[Kimura and Yasui]
We changed and confirmed the operating threshold of the alarm contacts within CC-10s for ithe ineterlocks as follows;
GVbsx & GVbsy :8.0 x 10^-6 Pa
GVitmx: 5.9 x 10^-5 Pa
GVitmy: 5.9 x 10^-5 Pa
GVetmy: 5.9 x 10^-5 Pa
As reported K-log(32354), the alarm contacts are designed to operate when the threshold is exceeded.
The threshold of GVetmx will change in this afternoon.
{Kimura and Yasui]
We opened gate valve called GVetmy near EYA at around 10:30 AM.
And we changed the operating threshold of the alarm contacts within CC-10 to 5.9x10^-5 Pa.
As reported K-log(32354), the alarm contacts are designed to operate when the threshold is exceeded.
We set the threshold at 0.1 lower than the previously agreed value of 6.0x10^-5 Pa.
It has been confirmed that GVetmy closes when the operating threshold is exceeded.
Nakagaki-san reported that,
So, no closing at 12.x10^-5 Pa that is the setting value was no problem.
However, I asked Kimura-san to set new trigger values for GV closing at 5.9*10^-5 Pa.
Yamamoto kun reported via slack,
Fig.3 shows the X-30 pressure. I and Uchiyama-kun also confirmed that the GV_eya was open. Anyway Uchiyama-kun closed EYA GV by using Switchbot manually, although Uchiyama-kun wondered whether it should be closed or not because this enhancement should be the increase of nitrogen that is harmless for a mirror.
I checked the vacuum level and cryocooler status for EY as Fig.1 and Fig.2. Actually, the vacuum level at EYA CC-10(EYC side) was enhanced, but it started decreasing as in Fig.2. The temperatures also started decreasing as in Fig.1.
I would like to propose,
because,
Nakagaki-san reported that,
So, no closing at 12.x10^-5 Pa that is the setting value was no problem.
However, I asked Kimura-san to set new trigger values for GV closing at 5.9*10^-5 Pa.
{Kimura and Yasui]
We opened gate valve called GVetmy near EYA at around 10:30 AM.
And we changed the operating threshold of the alarm contacts within CC-10 to 5.9x10^-5 Pa.
As reported K-log(32354), the alarm contacts are designed to operate when the threshold is exceeded.
We set the threshold at 0.1 lower than the previously agreed value of 6.0x10^-5 Pa.
It has been confirmed that GVetmy closes when the operating threshold is exceeded.
This morning, Yokozawa-san and Yuzurihara-san asked me to investigate the cause of the IMC lock loss because whole IFO lock loss seemed to be caused by IMC lock loss when seismic motion around 1Hz became larger during O4a. So I started the investigation.
As the first step, from the summry page of lock losses during O4a, I selected one event labeled "1 Hz" at 22:32:42 UTC on June 12 2023 and saw some channels related to IMC (K1:IMC-CAV_TRANS_OUT_DQ (IMC trans. power), K1:LAS-POW_PMC_OUT_DQ (PMC trans. power), K1:IMC-SERVO_SLOW_DAQ_OUT_DQ (PZT feedback signal), K1:LSC-MCL_OUT_DQ(MCE feedback signal)) and the states of the LSC_LOCK and IO guardians with ndscope.
I found that IMC and PMC trans. power oscillated at high frequency several seconds before the LSC_LOCK guardain went to the DOWN state (fig.1). This indicates that ISS lock lost at that moment. After the ISS lock loss, the BPCs seemed to be broken due to becasue the BPC error signals varied in the intensity noise and ITMs moved a lot. Then, the whole IFO fell down.
I chosed other 3 events (07:32:21 UTC June 15 2023 (fig.2), 05:58:22 UTC May 30 2023(fig.3), and 11:46:07 UTC June 09 2023(fig.4)) randamly and observed the same trend. Therefore, it is likely that ISS lock loss is the cause of whole IFO lock loss when 1 Hz seismic motion became larger.
However, I'm not sure of the reason why ISS lock was lost when 1 Hz seismic motion became larger for now. We need more investigation.
Yokozawa, Ushiba, Tanaka
related to klog31986 (REFL), klog31992 (MCE TRANS), klog32330 (IMC REFL)
The optical layout on the AS table is here. Optics name convention as below is followed as the ones in this layout. Working photo are saved in the KAGRA dropbox.
## Abstract
We checked the beam path (OMC {TRANS, REFL}, AS) on the AS table when X-arm was locked. The beam position of almost all of the optics seem to be fine in the OMC TRANS and AS path, except for the HWP3 in the AS path. We shifted the HWP3 and tweaked the mirrors in front of PDs and QPDs for the centering. We found that there are some ghosts around the beam dump and HWP1. We should dump them after the new dumpers are delivered.
On the other hand, the beam position on the periscope upper mirror in OMC REFL path seems to be shifted from the center and clipped. So we need to lower the position of this mirror. However, we have no time today. We will do it later.
## What we did
We checked the beam path (OMC {TRANS, REFL}, AS) on the AS table when X-arm was locked.
### OMC TRANS path
First, we checked the beam in the OMC TRANS path. We found that the beam power in this path was too low to see the beam postion on the optics with the IR camera. So we used a sensor card to check the beam position. The beam positions seem to be fine.
### AS path
Next, we checked the beam in the AS path. The beam position on HWP3 was shifted in the -Y direction. So we shifted HWP3 to the -Y direction. The beam postions on the other optics until the QPD path seems to be almost center. The beam positions on the mirrors in the QPD path seemed to be off center slightly. Therefore, we tweaked STM4 and ST12 for QPD2 centering and BS3 and ST3 for QPD1 centering.
After that, we tweaked the steering mirrors to center RFPDs and the AS camera.
Also, we found that there are some ghost beams around the beam dumper (fig.1) and HWP1 (fig.2). We should dump them after the new dumpers are delivered.
### OMC REFL path
Finally, we started to check the OMC REFL path. We turned off the ceiling light above the OMC booth and we took the photo of the periscope upper mirror in the path (Fig. 3). The beam position seems to be shifted from the center and clipped. So we need to lower the position of this mirror. However, we have no time today. We will do it later.
-----------------------
## Note
We left the SD card of the camera in front of the k1ctr9 PC. Do not forget to bring the card before using the IR camera.
[Uchikata, Narikawa, YamaT, Yuzurihara]
I have two reports related to the segment files.
(1) Last year, we missed the automatic generation of the segment files due to the failure of the cache file generation (it's due to the expire of the GPS package). The period is 2024/06/28~07/18. We regenerated segment files at System-B (Kashiwa cluster) and transferred them to Kamioka.
The regenerated files were uploaded to DQSEGDB with the huge help of the administrator, Robert. The regular upload to DQSEGDB has resumed since 2025/01/09.
(2) During the manual regeneration, we used a different conda environment in the Kashiwa cluster from the usual environment in Kamioka. Actually, the package version of `python-logo-lw` is different between them, and it caused the different format segments. This time, we regenerated the segment files in the conda environment, including the consistent packages.
For the countermeasure next time, I'm updating the script (/users/DET/tools/Segments/Script/make24hourSegment_igwn.py) to add the option to read the data via k1nds2.
This script is different from the script running every 15 minutes. So it will not affect the regular generation of the segment files.
I'm testing the script for several dates to validate the script. It took 5 minutes ~ 30 minutes to finish reading the one-day data vis k1nds2. We guess the possible cause is the overload of the HDD in the Kashiwa cluster.
[Kimura and Yasui]
As the temperature of EYC's 8K shield increased, nitrogen trapped on the shield surface was thought to have been released.
As a result, the pressure near Y-30 increased and exceeded the threshold value, causing the interlock device for GVetmx to operate and GVetmy to close.
After confirming that the gas released by Q-mass was nitrogen, the operating threshold of the GVetmx interlock device was changed to 1.2 x 10^-5 Pa.
The operating threshold before the change was 9.6 x 10^-6 Pa.
[Kimura and Yasui]
At around 15:00, after opening GVetmx, the interlock device for GVetmx was re-enabled.
The pressure gauge for the interlock device installed in X-30 indicated 3.1 x 10^-6 Pa with GVetmx closed and 6.1 x 10^-6 Pa after GVetmx was opened.
The threshold for the interlock device to operate was set at 8.0 x 10^-6 Pa.
With Misato Onishi
We conducted the integrating sphere calibration for Pcal-X.
The results are attached.
There were no unexpected findings.
We got the following message in JST,
We checked the cryocooler status as IY. Then we found the temp enhancement at 4K stages and reduction at 50K stages that were regarded as the unbalance cooling of cryocoolers as in the attached Fig.1. The temp of 8K shield once reached around 30.6K that is near the nitrogen adsorption temp and went up. So we can expect the nitrogen release as in IY.
By the way, these same phenomena (Temps onee reached around 30K and went up in IY and EY) are interesting. The nitrogen adsorption with thermal flow/exchange might affect the performance of cryocoolers.
For your information.
There are several types of beam dump holders:
The second ones would be more popular now in KAGRA, probably, but the first ones were used since TAMA (and also in LIGO? by Arai-san). We have manufactured some of the first one and they are stocked in the KAGRA tunnel anywhere. This might be smaller in size than that of the second one, so suitable to be installed in the narrow areas.
PRFPMI_RF_LOCKED was not yet recovered.
LSC seems to work but ASC does not.
It is probably due to the bad beam spot on ITMX, so fine initial alignment is necessary.
I performed initial alignment (thogh ITMX and ETMX was not moved), and tried to lock PRFPMI.
Though PRFPMI LSC was succeeded, maximum transmission power was about 100 for both arms, which implies that the beam spot on ITMX is not good.
After the success of LSC, I implemented ASC but lock was lost when PRMI ASC (WFS) was engaged.
It is probably due to the bad alignment to POP QPDs, so QPD centering would be also necessary if the beam centering to QPDs is still bad after fine adjustment of the beam spot on ITMX.
1. Perform initial alignment and recover the beam spot on ITMX as much as possible.
2. Perform QPD centering with PICO if necessary, especially for POP QPDs.
3. Search good actuator balances for ITMX BPC to keep the good ITMX beam spot.
The spectra of {MN,TM} Oplev signals in LOCK_ACQUISITION are shown in fig.1, blue lines are the spectra before the adjustment, and red lines are after the adjustement. As you can see, the MN TILT/LEN oplevs see the suspension motion in the TILT and LEN directions, respectively. Thanks to this, the 1 Hz peaks in TM PIT and YAW are well damped.
On the other hand, the peaks around 2 Hz in MN/TM oplevs can be seen. We may need to adjust the damping filters.
[Tanaka, Yokozawa, Ushiba]
We moved injection beam path of ETMX MN TILT OpLevs.
Also, we performed centering of MN OpLevs.
We checked the MN TILT OpLev beam at reciver side and found that the beam shape coming out from the viewports were already clipped (fig1).
So, we moved injection beam to recover the clipping, and then the QPD SUM values were also recovered.
We also checked the MN ROL OpLev beam and it didn't seem to be clipped, so we didn't move the injection beam for MN ROL OpLev.
After that, we centered all MN OpLevs (MN LEN, MN TILT, MN TRA, MN ROL, and MN VER).
