TITLE: 01/14 Eve Shift 00:00 – 08:00 (16:00-00:00), all times posted in UTC

STATE of H1: Observing

INCOMING OPERATOR: Corey

SHIFT SUMMARY: Quiet shift, one superevent. Locked and Observing 10 hours.

LOG: 

04:24 (20:24) Superevent S200115j

Jeff. B, Niko, Rahul

ETMY mode 20 (there are 3 modes very close this frequency, ETMY mode 20, mode 12 and mode 18, so one has to be careful while changing the gains) was rung up when we acquired the lock this afternoon (ndscope shows it was on an upward trend in the morning before we broke the lock). So, Jeff. B and I switched off the gain for the next 2 hours, which checked the rise in its amplitude. At 5 pm (local time), I increased the gain to 0.5, which it didn't like. So, I tried negative 0.5. This seems to be working for now (amplitude seems to be decreasing, even though slowly). So, lets keep this going.

Ops Shift Transition: 01/14/2020, Eve Shift 00:00–08:00 (16:00-00:00) - UTC (PT)

State of H1: Locked

Intent Bit: Observing

Weather: 10 mph wind

Primary 0.03 – 0.1Hz: 0.01 um/s

Secondary 0.1 – 0.3Hz: 0.3 um/s

Outgoing Operator: Jeff

Quick Summary: Locked and Observing for 2 hours.

 Rahul, Jeff B.   

After relocking found ETM-Y Mode #20 was at 4.40 and climbing. 
At 22:20 (14:20) set the gain from 0.5 to 1.0. The mode continued to ring up. 
At 22:24 (14:24) with the mode up to 4.44 set the gain to 0.0. The mode is slowly coming down. At 23:20 (15:20) it is down to 4.43 and dropping.   
 

On Friday, I copied in the R0 tracking of the main chain for all the quad models (alog 54416) from LLO. I also copied over the medm screen updates that LLO had started that go with these.  Since we're using it at both sites, and it already seems good, after chatting with Stuart at LLO, I created an ECR (E2000011) which Peter Fritchel has approved.

Today Dave restarted all the quad sus models, so we've got the new paths implemented now.  Before letting the suspensions out of SAFE, I set the gain on the new R0 paths to zero.

I also copied over the suspension Foton files that LLO uses (Anamaria and Stuart had recently checked them in to the userapps/sus/l1/filterfiles/ folder, so I just had to svn-up them).  During maintenance, I turned on both the ETM R0 length tracking paths, with the same configuration that LLO is using (alog 50901). 

I did a quick injection test by pushing on ETMX M0 at 0.15 Hz, and then turning on and off the R0 tracking.  As expected, when the R0 tracking is on the displacement between the chains (as measured by both the L2 and L1 witness channels) is varying much less.  I mostly used the same gain that LLO does (-20), and was able to turn it up by a factor of 2 without trouble.  But turning it up a further factor of 2 caused oscillations.  So, we probably have headroom of a factor of 2 in gain, but not a factor of 4. 

Robert and I are on the board for tomorrow afternoon to do more thorough injection tests, so hopefully we can say more quantitatively after that what kind of improvement we're seeing in the realm of scattered light.

Ed, Jeff K

ITMX and ITMY PUM driver chassis had the monitor boards switched out as per ECR E1900049.

• Both affected suspensions were taken to "SAFE" prior to work
• Both ISI's were put in "DAMPING" - to keep watchdogs from tripping while suspension was undamped.
• All cables were identified and power switches were turned off prior to disconnect and removal
• Both chassis were power tested on the bench before re-install
• Both chassis were sucessfully re-installed without incident

See the image where a bar spans between the HAM7 south Support Tube Ends sitting on Support Tube Extentions and strapped down to the piping platform.  In this state, the Crossbeam Clamp Cap has been removed such that the Crossbeam is just sitting on the Support Tube & jacks.  Most of the hardware has also been removed at the ends of the Crossbeam back at the HEPI housing (out of frame.)  Next step is removing the HEPI Housing and then the Crossbeam.

WP8509 SUS QUAD filters

Jenne, Dave:

New models were installed for h1sus[i,e]tm[x,y]. The order of restarts was:

h1susetmx, h1susetmy, DAQ.

At this point we didn't want to restart the ITMs, but following the install of the new INI files the ITM DAQ data were now shuffled, so we quickly restarted

h1susitmx, h1susitmy

DAQ Restart

Dave:

The DAQ was restarted for both the above SUS changes and to add the new VIOLIN Damping gain EPICS channels to the EDC.

The h1edc was restarted immediately following the DC restart.

Following the new DAQ restart procedure, all outstanding testpoints were cleared at this time.

I restarted all the FOM NDS clients to make them active again. I rebooted nuc21 (psl cams stuck) and nuc23 (asc error ndscope missing).

Free ram disk space on diskless servers

Dave:

I cleared the ram disk of large log files on all three diskless servers (h1build, h1cdsrfm and h1ecatmon0). Prior to the clearing, the disk usages were 46%, 81% and 100% respectively.

Restart ISCT6 SHG Trans digital camera

Dave:

CAM18 had been stuck since early Dec 2019. I restarted its server process on h1digivideo1 via the monit web page and the camera came back to life.

I just restarted the primary and redundant GDS calibration pipelines at GPS time 1263067200.  The latency had already climbed to ~10s since yesterday, likely due to long data dropouts that normally occur during Tuesday maintenance.  Note, this problem should no longer occur when gstlal-calibration-1.2.12 is installed.  See LLO aLOG 50942 for details about the latency bug fixes.

Attached is a dtt file with measured transfer functions from master outs to the newly installed noise monitors for ITMX.  I started but didn't finish making the same set of measurements for ITMY. 

I've updated the schedule file to include a series of DetChar hardware injections for early Wednesday morning (2 AM PST). To avoid having to update the schedule file several more times, I scheduled injections for this same time on 4 consecutive Wednesdays. 

Relevant lines in the schedule file:

1263117618 H1 INJECT_DETCHAR_ACTIVE 1 1.0 detchar/detchar-hwinj-schedule_LIGO-T1900555-PROPOSED_{ifo}_INJECTIONS-1251586818-390.txt
1263722418 H1 INJECT_DETCHAR_ACTIVE 1 1.0 detchar/detchar-hwinj-schedule_LIGO-T1900555-PROPOSED_{ifo}_INJECTIONS-1251586818-390.txt
1264327218 H1 INJECT_DETCHAR_ACTIVE 1 1.0 detchar/detchar-hwinj-schedule_LIGO-T1900555-PROPOSED_{ifo}_INJECTIONS-1251586818-390.txt
1264932018 H1 INJECT_DETCHAR_ACTIVE 1 1.0 detchar/detchar-hwinj-schedule_LIGO-T1900555-PROPOSED_{ifo}_INJECTIONS-1251586818-390.txt

Continued labeling the RF cables in the ISC racks.

Had one cable that was not crimped correctly pull apart. Cable was reterminated. Cable connects the output of the Diplexer (135MHz, ~40dB output) to the LSC I/Q Demodulator, CH4 (LSC_REFLAIR_B_135MHz).

I reset both PSL power watchdogs at 18:44 UTC (10:44 PST).  This completes FAMIS 10745.

Previously, this node would just go into FAULT when the beam was out of its range, but LLO suggested that we have an out of range state to track this better. These nodes are watched by the ISI_ETM{X/Y}_ST1_SC nodes, so I had to make slight modifications to the (userapps)/isi/h1/guardian/sei_config/sensor_correction/brs_states.py file as well.

The SUS Guardians will now read from the susconst.py file (in the h1 area), when going to the MISALIGNED state. This state will turn on the top stage TEST bank offsets to misalign the optic (conception of misalign state in alog17259), but previously it assumed that we have the correct values in that field already. This has bit us a handful of times because we also use the TEST bank for, well, tests. So, we will now have all of our misaligned values saved in the susconst.py to verify that we are  going to a consistent misaligned offset.

I also added a decorator in the idle ALIGNED and MISALIGNED states to check that the TEST Offsets are turned on/off when they should/shouldn't be. If there is a discrepancy, it will make a notification saying so. Our TMSs are an exception to this because we use them to move our spot around the point absorbers (only a local change). Since this is common code, I have already been in contact with LLO.

To sum up my changes:

(userapps)/sus/common/guardian/SUS3.py - Added a reference to a dictionary in susconst.py of misaligned values. Also added a decorator to check the state of the TEST Offset.

(userapps)/sus/h1/guardian/susconst.py - Added dictionary of misalignment offsets for every optic.

(userapps)/sus/h1/guardian/SUS_TMS.py - Changed to not use the decorator to check on the TEST offset.

(userapps)/isc/h1/guardian/ISC_LOCK.py - In the PREP_FOR_LOCKING state, it will reset the TMS TEST offsets to their misalignment values. It now imports susconst to get these values.

Philippe asked me to check the orientation of the vertex magnetometer because the signals were inconsistent with expected field directions. I found that the magnetometer was moved, probably during the October 2019 break, such that the Y direction of the magnetometer was off by about 45 degrees, pointing +Y, -X. I moved the tripod, matching the numbers on the leg to the numbered marks on the floor, and the proper orientation was restored. I did this at about 17:30 UTC.  The affected channels were: H1:PEM-CS_MAG_LVEA_VERTEX_X_DQ and Y and Z. I also checked the EY VEA magnetometer and it was properly oriented.