This morning, we continued the tuning activity started yesterday (69761). 
The initial TCS settings are summarized in the table below.

The CARM NULL locks were limited by a vertical oscillation of the PR at around 11 Hz, which was causing the ITF to unlock. Paolo and Manuel took care of the issue and managed to solve it.

At some point, we also experienced a few unlocks due to the green losing lock. This issue was investigated and solved by Piernicola.

At this point, since the B1p fringe was already quite dark (~0.016 mW), we decided to try a common step on both DASes in an attempt to increase the sideband power. We therefore decided to increase both the WI and NI DAS powers by 20% at 14:42 UTC (STEP 3)

Unfortunately, shortly afterwards, the ITF unlocked due to an earthquake, which prevented further lock acquisition attempts for some time (earthquake at 14:19 UTC).

Once the earthquake transient had approximately settled, we noticed that the sideband powers in CITF had decreased by about 10%.

ISC then attempted two lock acquisitions, but both failed at the same step (CARM offset, step 3/3).

Therefore, before retuning the lock acquisition, we decided to also increase both the WI and NI CH powers by 20% (STEP 4, at 16:25 UTC).

The sideband increased but not recovering the level before the STEP 3. Some attempts were made to reach CARM NULL, but we did not manage to reach it. The ITF always unlocked at the same step (CARM offset, between step 2 and 3).

The final CO₂ laser configuration is summarized in the table below.

We could not recover CARM_NULL_1F after the TCS steps: the lock parameters need a more careful tuning after these changes.

Presently, the lock breaks at step 2/3 of the CARM offset reduction.

We also undid the change in the /virgoDev/Automation/PyALP/PyUGF/algo.py, resetting the UGF servos step to ±\pm5%

Figure 1 and 2 shows two screenshots from the video of WI. One towards the end of the CARM NULL lock and well before. Jumping back and forth between the two pictures there is a faint absorber visible in the middle of the image. And that point disappears shortly after the unlock. 

Figure 3 shows that faint point circled.

Figure 4 and 5 shows the same type of analysis for the NI video. There I cannot notice any points, but it might be harder to notice on the color videos than on the black and white one.

In the following, the link to NI and WI thermal cameras videos of yesterday's lock acquisition starting around 14:39 UTC, CARM NULL (of around 40 seconds) and a brief time after, together with main ITF signals.

https://owncloud.ego-gw.it/s/NDrWctmKC5ebyb4 

As already observed (69742), there is no clear evidence of PAs.

ITF found locked in CARM_NULL_1F in COMMISSIONING mode.
The planned activity on: CARM NULL locking / DAS adjustment (Bersanetti, Mantovani) started at 07:00 UTC. Still in progress...

Another comparison of ITF signals during two different CARM NULL acquisitions:

• 11 September, around 15:17 UTC, less than 10 minutes long: without TCS [purple curves]

• 17 September, around 06:58 UTC, around 20 minutes long: with TCS, while tuning was ongoing [blue and orange curves]

In the plots containing two superimposed signals for each acquisition, the two reference curves are both shown in purple. They can be matched to the corresponding blue and orange curves by their relative vertical position: the upper purple curve corresponds to the upper curve in the TCS-on data, and likewise for the lower curves.

Overall, all the monitored signals appear to have improved after the initial TCS tuning, with the exception of the B1p 6 MHz balance.

The commissioning activity on CARM NULL locking /DAS tuning proceeded through the afternoon until 19:30 UTC #69761

The ITF was able to stay in CARM_NULL_1F for 20 minutes from 19:41 UTC to 20:01 UTC and another 21 minutes from 20:29 UTC to 20:50 UTC

21:00 UTC At the end of the shift it was left in LOCKED_ARMS_IR

Guard Tour

20:02 UTC

During the day, we will try to stabilize CARM NULL 1F with the goal to have the possiblity to tune the DAS directly at CARM NULL.

The starting point is the following:

We performed some tuning attempts. The first step regarded the CH. 

We increased CH powers by 18 % at 9.24 UTC.

At this point, we were able to reach CARM NULL 1F quite systematically. However, the ITF was repeatedly unlocking because of the high power on B1p, which exceeded the threshold for triggering the fast shutter (0.2 mW).

We therefore contacted DET to slightly increase the B1p DC threshold. Michal increased it from 0.2 mW to 0.25 mW.In parallel, in an attempt to reduce the B1p fringe, at 12:26 UTC we increased the WI DAS outer ring power by a factor of 2, assuming that the WI absorption is approximately twice that of the NI.

After that, we realized that the DIFFp loop had not been working properly since yesterday.

The issue was fixed, and we continued with the lock acquisition tuning. However, we then noticed another problem with the COMMp loop and decided to keep both the COMMp and SOFT loops open.

With this configuration, we were able to reach CARM NULL 1F and maintain the lock for about 10 minutes.
In the first figure, the two CARM NULL locks with and without TCS are shown. The situation seems to be improved at the sideband level.

Concerning the etalon, the sinusoidal behavior is clearly visible. The increase of 112 MHz observed in the upper-right corner of the plot shown in Fig. 2 is due to the change in the DAS settings.

 

ITF found in locked arms IR.

At around 7:30 UTC started the planned activity on CARM NULL locking /DAS tuning; the activity went on without major issues and it is still in progress at the end of the shift.

DMS / TCS

The TCS flags have been tuned and readjusted to take into account new configurations provided by the experts.

LSC

2026-09-16 07h51m06 UTC    berni 'NI:NI Set point [NI_RH_SET=19.85,rampTime=3600]' sent to LSC_Etalon_Acl

Another reason for transients in sideband power in the afternoon is that the inner ring of the WI DAS has been turned on during the maintenance this morning, and turned off around 14:30 UTC.

The operator instructions need to be updated, as unlike during O4, at the moment the inner ring of WI DAS is not used (https://logbook.virgo-gw.eu/virgo/?r=69742), so after the CO2 beam checks during the maintenance the flip mount needs to be left in the position that blocks the beam. 

Today we tried to recovery a stable lock in CARM null 1f in order to proceed with the TCS tuning, however we encountered some troubles during the shift.

We weren't able to find a stable lock up to carm null , unlocking most of the times at CARM step 3/3 (or randomly before), so right before carm null 3F.

At first we didn't find any clear explanation to the several unlocks. We tried to tune the sevaral phases and gains in order to make more robusts the steps, with no success, without even being able to engage the Diffp loop in drift control. At the end we found a correlation to a misfunctioning of the NI Etalon loop.

After the adjustement of the setpoint performed in the beginning of the afternoon (https://logbook.virgo-gw.eu/virgo/?r=69752) we noticed that the loop was still not properly working since it was not converging correctly to the setpoint as the temperature was still increasing above it, see fig.1.

A correlation was found with the Etalon misbehaviour and a change in the sidebands powers alonf the shift, see fig.2.

Expert was notified about the issue. Investigation will be performed to understand if the misfunctioning is at the level of the hardware or at the level of the control loop logic.

ITF found in COMMISSIONING Mode and LOCKED_CARM_MC_IR State, with activity on CARM_NULL/TCS DAS tuning ongoing. 
Activity concluded around 19:30 UTC.
ITF left in COMMISSIONING Mode and LOCKED_ARMS_IR State.
 

This morning we went to EER to investigate the extra noise on the EIB LVDT. The idea was to check the 10 kHz generator used for the modulation/demodulation.
However, when arriving, we noticed that the noise had actually been lower since the 26th of August (Figure 1).
On that day, the chiller of the Neovan electronics, situated right next to the rack of the EIB electronics, stopped working at 14:00 UTC. We went to EER starting from 14:15, but the noise was reduced at 14:40...
Given these new elements, we suspected that some of the cables of the EIB electronics could be loose and sensitive to being touched.
We checked them, and some of them were indeed quite loose. Touching them was introducing glitches in the control loop. We tightened them. It still seems quite sensitive.

Anyway, the noise seems to be similar to the level it was before the 14th of August (Figure 2: blue now, purple 14/08 before the noise appeared, red 17/08 with the extra noise, green on 27/08). It is worth noting that the horizontal LVDT now seems to be higher. It is not easy to draw clear conclusions since these signals also depend on the seismic activity. Also, the noise reduction seen in the RMS on the 26th is not so clear in the FFT.
We will keep monitoring the situation and will speak with Henk Jan about whether or not we should undo the changes he made to the control filters on the 21st of August.

All these tests were done with the injection down and the flip mirror at the output of the PMC flipped up.
When relocking the injection with the same BPC corrections, the beam appeared to be quite far away from the IMC cavity, particularly horizontally.
We suspected that the EIB position was different from the morning. This seems to be confirmed by the fact that the BPL corrections have changed, the beam seen by IBJM is in a different position and the corrections of the EIB actuators are different. This is the case even though the only EIB position that changed was the Tx position of the EIB, which was 1 µrad away from this morning (Figure 3).
We are quite surprised by this. It might be evidence that something is wrong with the position read by the LVDT (?).

6:00 UTC ITF found in COMMISSIONING mode and LOCKED_ARMS_IR state and set to MAINTENANCE mode and DOWN state

6:30 UTC DET_MAIN node set to SHUTTER_CLOSED and PAUSE, SDB2 angular and position loops open, DET photodiodes Vbias disabled for DET EER faulty power supply replacement (9 swapped with 31). At the end of the operation the offset of the B1_PD3 was large and was corrected by Romain. Metatron put back in standard state at 11:30 UTC #69748 #69754 (Berni, Gouaty, Masserot, operator)

7:00 UTC WI optical levers cabling check and Removal of WI MIR PSD power switch #69749 (Boldrini, Cavalieri, Majorana, Nocera)

8:20 UTC Reset of 10 kHz modulation for EIB LVDTs (Gosselin, Lagabbe)

10:09 UTC NI mirror temperature increase: NI etalon set to 19.78 ramp 1 #69752 (Bersanetti, Mantovani, Spinicelli, operator)

Other operations carried out during maintenance:

• Nitrogen refill (Vac group)
• CB cleaning (external company)
• TCS chillers refill (Menzione)
• TCS thermal cameras checks and power checks (operator)
• Installation of cameras in CEB (Berni, Kraja)
• DAQ room, INJ EER, DET fiber length checks (Di Biase, Kraja)
• Fire estinguishers inspection in CEB (Fabozzi)

OMC scan/NI single bounce configuration skipped due to other operations ongoing (DET, DAQ, INJ)

12:00 UTC ITF in COMMISSIONING mode and LOCKED_ARMS_IR state

After the morning operations the DET_MAIN node was found stuck and unable to reach SHUTTER_CLOSED.

First we checked that the shutter was indeed closed by looking at the image of the camera B1s and increasing the integration time at maximum: no light was visible.

We noticed that the offset on the B1_PD3 photodiode was large (about 0.45 uW). Therefore we corrected it in the configuration of SDB2_dbox_bench (reloaded at 11h39 utc).

After the correction of the offset the standard state of DET_MAIN could be recovered by requesting the following states:

* FORCE_CHECK_CLOSED

* SHUTTER_CLOSED

For comparison, here we plot the B1p beam shape in CITF respectively  without any TCS compensation, with only DAS (after ~1h) and with both DAS and CH.

The NI etalon loop was not working properly due to a too low setpoint (we could not raise to do not interfere with the commissioning activities), for unknown reasons the temperature of NI increased, see figure 1.

We decided to open the loop (to discharge the corrections) and re-closed with higher set point.

Francesco replaced the Hameg_SN09 with the Hameg_SN31(box containing the Hameg power supply, rear and front wiring) used by the LeftUp and LeftDown  DAQ boxes

After this operation,

• the SDB2_PowerSupply server configuration has been updated accordly and successfully restarted
• the SDB2 DAQ boxes have been successfully reconfigured 
• the Photodiodes and Cameras have been set to their expected state
• The DET_Img_CEB server, used to acquire the SDB2 cameras, have been  restarted 
• and the SDB2_LC and the SDB2_SBE loops were successfully closed 

The attached plot shows  the trend related to SDB2 DAQ boxes power supplies

The switch box has been restored, the fault was simply a cable not well connected.

The behaviour of the MIR signals is identical to what it was before the intervention (fig.1 before, fig.2 after). We also restored the damper gain to its previous value.

The glitches observed on MIR PSDF in May (fig.3, 69083) are still there, although their shape is slightly different.

We turned off and removed the switch box between the MIR PSD and the respective power supply on the WI payload.

The MIR signals are now offline, and the damper's gain was set to 0 beforehand.

We will try to repair the switch box and reinstall it before the end of today's maintenance.

The planned activty of the shift was:  TCS CARM_NULL_1F recovery ; it went on without major issues.

Activity concluded at 19:00 UTC; ITF left in locked arms IR.

The goal of this shift was to recover CARM_NULL_1F.

During the afternoon the ITF slowly continued to thermalize after the TCS actuators were enabled, causing the sidebands to rise in power at a steady rate that only started forming a plateau around 16:00 UTC (Fig.1).

This forced a continuous back and forth to adjust gains and phases, following this moving target, until the thermalization process stabilized.

Towards the end of our shift we managed to reach CARM_NULL_1F but could not stabilize this state, as a 11 Hz oscillation, slowly but steadily rising, eventually breakes to lock. The main culprit seems to be SRCL gain being too high, with the oscillation being wide enough to overwhelm the UGF servos, preventing them to adjust the gains. This oscillation is witness by a vast number of longitudinal and angular loops, making diagnosis more difficult.

Overall, the demodulation phases are adequate in 3F and good enough in 1F to allow the transition, but we could not complete their fine tuning before the end of the shift.

A fourth event occurred the 2026-09-13-11h28m55-UTC and was promplty recovered .

This morning, we first checked that the CO2 actuator powers were still consistent with the values set last week, based on the OSCAR simulations and summarized in the table below: 

Only a small adjustment of the NI CH power was needed, probably due to the recent interventions on the cooling system. At the end of the check, all the CO2 flip mirrors were closed again. 

A first attempt to reach CARM NULL was performed around 08:10 UTC, but it was not successful. No clear evidence of PA was observed on the ITM thermal cameras. 

We therefore decided to return to CITF and proceed with the switching ON of the CO2 actuators. CITF 1F was reached at around 08:44 UTC, with the following initial sideband levels:

• 112 MHz: 0.027 mW

• 12 MHz: 0.051 mW

At 08:45 UTC, the DAS flip mirrors were opened, with the powers reported in the table above. The evolution of the sideband levels was then monitored for almost two hours. A general worsening was expected with the DAS actuators ON alone. 

After about one hour, at 09:55 UTC, the 56 MHz and 6 MHz sideband levels had decreased by approximately 30% and 20%, respectively. We agreed to manually unlock the interferometer at 10:10 UTC. 

The ITF returned to CITF 1F at 10:20 UTC, with the following sideband levels:

• 112 MHz: 0.0195 mW (-27%)

• 12 MHz: 0.0355 mW (-30%)

The levels were then quite stable. The trend during this first period is shown in Figure 1. 

At 10:34 UTC, the CH flip mirrors were opened, again with the powers reported in the table above. The sideband levels increased almost immediately. Already after about 20 minutes, at 10:55 UTC, they had nearly recovered their initial values. 

After the lunch break, some adjustment of the PR angular loop gain was needed to allow the ITF to lock again in CITF 1F, which was achieved at around 12:05 UTC. The sideband levels then reached: 

• 112 MHz: 0.0305 mW (+12%)

• 12 MHz: 0.0544 mW (+7%)

The trend during this second period is shown in Figure 2.

Just before the Daily Meeting, the first attempts to reach CARM NULL started.

The activity continued with the afternoon shift.

ITF found in COMMISSIONING Mode and LOCKED_ARMS_IR State,
All times are UTC.
07:06 Start of commissioning activity - Turn on DAS/CH in CITF (Lumaca, Corubolo, Nardecchia, Bersanetti, Spinicelli).

The shift was dedicated to TCS activity with centra interferometer lock. Activity still ongoing.
ITF left in COMMISSIONING Mode and LCOKED_ARMS_BEAT_DRMI_1F State.