ITF found stable locked at LN2 in TROUBLESHOOTING mode.
ITF recovery completed. COMMISSIONING mode set at 17:05 UTC.
From 17:00 to 20:00 UTC - Vardaro performed the planned activity on FDS FC recovery. Lock acquisition stable, up to LN3_ALIGNED.
ITF left locked at LN3_ALIGNED in PREPARE_SCIENCE mode.
We starteed to realign the SQZ beam toward ITF.
Preliminary operation
SQZ_MAIN in SQZ_LOCKED_NO_FC, SQZ_FLT in GR_AA
Open SDB1 Shutter before the ITF lock
CC PLL stopped to work: we reduced the starting set point from 29585 to 29500.
Remove SQB1 Retroreflector: at 17:52:38. 300000 steps verticl positive. Then we relaigned mainly in vertical to optimize again the HD 4MHz mag. We measured about 8.7dB of ASQZ and 5.6 dB of SQZ
Removed SQB1 Shutter steps 500k downward at 18:05:21 UTC
HWP Rotated for Injection into ITF at 18:12:40 -32500steps
The interferometer unlocked at: 18:09 UTC
Opening of shutter and check alignment
New lock in LN2 at 18:25 UTC
B1_4MHz 0.18mV
Working on SQB1 M22 and M23 I managed to reach 1.6mV of magnitude but the signal is strange and also the CC error signal is strange. I don't manage to close the CC loop. i also disengaged the active phase noise cancellation loop but I did not improve.
I leave all the shutter open SQB1 and SDB1. Only the fast shutter is closed
The filter cavity is again realigned with IR. The back reflection of the FC toward the HD detector has the same direction of the beam reflected by the retroreflector, this means that the retroreflector is parallel to the FCIM and that the FC can be enabled and disabled but the beam propagation does not change.
Tuesday
We wanted to remove the retroreflector, but the SQB1 PicoMotors process was crashed. Thus tried to chek the SQB1 powersupply process and we noticed that the power supply was not reachable. We went in DET-Eroom and we reloaded the interface of all the power supplies. After that we could reach it from the process but the channel 1 (Pico motors) was setted at 0V. We came back in DET Eroom and we set the channel at 12V, After that we managed to switch on the picomotors process.
We removed the retroreflector by moving -330000 steps RetrorefY.
At that point we went in the state SQZ_LOCKED_NO_FC with SQZ_MAIN, we opened the SC shutter and we unlocked the filter cavity from the green lock. We saw some small flashes with the IR.
We locked the FC on Green and we started to scan the SC PLL frequency from 100MHz and 100.100MHz. This means that the SC frequency is scanned between 1260Mhz and 1260.1 MHz. We saw the TEM00 looking FCEB_IR_Cam. So we reduced the ramp to about 5 kHz around the TEM00. After that acting on SQB1_M21, SQB1_M22, SQB1_M23. M22 and M23 are after the retroreflector so they not impact with the alignment of the IR toward the ITF.
Once we reached about 10V with IR flashes we decided to try to keep the IR in resonance acting on SC PLL frequency and run the IR AA loop.
To do this we used the automation. We have to change three flags, two in SQZ_MAIN: fis_injection from true to false and sqz_flt from false to true and one in SQZ_FLT itf_fis_rr from true to false. Then we asked the state FLT_IR_AA_FMODERR tuned and we managed (skipping FMODERR) to relock everything. We noticed some issue in the DDS (periodical glitchise when we changed the frequency, we asked to Mateusz to solve this issue because it is typical after the restart of the DDS box) We left this state for the night.
Wednesday
We found everything aligned and we decided to check all the demodulation phases. For the green everything was still tuned. For the IR we had to reduce all the pahses by about 0.2 (for QPD) and 0.5 for the longitudinal phase.
We then decided to go towar the HD and we saw that the EQB1_HD_DIFF_4MHz_mag was lower than usual 7mV instead 8.5mV. So we unclipped the beam into the FI using SQB1_M23. Then we re-engaged the FC IR AA in order to do some iteration in both the direction of the FI. At the end we managed to have 8.35mV and about 8.7 dB of ASQZ. We then put again the retroreflector but the IR was completely clipped. We unclipped moving it in horizontal. Once unclipped we managed to obtain more or less the same level of ant-squeezing with and without the retroreflector.
Last step to recover was the FMODERR loop. We asked for a frequency counter to Flavio and we went in DET Eroom to measure all the frequencies of the generator involved in the loop. But for some reason the FMODERR loop was overestimating the length of the FC by 3mm so the FSR by 4Hz and the total frequency of the SC by about 10 kHz.
Thursday
Mateusz solved the issue of the DDS
In DET lab we centered the EQB1 GR Camera, We realigned the PD for IR AA and we centered the PD for GR PDH of the filter cavity.
We measured again all the frequencies of FMODERR (we left the frequency counter to heat up all the night and we also measured again the FMODERR generator calibration.
FMODERR Generator settings: FM modulation 600Hz, Amplitude 0.9Vpp and frequency: 78.919292 MHz
Then we measured everything with Frequency meter:
After that we tried to use also the active legth measurement of the FC to adjust the detuing of the FC and it worked at the first attempt.
Still to be done:
Phase scan with HD and FCIM
Characterize FIS with HD
Some oscillation in the GR IR transmission from FC (to be checked/fixed).
We continued with the ITF's recovery.
We tested the engagement of CARM_SLOW with different a different order of the executed commands, in particular me found that changing the loop's gain in parallel with the ENABLE switch allow for an engagement smooth enough to preserve the lock.
Looking at older versions of ITF_LOCK.py, this is the way the loop was engaged in the automation, before the blackout. Since this procedure is rather unusual, it was probably modified during the recovery.
We restored the older version of ITF_LOCK.py and tested the lock acquisition multiple times, successfully.
We also tried to remove the double phase change for DIFFp in carm_null, but the lock acquisition failed so we reverted the change.
We could reach LN3_Aligned, but the lock was killed by a glitch on the NI accerometer (fig.1), similar to those observed on the BS in the past. This happened at least one more time.
7:00 UTC At the beginning of the shift I found the ITF in CARM_NULL_1F, TROUBLESHOOTING. The weather during the night was stormy, which may have influenced the unlocks between 1:48 UTC and 4:04 UTC.
7:30 UTC Beginning of daily activities: FDS FC recovery (Vardaro), ITF recovery from CARM_NULL_1F (Boldrini, Pinto, Ruggi, Spinicelli)
10:14 UTC Earthquake in Greece, magnitude 4.9, 9km from the surface
14:43 UTC ITF in LN2
At the end of the shift I left the ITF in LN2 with recovery ongoing
This morning, NEN and NWN frequencies have been swaped with the WEN and WWN frequencies. We have now:
The ITF was found in Troubleshooting and in locked arms with the commissioning team working on the ITF recovery, see Status of the ITF.
In parallel, Marco worked on FDS FD recovery; all the activities were stopped at 20:20 UTC.
At 19:54 UTC the ITF achieved a stable lock in CARM_NULL_1F; the ITF will be left in that state for the night.
This morning there were few problems with the lock acquisition regarding a couple of critical points:
1. The hand-off of PRCL from 3f to 1f. It seemed that when the error signal was making the hand-off, the demodulation phase was stil changing causing unlocks. We checked in the logs and indeed that phase change was sent with a ramp of 20s. We hardcoded a ramp of 1 second in the Automation.
2. The engagement of CARM slow loop is very agressive and it causes also some unlocks. We commented these lines and we managed to stay locked in CARM NULL 1f, even with the good phases of the DIFFp (and we fine tuned the phase of PRCL in 1f). We are currently working in the engagement of this loop, which seems to be our stopper.
A look at long averaged coherence between the ground current monitor and hrec_gated, up to 250Hz - Figures 1 and 2:
The two checked times are: 30 Nov 2025 9:00, 6 hrs (current probe at ground rod on the top of electrical panel in front of INJ cryotrap); 22 Feb 2026 2:00, 6 hrs (current probe at ground cable of the BS base).
Similar coherences are found also with magnetometers - Figures 3 and 4:
ITF found in LOCKED_ARMS_IR State and TROUBLESHOOTING Mode.
All times are UTC.
07:04 Unlock at REDUCING_CARM_OFFSET_1_OF_3.
07:14, 07:24, 07:36, 07:46, 08:11 Unlock at LOCKING_CARM_NULL_1F due to saturation of SDB2_B1p_PD2 causing Fast Shutter to close.
08:00 FDS FC recovery activity started (De Laurentis, Vardaro).
ITF left in LOCKED_ARMS_IR State and TROUBLESHOOTING Mode with ITF recovery still ongoing.
Yesterday, after the weekly maintenance, we continued working on the full recovery of the ITF.
During the night before, the ITF has been put in Locked arm due to troubles in locking even in carm null. The issue was due to the demodulation phase of the Diffp (mainly the TX DoF ). However, as the previous day, we could survive during the transient only using a "bad" phase (slightly different from two days ago), with most of the signal on the Q quadrature.
Moreover, looking at the phase camera, we saw some jumps of the sidebands, similarly to what we had when the PR Chrocc was mistuned. Looking at the usual PR Chrocc observarbles, both current and heater temperature probes were higher then the values before power outage (see fig. 1). Thus, we decided to perform a small step down (~0.05V ), at 14h12UTC.
During the whole afternoon, we could stably reach CARM_null_1f, but the step on chrocc didn't show any macroscopic improvement to the cleanness of the locking signals.
We tried few times to continue to LN2, with success. However, as the day before, the locks in LN2 lasted only ~10 minutes, with the interferometer unlocking without a clear reason (some of the longitudinal DoF drifting away).
Finally, the lock acquisition continued to have problem during the night. The work on it will continue also today.
ITF found in CARM_NULL_1F in Troubleshooting mode and SQZ team at work on FDS FC recovery (Vardaro, Zendri).
Activity concluded at 20:00 UTC with ITF locked at LN2.
Unfortunately, after the activity, the ITF was very unstable. It unlocked sistematically at LOCKING_CARM_NULL_1F.
After several unsuccessfully locking attempts, according with commissioners, I relocked at LOCKED_ARMS_IR.
SBE
16:00 UTC - SPRB vertical position recovered via stepping motors.
Today while locked in LN2 we found the MAR actuation of SR and PR payloads slighlty mistuned (former values pre-shutdown were not saved). We measured the coupling and corrected the driving.
DSP cards properly saved afterwards.
Today, on site, we
-Switched on the output of the NCal power supply (which were switched off due to the large general power supply failure of Feb 28th).
-Replaced the NWN controller. The removed one is suspect of being wrongly parametrised (slow frequency ramp up observed on 14 Oct 2025) and we will check this at IPHC.
With the replaced controller, the NWN frequency ramps up much faster than with the old one (see picture). The NWN phase control also looks better and needs to be tuned further which we will probably do next Tuesday during maintenance.
Time details of the changes:
10:30 utc in WEB:
Power on WEB NCal power and restart all rotors at default frequencies
11:00 utc in NEB:
Successfully test restart of NEB NCals before changing NWN motor controller
Replace NWN controller by new one (we bring back the replaced one to IPHC)
Restart all NEB NCal power supply and restart all NEB NCals with success.
ITF found in Troubleshooting mode with arms locked.
At 7:00 UTC ITF in Maintenance mode, below the list of the activties communicated in control room:
| CH [W] | OUTER RING [W] | INNER RING [W] | |
| WI | 0.44 | 0.26 | 0.03 |
| NI | 0.61 | 0.59 | 0.09 |
At around 12:00 UTC we started to relcok the ITF; the commissioning team is now working at the lock recovery.
DET
SDB1 LC opened by the guardian; properly closed at 8:25 UTC.
The power emitted by the CO₂ WI auxiliary laser is shown in the attached figure.
As can be seen, the power emitted by the laser dropped to about half following the blackout.
Yesterday evening, taking advantage of Antonio’s intervention to replace the pump on the WI tower, I went to the TCS room to check whether a better operating point could be found by adjusting the duty cycle of the square wave (as we did in the past).
I tested different duty cycle settings and found that the duty cycle value that maximizes the power is 74%. However, the power level measured before the blackout could not be recovered.
This morning we replaced the failed pump on WI.
About the disturbances observed yesterday evening , the failed pump was offline but still connected to the controller (failure of the magnetic bearing controller, replaced just after the shutdown), we can't exclude noisy effects due to the faulty electronics.
The Imaging servers have been restarted with the release v17r0p7 aligned with the latest Fr(v8r49p0) and Fd(v8r65p1) librairies
Operations performed between 2026-03-10-06h17m33-UTC and 2026-03-10-06h20m37-UTC
After the WI pump failure, WI suspension is moving quite a lot, more with respect to the usual, fig.1.
I don't know if this is related to the failed locking attempts, tomorrow the problem will be better addressed.
ITF still in recovery - TROUBLESHOOTING at the beginning of the shift, which kept on going through the afternoon
Here the steps taken during the recovery and the edits to automation (Pinto, Spinicelli, Nardecchia, Was, Ruggi). LN2 was achieved three twice (18:30 UTC - vac pump failure, 19:57 UTC) lasting only a few minutes in both cases.
18:30 UTC Vacuum pump failure, fixed (Vacuum Team)
20:43 UTC ITF in LN2
20:56 UTC unlock, then unlock at LOCKING_CARM_NULL_1F (21:06 UTC), REDUCING_CARM_OFFSET_2_OF_3 (21:20 UTC), LOCKING_CARM_NULL_1F (21:30 UTC)
At the end of the shift, as instructed, I left the ITF in LOCKED_ARMS_IR - TROUBLESHOOTING, since it had not been possible to keep a stable lock in LN2 or to reach CARM_NULL_1F during the other locking attempts.
After the WI pump failure, WI suspension is moving quite a lot, more with respect to the usual, fig.1.
I don't know if this is related to the failed locking attempts, tomorrow the problem will be better addressed.
Today we continued the recovery of the ITF. We managed to reach LOW_NOISE 2.
At the beginning we noticed a lower power values on the 112 and 12 MHz sidebands, which was taken care of by Ilaria, acting on CH:
NI CH = 0.577W -> 0.708 W
WI CH = 0.324 W -> 0.416 W
At the end, we partially reversed the NI CH PWR increase: 0.708 W -> 0.630 W. (17.32 UTC finish of the actions).
After this, we had several issues in finding a good configuration of demodulation phases in order to get past the handoff from 3f to 1f in CARM null. In particular DIFFp was showing during the transient some misbehaviour (not entirely clear the rigth explanation), in which with good phases both I and Q (on TX DoF) were witnesses of some movement which ended up in a quick drift of the alignment and eventually the unlock. Curiously enough, the phase able to get past the initial transient was showing a ~45 deg mistuning. After several trials in which we noticed this behaviour, we implemented a double condition in the automation in which we use two phases, one 'wrong' for the transient and one good for the steady state (similarly to what we do for B4 56 MHz).
With this configuration we were able to lock the OMC, after a tuning of the OMC locking phases, and go up to LOWNOISE 2. However, after a while we unlocked due a to a WI tower pump failure. Antonio went to CEB to take care of it.
We leave the ITF locking on LOWNOISE2 for the night, to acquire some statistics on the acquisition.
N.B. the modification on the automation are the following:
ITF_LOCK.py: added lines 4681-82; 4685-86.
ITF_LOCK.ini, in the SHUTTER section: added the steady state DIFFp phases.
If the acquisition is robust enough, tomorrow we will proceed to fine tune all the several phases, if needed.
:)
Yesterday, after the weekly maintenance, we continued working on the full recovery of the ITF.
During the night before, the ITF has been put in Locked arm due to troubles in locking even in carm null. The issue was due to the demodulation phase of the Diffp (mainly the TX DoF ). However, as the previous day, we could survive during the transient only using a "bad" phase (slightly different from two days ago), with most of the signal on the Q quadrature.
Moreover, looking at the phase camera, we saw some jumps of the sidebands, similarly to what we had when the PR Chrocc was mistuned. Looking at the usual PR Chrocc observarbles, both current and heater temperature probes were higher then the values before power outage (see fig. 1). Thus, we decided to perform a small step down (~0.05V ), at 14h12UTC.
During the whole afternoon, we could stably reach CARM_null_1f, but the step on chrocc didn't show any macroscopic improvement to the cleanness of the locking signals.
We tried few times to continue to LN2, with success. However, as the day before, the locks in LN2 lasted only ~10 minutes, with the interferometer unlocking without a clear reason (some of the longitudinal DoF drifting away).
Finally, the lock acquisition continued to have problem during the night. The work on it will continue also today.
This morning we replaced the failed pump on WI.
About the disturbances observed yesterday evening , the failed pump was offline but still connected to the controller (failure of the magnetic bearing controller, replaced just after the shutdown), we can't exclude noisy effects due to the faulty electronics.
Also SC is aligned into the OPA. The heighest peak is 0.2% of TEM00
