The V1 I/O chassis on which HIB adapter card is still attached is stored at U8-11 of C2 rack. It will be kept as a spare chassis or be used on the test stand at the SK building. The V3 server on which IRIG-B and HIB host card are still attached is stored at U10-11 of the P7 rack. It may be used to upgrade V0 and V2 front-end servers (PX1, MCF, and IY0) if a number of the V4 servers becomes insufficient.
I measured the spectra of the FLDACCs in displacement (um) compared with the ACCs (L-4C). Plot 1 shows the spectra in READY mode; the IP was free. Plot 2 shows the spectra in ISOLATED mode; the IP was controlled by the LVDTs. The noise level of the ACC1 is larger than that of the other ACCs.
[YamaT, Nakagaki, Ikeda]
This work is related to K-Log#36654.
We used the new V2 IO chassis to measure the noise levels of the ADC and DAC.
For the DAC measurement, the Whitening Chassis used the S1402923 port (on the Board2 side).
Since the input side of this Whitening Chassis is a D-Sub 15-pin female connector, an adapter was required for connection.
During DAC measurements, the Whitening Chassis provided a 42 dB gain.
[Data]
/users/DGS/measurements/{ADC,DAC}/K1EX0/2026/0326_V2_IO_CHASSESS/
(Work on 25th, detail report)
[Takahashi.R, Dan.Chen, Hirata]
This is the detail report of as Takahashi-san's report in klog36657.
>2.Checked the height of the TM/RM
The center of the mirror was 2mm higher than the target on the SRM chamber.
1.We set up two laser levelers: green one on the +X side and red one on the -X side.
2.The green laser leveler was aligned with the mirror height target on the SRM chamber(pic1). The scratch line on the +X side is 1 mm above the green laser level line(pic2).
3.The red laser leveler was set on the -X side. The scratch line on the -X side is 1 mm above the red laser level line(pic3).
4.The height difference between green laser leveler and red laser leveler was 2 mm.(pic4,5)
Conclusion: The +X scratch line was 1mm higher than target, and the -X scratch line was 3mm higher than the target. Therefore, the center of the mirror was 2mm higher than the target on the SRM chamber.
>3.Released the breadboard and checked the height.
The result was almost the same as previous measurement.
1.Two 5 kg masses were left in the position where the mid-size baffle had been installed. (Mid-size baffle were 4.94kg and 4.931kg. see klog36531) Those masses were fixed by M6 studbolts and nuts.(pic6)
2.We checked the height of bread board using laser leveler aligned with the target on the SRM chamber. The measurement point was approximately the same as in the previous measurement.klog36531
3.The result are as follows.
・+X side (Green line) : 383.5mm.(pic7)
・-X side (Red line) : 380.0mm.(pic8)
*Note: The previous measurement values reported in klog36531 were reversed. Sorry for the mistake. Please check the pic7 and pic8 in klog3631)
4.We confirmed that nothing was blocking the oplev beam.(pic9,10)
>5.Checked the cables at the feedthrough flanges.
We checked in-vacuum cable connection. The result are attached.
We brought NCal Pylons back from pre-center room to Xend parking area.
We will bring them to Xend (Not clean room) on 4/6.
We turned ON the Pcal-X laser source and requested GRD to be LOW POWER state.
There was no problem.
I measured the transfer function from the H2+H3 actuator for the IRM dampre to the Oplev yaw "IRM_OLDAMP_Y" in higher resolution. The first mode frequency was 59mHz and the second mode frequency was 158mHz.
[Hirata, Dan, Takahashi]
This is a log of the work on the 25th. We performed the following tasks:
- Checked the FLDACC signals.
- Checked the height of the TM/RM.
- Released the breadboard and checked the height.
- Adjusted the gaps of the EQ stoppers and took the pictures.
- Checked the cables at the feedthrough flanges.
Details will be reported later.
[Nakagaki, Ikeda, YamaT]
This is similar work of ITMX (klog#36572) and ITMY (klog#36625).
A preliminary research about a situation of both end stations was reported in klog#36635.
I/O chassis of K1EX0 was replaced from V1 (S1604791) to V2 (S2416131).
The front-end computer was also replaced from V3 to V4, but it hasn't been moved from EX0 rack to EXV1 rack due to a miscommunication about laying MTP cables.
We cannot complete this work until purchasing some cables in April.
ADC/DAC noise measurement is planned to be done tomorrow.
-----
I/O chassis replacement
We replaced the V1 I/O chassis (S1604791) at U18-21 of EX0 rack as V2 I/O chassis (S2416131). In this time, all PCIe cards were moved from V1 I/O chassis to V2 I/O chassis in stead of preparing V2 I/O chassis in which attached spare PCIe cards. Though this procedure involves a risk of taking long time if we need to restore all due to some trouble with V2 I/O chassis, we determined that’s highly unlikely based on the past experiences. (Besides, there’s also the fact that it’s not realistic to prepare same number of spare cards as one of installed cards.) On the other hand, an advantage of this procedure is S-number of used ADC/DAC boards isn't changed.
Front-end replacement
In the original plan, the front-end computer was planned to be moved from the the EX0 rack (1F) to the EX1 rack (2F) as part of the upgrade from the V3 server to the V4 one. An combination of V2 I/O chassis and V3 server has a limitation about a number of attached PCIe cards and EX0 has many ADC/DAC boards. So replacing V2 I/O chassis at EX0 requires to upgrade also the front-end server. Moving front-end server from the EX0 rack to EX1 rack is the most important advantage of using V2 I/O chassis. Because of specification of the cables for V1 I/O chassis, we couldn't separate the servers from analog electronics in the 1st floor of the end stations. Another concern was the increasing number of daisy-chain segments and growing latency across various networks, including RFM.
For this reason, we tried to install a new V4 server in the EX1 rack at first. But a gender of MTP at the splice box was different from one in the discussion in advance. Because MTP pins are very thin and are easily broken, we designed the entire assembly so that the connector of cables which is wired by ourselves would be female. So MTP of breakout cable we have has female connector. The splice box at EX1 rack has unfortunately also female receptor and we couldn't connect them.
Finally, we gave up to install V4 server in the EX1 rack today and installed it at U13-14 in the EX0 rack after removing the V3 server. So the replacement of I/O chassis and front-end computer were finished but the upgrade task hasn't been completed yet. To proceed it, MTP breakout cables with male connector must be purchased. We can resume this task after April.
Side Note
Some SDF differences on the EX0 models remained.
k1tmsx
All differences are accepted in safe.snap (see also Fig.1)
k1calex
It comes from CAL_EX guardian was in FAULT and was paused. It's related to a shutdonw of laser itself (see also klog#36639). For such case, SAFE state should be defined as allowing to stop a laser output without stopping/pausing. Otherwise, any other maintenance cannot be down while the laser output is stopped. In this time, I manually removed all DAC output, then all values accommodated with the set point in safe.snap.
[YamaT, Nakagaki, Ikeda]
This work is related to K-Log#36654.
We used the new V2 IO chassis to measure the noise levels of the ADC and DAC.
For the DAC measurement, the Whitening Chassis used the S1402923 port (on the Board2 side).
Since the input side of this Whitening Chassis is a D-Sub 15-pin female connector, an adapter was required for connection.
During DAC measurements, the Whitening Chassis provided a 42 dB gain.
[Data]
/users/DGS/measurements/{ADC,DAC}/K1EX0/2026/0326_V2_IO_CHASSESS/
The V1 I/O chassis on which HIB adapter card is still attached is stored at U8-11 of C2 rack. It will be kept as a spare chassis or be used on the test stand at the SK building. The V3 server on which IRIG-B and HIB host card are still attached is stored at U10-11 of the P7 rack. It may be used to upgrade V0 and V2 front-end servers (PX1, MCF, and IY0) if a number of the V4 servers becomes insufficient.
[Kimura and Yasui]
On March 25, we restarted four duct-shield cryo-coolers called Yea and Yer in the Y-end mirro room at 14:00 as a test run.
And we turned on the cooling water unit for the cryo-coolers at the same time.
[Kimura and Yasui]
On March 25, we switched form TMP to the Ion pump of the #30 vacuum pump unit connected to the EYV .
After the activation of the #30 Ion pump, we turned off the #30 TMP, dry pump and an air conditioner.
Also, we switched form TMP to the Ion pump of the #35 vacuum pump unit connected to the beam duct in the 1st floor of Y-end mirror room.
We turned off the #38 TMP and #35 TMP with dry pumps.
Checking the temperature in the computer room, it has risen from the normal 22 to 25 degrees.
In addition, the discharge temperature of the No. 2 air conditioner is almost the same as the room temperature.
Hayakawa-san is now contacting the vendor to ask for the repair of the air conditioner.
The room temperature has been holding at around 25 degrees for the past 12 hours.
I moved a script to read the pcal laser status and put them as EPICS records from k1script1 to k1script0.
-----
A old script on k1script1 submitted commands to the laser control server via SSH. Then the laser server wrote a status to files on NFS and another script on k1script1 read those files to put the EPICS records. This implementation required many communication between the laser server and DGS servers and it's hard to check the command status because command execution via SSH returns a status code of SSH itself not a executed command in normal.
So I rewrote these script to get a status of command execution. Now the PcalLaserListener.py as the server application on the laser control server to wait requests from k1script[01] and PcalLaser.py as the client application on the DGS servers is launched on k1script[01]. As this change, command status (communication errror between laser itself) can become handled properly and the migration work for CAL related scripts on k1script was completed.
[Hirata, Dan, Takahashi]
After installing the new mirror, we released the suspension and the IP in the following order: TM, RM, IM, IRM, BF, F0, and IP. Then we tried to align the TM using the Oplev. Initially, the beam spot of the Oplev was at the upper right. We adjusted the horizontal position with the F0 yaw FR and the vertical position with the IM picomotor (moving mass). The TM was significantly deviated in the pitch direction (maybe due to the lighter new mirror), so 130,000 steps were necessary for the picomotor. We could find the beam spot on the QPD (Picture 1).
We closed the side hatch and the top chamber. The transfer function from the H2+H3 actuator for the IRM dampre to the Oplev yaw "IRM_OLDAMP_Y" was measured (Picture 2). Two modes at 60mHz and 160mHz can be confirmed.
More pictures: https://dac.icrr.u-tokyo.ac.jp/KAGRA/DAWG/CAL/pic/MIF/260324
As posted at the issue, the regular production of segment files at k1det1 has stopped between 3/21 UTC and 3/22 UTC. As I checked the log file, the time of stopping the segment production seems to be 2026-03-22 02:15:00 JST ~ 2026-03-22 02:30:00 JST. After that, the process to generate segment files restarted from 2026-03-22 14:45:00 JST.
This is coincident with the occurance time of k1nfs0 issue as reported in klog.
Since the interferometer was down at the time, I don't think many people would be interested in the segment information for this missing period. If that's the case, then ignoring this missing period is also an option.
[Kimura and Yasui]
On March 24, we restarted four duct-shield cryo-coolers called Xfa, Xfs, Yfa and Yfs in the center mirro room at 15:00 as a test run.
And we turned on the cooling water unit for the cryo-coolers at the same time,
[Kimura, Yasui, M. Takahashi and H. Sawada]
On March 24, as part of maintenance work on the cryogenic cooling units, we removed four valve units from the radiation shield cryo-coolers (IXC P-53, IXC P-55, IYC P-53 and IYC P-55).
The removed valve units were packaged and returned to the manufacturing plant, where they will be disassembled and inspected.
[Takahashi.R, Dan.Chen, Hirata]
We installed new SRM mirror.
・The First Contact on the AR side was peeled off before installation.
・Confirmed that the wedge mark was on the -X side.
・Installed the new SRM mirror and tightened the M4 screws to 1.5 Nm.
・Peeled off the First Contact on the HR side.
・Inspected the surface and found some dots near the center. (pic1)
・We removed the mirror and applied First Contact again. (pic2)
・We reinstalled the mirror and peeled off the First Contact again. The second trial was successful, and the dots near the center were removed. (pic3)
・Finally, we confirmed that the wedge mark was on the -X side again. (pic4)
We removed unused power supply units (PAN35-10A) from U1-4 of SRM rack.
These units were used as a pair of +/-18V supply before O3. They will be used for +24V supply for IO chassis at the both end stations.
We have now all essential stuffs to replace the IO chassis at the 1st floor of the both end stations.
Replacement for EX0 can be started soon.
For EY0, we are still discussing how to transport these stuffs around EY0 rack through EYA booth.
If mine entrance at Mozumi is available, it's better to transport stuffs also for EY1 (the 2nd floor) at the same time.
[Nakagaki, YamaT (Support), Ikeda]
This work is related to K-Log#36625.
We used the new V2 IO chassis to measure the noise levels of the ADC and DAC.
For the DAC measurement, the Whitening Chassis used an available port (on the Board2 side) from S1909738.
During DAC measurements, the Whitening level was set to 42 dB.
[Data]
/users/DGS/measurements/{ADC,DAC}/K1IY1/2026/0324_V2_IO_CHASSIS/
Unnecessary cables were also cleaned up.
Though Timing Fanout chassis at U38, IRIG-B chassis at U36-37, DAQ switch at U33 and KVM at U17 were also no longer necessary, they are still in ICV1 rack. They can be moved required location in anytime. Timing fanout can be used to fill a shortage of timing SFP ports in the server room. IRIG-B chassis is a new production lot, so replacing old production lot makes a work for planned power outage easy.
With Shingo
We requested Pcal-X GRD to be SAFE, then stopped the GRD.
Finnally turned OFF the Pcal-X laser source.
We turned ON the Pcal-X laser source and requested GRD to be LOW POWER state.
There was no problem.