Light Speed Measurements from Roemer and Bradley to the GPS System
| Scientific Paper | |
|---|---|
| Title | Light Speed Measurements from Roemer and Bradley to the GPS System |
| Read in full | Link to paper |
| Author(s) | Ralph Sansbury |
| Keywords | light speed, GPS, speed of light, instantaneous action at a distance, stellar aberration |
| Published | 2011 |
| Journal | Proceedings of the NPA |
| Volume | 8 |
| No. of pages | 11 |
| Pages | 507-517 |
Read the full paper here
Abstract
A survey of light speed measurements suggests that light speed can be calibrated to be equal to the standard speed of light assumption for distances up to 12000 miles about, but that for greater distances, such as that between the Earth and spacecraft, the results are not so clear. Also that our knowledge of the speed of light delay from stars is confounded with the speed of light delay between lenses and with the effects of light intensity at the lens or ccd array etc. Thus, the actual speed of light delay from astronomical distances is not as obvious as commonly believed.
Overview
Ralph Sansbury's paper is the observational half of a long-running programme in which light is not a travelling thing at all. There are, on his account, neither aether waves nor photons crossing the gap between emitter and receiver; there are only instantaneous electrostatic forces, and what we measure as the speed of light is the time a receiver takes to build up a detectable oscillation of charge under those forces. The delay, in his formulation, is not r/c but kr/c with k ≤ 1, and c2 is "a measure of elasticity of charge polarization inside electrons and atomic nuclei" rather than a propagation speed.
The immediate consequence is that delay should depend on signal strength as well as distance, since a stronger field drives the receiver above the noise floor sooner. Sansbury's survey therefore re-reads the historical measurements — Roemer, Bradley, Fizeau, Foucault, Michelson, planetary radar, GPS, the deep-space probes — asking in each case whether what was measured was propagation time or something happening in the detector, and whether source power was ever varied. His answer is that within about 12,000 miles the standard assumption can always be calibrated to hold, and that beyond it the evidence thins out badly. He compares the accumulation of properties assigned to light and to space-time with "the ever increasing number of Ptolemaic epicycles".
The argument
The experimental starting point
The paper's own evidence is a shuttered-photocell experiment: 15 ns light pulses, 2,000 per second, with a Pockels-cell shutter in front of a photocell thirty feet (30 ns) from the source. Blocking the photocell at the expected arrival time, ±5 ns, left the signal "nearly maximal"; blocking it during the emission interval left the signal zero. Sansbury reads this as showing the receiver responds to what the source is doing at the moment of emission, not to something arriving 30 ns later.
Charge inside charge
The mechanism requires structure within particles. Sansbury posits an orbiting charge of −2e about a core of +e inside the free electron, and an orbiting −e about a core of +2e inside the nucleus, moving at "necessarily superluminal velocities". Longitudinal forces from an emitter induce transverse charge polarisation in these internal orbits, transverse polarisation induces longitudinal, and the alternation builds up cumulatively until an above-noise oscillation appears in the receiver's free electrons. On this basis he claims to derive the magnetic force between parallel currents as an electrostatic dipole force, the existence of ground orbits and discrete energy levels, and closed-shell electron counts. Relativistic mass increase is likewise reinterpreted: an electron near c is not more massive but less responsive, its internal polarisation rate falling off.
His replacement for the Maxwell far-field formula keeps the same bracketed expression but inserts the factor kr in place of r in the retardation, so that after kr/c seconds the field at the receiver is what Maxwell's formula would predict at the reduced distance. If the earlier oscillations are indistinguishable from noise, the two formulas are observationally identical — but only at the intensities so far tested.
Re-reading the classical measurements
Roemer (1676). Sansbury revives Cassini's objection that the variation in Io's eclipse timings could come from changing vantage points as Jupiter, its moons and the Earth move in differently inclined planes. He notes the 1.31° inclination between Jupiter's orbit and the ecliptic, which over ~7 × 108 km amounts to some 107 km of out-of-plane displacement.
Bradley (1728). The 20.5 arcsecond annual aberration is analysed as a right triangle whose vertical side is ct and horizontal side ut, with u = 29 km/s; the quotient is tan(20.5″) and c comes out at 2.93 × 108 m/s. Sansbury's point is that t cancels, so the measurement fixes c without fixing where the delay occurs. He suggests it may occur entirely inside the 25.5-foot telescope — an implied delay of about 25.5 ns between objective and eyepiece — or, following T. Melvill's 1753 suggestion, "wholly inside the eye". Bradley's result, he concludes, measures "the time delay of increasingly faint light from the objective lenses of increasingly long telescopes a few meters from the eyepiece".
Fizeau and Foucault. The toothed-wheel measurement is re-read as a shuttering of the receiver, not a timing of transit: the delays could occur in the distant mirror, the lenses and the observer's eye. Foucault's finding that light in water is slower by a factor of about three-quarters becomes an interference between the primary field and secondary scattering from the water molecules, delaying the rise of the signal rather than slowing a wavefront. Young's double slit is likewise re-read as fields rising above detectability at slightly different times.
Planetary radar and deep space. Venus echoes are so weak that they must be dug out of noise by integration over minutes, and the returned sequence is chosen from among candidates as the one least like noise; "radar returns arriving earlier than expected are not examined". For the GPS system he grants that the 0.064 s delay over 19,310 km is internally consistent, but argues the transmitters are deliberately weak — "Why not at least a 100 watt GPS transmitter for nearly the same price?" — and that the delay is calibrated against Newtonian orbit computations rather than measured independently. He predicts that doubling emitter power and receiver sensitivity "could produce a halving of the light speed delay". The Pioneer anomaly is offered as a symptom: the craft appeared slightly too strongly attracted to the Sun, which he attributes to wrong assumptions about where the craft was when its signals were emitted.
Quantum phenomena without quanta
The last section applies the same instantaneous-force picture to atomic physics. Planck's constant is reconstructed as the product of an orbital kinetic energy and an orbital period: with r0 ≈ 10−10 m, ½mv02 ≈ 10−18 J, v0 ≈ 106 m/s and 1/f0 ≈ 10−16 s, giving h ≈ 10−34 J·s. Ground-state stability comes from the internal orbiting charges of electron and nucleus being "in synch", and quantisation from the requirement that their periods be commensurate. Helium's first ionisation potential of 24.6 V, which defeated Bohr's orbital model, is recovered by giving each helium electron an internal dipole moment 2es with a length s chosen to make the net force come out right; Sansbury calls this "exact quantitative proof". Finally he revives J. W. Nicholson's position that the radiation emitted during a transition is simply the average of the two orbital frequencies, arguing that this average is indistinguishable from the difference frequency, so that discontinuous emission and the uncertainty principle are unnecessary: "God does not play dice with the universe".
Assessment
The survey portion has real merit as history. It is true that Roemer's contemporaries, Cassini included, resisted his interpretation; true that Bradley's aberration constant fixes u/c without localising the delay; true that deep-space ranging involves layers of calibration and model-dependence; and true that no classical light-speed measurement systematically varied source intensity. Sansbury's insistence on asking what was actually measured, as opposed to what was inferred within an assumed framework, is legitimate methodology, and the paper is unusually specific about apparatus, powers and bit timings rather than gesturing at them. The arithmetic in the survey is broadly CORRECT: 19,310 km at c is 0.064 s as stated; 0.25 AU gives a 4.1-minute round trip on an 8.3-minute AU; 29 km/s divided by tan(20.5″) = 9.94 × 10−5 does give about 2.92 × 108 m/s; and 29 km/s is 1.1417 × 106 inches per second. The Fizeau line is the one internal slip: 8.633 km each way with the quoted 5.566 × 10−5 s reproduces his 3.102 × 108 m/s exactly, but the text gives the mirror distance as 8.67 km, which would yield 3.115 × 108.
The central proposal, however, is refuted by measurements the paper does not discuss. The claim that delay depends on received intensity is directly testable and has been tested: lunar laser ranging returns arrive at the same 2.5 s whether the return is a single photoelectron or many, and the round-trip time tracks the Moon's distance to millimetres over a range of return rates spanning orders of magnitude. Interplanetary radar delay to Venus varies from about 4.5 to 28 minutes across the synodic cycle exactly as the orbital geometry requires, and the astronomical unit derived from it agrees with independent spacecraft telemetry to parts in 109 — an agreement that a strength-dependent delay would destroy, since echo strength varies as 1/r4 over that same cycle while the fitted k would have to stay pinned at unity. Sansbury's own preferred anomaly has since been closed: the Pioneer acceleration was resolved in 2012 by Turyshev and colleagues as anisotropic thermal recoil from the spacecraft's radioisotope generators and electronics, with no revision of light-time assumptions.
The Bradley reinterpretation founders on a nineteenth-century experiment. If aberration were generated by the transit time between objective and eyepiece, filling the telescope tube with water — where light travels at c/n and the transit takes 33% longer — would increase the aberration angle by a third. Airy performed exactly this experiment in 1871 with a water-filled telescope and found the aberration angle unchanged. That result localises aberration outside the instrument and rules out the "delay inside the telescope, or inside the eye" reading, and the paper does not mention it. The Roemer reinterpretation fares no better: the eclipse residual follows the Earth-Jupiter distance with the synodic period and a full amplitude of about 16.6 minutes, whereas the out-of-plane geometry Sansbury invokes has the wrong period and orders of magnitude too little effect on line-of-sight distance.
Two internal problems are worth naming. First, the helium calculation is the clearest case in the paper of a fitted parameter doing the work the mechanism claims: the internal dipole length s has no independent determination and is set so that the force reproduces the measured 24.6 V, after which the agreement is described as proof. Second, the Nicholson revival can be checked directly and fails. In the Bohr ground state the orbital frequency is 6.58 × 1015 Hz and for n = 2 it is 8.22 × 1014 Hz; their average is 3.70 × 1015 Hz, whereas the Lyman-alpha transition frequency is 2.47 × 1015 Hz. That is a 50% discrepancy, not a difference "within experimental error". The average and the transition frequency converge only for large n — which is precisely Bohr's correspondence principle, a standard result, and the reason Bohr did not adopt Nicholson's position. Relatedly, the h = ½mv2 × period construction is only an order-of-magnitude identity: exactly, hforbital = 2 × the kinetic energy in the Bohr ground state, and writing every quantity as a bare power of ten conceals the factor of two.
Finally, the internal-structure premise conflicts with direct measurement. Electron-positron scattering bounds any electron substructure below about 10−18 m, and QED reproduces the electron magnetic moment to twelve figures on the assumption of a structureless Dirac particle; there is no room for an orbiting −2e inside it. Nor can relativistic mass increase be a charge-polarisation effect, since the same energy-momentum relation is measured for uncharged particles — neutrons in cold-neutron beams, neutral kaons whose decay lengths scale as γcτ — which have no net charge to polarise. Sansbury's programme is unusually consistent internally, and the shuttered-photocell result deserves an explanation; but as a general account of light delay it is contradicted by the ranging measurements it does not address.