The range has been ratty for the past three hours. TCSX guardian lost its laser lock point and briefly kicked us out of Observe couple of times. Injection guardian has been set to INJ_KILL since the EM follow up call.
Dan has reconfigured the seismon channel_access_client process on h1hwinj2 to be managed by systemd. This will automatically restart the process if it terminates. The script currently has a race condition bug, wherein during the clean up of the H1O2 directory of data older than 5 hours, the program trips up on the oldest file being deleted (even though it is not read). Until this bug is fixed, a crashed process will now be respawned. In any event, this feature was required no matter what caused the crash.
Sheila, Robert
When LLO was down Sheila and I tried adjusting IMC WFS offsets to see the effect on jitter coupling into DARM. We monitored jitter coupling with 2 injected IOO piezo peaks (290 Hz in pitch, 400 Hz in yaw), the permanent peaks around 300 Hz, and peaks from table shaking. We adjusted 3 DOFs and the best settings we found for them were:
DOF2_P : 200
DOF1_Y: 0
DOF2_Y: -1200
The first plot shows that we reduced the 2 big jitter peaks just above 300 Hz by a little more than a factor of 2. In addition, peaks in the 100-200 Hz region and just below 600 Hz were reduced. On the other hand, there seems to have been an increase in noise in the 40-80 Hz region and the 400-500 Hz region. But the coherence with jitter monitors did not seem to increase in these regions, so we interpret these increases as other noise, such as scattering, that increased by chance due to the alignment differences between the two states.
In addition to the IOO Piezo jitter lines at 290 (pitch) and 400 (yaw), we used a shaker to shake the PSL table (see second figure, bottom plot). As we adjusted the pitch offset, we found no improvement in the broad band coupling of the shaker injection or the permanent peaks above 300 Hz (although we did reduce the pitch piezo peak, which had been made very large relative to yaw in order to show in DARM), by 10. In contrast, the yaw adjustment reduced the peaks that were normally in the spectrum, and the peaks from shaking, by a factor of a little over 2, as well as the piezo yaw line at 400 Hz. We conclude that the interferometer alignment is such that yaw jitter coupling dominates. Also, the piezo injections are a good representation of general table shaking (second figure).
The adjustments didn’t change the inspiral range by much. But we would argue that changing to the settings we found is an improvement because the loudest part of our binary black hole mergers have been in the100-400 Hz region, which was improved, especially in the 300 Hz region.
We would like to try putting in the new settings and adjusting the interferometer alignment to minimize the low frequency noise that increased (and inject at BS HEPI), and we would like to try moving the spot positions to where they ended up, but without the IMC WFS offsets.
The offsets were in and the interferometer undisturbed: Feb. 16 23:09:00 - 23:19:30
Starting CP3 fill. LLCV enabled. LLCV set to manual control. LLCV set to 50% open. Fill completed in 207 seconds. TC A did not register fill. LLCV set back to 15.0% open. Starting CP4 fill. LLCV enabled. LLCV set to manual control. LLCV set to 70% open. Fill completed in 1113 seconds. LLCV set back to 34.0% open.
Raised CP3 LLCV to 16% and CP4 to 36% open.
Posting a couple of quick plots comparing LHO's duty cycle and environment between O1 & O2.
The first plot are histograms of the total wind for O1 & O2. Overall O2 has been gentle compared to O1.
The second plot is the wind duty cycle comparison in both percentile and absolute wind speed. LHO is doing better overall (I think the dip at low wind speeds is partly due to Snowpocalypse 2017), but we will know better after more run time. Spring is coming!
The third plot is the duty cycle for microseism. Still looking pretty good compared to O1, and we made it through the worst of the microseism.
The fourth plot is the wind duty cycle again, but I filtered the O1 data by when we were using 45 mhz blends in the Y direction on the ETMY ISI. We are clearly doing better with respect to wind duty cycle here.
TITLE: 02/17 Owl Shift: 08:00-16:00 UTC (00:00-08:00 PST), all times posted in UTC
STATE of H1: Observing at 69Mpc
INCOMING OPERATOR: Nutsinee
SHIFT SUMMARY: Observing for 14+ hours. No issues.
LOG: None
Observing for ~10 hours.
TITLE: 02/17 Owl Shift: 08:00-16:00 UTC (00:00-08:00 PST), all times posted in UTC
STATE of H1: Observing at 70Mpc
OUTGOING OPERATOR: Patrick
CURRENT ENVIRONMENT:
Wind: 5mph Gusts, 4mph 5min avg
Primary useism: 0.02 μm/s
Secondary useism: 0.41 μm/s
QUICK SUMMARY: Locked at NLN for ~11 hours, observing for past ~6 hours. No issues were handed off from Patrick.
TITLE: 02/17 Eve Shift: 00:00-08:00 UTC (16:00-00:00 PST), all times posted in UTC STATE of H1: Observing at 71Mpc INCOMING OPERATOR: Travis SHIFT SUMMARY: No issues. Remained in observing except to allow Sheila and Robert to run measurement when LLO lost lock. LOG: 00:19 UTC Restarted video2 01:31 - 01:57 UTC Dropped from observing to let Sheila and Robert excite the reaction mass at end X to test for scattering (LLO lost lock)
Still in observing. Commissioners have left. No issues to report.
I ran a copy of the filter files through foton to check for errors. Most had no issues, some had filters defined in the body which were missing in the MODULES block (filters which have been removed from the model). Two models reported errors:
Module DARM_AUDIO section 0: Mismatch betweeen design and coefficients.
Module MC2_M2_DRIVEALIGN_L2P section 3: Mismatch betweeen design and coefficients.
though looking at the corrected filter file and the original does not show a difference in coefficients, Jim is investigating.
The procedure followed to check for these issues is:
cd /ligo/home/david.barker/temp
rm /ligo/home/david.barker/temp/*
cp /opt/rtcds/lho/h1/chans/tmp/*.txt .
ls > /tmp/filterfiles.txt
rm /tmp/fotonresults
for i in $(cat /tmp/filterfiles.txt);do echo $i;foton -c $i >> /tmp/fotonresults; done
Now /tmp/fotonresults has all the issues foton found with the files, and /ligo/home/david.barker/temp contains the "corrected" files.
FAMIS 6885 ITMX ST1 V1 is high (first plot). Spectrum attached.
TITLE: 02/17 Eve Shift: 00:00-08:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Observing at 66Mpc
OUTGOING OPERATOR: Nutsinee
CURRENT ENVIRONMENT:
Wind: 14mph Gusts, 11mph 5min avg
Primary useism: 0.03 μm/s
Secondary useism: 0.49 μm/s
QUICK SUMMARY:
No issues to report.
TITLE: 02/16 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Lock Acquisition
INCOMING OPERATOR: Patrick
SHIFT SUMMARY: Commissioning since morning (bullseye detector + injection).
LOG:
16:29 Joe bringing a truck to chemical bunker
16:34 Karen driving to VPW
17:22 Take intent bit to Commissioning for Robert and Sheila
17:25 Christina opens OSB roll up door
17:44 Sheila+Robert to PSL/LVEA
18:08 Karen to H1 building
18:48 Robert out
19:02 Richard to find Sheila in LVEA
19:10 Sheila+Robert to PSL enclosure
Richard out
20:02 Jason+Betsy out (was at beer garden for an hour or so)
20:30 Robert out
20:35 Sheila to the rack by PSL
Bubba to MY
20:38 Sheila out
21:10 Robert+Sheila taking LLO down time to to injections
21:23 Marc to MY
22:00 Marc back
After working on the Instrument's Host box with Robert on Tuesday, we were getting resigned to pulling this instrument as few tests remained. Was going to swap the host box to test that as the source of the glitching as all our shenanigens got us bupkis--The 46 second glitching, the 0.02ish Hz comb, and the high noise shelf starting at 60Hz, all continued. At least for a few hours.
Now when I look, the time series glitches all but disappeared by about 2330 utc Tuesday, see first attachment. The second attachment is the spectra for the PEM STS and the ISI GND STS. The reference traces are when the glitching was occurring showing the Comb and the Shelf. So the current ADC channels don't have these features but noise floors still don't correspond with ISI GND STS: see below ~60mHz and above 10Hz. This is the signal state a few weeks ago when Krishna looked at which time I found at least one STS mass completely misscentered.
So I propose the following at an opportunity: go measure the mass centering and other diognostics on the Host Box. If there is anything suspicious, replace the Host Box.
The CKeditor component of aLOG (allows WYSIWYG editing of posts) has been updated per FRS#7114 ("Spell checker on alog doesn't appear to work anymore") requested changes.
Added 175ml to crystal chiller.
By Terra's request I have added a new button on the PI main screen that will launch a matlab script to plot the spectrum from 32758-32768Hz. Matlab has to be used because DTT cannot plot frequencies that close to the Nyquist Frequency, and python cannot be used because the Ubunutu machines have an old version of scipy without the welch function. The Debian machines do have an up-to-date version of scipy but operators use both machines. This is NOT a live spectrum, and should be used only to reference at what frequencies the modes are at. Continue to use the StripTools to monitor these PIs in real time.
A few notes:
Example output attached.
To clarify a few things:
The button will call a bash script in an xterm window, which will run a <i>matlab</i> script to display the spectrum. I may have added too many negatives and triple/quadruple negatives and made what was actually being used much too confusing.
I say that it is not live in the sense that it will not continuously refresh or rerun the mesurement like we would normally see when we use DTT. The plot that is displayed takes 90 seconds worth of data, with the gps start time in the title and (now) in the legend.
Awesome, thanks so much TJ!
As TJ said, this is to be used just as you would the PI DTT - to check the BandPass placements for MODE22 and MODE23. These are new modes to the PI medm, but they've broken lock several times in the past without us knowing, so lets keep an eye on them.
Sheila, Jenne, Mark, Richard, Fil, Daniel
We made some progress on installing a prototype of the bullseye detector that Mark has been building in the PSL today.
We will hopefully get the whitened signals into the ADC channels that are already set aside and connected to AS_D in the ASC model tomorow.
Fil and Richard Verified the cabling for this setup. Out of the whitening was a 9pin ISC cable 250 which ran to ADC0 AA chassis in ISC-C1 port 7. It was unplugged from that port and move to ADC5 AA chassis port6 channels 21-24. We still need to verify operation.
This morning I checked the signals from the bullseye detector using a scope in the ISC rack, and saw that all 4 channels had a large oscialltion at 52kHz. I got it off the table and Mark Richard and Fil spent some time in the EE shop, and saw that indeed, attaching a long cable to the detector made it oscillate. Mark changed the 50 Ohm series resistors to 100Ohms, which fixed the problem. While we were there Mark also confirmed with a laser pointer that the middle segment is on pin 2 of the connector.
While the detector was in the shop, Vaishali and I went back to the rack and checked which pins on the input to the whitening chassis showed up as which segments in the digital system. This was not as I had expected, the correct mapping is in this table:
| Head | pins | segment in AS_D |
| center | 2+7 | 3 |
| bottom | 4+9 | 1 |
| HAM1 side | 3+8 | 2 |
| anteroom side | 1+6 | 4 |
Robert and I went back into the PSL and put the detecor back on the table, and removed the lens so that we now have close to equal amounts of power in the center of the bullseye and in the outer ring. By the time we came out, the alignment seems to have shifted a little.
Keita checked on the whitening, and it seems to be working OK. We set dark offsets with the diode unplugged, one stage of whitening on and 24dB of whitening gain. We didn't measure the dark noise, so if anyone gets a chance to go back in it would be good to both measure the dark noise and try to readjust the alignment.
I went in to the PSL yesterday, the 14th, during the maintenance period as described in WP6522. I have done the following activities.
By the way, here is a picture of the bullseye setup.
