Hmm, as a GPS geek, I find it entertaining they decided to use the GPS time scale as the log reference. Makes sense, as using GPS time avoids leap seconds chaos. Putting it in straight seconds as opposed to the standard week number/Time Of Week setting makes printing and math easy even though the numbers are big. I assume your logging time scale is coarse enough that the roughly 100 microseconds granularity is OK...you're at 13 significant digits already. When I'm crunching GPS signals to nanosecond levels, I need to keep the seconds field under 10^6, and use some care to prevent accumulated rounding errors.
If I've remembered rightly they have looked in detail on the time standard issue. As you do. Here's a couple of the earlier papers on the topic, here and there and over there.
I think the basic concerns are - what happens if GPS goes down, plus an accuracy under 10us and preferably 1 us ( eventually ). At a foot-per-nanosecond that gives about 1000ft which thus constrains timing between interferometers for a mutually observed signal, and hence angular localisation of source in the sky.
Cheers, Mike.
( edit ) Most importantly :
Quote:
Local time is not used in LIGO (except for lunch).
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
If I've remembered rightly they have looked in detail on the time standard issue. As you do. Here's a couple of the earlier papers on the topic, here and there and over there.
I think the basic concerns are - what happens if GPS goes down, plus an accuracy under 10us and preferably 1 us ( eventually ). At a foot-per-nanosecond that gives about 1000ft which thus constrains timing between interferometers for a mutually observed signal, and hence angular localisation of source in the sky.
Well Mike, those were some very interesting links! Yes, I'm definitely a GPS and timing geek (I work for a defense contractor involved in GPS issues for the military). I agree that cheap commercial GPS timing sources are attractive, but the failure modes are worrisome. Personally, I use a Symmetricom XLi GPS-disciplined Rubidium source in my lab; I got the best quality Rb they offered. I have some Timing Solutions (now part of Symmetricom) gear in my lab as well. For ultimate accuracy, you could run your local Cs reference into a high-quality GPS time transfer receiver (I think Javad and Trimble are the best candidates for that at the moment). With that, you could perform carrier phase common view measurements with NIST and between Hanford and Livingston and match them up to each other to within a couple nanoseconds at worst. That approach is a bit cheaper than the "gold standard" for time transfer called TWSTT; that is what the NIST and USNO and other high end labs use to compare their clocks.
I wonder though...with you guys firing photons back and forth along the arms, maybe you can use some kind of modulation on that to transfer time and not worry about the variations in transport of those photons over fiber optics. Probably overkill though...
"Better is the enemy of the good." - Voltaire (should be memorized by every requirements lead)
Well Mike, those were some very interesting links!
Yeah, I thought you'd like those! I've got a bottom draw, or three, full of such stuff. :-)
Quote:
Yes, I'm definitely a GPS and timing geek (I work for a defense contractor involved in GPS issues for the military). I agree that cheap commercial GPS timing sources are attractive, but the failure modes are worrisome.
Not to put too fine a point on it, LIGO would be stuffed if GPS went down. We absolutely need gold medal standard correlation between IFO's for confirmation of GW signal existence, source characterisation and localisation.
Quote:
Personally, I use a Symmetricom XLi GPS-disciplined Rubidium source in my lab; I got the best quality Rb they offered. I have some Timing Solutions (now part of Symmetricom) gear in my lab as well. For ultimate accuracy, you could run your local Cs reference into a high-quality GPS time transfer receiver (I think Javad and Trimble are the best candidates for that at the moment). With that, you could perform carrier phase common view measurements with NIST and between Hanford and Livingston and match them up to each other to within a couple nanoseconds at worst. That approach is a bit cheaper than the "gold standard" for time transfer called TWSTT; that is what the NIST and USNO and other high end labs use to compare their clocks.
I think I got most of that. Let's hope that strategy is looked at.
Quote:
I wonder though...with you guys firing photons back and forth along the arms, maybe you can use some kind of modulation on that to transfer time and not worry about the variations in transport of those photons over fiber optics. Probably overkill though...
Outstanding! Now that's lateral thinking. They're already dithering/heterodyning with sidebands for cavity length adjustment/lock. Maybe Morse Code then? The ultimate telegraph. :-)
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Maybe Morse Code then? The ultimate telegraph. :-)
Could a fibre-optic cable be used between Hanford and Livermore?
Could that also act as a third arm of the interferometers?... (Or would thermal expansion/contraction physically obliterate anything useful being measured?)
Maybe Morse Code then? The ultimate telegraph. :-)
Could a fibre-optic cable be used between Hanford and Livermore?
It's Livingston, not Livermore...maybe you had Lawrence-Livermore Labs in mind.
Quote:
Could that also act as a third arm of the interferometers?... (Or would thermal expansion/contraction physically obliterate anything useful being measured?)
...and the amplifiers that you would need in between. Vacuum is the best choice, either as a hole in the atmosphere or in space (e.g. LISA).
Well, I think Marty was suggesting transmitting the time signal up the arms to the end stations in vacuo via modulation of the laser carrier frequency. Currently that is done fibre optically outside the vacuum space. There is hardware up each end which needs that signal.
For interferometry the interior of a fibre optic cable is way too noisy for gravity wave detection. There's dispersion ( different speeds for different frequencies ), absorption, lattice imperfections, phonons and other couplings, boundary effects and more. As Bikeman says you can't go 4km with fibre optic without a repeater.
As for third arms, I did ask way long ago about having an extra arm : like perpendicular to the other two. I recall the reply was to be useful it would have to be of similiar length, so at 4km a mineshaft or a tower to the sky isn't practical. It would be neat :-)
Of course, LISA will have three 'arms'.
Cheers, Mike.
( edit ) Or another laser in the beam tube? There's plenty of cross-sectional area.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
... For interferometry the interior of a fibre optic cable is way too noisy for gravity wave detection. There's dispersion ( different speeds for different frequencies ), absorption, lattice imperfections, phonons and other couplings, boundary effects and more. As Bikeman says you can't go 4km with fibre optic without a repeater.
You might need a few fibre amplifiers to go between the two LIGOs, but 4km should be a easy with no amplifiers needed if you choose your wavelength for the fibre.
Shame about there being too much noise with the fibre in the first place...
Quote:
As for third arms, I did ask way long ago about having an extra arm : like perpendicular to the other two. I recall the reply was to be useful it would have to be of similiar length, so at 4km a mineshaft or a tower to the sky isn't practical. It would be neat :-)
Mmmm... A 4km deep shaft might be stretching the drilling/mining a little, and even a Geo-sized 600m depth might still be stretching the finances somewhat!
... Unless that is the end stations can be remote controlled and minaturised so that only a drilled small diameter shaft is needed...
Quote:
Of course, LISA will have three 'arms'.
Likely a much better idea than trying to go "3d" here on Earth!
... Or, how about in the ocean?...
Quote:
( edit ) Or another laser in the beam tube? There's plenty of cross-sectional area.
You might need a few fibre amplifiers to go between the two LIGOs, but 4km should be a easy with no amplifiers needed if you choose your wavelength for the fibre.
True, but beware as the wavelength choice returns to you in terms of fringe shift and like issues ie. how well one can pickup a given spacetime deformation as a phase alteration.
Quote:
Possible problems of stray light interference?
If it is pumped out really well, plus beam dumps etc it should be good. Note that the entire vacuum space and surrounds was designed from the get-go with multiplicity in mind. As already demonstrated by H1 & H2.
I should add the time signal as now going via optical fibre is digital. This means at each repeater the signal is re-formed to stay within tolerances of whatever the distinct digital waveforms are set. Thus avoiding the no-mans land between signal definition ( a '0' or a '1' ? ) that is an error cause with a degraded signal.
Whereas the GW signal ( as converted to phase changes by the interferometer ) is a pure analog process.
The neat thing about three arms at mutual right angle is that actually gives you three two-arm interferometers : X & Y, Y & Z, X & Z. But Lisa effectively does much the same too.
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
4 km would be possible, but you'd have to have a floating observatory moored in pretty deep ocean. Just think of the fun and games they'd have trying to keep everything rigid and locked-in to within a fraction of a wavelength, especially for the two surface arms.
RE: Hmm, as a GPS geek, I
)
If I've remembered rightly they have looked in detail on the time standard issue. As you do. Here's a couple of the earlier papers on the topic, here and there and over there.
I think the basic concerns are - what happens if GPS goes down, plus an accuracy under 10us and preferably 1 us ( eventually ). At a foot-per-nanosecond that gives about 1000ft which thus constrains timing between interferometers for a mutually observed signal, and hence angular localisation of source in the sky.
Cheers, Mike.
( edit ) Most importantly :
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
RE: If I've remembered
)
Well Mike, those were some very interesting links! Yes, I'm definitely a GPS and timing geek (I work for a defense contractor involved in GPS issues for the military). I agree that cheap commercial GPS timing sources are attractive, but the failure modes are worrisome. Personally, I use a Symmetricom XLi GPS-disciplined Rubidium source in my lab; I got the best quality Rb they offered. I have some Timing Solutions (now part of Symmetricom) gear in my lab as well. For ultimate accuracy, you could run your local Cs reference into a high-quality GPS time transfer receiver (I think Javad and Trimble are the best candidates for that at the moment). With that, you could perform carrier phase common view measurements with NIST and between Hanford and Livingston and match them up to each other to within a couple nanoseconds at worst. That approach is a bit cheaper than the "gold standard" for time transfer called TWSTT; that is what the NIST and USNO and other high end labs use to compare their clocks.
I wonder though...with you guys firing photons back and forth along the arms, maybe you can use some kind of modulation on that to transfer time and not worry about the variations in transport of those photons over fiber optics. Probably overkill though...
"Better is the enemy of the good." - Voltaire (should be memorized by every requirements lead)
RE: Well Mike, those were
)
Yeah, I thought you'd like those! I've got a bottom draw, or three, full of such stuff. :-)
Not to put too fine a point on it, LIGO would be stuffed if GPS went down. We absolutely need gold medal standard correlation between IFO's for confirmation of GW signal existence, source characterisation and localisation.
I think I got most of that. Let's hope that strategy is looked at.
Outstanding! Now that's lateral thinking. They're already dithering/heterodyning with sidebands for cavity length adjustment/lock. Maybe Morse Code then? The ultimate telegraph. :-)
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
RE: Maybe Morse Code then?
)
Could a fibre-optic cable be used between Hanford and Livermore?
Could that also act as a third arm of the interferometers?... (Or would thermal expansion/contraction physically obliterate anything useful being measured?)
Keep searchin',
Martin
See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
RE: RE: Maybe Morse Code
)
It's Livingston, not Livermore...maybe you had Lawrence-Livermore Labs in mind.
...and the amplifiers that you would need in between. Vacuum is the best choice, either as a hole in the atmosphere or in space (e.g. LISA).
CU
Bikeman
Well, I think Marty was
)
Well, I think Marty was suggesting transmitting the time signal up the arms to the end stations in vacuo via modulation of the laser carrier frequency. Currently that is done fibre optically outside the vacuum space. There is hardware up each end which needs that signal.
For interferometry the interior of a fibre optic cable is way too noisy for gravity wave detection. There's dispersion ( different speeds for different frequencies ), absorption, lattice imperfections, phonons and other couplings, boundary effects and more. As Bikeman says you can't go 4km with fibre optic without a repeater.
As for third arms, I did ask way long ago about having an extra arm : like perpendicular to the other two. I recall the reply was to be useful it would have to be of similiar length, so at 4km a mineshaft or a tower to the sky isn't practical. It would be neat :-)
Of course, LISA will have three 'arms'.
Cheers, Mike.
( edit ) Or another laser in the beam tube? There's plenty of cross-sectional area.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
RE: RE: RE: Maybe Morse
)
Yep on all counts. Just a case of mistyping faster than the speed of thought!
Keep searchin',
Martin
See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
RE: ... For interferometry
)
You might need a few fibre amplifiers to go between the two LIGOs, but 4km should be a easy with no amplifiers needed if you choose your wavelength for the fibre.
Shame about there being too much noise with the fibre in the first place...
Mmmm... A 4km deep shaft might be stretching the drilling/mining a little, and even a Geo-sized 600m depth might still be stretching the finances somewhat!
... Unless that is the end stations can be remote controlled and minaturised so that only a drilled small diameter shaft is needed...
Likely a much better idea than trying to go "3d" here on Earth!
... Or, how about in the ocean?...
Possible problems of stray light interference?
Regards,
Martin
See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
RE: You might need a few
)
True, but beware as the wavelength choice returns to you in terms of fringe shift and like issues ie. how well one can pickup a given spacetime deformation as a phase alteration.
If it is pumped out really well, plus beam dumps etc it should be good. Note that the entire vacuum space and surrounds was designed from the get-go with multiplicity in mind. As already demonstrated by H1 & H2.
I should add the time signal as now going via optical fibre is digital. This means at each repeater the signal is re-formed to stay within tolerances of whatever the distinct digital waveforms are set. Thus avoiding the no-mans land between signal definition ( a '0' or a '1' ? ) that is an error cause with a degraded signal.
Whereas the GW signal ( as converted to phase changes by the interferometer ) is a pure analog process.
The neat thing about three arms at mutual right angle is that actually gives you three two-arm interferometers : X & Y, Y & Z, X & Z. But Lisa effectively does much the same too.
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
RE: ... Or, how about in
)
4 km would be possible, but you'd have to have a floating observatory moored in pretty deep ocean. Just think of the fun and games they'd have trying to keep everything rigid and locked-in to within a fraction of a wavelength, especially for the two surface arms.