[hamdani yusuf (OP)]

I'm pretty sure it can be done, but most of the devices that can do this accurately do so by measuring the wavelength and computing it from that using a known c. That doesn't work if you're trying to measure c.

What device do you think can measure the wavelength accurately with high precision?

Radar speed guns measure the frequency instead of wavelength.
https://en.wikipedia.org/wiki/Radar_speed_gun#How_it_works

Speed guns use Doppler radar to perform speed measurements.

Radar speed guns, like other types of radar, consist of a radio transmitter and receiver. They send out a radio signal in a narrow beam, then receive the same signal back after it bounces off the target object. Due to a phenomenon called the Doppler effect, if the object is moving toward or away from the gun, the frequency of the reflected radio waves when they come back is different from the transmitted waves. When the object is approaching the radar, the frequency of the return waves is higher than the transmitted waves; when the object is moving away, the frequency is lower. From that difference, the radar speed gun can calculate the speed of the object from which the waves have been bounced. This speed is given by the following equation:


where c is the speed of light, f is the emitted frequency of the radio waves and Δf is the difference in frequency between the radio waves that are emitted and those received back by the gun.

By combining the measurement results from both wavelength and frequency, we can confirm (or refute) the constancy of light speed in vacuum.

[alancalverd]

What device do you think can measure the wavelength accurately with high precision?

A diffraction grating is generally best.

[Bored chemist]

I'm pretty sure it can be done, but most of the devices that can do this accurately do so by measuring the wavelength and computing it from that using a known c. That doesn't work if you're trying to measure c.

If we generalise "light" to "the em Spectrum", then we can use microwaves. There are frequency counters that will let you measure microwave frequencies directly to about 8 digits.
If you choose the right frequency- that of a caesium clock- you can measure the beat frequency produced by mixing the incoming microwaves with those produced by the Cs atoms
Doing that lets you measure the frequencies to a few parts in 10^10 relatively easily.
Indeed, that's more or less how the radar traps work, but they are only interested in a ratio of frequencies, not an absolute measurement.

What device do you think can measure the wavelength accurately with high precision?

A diffraction grating is generally best.

Yes and no.
Over a relatively large range of wavelengths (say, visible light), a grating is best.
If you are looking to check a change in wavelength over a narrow band you are probably better off with an etalon
https://en.wikipedia.org/wiki/Fabry%E2%80%93P%C3%A9rot_interferometer

It is likely that astronomers have looked at the microwaves from distant stars (which are moving rapidly WRT us) using both frequency and wavelength measurements.
If the measurements didn't agree, they would have noticed.
So this experiment

By combining the measurement results from both wavelength and frequency, we can confirm (or refute) the constancy of light speed in vacuum.

has been done.
Relativity passed the test.

[alancalverd]

radar traps work, but they are only interested in a ratio of frequencies, not an absolute measurement.

Difference, actually. Much simpler detector, fr - fo generally lies in the audio spectrum, and is directly related to the speed of the target.

[Bored chemist]

Difference, actually.

No.
Delta f/f is a ratio and that's what's proportional to the speed.

If I tell you that the beat frequency is 1000Hz you can not use that to determine the speed.
If I tell you that the ratio of the beat frequency to the transmitted frequency is 1/10,000,000 then you can tell me the speed.
So, you need both the difference and the base frequency.

Radar guns measure a difference in two nearly identical frequencies, not the frequency itself, so you can't use one to just place in an unknown beam of light and get some kind of frequency count from it.
Again, I'm not saying there is no such device, just that I can't readily think of one.

If one of those frequencies is from a Caesium clock then you can ratio the difference from the standard against the standard and get a measurement. However, current tech doesn't have counters fast enough to do this with visible light (yet).

[Halc]

Difference, actually.

No.
Delta f/f is a ratio and that's what's proportional to the speed.

If I tell you that the beat frequency is 1000Hz you can not use that to determine the speed.

It does addition/subtraction of the two signals to get the beat frequency. If the BF is 1000Hz, I can very much use that to determine the speed since the signal was sent at a known frequency, so the ratio can then be computed.
Key word is that known frequency that the device sends out.  When measuring a random beam of light, the frequency is not known since it is what we're supposed to be measuring.

So, you need both the difference and the base frequency.

Which is why a radar gun works, because it has both, and why such a device cannot measure frequency of a light beam because it doesn't have either.

But point taken.  I can put a laser on a train, which puts out a known frequency. I have a similar laser on the platform putting out the exact same frequency.  I can use radar-gun technology to subtract the two signals to get a beat frequency which can thus tell me the frequency of the signal coming from the train. That's pretty much a direct measurement of frequency.  The we take that same light and measure its wavelength, and poof: we have a measurement of light speed utilizing a beam of light whose exact wavelength is not known ahead of time, but is close to one that is.

Side note: The radar gun does not take relativity into account. It wouldn't give accurate readings for cars moving very fast. They don't need to since their use-case is always very low-speed difference between the gun and the target and background.

[yor_on]

 hamdani yusuf " how to demonstrate that speed of light is still c when observed by train A as well as train B? "

Don't think it can be done. This is the way we define it. https://en.wikipedia.org/wiki/Fizeau%E2%80%93Foucault_apparatus and  https://www.britannica.com/science/light/The-Michelson-Morley-experiment

The results are observer dependent as far as I understand it. And that is a weird idea, we use a observer dependent definition as a 'constant', but it is a constant if every observer agree on the result of that experiment no matter how they are defined to move relative something else. https://www.phys.ksu.edu/personal/rprice/SpeedofLight.pdf

Then again, isn't that what we use for all 'repeatable experiments'? It doesn't matter for it what uniform 'relative' motion you are in relative something else, if A and B agree on the outcome then it becomes 'repeatable'. That one you can think of in terms of repeating a equivalent experiment inside differently moving 'black boxes' all in relative motion (different geodesics)  relative each other. As far as I get it that relative motion won't matter for if a experiment is 'repeatable' or not.
=

That light change energy, frequency etc, as in the sinks motion versus the source, doesn't change the speed it is found to propagate at.. (All of this presuming relative motion btw.) It also presume your experimental setup to be 'at rest' with the detector. In a 'same frame of reference' as one expression goes. That is also what makes 'c' 'observer dependent' as you collecting your instruments reading also becomes part of that frame of reference. To get it in any other way is to demand a absolute frame of reference, and that frame was what Einstein looked for until his death.

Another way to express it is in terms of 'locality'. It's a 'locally consistent constant'. The thing that is really weird about it is that you can use any light you like to then let it 'bounce' between mirrors. It will always come out as 'c'. The distance of the lights propagation before it, be it a far away star, or that stars motion relative earth doesn't matter for it. So the 'age' of the light doesn't matter (or the lights source if you like). And anyone between you and the source can do the same experiment getting the same results no matter their own 'relative motion'.

Einstein defined gravity as being a equivalence to acceleration. And the way you define a gravity is by using a scale. So, inside a black box you always will be able to define if you're in a 'acceleration' by weighting yourself. You can feel it without a scale too of course. But in that black box you won't know if you are on earth or in a constant, uniform, acceleration. (ignoring spin for this, which is a acceleration we don't want to introduce)

So, it can be defined as a local 'absolute' frame of reference possibly? But if so, so should the absence of weight be able to be defined as a 'absolute' local frame of reference too, shouldn't it? Using this you then find that all geodesics are the same, no weight to them. And that whatever uniform motion you might want to define to this geodesic has no meaning inside that black box.

[phyti39]

hamdani;

The parallel blue lines from the origin represent the time interval for 1 em cycle.
The speeds of A and B are exaggerated for clarity.
The vertical blue line is absorbed, in a shorter A timeline (blue shift), and a longer B timeline (red shift).
If t is the original period then,

tA=t*sqrt[(1-v)/(1+v)], and

tB=t*sqrt[(1+v)/(1-v)].

ns-hamdani.gif (4.53 kB . 457x485 - viewed 3432 times)

[Bored chemist]

When measuring a random beam of light, the frequency is not known since it is what we're supposed to be measuring.

If I wait a while, I may be able to solve this problem.
The current definition of the second involves a microwave frequency; but there are moves afoot to use optical frequencies.
For example, this clock uses 674 nm light
http://resource.npl.co.uk/docs/networks/time/meeting3/klein.pdf

 at 444 779 044 095 484.3  Hz
That's enough digits for most of us.
So (in principle) I can put one of these on train and another on the platform.
And I can mix the light signals from them and (for some suitable relative velocity) I will get a beat frequency in the radio frequency bands which I can measure with an off-the-peg meter.
That's not terribly practical, but in principle I can do it.

Of course that assumes that we start with the "right" wavelength- but that's not much of a problem.
The question posed in reply 23 uses a laser at 300 nM, but I think that's fairly arbitrary.

Hamdani Yusuf,
would you be prepared to accept using a laser at 674 nm instead of 300 nm?

If we do that then we can measure the frequencies to preposterous precision (in principle- In practice, I doubt that sort of equipment takes kindly to being put on a train and moved about.)

[alancalverd]

When measuring a random beam of light, the frequency is not known since it is what we're supposed to be measuring.

Then don't use a random beam. A sodium lamp will give you a nice spectrum with some well defined lines, or you can use mercury vapor around 300 nm. You can calibrate your etalon or grating on the train by using another such lamp. I guess most folk would go for a laser nowadays but I think the line spectrum of a low-pressure lamp is narrower than any UV laser.

[Bored chemist]

I think the line spectrum of a low-pressure lamp is narrower than any UV laser.

In general, it isn't.

[hamdani yusuf (OP)]

Hamdani Yusuf,
would you be prepared to accept using a laser at 674 nm instead of 300 nm?

No problem,  I used arbitrarily round number just to make calculation easier.

[hamdani yusuf (OP)]

It is likely that astronomers have looked at the microwaves from distant stars (which are moving rapidly WRT us) using both frequency and wavelength measurements.
If the measurements didn't agree, they would have noticed.

Those experiments assumed that interstellar medium has negligible effects, which may or may not be true. To get more reliable answer,  we need a controlled environment.

[alancalverd]

Hamdani Yusuf,
would you be prepared to accept using a laser at 674 nm instead of 300 nm?

No problem,  I used arbitrarily round number just to make calculation easier.

Excellent. And if you are prepared to go to microwave frequencies you can get an easily-measured Doppler shift on a moving train.

[Bored chemist]

Doppler radar guns have been used by police for decades.
They have been tested in court.
If there was any chance that the outcome was different from what was expected, someone would have challenged it.

Those experiments assumed that interstellar medium has negligible effects, which may or may not be true. To get more reliable answer,  we need a controlled environment.

The interstellar medium is a much  better vacuum that we can achieve in our experiments.

[alancalverd]

It strikes me that this solves the "one way speed of light" question that turns up here from time to time.

Let A and B have identical apparatus for sending, receiving and analysing an electromagnetic signal. Each measures the frequency and wavelength of a pulse from the other, and calculates c.

[Bored chemist]

It strikes me that this solves the "one way speed of light" question that turns up here from time to time.

Let A and B have identical apparatus for sending, receiving and analysing an electromagnetic signal. Each measures the frequency and wavelength of a pulse from the other, and calculates c.

Due to budget cuts there's only enough money for 1 set of apparatus, and a mirror.
Does that affect the outcome?

[alancalverd]

According to several correspondents to this forum, it might, and the possibility confirms their belief in the aether, flat earth and fairies.

[Bored chemist]

I can imagine scenarios where it might- for example, if you put a black hole there so its event horizon is just behind the mirror.
But most of the discussions are carefully framed to avoid weird things like that.

[Halc]

It strikes me that this solves the "one way speed of light" question that turns up here from time to time.

Let A and B have identical apparatus for sending, receiving and analysing an electromagnetic signal. Each measures the frequency and wavelength of a pulse from the other, and calculates c.

It does calculate c, but it does not measure the one-way speed of a given light pulse.  Similarly, Roemer first measured c using a one way method, something I occasionally point out to the absolutists. What he did is entirely valid, but measuring c using a one-way method is very different than measuring the one-way speed of light.
Ditto for the measuring the frequency and wavelength thing, both measurements yielding incorrect results in an absolutist interpretation of the universe. The joke is that they claim this is a simplification, and yet they are forced to use Einstein's relativistic (not abolute) mathematics when computing anything. I've never seen the computations of any complex system done the 'simpler' absolutist way. Fairies indeed.

Due to budget cuts there's only enough money for 1 set of apparatus, and a mirror.

If there's a mirror involved, it isn't a one-way experiment. Roemer used no mirror, but he did use a pair of clocks separated by a very large distance. It was the recent invention of the clock that made this first measurement of c possible. At his time, the most accurate clock for long term (months) measurement was still a sundial.

I can imagine scenarios where it might- for example, if you put a black hole there so its event horizon is just behind the mirror.

Again a 2-way measurement which will yield exactly c if you use Alan's method.
And a black hole is not necessary. There are reflectors placed on the moon for such purposes, and a light pulse sent there and back again makes the round trip at faster than light. If they did the same on the moon with a reflector on Earth, the round trip would take place at less than light speed. This is a different method of measuring c than Alan's method which doesn't rely on mirrors or round trips.  But the method assumes relatively stationary equipment, making it invalid for computing one-way speed under a view that does not assume the postulates of relativity.