The ultrafast photodiode for SC-CC beat note detection was misalignment a couple of weeks ago, and the signal was not recovered from remote since then On Friday I found a faulty BNC adapter at the photodiode output. After replacing the adapter and manual alignment, the signal was back again, and I could engage the SC automatic alignment loop. Fig 1 shows a locking sequence - the SQZ_EQB1_HD_RF_DEMOD signal measures the SC-CC beat amplitude with a negative slope of -25 mV/dB. I tuned the axial position of the two lenses of the SC telescope to improve the beat amplitude (and thus the mode matching on OPA); when doing so, one of the actuators ran into saturation, see FIg 2. I left the AA loop engaged over night; during the nigth, the actuator recovered the linear range and the beat amplitude went back to the optimal value, see Fig 3 and Fig 4.

To measure the mode matching of SC on OPA:

1) I sent the beam on the SQZ_EQB1_IR_PD_MONI photodiode

2) I set the OPA In scanning mode, and I temporarily removed the BAB by acting on the BAB/2f actuated mirror in the AEI squeezer box.

The TEM00 reflection peak is about 360 mV, see Fig 5. The peaks of the three strongest HOMs are about 7 mV (i.e. 2% of the TEM00), 3 mV (i.e. 0.8% of the TEM00), and 2 mV (i.e. 0.5% of the TEM00), see Fig 6.

After pumping, the position of F0 and F1 were found to be +4700um and +1800um while the bench ended up around +4200um. I have tried acting on Motor#3 to bring F0 around mid-range but the lower end-stop of the fishing rod was reached with F0 still at ~4400um.
This situation indicates lack of mass on the bench and will be easily fixed when venting the minitower. However, in such a configuration it's not possible to operate the vertical control. Therefore i acted on Motor#4 and lowered the bench to Y~ 0 in order to be able to close all the other loops. During these operations i could verify that both fishing rods now work properly. Here the list of following actions:

• LC Tx, Tz set-points changed and LC angular loops closed. LC controls are stable and LC-Coil-Total is less than 2V.
• X,Z orientation changed in the SBE process to match the LC's one. Now Z axis is pointing towards North and X axis is pointing towards West; this configuration holds for all suspensions in the QNR system.
• performed a new diagonalization of the SBE actuators; the residual coupling between the 3 DoF is again less than 10%. The old one, made before the recent installation of new cables, was not good anymore. The new actuation matrix indicates that the stiffness has increased significantly in X and Z.
• SBE loops (except for Y which is temporarily disabled) were successfully closed and they are stable. Optimization of the control filters will be done in the coming days.

I have left LC and SBE loops open for the night.

There were typos in the name of the angular degree of freedom in the previous entry. The following paragraph has been corrected:

We performed a OMC scan with MMT_M2_TY = +2770 steps at 12:36:00 utc:

TEM00 peak at 0.80 mW, order 1 at 0.0003 mW (0.04%), order 2 at 0.0087  mW (1.09%) > The mode matching seems to be slighty improved.

We performed a OMC scan with MMT_M2_TY = +3470 steps at 13:15:00 utc:

TEM00 peak at 0.80 mW, order 1 at 0.00065 mW (0.08%), order 2 at 0.0082  mW (1.03%) > The mode matching seems again slighty improved.

We performed a OMC scan with MMT_M2_TY = +4170 steps at 14:12:00 utc:

TEM00 peak at 0.785 mW, order 1 at 0.0004 mW (0.05%), order 2 at 0.0076  mW (0.97%) > The mode matching seems again slighty improved.

At this point we considered that a motion of +4170 steps on MMT_M2_TY was already quite large as it corresponded to a miscentering of the B1p beam on its camera of -800 um. We therefore decided to stop the rotation of MMT_M2 _TY at this level and we hold the new postion MMT_M2_TY = +4170 steps.

In my previous comment I wrote 'bad data received from outside'. Again, for this specific event this mechanism is not demonstrated (it was for other events, similar in the final effect). But it is possible, given the architecture of the suspension controls. SDB1 top stage control is involved in the complex network of Global Inverted Pendulum Control. It receives data 'from outside', namely PR top stage master board via TOLM system. PR receives data from several boards, and in particular data which are connected to global control output towards suspensions (GIPC is the control architecture which put the locking force in loop for top stage controls).

If a NAN enters a board in one position of this network, it is able in principle to reach boards quite far away, even if GIPC is not turned on, because a switch at zero is not able to stop it (NAN*0=NAN, in our codes). This is not simply theoretically possible: it happened several times.

Now I must say again that there is no evidence that this is the mechanism that put SDB1 suspension control board in a state that required a reboot. I'm just saying that, if somebody want to investigate on that event, the bad data received from outside has to be considered a trail to follow.

Yesterday we worked on the tuning of the mode matching of the single bounce beam (NI+BS aligned) with the OMC.

At the beginning of the shift we performed a scan of the OMC to check the status of alignment and mode matching (scan started at 07:39:20 utc, duration 5min):

TEM00 peak at 0.68 mW, order 1 at 0.058 mW (8.5%), order 2 at 0.0094  mW (1.4%).

Given the rather large misalignment, we locked the OMC and improved its alignment.

New scan performed after the realignment, at 08:26:20 utc (5min):

TEM00 peak at 0.77 mW, order 1 at 0.0095 mW (1.2%), order 2 at 0.0091  mW (1.2%).

The mode matching is therefore already better than what was achieved during the previous shift on OMC mode matching on March 27th (51233) when the best mode mismatch was 1.5%. Moreover the image of the B1 camera clearly indicate that the mismatch is dominated by a difference of waist or difference of waist position between the tangiential and sagital axes (clover shape).

We locked again the OMC and displaced the MMT_M2_TX dof. During all this activity the SDB1 bench TY was controlled in high bandwidth and the drift control was engaged. At each step we slightly realigned the OMC and checked the effect of the MMT_M2_TX dof on the power transmitted by the OMC (B1_PD3_DC) and on the channel B1_PD3_DC_norm_B1s corresponding to the transmitted power normalized by the reflected power, which is our main figure of merit.

From the previous table its seems that the figure of merit constantly improves. With MMT_M2_TX rotated in TX by -650 steps, we performed another OMC scan at 09:41:40 utc:

TEM00 peak at 0.785 mW, order 1 at 0.0009 mW (0.11%), order 2 at 0.0091  mW (1.16%).

The main improvement concerns the alignment, while the mode matching did not change very significantly. We hold the new TX position of MMT_M2.

We then tried different angular positions of the MMT_M2_TY dof, which, according to Zeemax simulations performed by Matteo Tacca, is the degree of freedom that should be the most appropriate to compensate a motion of MMT_M1_Z. Note that in the past shifts on OMC mode matching (27/03 and 11/03) we have already changed the MMT_M2_TY by +1370 urad. Therefore we start from this position.

We have already noticed that when the SDB1 bench TY is controlled in high bandwidth, a motion of MMT_M2_TY excites a large resonance around 0.95 Hz that is difficult to dump. Thus we decide the put the SDB1 TY control in low bandwidth. In this case the resonance is no longer excited when moving MMT_M2_TY which enabled us to keep the OMC locked while changing the angular position of the mirror. Results are summarized in the table below.

From the previous table, it seems that the fom constantly improved from +1970 to +2770 steps and then remained constant from +2770 to +4170 steps.

We performed a OMC scan with MMT_M2_TX = +2770 steps at 12:36:00 utc:

TEM00 peak at 0.80 mW, order 1 at 0.0003 mW (0.04%), order 2 at 0.0087  mW (1.09%) > The mode matching seems to be slighty improved.

We performed a OMC scan with MMT_M2_TX = +3470 steps at 13:15:00 utc:

TEM00 peak at 0.80 mW, order 1 at 0.00065 mW (0.08%), order 2 at 0.0082  mW (1.03%) > The mode matching seems again slighty improved.

We performed a OMC scan with MMT_M2_TX = +4170 steps at 14:12:00 utc:

TEM00 peak at 0.785 mW, order 1 at 0.0004 mW (0.05%), order 2 at 0.0076  mW (0.97%) > The mode matching seems again slighty improved.

At this point we considered that a motion of +4170 steps on MMT_M2_TX was already quite large as it corresponded to a miscentering of the B1p beam on its camera of -800 um. We therefore decided to stop the rotation of MMT_M2 _TY at this level and we hold the new postion MMT_M2_TX = +4170 steps.

Then we tried to improve further the mode matching by rotating again the MMT_M2_TX, that we brought to TX = -1000 steps, with the bench TY controlled again in high bandwidth. This position was reached around 14:54:40 utc.

We performed a scan in the final configuration at 14:55:20 utc:

TEM00 peak at 0.787 mW, order 1 at 0.0026 mW (0.3%), order 2 at 0.0072  mW (0.91%) > The mode matching seems again slighty improved, however the alignment is a bit degraded, therefore the conclusion is not so clear.

In any case it seems that we have reached a mode matching of the order of 99%.

While rotating the MMT_M2_TX/TY, the bench was under drift control, therefore the TX/TY angular positions of SDB1 changed. This induced large positive correction on the marionetta TX correction, that was compensated by displacing the SDB1 motorized counter-weight by +18000 steps.

At the end of the shift we updated the values of the new sepoints in SDB1_LC configuration.

Since the beam position on the B1p camera have changed by -800 um, we decided to update the TY setpoint of SDB2 in order to recenter the beam on the camera. The SDB2 TY setpoint was changed from 1100 to 900 urad (Fig.1). With this new setpoint we recover a beam position which is only 300 um away from the initial position on the B1p camera, and it remains close to the center on B5 camera. This should be acceptable for the photodiode centering.

We also recentered the B1p on the quadrants with -800 steps on the B1p_M1_H picomotor and +700 steps on the B1p_M2_V picomotor. B5 quadrant centering was not checked due to an issue of communication with the ethernet bridge ethmt3 preventing the opening of the shutters.

Problems encountered during the shift:

• In the process SDB1_PDU, the values of Socket 1 and Socket 2 that are displayed are always 0, even when we switched on the picomotors. We are confident that the drivers were powered since we could successfully actuate the picomotors.
• We could not communicate with the SDB2 quadrants. The Ethernet bridge ethmt3 is unreachable with a ping command. A similar problem had been encountered earlier this year (https://logbook.virgo-gw.eu/virgo/?r=50426). Several attempts to switch on/off the power supplies PowerUnit11 and PowerUnit14 did not restore the communication with this Ethernet bridge. Until this problem is solved, this will prevent to use the B1p and B5 quadrants. Below we report the list of commands sent in an attempt to restore the communication.

2021-05-08-19h41m21-UTC    info gouaty       Disable power for device Switch using PowerUnit11
2021-05-08-19h41m39-UTC    info gouaty       Enable power for device Switch using PowerUnit11
2021-05-08-19h43m06-UTC    info gouaty       Disable power for device Switch using PowerUnit11
2021-05-08-19h43m07-UTC    info gouaty       Enable power for device Switch using PowerUnit11
2021-05-08-19h43m28-UTC    info gouaty       Disable power for device Switch using PowerUnit11
2021-05-08-19h43m43-UTC    info gouaty       Enable power for device Switch using PowerUnit11
2021-05-08-19h47m10-UTC    info gouaty       'B1p QD2:Open B1p QD2 galvo loop' sent to SDB2_Quadrants
2021-05-08-19h47m12-UTC    info gouaty       'B1p QD1:Open B1p QD1 galvo loop' sent to SDB2_Quadrants
2021-05-08-19h48m03-UTC    info gouaty       Disable output of PowerUnit11
2021-05-08-19h48m41-UTC    info gouaty       Enable output of PowerUnit11
2021-05-08-19h52m50-UTC    info gouaty       Disable output of PowerUnit11
2021-05-08-19h52m54-UTC    info gouaty       Disable output of PowerUnit14
2021-05-08-19h55m16-UTC    info gouaty       Enable output of PowerUnit14
2021-05-08-19h55m21-UTC    info gouaty       Enable output of PowerUnit11
2021-05-08-20h00m43-UTC    info gouaty       Disable output of PowerUnit14
2021-05-08-20h00m48-UTC    info gouaty       Disable output of PowerUnit11
2021-05-08-20h01m31-UTC    info gouaty       Enable output of PowerUnit14
2021-05-08-20h02m03-UTC    info gouaty       Enable output of PowerUnit11
2021-05-08-20h05m51-UTC    info gouaty       Disable output of PowerUnit11
2021-05-08-20h06m20-UTC    info gouaty       Enable output of PowerUnit11
2021-05-08-20h07m52-UTC    info gouaty       Disable output of PowerUnit11
2021-05-08-20h08m20-UTC    info gouaty       Enable output of PowerUnit11
2021-05-08-20h12m45-UTC    info gouaty       Disable output of PowerUnit11
2021-05-08-20h14m21-UTC    info gouaty       Enable output of PowerUnit11
2021-05-08-20h18m13-UTC    info gouaty       Disable output of PowerUnit11
2021-05-08-20h18m21-UTC    info gouaty       Enable output of PowerUnit11

2021-05-09 11h46m08 UTC    gouaty Disable output of PowerUnit11

2021-05-09 11h46m36 UTC    gouaty Enable output of PowerUnit11

The goal of the shift was to investigate the issues of the dark fringe degradation while reducing the CARM offset.

Since the brightening of the fringe for little CARM offsets has been such a recurrent and critical problem in the past shifts, our idea was to arrive at a position that is known to be stable (offset 300 Hz) and investigate the available signals for the control of the arms for possible hand off candidates.

While the arms behaved very well, the CITF required to be pre-aligned a couple of times before we managed to lock it at around UTC 18:00:00.

We searched for a hand off candidate for DARM by injecting a line @ 70 Hz. We found that LSC_B7_B8_DIF_NORM shows good coherence and signal to noise ratio (Fig.1).
For CARM we did the same with the 1111 Hz line. We found that LSC_B7_B8_SUM_SQRT shows good coherence and signal to noise ratio (Fig.2)

Once we found these candidates for the hand off, we measured coefficients and signs with their TFs with DARM and CARM INPUTs respectively and proceeded to the hand off.

We performed the DARM hand-off at 20:23:30 UTC (Fig.3) and then again at 20:54:00 UTC. We performed the hand off for CARM at 21:03:00 UTC (Fig.4). Both hand off were achieved very smoothly, and we proceeded to reduce the CARM offset. The CITF unlocked shortly after we arrived at offset = 100 Hz. Very noticeably, the well known issues of the two spots on B1p didn't manifest while reducing the CARM offset with the new inputs.

This solution can potentially be improved greatly by implementing appropriate filters for the hand off, so it offers handles to improve the stability of the CITF that will be further explored next week.

We leave the ITF with the arms locked on the IR.

The Shift was dedicated to the planned ISC activity, carm offset, carried out by Casanueva and Boldrini.
In parallel the SQB2 bench was put under Vacuum (see entry #51701).
ISC Activity concluded at 21:25 UTC, infrared cavities left locked.

DAQ
(07-05-2021 18:05 - 07-05-2021 18:30) From remote
Status: Ended
Description: At 18:00 UTC the signals related to ENV and ACS of the various scientific Buildings became unavailable (see attached plot).
Actions undertaken: I contacted the DAQ Oncall (Masserot) that from remote was able to solve the issue. The data were available again from 18:30 UTC.

ISYS
(07-05-2021 19:45 - 07-05-2021 01:00) From remote
Status: Ended
Description: From 18:55 UTC The PSTAB became noisy (red flag on the DMS). Under request of the commissioning crew I tried to open and close the related loop, but this action did not solve the issue.
Actions undertaken: I contacted the INJ Oncall (Chiummo) that investigated the problem. Unfortunately, he couldn't restore the nominal condition, further investigation needed.

During the period at 150 Hz offset, the B1p_56MHz_I signals was compared with the DARM_FREQ signal, as a possible handoff candidate. The coherence between the two signals is high as shown in figure 1 fo this comment.

The aim of the shift was to further progress in the CARM offset reduction procedure.

Both arms green lasers were locked at  7.10 UTC but the ALS_NEB_PD_REFL signal (reflection green laser signal) showed too large power fluctuations for the lock to be stable (see figure 1 and 2 for a comparison between the West arm and North arm signals fluctuations).  These fluctuations were related to the SNEB oscillations.  In particular, glitches on SNEB_LC_TY caused spikes in the reflection PD signals and, ultimately, unlocks.

The issue was solved by moving the SNEB_LC_Z from -310 to -280 um. While this removed the large spikes on TY and the North green laser unlocks, the fluctuations on ALS_NEB_PD_REFL remained highly correlated with SNEB_LC_TX, with a high coherence peak at ~0.5 Hz.

After this, we proceeded with locking CARM and DARm on the green signals 8.23 UTC, and then 2 kHz of CARM offset were subtracted.

The DRMI was then aligned and locked at 8.44 UTC and a successful handoff to the 3f signals was first obtained at 9.04 UTC.

We proceeded reducing the carm offset. Several attempts of reducing the carm offset were performed during the shift, all concluded with unlocks. The B1p and B4 DC triggers were disabled each time 500 Hz of offset was reached. The lowest carm offset reached was 150 Hz at 9.18.40 UTC. THis lock lasted ~30 seconds and then the DRMI unlocked simultaneously with a power spike on the B7 and B8 signals, indicating a sudden passage trough the IR arm resonance (see figure 4). No particular glitches were shown on the CARM and DARM signals in that moment. Following, the list of the other carm reduction attempts unlocks:

During most of these unlocks, the shutters on B2 PD1, B2 PD3, and B7 closed and required a rearm of the PDs.

During the shift the CARM and DARM loops unlocked 9.28 11.00, 12.20. Most of the times they relocked immediately without particular issues, except for two periods in which high frequency noise tampered with the lock stability (between 7.50 and 8.20 for WEB and between 12.40 and 12.55 on NEB).  At 12.55 the NEB BPC tilt x setpoint was changed from 0.28 to 0.32 to recover the stability of the NARM ALS lock.

The activity continued in the following shift.

This morning has been dedicated to ISC activity, the work went on all the shift without any major problem and will go on in the afternoon...
 
Others activities carried out :

- SQB1 closing and evacuation;
- SQB2 opening and intervention on multisas;

Electricity
at 6:39UTC we lost the data and the connection over the net with the TB_UPS3 (experts informed).

SBE
at 7:37UTC SNEB has been moved over the Z direction solve the problem of noisy TY (#51695).

As requested by the Polgraw Group, around 11:30 UTC I swopped the two sensors 5.1 and 5.7 in chain 5 because of malfunctioning of sensor 5.1 (input ethernet port of the sensor was not working fine). They will perform further testing operations remotely.

The filter cavity end bench has been installed. The distance between the top of the bench and the mid of the flange has been set to ~10 cm. It can be further adjusted if needed.

OB DMS thresholds have been updated for F4. All SAT motors are now off.

The shift was dedicated to the planned ISC activity on Carm offset. It has been carried out by Bersanetti and Boldrini.
It went on till 21:00 UTC without major problems for the whole shift.

DAQ
15:03 UTC FbTrend_10800 process crashed. Killed via shell and restarted via VPM.

ISYS
PSTAB Noisy (PSTAB_PD2_AC_MONIT). Val 0.15 instead 0.01. According with Chiummo and ISC team, I opened and closed PSTAB via VPM. Problem temporary fixed.

On-line Computing & Storage
Trouble with Olwin session: during the actions to restore the vertical position of SDB2 the Olwin session closed several times making the operation difficult.

SBE
21:00 UTC - SBE SDB2 vertical position restored via Stepping Motors.
Unfortunately during the restoring the Actuator ACT_F0H0 started to drift down close to saturation. I tried to contact The SBE Sys Manager without success. I called Bertolini who was acting from remote at the FCIM Mot Switch, and he provided to restore the standard state. Operation concluded at 21:52 UTC.

SQZ
16:10 UTC - Activity concluded by Vardaro, Tacca. SQB2 in Vacuum, SQB1 still in air (task postponed to tomorrow).

SBE
(06-05-2021 23:15 - 06-05-2021 23:45) From remote
Status: Ended
Description: SDB2_ACT_F0H0 Actuator behavior

Friday 30th april SQB2 was put under vacuum. Alessandro relaized that motor 3 under vacuum was stucked in both the directions.

We had the impression that the screw of the motor 3 had the same problem we found out one month ago in the motor 3 of SQB1. Today we open the dome in order to replace the screw but we realized that it was moving perfectly in both the direction both by hand and by moving the motor.

We put again the motor 3 and 4 at 3/4 of its range toward down, we closed the dome and we asked again the vacuum team to evacuate SQB2.

We will try to move again motor 3 in vacuum. If we will find again int stucked we will replace the screw.

Goal of the shift was to continue to decrease the CARM offset.

We began the shift by improving yet again the alignment of the green laser in the West Arm, as the VCO kept locking on a higher order mode. After this correction, the arms locked smoothly and quickly for the rest of the shift.

Similarly, the CITF behaved well enough to attempt a CARM offset reduction several times during the shift, and we managed to get to 80 Hz away from the IR peal. We observed a few interesting things during these attempts:

• as suggested by the morning crew, we tried to improve the degradation of the dark fringe while approaching the IR peak by moving the mirrors of the arms. In particular, we tried to chase the separation of the spots on B1p by moving the WE on TX and TY with steps of 0.2 urad. The CITF lock and the arms were capable of coping with this somewhat aggressive intervention quite well, until the arm unlocks, and this approach helps in controlling this problem at least temporarily, since at no point the spots ceased to move away form each other.
• towards the end of the shift, we accidentally sset the CARM offset to 0.01 Hz (meaning ~20 kHz away from the IR peak) in about 10 seconds. While doing so we crossed the IR resonance on the arms, but  we did it so quickly that both the green and the CITF lock survived this process, and the shutters on B7 and B8 didn't close. This suggests that the ITF is robust to quick CARM offset changes. To test this, we "jumped" back, going to CARM offset = IR peak+150 Hz in 10 seconds, crossing again the resonance, and the lock survived once more.

During our several attempts we somewhat refined some procedure to get down to low CARM offsets:

• usually, with even a fine pre-alignment of the CITF, it is ok to set the CARM offset 2 kHz away from the IR resonance; however, when only the CITF unlocks and tries to relock immediately again, putting back the same offset is not enough as the PR alignment loops greatly improve the recycling gain. As a rule of thumb, a good offset is 3 kHz instead, and the figure of merit is flashes on B7 and B8 being lower than ~ 1mW;
• the CARM offset reduction can be done with no manual interventions down to 500 Hz; at this point, although still premature but safer in case of brightnening of the fringe, we can disengage the triggers for B1p_DC and B4_DC; the way to to do that is use the following commands, which are already reverted by the DOWN state of the DRMI_LOCK Metatron node:
  • LSC.B1p_DC_TRIG = 1
  • LSC.B4_DC_TRIG = 1
• at the same 500 Hz offset it is often needed to start reducing the MICH gain, by a factor ~ 33%, and the PRCL gain of a factor ~ 25%;
• moving then to 150 Hz can be done by reducing both the MICH and PRCL gains to an overall 50% of their initial values;
• during the shift we also did some tests increasing the PR_TX loop gain by 50% (already from the beginning); its relevance has to be studied a bit more, so for the time being such value was not updated in the Metatron configuraion file.

Minor things happened during the shift:

• we had an unlock of the IMC, and after its recovery we opened and closed the PSTAB loop to recover it;
• we suffered an unlock due to some missing data from the NI (F7 and PAY flags becoming grey on the DMS);
• we noticed the vertical correction of SBE_SDB2 constantly increasing up to - 5V, but Nicola could recover it.
• during the shift we had to re open the B2_PD1, B2_PD3, B7_PD1 shutters many times

We leave the arms locked on the IR beam.