruments at the surface and their own light sources, the setting Miller and Galaev actually measured in. Status of each: ring laser: done GNSS: preliminary fibre: literature only ROMY tilted rings: data blocked

1 · Wettzell G ring laser — a 4 m square interferometer with its own laser, at the surface

The G ring at the Geodetic Observatory Wettzell is a 4 m × 4 m helium-neon ring laser mounted on a granite monument, 5 m below ground under a 35 m earth mound, at latitude 49.14° N. Two beams travel the square in opposite directions; the beat between them, the Sagnac frequency of about 348.5 Hz, is proportional to the rotation rate of the ring about its normal. It is the Michelson–Gale experiment of 1925, which Sungenis and Bennett cite as evidence for ether, built to a precision of parts in 10⁹ and run continuously. Schreiber, Kodet and Klügel published 137 days of one-minute raw observations (Aug 2022 – Jan 2023) on Zenodo, with their reduction program alongside.

What an ether flow would do to it

Why a sidereal period is expected here anyway

Two ordinary effects give a ring laser a term with exactly the sidereal period: the K1 tide, the main daily constituent of the lunisolar tide, which tilts the monument by nanoradians and so changes how much of Earth's rotation projects onto the ring; and the diurnal polar motion (Oppolzer terms), the daily circle the rotation axis traces in the ground frame. Both are predicted by models that know nothing about the ring, and the solar part of the diurnal tide scales with the Sun's declination, so the combined daily amplitude is smallest at the equinoxes and largest at the solstices. A sidereal-period signal is therefore not by itself an ether fingerprint; the test is whether its amplitude, seasonal behaviour and phase follow the tidal prediction or Miller's and Allais's.

In a Galilean ether, light moves at c relative to the ether. For a closed loop at rest in the lab, the round-trip time difference between the two directions is (2/c²) ∮ W·dl, which by Stokes' theorem is proportional to the curl of the ether velocity field threading the ring. A rigid rotation of the Earth through a resting ether is one such curl, and that is the Sagnac effect the ring measures. A uniform wind has no curl and does nothing. But Galaev's ether is not uniform: his whole height effect is a vertical gradient of the horizontal wind, which he measured at 8.6 m/s per metre near Kharkov, and his terrain model has flows steered by hills and buildings over hundreds of metres. Any such gradient of a 200 to 1,400 m/s wind over metres to kilometres is a curl of order 0.1 to 10 per second. The Earth's rotation is 7.3×10⁻⁵ per second. A ring laser would read Galaev's gradient as an apparent rotation ten thousand to a hundred thousand times faster than the Earth, varying with the sidereal day as the wind swings around.

Result. The G ring's Sagnac rate carries a once-a-day term that is locked to the sidereal day: read on a 24-hour clock its phase drifts by -131° over the record, against -118° expected for a star-locked signal, while on a sidereal clock it holds still (-13°). Its amplitude is 26 to 63 parts per billion of Earth's rotation, smallest in the second half of September and largest in mid December. That is the fingerprint of the diurnal tide, whose amplitude follows the Sun's declination (zero at the equinoxes, largest at the solstices) and whose main line, K1, has exactly the sidereal period. The operators' independent tide-and-polar-motion correction removes about 85% of it, leaving at most 10 ppb. Read as an ether curl threading the ring, the uncorrected term bounds it at 6.9e-12 per second; Galaev's measured gradient is 8.6 per second, 1e+12 times larger. Allais's seasonal claim, a sidereal amplitude largest near the September equinox, is the opposite of what the ring shows.

Daily means of the Sagnac frequency, as deviation from the 137-day mean in parts per billion. The slow wander in the raw series is instrumental (backscatter and null-shift drift); the corrected series is what the operators use.

Top: amplitude of the daily term in successive 15-day windows, raw and after the operators' correction, with the shape of the solar diurnal tide for comparison. Bottom: the phase of the same term read on two clocks. A star-locked signal drifts on the solar clock and stands still on the sidereal clock, which is what the data do. Method note: 137 days cannot separate the 24 h and 23 h 56 m periods in one fit, so an earlier version of this page that quoted separate solar and sidereal amplitudes was ill-conditioned; this phase test replaces it.

Raw Sagnac rate binned by sidereal phase over the whole record. The wave is tens of ppb; a Galaev-scale ether gradient would be 10¹⁰ to 10¹¹ ppb.

What this tests and what it does not. The G ring is horizontal, so it reads the vertical component of the curl, i.e. horizontal shear of the wind, not the vertical gradient Galaev measured directly. The ether field he describes, steered by terrain into corridors and vortices at the kilometre scale, implies horizontal shear of the same order as his vertical gradient; a perfectly stratified wind with zero horizontal variation would evade this ring. The tilted rings of ROMY in Munich read the vertical gradient directly and would settle that case; their channels are listed as open on the Munich FDSN node but the waveform service returned no data for any request during this work, so that test is pending. The ring is also 5 m underground: Galaev reports that a brick building reduced his signal three to four times, so the shielding reply here is a factor of a few against a discrepancy of 10¹². The DC Sagnac rate agrees with Earth's rotation to 4×10⁻⁴ using the nominal ring dimensions, but the ring community calibrates the scale factor against Earth's rotation itself, so the DC comparison is not an independent test and is not used.

2 · GNSS short baselines preliminary

Two GNSS receivers a few kilometres apart compare the arrival of the same satellite signal at two ground antennas, so the measurement is ground to ground even though the lamp is a satellite. A Galilean wind changes each one-way arrival by (W·K/c) × (b·K/c), a few tens of centimetres on a 10 km baseline for 10 km/s, varying with the satellite direction. Method: process a day of public IGS data with RTKLIB in static mode, then inject a small synthetic wind (±200 m/s, Miller's apex) into the rover's carrier phases, reprocess, and use the change in the post-fit residuals as a matched-filter template for the real residuals. Data from the BKG IGS archive.

pair	baseline (km)	residuals used	residual RMS (mm)	ambiguities fixed	wind along Miller apex (m/s)
CEBR–MADR	10.4	31531	19	66%	+72 ± 16
KIR0–KIRU	4.5	46708	15	81%	+496 ± 26
WTZR–WTZZ	0.0	50776	6	87%	+nan ± nan

These numbers are not yet a result and must not be read as detections. The two pairs disagree with each other by far more than their formal errors (which only count random scatter), the per-satellite residuals are three to five times larger than a clean 5 to 10 km solution should give, and GNSS multipath repeats with the satellite geometry every sidereal day, which is exactly the kind of pattern a sky-fixed template can pick up. Read them as systematics-limited upper limits of a few hundred metres per second: still a factor 20 to 100 below Miller, but not yet below Galaev. The pipeline runs end to end and the injection recovers its own signature; what remains is tuning RTKLIB (or replacing it) until the residual floor is a few millimetres, then running many days and pairs. At that floor the expected sensitivity is a few metres per second per baseline-day. Left as the next step.

3 · Fibre-optic one-way tests literature

The cleanest terrestrial first-order test with a local source is a one-way light signal between two separated clocks. Krisher et al. (Phys. Rev. D 42, 731, 1990) compared two hydrogen masers 21 km apart at JPL's Goldstone complex through a buried optical fibre over five rotations of the Earth and bounded any direction-dependent variation of the one-way speed at Δc/c < 3.5×10⁻⁷, about 100 m/s. That is already below Galaev's radio value at 42 m (1,414 m/s) and his optical value at 4.75 m (435 m/s), and four orders below Miller. Modern metrology fibre links between national labs are a thousand times more stable, but their raw data are not public. A buried fibre is a pipe, which Galaev would argue guides the ether; his own tube experiments found the flow inside a tube reaches the outside speed within seconds, so the shielding reply does not obviously help here either.

4 · Summary ledger

test	where	light source	order	quantity	claimed	found
VLBI (main page)	free ether, 100s km up	quasar noise	first	wind W	10,000 m/s (Miller)	0.3 m/s
G ring laser	surface, 5 m under	own HeNe laser	first (Sagnac)	curl of W	~8.6 /s (Galaev gradient)	< 7e-12 /s
Fibre one-way (Krisher 1990)	surface, buried	own laser	first	Δc/c	200–1,400 m/s (Galaev)	< 100 m/s
GNSS short baseline	surface	satellite	first	wind W	10,000 m/s	pending

Code and results: quest haven/aether-terrestrial_QST-0000059 on scouts-git. Raw ring-laser data © Schreiber, Kodet, Klügel 2023 (Zenodo 8146444, CC-BY). JSON results: gring.json.