The bad behaviour of SR angular control can be better understood adding a plot of the longitudinal correction. Saturation of actuators and bad balance of actuators can cause angular oscillations when they occur together. This is a typical one year old situation, occuring when the lock acquisition strategy was far from being well tuned. An activity on SR actuators balancing could mitigate the problem of angular stability, but very likely will not fix the problem of lock acquisition.

The planned ISC activity on New TCS working point exploration (Casanueva, Pinto), started at 14:00 UTC and went on without major problems for the whole shift.
At 21:00 UTC Taranto (TCS) started from remote the NI RH switch on for CH centering measurement.
TCS activity still in progress.

No particular operations to report.

Parallel activity:
SGD: Automation (Bawaj, Holland)

The first part of the shift was dedicated to continuing with the morning activity of exploring the input beam alignment. The target was to improve the shape of the sidebands, that showed a vertical alignment mode. An offset was added to the BPC shift DOFs and we checked the behaviour of the dithering error signals and the input mirror centering error signals. A dedicated entry will be posted.

Then we tried to recover the lock of the CITF and progress with the lock acquisition. HOwever, we were not able to lock the CITF in a stable way. The few locks we had were achieved by misaligning the SR significantly (10urad). We were not able to realign the SR and keep a stable lock. Moreover, at some moments the SR local controls were showing a big movement of the SR, of several urads, without any loop being closed on it. The maximum sidebands we reached were 0.06 for the 12MHz and barely 1mW for B4 DC. Figure 1 shows a trend of the CITF lock attempts, with the SR local controls as well.

We contacted Claudia at the end of the shift and we leave the ITF in single bounce for the TCS activity.

We continued the work on the FCEB bench. 

1. Installed another lens in the IR path to collimate the beam

In the previous configuration, the IR PD is installed on the focal plane of the 500mm lens, this beam has a quite large divergence, so we suspected the low power on the IR transmission was due to the large beam size on the PD. So we installed a 100mm lens and extended the IR path much longer, the new version of the scheme and the picture is shown in pic 1 and pic 2. 

We could reach a maximum of 11V IR transmission after this adjustment. Still few volts are missing compared to two weeks ago, which we think could be due to the bad alignment of the IR filter. Further adjustments will be done.

2. Changed the green beam splitter from 90:10 to 10:90 (R:T)

Today the new beam splitter arrived, so we changed it and realign the green PD and the PSD. Now we have about 7.7V of green transmission. The filter in front of the PSD has been changed to ND = 1 from ND =2. 

The triggers for the locking, angular control loops, and the offload has been restored to the old value(5V). 

At the beginning of the shift, the ITF was in SINGLE_BOUNCE_WI: after a brief realignment the infrared cavities were relocked at 6:21 UTC.
The shift was dedicated to the planned ISC activity, locking and ITF characterization, carried out by Mantovani and Boldrini. The work proceeded during all the shift without major issues.
From 7:20 UTC to 7:40 UTC, under the request of the commissioning Crew, I performed the following Scan:

7:20 UTC - CARM_SET_scan(2)
7:30 UTC - SCAN_TEM0(5)
7:34 UTC - SCAN_TEM2(10)

Activity concluded at 14:10 UTC, infrared cavities left locked.

Other Activities carried out during the shift:
SGD: SGD: RTL optimization (Guo, Ding).

DAQ
The SQZ team reported that, since the 26 of June, the ENV signals related to the Filter cavity End Building (FCEB) were no more available. Upon investigation, I found that the related Slow Monitorning Box (ENV-FC-END) was not reachable anymore.
Thanks to the intervention of Cortese the connection was re-established at around 9:35 UTC and, after a restart of the process EnvServerSqz, the signals were available again.

Summary of settings:

Yesterday, at 21.28 UTC, NI and WI DAS powers were changed from the “28th June” settings to “29th June” settings. 

HWS-INJ acquisition was started at 21.18 UTC (before powers change) to control the transition. The acquisition lasted for 4 hours and showed the curvature change in Fig.1.

Actually, the OMC FSR should 834MHz not 864MHz. This makes the spacing of the modes in MHz in the figures is 3.5% wider than it should, due to the wrong conversion of time into MHz.

The spectrograms of the microphone (ENV_EDB_MIC) and accelerometer (ENV_EDB_ACC) make in evidence the acoustic and seismic noise variations in DET-Lab after the last actions to set-up the supply inverter frequency of that air handling unit (AHU), Figure 1 and Figure 2. In both spectrograms, a significant noise reduction is visible after each inverter frequency reduction action. The rms signal of the microphone shows a noise reduction of about 40% in the band (1-16)Hz and of about 70% in the band (16-512)Hz having set up (step by step) the supply inverter frequency from 45 Hz to 35 Hz, Figure 3.
Moreover, the evidence of noise reduction is visible also in the other attached Figures concerning the acoustic and vibration sensors, also in the squeezing area (ENV_EQB1*). General important noise mitigation was achieved in the bandwidth ~(10-100) Hz, Figure 4, 5, 6,7, 8.

More analysis will follow.

I have looked at the downgoing scan (so it is further away from the steps of the fast scan which will make the relation between the temperature of the OMC, and the temperature of the copper plate measured by the thermistance more linear). Starting from 09:32 UTC.

Also roughly calibrated time into MHz by using the OMC FSR spacing of 864MHz. And then indentified a few modes using the camera image (but I did not check all of them). The power shown in the following figures is B1 PD3 multiplied by 1000, which should be the power transmitted by the OMC, as B1 PD3 looks at a 0.1% pick-off.

Figure 1 shows the FSR scan around the order 0, 1 and 2 modes. WIth the carrier TEM00 highlighted in red. The tick markes show the 6MHz and 56MHz, LSB and USB.

Figure 2 shows the same with tick marks for the 56MHz LSB order 0, 1 and 2 modes.

Figure 3 shows the same with tick marks for the 6MHz USB order 0, 1 and 2 modes.

Figure 4 shows the same with tick marks for the 6MHz LSB order 0, 1 and 2 modes.

Figure 5 shows the same with tick marks for the 56MHz USB order 0, 1 and 2 modes.

Figure 6 confirms that on the phase camera the LSB (both for 6MHz and 56MHz) has much higher power  ( by a factor few) than the USB.

The mode mismatch mode (order 2) is much higher in relative power for the USB (both 6MHz and 56MHz) at ~110% of the TEM00, than for the LSB (both 6MHz and 56MHz) that are at about 50% of the TEM00 opwer.

/users/mwas/OMC/OMC_scan_20220630

ITF State: SINGLE_BOUNCE_WI
ITF Mode: Commissioning
Quick Summary: IMC and RFC locked; all Suspensions loops closed, all SBE loops closed.
Activities ongoing since this morning: Infrared cavities relock in progress.

The goal of the activity is to use the arm cavities as diagnostic FP to measure HOMs of the sidebands which resonate into de CITF.

We locked the DRMI with a positive CARM SET (3000 Hz) and made a scan of the CARM SET toward positive values. See pic1.

We used the information on the position of the SB in the arm cavity spectrum (pic 2, 3) and the spacing of HOM to identify SB HOM during the arm scan with locked CITF. (pic 4)

The expected position of the SB HOM agrees well with the measurement. We could double check the identified HOM by looking at the shape of the transmitted beam on B7 Cam.  A video of the B7Cam during the scan can be seen here. (link)

In particular we focused on the HOM2 which can be used to estimate the mismatching of the SBs. As can be seen from the video, for the four SB HOM2 the brightest peak is a mismatching/astigmatic mode (pic 5). (Note that there is a splitting of the resonace frequency of HOM of the same order due to astigmatism of the cavity),

In order to assess the mismatch we cannot measure directly the height of SB fundamental modes, as CITF unlocks when they become resonant in the arms.

We can instead compare the power of the SB mode of order 2 (P_HOM2_CITF) measured on B7/B8 PD during the CITF CARM set scan and the height of the SB TEM00 measured on B7/B8 PD during the simple arm scan (P_TEM00_SC)

We have the following relations:

P_HOM2_CITF = 1/2 * P_in * G_PR * T_ARM * MM

P_TEM00_SC = 1/2 * P_in * T_PR * T_ARM

Where P_in is the input power on the BS, G_PR is the SB recycling gain, T_PR is the transmission of PR, T_arm is the arm cavity transmission and MM is the mismatching.

The mismatch can be then computed as  MM = (P_HOM2_CITF/P_TEM00_SC)*T_PR/G_PR

We will do as slower scan around HOM2 to have a more precise estimation of the peak height.

--------------------------------------------------------------------------------------------------------------

Other notes

1) We noted that locking DRMI with the opposite offset wrt the usual one brings it to a lower level of 12MHz_mag (roughly 0.08 instead of 0.1). The power could be fully recovered by adding an offset on the PR angular TY angular loop.

2) In order to have a more precise estimation of the peak position during the scan we made available a new CARM SET channel sampled at 10kHz  (LSC_CARM_SET_FS).

This afternoon we continued the work of yesterday evening, that is to recover the lock acquisition and study this particular ITF configuration. Here the main changes we did and what we could observe:

• initially we left unchanged the DRMI lock, both on 1f and 3f; we could see that the PR alignment loop could already benefit from an offset on TY and we put it (-0.005); later, we added this possibility in the Automation, and the offset is set just before the finalization of the DRMI lock in 1f;
• the DRMI lock itself was really smooth for more than one half of the shift, while in the last 2 hours it became much more difficult;
• the CARM handoff to the infrared on the MC was working without any issue; initially, we performed the DARM handoff by hand following the same methodology as yesterday: increase the arm transmitted power to 40 mW, then perform the DARM handoff to B1p_56MHz_I, then change the control filter; DARM calibration was set to 2.2e5;
• then we verified the parameters and reached the same step with the automation, which also brought us to 50 mW arm powers, as usual; we immediately saw a drop of the DARM gain by basically 50%; we reverted the last step and moved back to 40 mW, but we saw no change; the DPHI monitor also showed a change; even after retuning the phase (at later steps) did not cure this, and eventually we put the DARM calibration to 5e5 for the rest of the shift (Figures 1 and 2 the trend, Figure 3 the DARM OLTF after the drop);
• given that the DARM handoff happens now at 40 mW arm powers, we decided to move all the other CARM offset steps: now, the ramp after the DARM handoff stops at 75 mW instead of 50; STEP1 is now 100 mW instead of 80 and STEP2 is 200 mW instead of 160;
• we managed to get to STEP3 many times, but we suffered many unlocks, mostly at STEP2 and STEP3; the main causes were bursts of noise at 150 Hz and oscillations at around ~12-13 Hz (Figure 4); we cured the former by tuning the gain of the CARM loop injecting noise on the RFC, while the latter are still unexplained (Figure 5, CARM OLTF before the gain increase); we also had one unlock thanks to a 9 Hz oscillation and one at a much lower frequency, around 1.2 Hz;
• during the shift we consistently had a PR_TY offset, and we generally increased the gain of such loop during the CARM offset reduction; to recap, we have now three places where we put offsets on PR before CARM_NULL_3F: just after closing the control at DRMI 1f, between the CARM handoff on IR and the DARM one; again, at STEP3.

We left the ITF in the Single Bounce (WI) configuration, which we added in ITF_LOCK (in the "all" menu).

Several flaws in the code were removed. The most important were:

- log flooding issues;

- cm_send flooding.

For the first problem, log writing was removed from the states run() and some were substituted with notify()

For the second problem, a boolean flag was introduced whenever a risk of sending multiple cm_send was present.

Automation tests were difficult due to frequent INJ relocks and after some time they were moved to the next automation shift.

The planned ISC activity on locking and ITF characterisation (De Rossi, Capocasa) started at 07:00 UTC and went on without major problems for the whole shift.
No particular operations to report.

Parallel activity:
SGD: RTL optimization (Guo, Zhao)

Here a plot of the slow scan (with OMC temperature going up) over a whole FSR with images of the main peak.

Not sure how to interpret this.

We have performed the analysis of the throughput of the INJ system in the period of 6th May (after the PR centering activity: eLog 55725) to 28th June. The last throughput analysis done is reported here (eLog 55579).

The first attached graph shows the power trend of the transmission of the Mode Cleaner (top graph), the throughput of the injecton system (middle graph), i.e., is the ratio between the calculated power at the output of the PMC (we have estimated 7% losses from the output of the PMC to the output of EIB) and the power at the input of the interferometer (which is estimated to be 7.5% less w.r.t the output of the IMC due to SIB1 losses), and the throughput losses of the Mode Cleaner cavity, i.e., intracavity losses (bottom graph).

On the other hand, the second graph shows the mismatch of the beam to the IMC cavity. We can see that from the last time the mismatch has increase by roughly 0.5%, while the throughput losses have decreased by around 1.5%.

This morning I made some scans of the OMC in CITF as requested by the commissioning crew.

The slow shutter was opened at 08:49:25 UTC.

Figure 1 shows a time plot of all the scans, as one can see I first did "fast" scans (by mistake) between 08:52 UTC and 09:03 UTC (fig 2) and then I did a "slow" scan (temp going up then down) between  09:21:30 UTC and 09:42:30 UTC (fig 3).

Finally, the slow shutter was closed at 09:45:45 UTC.

The photodiodes B1 PD1 and PD2 remained closed during the scans.

Using the angular lines that are injected to the PR mirror I have computed the optical gain of the corresponding error signals. The calibration is based on the Local controls readout signals of the PR mirror, which give the information in radiants already. The calibration has been made by making the transfer function between Sc_PR_MIR_T* and the corresponding error signal. We have checked both the 6MHz and the 18MHz signals. The coherence and optical gain is shown in Figures 1 to 8.

IN order to make a comparison with simulation, we have normalized the optical gans by the total power reaching the quadrants and then we have combined the information from both quadrants (sqrt(QD1^2 + QD2^2)), to remove the uncertainty on the gouy phase of the sensors. All the information is summarized in the following table:

Yesterday, at the end of the shift we made a last trial with a bit less gain, and the hand-off to the IR control filter just after. This procedure worked very well, and we got to a stable state that we unlocked by hand. Figure 1 shows the time plots of the hand-off: first we turn off the line of DARM, then we make the hand-off and just after we pass the filter to the IR one. FIgure 2 shows the fft of the final state and Figure 3 shows an fft done on the previous trial, that we just did the hand-off. Notice that a 74 Hz line with sidebands is present even when the DARM line is off.

ITF State: ITF Locked on IR
ITF Mode: Not Locked
Quick Summary: IMC and RFC locked; all Suspensions loops closed, all SBE loops closed

Activities ongoing since this morning: No activities ongoing

The shift was dedicated to the planned ISC activity, Locking and ITF characterization, carried out by Casanueva and van Dael (see report #56297). The work proceeded for the whole shift without major issues.
Activity concluded at 21:40 UTC, infrared cavities left locked.

Other activities carried out during the shift:
SGD BPC with all DoF (two PSD), carried out by Guo.

DAQ
FdWRawFull stopped working at around 21:20 UTC. Process restarted via VPM at 21:38 UTC.

The goal of the shift was to recover the interferometer after changing the RH power.

Locking the CITF went very smoothly (we locked within 1 minute on the first attempt) and we didn't have to tune any phases except for B2_169MHz (minor change in demodulation phase) while doing the hand-off of CARM from the beating signals to the IR. After that we however got stuck trying to do the hand-off of DARM to the IR. We observed several things here:

• We noticed that B1p_56MHz demodulation phase was drifting quite a bit. We left the interferometer locked for 10 minutes before doing the hand-off of DARM to observe the drift of the demodulation phase and attached are three different figures (last three) taken at different timestamps in the lock of the B1p_56MHz_I and B1p_56MHz_Q channel. We tune the phase by looking at the DARM line at 74 Hz and the figures show that the peaks of the I and Q channels change over time, which corresponds with the drift of the optimal demodulation phase in the LSC_B1p_DPHI channel that we saw. Nevertheless when we look at the entire spectrum to determine the optimal demodulation phase then the found demodulation phase seemed to be rather consistent. The question therefore is whether tuning the phase on the DARM line is a good figure of merit at this stage for finding the correct demodulation phase.
• We noticed that going from 20 mW to 40 mW (On B7 & B8) by reducing the CARM_offset that the sign of the B1p_56MHz error signal would flip 180 degrees. 
• At 40 mW the optimal demodulation phase seemed to be consistently good and not drift anymore. We therefore used the same demodulation phase for the last few hours of the shift and we didn't observe the drift of the demodulation phase anymore.

We tried to do the hand-off therefore at 40 mW which after some tuning of the gain seemed to work for just a few seconds. After optimizing the hand-off (tuning the gain and ramptime), we managed to do a hand-off lasting roughly 40 seconds after which we unlocked while playing with the gain of DARM (see first figure). We made one more attempt at the hand-off after this which lasted about 10 seconds after which we concluded the activity.

We leave the arms locked on the IR.

Since yesterday we worked on the recovery of the system after the rearrangement of the filter cavity end bench.

1. Recovery of the SC alignment

We first tried to recover the SC alignment into the filter cavity. The SC was quite easy to find and with IR AA we could reach 5V in transmission. To investigate what is the reason for this low power, we did the following attempts:

1. Check the possible clipping of the IR PD. We manually scanned the mirror angle in front of the PD, and we could see in both x and y direction, that there is a period the power stop changing which means the beam is well inside the sensor.

2. Slow scan of the SC frequency to check the higher-order mode, it seems the TEM01 and TEM02 modes are both quite small. 

3. We moved the IR PD along the beam, in this way we could reach 8V. It seems the problem would be the beam size on the sensor, but this is not consistent with what we saw in step 1. 

I did a rough lock-unlock measurement to check the RTL, it seems the RTL is around 70ppm which is quite similar to what we have in the past. We will try to repeat this measurement in the following shift, if the RTL is fine, the low power on the IR PD maybe only caused by some local reasons on the FCEB bench.

One strange thing we noticed tonight is that the IR seems to oscillate with 0.5Hz frequency, which may be from the suspension. The possible reason could be the temperature in the clean room change due to our presence there. But during the work on Sunday, we unplugged and replugged the cable of the sensors on the end bench, it seems we could not get any signal after that(pic 1), we will inform the responsible person tomorrow about this. 

2. Recover the green pointing loop

Then we tried to recover the green beam pointing loop. To be noted, the power on the PSD was reduced to about 2V instead of 13V after the bench was rearranged, and an ND= 2.0 filter was installed in front of it. The new sensing matrix was made in the same way as usual. The new driving matrix inserted in the Acl is shown below:

ACL_MATRIX_BEGIN GR_BPC2_DRIVING       1    PSD_X_NORM         PSD_Y_NORM 
ACL_MATRIX_CH    GR_M7_BPC2_X       ""      66.94            -11.38          
ACL_MATRIX_CH    GR_M7_BPC2_Y       ""      -9.11            -66.48      
ACL_MATRIX_END   GR_BPC2_DRIVING

The new set points of the loop are 

ACL_CONST_CH    GR_M7_BPC2_X_SET    ""    1    LOOP_FREQ        1.65
ACL_CONST_CH    GR_M7_BPC2_Y_SET    ""    1    LOOP_FREQ        1.95 

The loop was closed with the new settings, we will check the performance of the loop tomorrow. 

In pic 2, we show the picture of the end bench now. 

In Figure 1 and Figure 2 (frequency zoom up to 40 Hz), it is shown the acoustic noise in the NEB room during the last six months when acoustic soundproofing of the HVAC air ducts was performed.
Focusing on the work performed on the supply air duct in the experimental room, at first glance there is no significant reduction in acoustic noise related to the interventions. It is worth noting that:

• in  Figure 3, a sligth acoustic reduction is visible at high frequency (> 200 Hz); this occurred on March 15th when the company was completing the installation of the sound-proofing panels along the vertical part of the supply air duct. In the bottom panel, the acoustic rms in frequency band (512-2048) Hz is reduced by ~14%;
• looking at the acoustic rms in the frequency (1-16) Hz and (16-512) Hz, in some cases, the reduction of acoustic noise floor occurs in coincidence of the overpressure (INF_NEB_PRES) reduction in the experimental hall, as observed a few months ago (#54893) due to an accidental opening of the door between the SAS area and the experimental hall. It looks like that the overprussere level could affect the acoustic noise floor, see Figure 4;
• in Figure 4, it is observed a worsening of acoustic noise in the hall coinciding with the last actions (May 3rd and May10th) on air ducts but it is not clear if it depends on the installation or the increase noise floor.