Motivation: Tommorrow we will try and optimize the input mode matching. This is based on elog 67078 which concluded that the input mode matching was suboptimal. 

In May last year, the west arm test masses were changed. So, I will not summarize any measurements prior to that date. Clément and I did some measurements during the west mirror swap, and those are summarised in VIR-0907A-25.pdf

June Measurement (mwas, elog 67078)

In June, Michal et al. changed the RH powers. 

North: 7.4 W -> 3.4 W
West: 8.1 W -> 4.0 W

At the same time, a noise peak in DARM shifted from 10700 Hz to 11500 Hz. 

December Measurements (cjacquet et. al.)

In December, several measurements were carried out. In all cases, the RH settings were:

NE Coil 1 16.143 V,  0.265 A, NE Coil 2: 16.1425 V, 0.263 A

WE Coil 1: 17.5745 V, 0.293 A, WE coil 2: 17.5728, 0.297 A

Voltage uncertainty +/- 0.0001 V; amperage uncertainty +/- 0.001 A. This was stable across all measurements (as report by in loop sensors, so caveats apply). 

This amounts to 8.53 W in the north arm and 10.36 W on the west arm (note this was tuned again in as of Thursday 22nd at 17h00 it was 8.65 W and 10.19 W).

NB: In all cases I take the FSR to be 49968 Hz. 

December 2nd (Elog 68300)

Clément and I took some data, but no results were posted and I have no reanalyzed the data.

December 15th (elog 68385)

Scans were taken using a CARM offset. This means there was no 1064nm beam and so no thermal lensing, the method is described in the Opt Char Measurement guide. The ring heaters were set to the above values. Using a quick python script, I found the 2nd order mode separation to be 11,731 Hz or 0.235 fractional FSR. See attachment 04. Note that I used a "downward" scan, so the frequencies in the graph need to be read as FSR - f,

My quick python script was done "blind" to validate Cléments results quasi-independantly, but I present it here as part of a coherent narrative. 

December 19th (elog 68417)

On December 18th the cryotrap was closed, in the morning, it was opened and the effect the on the mirror observed. As the mirror cooled down, A shift from a fractional FSR position of the second order mode from 0.241 -> 0.236 was observed. This was 12,042 Hz -> ~11731 Hz. Details can be found in his elog.

December 22nd (elog 68423)

On December 22nd, the RH's were tuned off in both arms. Aat the start of the measurement the RH's were in their nominal configuration, as listed above. Details in Cléments log. For both arms, the shift was 0.235 (11,742) -> 0.205 (10,243 Hz).

Warning: What is strange here is that the mode separation frequency is going down when the RH's are tuned off. This is the opposite to in the june measurement.

Summary of measurements so far
Putting these in a line we have:

• 10,235 Hz (No ring heater + cryotrap)
• 11,724 Hz (Ring heater + cryotrap)
• 12,042 Hz (Ring heater + no cryotrap)

Estimating hot curvature - Part 1, Estimating the effect of the RH

In VIR-0907A-25.pdf, on page 8, Clément gives the LMA measurements as:

LMA_measurements = dict(NI=1424.6,NE=1692.6,WI=1425.5,WE=1695)

Ideally, we would have taken data with no cryotrap & no ring heater. Then we could compare that gouy phase number directly against the LMA measurement. We don't have that, so using the LMA measurements for the input mirror and the second order mode seperation frequency, I solved for the end mirror RoC. In the North arm, it is 1653 m with the no RH + cryostat and 1678 m with the RH + cryostat. In both cases it comes out to +25.0m. 
 

Warning: This implies that the RoC seen by the mirror increased when the RH was turned off, i.e the mirror got flatter/less curved. This is not what I would have expected. 

Estimating hot curvature - part 2, Effect of cryotrap & true curvature.

There is a descrepancy between the LMA measurement and the in situ measurement. Our leading explaination for this is a systematic error in the measurement at LMA, which would be common for all mirrors. Calculating the mode separation frequency for several different offsets, we can find an intersection at -17.5m. See attachement. 

We can then recompute the effect of the ring heater, with the static offset applied. We compute the effect of the RH to be +24.9m on the ETMs and the systematic offset with the LMA measurement to be -17.5m.

The NE RH was set to 8.53 W, therefore, 2.92 m/W
The WE RH was set to 10.36 W, therefore 2.40 m/W
The analysis for parts 1 and 2 is in this python file https://git.ligo.org/IFOsim/Finesse_playground/-/blob/21e5f39423fc9cc7245d5c8b5796b48d5617aee0/aaron_jones/2026_virgo/03a_utils.py

Estimating hot curvature - part 3

We can then recalculate the curvatures for the RHs after the TCS tuning last thursday. The method is shown in this file and relies upon Ilaria's TCS tuning documents:

The expected second order mode freq is: North: 13082 Hz West: 13206 Hz.

We can then explore the range of 2nd order mode frequencies as a function of absorbtion and RH power (caveat that I'm not totally sure on Clément's measurements). This is shown in the attachements. 

A task that was no slot to fit in was to check if after the tuning of the past two weeks something significant changed for the LN3 SR misaligned sensitivity comparable to the O4 science run. 

I have been in front of my computer when the interferometer was relocking, so I went into commissioning and changed the end state to LN3 SR misaligned. It stayed locked in that state between around 12:00 and 12:30 UTC, and after that unlocked by itself. I have reverted then the end state to LN2, so that SR could realign, and then someone (or the automation) changed the end state to LN3 SR aligned after 30 minutes, which was the right thing to do.

Before starting the magnetic line and magnetic sweep injections today, we performed a few tests at WEB, since the last lines injection (#68547) did not work properly.

We injected a magnetic sweep via the Metatron node and a magnetic line via VPM; however, in both cases no magnetic field was generated. Both the coil switch-on and the DAC signal sent to the coil worked as expected. At the moment, we have not identified the origin of the issue. During the maintenance on Tuesday, we will perform a check at WEB to inspect the setup and investigate the problem.

The injections started at ~11:20 UTC and ended at ~18:30 UTC and the output files are available in (/virgoData/NoiseInjections/MagneticInjectionsO4/output):

Far-field sweep magnetic injections

• CEB--> MagneticSweep_NOISE_MAG_CEB-1453288903.txt
• NEB --> MagneticSweep_NOISE_MAG_NEB-1453289832.txt

Low frequency lines injection (lines parameters are reported below)

• NEB--> MagneticLine_NOISE_MAG_NEB-1453290488.txt
• CEB--> MagneticLine_NOISE_MAG_CEB-1453302290.txt

FreqList = 2.577 3.677 4.677 5.777 6.777 7.477 8.977 10.577 11.977 13.577 15.377 17.577 19.177 22.577 24.577 
DurList = 900 900 900 900 900 900 900 900 600 600 600 600 600 600 600
Amp= 2.0 2.0 2.0 4.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0

**** The interferometer was left in PREPARE SCIENCE, LOW_NOISE_3_ALIGNED with AUTORELOCK_FAILSAFE ****

I think not only the DCP poor stability matters, but also the fact that we are not working in the maximum. The attached figure shows  SRCL vs DCP correlation and the presence of an offset. The way we compute SRCL set is an indirect estimation of optical spring, assuming that the power in the arms is correlate to DARM line only by radiation pressure, but any other coupling mechanism could create an offset.

I think we could try to demodulate the DCP wrt a line on SRCL and use it to define the optimal SRC longitudinal working, as we do for the angular working point. The line could stay below 10 Hz as for the angular lines.

Figure 1. Increasing the DARM line by a factor 2 has been effective at making the optical response pole (DCP) more stable. It also reduced the SRCL INPUT fluctuations. 

Figure 2. The DCP frequency and SRCL INPUT after the change (purple) are less correlated and have a lower spectrum below 1mHz than the previous night (blue). So there is improvement, but SRCL INPUT still dominates the slow fluctuations of the DCP with a coherence of ~60%. 

Having a larger DARM line create a large bump at 74Hz in the sensitivity. This is acceptable during the current commissioning period, but wouldn't be acceptable during a science run. In the long term a DARM line at lower frequency should provide a better signal for SRCL, as the radiation pressure effects scale as 1/f^2, so should be much larger and clearly visible in B7 for a lower frequency DARM line.

ITF found in LN3_ALIGNED with EDB OMC lock activity ongoing (Was)
20:00 UTC end of activity
21:00 UTC etalon ramps started (operator)
ITF left in PREPARE SCIENCE, LN3_ALIGNED with AUTORELOCK_FAILSAFE

Background: We are trying to assess the matching between the input beams and the arm cavies, to support an effort of the INJ on monday to optimize the mode matching. Mismatch between injected beam and the arm cavities creates second order mode content in the cavities. The frequency noise means there is always a small sideband at ~ 10 kHz, once scattered into the second order mode, it is resonant. This shows up in DARM. 

For instance, see the attached plot "B1 over SSFS". We can see a small bump in the DCPDs at ~ 10700 Hz. In the bottom panel, we can see high coherece between the freq. noise and the DCPDs, strongly suggesting that this is freq noise. We can also see a phase shift as we go through the resonance of the cavity. 

Objective: The objective of this work is to create a calibrated Finesse model that is representaitve for the PRM transient measurement. 

Assets: The finesse code can be found here. You'll need some files from the Virgo DAQ https://git.ligo.org/IFOsim/Finesse_playground/-/tree/eabd12fdbe33ff868ec6dcac85bb72b8c6c664b0/aaron_jones/2026_virgo. All the datadisplay templates can be found in my home directory, then 2025_01_22_PRM_RH, its readable by everyone.

Method: I set up a simple model, no RF locks and pretuned it. I then injected obtained the freq noise to strain transfer function, following my earlier method: https://finesse.docs.ligo.org/finesse-40m/examples/mody_input.html

I then measured the strain data and SSFS error signal, as shown in the attachment SSFS_Err_pre. The SSFS error signal was not calibrated, but a line is injected at 227 Hz. You can see it clearly in the top left plot. This line has high coherence (0.89) with the same line in the Hrec. I used this as a calibration factor. Then I projected the error signal through my model to obtain an estimate for the freqeuncy noise coupling into DARM. There is also a calibtation line at 3323 Hz, but the coherence was only 0.60 so I didn't use it. This is a big caveat, and so I have not been able to rule out an unexpected frequency dependance. 

Results: The plot showing the freq noise is attached the freq noise projected into DARM. 
Conclusions: No conclusions, this was a preparation exercise. However, 

Next steps: Now the calibration is done, I can retake data with high density around the 10700 Hz area. Then I will vary PRM RoC and see if I can replicate the frequency noise coupling in Finesse. 

Figure 1 summarizes the realignment of the EDB OMC using the USB TEM00, and then the lock on the carrier HOM between order 2 and 9

14:35 UTC Locked on EDB OMC 56MHz USB TEM00. Power around 0.5mW on B1t.
Realignign the OMC
15:06 UTC (6min) lock on USB TEM00, between 0.55mW and 0.6mW. With low
jitter peaks at a few hundred Hz

15:22 UTC (15min) lock on carrier order 2 mode, unstable, jumping back - Figure 2
and forth between the vertical and horizontal mode, the later having
double the power

15:45 UTC (10min) lock on carrier order 3 mode, horizontal peak - Figure 3

16:03 UTC (10min) lock on carrier order 4 mode, horizontal peak - Figure 4

16:22 UTC (10min) lock on carrier order 5 mode, vertical peak

16:37 UTC (10min) lock on carrier order 5 mode, horizontal peak that has 3 times more power

16:58 UTC (10min) lock on carrier order 6 mode, horizontal peak - Figure 5

17:18 UTC (10min) lock on carrier order 7 mode, horizontal peak

17:48 UTC (10min) lock on carrier order 8 mode, horizontal peak - Figure 6

18:08 UTC (20min) lock on carrier order 9 mode, horizontal peak - Figure 7

Restarting the EDB OMC scan with  freq 0.0005Hz, ampl 0.5, offset 23.27

19:00 UTC increased DARM line amplitude (74.4Hz) by a factor 2 and reduce OS calibration by a factor 2, using a Cm command.

Figure 8. The change in DARM line clearly broke the LSC_DCP_moni_mad_cal monitor of the DCP frequency. The Hrec is still working. Note that the Hrec measurement of the DCP follows almost exactly the SRCL_INPUT signal, after the factor 2 increase in DARM line amplitude it appears more stable. To be confirmed with a few more hours of data if the DCP really becomes more stable with a higher SNR for the optical spring signal that control SRCL set point.

ITF found in LN2, at 7:08 UTC I locked in LN3_ALIGNED as requested by Romain.

The ITF unlocked twice from LN3_ALIGNED thus Romain measured the impact of OMC mode matching on the interferometer in LN2; activity concluded at around 13:30 UTC.

At 14:20 UTC Michal started to work in LN3_aligned.

Software

CoilsSbWE restarted at 10:18 UTC after a crash

This morning we scanned the longitudinal position of the SDB1 bench in order to study the impact of the OMC mode matching on the interferometer figures of merit with the ITF locked in LN2.

The conclusion of the shift is that we were able to improve the optical gain by about 3% when displacing the bench by -3000 um. We tried a further bench displacement but without clear conclusion on the optical gain and BNS range, thus we came back to the previous position.

Since at the beginning of the shift the ITF did not reach LN3, we decided to work in LN2. Maddalena engaged manually the high bandwidth control of BS.

Activity starting at 08h45 utc with ITF in LN2 and BS full bandwidth engaged.

Initial position of SDB1 bench is : SDB1_LC_Z = 3708 um , Sa_OB_F0_X = 2600.

Start a ramp of -3000 um at 08h53 utc (Fig.1). We observed that the optical gain has increased by about 3% while moving the bench  by- 3000 um. The BNS range also seems to have improved (by maximum 1Mpc) but it is not very clear if this is linked to the motion of SDB1. The DPC frequency was rather stable. From Fig.2 we can see that the decreasing power trend on B1p and B1s was already present during the previous lock. It is likely to be still the transient of the etalon effect. While scanning SDB1 bench we also observed a decreasing of the B1s DARM signal, which clearly indicate an improvement of the mode matching.

ITF unlocked at ~8h39 utc. The ITF was relocked in LN2 and Camilla engaged the BS in highbandwidth. We started collecting reference data from 10h14 utc. Unfortunately we noticed later that the B1p QD1 galvo loops were open. So we closed them at 10h21m43 utc.

At 10h32 utc we restart the SDB1 Z scan, moving down by -1300 um. Since we started observing a degradation of the BNS range and some fluctuations on the optical gain (while the B1s DARM signal was getting close to the zero crossing), we stopped the SDB1 scan, at the position corresponding to SDB1_LC_Z = 90 um. Looking at the trends on Fig.3, it seems that the degradation of the BNS range and the CMRF is not linked to the bench motion, as the situation seems to get back to normal after a moment.

We restart the scan at 11h11 utc,  moving by about -700 um (see Fig.4). Another ITF unlock at 11h22 utc. It seems that we have reached the zero crossing of the B1s DARM signal. Besides this, we do not see any improvement (or any clear degradation) correlated with the last scan of the bench. 

ITF relocked in LN2 at 13h16 utc. We bring back the bench at the position corresponding to OB_F0_X = -400 um (and SDB1_LC_Z = +1185 um ) where we have found a better optical gain this morning. This position is reached at 13h30 utc. We will now use it as new nominal position for SDB1.

New SDB1 TX/TY setpoints (TX=180, TY = 48) updated in SDB1_LC configuration file at 13h37 utc.

Figure 1. The effect of the ring heater was clearly visible in the arm power, with a decrease in arm power. There was no associated peak in the B1p power. However, superposed to it there is a trend of increase in the B1p power that started before the PR RH was turned on, and continued afterwards. Also the difference in power has not returned to its starting state.

Figure 2. A likely explanation is that the etalon loop has been changed yesterday, which resulted in a transient during the night spanning 0.07 degrees (1/3 of an etalon fringe?). We have seen the etalon loop affect the arm cavity power, especially increase the difference between B7 and B8. So the steady state difference in arm powers after the PR RH transient is likely due the etalon, although there is also the tail of the PR RH cooling, this will complicate a quantitative analysis of this data.

Figure 3 The B1p and B1s power also increased during the night, on a slow time scale, likely due to the etalon. The order 3 mode has increased by a factor 4, and we have seen last week that the etalon can do that for a half a fringe variation.

This morning I have reenabled the EDB B1t photodiode by hand, it must have turned off during the unlock around 6:00 UTC.

ITF found locked at LN2 in COMMISSIONING mode and DAS power tuning activity in progress.
At 14:50 UTC Gherardini and Tringali switched ON the PR RH power supply.

The TCS activity went on without major issues till 20:00 UTC. ITF still locked at LN3.

Following Michal’s indication, at 19:18 UTC, in order to make the effect induced by the ring heater more evident, we increased the power to 3 W (V = 10.6 V).

The PRM ring heater was then switched off at 21:14 UTC.

In order to evaluate the impact of DAS tuning, performed yesterday, on the mystery noise Vs B1p_DC, I tried to find two gps relatively close in time, corresponding to different TCS working point, having a clear difference in terms of B1p_DC, but the same value of DCP and other parameters relevant for the sensitivity.

My best choice has been:

1453039518 - 14:05 UTC - blue data

1453058818 - 19:26 UTC - red data

At lower B1p corresponds lower noise. The difference in terms of floor is confined in a relatively narrow band (80 - 130 Hz); other differences seem more related to structures going up an down. The variation of floor is compatible to a rescaling of the mystery component by about 10 - 15 %.

Doing this analysis with more data and statistics is a bit difficult because a similar variation of noise occurs when the DCP chenges by a relatively small amount, like 20 Hz, which is the normal amplitude of fluctuation even when the alignment stability is relatively good.

In preparation for the input mode-matching measurement, Michal asked us to switch on the PR ring heater at the end of the DAS tuning activity.

To do this, Fabio, Maria, and Aaron went onto the platform in the CB to switch on the PR-NI ring heater power supply.

After completing the DAS outer rings power tuning, with the ITF in LN2, the PR ring heater was switched on at 1 W (V = 6.1 V) at 16:56 UTC.

Today, we continued the investigation of the DAS working point power started yesterday (68543). Between last night and this morning, the ITF unlocked and relocked twice without issues. Therefore, the lock acquisition seems to work properly with the DAS working point left for the night.

With the ITF in LN2 (and BS control in full bandwidth) we performed three different steps:

STEP 10 | 08.33 UTC:  increase of the NI DAS outer ring by +10%. (pickoff from 0.636 W to 0.691 W)

STEP 11 | 09.42 UTC: decrease of the WI  DAS outer ring by -10%. (pickoff from 0.273 W to 0.247 W) 

STEP 12 |11.05 UTC:  increase of the NI  DAS outer ring by +10%. (pickoff from 0.691 W to 0.753 W) 

For each step, it was quite difficult to assess whether it was beneficial or not (see Fig.1).  The most significant effect was a loss of sideband gain, so after discussing with Michal and Madda, we decided to revert all steps in order to recover the pre-shift condition. All steps were reverted at 13:08 UTC.

STEP 13 | 13.08 UTC:   WI DAS OUT +10% and NI DAS OUT - 20% (restore the starting configuration) 
At 13:27 UTC, the ITF unlocked due to an earthquake . We exploited the unlock to check the DAS powers by closing the flip mirrors one at a time  (completed at 13.48 UTC).
The ITF relocked in LN2 at 14.20 UTC.

During the transient of STEP 10, we observed an initial improvement in the BNS range and on both B1p and B1s, therefore, we performed an additional step:

STEP 14 | 15.05 UTC UTC:   NI DAS OUT +5%.
All steps performed today are schematically represented in Fig.2. 

The DAS outer rings power tuning can be considered completed. 

Hence, the reference powers from now on are as reported in the table.

The work on TCS (DAS tuning) went on in the morning with the ITF locked in LOW_NOISE_2.
13:27 UTC unlock due to an earthquake in Kamchatka (magnitude 6.3)
14:19 UTC relock to LN2

Sub-systems reports

Air Conditioning

• TB heater stopped, backup switched on, spare part shipped today, it will arrive between tomorrow and Saturday (#68546, Pezzimenti, Soldani)
• WE heater malfunction, fixed (Soldani)
• FCIB chillers 1 and 2 malfunction, fixed (Andreazzoli)

Yesterday night (Jan 21), the magnetic line injections were performed via the ENV METRATON node in NEB and WEB.
Regarding the CEB magnetic lines injection, this morning I found that it was still ongoing. In principle, it should have been completed during the night, but there was an interferometer unlock (during NEB injection). I stopped the CEB lines injection this morning since the interferometer had been unlocked to allow the start of the scheduled shift activities.

The output files are /virgoData/NoiseInjections/MagneticInjectionsO4/output:

• MagneticLine_NOISE_MAG_CEB-1453096947.txt
• MagneticLine_NOISE_MAG_NEB-1453060711.txt
• MagneticLine_NOISE_MAG_NEB-1453073344.txt
• MagneticLine_NOISE_MAG_WEB-1453085146.txt

The following report has been submitted to the On-call interface.

On-call events -> Air Conditioning

Title: Caldaia GPL TB in blocco

Author(s): Pezzimenti

Details:

Ricevuto allarme sms da dms per caldaia gpl tb in blocco. In contatto telefonico con Soldani, avviato caldaia caldaia di backup a gasolio.

* Note that any files attached to this report are available in the On-call interface.

Tonight I found the automation stuck in LOCKING_DC_READOUT, because the new command to enable EDB_B1t was not working:

2026-01-21T22:52:40.581Z ITF_LOCK [LOCKING_DC_READOUT.run] USERMSG 0: EZCA CONNECTION ERROR: Any Other Error: Could not get value from channel: SDB_B1t_PD1_VBias

Looking at the error, META_library and the new configuration, one hypothesis is that the function that enables Vbias extracts the name of the bench from the first segment of the process name, but in this very specific case the name of the process and the name of the bench do not share the same segment (SDB and EDB respectvely), so the standard function cannot work and something specific needs to be devised.

I commented the engagement of EDB_B1t and the lock acquisition could proceed. To be debugged tomorrow.

Another long standing bug is that the autorelock selected LOW_NOISE_2 instead of LOW_NOISE_3_ALIGNED as target state. This used to happen with LOW_NOISE_1 and LOW_NOISE_3 as target, but the recent increase of states made the bug scale up the wrong request as well.

The configuration for EDB_B1t has been added to META_library, and the engagement of the PD has been added just before reaching LOCKED_DC_READOUT. To be tested at the next acquisition.

As agreed, we started today’s DAS tuning in LN2 (8:17 UTC).

The actions performed on TCS DAS outer rings are summarised in the table below.

SUMMARY TABLE

The STEP 1 (WI OUTer ring +10%) seemed to cause a slight increase of the SBs and a deterioration of B1p, B1s and CMRF. 

Due to this, we decided to try to act on the other arm, but the STEP 2 (NI OUTer ring +10%) led to SBs getting lower, while B1p and B1s don’t seem to be changed. Furthermore, we noticed that BNS range worse starting from STEP 1.

We agreed to revert STEP 1, to see the effect of only STEP 2, doing STEP 3 (WI DAS OUTer ring -10%). As a result, SBs continued to go down, while other figures of merit  (B1p, B1s, BNS range) came back to this morning's values. 

Since the SBs decrease seemed to be correlated with STEP 2, we tried to revert the action on N arm with STEP 4 (NI Outer ring -20%). B1p and B1s remain untouched, while SBs and BNS range get better and then worse. 

Before making any other change, we agreed to try an unlock and repeat lock acquisition.

12.03 UTC ITF unlocked.

12.42 UTC ITF locked in LN2.

Since we would like to understand if some other trend was going on in the meanwhile, we made STEP 5 (NI OUTer ring +10%) to restore the thermal actuation condition of this morning. 

After around 1h and half we noticed that all the figures of merit almost reset the values to the one of this morning, except for SBs: after some investigations, Maddalena noticed that it could be related to Etalon loops (see Figure 1: behaviour of SBs level VS WI/NI temperatures from monday, as reference; the SBs level minima are reached, at different moments, between 10 and 14 UTC in the past days).

With STEP 6 (WI OUTer ring -10%) we start to explore the last direction left with the WI DAS. The effect was quite evident on SBs and also on B1p and B1s, all getting better.

Since B1p and B1s get slightly better but SBs lower dramatically, and the BNS range as well, we agreed to perform another step, the STEP 7 (NI OUTer ring -10%), to recover the SBs keeping the lower achieved values of B1p and B1s. 

STEP 8 (NI DAS OUTer ring +10%) and STEP 9 (NI DAS OUTer ring +10%) were performed because the previous direction does not seem to be the right one. After these two steps, BNS range recovered, SBs improved and other figures of merit remained pretty stable.

Michal noticed that the MICH control loop was not working properly and corrected it manually just before letting the ITF go in LN3_ALIGNED (locked in this state at 19.50 UTC), in order to start the ENV low-noise injections.

In Figure 2 the general behaviour of the main signal during all the steps performed and Figure 3 a zoom of the last hours are shown.

The thermal configuration left for the night is reported in the table below.

Today I changed, for LOW_NOISE_3_ALIGNED, the amplitude of the arms dithering lines and the darm LF line (74.4 Hz) at the same level as in LOW_NOISE_2. The OS/DCP calibrations and OS gain have been changed accordingly. To be tested later once we move forward from LOW_NOISE_2.

The PD flag in the DMS had been updated to include a check on the EDB B1t photodiode Vbias. The flag should turn red whenever the Vbias of this photodiode is not engaged after DC readout.

The DetMoni process was restarted at 16h45 utc.