4 ms·

Having extra detectors also helps hugely with uptime. We're currently only active and collecting clean data ~50% of the time for each detector. We need at leas

Having extra detectors also helps hugely with uptime. We're currently only active and collecting clean data ~50% of the time for each detector. We need at least two detectors active for a detection, and need three detectors for this kind of beautiful directional isolation. So having extra detectors decreases our skymap areas significantly, which helps in EM followup, but it also significantly reduces GW detector network downtime. And it allows for graceful degradation; with 5 detectors, you can afford to have a couple detectors down and still detect an event with good direction reconstruction. We also have tons of downtime for upgrades. Now that our second observing run is over, we have a ~1.5yr upgrade cycle. There's no point in running VIRGO without LIGO (except for engineering reasons); having more detectors will increase flexibility in this regard, too. When KAGRA and LIGO India come online, we should see much better skymaps, but we will also see much higher uptimes (and hence higher detection rates). It's especially exciting to see improved direction reconstruction given that VIRGO has faced setbacks this year with their mirrors. I wasn't expecting to see such beautiful skymaps until our third observing run! Needless to say it was an exciting summer :) - grad student working on LIGO

saganus 9y ago
1.5 years for an upgrade? Can you expand on why it takes so much time to do an upgrade? I'm guessing it's due to the complexity of the system, kinda like with the LHC? or is it something different? I mean, having equipment that expensive on downtime for more than a year seems like a lot!
- sleavey 9y ago
  The required time is due to how complicated the instruments are, and how hard it is to tune them such that they achieve the sensitivities they are designed to achieve across such a wide frequency band. In the downtime until O3, new optics are to be installed. As these are 40kg each and suspended from monolithic fused silica fibres (each a couple hundred microns thick), this process is complicated and requires great care to avoid damaging either the fibres or the optics during installation. In addition, both LIGO and Virgo are going to get squeezed light sources installed, which allow for reduction of limiting quantum noise at high frequencies. These are immensely sensitive to light loss in the optical path, so commissioners need to be very careful to minimised scattered light and ensure good alignment. After making all of the planned upgrades, the interferometers then need to be recommissioned: all of the control loops (of which there are thousands) need to be re-tuned from their settings as of now, given the changes in the behaviour of the interferometer due to the upgrades. This takes a lot of person power and a lot of time to get right - for the first observing run this process was basically also 1.5 years.
  - saganus 9y ago
    Wow! Great explanation. Thanks! I just can't start to fathom the complexity of these instruments. Also... how cool that you get to work on such interesting field and with such cool "lab" devices!
  - ssijak 9y ago
    Could be easily put in a SF novel about 3000 something year and I would not blink.
  - stefco_ 9y ago
    Another point worth making: we gain more by upgrading than by running. Since our search volume is naturally the cube of our max search distance, a twofold increase in range yields and eightfold yield in search volume, and hence event rate. Put another way, if you can get better than a 30% increase in search distance by upgrading half of the time and running the other half, you're doing better than breaking even (1.3^3 > 2) because your increase in event rate more than makes up for the downtime. This is why it actually makes more sense for us to do these thorough, ambitious upgrades than to just leave the detectors on all the time.
    - jcims 9y ago
      Has anyone calculated whether or not LIGO et. al. could detect the mass energy conversion of a nuke going off?
      - stefco_ 9y ago
        A nuke wouldn't emit measurable GWs, but it would emit seismic waves. There are seismometers that are highly optimized for this task; LIGO, on the other hand, has extremely advanced damping technology designed to filter out seismic and other environmental vibrations. I don't know the details of nuclear-detonation induced seismic activity, but it seems to me that LIGO would be a very inefficient observational tool. That said, I don't know if anyone has looked at our data in the vicinity of e.g. DPRK nuclear tests to see if any nuclear detonations are visible in the strain timeseries.

saganus 9y ago

1.5 years for an upgrade? Can you expand on why it takes so much time to do an upgrade? I'm guessing it's due to the complexity of the system, kinda like with the LHC? or is it something different? I mean, having equipment that expensive on downtime for more than a year seems like a lot!

sleavey 9y ago
The required time is due to how complicated the instruments are, and how hard it is to tune them such that they achieve the sensitivities they are designed to achieve across such a wide frequency band. In the downtime until O3, new optics are to be installed. As these are 40kg each and suspended from monolithic fused silica fibres (each a couple hundred microns thick), this process is complicated and requires great care to avoid damaging either the fibres or the optics during installation. In addition, both LIGO and Virgo are going to get squeezed light sources installed, which allow for reduction of limiting quantum noise at high frequencies. These are immensely sensitive to light loss in the optical path, so commissioners need to be very careful to minimised scattered light and ensure good alignment. After making all of the planned upgrades, the interferometers then need to be recommissioned: all of the control loops (of which there are thousands) need to be re-tuned from their settings as of now, given the changes in the behaviour of the interferometer due to the upgrades. This takes a lot of person power and a lot of time to get right - for the first observing run this process was basically also 1.5 years.
- saganus 9y ago
  Wow! Great explanation. Thanks! I just can't start to fathom the complexity of these instruments. Also... how cool that you get to work on such interesting field and with such cool "lab" devices!
- ssijak 9y ago
  Could be easily put in a SF novel about 3000 something year and I would not blink.
- stefco_ 9y ago
  Another point worth making: we gain more by upgrading than by running. Since our search volume is naturally the cube of our max search distance, a twofold increase in range yields and eightfold yield in search volume, and hence event rate. Put another way, if you can get better than a 30% increase in search distance by upgrading half of the time and running the other half, you're doing better than breaking even (1.3^3 > 2) because your increase in event rate more than makes up for the downtime. This is why it actually makes more sense for us to do these thorough, ambitious upgrades than to just leave the detectors on all the time.
  - jcims 9y ago
    Has anyone calculated whether or not LIGO et. al. could detect the mass energy conversion of a nuke going off?
    - stefco_ 9y ago
      A nuke wouldn't emit measurable GWs, but it would emit seismic waves. There are seismometers that are highly optimized for this task; LIGO, on the other hand, has extremely advanced damping technology designed to filter out seismic and other environmental vibrations. I don't know the details of nuclear-detonation induced seismic activity, but it seems to me that LIGO would be a very inefficient observational tool. That said, I don't know if anyone has looked at our data in the vicinity of e.g. DPRK nuclear tests to see if any nuclear detonations are visible in the strain timeseries.