There's been a fair amount of triple-coincidence time during the past day. But the weather's been on the blustery side during the few days prior. At Hanford, when tumbleweeds aren't being baled during the day, they can be heard at night being blown by the winds against parts of the buildings. Of course, when the wind is that bad, it's difficult to maintain lock, tumbleweeds or not...
There was also this entry (from a couple days ago) regarding H2's ETMy:
Quote:
H2 spontaniously had another ETMY event. They are happening with increasing frequency: we're up to 3 times in the past 6 days.
I performed the standard prescription of ringing etmy up to 700 in pos, then returning to normal locking. Hugh got it to CM and all evidence of the problem was gone.
I'll hazard a guess that it's microscopic contaminant stuck between two mating surfaces in the suspension somewhere... but there are many things to consider:
Quote:
On Sunday night, H2's ETMy exhibited it's BAD NOISE again. This occasion was special in that the noise appeared while the ifo was Powered UP and in Science Mode (#1902). There was thought of an injection being the cause of the event because there was a Burst Injection around the time H2 went bad.
Trends (attached) show that the Burst Injection (11/20 4:00:16utc) doesn't line up w/ the BAD NOISE event. AS_Q (via ETMy) became noisier at 4:06:01utc. So the injection and our noisy ETMy are NOT correlated.
I looked at 100s of other signals around the time the BAD NOISE re-appeared, but I couldn't find anything obvious---darn!
Also, a tremendous amount of work continues to be done with the Control and Data Systems (CDS) at Hanford, to replace one of the primary servers. So far, using the newer hardware, there have been quite a few comments on how much faster systems can be rebooted. Even the automated Figures of Merit take less time to produce:
Quote:
As proof that the new hanford1 is faster, roboscimon now
takes 10 minutes compared to 20 for the previous hanford1.
So now roboscimon (started at 8am) posts before robofom
(starts at 8:15am).
Hopefully this also results in faster recovery times after a lock loss... :)
I found quite a few interesting plots. For the first one, I've included the code that sifts through the data and generates the chart. Does anyone know if the code is MATLAB? (I'll try and find out)
Averaging the duty cycle over the course of a week, and then plotting the averages for the dates shown (from the code you can pick out that it is for Hanford's H1).
The next two just show the plots that were generated (omitted the code for them), which show the progress in the range pretty well. The first one is for H2's inspiral range, and the second one has L1, H1, and H2 all plotted together.
Meanwhile, work continues with the CDS at Hanford (porting files, etc. to the new server), and H2's mirror in the Y-arm (H2 ETMy) continues to give problems from time to time, and it's not easy for me to understand much of what's being looked at, but a bit of ringing is still the remedy, and the interferometer goes through the stages to regain lock without too much delay.
And here's the latest plot of the range. Never seen Hanford's H1 and H2 looking better! :)
Here's the latest plot of the NS/NS Inspiral Ranges. I added notes to show the corresponding elog entries, to show what the trouble was.
Note 1:
Quote:
2K:
---------------
Locked in SM approx. 6 hours of the shift. Range was about 7.3 MPc.
We had a lockloss at approx. 22:30, which gave us some difficulty in getting back to a good alignment. Y arm in particular was in bad shape, and the AS port didn't look right. The small optics didn't show much inmon drift. (We had suspicions of thermal drift from the cold weather/ heater stage startup.)
It wouldn't lock state 3s, so ran an alignment including some manual tweaks on FMY pitch. The ifo grudgingly engaged wfs after a second try, and we gave it wfs and qpd relief before going back to sm. We may still have some lingering alignment issues. Range is presently at 7.5 MPc.
Note 2:
Quote:
H2's ETMY strikes again!
or "Old Jalopy Effect"
Justin
H2-1973 ended when Keita [Keita subsequently mentioned it was John] noticed that the asq spectrum jumped up several orders of magnitude in 100 Hz region. We did some quick forensic analysis, then broke lock and followed the standard recipe. Luckily the gods were on our side and we recovered to SM in record time (about 25 minutes).
And the summary for the Scimon Owl shift looked like this:
Quote:
Reasonably quiet shift.
H1 locked entire shift, dropped out once by AS trigger at the same time H2 dropped lock due to quake (see Justin's elog).
H2 lost lock three times. The first lock loss was nasty in that the alignment was completely screwed (see Mark's elog). The second one was due to the quake mentioned above.
The third one was due to the return of the dreaded noise. We had to kill the lock and shake ETMY, but Justin had done a respectable job and the downtime was minimal.
No injection during the shift.
A few big, short, broad glitches in H1 (but not in H2 and L1), which was easily visible in AS_Q spectrum, but not in AS_Q time series. Cause unknown. See attached for the biggest thing. Maybe the same thing as Fred talked about yesterday. In the time series (which is unfiltered), you really cannot tell the difference between red, blue, green and the brown, but the glitch is only in green segment.
Some alarm about H2 EY OL FOM and H1 MC FAST.
Some vacuum things are red, but I don't know what it is (see JohnM's elog).
I didn't check to see what happened at Livingston, guessing without looking that it was trains (seismic). The time in Science Mode looks better than 90%!
Things are going well at Hanford. The duty cycle numbers for 12/3 are posted, and H1 was 94.2%, and H2 was 95.2%, and for Livingston, L1 was 82.0%. Lots of seismic activity at Livingston, with some more logging going on. Also more baling activity at Hanford, but there, the interferometers were still able to maintain lock, and so the segments were flagged for 'vehicle activity'.
The data collected during a period of time in Science Mode is referred to as a 'segment', and so the S5 Science Run will continue until all the segments (for L1, H1, and H2) add up to a full year of observation time, in triple coincidence. There are a couple of projections for when S5 will be completed, and both are in the third quarter of '07, IIRC. (I'll try to show the charts for both projections, in a later post.)
The long segments for H1/H2 were finally ended by seismic events. Here's what it looks like to one of the various seismic sensors when a helicopter flies over the observatory:
The spectrogram covers about 4 minutes, and the helicopter's noise is seen mainly in the 15 Hz level (and multiples of that).
Some of you may already have noticed the link on the Server Status page (whose link is on the E@H main page) called Graphical display of S5 progress. There you will see 4 charts: Registered Users/Hosts in past 24h, Active Hosts, Percent Done, and Floating Point Speed. And if you've checked recently, then you know we're getting close to being 70% done with S5.
I think this pertains, however, to the 'S5 R1' search (for spinning neutron stars) that we're currently crunching. It's not quite the same as the 'S5 Science Run'. Hopefully I'm explaining the difference correctly.
Most of us, conscientious crunchers that we are, have probably seen the Nov. 6 news item on the main page, highlighting 50% completion of the first S5 search. And along with the news was the estimate of Feb. 10, '07, for sending out all remaining workunits.
But if you haven't visited the Cruncher's Corner, then you may be interested to know that RandyC has a thread for submitting your own estimate (Post your S5R1 completion estimates here), and that Randy's (quite sophisticated) estimate is for the end of January, possibly the 26th. I'm happy to say that I checked today's Percent Done chart (showing the last 2 weeks), and came up with this:
So with 2% more crunched for every 3-day's worth of crunching, then to do another 33% will take 49.5 days. Starting at Dec. 7 and counting the days, I too came up with a completion time of Jan. 26 ( ~ midday :)).
So how much longer until a whole year's worth of 3C (triple coincidence) data has been collected?
The pre-run estimates for the duration of S5 was 18 months, or 548 days. Of course, as time progresses, it's possible to get a better idea of how long the run will take. For example, at the half-way point, you could figure that the second half will take just as much time as the first half, basing your estimate on the whole run so far. But this estimate includes downtime for commissioning work that won't have to be repeated during the second half of the run. So a better estimate might be to just consider a more recent interval of time, like the last 100 days.
Turns out, if you base your estimate on all available data points so far, then the predicted duration for S5 is 781.24 days, and hence would be completed by 12/5/07 at 21:45:35 UTC. But when using the most recent 100 data points, then the predicted duration is 730.03 days, completing S5 by 11/04/07 at 16:43:11 UTC. These figures were last updated on 11/29/06, and you can see plots to go with them here: Forecast for the end of S5
I found quite a few interesting plots. For the first one, I've included the code that sifts through the data and generates the chart. Does anyone know if the code is MATLAB?
It's definitely not Matlab. I am not completely certain, but I think it's Mathematica.
Sorry for my prolonged absence, it wasn't intended.
I've been travelling, doing exams, had hardware failures at home and work, and more recently been helping out a mate with his hay. This is not to forget the bloody drought and damn bushfires!
( .... I see I'm not alone with hardware gremlins! :-) )
( .... it's Mathematica, by the way )
( .... on the upside, I've got a P4 1.8G dual core now! )
Thanks to Chipper for holding the fort, as usual he has found some treats!
Now I'll see if I can get back up to speed on the logs..... :-)
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
Corey Gray & Justin Garofoli ( Hanford ), good men that they are, have started a visual library of LIGO operations here. There are links ( thank you Chipper ) from the home pages for Corey and Justin. Both pages have tons of other stuff and well worth a look!
Those white tanks in these shots are described as "LN tanks" in the Hanford log ( 04/12/2006 ). I'd take that as Liquid Nitrogen which would be used for various cooling duties: boils at -195.8°C at standard atmosphere pressure, stays liquid under pressure ie. in a strong tank, presumably with some insulation and a relief valve.
For any of you absolute tech-heads there's an interesting discussion ( Hanford 08/12/2006 "New PMC locking procedure" ) about piezo-electrics. Basically some crystals alter their electrical properties depending on the pressure they are under. So you can vary a voltage and change it's shape, or conversely if it's shape is changed then the voltage across it will vary. Transduction. So one of these in the laser mode cleaner is acting as a barometer, being outside the vacuum space!! As the weather systems travel across the Hanford site this alters that element of the mode cleaner and if not properly set up could cause a drift in the optics. A similiar effect seen at Livingstone ( 30/11/2006 ) - "The PMC voltage is getting low, probably associated with the cold front passing by. I'm relocking it."
As Chipper says, science within the science.... :-)
An unusual visitor turned up at Livingstone on 30/11/2006 : "It was VERY windy. There was a young woman wandering around the site with some flashy red slippers and a small dog, asking for directions to Kansas. We told her to turn right at the x-end and keep walking." :-)
I always look for Dan Hoak's ( Livingstone ) witty bits - ' "What we cannot speak about we must put in an ops log," according to Ludwig Wittgenstein.'
H2 continues to be fickle. Rumbling Mt St Helens, Mexican Gulf weather, and a variety of more distant earthquakes all continue to niggle the IFO's.
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
At Hanford they have rigged up the following ( excerpt from 10/12/2006 ):
Quote:
Robert had set up a view with DataViewer strip chart of H1:SUS-ETMY_COIL_LL to look at connection between
microseism and upconversion (10-140 Hz)
with microseism generated off of Iceland/Greenland, most microseismic effect
is on the Y arm
It is easy to see that envelope of coil current, when high, corresponds to times of high upconversion
For a while, he and I try to see if the 8 sec periodicity of microseism-driven coil current
shows up (with proper phase) in upconversion. Doing by eye and voice is inconclusive.
Vagesh launches 4k_Rock_ifo, a program that plays AS_Q through speakers
We unplug the tweeter, turn up bass, to emphasize low frequencies
00:21 heard H2 lose lock while listening to low-frequency H1:AS_Q on woofer
H2 lost lock due to 1-3 and 3-10 Hz seismic "bump" (i.e., local)
H1:AS_Q woofer signal sounds like snoring, from the pace of the amplitude modulation
or a low-pitched modulated rumbling
01:31 really loud upconversion event, huge in H1 but showed up some in H2 also
01:38 another brief loud bump
02:07 looking at H1:DARM_CTRL display on standard dataviewer screen
Rumbles are loud and strong when time derivative of DARM_CTRL appears high
(hard to be sure this is the correct relationship,
since we are using two different digital programs,
Data Viewer and 4k_Rock_ifo, with unknown differences in latency)
** Does anyone know whether latencies are short compared to 1 sec or not? **
Can this be "micro-wiggles"/scattering/speckle?
02:41 three very loud bumps within about 10 sec
03:21 several loud bumps, didn't see anything obvious on Data Viewer
(ETMY_COIL_LL saturated at ~35000 counts around the same time, but not exactly the same time)
I reckon they should stream it!! This is visceral physics... :-)
Even better - whack on a VR helmet, headphones and a data glove or two! [ + a prochlorperazine/metoclopramide/scopolamine patch ]
This entry was made by Peter Saulson, who has written the excellent book "Fundamentals Of Interferometric Gravitational Wave Detectors".
[ published by World Scientific Publishing, available at say Amazon for the terrific price of USD $91.00 ]
I am ploughing through my copy, and boy does it help give an overall perspective, and some sensation of the detail of the logs too. It has made clear to me why LIGO needs to be the collaborative effort that it is.
I'll stop toadying now!!
Perhaps the gravity music, the Earth moving, brought out this: "A mouse made a few rounds of the control room during the shift."
Livingstone had low duty cycle ( 32.3 % ) due to microsiesm "STS 0.1-0.35Hz peak value was ~10um/s"
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
Both IFO's are struggling with lock probably due to this storm:
Livingstone achieved less than 10% science mode, despite some valiant efforts to the contrary:
Quote:
1351 The AS_TRIGGER tripped and took the IFO out of Science Mode.
1352 The trigger has been reset, and the IFO is back in Science Mode.
1352 The trigger tripped again.
1353 Lost lock. The microseismic noise is decreasing, but is still high,
averaging around 2 to 3 microns/second.
1415 The IFO is locked and in Science Mode. This is segment S5-03939.
The laser power is 5.6 Watts.
1416 The AS_TRIGGER tripped and took the IFO out of Science Mode.
1418 The trigger has been reset, and the IFO is back in Science Mode.
This is segment S5-03940. The Trigger Threshold has been increased
to 15.
1602 The AS_TRIGGER tripped and took the IFO out of Science Mode.
1604 The trigger has been reset and the IFO is back in Science Mode.
This is segment S5-03941. The Trigger Threshold has been increased
to 20.
1604 The AS_TRIGGER tripped and took the IFO out of Science Mode.
1612 The trigger has been reset, and the IFO is back in Science Mode.
This is segment S5-03942. The Trigger Threshold has been increased
to 25.
1612 The AS_TRIGGER tripped and took the IFO out of Science Mode.
1614 The trigger has been reset, and the IFO is back in Science Mode.
This is segment S5-03943.
1615 The AS_TRIGGER tripped and took the IFO out of Science Mode.
1632 The trigger has been reset and the IFO is back in Science Mode.
This is segment S5-03944. The Trigger Threshold has been increased
to 40.
1635 The AS_TRIGGER tripped and took the IFO out of Science Mode.
1638 The trigger has been reset, and the IFO is back in Science Mode.
This is segment S5-03945. The Trigger Threshold has been increased
to 60.
1640 The AS_TRIGGER tripped and took the IFO out of Science Mode.
1642 The trigger has been reset, and the IFO is back in Science Mode.
This is segment S5-03946. The Trigger Threshold has been increased
to 100.
1748 Lost lock, probably due to a seismic spike at the X end.
What's happening here is that excessive swingng of our test masses/mirrors has increased an output signal of the IFO.
AS means the Anti-Symmetric port of the IFO, which is hopefully where the light level is kept to a minimum. Otherwise known as the dark port.
[ .... give in to the hatred and come over to the dark port, Luke ..... :-) ]
An exceptionally complex feedback system attempts to keep the Fabry-Perot cavity section in resonance, and if AS exceeds a preset level ( AS_TRIGGER ) then I think said feedback is no longer attempted. ( I don't know if the mirrors swing free at this point or whether they are even clamped ). One hopes the energy of such motions disappates, and less driving by disturbances occurs, so that when the trigger is reset Science Mode can be re-acquired.
As you can read, the threshold for opting out of resonance was increased about ten fold over a four hour period. Generally only a couple of minutes went by before SM came by again.
Robert Schofield at Hanford has an interesting discussion of the effect of ocean waves striking the shore, with a good link if you like animated maps:
Quote:
Predicting the microseismic peak?
A couple of people have asked me for the web site that I was basing predictions of the microseismic peak
on so I thought that I would put it in the elog:
The model predicts global wave heights for a week in advance (Hs files). It also predicts the wave period
(Tp files). So, if you want to know the source of your microseism, find the period of the troubling peak in
the seismometer spectrum, double it, and look for a region with waves that have that period at
present. Be sure to check both the Atlantic map and the Pacific map (lower on the web page). If you want
to predict how long the microseism problem will last, look at the predicted wave heights in the future for
the storm producing the microseismic peak. Bigger waves put more energy into the ground. Of course this
doesnt take into account the need for standing wave fields in order to couple to the ground, but, at least
for the 3 main sites, off Greenland, off Alaska and off our Pacific coast, big waves always seem to couple
to the ground pretty well.
Anyway, based on this web site, I predict that both LHO and LLO will have a low-microseismic period next
weekend. But there will be big waves off of the pacific coast mid to late this week. We'll see......
Now you can see why some are looking forward to a space based array .... mind you it's also then harder to repair/fix/upgrade!!
There's been a fair amount of
)
There's been a fair amount of triple-coincidence time during the past day. But the weather's been on the blustery side during the few days prior. At Hanford, when tumbleweeds aren't being baled during the day, they can be heard at night being blown by the winds against parts of the buildings. Of course, when the wind is that bad, it's difficult to maintain lock, tumbleweeds or not...
There was also this entry (from a couple days ago) regarding H2's ETMy:
I'll hazard a guess that it's microscopic contaminant stuck between two mating surfaces in the suspension somewhere... but there are many things to consider:
Also, a tremendous amount of work continues to be done with the Control and Data Systems (CDS) at Hanford, to replace one of the primary servers. So far, using the newer hardware, there have been quite a few comments on how much faster systems can be rebooted. Even the automated Figures of Merit take less time to produce:
Hopefully this also results in faster recovery times after a lock loss... :)
I found quite a few
)
I found quite a few interesting plots. For the first one, I've included the code that sifts through the data and generates the chart. Does anyone know if the code is MATLAB? (I'll try and find out)
Averaging the duty cycle over the course of a week, and then plotting the averages for the dates shown (from the code you can pick out that it is for Hanford's H1).
The next two just show the plots that were generated (omitted the code for them), which show the progress in the range pretty well. The first one is for H2's inspiral range, and the second one has L1, H1, and H2 all plotted together.
Meanwhile, work continues with the CDS at Hanford (porting files, etc. to the new server), and H2's mirror in the Y-arm (H2 ETMy) continues to give problems from time to time, and it's not easy for me to understand much of what's being looked at, but a bit of ringing is still the remedy, and the interferometer goes through the stages to regain lock without too much delay.
And here's the latest plot of the range. Never seen Hanford's H1 and H2 looking better! :)
Here's the latest plot of the
)
Here's the latest plot of the NS/NS Inspiral Ranges. I added notes to show the corresponding elog entries, to show what the trouble was.
Note 1:
Note 2:
And the summary for the Scimon Owl shift looked like this:
I didn't check to see what happened at Livingston, guessing without looking that it was trains (seismic). The time in Science Mode looks better than 90%!
Things are going well at
)
Things are going well at Hanford. The duty cycle numbers for 12/3 are posted, and H1 was 94.2%, and H2 was 95.2%, and for Livingston, L1 was 82.0%. Lots of seismic activity at Livingston, with some more logging going on. Also more baling activity at Hanford, but there, the interferometers were still able to maintain lock, and so the segments were flagged for 'vehicle activity'.
The data collected during a period of time in Science Mode is referred to as a 'segment', and so the S5 Science Run will continue until all the segments (for L1, H1, and H2) add up to a full year of observation time, in triple coincidence. There are a couple of projections for when S5 will be completed, and both are in the third quarter of '07, IIRC. (I'll try to show the charts for both projections, in a later post.)
The long segments for H1/H2 were finally ended by seismic events. Here's what it looks like to one of the various seismic sensors when a helicopter flies over the observatory:
The spectrogram covers about 4 minutes, and the helicopter's noise is seen mainly in the 15 Hz level (and multiples of that).
Some of you may already have
)
Some of you may already have noticed the link on the Server Status page (whose link is on the E@H main page) called Graphical display of S5 progress. There you will see 4 charts: Registered Users/Hosts in past 24h, Active Hosts, Percent Done, and Floating Point Speed. And if you've checked recently, then you know we're getting close to being 70% done with S5.
I think this pertains, however, to the 'S5 R1' search (for spinning neutron stars) that we're currently crunching. It's not quite the same as the 'S5 Science Run'. Hopefully I'm explaining the difference correctly.
Most of us, conscientious crunchers that we are, have probably seen the Nov. 6 news item on the main page, highlighting 50% completion of the first S5 search. And along with the news was the estimate of Feb. 10, '07, for sending out all remaining workunits.
But if you haven't visited the Cruncher's Corner, then you may be interested to know that RandyC has a thread for submitting your own estimate (Post your S5R1 completion estimates here), and that Randy's (quite sophisticated) estimate is for the end of January, possibly the 26th. I'm happy to say that I checked today's Percent Done chart (showing the last 2 weeks), and came up with this:
So with 2% more crunched for every 3-day's worth of crunching, then to do another 33% will take 49.5 days. Starting at Dec. 7 and counting the days, I too came up with a completion time of Jan. 26 ( ~ midday :)).
So how much longer until a whole year's worth of 3C (triple coincidence) data has been collected?
The pre-run estimates for the duration of S5 was 18 months, or 548 days. Of course, as time progresses, it's possible to get a better idea of how long the run will take. For example, at the half-way point, you could figure that the second half will take just as much time as the first half, basing your estimate on the whole run so far. But this estimate includes downtime for commissioning work that won't have to be repeated during the second half of the run. So a better estimate might be to just consider a more recent interval of time, like the last 100 days.
Turns out, if you base your estimate on all available data points so far, then the predicted duration for S5 is 781.24 days, and hence would be completed by 12/5/07 at 21:45:35 UTC. But when using the most recent 100 data points, then the predicted duration is 730.03 days, completing S5 by 11/04/07 at 16:43:11 UTC. These figures were last updated on 11/29/06, and you can see plots to go with them here: Forecast for the end of S5
Hi Chipper! RE: I
)
Hi Chipper!
It's definitely not Matlab. I am not completely certain, but I think it's Mathematica.
Regards,
Bert
Somnio ergo sum
Sorry for my prolonged
)
Sorry for my prolonged absence, it wasn't intended.
I've been travelling, doing exams, had hardware failures at home and work, and more recently been helping out a mate with his hay. This is not to forget the bloody drought and damn bushfires!
( .... I see I'm not alone with hardware gremlins! :-) )
( .... it's Mathematica, by the way )
( .... on the upside, I've got a P4 1.8G dual core now! )
Thanks to Chipper for holding the fort, as usual he has found some treats!
Now I'll see if I can get back up to speed on the logs..... :-)
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
Corey Gray & Justin Garofoli
)
Corey Gray & Justin Garofoli ( Hanford ), good men that they are, have started a visual library of LIGO operations here. There are links ( thank you Chipper ) from the home pages for Corey and Justin. Both pages have tons of other stuff and well worth a look!
Those white tanks in these shots are described as "LN tanks" in the Hanford log ( 04/12/2006 ). I'd take that as Liquid Nitrogen which would be used for various cooling duties: boils at -195.8°C at standard atmosphere pressure, stays liquid under pressure ie. in a strong tank, presumably with some insulation and a relief valve.
For any of you absolute tech-heads there's an interesting discussion ( Hanford 08/12/2006 "New PMC locking procedure" ) about piezo-electrics. Basically some crystals alter their electrical properties depending on the pressure they are under. So you can vary a voltage and change it's shape, or conversely if it's shape is changed then the voltage across it will vary. Transduction. So one of these in the laser mode cleaner is acting as a barometer, being outside the vacuum space!! As the weather systems travel across the Hanford site this alters that element of the mode cleaner and if not properly set up could cause a drift in the optics. A similiar effect seen at Livingstone ( 30/11/2006 ) - "The PMC voltage is getting low, probably associated with the cold front passing by. I'm relocking it."
As Chipper says, science within the science.... :-)
An unusual visitor turned up at Livingstone on 30/11/2006 : "It was VERY windy. There was a young woman wandering around the site with some flashy red slippers and a small dog, asking for directions to Kansas. We told her to turn right at the x-end and keep walking." :-)
I always look for Dan Hoak's ( Livingstone ) witty bits - ' "What we cannot speak about we must put in an ops log," according to Ludwig Wittgenstein.'
H2 continues to be fickle. Rumbling Mt St Helens, Mexican Gulf weather, and a variety of more distant earthquakes all continue to niggle the IFO's.
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
Who said LIGO doesn't
)
Who said LIGO doesn't rock?!!
At Hanford they have rigged up the following ( excerpt from 10/12/2006 ):
I reckon they should stream it!! This is visceral physics... :-)
Even better - whack on a VR helmet, headphones and a data glove or two! [ + a prochlorperazine/metoclopramide/scopolamine patch ]
This entry was made by Peter Saulson, who has written the excellent book "Fundamentals Of Interferometric Gravitational Wave Detectors".
[ published by World Scientific Publishing, available at say Amazon for the terrific price of USD $91.00 ]
I am ploughing through my copy, and boy does it help give an overall perspective, and some sensation of the detail of the logs too. It has made clear to me why LIGO needs to be the collaborative effort that it is.
I'll stop toadying now!!
Perhaps the gravity music, the Earth moving, brought out this: "A mouse made a few rounds of the control room during the shift."
Livingstone had low duty cycle ( 32.3 % ) due to microsiesm "STS 0.1-0.35Hz peak value was ~10um/s"
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
Both IFO's are struggling
)
Both IFO's are struggling with lock probably due to this storm:
Livingstone achieved less than 10% science mode, despite some valiant efforts to the contrary:
What's happening here is that excessive swingng of our test masses/mirrors has increased an output signal of the IFO.
AS means the Anti-Symmetric port of the IFO, which is hopefully where the light level is kept to a minimum. Otherwise known as the dark port.
[ .... give in to the hatred and come over to the dark port, Luke ..... :-) ]
An exceptionally complex feedback system attempts to keep the Fabry-Perot cavity section in resonance, and if AS exceeds a preset level ( AS_TRIGGER ) then I think said feedback is no longer attempted. ( I don't know if the mirrors swing free at this point or whether they are even clamped ). One hopes the energy of such motions disappates, and less driving by disturbances occurs, so that when the trigger is reset Science Mode can be re-acquired.
As you can read, the threshold for opting out of resonance was increased about ten fold over a four hour period. Generally only a couple of minutes went by before SM came by again.
Robert Schofield at Hanford has an interesting discussion of the effect of ocean waves striking the shore, with a good link if you like animated maps:
Now you can see why some are looking forward to a space based array .... mind you it's also then harder to repair/fix/upgrade!!
This thread will continue here.
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