E. Hall, S. Pandey, S. Dwyer, L. Dartez

We took another look at the new DARM configuration this evening. Evan had a suggestion for a new test: move into NLN_ETMY, pause here before switching back to ETMX in the new configuration (but after all the new filters have been put into place) to take measurements of the new ETMX OLG before activating it. I took us into NLN_ETMY using the guardian then manually ran the instructions in NEW_DARM to set up the ETMX locking filters but paused before switching arms (so we were still locked with ETMY, notably without the MICH feedforward which remained inactive for the rest of our testing).

We measured H1:SUS-ETMX_L3_LOCK_L_IN1 / H1:SUS-ETMX_L3_LOCK_L_IN2 and H1:SUS-ETMY_L3_LOCK_L_IN1 / H1:SUS-ETMY_L3_LOCK_L_IN2 (DTT template paths are below) to isolate the OLG for the new DARM path on the x-arm. Some post-processing of the data is needed to do this. We'll follow up with that.


Finishing the Transition to the new state:
After the measurements above (but before we had a chance to look at the data) we tried slowly bringing up ETMX slowly starting with an L3 LOCK gain of 0.01 (and an ETMY L3 LOCK gain of 1). We immediately lost lock. This indicates that there is something severely wrong with the new DARM configuration as it's currently installed. Most likely, this is not something as subtle as an instability. We'll need to walk through these steps more carefully next time.


DTT Templates:
H1:SUS-ETMX_L3_LOCK_L_IN1 / H1:SUS-ETMX_L3_LOCK_L_IN2: /ligo/home/louis.dartez/projects/20240311_new_darm_investigations/SUSETMX_L3_SS_new_darm.xml
H1:SUS-ETMY_L3_LOCK_L_IN1 / H1:SUS-ETMY_L3_LOCK_L_IN2: /ligo/home/louis.dartez/projects/20240311_new_darm_investigations/SUSETMY_L3_SS_new_darm.xml

Relevant logs:
- Old Matlab model revived (LHO:75584)
- Unsuccessful attempts in early 2024 (LHO:75308)
- Times spent successfully in new DARM state (LHO:75631)
- Calibration attempts in new DARM state (LHO:74977)

Naoki, Nutsinee

Today SQZ_MANAGER failed to go inject freq independent squeezing multiple times. There were a couple of issues:

1) FC ASC engages when the Q error signal went below a certain threshold but didn't actually grab lock. We haven't fixed this one so the problem can repeat. Hopefully we will fix this tomorrow.

2) SQA_MANAGER waited 2 minutes for SQZ_FC to go to the correct state. See attached screenshot of SQZ_MANAGER, line 532. Without checking to see what's actually going on if the SQZ_FC guardian didn't reach its nominal state SQZ_MANAGER will just close the beam diverter and go to SQZ_READY_IFO. SQZ_FC worked fine most of the time but it just happened to take a little more than 2 minutes to reach its nominal state. The logic needs to be fixed with a loop that checks for FC_LOCK guardian state. For now we bypass the issue by increasing the wait time to 3 minutes.

Quiet time between

PDT: 2024-03-11 19:11:02.319941 PDT
UTC: 2024-03-12 02:11:02.319941 UTC
GPS: 1394244680.319941

and

PDT: 2024-03-11 19:21:42.276196 PDT
UTC: 2024-03-12 02:21:42.276196 UTC
GPS: 1394245320.276196
 

Used this time to run BruCo: https://ldas-jobs.ligo-wa.caltech.edu/~gabriele.vajente/bruco_1394244680_STRAIN_CLEAN/

Notable coherences:

• DHARD_Y at frequencies below 30 Hz [so we might need to tune A2L]
• to a lower extent, CHARD_Y below 30 Hz

OMC-REFL_A_LF shows low-ish but broadband coherence with DARM above 30 Hz, and this is suggestive of the excess noise we see now w.r.t. O4a. Evan suggests that this could be 45 MHz sidebands amplitude noise, that dominates the OMC reflection, and has a small transmission through the OMC

[Gabriele, Elenna]

We have updated the LSC feedforward, and made fantastic use of Gabriele's new interactive fitting tool. All templates and code found in [userapps] isc/h1/scripts/feedforward

We followed the usual procedure:

• leave SRCL FF on, turn off MICH FF, disengage the input and turn off the FF filter. Measure using the MICHFF_excitation_etmypum template
• repeat a similar process for SRCLFF shaping using the SRCLFF_excitation_etmypum template
• Interatively measure SRCL and MICH FF by leaving on the old filters and injecting using the MICH_excitation and SRCL_excitation templates
• create files to fit by updating the LSC_FF_PrepareData.ipynb, make sure to update the filter zpks for the current filters
• Open Gabriele's fitting tool: run ipython, then these commands: > from interactivefitting import *, > InteractiveFitting()
• Have fun fitting the new filters! (the tutorial for this tool is a bit too detailed for this alog), added: see https://dcc.ligo.org/LIGO-T2400013 for the code and a tutorial on how to use!

Attached are screenshots comparing the DARM/SRCL and DARM/MICH injections with no feedforward, old feedforward, and new feedforward.

The new filters are in FM3 for both. Guardian code is updated. No SDF updated for now.

We think this is a new record, with about factor 500 suppression of MICH :)

[Gabriele, Louis, Evan H., Swadha]

After tuning the LSC feedforward and updating the calibration, we took a DARM quiet time to compare with the O4a sensitivity. With the updated calibration, our range sits around 140 Mpc. The attached screenshot shows a comparison between cal delta L on January 7th to today. Louis walked me through updating the DARM calibration for both traces to make sure it was accurate. I chose Jan 7 because it appeared to be a time of a recent, decent sensitivity in O4a (around 160 Mpc). Our low frequency noise does not appear to have changed significantly. The reduction in range seems to come from the mid-range region where adjustments in the squeezing would probably help.

Template saved in /ligo/home/elenna.capote/DTT/DARM_compare_O4a_b.xml

J. Kissel, L. Dartez

Jeff ran the calibration measurement suite. We processed it according to the instructions here. I then updated the CAL_DELTAL_EXTERNAL calibration using the new report at /ligo/groups/cal/H1/reports/20240311T214031Z.

Matt, Stefan, Jennie W

Overall: We adjusted the OMC input matrix resulting in a factor of ten reduction in the OMC suspension drive.

The procedure is as follows:
1) Start by calculating the sensing matrix which maps OM3 and OMC degrees of freedom to QPD A and B changes for one of the suspension degrees of freedom (pitch or yaw).

2) Calculate the inverse of the sensing matrix, which will give you a possible input matrix, mapping QPDA and B changes to OM3 and OMC changes which we use for feedback. The first row of the matrix maps only to OM3, for example, which we chose to be our primary degree of freedom with the higher bandwidth. The trouble we were finding before is that the inverse of our sensing matrix yields a strong degeneracy between the two degrees of freedom, and so we push our second row in the direction that cancels most of the noise in the error signal, and also reduces the degeneracy between the error signals.

This procedure can then be repeated for the other degree of freedom (pitch or yaw).

The results of the noise reduction can be seen in the time domain as well.

Jennie, Craig

Today we reran the mod-depth up down tests.  We are not fully trusting these numbers because we suspect we are clipping a bit on IM4_TRANS, which might artificially raise all of our estimated PRGs, including those for RF9 and RF45.
Additionally, there was obvious thermalization still happening during the measurement.  Still, the first pass measurement is useful for up to 10% estimates?
We should rerun after we recenter on IM4.

PRGs
9 MHz PRG = 85.6
45 MHz PRG = 36.6
Carrier PRG = 50.4

REFL ratios
9 MHz reflection ratio = 0.309
45 MHz reflection ratio = 0.268
Carrier reflection ratio = 0.055

Table of relative powers
Channels
9 MHz
45 MHz
Carrier
H1:IMC-PWR_IN_OUT16
0.013
0.015
0.972
H1:IMC-IM4_TRANS_NSUM_OUT16
0.013
0.015
0.972
H1:LSC-REFL_A_LF_OUT16
0.065
0.065
0.870
H1:LSC-REFL_B_LF_OUT16
0.063
0.062
0.874
H1:LSC-POP_A_LF_OUT16
0.022
0.011
0.967
H1:ASC-POP_A_NSUM_OUT16
0.021
0.011
0.968
H1:ASC-POP_B_NSUM_OUT16
0.021
0.011
0.968
H1:ASC-AS_C_NSUM_OUT16
0.181
0.521
0.298
H1:ASC-OMC_A_NSUM_OUT16
0.174
0.637
0.189
H1:ASC-OMC_B_NSUM_OUT16
0.180
0.565
0.255
H1:ASC-X_TR_A_NSUM_OUT16
0.008
0.009
0.983
H1:ASC-X_TR_B_NSUM_OUT16
0.008
0.009
0.983
H1:ASC-Y_TR_A_NSUM_OUT16
0.009
0.009
0.983
H1:ASC-Y_TR_B_NSUM_OUT16
0.009
0.009
0.983

EDIT:
We forgot where the POP beam diverter was, so I'm also posting some MEDMs and Guardian code snippets of where and how to reopen that beam diverter.  We have left the POP beam diverter open for now.

Trent, Georgia

We wanted to see how the arm cavity higher order modes changed with squeezing phase, so we plotted Dhruva's data from 19 February 2023 zooming in on the appropirate frequency range. See Dhruva's post for more info about the data. We have two plots, one at higher frequency (10200Hz - 10900Hz) (02 20 modes) and one at lower frequency (5000Hz - 5700Hz) (01, 10 modes).

Note that for the high frequency plot, the y-arm HOM is degenerate with the acoustic modes of the test masses (at ~10.4kHz) and the acoustic modes dominate the signal at this frequency. Therefore, we can't say much about the squeeze angle dependance of the y-arm HOM but the x-arm HOM amplitude does not seem to change with the squeeze angle.

The low frequency plot also does not show a significant change in the HOM amplitude.

The list below gives the GPS times and the corresponding curve colors.

Ref 0: 1360822152 (Red)

Ref 1: 1360822844 (Blue)

Ref 2: 1360823864 (Green)

Ref 3: 1360824833 (Brown)

Ref 4: 1360825873 (Pink)

Ref 10: 1360834978 (Teal)

A nice DARM video is found in this post which shows the arm HOM shifting in frequency as the test masses thermalize.

TITLE: 03/11 Day Shift: 15:00-23:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
INCOMING OPERATOR: None
SHIFT SUMMARY:

15:00UTC Started off the day in a Manual Initial Alignment

For the intial alignment this morning, yarm was found by hand and we had to lower the locking threshold from 0.7 to 0.5 from the ALS YARM PHD medm screen (this was reverted shortly after). To get MICH_BRIGHT to lock I had to move the BS quite a bit in Pitch (>2 microradians), for SRC locking guardian prompted me to center SR2 by hand so I misaligned SRM and adjusted SR2 to center it on AS-AS_C, which I had to move a lot in pitch again (312.2 -> 292.7).

16:30UTC After finishing IA Shiela start to work on the OMC suspension IM alignment / saturation issue

18:20 ISI HAM7 WD tripped, probably from SQZ rack workpause

Struggled to lock DRMI or PRMI even after CHECK_MICH, so we started another IA at 18:50UTC finished at 19:14UTC, Locked DRMI on the first try after this IA

Lockloss at 21:14

Struggled to get good buildups with PRMI, ASC would kill it. I tried to hold at PRMI_LOCK and touch up the BS and PRM by hand looking the slow building ups scope but I was not able to do it effectively/fast enough?

I'm starting another IA at 22:25UTC which is still ongoing as of 00:0UTC and almost finished

  LOG:         

https://ldas-jobs.ligo-wa.caltech.edu/~lockloss/index.cgi?event=1394230513

DCPD saturation then LL

I updated the LSC CAL CS filters with changes to the MICH, PRCL and SRCL actuation. This ensures that we can accurately use the CAL CS channels to understand the LSC noise coupling to DARM, among other benefits.

For PRM and SRM (PRCL and SRCL actuators), the big change was Gabriele's improvement of the offload filters on M1. I rearranged filters in the CAL CS bank to line up with the filters as they are in the mirror locking bank. All the changes occured in the CAL CS control path M1 filters for PRM and SRM- I moved the M2M3lock and sus_um filters to be in FM3 and FM4 respectively, and rearranged the current offloading filters into FMs 1, 2, 6, 7 and 10 to match PRM/SRM offloading structure. As a reminder, PRCL and SRCL actuation occurs from M3 and M1, and there is a gain of 10 in the M2 locking bank but the lock outswitch is OFF.

For MICH, I incorrectly included an M3 locking filter, the 200:1 filter, but Sheila showed me that cancels out something in the ISC input signal bank, so that is not necessary to account for. More importantly, I engaged the sqrt(1/2) filter that accounts for our MICH actuation from the beamsplitter. I also rearranged filters according to their placement in the beamsplitter locking banks.

I also found some output switch off in the SRM CAL CS bank, so I engaged that.

I have SDFed these changes accordingly, see attached screenshots.

I have also attached screenshots of the LSC CAL CS filter banks for the BS, PRM and SRM control. Reminder, these settings are SDFed in SAFE only.

Using alogs 47217 and 45734 as a guide, we measured/calculated a new OMC-DCPD_MATRIX.

Note: this might have to be redone if we update the anti-TIA filtes for the DCPDs, see alog 76228.

We measured the coherent signal amplitude ratio g = PD_A/PD_B above 30Hz (i.e. above the TIA transfer function imbalance), so the matrix is optimized where we actually care.

g=1.0471

We measure the shot noise signal amplitude ratio h = PD_A/PD_B above 400Hz to avoid thermal noise and correlations from the feedback.

h=1.0247

This produced the new output matrix (see secript)

0.99861 1.00139
0.97725 -1.02328

For reference, old matrix:

0.99963 1.00040
0.98198 -1.0184

The new matrix was loaded, but not yet tested - the IFO lost lock due to 43kts wind gust.

Meashured the transfer function between DCPD_A and DCPD_B usinng a broadband injection. There is a slight frequency-dependent disagreement (0.2dB) between the two diodes, right around 25Hz, where the TIA has its complex poles.

It looks like  one or both Anti-TIA filters are slightly mismateched. We won't tweak it today because we don't know with one is the culprit.

Naoki, Vicky, Nutsinee, Matt

We recovered the FDS and got 4.5dB squeezing in IFO with 40 times more CLF power as shown in the attached figure. 

First we restored the ZM1/2/3 and FC1/2 OSEM positions to the values in O4a. Then we found the FC green trans in camera and PD. We aligned FC1 and FC2 and successfully locked the FC green. We went to SQZT8 and centered the green camera and green QPD.

Since we aligned the green QPD, the green QPD offset value is not valid anymore so we removed the beam spot control from FC ASC. We made a flag for the beam spot control in sqzparams and it is set to False now. We also set the ADF servo flag to False since the ADF demod phase might not be correct. After we figure out the optimal green QPD offset and ADF demod phase, we should revert them.

We reduced the FC IR gain from -3.5 to -0.1 and reduced the FC ASC gain from 0.1 to 0.005.

We have not done any PSAMS scan so we will do it next week.