ETMY went to PAY_TRIPPED at 20:28:58 JST.

It was caused by a loud glitch on the photosensor for MN_V2 as shown in Fig.1 and Fig.2 which is just a zoom up of Fig.1.
Because other sensors (V1, V3, H1, H2, H3) didn't detect such a loud glitches, it shouldn't be a real motion.

I have no enough time to check more detail today. So I keep it in PAY_TRIPPED with disabling voice notification by guardian.

Around 18:30, I changed the setting temp of #1 precision cooler for the PSL room from 21.86 to 21.88 because the temp in the PSL room is decreasing because of the temp decrease in the corner station area as the Figure.

I confirmed that the new matrix (and damping) works well at this moment and IFO can reach OBSERVATION state without any problems.

• https://klog.icrr.u-tokyo.ac.jp/osl/?r=22084
• https://klog.icrr.u-tokyo.ac.jp/osl/?r=24711
• https://klog.icrr.u-tokyo.ac.jp/osl/?r=29334 : Reinforced the foot flange of the STM2 pedestal.
• No sticking out stop screws on the STM1 holder's face. https://klog.icrr.u-tokyo.ac.jp/osl/?r=21655.

I found that the strange sticking out stop screw on the STM2 holder's face did not find in 2018 https://klog.icrr.u-tokyo.ac.jp/osl/?r=5943. And the STM2's mirror holder seems Newport/New Focus 8822-AC-UHV. Looking at the web site here, I think this part should not be sticking out. And also, some photos show that the corresponding screw's head seemingly sinks into the PEEK-like small part, while the design drawing would indicate that such sinking would not happen. Seems not so healthy... If the PEEK-like part was broken due to the "too-much" screwing, and the breaking flagments are in the holder, that would cause some resonances.

The link to the page is here.

FYI, the latest STM2 photo can be seen in klog29410.

Several concerns are:
1. Clamp is too short to clamp the post that has a significantly larger thickness compared with the clamp.
2. The post extends slightly beyond the edge of the optical table.

I personally quickly checked what happend around 2022 Jan 17 15:10 JST, since when the 128 Hz peak appearing even in IMMT1T QPDs, according to the investigation in 24703. As far as I found in klog, only the post 19489 was directly related to this area.

What I find from this klog post are:

• According to the photo attached to 19489, there were no beam dumps attached to STM2. So the beam dump structure would not be related to this peak.
• The pedestal and the mirror holder should be identical ones we are now using. What they did are (1) moved the STM2 assembly a little bit, and (2) rotated the STM2 mirror holder.
• So, the pedestal structure itself might not be the main cause.
• So, the existence of the electric wires seen in the photo in concern might not be the main cause.
• So, the mirror holder structure itself might not be the maincause.
• The way to fix this foot to the optical table is bad, but this is still so bad. So I do not think this is the main cause.
• Assume beam clipping in IFI would happen due to the work regarding STM2. But it seemed they used some irises to preserve the input beam axis. They reported "...We put irises after MC, before IMMT1, after IMMT2 and at the POP table." and "After STM2 was moved, we adjusted the orientation of STM2 to put the beam through the irises." Reading this, if this is true, the STM2 at the new position can steer the beam passing through the iris before IMMT1 and after IMM2, and at the POP table (??). It would be natural to think, no NEW clipping in IFI happened if the beam path is so binded. They mentioned "We also checked the beam spots inside the Faraday and found no beam clipping.", but I think it is difficult to declare such a thing definetely with the real world. But anyway, this would not affect the my conclusion.
• They rotated the STM2 mirror holder. Looking at the pdf file in 19489, the photo in 19489 would be taken before rotating the STM2 mirror holder. Actually, in the recent photo in 29342, surely the STM2 mirror holder is rotated. But how can we think only rotating the mirror holder cause such peak generation??
• By the way, I found something in the photo in 19489. The the front face of the STM2 mirror holder has three "holes" near the mirror, and one of them (at the most right position) seems to have a stop screw-like part. I am not sure well what is this for. I can assume this part may come to the top of the mirror holder now. Anyway, this type of mirror holder has unnecessarily number of parts to support the mirror (like some PEEK? parts behind the mirror). These too much parts may lower the resonant frequency of the whold mirror holder.

To recover the suspensions, I modified ETMY_IM_OSEM2EUL matrix from the one in fig1 to the one in fig2.
In this modification, Y2L, Y2T, and L2Y coupling cannot be canceled, so these coupling would be increased a lot but it would be fine because these signals are only using for NBDAMP and Y resonances are far from the L/T resonances.

After the modification, I requested the LOCK_ACQUISITION state and confirmed ETMY can reach LOCK_ACQUISITION state without problems.
Since I haven't checked the spectra of the signals, it would be better to measure it and compare with the one before IMH1 photosensor trouble.

Note:

Since DGS maintainance is ongoing, I also don't chack IFO can be locked with this new configuration.
So, it should be also checked soon.

ETMY was TRIPPED at 11:42:15 JST today (fig1).
The reason seems that strange behaviour of H1 photosensors, so I disabled watchdog for IMH1 photosensor.
Since this photosensor is used for NBDAMP_{L1,T1}, it should be consider how to damp these two modes.

Note:

I am not so sure the exact reason why the sensor became strange but we have had many experiences that sensors became strange during cooling.
So, it would be due to thermal stress on the cables inside the cryostat due to cooling.

[Ushiba, Komori]
We updated the estimation of the KAGRA suspension thermal noises.
These updates are based on the Q-values measured in klog:32009 and the coupling constants measured in klog:32004

For the pitch thermal noise, I assume that the dominant contribution originates from IX due to the significant beam mis-centering of approximately 1 cm.
The plotted pitch thermal noise is calculated based solely on the IX pitch Q-value of the TM and the 1-cm mis-centering.

For the roll thermal noise, I noticed a mistake in my previous estimation, where I neglected the dynamics of the MN and used an inaccurate moment of inertia for the IM and TM.
To correct this, I used the moment of inertia values provided in the list consistent with the SUMCON model referenced in klog:32004.
From the transfer function measurement of the IX MN roll to DARM, which was found to be 2.8e-4 m/rad, I derived the roll-horizontal coupling (RHC) of the TM to be 1.8e-4 m/rad by accounting for the moments of inertia of the MN, IM, and TM.
Based on the measured peak heights from all test masses, the RHC values are similar across all four test masses.

The vertical thermal noise is estimated using the measured vertical-horizontal coupling (VHC) value of 1/300.

The black solid line in the first figure shows the current best estimate of the total suspension thermal noise.
Among these components, the vertical and roll thermal noises are significantly smaller than the horizontal and pitch thermal noises.

Additionally, I estimated the binary neutron star (BNS) range improvement by removing the suspension thermal noise.
If this noise source is eliminated, the DARM sensitivity would resemble the cyan line shown in the second figure.
The BNS range is expected to improve by only ~20%, indicating that other noise sources also need to be addressed.
Even if the actual suspension thermal noise is slightly higher than this estimation, the improvement in the BNS range would remain modest and unlikely to exceed a factor of two.

Abstract:

I measured the coupling from ITMX V/R motion to DARM by using V resonance at 10Hz and R resonance at 22.7Hz.
Measured coupling values are 3.4e-3 m/m and 1.4e-4 m/rad for V and R, respectively.

Detail:

To evaluate the thermal noise from vertical/roll modes, I measured the coupling from vertical/roll to DARM by using ITMX.
I excited 22.7 Hz resonances for R and 10 Hz for V by using MN actuators to measure the coupling.
To monitor the vertical/roll motions, MN OpLevs are used.
During the measurement, I engaged the resonant gain (FM4 of DARM2 filter, filter shape was changed every measurement) at measured frequency to reduce the RMS of error signals.

Top right figure in fig1 and 2 shows the TF from local sensors to DARM displacement.
Since MNV/R OpLevs were calculated into the unit of um and urad,respectively, the TF gain directly represents the coupling from MNV/R motion to DARM.
Note that it is not the coupling factor from TM motions to DARM, so it is necessary to estimate the ratio between MN motions and TM motions.

Obtained values are as follows:
MNV: 6.26e-3 m/m @ 10 Hz resonance
MNR: 2.79e-4 m/rad @ 22.7 Hz resonance

Roughly speaking, the difference betweenTM motion and MN motion are decided by mass (or moment of inertia) of TM and IM+MN.
The mass and moment of inertia ratio between TM and IM+MN are 0.54 (22.88(TM)/(20.13(IM)+22.53(MN))) and 0.50(0.14(TM)/(0.095(IM)+0.14(MN))), respectively.
So, the couplings from TM motion to DARM are
TMV: 3.4e-3 m/m
TMR: 1.4e-4 m/rad.

Note:

Measured TMV to DARM coupling is very close to the ideal limit (1/300=3.33e-3 m/m), so it would be difficult to reduce the coupling even if the vertical thermal noise is a problematic.
Used mass and moment of inertia value is the one in the SUMCON model named TypeA180429_300K.m.

Since ETMX NBDAMP L1/T1 seem to be oscilating, I reduced the gain by a factor of 2 (I added a gain of 0.5 in FM1 of both L1 and L2).
After that, oscillation seems to stop.

Since cooling is gradually proceeding, it would be better to have VIS campain before New Year holidays.

For the estimation of V2L coupling of Type-A suspensions, I implemented the MNV OpLev calibrations based on the values written in OpLev summaries (JGW-G2113192-v4 and JGW-G2214012-v2).
Figure 1 shows the filter banks, which I implemented the calibration factors.

Here are the image analysis results of the taken images.
Note that "Before" : 2024/12/19 8:26~8:32 and "After" : 2024/12/19 10:57~11:03

• ITMX
  • original(before) -> (after)
  • fitting result (before) -> (after)
  • overlay with previous reference center (before) -> (after)
• ITMY
  • original(before) -> (after)
  • fitting result (before) -> (after)
  • overlay with previous reference center (before) -> (after)
• ETMX
  • original (before) -> (after)
  • fitting result (before) -> (after)
  • overlay with previous reference center (before) -> (after)
• ETMY
  • original (before) -> (after)
  • fitting result (before) -> (after)
  • overlay with previous reference center (before) -> (after)
• For ETMX, I updated the reference positions. I also re-draw the yellow line shown at the k1mon4. For other mirrors, we don't need to update the reference positions.

I changed the setpoint of the heater from 25°C to 27°C at 15:06 JST.