At least from the camera footage, the two units on the shutter side appear to be attached to each other. On the parking-lot side, it is completely dark, so we cannot check them with the camera either.

Unlike the air conditioner in the data analysis building, the mounting surface in the KAGRA site computer room is flat and smooth, allowing the adhesive to make very good contact with the surface.

The air conditioner in the analysis building appears to have fallen because the mounting surface was neither smooth nor flat/curved, and the gel was attached using the adhesive surface that originally came with the gel material.

We will use sticker tape for them.

[Kimura, Yasui, H. Sawada and Hokushin's worker]
 From Sep. 28 through Sept. 29, we carried out replacement work on the air compressor installed in the corner near the Central Mirror Room BS.
After operational testing, the replaced compressor was turned off and is currently in standby mode.
　The connection to the compressed air supply piping and the operating procedures for the compressor will be determined following discussions during the regular scheduling meeting.
 Please refer to the attached photos for images of the work in progress and the exterior of the replaced air compressor.
An overhead crane and a lift truck were used during the replacement work.
 

Kimura, Yasui

We successfully switched the vacuum pumps in the center area.

It seems that there was no significant outgassing from IPs during this time.

In the Y-end area, we are currently working on switching #35 and #36 IPs.

We plan to continue with the remaining steps in the procedure from tomorrow onward.

NOTE:

We need to request the SAFE state not only for SR2 but also for SR3 when we stop (maybe we start) the #4 TMP.

Also, please do not forget to start the cooling water chiller when you start the #5~#8 TMPs.

I checked the TF measurement results for the health check and found strange behaviour.
Since there are so many results different from the reference, I just pick up several results.

Figure 1 shows the TF measurement results of BF GAS.
The coherence at entire frequency band is lower than before.
I'm not so sure the exact reason but it might imply that the BF GAS is rubbing somewhere.

Figure 2 shows the TF measurement results of IM OSRM V3.
Not only the overall gain but also the TF shape are different from the reference.
I'm also not so sure the reason but this results also imply that the suspension rubbing somewhere.

I checked the BF GAS behaviout when moving SF GAS (fig3) and no strange motion can be observed at BF stage.
This reveals that the suspension is not rubbing on the security flame.
So, we need to check if the suspension is internally rubbing or not.

I checked the TF measurement results for the health check (fig1 - fig33).
All TF seems healthy.
Athough resonant frequency of SF GAS TF measurement was shifted from the reference, this can be already seen in the previous health check results.
So, this result may not be so problematic

I checked the TF measurement results for the health check (fig1 - fig33).
Although the IM OSEM H1 and V2 have smaller gains because the OSEM positions are close to the edge of the linear range, this does not seem to be due to the power outage.
As the OSEM positions have not been changed since the last vacuum evacuation, these results may be acceptable.
Other than that the suspension seems healthy.

The health check after recovery from the power outage was performed. The data for the IP TFs were not recorded.

• The gains of the TFs in IM H1, H2, H3, V1, V2, V3, L, R, T and Y were a little (< 5 dB) smaller than the references.
• The gains of the TFs in TM P and Y were larger than the references (~10dB).
   

The health check after recovery from the power outage was performed. Though the gains of the TFs in IM H2, H3, V2, V3, P, R, T, V, and Y were a little (< 5 dB) smaller than the references, the measured TFs look healthy.

The health check after recovery from the power outage was performed. The data for the IP TFs were not recorded. Though the gains of the TFs in IM were a little (< 5 dB) smaller than the references, the measured TFs look healthy.

Hirose, Ushiba(remote), Miyakawa

1. IMC ASC

The PZT for the IMC ASC was saturated. It is sometimes possible to avoid ASC saturation by changing the IMC LSC offset, but this time the required correction was outside the available range. We entered the PSL room and offloaded the PZT.

The pitch of PZT1 was saturated at +75 V. Hirose-san turned the adjustment screw clockwise and, with the control-loop gain reduced so that the feedback signal was almost zero, adjusted it until the error signal was close to zero.

After restoring the gain and closing the feedback loop again, the output signal to the PZT became close to 0 V, and the ASC control started working properly.

2. No light visible from the FNC

Since the green light was not visible on the POP or POS tables, we checked the green Fiber Noise Cancellation (FNC) paths on the PSL table.

The behavior was similar for both X and Y. In both cases, when we blocked the path to the fiber in one arm of the Michelson interferometer, no returning light was observed from that arm. The light from the other, short arm was returning, but its power was too low to be detected by the Michelson PD.

In fact, the optical power just before the Michelson BS itself appeared to be quite low. Sufficient optical power was present farther upstream, closer to the light source.

Since the same behavior was observed for both X and Y, the cause was expected to be something common to both paths. Ushiba-kun noticed that there might be a problem with the AOM, so we left the PSL room and checked the VCO signal generator for the AOM at the very bottom of the ALS2 rack.

Actually, the signal generator settings had been reset due to the power outage. After setting it to 70.00 MHz and 13.0 dBm and turning on the RF output, light became visible on the Michelson PD, and the FNC could also be locked successfully.

The health check after recovery from the power outage was performed. The measured transfer functions look healthy.

I changed the setpoint for the F1 GAS from -3700 to -3460, so that the MN V Oplev can be zero.

I offloaded the F1 and F3 GAS filters with the FRs.

I changed the setpoint for the BF GAS from 3490 to 3860 so that the MN V Oplev can be zero. The BF GAS was offloaded with the FR.

I changed the setpoint for the F1 GAS from -3150 to -2200 so that the MN V Oplev can be zero. The F1 GAS was offloaded with the FR.

Though I tried to offload the F0 GAS with the FR, the FR didn't work well. So I changed the setpoint for the F0 GAS from 4527 to 4250 temporarily to avoid saturation.

I offloaded the F0 and BF GAS filters with the FRs.

PLL can be locked by the guardians without any problems after recovering IMC.

M.Takahashi, Yasui

We tried to switch vacuum pumps from TMP to IP at the center area, following the procedure prepared by Uchiyama-san.

14:40 requested the SAFE state to the PR2

14:49 started #2 TMP 

14:52 opened GVifi

14:56 started #1 IP

15:02 opented GVsrm

15:04 started #14 IP

15:07 connected to the #2 TMP and requested the LOCK_ACUISITION state to the PR2

15:14 started #13 IP

15:15 started #3 IP

15:21 started #8 IP

15:24 started #7 IP

The pressure values in each T-tube seem to be low enough, but we haven't connected these IPs to the vacuum duct yet, in case there is still some outgassing from the IPs. So we will continue pumping down overnight with TMP in parallel.

We plan to continue the work tomorrow, following the remaining steps in the procedure.

Following the procedure described in klog:35425, I set the frequencies and amplitudes of the LOs for IMC, PLLX, PLLY, Gr X PHD, and Gr Y PHD around the IOO rack to the specified values, and turned on their RF outputs.

• Only the PLLY LO had kept its settings from before the power outage. All the other signal generators had been reset when the power was lost.
• The PLLX LO had been set to remote-control mode, so I first disconnected the LAN cable, switched it to the local mode, configured the settings, and then reconnected the LAN cable.

I then entered the PSL room and checked each circuit.

• Most of the electronics automatically returned to their previous states after the power was restored, and there did not seem to be any particular problem. I also confirmed that the HV amp was actually providing approximately +/-191 V.
• I monitored the output of the SR560 just before the signal goes to the laser PZT with an oscilloscope. I confirmed that some signal appeared when the common-mode servo attempted to acquire lock.
• I also monitored the output of the HV amp just before the signal goes to the wide-band EOM with an oscilloscope. Again, I confirmed that some signal appeared when the common-mode servo attempted to acquire lock.

Since the IMC still did not lock properly, I left the PSL room and checked the signal generators at the IOO rack again. I then realized that I had turned on the LO for the main interferometer instead of the one for the IMC. After correctly configuring and turning on the IMC LO, the IMC locked successfully.

In the end, simply looking at the feedback signals does not allow us to distinguish whether they are the correct signals or spurious ones. If necessary, it would be better to check the error signals as well.

[Miyakawa, Ushiba (remote)]

Abstract:

We found that the signal generators for IMC LO (13.77 MHz, 7.4dBm) was OFF.
After turning on it, IMC can be locked.
However, IP1 PIT output was saturated, so it might be necessary to offload it at some point.

Detail:

Miyakawa-san checked the circuits related to the IMC lock (detail will be posted later by Miyakawa-san) and found that signal generators for IMC LO was turned off.
After recovering the LO, we tried to lock IMC and it can be locked without any problems.

After that, IMC ASC was engaged and IO guardian reaches PROVIDING_STABLE_LIGHT state.
However, one of the PZTs (IP1 PIT) is saturating.
Since all workers has already been out from the mine, we keep it as it is.
So, it might be necessary to offload it at some points.

The health check after recovery from the power outage was performed.

• The gains of the TFs in IM L and T were smaller than the references (< 10 dB).
• The gain of the TF in IM V was much smaller than the reference (> 10 dB).
• The TFs in BF V1, V2, and V3 were different from the references.
• The TF in BF V was different from the reference.
• The TFs in F0, F1, and BF GAS were different from the references. The platform may be touching the stoppers vertically.
• The F3 ACT is not working.
• The TFs in IP ACCL and ACCT were different from the references.

The health check after recovery from the power outage was performed.

• The gains of the TFs in IM L, P, R, T, V, and Y were much smaller than the references (> 10 dB).
• The gains of the TFs in MN H1, H2, H3, V1, V2, and V3 were larger than the references (< 10 dB).
• The gains of the TFs in MN L, R, T, and V were much smaller than the references (> 10 dB).
• The TFs in BF V1, V2, and V3 were different from the references.
• The TF in BF V was different from the reference.
• The TFs in F0, F1, F3, and BF GAS were different from the references. The platform may be touching the stoppers vertically.
• The F2 LVDT-ACT is not working.
• The TFs in IP ACCL and ACCT were different from the reference.