IM Temperature seemed to be reaching the equilibrium temperature because of the non-operation of the REF4 cryocooler.

The temperature of 50K REFARM_HEAD seemed to show a tendency to slightly increase again.

Finally, the EX 50K REFBRT HEAD temperature seemed to show a decreasing tendency?

The temperature of BRT2 was clearly below the critical temperature. Then the vacuum level at EXC/T also decreases more.

I checked the Pcal-Y beam positions on ETMY with the new semi-automation code (klog33767).
There was no problem.

Beam position information was updated:
 

I changed the voltage from 10.6 to 12.7 to 12.2 because the temperature started decreasing again after the second head temperature (4K_REF4_4K_HEAD) reduction of REF4 cryocoolers.

[Aso, Komori]

We attempted to detach the IX main chain from the recoil mass chain by applying combinations of offsets to MN pitch (±4000), roll (±4000), and yaw (±2000).
To avoid frequent adjustments of the MN oplev position, we monitored the IM transfer function, which is typically measured during the suspension health check in adding these offsets.

The red line in the attached figure shows an example of the measured transfer function from IM Y to Y.
For comparison, the blue and green lines represent the transfer function in a healthy state and the one measured just after contact occurred (as shown in klog:33725).

Despite trying various offset combinations, the transfer functions remained similar to the red line, suggesting that the suspension could not be detached using these offsets.

Aso, Komori, Tanaka

This morning, Aso-san and Komori-san tried to perform releasing ITMX by applying the DC offset. Sorry, I did not join this work this morning so I'm not sure of the detail (They will report about this later, I hope.). However the situation seems not to be changed, unfortunately.

From this afternoon, we tried to identify the touching point by measuring DC responces from the excitation to oplev signals, and by comparing IX's responces with IY's ones. The reason why we used oplev signals is that the calibration factor of the photosensor responce may be changed by cooling and the amount of the change may be difference between IX and IY. As for MN, we measured the DC responces from MN_SUMOUT_{L,P,R}_OUT to MN_MNOLDAMP_{L,P,R}_IN1 at 300mHz. And we measured the DC responce from MN_SUMOUT_Y_OUT to MN_MNOLDAMP_Y_IN1 at 100mHz because 300mHz was yaw resonance frequency so we avoided it. As for IM, we measured the DC responces from IM_SUMOUT_{L,P,Y}_OUT to TM_WIT_{L,P,Y} at 100 mHz. At that time, we centered the MN(TM) oplev with when we measured the responces with MN(TM) oplev with BF Yaw DC control and moving masses for PIT/Roll.

Fig.1,2,3,4 show the DC responces about MN {L,R,P,Y} and Fig.5,6,7,8 show the timeseries of IX/IY MN oplevs {L,R,P,Y} when we excited them. Fig.9,10,11 show the DC responces about IM and Fig.12,13,14 show the timeseries of IX/IY IM oplevs when we excited them. Following table is the summary of the DC responces at the excitation frequency and the difference between IX and IY.

According to the result, the discrepancy of the responces of IM horizontal directions seem to be larger than MN ones. So somewhere IM in horizontal direction is touched?

Abstract:

Touching of the suspension might start from May 6, which is earlier than the last day when we could lock PRFFPMI (May 7).
Since the drift of the photosensor signals due to cooling is large, it is hard to say suspensions drift to which direction.

Detail:

To investigate where is the suspisious touching point, I checked ITMX photosensor signals.
Figure 1 and 2 show the time series data of MN and IM photosensor signals (euler coordinate), respectively.

In both cases, slope of the drift changed at the left T cursor (around 19:00 of May 6).
It means suspension might be touched each other even May 7 (the last day wee could lock PRFPMI).
Also, we can see large jump of MNT photosensor signals around 2025/5/9 0:30 JST (right cursor in fig1) while the other DoF doesn't change.

Since there is a large drift due to the efficiency change of photosensors (LED light emmission efficiency increases when temperture becomes low), it is difficult to say where is a suspicious position for touching.

Fig.1 is the temp data in the cryopayload when I applied heat in IM from ~ 8:00 am.

As we expected, IM heat could easily transfer to MT because they are directly connected using heat links. Also, MTR temp changed as IM and MT. Is this as we expect or strange??

When the moving mass was moved, maybe by Yokozawa-kun around 8:20 am, the temperatures in IM, etc. became 40 K ~ 60 K (~ 120K) for a moment and decreased rapidly.

But why is the temperature of IM highest? The moving mass was set in MT?

Can we believe these temp values? It could be because of the electrical interference? MT temp data became zero, maybe during moving the mass.

I adjusted 7V -> 7.2V -> 4V > 6V -> 6.5V now.

Fig.1 shows the temperatures in the IX cryopayload when I applied heat in IM in last night.

IM responses were caused by the change of applied voltage(current) to the heater. However, almost no reactions in other places. This could mean that one of the contacted materials is a low thermal conductive material such as plastic/resin-like?

With Shingo Hido

We have developed a semi-automated (guided) Python script to assist with the photo capture and analysis for Pcal beam position estimation.

The script has been tested and confirmed to work as expected, although there are still many areas for improvement.
Shingo Hido kindly tested the script in a trial run.

Details and usage instructions are documented here:
https://dac.icrr.u-tokyo.ac.jp/KAGRA/DAWG/CAL/PCAL/Tcam_know/auto_Tcam

I applied current to the IM heater on IX payload because it reached 38K.

The voltage was 7.0 V.

Aso, Komori, Tanaka

## Abstract

We scanned MN by applying the DC offset in OPTICALIGN from 800 to -3600 cnts in yaw and from -1700 to 1000 cnts in pit after removing the resisters on MN coil drivers to release ITMX. However, the situattion seems not to be changed, unfortunately. We try to scan it in the other region (over 800 in Yaw?) tomorrow.

## What we did

First, we removed the resistor between MN coil drivers and MN coils to increase the range in the SAFE state of ITMX. Then, we changed the state from SAFE to ISOLATED and turned off the comp gain filters in FM9 of MN_COILOUTF.

Then, we performed the centering of TM TILT OPLEV QPD by changing the setpoint of BF DC control in Yaw direction and by moving the moving mass in PIT direction. This time, the moving mass labeld PITCH_AR seems not to be responded, also the mass labeled PITCH_HR seems to be responded only the negative direction. Fortunately, the direction was the same as which direction we want to move. So we moved the mass labeled PITCH_HR to centering. (We don't confirm that whether the mass was at the edge of the range in the positive direction.)

After that, we performed the centering of MN TILT OPLEV QPD by moving the stage manually. Fig. 1 and Fig. 2 show the value of the micrometer in PIT and Yaw before centering, respectively.

Next, we moved MN in negative Yaw direction by applying DC offset in OPTICALIGN and checked the PIT/YAW couplling in MN and TM oplev changed or not. We applied the DC offset by 200 cnts. Whenever the TM oplev beam was out of the range, we centered the QPD by adjusting the setpoint of BF DC control or by moving the moving mass. Whenever the MN oplev beam was out of the range, we centered the QPD manually. At last we moved MN from 800 cnts to -3600 cnts in terms of OPTICALIGN, but the situation seems not to be changed. 

We performed the similar trial for PIT. In this time, MN and TM Yaw were at each edge of each QPD. In this state, we scanned OPTICALIGN from -2000 to 1000. However, the situation also seems not to be changed.  

During these trial, we found there seems to be hysterisis.

Fig.3 and Fig. 4 are ones after the trial in PIT and YAW, respectively.

We try to scan it in the other region (over 800 in Yaw?) tomorrow.

Abstract

In the calibration rehearsal (klog#33609), we had ~11% inconsistency on the actuator efficiency ratio of PRM and BS from the past measurement (klog#30091).
We haven't done follow-up measurements for this issue yet. (We can resume it after coming back to MICH.)
Because our result also have inconsistency with the feedforward gain for MICH2PRCL (klog#33695), I checked past measurement to decide feedforward gain.
Then, I found a possibility that current MICH2PRCL gain is biased as 10-15% by the effect of a suppression by PRCL OLTF.
 

Details

According to the filter name of FM9 on K1:LSC-MICHFF2, feedforward gain is based on the measurement results on Dec. 26th, 2023 (klog#28094) and related measurement files are /users/Commissioning/data/MICH/2023/1226/TF_MICH2DARM_MICHFFOFF_20231226.xml for MICH2PRCL and /users/Commissioning/data/PRCL/2023/1226/TF_PRCL2DARM_MICHFFon_PRCLFFoff_20231226.xml for PRCLout2PRCL.

All MICH related test points are the out-of-loop channel of the PRCL loop. On the other hand, there is no out-of-loop channel at the downstream of the feedback point in the measurement file of PRCL. So the feed-forward gain seemed to be computed as
(PRCL_IN1/PRCL_OUT) / (PRCL_IN1/MICH_OUT) = A_prm / A_bs * (1 + G_prcl).

Though I tried to find measured OLTF around 2023/12/26, there was no measurement during 3 months before and after. So I tried to reproduce PRCL OLTF from error and feedback signals just after the measurement on that day. As shown in Fig.1, estimated UGF is ~13Hz. Open loop suppression by this OLTF is shown in Fig.2 and we can see suppression effect can be ignore above ~100Hz. On the other hand, measurements for FF were done with 50-200Hz. Blue, red, and green curves in Fig.3 represent a TF from PRCLout to PRCL, from MICHout to PRCL w/o G_prcl correction, and from MICHout to PRCL w/ G_prcl correction.

Actuator efficiency ratio by these TFs are also shown in Fig.4 and there is ~13% difference in the estimated results between without (~24.5) and with (~27.9) correcting G_prcl suppression. A result with correction of G_prcl suppression is consistent with the results of calibration measurement in klog#33609 (~27.2). And also, current FF gain in MICHFF2 (corresponds to ~25.7) is close to a result without a correction of G_prcl suppression. So current FF gain for MICH2PRCL may have a ~13% bias (though it might not be so serious from the view point of IFO lock).

Because an estimation in this post is based on the reproduced OLTF, it's difficult to believe values in this post. But this result suggest that it's better to validate a feedforward gain by multiple measurements. In addition for same measurements which will be done in future, G_prcl suppression will not be a negligible because PRCL UGF is increased in klog#33437 and klog#33627.

With Shingo Hido

We performed the integrating sphere calibration for Pcal-X.
The results are attached (PDF file).

Based on these results, we recalculated the relevant online parameters and updated them accordingly.
The updated EPICS channel values are summarized below (rounded to four significant figures):

Notes

• There were no significant changes in the RxPD or TxPD values.
• The background (BG) values were updated using only today's measurements. The previous values were based on averages over multiple past measurements. This update avoids the influence of older values that were set using different methods (see klog33465).
• We also started updating K1:CAL-PCAL_EX_2_INJ_V_GAIN, a parameter used to equalize the injected sine wave amplitudes from path 1 and path 2 at the ETM. This parameter had not been updated for a while.
• Prior to this Pcal-X calibration, we performed a beam position check using the Tcam. Starting from this update, we have begun reflecting the measured beam positions in the corresponding online parameters.

As a reference, we performed the same procedure on the IY cryopayload.
The result is shown in the attached figure.

The response ratio between the MN and TM in IY is approximately 1.2, compared to 1.6 in the IX case.
Additionally, the response in IY is quite linear and reproducible, and the response amplitude is significantly larger than that observed in IX.

[Yokozawa, Aso, Komori]

Abstract:

The point of contact may be located around the IM or TM, rather than the MN, based on the measurement that the DC response of the TM pitch is smaller than that of the MN.

Detail:

We attempted to identify the point of contact in the IX cryopayload.

First, we performed a rough centering of the beam spot on both the MN and TM QPDs by adjusting the moving mass on IX MN.
At that time, the oplev pitch signals in the unit of error function were approximately –0.3 and –0.2 for the MN and TM, respectively.
Since the MN yaw error signal was 0.997 and we do not have a system like the moving mass to adjust the yaw degree of freedom, we could not perform the same thing in yaw unless we relocate the MN QPD inside the tunnel.

Next, we applied several offsets to the MN optic alignment, ranging from –300 to +300 in steps of 100, and measured the resulting oplev values for both the MN and TM, as shown in the attached figure.
The MN response was consistently larger than the TM response, suggesting that the point of contact is around the IM or TM.

It should be noted that the responses were not linear (larger positive offsets resulted in larger DC responses), and the oplev values themselves were not reproducible.
However, the trend—namely, that the MN response was always larger—was reproducible.