Wang / Sagnac Devices
in [Relativity]
|
Prev: Relativity ring problem - what shape is this?
Next: BUY CHEAP TEXTBOOKS | College Textbooks | Used Textbooks |
From: Tom Roberts on 30 Oct 2009 10:24 Darwin123 wrote: > On Oct 27, 10:02 am, Tom Roberts <tjroberts...(a)sbcglobal.net> wrote: >> Jonah Thomas wrote: >>> I haven't seen an argument yet why there shouldn't be frequency >>> differences in Sagnac. >> A light source emits periodic waves, but as I have discussed before, one >> must consider a short light pulse. > One can validly describe the system in terms of wave packets. Using wave packets can be problematical, if the dispersion is large within the bandwidth of the packet. I misspoke slightly: "short light pulse" can be replaced by "short region of a light wave". > This > is what you are describing. No. I described successive wavecrests of a monochromatic light wave. > The spectral bandwidth of each pulse is > not effected by Doppler shift. For very short pulses with very large bandwidths, Doppler shift does affect the bandwidth; this is usually minor. > I think this spectral bandwidth is what > you mean by saying the "frequency is not changed." Not at all. I meant the frequency of a monochromatic light wave. Not its bandwidth. > However, the > carrier wave frequency will be changed by Doppler shift. Yes, if Doppler shift applies. For the case I described it doesn't apply. That was the point: I EXPLICITLY showed that the frequency at the detector is the same as the frequency at the source. This is true for any rigid apparatus that constrains the light paths to not change in length, no matter how it moves as long as accelerations are small. But for an idealized circular Sagnac interferometer, as long as the light paths remain rigid the acceleration due to rotation can be large, as it is orthogonal to the light paths. Tom Roberts
From: Darwin123 on 1 Nov 2009 16:49 On Oct 30, 9:24 am, Tom Roberts <tjrob...(a)sbcglobal.net> wrote: > Darwin123 wrote: > > On Oct 27, 10:02 am, Tom Roberts <tjroberts...(a)sbcglobal.net> wrote: > >> Jonah Thomas wrote: > >>> I haven't seen an argument yet why there shouldn't be frequency > >>> differences in Sagnac. > >> A light source emits periodic waves, but as I have discussed before, one > >> must consider a short light pulse. > > One can validly describe the system in terms of wave packets. > > Using wave packets can be problematical, if the dispersion is large within the > bandwidth of the packet. > > I misspoke slightly: "short light pulse" can be replaced by "short region of a > light wave". > > > This > > is what you are describing. > > No. I described successive wavecrests of a monochromatic light wave. > > > The spectral bandwidth of each pulse is > > not effected by Doppler shift. > > For very short pulses with very large bandwidths, Doppler shift does affect the > bandwidth; this is usually minor. > > > I think this spectral bandwidth is what > > you mean by saying the "frequency is not changed." > > Not at all. I meant the frequency of a monochromatic light wave. Not its bandwidth. > > > However, the > > carrier wave frequency will be changed by Doppler shift. > > Yes, if Doppler shift applies. For the case I described it doesn't apply. That > was the point: I EXPLICITLY showed that the frequency at the detector is the > same as the frequency at the source. This is true for any rigid apparatus that > constrains the light paths to not change in length, no matter how it moves as > long as accelerations are small. But for an idealized circular Sagnac > interferometer, as long as the light paths remain rigid the acceleration due to > rotation can be large, as it is orthogonal to the light paths. > > Tom Roberts You are right and I was wrong. However, I am not sure where my mistake was. Some articles on the Sagnac effect refer to "transverse" and "longitudinal" Doppler shifts. Supposedly, the "transverse Doppler shift" only occurs in Lorentz invariant systems, not in Galilean invariant systems. I will look at these articles more closely to see what exactly they meant by "Doppler shift."
From: Tom Roberts on 2 Nov 2009 09:43
Darwin123 wrote: > Some articles on the Sagnac effect refer to "transverse" and > "longitudinal" Doppler shifts. Supposedly, the "transverse Doppler > shift" only occurs in Lorentz invariant systems, not in Galilean > invariant systems. I will look at these articles more closely to see > what exactly they meant by "Doppler shift." Any sensible article will use "Doppler shift" to mean the change in frequency or wavelength of a light ray (or EM signal) between its emission from a moving source and its observation by a detector at rest in a given inertial frame. This depends not only on the speed of the source, but also on its direction of motion relative to the line-of-sight from source to detector. Transverse Doppler shift refers to the fact that in SR, the Doppler shift does not go to zero when the source is moving exactly perpendicular to the line-of-sight to the detector. Indeed, it is essentially the same as "time dilation". This is absent in pre-SR theories, which give zero Doppler shift for such transverse motion. Tom Roberts |