[Kimura, Tomaru, M. Takahashi, H. Sawada and R. Takahashi]
"Inspection results for the sealing surface of the +Y side flange"
A vacuum leak in the range of 1×10⁻⁹ Pa·m³/s was detected, so we inspected the sealing surface of the side flange on the +Y side.
As a result, we found a slight scratch on the elastomer surface.
Based on the location of the scratch, we believe it is the cause of the vacuum leak.
After installing a new elastomer on the sealing surface of the +y side flange, the flange was closed.
The flange fastening bolts are currently in a temporarily secured state.
"Inspection results for the sealing surface of the SRM top flange"
After removing the SRM top flange, the condition of the flange sealing surface was inspected.
No dirt or scratches were found on the surface of the elastomer.
Therefore, it was decided to reuse the SRM top flange as a seal.
Date of work:
23, Apr, 2026
Workers:
Kimura and Tomaru
Abstract:
We pressurized SRM atmospheric pressure for the investigation of outgassing source in SRM (k-log 36811).
The following is a record of the opening and closing of valves and the pressure of the pressurization process.
The same process is used for k-log 26030. It is noted here for reference.
~9:50 GVoomt Close
Set safety valve to SRM vacuum pump unit
Turn on main SW of TPM (Not acceration, only raise the rotor blades to protect the TMP)
9:55 Pressurization was started by G-2class grade air (klog-25912). Cylinder is filled with 7 m³ of compressed dry air.
Pressure inside SRM was 1.7 x 10^1 Pa Cylinder pressure 15.2 MPaG
10:00 " 2.1 x 10^3 Pa
10:22 " 3.7 x 10^3 Pa
10:33 " 4.2 x 10^3 Pa Cylinder pressure 14.2 MPaG
10:40 " 7.8 x 10^3 Pa
10:55 " 1.6 x 10^4 Pa Cylinder pressure 11.9 MPaG
11:21 " 3.3 x 10^4 Pa " 8.9 MPaG
11:30 " 3.9 x 10^4 Pa " 7.9 MPaG Increased air flow (Secondary pressure 1.1 kg/cm^2G)
11:34 " 4.1 x 10^4 Pa " 7.5 MPaG
11:54 " 5.3 x 10^4 Pa
12:12 " 6.4 x 10^4 Pa
12:12 " 6.4 x 10^4 Pa
12:19 " 6.9 x 10^4 Pa
12:27 " 7.4 x 10^4 Pa Cylinder pressure 2.0 MPaG
12:34 " 7.8 x 10^4 Pa " 1.1 MPaG
Change to a new cylinder filled with 7 m³ of compressed dry air
12:41 " 7.9 x 10^4 Pa Cylinder pressure 15.2 MPaG
12:45 " 8.1 x 10^4 Pa " 15.1 MPaG
12:53 " 8.4 x 10^4 Pa
12:58 " 8.9 x 10^4 Pa
13:02 " 9.0 x 10^4 Pa " 13.2 MPaG
13:04 " 9.1 x 10^4 Pa
13:07 " 9.4 x 10^4 Pa " 12.6 MPaG
13:10 " 9.6 x 10^4 Pa
13:11 " 9.7 x 10^4 Pa
The pressurization process is terminated because the pressure in the SRM has reached almost atmospheric pressure at 13:11.
Based on the volume of dry air, the internal volume of the SRM was estimated to be approximately 9 m³.
mcrack, iya, ixc
They are AC power supply, but the no signals in display, I touched the AC tap, they are recovered.
Maybe after the blackout, the AC tap connected unstable
sr3, srm
After changed the battery, they recovered.
I checked the IM transfer functions after closing the chamber. The behavior did not change after opening the chamber to remove the geophone pods.
I checked the IM transfer functions after the pressure reached 8E4 Pa. The behavior returned to the situation before starting the evacuation.
[Washimi, Tomaru, Kimura, M.Takahashi, Sawada, R.Takahashi]
We opened the X+ side hatch and the top belljar. We could not find any dirty items inside the chamber. Pictures are here.
We removed the geophone pods. The weight of each pod was 12.6kg. After closing the chamber, we measured the IP transfer functions. The resonant frequencies became 0.16Hz for L and T, and 0.4Hz for Y.
KAGRA Pcal-Y updates (2026/04/23)
Workers: Dan Chen, Misato Onishi
We performed monthly Pcal-Y calibration on 2026/04/23.
After the calibration, we updated EPICS parameters related to the Pcal-Y system. No issues were found.
| EPICS Key | Before | After | Δ (After − Before) |
|---|---|---|---|
| K1:CAL-PCAL_EY_1_OE_R_SET | 0.98977 | 0.98970 | -0.00006 |
| K1:CAL-PCAL_EY_1_OE_T_SET | 0.98977 | 0.98970 | -0.00006 |
| K1:CAL-PCAL_EY_1_PD_BG_RX_V_SET | -0.00492 | -0.00468 | 0.00024 |
| K1:CAL-PCAL_EY_1_PD_BG_TX_V_SET | 0.01218 | 0.02250 | 0.01032 |
| K1:CAL-PCAL_EY_1_RX_V_R_SET | 0.50276 | 0.50286 | 0.00010 |
| K1:CAL-PCAL_EY_2_INJ_V_GAIN | 0.51836 | 0.52009 | 0.00173 |
| K1:CAL-PCAL_EY_2_OE_R_SET | 0.98609 | 0.98588 | -0.00020 |
| K1:CAL-PCAL_EY_2_OE_T_SET | 0.98609 | 0.98588 | -0.00020 |
| K1:CAL-PCAL_EY_2_PD_BG_TX_V_SET | 0.01466 | 0.02476 | 0.01011 |
| K1:CAL-PCAL_EY_2_RX_V_R_SET | 0.49724 | 0.49714 | -0.00010 |
| K1:CAL-PCAL_EY_WSK_PER_RX_SET | 1.84228 | 1.84385 | 0.00156 |
| K1:CAL-PCAL_EY_WSK_PER_TX1_SET | 0.33354 | 0.33366 | 0.00012 |
| K1:CAL-PCAL_EY_WSK_PER_TX2_SET | 0.90266 | 0.90553 | 0.00287 |
For the ETMY, actually, the mirror position changed, but smaller than the difference from Feb. to Apr. gif.1.
And oplev scattering image moved with the mirror position.
Thank to klog36803, the picos are available. So I adjusted the Pcal-Y beam position.
The TM positions on the Tcam photo as the ref for Pcal were also updated:
| [pixel] | Pcal-X horizontal | Pcal-X vertical | Pcal-Y horizontal | Pcal-Y vertical |
| Before | 1938 | 1420 | 2240 | 1035 |
| After | 1942 | 1410 | 2250 | 1000 |
The factor converting mm to dot is 11.3[pixel/mm]
Through this work:
the Pcal-Y path 1 was moved down by ~4 mm on ETM = by ~8 mm on RxPD, and
the Pcal-Y path 2 was moved down by ~2 mm on ETM = by ~4 mm on RxPD, and
the RxPD output recovered by ~1.5%.
[Kimura, M.Takahashi, H. Sawada, R. Takahashi and Tomaru]
On April 22, we continued conducting leak tests on the SRM vacuum vessel.
Since vacuum leaks were suspected due to cracks in the weld lines,
we performed leak tests by blowing helium gas onto the weld lines and using the hood method; however,we were unable to locate the leak. (See attached photo.)
The build-up test was conducted with the SRM and OMMT sharing a common vacuum, and the pressure was recorded using the OMMT’s pressure gauge.
The plot of the build-up test is attached.
The total duration of the build-up test was 935 minutes.
The estimated leak rate from the build-up test was 9.7 × 10⁻⁴ Pa·m³/s.
The slope of the pressure rise was approximately 1.
As reported in klog#36789, Pcal picomotors didn't work.
We found that it somehow occurs by the communication error between picomotors in the old PICO network and the Debian13 system.
Pcal picomotors are required for proceeding tomorrow's calibration of the integration sphere in YPcal.
So I downgraded k1script1 which is the secondary script server and processes for Pcal picomotors were moved back to it as the first-aid.
Another processes which works fine also on the Debian13 system are still on k1script0 which is the primary server.
Modification of the picomotor script, moving Pcal picomotors to the PICO network and/or update of picomotor driver firmware is required to make old script server retired.
-----
A reason why Pcal picomotors didn't work was a mismatch of the communication mode of Telnet between the picomotor script and its driver. Though picomotor drivers seem to support the char-mode, the picomotor script on the Debian13 system tries to connect as the line-mode for picomotors in the old PICO network in some reason. This issue didn't occur for picomotors in the PICO network. Script servers can access the PICO network directly. On the other hand, the old PICO network can be accessed only via k1gate. This is an only difference between these two networks. But a difference in L3 connection doesn't affect a mode negotiation of Telnet in normal. Because all picomotors had worked fine on the old script server and picomotor drivers in the old PICO network was able to be accessed from k1ctr which was the Debian12 system, it seems an issue only on the Debian13 system.
Although we haven’t identified the exact cause of this issue yet, I downgraded k1script1 which was the secondary script server and moved only the problematic picomotors on that server because Pcal picomotors are required for tomorrow's integration sphere calibration at YPcal. Another all processes which have no issue are still on k1script0 which is the primary script server.
The picomotor script connect picomotor drivers by using python3-socket package instead of telnet package. So IAC negotiation which is done at the beginning of Telnet connection must be implemented by ourselves. I'm not sure the firmware on picomotor drivers are kept modern ones, but anyway a versions of socket libraries and a python3-socket package on the Debian13 system is bumped up from ones on the old Debian systems. For this reason, some kind of incompatibility on a specification of Telnet negotiation expected by Debian13 and picomotor drivers. To make the old script server retired, this issue must be solved.
I changed the setpoint for the F1 GAS from -3170 to -3700, and the F2 GAS from 1280 to 1240, so that the MN V Oplev can be zero.
I changed the setpoint for the F2 GAS from -1109 to -1198 so that the MN V Oplev can be zero.
I checked the IM transfer functions after the TMP stopped. The situation was not changed (The DC gain of transfer functions is smaller than the reference, except for L).
[Tanaka, Hirose, Saito]
To align the optical path of the sub-laser beam with the IR beam on the POS table, we reinstalled two irises and performed alignment. In addition, we prepared mirrors, a BS, lenses, and ND filters to combine the sub-laser beam and the IR beam on the POS table and inject them simultaneously into the RFPD.
- Since the alignment in klog:36773 was not successful, we changed the positions of the irises and the mirrors used for alignment (Photo 1). To more precisely match the optical path of the sub-laser beam with the IR beam on the POS table, we moved the position of the second iris (in klog:36773) to 2125 mm from the sub-laser. We adjusted the setup so that the IR beam on the POS table hit the center of the mirror located after the second iris, and then fine-tuned the iris so that the beam passed through its center. Using the second and third mirrors after the second iris, we aligned the sub-laser beam so that it passed through both irises.
- We prepared mirrors, a BS, lenses, and ND filters to combine the sub-laser beam and the IR beam on the POS table and inject them into the RFPD. The lenses are used to make the beam diameter smaller than the RFPD sensor diameter and to match the beam sizes of the sub-laser and the IR beam on the POS table at the RFPD location. The ND filters are used to reduce the sub-laser power entering the RFPD, since the maximum allowable input power to the RFPD is 1 mW.
Removed time segment is [1300000000, 1350000000)
These data is available on Kashiwa.
There was no undelivered data to Kashiwa in that time segment.
From my eye,
1. TCam itself became upper direction (and some rotation?)
This issue come from the image of the beam duct shadow and oplev scattered signal.
2. Mirror itself became upper direction
Due to the cooling down the duct shield, mirror position would be upper direction, but the amount of movement would be smaller than movement of 1.
We will wait the Takahashi-san work.
3. PCal beam became more upper direction
Compared with the movement of 1. and 2. both PCal beams became more upper direction.
Dan-san described that the beam position of the Rx PD became upper position from hardware condition.
We should consider again after the Takahashi-san work,
one possibility is that "the EYA chamber tilted due to the duct shield cooling".
We will search the past log around the cooling of the previous EYC duct shield cooling
-> I summarized the previous TCam image
klog31367, start the duct shield cooling 18th Oct., 2024 gif.2.
gif.3. showed the 296 K -> 250 K small rotating?? by checking the oplev scattered beam
I changed the setpoint for the BF GAS from 3860 to 3490 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 -2440 to -3150 so that the MN V Oplev can be zero. The F1 GAS was offloaded with the FR.
For most of the components, the shape(or structure) of the images didn't change, the brightness changed.
One notification is that the some bright shadow at the left top part of mirror changed so much depending on the beam position.
Next, changed the mirror height and evaluate the structures of the bright points.
(Changing the BS alignment with Yarm initial alignment)
The results will be appeared later
For most of the components, the shape(or structure) of the images didn't change, the brightness changed.
One notification is that the some bright shadow at the left top part of mirror changed so much depending on the beam position.
Next, changed the mirror height and evaluate the structures of the bright points.
[Kimura, Yasui, R. Takahashi, M. Takahashi and H. Sawada]
Test period: 2026/04/20~2026/04/21
A vacuum leak test was conducted on the SRM.
A leak of approximately 1×10⁻⁹ Pa·m³/s was detected at the side flange on the +X side of the SRM.
On the other hand, the results of the SRM build-up test showed a rate in the range of 1×10⁻⁴ Pa·m³/s.
Since the difference in leakage rate was significant, a leak test was conducted by blowing helium gas onto all flanges installed on the SRM; however, no new leak points were found other than the side flange.
For reference, a plot of the pressure rise during the build-up test is attached.
Based on the slopes of the straight lines on the double-logarithmic graph (SRM ~ 1, OMMT ~ 0.2), it was determined that the pressure rise in the SRM was due to a leak, while that in the OMMT was due to outgassing.
The build-up value for the OMMT was in the range of 1×10⁻⁷ Pa·m³/s.
[Kimura, M.Takahashi, H. Sawada, R. Takahashi and Tomaru]
On April 22, we continued conducting leak tests on the SRM vacuum vessel.
Since vacuum leaks were suspected due to cracks in the weld lines,
we performed leak tests by blowing helium gas onto the weld lines and using the hood method; however,we were unable to locate the leak. (See attached photo.)
The build-up test was conducted with the SRM and OMMT sharing a common vacuum, and the pressure was recorded using the OMMT’s pressure gauge.
The plot of the build-up test is attached.
The total duration of the build-up test was 935 minutes.
The estimated leak rate from the build-up test was 9.7 × 10⁻⁴ Pa·m³/s.
The slope of the pressure rise was approximately 1.
[Kimura, Tomaru, M. Takahashi, H. Sawada and R. Takahashi]
"Inspection results for the sealing surface of the +Y side flange"
A vacuum leak in the range of 1×10⁻⁹ Pa·m³/s was detected, so we inspected the sealing surface of the side flange on the +Y side.
As a result, we found a slight scratch on the elastomer surface.
Based on the location of the scratch, we believe it is the cause of the vacuum leak.
After installing a new elastomer on the sealing surface of the +y side flange, the flange was closed.
The flange fastening bolts are currently in a temporarily secured state.
"Inspection results for the sealing surface of the SRM top flange"
After removing the SRM top flange, the condition of the flange sealing surface was inspected.
No dirt or scratches were found on the surface of the elastomer.
Therefore, it was decided to reuse the SRM top flange as a seal.
I offloaded the following GAS filters with the FRs.
- ITMX: F1 GAS
- ITMY: F0 and BF GAS
- ETMX: BF GAS
- ETMY: F0, F1, F3, and BF GAS
- BS: F0 GAS
- SR2: F0 GAS
- SR3: F0 GAS
Abstract:
I implemented GAS modal damping for PR3 and checked the performance.
Though the best configuration for low frequency noise reduction and high frequency noise reduction is diferent, M1 modal damping with IMV damping would be good as well as the case of BS.
In addition, sensor correction doesn't seem to work well in the case of using GAS modal damping.
Detail:
I implemented modal damping for PR3 as well as BS, SR2, and SR3 (klog36755, klog36780, and klog36781).
Figure 1 and 2 shows the sensor and actuator matrices for PR3 GAS modal damping.
The measured data are stored at /users/VISsvn/TypeBp/PR3/{Spectra,TF}/Measurements/2026/0421/.
Figure 3 shows the OLTF of GAS modal damping for the first mode.
After implementing the modal damping, I measured IMV OSEM spectrm with thevarious configuration.
Figure 4 shows the result (see legend for the configuration of each color measurement).
Current LOCK_ACQUISITION state induces large noise at high frequency while it has the best performance at microseismic band.
If only IMV damping is engaged, the noise level at high frequency is better as well as no control case while the peak at microseismic band is larger.
If using the modal damping and IMV damping at the same time, the performance is middle of the above two condition.
Though I'm not so sure which configuration is the best for LOCK_ACQUISITION and OBSERVATION, M1 modal damping with IMV damping would be good because this configuration balances the noise level at low frequency and high frequency.
Anyway, if the suspension motion disturbs the lock acquisition of DRMI and RSE, the modal damping is worth testing to reduce te suspension motion of PR3.
