[David Cooper]

Hi David,
I think it would be possible to carry out the experiment. Would you be  (or would you know somebody ) interested in helping to set it up?

I'd certainly like to be involved in setting it up if once the numbers have been properly crunched it looks as if it would work, but I don't think the right crunching has been done yet. I've thought of a simpler way to do that crunching, so here's the new approach. We need to crunch the numbers twice: the first time with the assumption that the clocks are stationary and that the rod is length contracted, and the second time with the assumption that the rod is stationary and that the clocks are moving past it instead, which means the clocks are closer together instead. When ends of the rod pass clocks, light is sent out towards the other clock, but on the first run, clock A starts running before clock B because the rod is shorter than the distance AB, and on the second run, clock B starts running before clock A because the rod is longer than the contracted distance AB.

So, what we need to know is this: what does clock A read when light reaches it from clock B, and what does clock B read when light reaches it from clock A. If the numbers match up for the first and second run, then the experiment cannot detect the one-way speed of light, but if they don't match, it is likely that you will get a Nobel prize. I'm sure this will have been explored a hundred years ago though, so I think they'll match.

[David Cooper]

Let's try again.

A photon has no idea whether it is travelling towards or away from you. Until it reaches you, it doesn't even know if you exist. Therefore its speed in vacuo is independent of direction.

Now suppose that the vacuum is in fact filled with aether, which is in some way essential for the propagation of light,  and we are travelling through it in some direction. The measured two-way speed will depend on whether we are measuring "into wind" or "across wind", and the Michelson-Morley experiment (and its descendants) have clearly and consistently shown no difference.

Correct: they measure the two-way speed of light and always show that to be c. That doesn't get us any further on as everyone appears to agree on that point already.

[alancalverd]

The point is that if the one-way speed was c + x and c - x in the other direction,

(a) the measured two way speed would always be less than c

(b) the measured two-way speed would vary with the initial direction of propagation 

(c) Maxwell's equations would not give the measured value of c

(d) indeed Maxwell's equations would not describe a selfpropagating electrromagnetic wave at all

(e) you would have to invent some means by which a photon would know whether it was travelling towards or away from you (and explain why "you" were of cosmic significance)

(f) the energy of the photons emitted from pair production and annihiliation would depend on the direction of travel of the initiating photon that produced the pair

and so forth. In a few words, if c depended on the direction of propagation, a whole lot of what we know from experiment would not be true. If c does not depend on the direction of propagation, the one-way speed must be  the same as the two-way speed.

But if you want a single defining experiment, measure radiation pressure.

[David Cooper]

The point is that if the one-way speed was c + x and c - x in the other direction,

(a) the measured two way speed would always be less than c

Not when your clock runs slow - your movement through space cancels out the difference.

(b) the measured two-way speed would vary with the initial direction of propagation

You need to elaborate on that to spell out where there's a problem.

(c) Maxwell's equations would not give the measured value of c

(d) indeed Maxwell's equations would not describe a selfpropagating electrromagnetic wave at all

Are you sure there isn't a distance term being smuggled in somewhere which is calculated based on the assumption that you aren't moving through space? And if there's no distance involved, how would the equations have any relation to relative position or to movement through space?

(e) you would have to invent some means by which a photon would know whether it was travelling towards or away from you (and explain why "you" were of cosmic significance)

Why would it need to know anything? If a photon's moving away from you, you simply won't encounter it.

(f) the energy of the photons emitted from pair production and annihiliation would depend on the direction of travel of the initiating photon that produced the pair

Again you need to elaborate. Where's the problem?

and so forth. In a few words, if c depended on the direction of propagation, a whole lot of what we know from experiment would not be true. If c does not depend on the direction of propagation, the one-way speed must be  the same as the two-way speed.

All the experiments run into the same issue - movement through space is hidden and none of those experiments are capable of revealing that movement. Anyone who thinks they should be able to identify movement through space does not understand relativity adequately.

But if you want a single defining experiment, measure radiation pressure.

What use is that? Again you're being light on detail in the hope that your assertions will be left standing purely because there's not enough substance to them to attack. The faster you run into radiation, the higher its energy appears to be, so that would be measured as a higher pressure.

[guest4091]

If light signals moving at c, in opposite directions around the earth at the equator, return to their starting positions at different times, then with a rotation speed of o mph, 1 circuit requires .1333 sec. With a rotation speed of 1000 mph, the westbound signal arrives 200 ns earlier, the eastbound signal arrives 200 ns later. Knowing the time and distance should allow comparison of 1-way speeds. This experiment involves an absolute frame whereas in SR there is none.My statement in #4 was within the SR environment.

[David Cooper]

If light signals moving at c, in opposite directions around the earth at the equator, return to their starting positions at different times, then with a rotation speed of o mph, 1 circuit requires .1333 sec. With a rotation speed of 1000 mph, the westbound signal arrives 200 ns earlier, the eastbound signal arrives 200 ns later. Knowing the time and distance should allow comparison of 1-way speeds. This experiment involves an absolute frame whereas in SR there is none.My statement in #4 was within the SR environment.

SR depends on tolerating contradictions. You pick the frame of reference in which you're stationary and assert that the speed of light is the same relative to you in all directions, but you ignore the fact that it can't then be the same in all directions relative to anything else that's moving relative to you. When you deal with something moving relative to you, you then switch to the frame in which it is stationary, and then you repeat the same assertion about light moving at the same speed in all directions relative to it while ignoring the fact that it then won't be the same in all directions relative to you any more. That is the ridiculous, irrational game SR people play time and time again. The fibre-optic-cable-round-the-earth thought experiment helps to show up the stunt they're trying to pull - it takes up the assertion that the speed of light is the same in all directions at any point (relative to that point) and applies that at all points on the circuit, then shows that if the assertions were all true, the light pulses would have to return to the timer simultaneously even though the path is longer one way than the other. But irrational people generally believe they are rational and appear to be incurable, so they blunder on with a disproven theory regardless.

[Edited to tighten up the wording.]

[Kris Kuitkowski]

Hi David,
I'm going for short holidays tomorrow. When I'm back, I'll do some further thinking how we can approach practical aspect of carrying out the experiment.

[David Cooper]

Hi Kris,

If we run the experiment with the clocks stationary and the rod moving, the clocks will count how long it takes for light to reach them from the other clock after the rod starts them running. If we view this from a different frame of reference, we must see the clocks count up the same number of ticks. All frames of reference are believed to act exactly as if they are the one that represents an absolute frame, so no matter how fast the experiment runs through space, the clocks should count up the same values every time you run the experiment and make it impossible to measure the speed of the apparatus through space. Everything happening in the experiment must conform to the normal rules, so it cannot possibly work. And yet somehow, it still sounds as if it should work. There's something really weird going on that needs to be understood. Why does the thought experiment's argument sound so convincing? That's what I'm trying to explore, and I'll keep working on it until I can explain it.

[guest4091]

The SR clock synch method is to send an em signal from the center of the spacing of two clocks with a selected time for the center clock. This method synchronizes two clocks in the same frame and establishes a relative axis of simultaneity (red) for that frame at its current speed. It does not synchronize clocks from different frames.
The first graphic should represent the proposal in Kris #15.
https://app.box.com/s/79xzoqaefscp0meiuq1hlhi2ddvqw3n7
https://app.box.com/s/mun2wugdjlos0hqh7wxubmfcitjub04b
The second graphic shows synchronized clocks for A moving at .3c and B moving at .6c, relative to earth. It also shows the reciprocal observations of slow clocks. With both assuming equal path lengths for light transit times, the clock reading of 1.00 occurs at 1.075 on their local clock.
Speed of B as calculated by A = .366.
1/gamma = .931.
1.075(.931) = 1.00 (removing effect of td)

To establish synchronization between frames, each clock must respond to a signal from the other by sending the current time and one clock adjusted. Here is an example using the second graphic, but advancing the A-clock .32. All times are local.
A send 1.00 to B
B send 1.00 to A
A receive at 1.79
1.00/.931 = 1.074      (adjust B-time for td)
(1.79+1.00)/2 = 1.395   (calc mid point)
1.074 -1.395=- .32      (amount A must adjust its clock)

[guest4091]

david cooper #45
https://app.box.com/s/zlhkh5g3fv56n2t0rd1obd576prd95kg
graphics describe much better than words.
The graphic is a speed plot, thus c is constant at the same angle, and relative speed results in different distances traveled.
Calculations used 186000 mps.
Rotation speed is exggerated to show the tiny differences.
Why would someone be so hostile toward a theory?

[alancalverd]

Imagine a solid cylinder, 30 cm in diameter and 30 cm long. Drill  hole parallel to the axis, near the circumference. Spin the cylinder on its axis. Light can pass through the hole but will be attenuated as you increase the rate of rotation, unless you skew the hole slightly.

Assume you can spin the cylinder at 12,000 rpm (not difficult). It is left as an exercise to the reader to determine the skew angle required to maximise the amount of light that can pass through the hole. Make the experiment easy by drilling lots of holes. Now measure the one-way speed of light from the skew angle and rotation speed that produces the maximum transmitted intensity. 

This "skewed hole" method is routinely used for analysing neutron beams and selecting neutrons of a desired speed, but needs a bit of engineering revision to cope with photons moving at c. Fortunately visible photons can be stopped by very thin sheets of metal, so we can replace the solid cylinder by an axle, maybe 1 km long, with thin discs at each end. perforated by circumferential holes. Put the whole thing inside an evacuated tube and use a laser source. You now have a feasible, updated and unidirectional version of Fizeau's classic 1848 experiment.   

[David Cooper]

Fortunately visible photons can be stopped by very thin sheets of metal, so we can replace the solid cylinder by an axle, maybe 1 km long, with thin discs at each end. perforated by circumferential holes. Put the whole thing inside an evacuated tube and use a laser source. You now have a feasible, updated and unidirectional version of Fizeau's classic 1848 experiment.

How are you coordinating the two discs at the end of your axle to keep them aligned? No matter how rigid your axle is, it will twist due to delays in the communications of forces between atoms, so it adjusts for the speed it's moving through space. If you get rid of the axle and have a separate axle for each disc 1km apart and use radio signals to coordinate their rotation to try to stay aligned with each other, you should realise that they do not stay in sync, and the communication delays that cause them to deviate from the alignment you want them to have apply equally to the functionality of your 1km-long axle.

[David Cooper]

david cooper #45
https://app.box.com/s/zlhkh5g3fv56n2t0rd1obd576prd95kg
graphics describe much better than words.
The graphic is a speed plot, thus c is constant at the same angle, and relative speed results in different distances traveled.
Calculations used 186000 mps.
Rotation speed is exggerated to show the tiny differences.
Why would someone be so hostile toward a theory?

Your diagram shows the speed of light being different in opposite directions relative to each point on the circuit. Why would anyone not be hostile towards the presentation of an irrational theory as a rational one?

[alancalverd]

Here's a simple experiment.

Most airplanes carry a gadget called Distance Measuring Equipment (DME). It is a radio transceiver. About once per second it emits a coded radio pulse. If a DME ground station detects the pulse, it waits 50 microseconds then retransmits the code. The DME receiver subtracts 50 microseconds from the time between transmission and reception, and calculates the distance to the ground station, assuming c is independent of direction. It's a pretty old (1945) system that could be made a lot more accurate nowadays*.

Now suppose the discrepancy between out and return speeds is 2x in one direction, and 2y and 2z in the perpendicular directions. Set up a "precision DME" with three ground stations at known distances d (use a surveyor's chain) in the x, y and z directions, and measure their apparent distances d' with the DME

d_x'  = d_xc²/(c² - x²) and similarly for y and z.

If x, y and z are not all zero, there will be a discrepancy (d' - d) in at least two measurements.

*DME works over distances of up to 200 miles if you are lucky, and reads to +/- 0.1 mile even using thermionic valve technology. No point in making it more accurate for aviation use because the plane is moving at around 0.1mile/second anyway!

[alancalverd]

How are you coordinating the two discs at the end of your axle to keep them aligned?

Drive the axle from its midpoint, so the discs have equal phase lag from the driving point, and wait as long as you like for any second-order phase lag to dissipate.   

[David Cooper]

How are you coordinating the two discs at the end of your axle to keep them aligned?

Drive the axle from its midpoint, so the discs have equal phase lag from the driving point, and wait as long as you like for any second-order phase lag to dissipate.

If the apparatus is not moving through space, the holes in the discs can be placed so that if light passes through one it will be able to pass through the other one too, but only in one direction - if it tries to go through in the opposite direction it will be blocked. We can use light from a laser so that we can use that light and avoid light travelling on other paths. If we now put the apparatus in a rocket and accelerate it to relativistic speed, the light from the laser will pass through it much more quickly, which makes it look as if the light should no longer be able to get through the second hole as that disc hasn't rotated into the right position in time for that to be possible, but movement through the fabric of space affects the coordination of the discs such that the lead one lags behind where you expect it to be and the rear one is ahead of your expectations in terms of its rotation, thus adjusting the apparatus to maintain the ability of the laser light to pass through both holes.

Keep the rocket moving fast but turn it around to move along backwards and what happens to the discs now? The experiment's rear disc now lags (because it is leading the way through space) while the experiment's front disc rotates ahead of schedule (because it's trailing the other disc on its way through space) and the result is that the light still passes through both holes even though it takes longer for it to travel from one hole to the other.

This auto-correction works well for low speeds of travel through space, but it looks as if it might break down at higher speeds, so perhaps this experiment could in theory identify the one-way speed of light if we could move the apparatus fast enough or improve its precision. However, it's really equivalent to a light clock with the mirrors at either end on rotating discs which always put the mirror in the right place just in time to hit the pulse of laser light on what for them is a tick of the clock (or half tick). The coordination of that should not go out of sync just from turning the clock round by 180 degrees - the rotations of the discs should automatically be adjusted to keep the mirrors hitting the light pulses because each rotating disc is directly equivalent to a clock. For that reason, I think the experiment cannot work no matter how quickly you move it through space, even though it feels as if it could work when you consider how long it could take light to travel through the apparatus in the same direction the apparatus is moving and compare that with how quickly light goes through it in the opposite direction. I can see now why the people I know who looked at this experiment last year spent seven months working on it before they decided it couldn't detect the one-way speed of light - there's something counter-intuitive going on with it.

[David Cooper]

Set up a "precision DME" with three ground stations at known distances d (use a surveyor's chain) in the x, y and z directions, and measure their apparent distances d' with the DME

Are you taking into account length-contraction on your surveyor's chain, and how do you handle the variation in the actual distances as the Earth rotates and moves through space?

[alancalverd]

Perhaps you would like to set out your definition of velocity? Or distance? Or time?

Anyway, if the earth's rotation or movement through space makes any difference to the propagation speed of light, you will find that  d_x', d_y' and d_z' vary with time and in relation to each other.

movement through the fabric of space affects the coordination of the discs such that the lead one lags behind where you expect it to be and the rear one is ahead of your expectations in terms of its rotation, thus adjusting the apparatus to maintain the ability of the laser light to pass through both holes.

Same argument applies: you should measure different rotation speeds for maximum illumination depending on the orientation of the axle relative to the direction of travel . No need for a rocket - with the dimensions I have suggested, the rotation of the earth will suffice. 

[David Cooper]

Perhaps you would like to set out your definition of velocity? Or distance? Or time?

Anyway, if the earth's rotation or movement through space makes any difference to the propagation speed of light, you will find that  d_x', d_y' and d_z' vary with time and in relation to each other.

There is no way that whatever it is you're doing there is going to show anything up, so this is a wild goose chase. But if you insist on following it through, I need to know what you're talking about and at the moment I can't make sense of it.

Now suppose the discrepancy between out and return speeds is 2x in one direction, and 2y and 2z in the perpendicular directions. Set up a "precision DME" with three ground stations at known distances d (use a surveyor's chain) in the x, y and z directions, and measure their apparent distances d' with the DME

What exactly are these 2x, 2y and 2z bits supposed to be? You've got radio signals moving on round trips there and back and a slowing of clocks due to movement which cancels out any delays caused by the whole system moving, so how are you going to pick up any discrepancy?

movement through the fabric of space affects the coordination of the discs such that the lead one lags behind where you expect it to be and the rear one is ahead of your expectations in terms of its rotation, thus adjusting the apparatus to maintain the ability of the laser light to pass through both holes.

Same argument applies: you should measure different rotation speeds for maximum illumination depending on the orientation of the axle relative to the direction of travel . No need for a rocket - with the dimensions I have suggested, the rotation of the earth will suffice.

Again there is no way this experiment can pick up its movement through space - it is directly equivalent to a pair of clocks bouncing light back and forth between each other. Imagine a clock with a hand that rotates like on an old clock. On the end of the hand is a mirror, but the clock is covered by a disc with a hole in it to allow the 12 position on the dial to be visible. The mirror appears in that hole once per rotation of the clock's hand, and it appears there every time the light pulse arrives from the other clock at the far end of the apparatus. We set it up with the apparatus at rest with both clocks putting their mirror behind the hole at the right time to intercept the light, and they bounce it back and forth continually. (We can boost it by having a laser shine through one of the mirrors which can be semi-silvered.) If we move the apparatus about, we can accelerate it this way and that, and twist and turn it, but nothing we do will stop those mirrors reaching the holes at the times when the light pulse arrives there - it all automatically corrects and keeps everything in sync, so long as the clocks are sufficiently accurate. If there's a need to maintain the synchronisation of the clocks by sending signals between them, this can be done with radio signals or by connecting them with a long axle - both will do the same job and keep them in sync with the light.

[alancalverd]

Please follow the maths in my previous replies, which shows how any difference between out and return speeds  always leads to a mean round-trip speed less than the maximum. Therefore any discrepancy, however caused, will be revealed by comparing the measured two-way speed in orthogonal directions. If there is no discrepancy, then the one-way speed must equal the two-way speed.  No physics, no relativity, just algebra.