“We don’t really know the speed of light”

[andrew s]

36 minutes ago, vlaiv said:

I don't think there is a problem.

All I can say is the consensus is it can't be done.  Without a detailed operational analysis of how exactly it would be done it's impossible to say. For example, the slots are moving in a rotating frame of reference so how does that impact the angle? How would you check it?  It's beyond my abilities to compute the effect.

Have a look a the section on "experiments in which light follows a unidirectional path"  here . 

Regards  Andrew 

 

Edited February 19, 2023 by andrew s

[vlaiv]

19 minutes ago, andrew s said:

All I can say is the consensus is it can't be done.  Without a detailed operational analysis of how exactly it would be done it's impossible to say. For example, the slots are moving in a rotating frame of reference so how does that impact the angle? How would you check it?  It's beyond my abilities to compute the effect.

Have a look a the section on "experiments in which light follows a unidirectional path"  here . 

Regards  Andrew 

 

Ok, I think I understand where "the catch" is - in the case of above experiment.

It might not be so obvious because we have what we perceive as "single piece of equipment" and our brain tells us that for one piece of equipment "same conditions" apply - but if we rephrase experiment in different way - it becomes obvious that we need to do slow clock synchronization.

Say that instead of rotating discs with notch in them - we simply have mechanical shutters that are independent - but are set to open "at certain interval, one after another".

Such devices would need to be synchronized to actually open / shut in sequence with certain delay viewed from our frame of reference.

Implicitly, such synchronization happens when we start to spin our device - that motion is in fact not transferred instantaneously from one end to another - it is after all mediated by same mechanism of EM force between atoms in the machine, so it takes some time for motion to arrive at the other end.

 

[Brutha]

8 hours ago, vlaiv said:

Implicitly, such synchronization happens when we start to spin our device - that motion is in fact not transferred instantaneously from one end to another - it is after all mediated by same mechanism of EM force between atoms in the machine, so it takes some time for motion to arrive at the other end.

This is measurable though. You have a sensor next to each disk, that triggers a signal when the notch passes, and a device in the middle that checks the difference. Then you slowly ramp up the rotation till you get the timing difference you want.

EDIT - oh wait, I’ve assumed isotropic speed of electrons in the wire, haven’t I?!

Edited February 20, 2023 by Brutha

[andrew s]

1 hour ago, Brutha said:

EDIT - oh wait, I’ve assumed isotropic speed of electrons in the wire, haven’t I?!

Yes. It's a bit like trying to design a perpetual motion machine. Cannot be done.

Of course the impossibility is now build into the SI system of units.

Regards Andrew 

[Brutha]

2 hours ago, andrew s said:

Yes. It's a bit like trying to design a perpetual motion machine. Cannot be done.

Of course the impossibility is now build into the SI system of units.

Regards Andrew 

Thanks - now I get it!

As @vlaiv said, it takes time for any change to propagate along the apparatus, and this will depend on the speed of light in that direction.

If the speed of light is faster coming down the tube towards me, it will also take longer for the rotation to move up the tube.

So, if I've understood correctly, my experiment is also actually measuring the *two way* speed of light!

[vlaiv]

2 minutes ago, Brutha said:

So, if I've understood correctly, my experiment is also actually measuring the *two way* speed of light!

To be honest, I'm slightly confused about all of it and here is why:

At first I thought that forward and backward speed "add up" to total two way speed - but on closer scrutiny - it does not seem to be the case.

Let's analyze what is happening. In two way speed experiment - two speeds need to add up to constant value, so if speed in one direction is greater - it must be smaller in opposite direction, right? (or otherwise we have same speeds in both direction). That is the case we are claiming we can't distinguish.

If we start rotation of machine on one end, we can distinguish two cases - we can either rotate it clockwise or anticlockwise. These two cases differ in what sort of effect they produce.

One of them will be reducing angle and other will be enlarging angle between notches due to limited speed of light (limited propagation speed of rotation due to EM interactions between atoms).

If we rotate so that angle reduces - we will measure higher return speed of light. If we rotate so that angle increases - we will measure lower return speed of light.

In either of the two cases it holds that angle difference will be larger if forward speed is smaller (start of machine will have more time to spin while rotation propagates to the end of it).

We also have following parameters that we can use to play around with setup:

Driving of rotation does not need to be on one of the ends - it can be placed arbitrarily along the shaft of the machine - and if we have different propagation speed in each direction - there will be difference in how each notch leads or trails depending on all of that.

We can also do "symmetric" setup - we can use same machine setup, driven from the center (or from each of the sides as a control) with two light sources and two detectors that measure speed of light in each direction.

We would then need to explain why is angle from left to right larger than from right to left if there is asymmetry in speed of light in these directions.

Finally - what we haven't considered is fact that notches are in non inertial frames of reference, but I can't imagine how could that be important as we can reduce rotation speed arbitrarily by precision notch separation and by increase in machine size.

[Brutha]

8 minutes ago, vlaiv said:

 

If we start rotation of machine on one end, we can distinguish two cases - we can either rotate it clockwise or anticlockwise. These two cases differ in what sort of effect they produce.

Are those actually different though in the end? Since in one case for the experiment to work the angle difference has to be negative, but in the other positive? 

I.e. if we rotate clockwise, I need the notch at the bottom to be e.g. 1 degree "behind" (as per a clockface assuming we are looking up the tube) the notch at the front.  But if I rotate anticlockwise, the notch at the bottom has to have the angle 1 degree "ahead".

[vlaiv]

2 minutes ago, Brutha said:

Are those actually different though in the end? Since in one case for the experiment to work the angle difference has to be negative, but in the other positive? 

I.e. if we rotate clockwise, I need the notch at the bottom to be e.g. 1 degree "behind" (as per a clockface assuming we are looking up the tube) the notch at the front.  But if I rotate anticlockwise, the notch at the bottom has to have the angle 1 degree "ahead".

Yes, you are right.

Rotation direction must match order of notches for light to pass in one direction. This also rules out symmetric case that I was talking about - as this "light filter" seems to be uni directional.

[vlaiv]

Well - then it is easy to explain why it is two way experiment after all.

Rotation will always tend to "close" the gap.

Smaller forward speed will close the gap more - which will imply higher return speed (straighter path means light must travel faster to get thru), and vice verse - higher forward speed will impact gap less thus return speed can remain smaller.

 

[vlaiv]

Ok, here is another experiment - let's see if we can spot the problem in this one:

Setup consists out of 5 stations all on the same line and equally spaced (well stations 1 and 5 need not be, but for simplicity let's say they are).

Station 1 and 5 only need to do two things:

a) shine a laser one towards the other

b) record if laser light is detected

Station 3 needs to only send light pulse periodically towards stations 2 and 4 (sync pulse).

Stations 2 and 4 need to do the following:

They each have two gates, upon detection of sync pulse they each need to open one gate for short period of time and then close it. Then they both wait agreed upon time period and repeat opening for short period of time with second gate.

Only difference is which gate is opened first.

Gates are of course aligned with the line and sit in laser paths.

Question is:

Will both Stations 1 and 5 either detect or not detect laser light depending on agreed upon time if speed of light is asymmetric?

[Brutha]

51 minutes ago, vlaiv said:

upon detection of sync pulse

Won't the sync pulse also be affected by the change in speed of light?

[vlaiv]

14 minutes ago, Brutha said:

Won't the sync pulse also be affected by the change in speed of light?

Yes of course - both speed of light to be detected and sync pulse change depending on direction - question is does this cancel out?

 

[Brutha]

49 minutes ago, vlaiv said:

Yes of course - both speed of light to be detected and sync pulse change depending on direction - question is does this cancel out?

 

I think so! But we'd need to do some examples with maths to be sure I think!

[vlaiv]

26 minutes ago, Brutha said:

I think so! But we'd need to do some examples with maths to be sure I think!

Well, if I do the math - it turns out that it will only work if speed is equal in both directions

a - some part of speed of light - speed of light is set to 1 and distance is set to 1 between stations (for ease of calculation), so in one direction speed of light is a, while in opposite it is 2-a (say 0.8c and 1.2c - it must add up to 2c so that result of measurement of speed of light in round trips is always c)

t0 - initial time

ta1 = t0 + 1/a
(time of first gate at Station 2)

ta2 = t0 + 1/a + dt
(time of second gate at Station2 - after some pre agreed time dt)

tb1 = t0 + 1/(2-a)
(time of first gate at Station 4)

tb2 = t0 + 1/(2-a) + dt
(time of second gate at Station 4)

i1 = tb2 - ta1 = t0 + 1/(2-a) + dt - t0 -1/a = 1/(2-a) - 1/a + dt
(interval of time for light to travel in one direction)

i2 = ta2 - tb1 = t0 + 1/a + dt - t0 - 1/(2-a) =  1/a - 1/(2-a) + dt
(interval of time for light to travel in opposite direction)

i1 = 2/a = 1/(2-a) - 1/a +dt => dt = 2/a + 1/a - 1/(2-a) = 3/a - 1/(2-a)
(at what dt will light make it in one direction?)

i2 = 2/(2-a) = 1/a - 1/(2-a) + 3/a - 1/(2-a)
(will it make it in other direction in same time?)

2/(2-a) = 4/a - 2(2-a)

1/(2-a) = 2/a - 1/(2-a)

2/(2-a) = 2/a

a = 2 - a

a = 1

Yes, only if a=1 and 2-a also = 1 - or speed of light is equal in both directions.

[Brutha]

2 hours ago, vlaiv said:

a) shine a laser one towards the other

Question - are these supposed to happen at the same time, t0? If so, how do ensure that?

[vlaiv]

12 minutes ago, Brutha said:

Question - are these supposed to happen at the same time, t0? If so, how do ensure that?

t0 is just time at which Station 3 sends sync pulse, it's not really important as it cancels out in most of the calculations.

Laser light can be continuously on.

It will either be blocked at gate at Station 2 or gate at Station 4.

Gates must be opened in just the right time for short duration (think of your example with spinning notches - open gate = aligned notch) for light to pass thru to the other side.

No need to track when the light pulse starts or ends - it is just a detection of light that counts - either it passes at some moment or it does not pass at all.

[Ouroboros]

@vlaiv Your method looks good to me. However, I have been reading a bit about the various cunning schemes that have been proposed to measure the speed of light in one direction only. Without exception they have all been shown to have really measured the speed of light in both directions. I’m not able to work out why or whether your scheme is any different, but it is sufficiently simple to imagine that someone would have come up with this or similar if it were a viable runner.  On the hand … who knows, maybe you’ve cracked it.  

[vlaiv]

15 minutes ago, Ouroboros said:

@vlaiv Your method looks good to me. However, I have been reading a bit about the various cunning schemes that have been proposed to measure the speed of light in one direction only. Without exception they have all been shown to have really measured the speed of light in both directions. I’m not able to work out why or whether your scheme is any different, but it is sufficiently simple to imagine that someone would have come up with this or similar if it were a viable runner.  On the hand … who knows, maybe you’ve cracked it.  

Well, above does not actually measure speed of light in any direction.

It is intended to show if speed of light is equal in both directions or not. Showing that it's equal and using two way measurement combined ensures that we know speed in any direction (as it is the same) - it is like measuring one way speed.

There might be a catch in setup though - how do we determine if Stations are equally spaced. For great enough distances we must use speed of light - at least it seems.

I've also came up with modified setup that does not need any such measurement - but rather simple trigonometry. Stations won't be aligned on a line - but rather at corners of equilateral triangle (for simplicity).

Middle station will use mirror to bounce lasers towards respective stations - and all will use simple 60 degree angle as alignment tool. Actual length of equilateral triangle is not important either if we perform series of tests with ever increasing delay between gate openings.

 

[vlaiv]

5 minutes ago, vlaiv said:

Well, above does not actually measure speed of light in any direction.

I have to be more specific about this statement.

If we accurately measure (somehow) distance between stations and space them equally - then we can use above method to measure speed of light (although not very efficiently - one might call it brute force method).

If we don't measure distance between stations and use trigonometry to ensure that distance is the same but unknown - then we can't measure speed of light - but we can ensure that it is equal in different directions.

[Grogfish]

This has been an interesting topic to read up on because I don’t remember ever hearing about this at uni - so it’s a nice new problem to think about.

Given some of the very famous names to have looked at this, I think it’s a case of “only fools go where angels fear to tread”, but I’m game!!

I’ve been wondering if some indirect methods might work to avoid the need to measure the round trip time. And it was interesting to see that Mossbauer appears to have experimented in this area, as I wondered if his effect might be useable. However, I suspect he would have investigated it himself if there is!
 

A different idea that comes to mind is to consider the angular distribution of photon energies from bremmstralung radiation. If it could be shown that the energy distribution curve is the same at different angles then I think that would indirectly demonstrate the speed of light would have to be the same in different directions. In fact, it could be that you see the change in the shape of the energy curve at different angles particularly at low energies; as you’d effectively be measuring the average speed of c across small section of the electron travel path. / but I guess that would really depend on how c varies.

I think in practice though it might be hard to get precise enough measurements to constrain the variability of c as much as would be nice.

And may be the Achilles heel of this method would be finding a conceivable energy detector that itself wouldn’t be effected by the same change in light speed along the same vector?

Any thoughts on this? Am I barking up the wrong tree?

 

[Gfamily]

I have a feeling that Noether's Theory states that angular momentum would not be conserved if systems were not symmetrical under rotation. 

If c varied by angle of propagation,  rotational symmetry would be broken and angular momentum would not be conserved. 

Since angular momentum is conserved, i am confident that c does not vary by angle of propagation. 

I don't expect anyone will be willing to accept my view without looking into it further, but if anyone can give a rigorous rebuttal I'll be happy to agree. 

[vlaiv]

22 hours ago, Gfamily said:

I don't expect anyone will be willing to accept my view without looking into it further, but if anyone can give a rigorous rebuttal I'll be happy to agree. 

More I think about this argument - more I'm convinced that is in fact true and very good explanation why speed of light must be the same in all directions.

I think that same can be concluded (or above argument augmented) from relation of speed of light to magnetic permeability and electric permittivity constants.

[andrew s]

22 hours ago, Gfamily said:

I have a feeling that Noether's Theory states that angular momentum would not be conserved if systems were not symmetrical under rotation. 

If c varied by angle of propagation,  rotational symmetry would be broken and angular momentum would not be conserved. 

Since angular momentum is conserved, i am confident that c does not vary by angle of propagation. 

I don't expect anyone will be willing to accept my view without looking into it further, but if anyone can give a rigorous rebuttal I'll be happy to agree. 

This would be true in an asymptomatically flat space time. 

However, this is not the case in our Universe where GR tells us it is not flat.

This means the traditional conservation laws only hold in small regions. They don't hold globally. For example time reversal symmetry fails along with the conservation of energy.

Locally, in a small enough region it is flat and  SR  hold along with the normal conservation laws. 

Locally we always measures the speed of light to be c and experiments also show it is isotropic to within the limits we can measure to. 

Regards Andrew 

Edited February 23, 2023 by andrew s

[jetstream]

Heres a post that adds zero to the thread.

I dont understand any of it but I sure like reading about this.

[Deadlake]

I’ve not read the whole post, however ways of measuring the velocity of light are 

1) Reflection, where a mirror is used to return the light.

2) Synchronised clocks reporting the velocity of light between two points?

And the aim is to measure the velocity of light one way?

Well one way (no pun) of measuring the one way velocity is to have light bend back on itself, just need a black hole to bend time/space and light will loop around and come back. The astrophysics are using this technique to find out new physics around the photon rings around back holes….