Ibrahim, Oli, Jeff, Betsy, Joe, Others

Summary:

• The BBSS Pitch Degree of Freedom is slowly drifting downwards in counts. This indicates a pitch instability in M1. We think it's the M1 Blade heights.
• The drift can be seen in Pitch only, showing up on the F1 OSEM only, the only OSEM above the center of mass on the face.
• We know it is not the electronics nor the OSEM itself. We took F1 out for a few days and there was no drift at its OLV.
• We know it's not the diurnal temp-sag which is day-to-day (as some prev. alogs believed)
• This is not a new issue and has happened consistently since the last RAL visit with one configuration exception (see below).

Relevant Alogs:

alog 79079: Recent Post-TF Diagnostic Check-up - one of the early discoveries of the drift and pitch instability.

alog 79181: Recent M1 TF Comparisons. More recent TFs have been taken (found at: /ligo/svncommon/SusSVN/sus/trunk/BBSS/X1/BS/SAGM1/Data on the X1 network). We are waiting on updated confirmation of model parameters in order to know what we should correctly be comparing our measurements to. We just confirmed d4 a few days ago following the bottom wire loop change and now seek to confirm d1 and what that means with respect to our referential calibration block.

alog 79042: First investigation into the BOSEM drift - still operating erroneously under the tmperature assumption.

alog 79032: First discovery of drift issue, originally erroneously thought to be part of the diurnal temperature driven suspension sag (where I though that blades sagging more than others contributed to the drift in pitch).

Hypothesis:

We think that this issue is related to the height of the blades for these reasons:

1. The issue was fixed when we lowered all blades from the calibration block's "nominal" or zero by -1.5mm with all 4 blades roughly close to this number (avg -1.5mm)
2. The issue came back when we attempted to fix the S-shaped M1 blade tip by correcting the extra swivel it needed to have in order to stay at the same height. (Joe recommendation to Betsy)
3. Oli and Jeff have a d1 investigation in alog 76071 overlays different P to P model TFs when the blade heights are above/below their physical D (called FD in the attached plots).
   1. Interestingly, there is a new mode at roughly 1.9Hz when d is above the model's physical D by +-4mm. This mode is confirmed to not be cross coupling. Our recent TFs don't have them but TFs with the drift earlier do - I think this is a red herring.
   2. More clearly, the attached file shows overlays from different d1 sizes (Pitch).
4. While the F1 blades are at an avg height of -1.5mm below nominal calibration block height, the spread between the individual blades is higher than before, with the problematic "soft/S" blade measuring at only -1mm. There's another blade at -1.8mm. This is the only difference between our current drifty -1.5mm avg and the non-drifty -1.5mm avg is the spread of each indiv. blade height. At this point, I'm interested in seeing how the spread of individual blades affects the drift effect in addition to just an everage d1 drop - could it be a combo of these effects? We can investigate the latter by playing with the model and the former by emperically measuring the drift itself.

Our Units:

We need to know how the calibration block converts to model parameters in d1 and whether that's effective or physical d1 in the model. Then we can stop using referential units.

To further investigate, we have questions:

1. What is the "sensor calibration block" calibrated to? Physical D (Center of Mass to blade tip) or Effective D?  What are these values? We just want to find a model way to test parameters rather than the cal block or the shim methods to our model since right now we're going off potentially old information.
2. Could differences between the 4 individual blades be causing a drift this stark? (i.e. it's not a net d1 height issue but a blade to blade height issue or a combo). I'm thinking this may be the case since we have two equal net heights (-1.5mm avg) with the only difference being the spread of the indiv. heights.

Attachments:

Naoki, Vicky -

Had an SHG relocking issue just now, when the squeezer briefly dropped lock at 2024 Sept 5 22:46:16 UTC, for the PMC to relock (its PZT bottomed out, routine issue).

SHG guardian went into a weird locking loophole which we had not seen before, summarized in screenshot. The SHG IR trans PD locking beatnote strength goes high when the SHG is unlocked, aka see the H1:SQZ-SHG_TRANS_RF24_DEMOD_RFMON signal. Its threshold is nominally at H1:SQZ-SHG_TRANS_RF24_DEMOD_RFMAX = 0. But the SHG_GRD has a hardfault state that brings guardian down if this threshold is exceeded, so GRD would try to LOCK, then see this error message for RF power level overload, then it would go DOWN, and its stuck.  See SHG guardian logs.

#FIXME: resolve this issue. Ideas- either remove this race condition from SHG guardian (send to IDLE if in fault?), changing RF threshold, etc.

To fix it this time, I manually changed the threshold H1:SQZ-SHG_TRANS_RF24_DEMOD_RFMAX = 5 (threshold was at 0, the demod beatnote was at 3.3), then brought SHG_GRD to LOCKED (worked fine), then reset threshold back H1:SQZ-SHG_TRANS_RF24_DEMOD_RFMAX = 0.

#TODO: decide whether this is a problem that needs fixing or just a weird one-off issue. Trending back, the RF demod power basically always goes high when unlocked, which often triggers this race condition. I'm not sure like 1) why was a problem this time, or 2)  why it is not a problem every time.

We made some improvements today in the sensitivity, going from about 151 Mpc on GDS CLEAN to about 158 Mpc. However, our best range from April 11th (DARM FOM reference pre-OFI disaster) is around 165 Mpc. I made a comparison of that time and now with today's commissioning improvements to see where we are still missing range. I have attached the four plot results from the darm_integral_compare results (see alog 76935 for directions).

The range integrand plot makes it much easier to see that we are still missing sensitivity around the mid-frequency band. However, the sensitivity difference shows that we lose 5 Mpc of range by 40 Hz as well. Much of this range loss seems to come from a variety of peaks that have appeared since the OFI vent, such as the 20 Hz peak. We lose another ~3 Mpc between 40-200 Hz.

I ran a bruco with GDS CALIB STRAIN CLEAN on high range time after commissioning today: post-commissioning bruco

It looks like many of these new low frequency peaks (like the large 20 Hz peak) are well witnessed by things like PSL accelerometers, indicating that they could be from jitter: PEM-CS_ACC_PSL_TABLE1_Y_DQ

Generally, there is a lot of jitter coherence, and given that this is the CLEAN channel, that's probably a sign that the jitter cleaning could be improved, maybe making use of other witness channels if the current witnesses are insufficient to subtract the noise.

A peak at 30 Hz has some coherence with MAG sensor channels, here is one: PEM-CS_MAG_LVEA_VERTEX_X_DQ

Right around 35.4 Hz, there is a lot of coherence with various ISI HAM6 sensors and OMC ASC sensors. For example: ISI-HAM6_GS13INF_V1_IN1_DQ

There is also still a large amount of LSC REFL RIN coherence up to 1 kHz: LSC-REFL_RIN_DQ

I think we should test the PRCL offset again, especially because this will help reduce the CHARD Y noise coupling (ASC-CHARD_Y_OUT_DQ) and will also possibly help this HF noise (frequency noise? intensity noise?)

SRCL is better than before, but maybe has more room for improvement between 10-25 Hz: LSC-SRCL_OUT_DQ

DHARD Y coherence is low, but still present, so we should be careful with the WFS offset: ASC-DHARD_Y_OUT_DQ

There is still PRCL coherence: LSC-PRCL_OUT_DQ which is likely coupling through a combination of CHARD Y, SRCL, and LSC REFL RIN. Again a PRCL offset will help. Other strategies are to check POP phasing, POP sensing, etc. Reminder: PRCL feedforward failed, so we need to consider other avenues for noise reduction.

To summarize some strategies to get back to April sensitivity:

• check if squeezing can be further optimized in the bucket
• work on improving jitter subtraction, and consider possible low frequency jitter as well
• investigate and improve PRCL coupling through the ASC and possible offsets causing excess frequency or intensity noise
• continue to hunt other low frequency peaks, especially those that are new from the OFI vent
• fiddle with SRCL FF a bit more

Editing because I went back to check the previous PRCL offset work and found this comment: 76818, in short, we can fix the REFL RIN coherence, but it has no effect on the sensitivity. However, it can improve CHARD Y noise, although at the time I don't think we were limited by CHARD Y enough to see the low frequency benefit.

TITLE: 09/05 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Observing at 147Mpc
INCOMING OPERATOR: Ibrahim
SHIFT SUMMARY:

The first half of the shift was devoted to calibration measurements + a little over 3hrs of commissioning time.  H1 then had a lockloss (after 16hrs & almost reaching 160Mpc).  45min into the next lock and have had range data point just under 158Mpc.
LOG:

• 1530 OUT of Observing for:
  • Quick OFFSET change to SRCL1 offset (from -175 to -400 for H1:LSC-SRCL1_OFFSET)
  • 1530-1559Thurs Morning Calibration
  • 1559-1917 Tranisition to Commissioning (until 1900utc)
• 1529-1557 cleaning (karen)
  • 1559 Heading to woodshop (karen)
  • 1729-1748 H2 Elec Room cleaning (kim)
• 1534 PEM measurements of fan audio (sam)
• 1841-1851 EQ mode activated thru M5.5 Iceland EQ that H1 rode through during a 2nd calibration at end of Commissioning Period.
• 1851-1934 MX measurements (janos)
• 1917 Back to OBSERVING
• 2114 LOCKLOSS (after 16hrs, and almost touching 160Mpc)
  • 2141-2201 INITIAL ALIGNMENT, since PRMI did not look good
• 2203-2218 Vac checks near in X-Beam manifold area (gerardo, jordan)
  • Gerardo noted that the LVEA lights had been left on in the LVEA....since Tues?  He turned them off remotely (and I went out to confirm the LVEA lights were OFF.  (I also turned OFF the Cleaning Room lights---room that leads to CER.)
• 2242 PCal measurement (francisco)
• 2246 Back to OBSERVING

TITLE: 09/05 Eve Shift: 2330-0500 UTC (1630-2200 PST), all times posted in UTC
STATE of H1: Observing at 147Mpc
OUTGOING OPERATOR: Corey
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 8mph Gusts, 3mph 3min avg
    Primary useism: 0.02 μm/s
    Secondary useism: 0.18 μm/s
QUICK SUMMARY:

IFO is in NLN and OBSERVING as of 22:46 UTC

This is a request from Vlad a long time ago. 

In 73697, Vlad changed the bandpass filter in H1:SUS-ETMX_PI_UPCONV_UC3_SIG to steeply cut off the lower sideband (80kHz-300Hz) for 80.3 kHz PI. The new filters are FM1 and 2 (80to81kHz_a and 80to81kHz_b).

Vlad asked me to implement a similar filter for other PIs. I changed the bandpass filter in H1:SUS-ETMY_PI_UPCONV_UC7_SIG for 10.4kHz PI. The new filters are FM4 and 5 (10to11kHz_a and 10to11kHz_b) and old filter is FM1.

The first and second attachments show the new and old bandpass filters. The new filter has -144dB gain for lower sideband (10kHz-430Hz), while the old filter has -40dB gain. The SDF is accepted.

This is the 3rd calibration of the day.  The one taken post-commissioning was not usable/had error.  So received request to run another Calibration.

Measurement NOTES:

• Following instructions (link)
• H1 at NLN for for over 13.25hrs
• Calibration_Monitor medm screenshot is attached
• 2003utc (1409596056 gps) Out of Observing, ISC LOCK to NLN_CAL_MEAS & Observatory Mode to CALIBRATION
• 2003-2008  Broadband measurement run
  • Measurement at:  /ligo/groups/cal/H1/measurements/PCALY2DARM_BB/
•  2008-2031 Simullines measurement
  • gps START time: 1409602151.842028
  • gps END time:1409603535.426098
  • Files writtten to subfolders here:  /ligo/groups/cal/H1/measurements

-Mattia, Sheila

We have written a python script to compute the full matrix of power spectral densities and cross-power spectral densities between a given channel, i.e., DARM, and a set of auxiliary channels. The code can be find at this repository https://git.ligo.org/mattia.emma/cross_psd  which includes a README file describing how to run it.

The main arguments the user has to pass are the start time (in GPS time) and length of the data to retrieve from gwpy, a list of channel names and the starting frequency for the strain plots.

The code creates five different types of plots using the coherence and cross-power spectral density matrix. The final result of the code is a coefficient for each frequency value expressing the algebraic sum of the contributions of all the auxiliary channels to DARM considering the cross-power spectral density terms. It also computes the coherence between the single auxiliary channels and the DARM channel, which are the diagonal terms in the cross-power spectral density matrix.

The five kinds of plots are:

1. The cumulative coherence. The sum of the coherence coefficients of the single auxiliary channels with DARM. As we add more and more channels to the sum, one can notice from the plot that the value of the cumulative coherence goes above one, which is unphysical. This motivates the inclusion of the cross-power spectral density terms to account for the correlation between the auxiliary channels.
2. The cumulative strain contribution. Plot of the DARM strain and the iterative contribution of the selected set of channels to the strain. For example 3_<Channel_name> means that the strain contribution was computed using three auxiliary channels and the added channel compared to the previous plot (2_<Channel_name>) is <Channel_name>. This plot has only two lines.
3. Strain_<number>.   Similar to 2 but with as many lines as in <number> plus the DARM strain to show how increasing the number of included channels saturates the DARM strain.
4. Strain_comparison_<channel_name_1>_<channel_name_2>. Similar to 2 but including  the DARM strain and only two lines. One showing the cumulative contribution to the strain until <channel_name_1> and one adding to this <channel_name_2>.
5. Single_coherence_<channel_name>. A plot of the coherence between the DARM channel and the <channel_name>.

All of these plots can also be created using as a main channel any auxiliary channel instead of DARM, e.g., if one would like to study the correlation between auxiliary channels. Each plot name also includes the start and end GPS time of the data used for them.

Comments are welcome. As a next step we would like to implement interactive plots to allow the user to include/exclude lines from the plots.

Ryan C, Rahul

We have finished the assembly of the 10^th Ham Relay Triple Suspension (HRTS) for O5, which is being assembled and characterized in the stagings clean room lab upstairs. In June we reported the assembly and testing of first five Freestanding HRTS suspension (see LHO alog 78574 and 78711), since then we have built an additional five of them, thus bringing the total count to ten for both the sites (with two more to go, total twelve required between LHO & LLO which includes one spare at each site).

I have attached several pictures (attachment01, attachment02, attachment03, attachment04) below which shows the assembled & locked HRTS stored in the lab on the flow bench in the clean room. Picture01 shows an HRTS with BOSEMs and cables attached, ready to be characterized on test stand (reference pic is shown here).

In this round of HRTS assembly work we have assembled three Freestanding configuration, one Suspended version (shown in the attachment05 below) and one OM0 (attachment06). The Suspended version of HRTS will be attached to the new BBSS (beam splitter) in O5 and Om0 will have bottom mass (optic) actuation using AOSEMS.The top mass will be controlled by 6 BOSEMs which is common for all types of HRTS.

After finishing the assembly work, we balanced all three stages of the suspensions for all six degrees of freedom. This involved lowering and matching  blade tip height and angle on the top stage (2 blade springs) and on the top mass (4 blade springs). In all the cases, optic's lowest edge was lowered to a height of 40.5mm (+-0.5mm) from the bottom of the cage. The PUM and Top Mass height was adjusted based on the scribe lines on the structure.

Given blow are the details (OLV, offsets and Gain) of the six BOSEMs attached the HRTS.

1. Structure no. 07, Configuration: Suspended

Suspended masses: Top Mass = 755gm, Penultimate mass = 802gm, Dummy optic = 300gm

BOSEMs s/n D060108-E. S1900741, S1900749, S1900622, S1900662, S1900637, S1900613.

2. Structure no. 06, Configuration: OM0

Suspended masses: Top Mass = 755gm, Penultimate mass = 802gm, Dummy optic = 301gm

BOSEMs s/n D060108-E. S1900726, S1900723, S1900746, S1900732, S1900742, S1900747

3. Structure no. 04, Configuration: Freestanding

Suspended masses: Top Mass = 758gm, Penultimate mass = 802gm, Dummy optic = 301gm

BOSEMs s/n D060108-E. S1900749, S1900722, S1900724, S1900735, S1900740, S1900744

4. Structure no. 05, Configuration: Freestanding

Suspended masses: Top Mass = 755gm, Penultimate mass = 803gm, Dummy optic = 300gm

BOSEMs s/n D060108-E. S1900730, S1900727, S1900750, S1900738, S1900743, S1900734

5. Structure no. 09, Configuration: Freestanding

Suspended masses: Top Mass = 758gm, Penultimate mass = 800gm, Dummy optic = 300gm

BOSEMs s/n D060108-E. S19007309 S1900725, S1900733, S1900736, S1900803, S1900728

********************************************

Note- The test results of the suspension will be posted below as comments.

Betsy asked me to take a look at doing B&K measurements of the BBSS being assembled in the staging building, so I reminded myself how to use the B&K and took measurements this morning. I'll attach a photo shortly, but I put the accel on the very bottom of the cage, with +X axis aligned with optic longitudinal, Z aligned with is optic vertical, Y is optic transverse. Plots are titled with the accel axis that I hit on the suspension cage, while the legends are labeled with the accel axis response, so "BBSS X Meas" shows the X,Y,Z tf from hitting the cage along the accel X axis. These measurements are a kind of a mid point of the assembly, looks like most of the parts were there, but the bottom stage osems didn't have flags, I didn't see any vibration absorbers.

I think the most concerning thing I see is this 65-ish hz X mode on the first plot. It's the biggest peak and is in a band that could potentially really limit some of the ISI loop gains, if it's not well damped on the table.

A while ago I made a script to plot these measurements, but didn't explicitly say how to run it. Script is quick_plot.py is userapps/sys/h1/scripts/bruel_and_kjaer:

jim.warner@cdsws22:~ 0$ userapps
jim.warner@cdsws22:release 0$ cd sys/h1/scripts/bruel_and_kjaer/
jim.warner@cdsws22:bruel_and_kjaer 148$ ./quick_plot.py -t "BBSS Z Meas" -f "/path/to/data/BBSS_Z_meas.csv"
 

Vicky, Sheila, Naoki

First we tried SRCL offset of -400, which looked good in yesterday's FIS SRCL offset measurement 79903. We took the calibration measurement with SRCL offset of -400 in 79911, but Louis reported in the mattermost that there is a large optical spring in the sensing function. Also, FDS with SRCL offset of -400 is worse than nominal. Vicky will add more plots for this.

Then we decided to change the SRCL offset to -290 and optimized FC detuning. This improved the sensitivity below 100 Hz as shown in the first attachment and improved the range by ~5Mpc. The optimal FC detuning changed from -34 Hz to -28 Hz and this could be because of SRCL offset change and arm power change.

After FC detuning improvement, we took the calibration measurement with SRCL offset of -290 in 79922, but the measurement did not make sense according to Louis so we took an another calibration measurement with SRCL offset of -290 in 79928. 

This was a fast test right at the end of the commissioning period. I tried changing the offset on the AS WFS to see if that changes the low frequency sensitivity of DARM. Short answer: no change in sensitivity without the offset, and no improvement in sensitivity with a higher offset, which yesterday's tests suggested would help (79904).

I forgot to save the reference for the higher offset test (it was exactly the same as the others), but here is a comparison of the current offset (-0.15) and zero offset. The April 11 reference trace is also shown. We are very close to our best low frequency sensitivity in April (pre-OFI problems).

I think that if it helps the SQZ ASC to have no offset, we should turn this offset off. DHARD Y coherence is now very low even with the offset off.

Dropped out of OBSERVING at 1530utc (830amPT), for calibration measurements and then went immediately into Thurs COMMISSIONING time.  Commissioning ended 1917utc (1217pmPT).

Sheila, Dave:

On Sheila's request, I have edited the generate_ifo_range_medm.py script to add a new line for the H1:CDS-SENSMON2_BNS_EFF_RANGE_CLEAN_MPC channel (plus its associated GPS channel).