TITLE: 06/16 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Lock Acquisition
SHIFT SUMMARY: Shift was full of earthquakes, keeping us down for several hours.

Lock #1:

• Dropped Observing from 16:08 to 16:44 for calibration measurements: alog 70520
• Lockloss @ 16:54: alog 70521

Lock #2:

• 4.9 earthquake from France, lost lock while acquiring DRMI
• Relocking went unaided; NLN @ 18:29
• Lockloss @ 18:40 (before Observing) from backscatter test at HAM3: alog 70528

Lock #3:

• Lost lock at TURN_ON_BS_STAGE_2

Lock #4:

• Made it to MOVE_SPOTS before a 6.5 quake from Tonga hit, holding in DOWN until it passes

Lock #5:

• Finally was able to lock after ground motion calmed down
• Relocking is going unaided (needed to CHECK_MICH_FRINGES and PRMI)

Handing off to Corey, H1 currently at MAX_POWER.

LOG:

TITLE: 06/16 Eve Shift: 23:00-07:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Earthquake
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 18mph Gusts, 14mph 5min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.09 μm/s
QUICK SUMMARY:

Arrived to see h1 working on getting back to NOMINAL LOW NOISE due to a big earthquake.

HAM6 OMC camera is still not on nuc30.

It's a little windy, but still better than earlier this week!

I have found a couple cases lately where SEI_ENV was switching out of earthquake mode too quickly. Peakmon was still around 1000nm/s, fairly high, and SEI_ENV would switch back to the nominal state. During the winter, I had to increase this transition threshold from 600 to 1100 because the high microseism was fooling the SEI_ENV guardian. After that, I made an adjustment to the bandpass filter peakmon uses, to better filter out the microseism (here), but didn't reassess the SEI_ENV thresholds. I've set it back to 600 for now, I'll try to figure out if we can live with this threshold wrt the filter I installed in January.

This shouldn't really affect my alog about the earthquake robustness. This threshold affects the transition out of the earthquake mode, there is a separate threshold of 400nm/s for the transition into the eq mode with seismon early notifications, and we are mostly losing lock before even reaching that threshold.

J. Kissel

For quick reference -- the current HAM-ISI blend filters are:
    (X, Y, Z) are all using the "comp_250" blend filter pair, whose blend frequency is 0.18 Hz (or 180 mHz -- hilariously *not* 250 mHz...)
    (RX, RY) are both using the "many_notches" blend pair, whose blend frequency is 0.375 mHz
    (RZ) is using the "fc_yaw_filt_lp." 

The recent ECR E2200473 that improves the performance of switching between blend filters also completely revamped the naming scheme for the filter banks which house the blend filters. That means older, extremely useful code like 
    /ligo/svncommon/SeiSVN/seismic/Common/MatlabTools/plot_current_blends.m
which converts the "as installed" foton representation of the blend filters into their design-intent, dimensionless "complementary" form broke.

Not having that code forced the infrequent viewer / non-designer to cut corners and only look at the displacement sensor blend, and either waive their hands guessing at the blend frequency / frequency-shaping or invent their own metric for "at what frequency do we start to get inertial isolation," e.g. as in LHO:70313.

I've now fixed the above mentioned code for the HAMs -- see attached collection of plots which show the 4 pairs of blend filters which are in play on ALL HAM ISIs at LHO (ISI HAMs 2 - 8).

Along the way, I couldn't help myself but to improve the plots:
    (1) I've added an explicit indication of the the blend frequency,
    (2) tightened up the frequency range and magnitude limits so you can quantitatively read off the magnitude better
    (3) For the phase, we used to plot both the individual filter's phase and the sum, but really only the sum matters. The sum of the HP and LP pair's deviation from 1.0 magnitude and 0.0 phase is a metric for *lack* of complementarity, and thus 
        (a) how much the open loop gain shape is impacted for *this* pair of filters and 
        (b) how much the phase will *change* when we switch to another set of blends with different [lack of] complementarity)

The code still needs further updates to handle BSC-ISI ST1 sensor blends, but I leave that for another day, or the initiative-full reader.

Lockloss @ 18:40 UTC from placing the retro-reflector into a viewport on HAM3 for a backscatter test.

I have added a disk-usage section to the DAQSTAT system. It will report a warning if the disk usage is between 93-95% and an error is greater than 95%.

The disk usage information is updated hourly by cds_report running on cdsmanager.

DAQSTAT was restarted several times yesterday and once today after lock-loss to install the new code.

Fri Jun 16 10:03:52 2023 INFO: Fill completed in 3min 52secs

Jordan confirmed a good fill curbside.

During a lock-loss around 10am localtime we took another device out to the CER switch to check that the port that cam15 is plugged into is live.  It was able to power another device.  So referring the next steps to Filiberto when he is available to go to ham6 and find/replace the camera.

Lockloss @ 16:54 UTC, possibly from CSOFT_P ringup. I don't notice excess ground motion or other immediate cause.

After having been locked for 3+ hours, I started a "sensing function only" calibration measurement following the instructions on the TakingCalibrationMeasurements wiki. This measurement took ~30 minutes after taking H1 out of Observing mode and into NLN_CAL_MEAS.

Attached are a screenshot of the calibration monitor screen prior to the measurement and the generated calibration report: /ligo/groups/cal/H1/reports/20230616T161654Z/H1_calibration_report_20230616T161654Z.pdf

FAMIS 23810

Noise on EX_FAN2_570_2 seems to have slightly increased 3 days ago.

All other fans look good.

TITLE: 06/16 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Observing at 138Mpc
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 13mph Gusts, 9mph 5min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.08 μm/s
QUICK SUMMARY: Taking over from Austin; H1 has been locked and observing for 2 hours.

• CDS, SEI, SUS, dust monitors all okay
• Recent earthquakes, but peakmon is low

TITLE: 06/16 Owl Shift: 07:00-15:00 UTC (00:00-08:00 PST), all times posted in UTC
STATE of H1: Observing at 137Mpc
SHIFT SUMMARY:

- Lock #1:

• Acquired NLN @ 22:30, OBSERVE @ 1:43
• EX saturation @ 7:00/8:55
• 6.0 EQ from Tonga - 8:29, EQ mode activated @ 8:36, @ 8:46 SEI went to CALM according to verbal, but the SEI grd is still in EQ mode, maybe a verbal issue?
• Lockloss @ 4:35

- Waiting in DOWN for ground to settle, before starting an IA - which went smoothly

- Lock #2:

• Locking was fully automated
• Acquired NLN @ 10:54, OBSERVE @ 11:04
• PI 31 ringup @ 10:57 - guardian looks to have damped the mode on its own
• 11:05 - another EQ from Tonga, reading 5.7 according to seismon, hopefully we can ride this one out - EQ mode activated @ 11:27
• Lockloss @ 11:48

- Lock #3:

• H1 struggled a bit with ALS (probably due to residual ground motion), but was able to get to NLN entirely automated @ 12:51, OBSERVE @ 12:59

- Leaving H1 to Ryan S. locked and in observing

LOG:

No log for this shift.

LOCKLOSS @ 9:07, a 5.7 EQ from Tonga. Ground motion was still ringing down when this EQ hit, peakmon was ~1000 counts at the time of the lockloss. There was also what looked to be an ASC ringup for MICH P/Y on nuc29, but I'm attributing this to the earthquake. Peakmon is already starting to drop so I'm going to give it a few minutes and try relocking.

Here is a list of the things we need to change to return as closely as possible to a desirable 60W configuration.

• Ring heaters: 1W ring heater power on both ETMX and ETMY. No change to the current ITM ring heater settings because we are running at the same settings as 60W
• CO2 settings: 1.7W annular on both ITMX and ITMY (we are currently running with these settings). Depending on the amount of time we have to do TCS additional commissioning, it might be worth some time to see if we can improve frequency noise by changing these settings differentially (increase CO2X power). See this alog comment for more details: alog 68842
• OM2 TSAMS: no change
• ASC, control design: I don't believe we need to make any changes to the loops. There has been significant update to the recycling cavity loop design at 76W, but these are power independent. The HARD loops are running with loop compensators designed for about 400 kW of power. I think dropping to 380 kW should not negatively impact the loops, as the hard loops have been stable through thermalization which can change the arm power by 30 kW. The soft loop compensator design is the same. We may even be able to reduce some soft loops gains, as most of the soft loop gains were increased to compensate for the 430 kW operation.
• ASC, HAM1 FF: pending another 5-10 minute on/off test to collect data, we can either use the current filters, or revert to ones that I tuned on April 6 while at 60W. See alog. The filters are named by the date they were installed (so just turn on the ones labeled 0406). We did not need HAM1 FF in the yaw loops at 60W, so we can turn off all those filters first and then check the necessity again.
• LSC, control design: we should consider reverting the PRCL control loop design. Gabriele designed a new loop that could manage the gain loss at 430 kW in addition to adjusting the loop gains. We will likely still need to adjust the PRCL loop gain throughout the lock at 60W. Once we go down in power we should dedicate some commissioning time to determine the gain loss trend and commission accordingly. We should also check that the UGF of MICH and SRCL are where we expect them to be. The PRCL1 filters are FM2, 3, 4, 5, and 6 with a gain of 11.6 and none of the PRCL2 filters were in use.
• LSC, feedforward: we will need to tune the LSC feedforward. We had been running with an outdated LSC feedforward with the plan to remeasure and update it once we settled on a final configuration. All recent updates occurred at 76W. The filter for SRCL was FM2, labelled EDH used with a gain of 1. The filters for MICH were FMs6-9, however FM9 has been overwritten by a new filter. I will use the archive to recover it. The MICH filters used a gain of 0.97.
• CARM gain: the CARM gain was increase by 3 dB to counteract the loss of gain that occurred during thermalization. We should remeasure CARM and determine if we still need that gain increase. alog 69152
• ISS second loop gain: we reduced the ISS second loop gain to -5 dB. We should remeasure the loop and determine a better gain setting for 60W. alog 68789
• SRCL offset: we had been running with an offset of -175. This was also with the previous LSC-POP_RF45 whitening at 21 dB. We could revert the whitening change as well if we think it's better for noise considerations
• DARM offset: 20 mA, not sure if we want to revert this value

If you are looking for a timestamp to determine when the power change occurred, the last full lock at 60W was on April 6 from about 17:00 UTC to April 7 3:30 UTC.

Naoki, Vicky

To try damping the 80 kHz PI recently causing locklosses (LHO:70434), today we installed a new path on PI28, with drive sent to ETMX. To phase-lock the ESD damping drive to the PI mode, we bandpassed the DCPD signal around 80.296 kHz (foton here). This frequency was chosen based on the DCPDs full-spectrum signal around 80.3 kHz, which today we saw the 3 peaks in red here between 80.295-80.301 kHz in full lock. It seems like our problem is the bigger peak around *296, where pink cursors are centered. We are trying to damp on ETMX first, since it seems like PI29 80kHz damping on ETMX could impact this mode.

The new path for PI28 has been updated and damping is guardianized, but this PI28 damping is untested, and there are no verbal alarms for PI28 yet.

Summary with the current status of PI damping:

SDFs were recoinciled after, see screenshots -- mainly, we un-monitored guardian-controlled things like the damping phase and PLL integrator.

Commissioning period,

The Ham1 FF was turned off for testing on 06/14/23 21:32:15. It stayed off for 5 minutes and was turned back on at 21:37:15.