Turned on new MICH FF at 19:45UTC, details and plot in 76956. Updated safe and observe sdf's and ISC_LOCK.

Took SRCL measurements 19:50-20:05UTC.

Gabriele suggested that we rerun Sheila's DARM comparison scripts (76768) to look at time from last night. I used the instructions that Sheila wrote in LHO:76768.

The two times I used are from Gabriele's DARM comparison (listed below), however I just used 600s segments instead of the full 3600s. Also, these plots compare DELTAL_EXTERNAL since we were having nds issues on-site and couldn't fetch GDS-CALIB_STRAIN for a few hours.

O4a GPS 1389058533
ER16 (last night) GPS 1396241443

Attachments:

comparison.png: DARM comparison shared by Gabriele in Mattermost showing that we are almost back down to the O4a level of noise near 100Hz.

compare_darm_spectra.pdf: DARM spectra comparison showing that we're currently pretty much right at late O4a noise levels above 70Hz. Our noise is better from about 18Hz to about 33Hz but our noise floor is higher now 50-70Hz.


compare_darm_range_integrand.pdf: BNS range integrand comparison

compare_cumulative_range.pdf: comparison of the cumulative range at both GPS times. BNS range is matched between the two times below about 18Hz and right at 150Hz (by eye). If I'm interpreting this plot correctly, we're doing better than late O4a now between 20 & 70Hz but slightly worse everywhere else. After talking to Sheila about how to interpret this plot: we are doing better from 20Hz to around 35Hz, then losing range 35Hz to ~70Hz. The improvements are likely from the new DARM offloading but those improvements are being negatively affected by other low frequency noise we're still trying to figure out.

H1 has been out of Observing most of the morning for mostly Squeezer measurements; will return to Observing at 4pm local time (or sooner).

Jennie, Jenne, Sheila

I pushed Jenne's updated cleaning but cannot check if this is better or worse until our problems getting data from nds2 are fixed.

I ran the following:

cd /ligo/gitcommon/NoiseCleaning_O4/Frontend_NonSENS/lho-online-cleaning/Jitter/CoeffFilesToWriteToEPICS/

python3 Jitter_writeEPICS.py

I accepted these DIFFS in OAF model in OBSERVE.snap but we might have to revert them before finish of the commissioning period today if we find out the cleaning is worse.

GPS 139627823 - 16 mins quiet time from last night.

11:12:41 UTC - 11:18:38 UTC quiet time just before cleaning implemented.

11:19:55 UTC new cleaning drops.

11:31:30 UTC end of quiet time.

Thu Apr 04 10:05:02 2024 INFO: Fill completed in 4min 58secs

Gerardo confirmed a good fill via camera. TCmins close to trip temps, it is chilly outside, 40F (4C)

Here are two plots which show the same thing:  the excess low frequency noise noted yesterday 76924 has gotten better but we are still worse at low frequency than we were April 2nd.

The first plot is just a DARM comparison (without cleaning, CAL DELTA L), the second is a plot of the low frequency blrms (  BLRMS 2 is brown is 20-29 Hz, BLRMS 3 is pink is 38-60 Hz).

The third plot shows the same blrms against a ground motion blrms, ground motion doesn't seem to be an explanation.

TITLE: 04/04 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Observing at 152Mpc
OUTGOING OPERATOR: TJ
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 5mph Gusts, 3mph 5min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.35 μm/s
QUICK SUMMARY:

Nice night for H1 on this rainy morn.  H1's currently been locked for almost 7.5 hrs (currently in observing).  There was about 90min of non-observing time for a lockloss, but it looks like H1 came back A-OK (with only a handful of locklosses on its way back up).

Additional Note:  Squeezer team was here early to jump on O4Break-COMMISSIONING shortly after 8am local.

Artem, Camilla, Louis, Oli

Charge measurements analysis code (see InLock_Charge_Measurements wiki) has been translated from Matlab to python with following significant changes:

* Channels H1:SUS-(QUAD)_L3_ESDAMON* are used, which are already calibrated to voltages and therefore don't require additional filters in the code;
* At the same time channel H1:CAL-DELTAL_EXTERNAL_DQ does require calibration, which was not applied previously and now it is ("6-pole, 6-zeros transfer function");
* Transfer function "from m DARM to N force" was pulled from SUS model every time in Matlab code (dampjngfjlters_QUAD_2014-11-10_LLO_model.mat); now it is saved to static text file and loaded from there;
* Uncertainties are added for all coefficients.

Attached to this post is comparison of "old" and "new" analysis results. First thing that jumps out is that numbers are completely different. However this is not necessarily a bad thing. Here's a quote from Jeff:
"I don't really have an intuitive map between the units of the in-lock charge measurement plots vs. well, anything. My instinct is to gravitate to what the calibration group creates - the unit-full numbers for the ESD actuation strength in units of (gamma-alpha) = [ N/V^2 ] (using the notation of Eq. 10 from [1]). That group thinks the latest best value for that number, as of May 10th 2023 was 2.545e-11 N/V^2 ( see the text table at the bottom page 11 of LHO:69696 ). And yet, the ETMX results from these in lock charge measurements report that the value is around (1650 - 4100) = -2.450e+3 [N/V^2]. Do we understand that 14 orders of magnitude discrepancy?"

[1] LIGO-T1700446

With the new scripts the value I get for a measurement around that (May 9 2023) time is gamma-alpha = -2.973887691585383e-11.

Besides this though, there are differences:
* Signs appear to be flipped, currently I don't  understand why. 
* Also some measurements (particularly for ITMX for some reason) have very low coherence in my calculations, I had to lover threshold from 0.25 used in Matlab code in order to process those. In principle, low coherence should be reflected in uncertainty, but I guess up to a point, for too low values this does not hold anymore.

All scripts are available here: https://git.ligo.org/artem.basalaev/inlock_charge_measurements, hopefully accessible by everyone. Note that I can't update scripts in the control room anymore since I don't have remote access. The versions currently available there are buggy early versions of my scripts (but also old Matlab versions are still there). As of now, use gitlab link above instead if you want to try running them, also in the control room.

TITLE: 04/04 Eve Shift: 23:00-07:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Lock Acquisition
INCOMING OPERATOR: TJ
SHIFT SUMMARY: Locked most of the shift, but lost lock after 7 hours. The alignment doesn't look too bad, especially after running an initial alignment.
LOG:

23:00UTC Detector relocking and at OMC_WHITENING
23:01 NOMINAL_LOW_NOISE
23:10 Observing - SDF diff for TCSCS reverted(attachment1)

23:13 Pushed out of Observing by sqz fc losing lock
23:31 Back into Observing

03:36 Earthquake mode activated due to incoming earthquake from Japan
04:16 Seismic to CALM

06:10 Lockloss (76944)
06:46 Started a manual initial alignment due to PRMI only wanting to catch on a diagonal 0 1 mode despite my adjustments to PRM and BS
06:59 Finished initial alignment, relocking

Lockloss 04/04 06:10UTC

Currently Observing and have been Locked for 5.5 hours

Closes FAMIS#26291, last checked 76669

Corner Station Fans (attachment1)

All fans are looking normal and within range.

Outbuilding Fans (attachment2)

MX_FAN1_370_1 has started again(see last week's) with the jumps in noise (or I guess jumps down). These jumps in noise are still well within range. All other fans are looking normal and are within range.

Closes FAMIS 26441

BRS Y looks fine

BRS X was fine for majority of last month, had a hiccup a day or two ago but is now back in its nominal state. alog 76888 shows that it was touched yesterday so that may be it - tagging Seismic.

Naoki, Eric, Camilla

We tried the PSAMS coarse scan from 0 to 200 with 100 step for two PSAMS. So 9 PSAMS setting in total. The nominal PSAMS is 100/100 and we tried 0/0. Unfortunately, when we tried 100/0, the lockloss happened. For each PSAMS setting, we ran the SCAN_ALIGNMENT with asqz-optimized and took 5+5 minutes quiet sqz/asqz data. The attachment shows the result. The 0/0 has much worse anti squeezing and large frequency dependence compared to 100/100. However, we are not sure if the mode matching is really worse or alignment is bad with 0/0 although we ran SCAN_ALIGNMENT. We will try to compensate the ZM alignment change caused by PSAMS change tomorrow.

no sqz (10 min)

PDT: 2024-04-03 11:40:00 PDT
UTC: 2024-04-03 18:40:00 UTC
GPS: 1396204818

asqz with 100/100 (5 min)

PDT: 2024-04-03 12:25:07 PDT
UTC: 2024-04-03 19:25:07 UTC
GPS: 1396207525

sqz with 100/100 (5 min)

PDT: 2024-04-03 12:34:10 PDT
UTC: 2024-04-03 19:34:10 UTC
GPS: 1396208068

asqz with 0/0 (5 min)

PDT: 2024-04-03 13:13:17 PDT
UTC: 2024-04-03 20:13:17 UTC
GPS: 1396210415

sqz with 0/0 (5 min)

PDT: 2024-04-03 13:33:50 PDT
UTC: 2024-04-03 20:33:50 UTC
GPS: 1396211648

Ran the data for the In Lock SUS Charge Measurements that ran this morning. Like Camilla said (76895), the excitations only worked for the ETMs, so although I am attaching all four plots, note that the last plot point for the ITMs is March 13th (ETMX, ETMY, ITMX, ITMY).

J. Oberling, R. Crouch, T. Guidry

Since the last update we have tried a couple of things.

First, we attempted to align the FARO to the BSC2 chamber directly.  We placed the FARO in the biergarten and shot the chamber-side door flanges.  We used the SMR to sweep along the outside of the +X side of the +Y door flange, and the +Y side of the +X door flange.  The shots were done as circles, which we could then place a point at circle center to represent the center of the door flange.  The hope was we could then project lines to represent the X and Y axes, and use that to align to our origin using a Plane/Axis/Center Point alignment routine.  I won't go into much detail, as it didn't work.  We used this alignment to look at PSI-6 in the biergarten, and it was off by over 1cm; we then looked at LV-35 and that one was off by even more.  I don't remember exact values as we didn't save screenshots since the devaiations were so large.  Let's say that if these deviations were true, I'm not sure how we would have a functioning, aligned, and alignable IFO.  So, that experiment didn't work, moving on.

Today we decided to try using height marks 901, 902, and 903 to set our Z axis alignment.  We looked at marks 901 and 902 with an autolevel and compared them to our BSC2 door flange scribe average we had used the water tube level to measure (alog 75974) and the listed local Z coordinate from T1100187.  Results:

• 901:
  • T1100187: -54.7 mm
  • Water level: -55.2 mm
  • Difference: -0.5 mm
• 902:
  • T1100187: -62.7 mm
  • Water level: -62.9 mm
  • Difference: -0.2 mm

Not bad, and definitely not the almost +4mm the FARO was measuring for these marks w.r.t. BTVE-1.  If we use our differential height survey from alog 75771 we can also calculate the local Z axis coordinate for mark 903, which comes out to -79.8 mm (versus the listed coordinate of -79.9 mm from T1100187), and again not the almost +4mm reported by the FARO when measured w.r.t. BTVE-1.  It's looking more and more like there was some error in the Z axis position of BTVE-1; whether that error was in setting the monument itself or in setting BSC2 w.r.t. to BTVE-1 we cannot say.  The fact of the matter is that when compared to BSC2 as it sits right now based on our water level survey, the height marks we've looked at are pretty close to their listed coordinates and BTVE-1 is definitely not.  We also happened to do the same differential height survey with BTVE-1 (also in alog 75771), so we can calculate a local Z coordinate for BTVE-1 based on our height mark 903 (which we have now tied to BSC2 Z=0).  Doing so gives a local Z axis coordinate for BTVE-1 of -1060.3 mm (Z₉₀₃ - deltaZ_BTVE-1/903 = -79.8 - 980.5); this becomes -1060.9 mm once we convert to our global coordinate frame (which is decidedly not the -1057.2 mm all of the old documentation lists it as).

Now that we have local Z coordinates for 901, 902, 903, and BTVE-1 that have all been tied to BSC2 Z=0, we used these to set a new alignment.  First, we need global Z coordiantes for our 3 height marks.  We got these by using a autolevel to set a magnetic nest inline with the height mark to be measured.  The FARO was put into an intial alignment using BTVE-1, PSI-1, PSI-2, and PSI-6 (X and Y are generally pretty good, Z is off but we don't care yet), and we then placed the SMR on the nest we placed inline with 903.  From this we got an X/Y coordinate for the nest that we could use to calculate our global Z.  Once this was done we added 903 into the alignment routine, and removed PSI-1, PSI-2, and PSI-6 for the Z axis fit.  we then moved the FARO to a position where we could see 901 and 902 and shot them in the same way we did 903.  Ultimately, we ended up with an alignment using BTVE-1, PSI-1, PSI-2, and PSI-6 for X and Y and 901, 902. and 903 for Z.  The coordinates used for the height marks were (format is [X,Y,Z]):

• 901: [1075.8, 12366.1, -55.7] mm
• 902: [1078.9, 24938.8, -63.3] mm
• 903: [1082.6, 37455.6, -80.0] mm

Tyler has a screenshot of the results of the alignment that he'll post as a comment to this alog.  One thing to note here: We were not using the new BTVE-1 Z as part of the Z axis alignment, only marks 901, 902, and 903.  It's somewhat comforting to see the Z axis coordinate of BTVE-1 as calculated by the alignment routine as close as it is to what we think the nominal shoud be based on our water level survey of BSC2.  The PSI monuments are all out, especially PSI-2.  I'm not sure how concerning this is, as we don't trust the Z coordinates of the PSI monuments anyway.

One caveat here, all of the height marks we used are almost in a line down the Y axis, with very little deviation in the X axis coordinate (a deltaX of less than 7mm across ~25m of the Y axis).  I'm not entirely confident we're picking up the global tilt of our X axis with this configuration.  We want to take this alignment and test and refine it; first thing we can do is get some more height marks spread along the X axis to hopefully catch the global X axis tilt.  More to come!

Kevin and Vicky helped to get the old range comparison plots from the noise budget working with some updates to the noise budget.  The first plots show the spectra, range integand and the cumulative range for a comparison between O4a and a few days ago.  This is range calculated by gwinc, the sensmon range is shown in the legend. 

Looking at the darm spectra and the range difference plot here, you can see that we gained about 15Mpc of range from low frequency improvements including DARM offloading.  Above 40Hz our recent sensitivity has been worse, we lose 15 Mpc from the decreased sensitivity beween 40-100 Hz and another 7 or so from the worse sensitivity above 100Hz. 

Last night we had slightly better BNS range after the squeezer alignment work 76757 , the same comparison of O4a to last night shows that we are loosing less range from 65-300 Hz, as shown in this plot as well. 

I've also made this comparison for no squeezing, using times from 76537 and 76540: here.  This shows that without squeezing the mid frequency sensitivity (30-70Hz) has gotten worse which is costing roughly 10Mpc of range.

Elenna's sensitivity comparison from a few weeks ago contains helpful links to recent changes: 76449

All three times we tried to move the input beam (76534, 76607 and yesterday), we caused a lock loss when moving in the yaw direction.

Approximate lock loss times: 1394946678 1395096385 1395535842

All lock losses appear to show the same behavior:

• IMC transmission drops in a "small" step before anything else happens. IMC transmission then stays constant for 120 ms, then drops quickly
• This corresponds to a "glitch" in REFL_SERVO_ERR, the start of a "ramp" in REFL_SERVO_OUT, and a discrete step in LSC error signals (PRCL, MICH, SRCL)
• Shortly after that, all signals go crazy, but there is no evident oscillation

Those lock losses are indeed very fast, and this seems to point to a CARM problem.