We found small metal pieces in the water vessel of the chiller. These metal pieces might came from inside the beam dumper as remnants after making a water path by milling.

The tube connection information is https://klog.icrr.u-tokyo.ac.jp/osl/?r=22249.

Precisely speaking, Chiller output -> Yellow tube -> Shutter BD -> Transparent tube -> PMC REFL BD -> Yellow tube -> Adjuster BD -> transparent tube -> Chiller return.

So, 

• remove the yellow tube at adjuster BD, then check the water flow.
  • If OK, the adjuster BD and its downstream transparent tube have stuck.
  • If no flow, the stuck point is at upstream.
    • remove the transparent tube at PMC REFL BD
      • If OK, the PMC REFL BD and its downstream yellow tube have stuck.
      • If no, the stuck point is at upstream: Shutter BD and its upstream yellow/downstream transparent tube.

By the way, do you know why the water-cooling type beam dumper was selected? The air cooling type beam dumper for 50W and 150W is enough for our purpose.

[Tanaka, Yamamoto, Yuzurihara, (miyoki, uchiyama: remote)]

[Summary]

There seems to be a stuck in the water circulation path for the beam dumpers(BDs). So these beam dumpers were not cooled by water from the chiller. As a result, the temp at the beam dumper at the REFL increased, then the interlock was triggered by its thermometer.

[What we did]

Firstly, we checked the water chiller for the beam dumper. The water level was slightly lower (The top of the metal circulating tube inside the water vessel was slightly above the water level). The chiller itself was on, and the temp setting was 15C(OK). At this moment we did not check whether the water was circulated or not.

Secondly, we checked the interlock box near the entrance door of the PSL  room. We found the interlock was activated with a red LED light. Thirdly, we checked the sealing of the thermometer on the back of the beam dump. The contact of the thermometer seemed to be good.

Then we reset the interlock successfully. After that, we confirmed that the warning message in the FB control application has not vanished and the temp shows ~50C. Then we closed the application and restarted it. The warning has vanished. However, the temp showed ~3C and it became ~ 50C again after pushing the enable button in the application. So, we closed the application again, and made off/on procedure for the FB laser amp itself. After starting the application, the temp showed ~20C and we could keep this temp after enabling the FB laser.

We increased the current to obtain 10W output.

After that, we started the investigation again on the water chiller system for BDs. Tanaka-kun noticed that the air bubble in the water tube did not flow at all. Also he touched the REFL beam dumper and he sensed relatively hot. Then we suspected that there was no water flow in the chiller system. We tried on and off the chiller. Tanaka-kun noticed that the air bubble position shifted a little bit every on/off action and that the tube itself showed a little bit of trembling. However, we could not realize the smooth flow of the air bubble.  Then we checked whether there was returning water flow just before the chiller or not even if the chiller seemed to try to flow water (We sensed some vibration in the chiller itself after pushing the pumping enable button). Then we confirmed that no water returning.

Firstly, we suspected there was something wrong in the chiller itself. So we replaced the chiller with a new one. However, there was also no returning water from the path in the PSL room. To check the chiller itself, we replaced the tube for going from the chiller to the PSL room with a new one, and checked whether there was water output or not. Then we got water flow output. So we concluded that there seemed to be a stuck in the tube-path for BDs in the PSL room.

Because stuck search in the PSL room is a hard work, we gave up today's recovery work.

The FB laser amp was left to be stopped. The seed IR was on.

[to do]

To check the stuck in the tube path in the PSL room, we need to 

• change from OBS mode to Working mode for the PSL room for the long activities.
• remove all parts on the tube path including BDs not to spoil the PSL room with water
• check the stuck point in each part.

If we want to recover the FB at 20W output temporally, we can put air cooling type power meters (50W max and 150W max that we have) instead of these water cooling type beam dumpers. In this case, the interlock system will be temporarily disabled.

YamaT-san reported the laser down again around 19:00. 

I checked the chiller for the FB laser, and I confirmed that it was working. So, we suspected the same interlock trouble as the previous down. In addition, I suspected that,

• the contact trouble of the thermometer on the water cooling type beam dumper.
• interlock switch troubles
• less water condition for the cooling water circulator for the beam dumper.

According to Yuzu-san's information, the laser seemed to be healthy around 18:00 (JST).

Around 15:00??, YamaT-kun performed all DAQ restart.

Some electrical influence from the DGS system and the interlock system through the grounding might exist. However, in the previous down case, the laser down happened before the DGS maintenance activities.

Anyway, I asked Tanaka-kun, YamaT-san and Yuzu-san to enter the PSL room tomorrow, and to check the above 3 items, and to restart the laser again.

I measured noise spectra of the three error signals (REFL 45, 56, and 17) directly reflected from PRM to estimate the sensing noise limit of these channels, particularly that of CARM.
Ideally, the sensing noise is completely limited by the shot noise.

The attached figure shows the result with some light.
In addition, it is compared to that without light to measure the electrical dark noise.

At high frequencies, the dark noise remains a dominant factor, which should be the shot noise.
We need to increase the input power to the PDs.
There are large excesses at low frequencies compared to the dark noise.
It might be because the optics inside the cahmber around IFI are still in air, so I will perform the same measurement after vacuum pumping.
Another potential reason is the amplitude modulation caused by detuning of the RF sidebands.
I will check the spectra after the fine tuning again, as described in klog:29385.

Furthermore, even-number harmonics of 60 Hz are prominent in the spectra with light.
Since there are no harmonics in the dark noise measurement, it should be generated by the IMC control or unknown electrical coupling peculiar to the configuration with light.

After vacuum pumping, I will calculate the CARM sensing noise limit in the unit of Hz/Hz and the projection on the DARM sensitivity.

[Tanaka, Hirose, Komori]

Abstract:

We adjusted the common offset immediately after the input of the IMC LSC common mode servo (-5.0 ± 0.3 mV with 14 dB input gain) and the fundamental frequency to generate multiple RF sidebands (5.6243667(1) MHz) to keep the carrier and the RF sidebands just on resonance of IMC.
As the result, we achieved a precise estimation of the IMC length, accurate to 1 µm, L_IMC = 53.302438(1) m.

Details:

During the previous measurement in klog:29084, residual peaks at the modulation frequency of 1.023 kHz persisted in either I-phase or Q-phase demodulation signals even after tuning the RF sideband frequency.
We attribute this to an extra offset in the IMC length control, causing detuning of the carrier and either the upper or lower sideband.
To address this, we experimented with combinations of LSC offset tuning and sideband frequency adjustments.

The setup is the same as klog:29084.
Initially, we adjusted the sideband frequency to equalize the residual peak heights in both demodulation signals.
Subsequently, we fine-tuned the common offset of the IMC LSC common mode servo.
This method effectively reduced the height of both peaks simultaneously, although we have not understood yet an unexpected swap of the I-phase and Q-phase peaks with minor adjustments to the sideband frequency (approximately 10 Hz).
After several iterations of this procedure, the residual peaks nearly vanished (the red and blue lines), yielding the results described above.

However, one hour later, upon rechecking the peak height, we observed the reappearance of residual peaks (the magenta and cyan lines).
We were unable to eliminate these peaks by solely adjusting the sideband frequency and common offset, suggesting that both parameters had drifted during this hour.
The drift in IMC length may originate from the re-locking of the IMC, leading to differences in the locked point by a few um, and thermal expansion of the IMC mirrors and suspensions.
The offset drift may arise from electrical circuit and residual amplitude modulation due to slight mismatches in the input polarization to the EOM, caused by temperature drifts.
We must consider strategies to compensate for these drifts and assess their impact on interferometer sensitivity.

After the today's acceptance check for anticipated closing IFI-IMM-PRM next week, here uploaded some photos to show the ISS beam is passing through about the center of the relevant viewport window; the photos taken the last week.

[Komori, Tanaka, YokozaWashimi]

We performed the Hammering test for the IFI stack (+X, +Y side stack).

• optical table -> optical table
• top stack -> optical table
• middle stack -> optical table
• bottom stack -> optical table
• base plate -> optical table
• ground -> optical table
• nothing -> optical table (as a reference)
• base plate -> base plate
• ground (leg) -> base plate
• ground (leg) -> ground
• ground (concrete) -> ground

I just noticed that I made a mistake in the number of outputs of the Whitining Filter in the mini-rack. The number of Whitining Filter outputs on the mini-rack is 8, not 4.
kTanaka-san just checked the empty input ports of ADC in IOO0, IOO1 rack.
The following are the empty input ports.

• IOO0 rack, ADC0: 8-19ch (FIG1)
• IOO0 rack, ADC1: 24-31ch(FIG2)
• IOO0 rack, ADC2: 4-11ch 20-27ch (FIG3, FIG4)
• IOO0 rack, ADC3: no the empty input port(FIG5)
• IOO1 rack: no the empty input port(FIG6)

For the mini-rack, I would like to use 4-11ch and 20-27ch in ADC2 and additionally 8-15ch in ADC0 and 24-31ch in ADC1.
I will check with others to confirm that we can use this port as an addition. Also, do additional cabling.

 Komori, Tanaka

We confirmed that the beam positon of the reflection from PRM which is in the "MISALIGNED_BF" state seems to be almost the center of the HP beam dump. (see attached movie) 

## what we did

• When IMMT1 and 2 were the ALIGNED state and PRM was the MISALIGNED_BF state, we locked IMC with ASC.
• Then we stopped the ASC by inputting 0 in the IMC ASC input gain.
• We confirmed the beam position on the HP beam dump and found it is almost the center of the dump.  

[Ikeda, Takahashi]

We checked the OSEMs again. We took pictures of the OSEMs with a 360º camera (THETA). The flap for the OSEM#1, #4, and #5 were rotated by 40~50°.

[Hirata, Dan Chen-san]

We recovered PR3 suspension. Oplev position is around center, and IM V1 OSEM value (K1:VIS-PR3_IM_OSEMINF_V1_INMON) is about 6200.

1. Firstly we moved IM pitch and yaw by IM Pico, and recovered Oplev position to 0. We moved pitch -23000, yaw -700.
2. IM V1 OSEM was still around 9900. So we started to adjust IM V1 OSEM.
3. Locked suspended breadboard.
4. Locked IRM.
5. Adjusted IM V1 OSEM. 
6. Released IRM, and confirmed IM V1 OSEM value(K1:VIS-PR3_IM_OSEMINF_V1_INMON). It is about 6200.
7. Checked that Oplev is around center.
8. Released suspended breadboard.
9. Closed chamber, and confirmed Guardian can go to "aligned" state(Pic1). After that, we changed it to "Safe".

posted by Miyoki instead of Kimura-san for the past activities on 24th April.

The gasket surface of the removed IFI flange was visually inspected. Based on the visual inspection results, the following three points are estimated to have caused this flange to leak.
1. scratches on the gasket surface of the flat flange with copper flange (thin vertical scratches can be observed when shining a light on it) One location
2. Scratches on the metal gasket surface (thin vertical scratches can be observed when illuminated by light) One scratch
Possibly traces of 1.
Uneven traces on the gasket seal surface If the gasket is tightened properly, the trace will be a circle of uniform width. Since the traces on the actual product are narrow and wide, there is a high possibility that the initial tightening of the claw clamps was not uniform.

Here are the countermeasures.
Since the instructions call for the flat flange with copper flange to be returned to the original mirror plate instead of using the copper flange, repair of the scratches on the flat flange gasket surface will not be performed. Instead, a visual inspection of the original mirror plate gasket surface is performed before installation. In addition, the gasket will be changed from a metal gasket to an elastomer gasket.

We connected the new cables to the photosensors and confirmed that the sensor is working.
Also, we checked the signals while touching the cables around BF but there seems no glitch.
So, new cables seems working well.

We will tie up the cables and fix them onto the payload.

[Hirata, Dan Chen-san]

We took photos for PRM payload earthquake stops. I uploaded today's photo to KAGRA dropbox.

During this work, we found two earthquake stops for test mass AR side were very close to the mirror surface(within 1mm?). We talked with Takahashi-san and Ushiba-san, and decided to withdraw them.

I checked the motion of the F0Y FR. I operated the stepper motor from -4563 step by 1000 steps. The BF Y signal didn't change until -16563 step. When I added one more -1000 step, the signal was changed from -770 to -850. Though the signal went back to -770 by 1000 steps, it didn't go in the plus direction anymore. The motor is working, but the motion of the wire receptacle on the bearing is not smooth due to the large friction.

[Ushiba, Tamaki, Komori, Takahashi]

We continued the photosensor recovery. The replaced cables were fixed onto the bottom of BF, the cable anchor (small hexagon), and the suspension rod with cable ties.

sorry, the legend was clipped.

It is same for all plots.

I plotted the high-resolution (3-hour data, 128s FFT) ASDs and Coherences for the geophones and ACCs on the OMC chamber, for the night time before vacuum breaking.