Thank you for measuring the surface of sapphire parts.

Is it possible to mask the edge of 1.5mm from the measurement results of HCB surface of sapphire ears?
Our specification of the flatness for HCB surface is 77*27 area of 80*30 surface, so it would be helpful to compare the result with the data from the company.

The monitor page (https://gwdet.icrr.u-tokyo.ac.jp/~controls/slowmonitor/GAS/) maintained by Washimi-san might be helpful to find the saturation quickly. How about asking the daily (or weekly) check for the shift operator?

I checked GAS filter output of all suspensions and found that PR3 SF GAS and SRM F1 GAS is close to the saturation.
Figure 1 shows the GAS filter output at IOP models (left:PR3 SF, right:SRM F1).

In addition, ITMY GAS output is also close to the limit (output is more than 85% of ful range, fig2).
So, it would be better to offload them in the next maintainance day.

We turned off the OBS INTENT around 9:09 JST and on it around 10:49 JST for SR3 GAS offload work (klog34492).

I accepted the following difference of SDFs (see related klog34492)

I offloaded the F1 GAS with the standalone FR.

We also used 2 microscopes to have a look on the surface of the blade.

• Keyence VR3200
• Nikon NM-40

I attached the pictures taken by the microscopes.

What we saw

• Numerous linear scratches were observed on the surface. These are clearly visible even to the naked eye.
• The sloped edges around the hole are noticeably rough.
• The rim of the hole appears jagged

I just compared BNS and IR/GR XY trans Norm for the recent 22 days as Fig.1.

• GRX is almost constant except for a slight drop at some point, while GRY increased at some point and has remained almost constant. On the other hand, IRXY keeps decreasing. The relative alignment for IR and GR is now aparting with each other?

With Yoich Aso and Toshihiro Tsuzuki (NAOJ ATC)

We measured surface figures of a sapphire ear (made by K), a nail head (made by S) and blade spring (made by S).

Measured surface:

• Sapphire ear
  • Surface for HCB (figure 1, 2: upside down)
  • Surface for Ga bonding (figure 3)
• Sapphire nail head
  • Surface for Ga (A) (figure 4)
  • Surface for Ga (B) (figure 5)
  • (We can not distinguish A and B)
• Sapphire blade spring
  • Top surface (figure 6,7)
  • Bottom surface (figure 8,9)
  • (We could not distinguish top and bottom at the measurement time.)
  • (Side surfaces could not be measured)

I attached quick results.

F1 GAS of SR3 is now being saturated (previous saturation was reported in klog#34102).
TPs of 2kHz and 65kHz are shown in Fig.1.
Saturation is still caused only in high frequency due to the up-sampling process.

Figure 2 shows the 1-month trend of GAS output and temperature around SR3.
Slightly higher temperature (same as around 25 days ago) seems to be better for F1 GAS output.
Anyway, it's better to do the offload work (by height adjustment or temperature adjustment?) to remove DAC saturation.

IM temperatures during GPS time = 1435056520s - 1435096620s

IX, EX, IY, EY ~ 93, 44, 42, 58 [K]

Apart from the control bandwidth. Does PR3 oplev QPDs have IR filters set?? If no, it might be contamination from the main IR scattered light.

Yokozawa-san,

Since current PRC2 ASC signals are fedback to the setpoint of PR3 OpLevs, the effect of WFS signals through the feedback is smaller than that of OpLev signals.
So, if ASC signals contaminates DARM, it should be also seen in PR3 OpLev signals.

I don't think OpLev feedback affects at such high frequency but could you check the signals?

I used longer data to estimate the Q value of the violin modes.
This time I could set 4096s as the fft length.
Used data: GPS time = 1435056520s - 1435096620s

The results were obtained as follows:

• The fitting performed better than before, primarily because more data points around the peaks could be used for fitting.
• No peaks showed unreasonably large Q-values.
• Most of the peaks have Q-values around 10^5.
• A figure summarizing the fitting results for all peaks is attached. Peaks marked with an "X" indicate cases where the number of available data points was small or where the fitting clearly did not fit well.
• All fitting results have been uploaded to Dropbox: link

Operators name: Tamaki, Yasui

Shift time: 9:00-17:00 (JST)

Check items:

• CRY: The temperature of EX_REFBRT_HEAD was rapidly increased.  Consequently, the temperatures of REFBRT_BAR, BRT1 and BRT2 also started increasing, and BRT3 and BRT4 are expected to follow. 

  (According to the manual, we reported this issue to Ushiba-san and Kimura-san after the regular check at 10:00 and 16:00, because we observed a temperature increase of more than 10 K compared to the previous values.)

• VAC: The pressure values of TMSX and EXC slightly increased due to the temperature rise mentioned above, but this is not caused by H2O.
• TEMP: There was no issue.

IFO state (JST):

9:00 OBSERVING

12:50 LOCKLOSS due to the earthquake

13:16 OBSERVING

13:26 LOCKLOSS due to the earthquake

13:58 OBSERVING

14:06 LOCKLOSS due to the earthquake

15:13 OBSERVING

17:00 OBSERVING

[Shu-Wei Yeh, Chia-Jui Chou]

We applied ICA to clean KAGRA strain data from 1435106958 to 1435111758 (00:49–02:09 UTC, June 28, 2025) [Ref. klog 34400], using a cleaning duration of 1200 seconds. The segment was divided into four intervals for testing noise subtraction: 1435106958, 1435108158, 1435109358, and 1435110558. Here, we present the results for the interval starting at 1435108158.

The central frequencies (f_c), step-by-step improvements, and the overall inspiral range improvement are listed below. We observed effective noise subtraction in the June 28 data.

For the remaining denoising results and the list of witness channels used, please refer to the attached slides here,  JGW-G2516738-v2.

The next step is testing DeepClean's performance on the June 28 data.

Operators: Hido, Yasui

Shift time: 9:00-17:00 JST

Check Items: There was no issue on the regular check of VAC, CRY, and TEMP.

(Ikeda-san’s script for comparing cryocooler temperatures was very helpful!)

IFO state (JST):

9:00 OBSERVING

12:33 LOCKLOSS

(While recovering the IFO, there was an earthquake near the Tokara Islands around 12:50.)

13:45 OBSERVING

17:00 OBSERVING