Proposed experiment for detection of absolute motion
| Scientific Paper | |
|---|---|
| Title | Proposed experiment for detection of absolute motion |
| Read in full | Link to paper |
| Author(s) | Gurcharn S Sandhu |
| Keywords | Relativity, isotropy, clock synchronization, absolute motion, pulsed laser. |
| Published | 2010 |
| Journal | Physics Essays |
| Volume | 23 |
| Number | 3 |
| No. of pages | 11 |
| Pages | 442 - 450 |
Read the full paper here
Abstract
According to special theory of relativity (SR) all motion is relative and existence of any privileged or absolute inertial frame of reference, which could be practically distinguished from all other inertial frames, is ruled out. However, we may define an absolute or universal reference frame as the one which is at rest with respect to the center of mass of the universe and assume the speed c of propagation of light to be an isotropic universal constant in that frame. Any motion with respect to such a reference frame will be called 'absolute motion'. The proposed experiment establishes the feasibility of detection of such an absolute motion by measuring the up-link and down-link signal propagation times between two fixed points on the surface of earth. With current technological advancements in pulsed lasers, detectors, precision atomic clocks and computers, feasibility of the proposed experiment has been confirmed. Successful conduct of the proposed experiment will initiate a paradigm shift in fundamental physics.
Overview
Sandhu, writing from Mohali in Punjab and publishing in Physics Essays, proposes a concrete terrestrial experiment to detect motion with respect to a preferred frame. He defines a Universal Celestial Reference Frame (UCRF) — at rest with respect to the centre of mass of the universe, non-rotating with respect to the celestial background — and postulates that the speed of light is isotropic in that frame and in no other. Motion relative to the UCRF is "absolute motion".
The measurement he proposes is deliberately narrow. It does not attempt to measure the one-way or two-way speed of light, does not require the baseline distance to be known, and uses no interference or moving mirrors. Two microwave towers or tall buildings about 30 km apart, each carrying a pulsed laser, an avalanche-photodiode detector, a caesium clock and an event timer, exchange single laser pulses; only the difference between the up-link and down-link propagation times is analysed, and only its variation over a sidereal day. Everything turns on how the two clocks are synchronised — they are set side by side, calibrated over 24 hours, and then carried apart — and the paper is candid that this is the crux: "An important part of this experiment consists of mutual synchronization of the two clocks."
The argument
Absolute time gives a measurable asymmetry
Two co-moving points A and B, separated by D along their common direction of motion, move at U through the frame in which light travels at c. The chasing pulse covers D + UTu = cTu; the returning pulse covers D − UTd = cTd. Eliminating the unmeasured D:
U/c = (Tu − Td) / (Tu + Td) (eq. 7)
"the ratio U/c depends on the ratio of the difference between up-link and down-link signal propagation times to the total round trip propagation time." Sandhu then presses a consistency argument: if time is absolute, the four recorded clock readings are one fixed set of numbers, yet referring the same pair of craft to the galactic or barycentric frame assigns them different velocities U1, U2. Equation (7) cannot hold for all of them at once, so "with 'absolute time', c cannot be the same isotropic universal constant in all reference frames" — the isotropic value picks out one frame, and that is the identifying mark of the UCRF.
Why relativity makes the effect undetectable
Section II works the same problem inside special relativity, and reaches the opposite conclusion honestly. If clocks A and B are Einstein-synchronised in their own rest frame K′, then T′u = T′d = D′/c by construction. Transforming the four events into a frame K in which the pair moves at U gives Tu = γ(T′u + UD′/c2) and Td = γ(T′d − UD′/c2), and dividing the difference by the sum recovers equation (7) in the form (23). But Sandhu emphasises that Tu and Td here "have not been physically 'measured' … but only 'computed' through Lorentz transformation", and cannot be obtained from the measured primed values without already knowing U. His conclusion is a clean statement of the circularity he wants to break: "if we begin by assuming the validity of the second postulate of SR, we cannot detect absolute motion because successful detection of such absolute motion will itself invalidate the second postulate."
The experiment
The two stations are synchronised in proximity to about 1 ns, then separated; any residual offset TSE is carried explicitly through the algebra, giving
Uab/c = (Tu − Td − 2TSE) / (Tu + Td) (eq. 30)
Sandhu notes two protective features. Because only the difference is used, "all constant hardware delays and atmospheric signal propagation delays will get cancelled out"; and because a common drift in clock rate multiplies Tu and Td by the same factor k, it cancels from the ratio. He specifically rejects synchronising the clocks to UTC through GPS, since the GPS links "are expected to get differentially affected by the Sagnac effect associated with their absolute motion", which would repeatedly re-absorb the very offset being sought.
Part I uses a west–east baseline, part II a south–north one. With U at polar angle θ and the site at latitude L, geometry gives the east–west modulation
Tu − Td = (U/c)(Tu+Td) sin θ sin[2πt/86164] − 2TSE (eq. 35)
a sinusoid at the sidereal period whose amplitude Ae gives U sin θ and whose mean isolates −2TSE. The north–south run yields a curve whose mean Mn gives U cos θ cos L. Combining them (eqs. 43, 44) delivers both the magnitude of U and its direction, and the phase of the sinusoid fixes the right ascension of the polar plane containing U.
Expected magnitudes
For D ≈ 30 km the round trip is about 200,000 ns. Summing the Earth's orbital motion (~30 km/s), the Sun's galactic motion (~220 km/s) and the Galaxy's motion with respect to the CMB frame (~500 km/s), Sandhu takes U in the range 300–600 km/s and expects (Tu − Td) of order 100–200 ns — far above the ~1 ns error budget he allows. The Earth's own rotation, at 0.46 km/s, is dismissed as insignificant against U.
Assessment
This is a well-posed experimental proposal, and it deserves credit for several things that similar papers get wrong. The algebra is correct throughout. Equation (7) follows exactly from equations (3) and (6): (c−U)Tu = (c+U)Td gives c(Tu−Td) = U(Tu+Td). The order-of-magnitude estimates also check out: 30 km gives a two-way time of 2×3×104/(3×108) = 2×10−4 s = 200,000 ns, and at U = 450 km/s the full-amplitude difference would be (1.5×10−3)(2×105 ns) ≈ 300 ns, consistent with the 100–200 ns quoted once the projection factor is included. The rotation term is likewise correctly judged small: for v = 0.46 km/s the same formula gives 2Dv/c2 ≈ 0.3 ns, which is the Sagnac contribution of a 30 km east–west terrestrial baseline. Sandhu's insistence that atmospheric and hardware delays cancel is also sound, since the two pulses traverse the same reciprocal path in opposite directions; and his reason for refusing GPS synchronisation is technically correct, because GPS time transfer explicitly applies a Sagnac correction computed in the Earth-centred inertial frame, which would indeed swallow the signal.
Section II is the paper's best part, and it is fair to both sides. Sandhu does not claim relativity is self-contradictory; he shows that within relativity the effect is unobservable by construction, and correctly identifies that the whole question reduces to which synchronisation convention is physically privileged. What his experiment actually tests, then, is not the existence of an aether directly but whether synchronisation by slow clock transport agrees with Einstein synchronisation. Under absolute time, transport preserves absolute simultaneity and equation (7) yields a signal; under relativity, transported clocks reproduce Einstein synchrony in the limit of vanishing transport speed — Eddington's theorem — and Tu = Td exactly. That is a real, quantitative discriminator, and stating it that sharply is a genuine contribution.
The difficulty is that this discriminator has already been run, and the paper does not engage with the record. Sandhu asserts that "the like of which has not been conducted by anyone as yet", but transported- and independent-clock tests of one-way light-speed isotropy have a long history. Krisher and colleagues at JPL (1990) compared hydrogen masers linked by a 21 km fibre and bounded any first-order anisotropy at Δc/c < 2×10−7. Riis and colleagues (1988) used two-photon absorption in a fast atomic beam to test the equivalence of slow-transport and Einstein synchronisation directly, again null. Wolf and Petit (1997) used GPS carrier-phase data to bound the first-order term at the 10−9 level. Sandhu's expected effect is U/c ≈ 1.5×10−3 — four to six orders of magnitude above these limits. He does discuss the closest precedent he found, Kozynchenko's 2006 one-way proposal, but rejects it on resolution and atmospheric grounds rather than reporting a result.
There is a deeper interpretive point the paper passes over. Modern anisotropy limits are usually expressed in the Robertson–Mansouri–Sexl test-theory parameters, and within that framework a Lorentzian aether theory — a preferred frame plus real length contraction and time dilation — makes exactly the same predictions as special relativity for every experiment, including this one. So even a null result would not refute the UCRF; it would show only that the UCRF is unobservable, which is precisely Lorentz's own position. Sandhu's proposal is therefore not a test of whether a preferred frame exists, but a test of whether Newtonian absolute time survives clock transport. Framed that way it is decidable, and the answer already on record is that it does not.
Two practical caveats are understated. First, the constancy of TSE over the run is load-bearing: a caesium standard of the class cited drifts by of order a nanosecond over a day, and while that is small against a 300 ns signal it is not against the "within one ns" budget claimed for everything else, and any drift with a diurnal component sits directly on top of the sought sidereal sinusoid. Second, the two clocks sit at the tops of two different towers; the paper assumes equal gravitational potential, but a height difference of even 10 m produces a fractional rate difference of gh/c2 ≈ 10−15, accumulating ~0.1 ns per day — again small against the predicted signal, but comparable to the stated error budget, and it needs surveying rather than assuming. Neither of these would defeat the experiment; both would need to be in the error analysis before a null could be quoted as a bound.
In sum: the arithmetic is right, the instrument list is realistic, the error-cancellation arguments are valid, and the logic connecting the second postulate to the undetectability of the effect is stated more clearly here than in most papers of this kind. What the paper lacks is the literature — a measurement of the same quantity, at far better sensitivity, has been made repeatedly, and a preferred frame of the Lorentzian type would survive either outcome.