Takano, Fujimoto, Tanaka

## Abstract

We found that SRC flash light injected to ITM MN photosensors when PRM was MISALIGNED_BF. In this state, if IMC output power reaches to ~4 W, the flash disturbs the ITM MN DAMP control and kicked ITM. Therefore, we cannot increase the laser power when SRM is aligned for now.

## Detail

This noon, during PLL work, we increased the IMC power from 1 W to 10 W to increase SNR of PLL beat signal by increasing the main beam power on POS table. After that, the power values at AS and REFL PDs got 0 (fig.1). Also, the beam spot on AS camera was disappeared although the VIS guardian seems to be not changed as long as VIS_OVERVIEW screen says. This time, since we used SRX cavity before increasing IMC power, ITMY, ETMX and ETMY were misaligned and others including SRM were aligned. We found that the behaviors of DC powers on AS and REFL PDs and the flash in POP RF17I seems to got strange when IMC power surpassed ~4 W (fig.2).  

After we got back to controll room, we found that ITMX oscillated too largely although the ITMX guardian saied ITMX is "LOCK_ACQUISITION". MN controll feedbacks' values seem to got too much and saturated.  So this is reason why the beam at AS and REFL disappeared. So we started to investigate ther reason why ITMX oscillated by increasing laser power to ~4W. On the other hand, This early morning, Dan-san performed an initial alignment. During the alignment, he increased the power to ~5W. However, he did not reported such an issue in his klog and ITM seemed to be stable in this morning. The difference between the early morning situation and this noon situation is that SRM is aligned or not. So one of possibilites is that SRC flash disturbs the ITM control with more than 4 W IMC output.

First, to distinguish the cause of this issue by local control itself or by increasing the power when SRM is aligned, we investigated there is something wrong in the local control by requesting ITM guardian to go to the upper state one by one (ISOLATED -> DAMPED -> ALIGNED -> LOCK_ACQUISITION). This time, SRM was aligned and IMC output power was 1 W. ITMX got the LOCK_ACQUISITION state without any oscillation. Therefore, this cause may not be the local control itself.

Second, we tried to monitor what happened in the ITMX control when we increased IMC output power with aligning SRM. So we requested PRM guardian to the MISALIGNED_BF state. Just after PRM got the state, MN photosenser got glichy. Fig.3 shows the time series around the moment when PRM got to MISALIGNED_BF. In top two rows, the error signals of the damping control with photosensers in each DoF and in bottom rows, IMC output power, PRM guardian state, and the SRC error signal, respectively. As you can see, after PRM moved from MISALIGNED state (GRD-VIS_PRM_STATE_N = 1400) to MISALIGNED_BF (STATE_N = 1600), photosensor error signals got glichy when IMC power was 1 W and SRC is flahing. In this state, we misaligned SRM manually so that SRC flash could not be observed. Then the gliches in photosensor signals were also disappeared (fig.4). Moreover, Fig.5 shows the timeseries when we increased the power to ~10 W in this noon. As the power increased, the gliche size increased. At last, MN control Therefore, the SRC flash injected into photosensors (especially H1, H2, V2, V3 judging from fig.6) and the flash disturbs ITMX MN control with photosensors. 

Similarly, this glitch issue were observed in ITMY photosensors (fig.7).

PRM must be in the MISALIGNED_BF state when we increase the power for the safety reason. So we cannot increase the power with SRC flash for now.

## Discussion

We also found that the glitch size seems to depend on the PRM alignment (Fig.8). When PRM is In MISALIGNED_BF state (N=1600) where PRM misaligns largely in Yaw and goes out of the oplev range, the glitches are in photosensor signals. However, when PRM is in the MISALIGNED state (N = 1400), where PRM is misaligned smally in Pit but PRM is in the oplev range, or the LOCK_ACQUISITION state (N = 2000), where PRM is aligned well, the glitches are not observed. Therefore, the SRC flash goes back to PRM, and reflected at PRM. In MISALIGNED_BF state, we are not sure of this reflection light. It may hit somewhere and scatter. This scattered light injects to ITM photosensor. This is one possiblity. However, we have not confirmed that the glitches are still not in photosensors when the power is increased with good PRM and SRM alignment. Also, we have not checked the relation between the misaligned direction. At any rate, we need more investigation.

## Discussion 2

By the way, during this investigation, we found that IR light seems to be injected in SRM oplev signal even though the IR filters (according to Dan-san) were installed in front of oplev QPDs. 

## Note

We should not trust what the guardian says completely.

[Kimura]

 I have shut down all but a few of the FFUs around the OMC, SRM, and SR3.
Attached are photos of the FFU control panels taken before and after the shutdown.
 One FFU on the OMMT side, one on the SRM side, and two on the SR3 side are currently operational.
This situation is the same as it was before the SRM was opened.

We implemented the LSC and ADS guardian for SRX/Y.

LSC

We kept the control filters and gains we used yesterday. For now, we only need these filters: FM5 (z3p500) in SRCL1, FM6 (ELP300) and FM9 (DB8) in SRCL2. We deleted other filters and added a new one in FM1 (temp) SRCL1, which has a gain of -800.

With these filters, we added new states in the VERTEX guardian to lock SRX/Y with POP17I. AS PDA1 DC was used to monitor the resonance condition. The SRY control by the guardian was done successfully, and the OLTF looked almost the same as yesterday.

We also implemented the 3f control using REFL51I. When SRY was locked with POP17I, we measured the relative gain between POP17I and REFL51I, and found that the latter was 30dB lower. We created a new filter with a gain of (-1) * -30dB named 'SRC3f' in SRCL2 FM2 to compensate for the gain difference. The negative sign is necessary to lock the SRC in a condition different from that with POP17I: with REFL51I, the carrier is anti-resonance, and the f1 sideband is on resonance. Before trying to lock with the VERTEX guardian, we noticed that REFL51I has a decently large offset, even though we removed it yesterday with the laser shutter closed. It seems the offset value changed quickly, but we are not sure why. For now, we manually added an offset in SRCL1 and closed the control by the VERTEX guardian. This time, the AS DC power was almost at its minimum, indicating the SRC is locked to the sideband. Once we locked the SRC to the sideband, we tuned the demodulation phase of AS34 to monitor the buildup of the sideband, so that AS34I is near the maximum and AS34Q is fluctuating around 0. The phase was now set from -130° to -20°.

ADC

The ADS control in the ASC_LOCK guardian was implemented. We created new states there, and the ADS for SRM alignment was turned on when SRC was locked with POP17I. The state name was meant for SRY, but in principle the ADS should work with SRX (we haven't tested yet). When we locked the SRY with POP17I, we requested the ADS for SRM and confirmed the AS DC power increased, indicating the alignment had improved.

For now, when the ADS is stopped, the output value remains, and it is not offloaded. We need to add the offloading function in guardians later.

Summary

• I manually misaligned SRM and then performed the initial alignment following the procedure in klog36759.
• No significant problem was found during the alignment.

Manual misalignment of SRM

First, I manually misaligned SRM. In the current initial alignment guardian, this step does not seem to be performed automatically. In addition, the MISALIGNED state of SRM has not yet been set up (klog36888). The current SRM state was ALIGNED.

I checked K1:LSC-POP_PDA1_RF17_I_MON and found flashes with an amplitude of approximately ±0.9. Similar flashes were also seen in K1:LSC-POS_PPOL_DC_OUT_DQ. This PD is the one temporarily installed at POS a few days ago.

To remove these flashes, I changed the oplev SET values of the SRM TM stage in both PIT and YAW. (fig_001)

XARM alignment

• Requested IRX_LOCKED.
• Requested GRX_LOCKED_WITH_IRX.
  • The lock acquisition did not succeed easily at first.
  • However, I then noticed that the GR shutter was not open. After opening the shutter, the guardian reached GRX_LOCKED_WITH_IRX.
• Requested ALIGNING_XARM.
  • While waiting for the ADS signals to settle, I checked the beam positions on EX and IX with cameras.
  • I slightly adjusted the PIT of ITMX to move the beam position on ETMX. The beam was moved upward.
  • I waited until the ADS signals became stable.
• Requested RECORD_GOOD_VALUES_XARM.
• Offloaded.

YARM alignment

• Requested IRY_LOCKED.
• Requested GRY_LOCKED_WITH_IRY.
• Requested ALIGNING_YARM.
  • In this state, ADS loops are engaged for ETMY, ITMY, and SR3.
  • I moved BS to adjust the beam position on ETMY.
  • The BS alignment was changed from (P, Y) = (8.0, -44.7) to (P, Y) = (8.5, -41.4).
  • I waited until the ADS signals became stable.
• Requested RECORD_GOOD_VALUES_YARM.
• Offloaded the alignment.

OMC alignment

I followed the procedure described in klog36759.

• Changed the PRM state to MISALIGNED_BF to increase the laser power.
• Closed the GRX and GRY shutters.
• Tweaked the OMMT1 angle using the coil-magnet actuators to find the beam on the OMMT2 trans QPD.
  • Even before changing anything, the beam was already visible on OMMT2_TRANS_QPDA1_DC.
  • I centered the beam on the QPD.
  • The OMMT1 alignment was changed from (P, Y) = (-22600, -3900) to (P, Y) = (-25300, -3900).
• Changed the OSTM state to MISALIGNED_FOR_LOCK_ACQ to avoid unnecessary flashes of the OMC.
• Increased the laser power to check whether the obtained beam on the QPD was really the IR beam.
  • I rotated the HWP of PLS.
  • The HWP angle was changed from 149.996 to 161.998.
  • The beam on the QPD responded as expected, so it was confirmed to be the IR beam. (fig_002)

Restoration

• I restored the IR laser power to the original value of 1 W. During this process, the lock was lost, but after relocking, the laser power was automatically restored to 1 W.
• I restored the OSTM state to LOCK_ACQUISITION.
• I restored the PRM state to MISALIGNED_FOR_LOCK_ACQ.
• I restored the PIT and YAW oplev SET values of the SRM TM stage to 0. After this restoration, the flashes appeared again.

[Takano, Tanaka, Fujimoto]

Summary

We measured the carrier buildup factor (ratio of on-resonance to off-resonance power) and the carrier reduction factor (ratio of off-resonance to single-path power) for the four SRY transmission ports: REFL, POP, POS, and AS.
In particular, the measured carrier buildup factors were:

• Design value: 7.3
• REFL: 5.0
• POP: 1.3
• POS: 1.6
• AS: 5.0

The REFL and AS values are somewhat lower than the design value, and POP and POS show significantly larger mismatches.
These mismatches could be caused by additional losses inside the SRY or optical offsets in the PD signals.
One possible origin is s-pol to p-pol conversion caused by birefringence in the ITMY substrate.

Details

As described in our previous klog (#36881), the true reflection port of the SRY is the port transmitted toward the ITMX side, while the interferometer REFL port acts as a transmission port for the SRY. Therefore, the currently available transmission ports for monitoring the SRY carrier buildup are REFL,POP, POS and AS.
For these ports, we performed:

• Measurement of the carrier buildup factor (ratio of on-resonance to off-resonance)

• Check for offsets in the PD outputs

• Measurement of the carrier reduction factor (ratio of off-resonance to single-path)

We also discussed the discrepancies between the measured values and the design values.

Measurement of carrier buildup

Fig. 1 shows the DCPD time-series data for each transmission port (REFL, POP, POS, AS) of the freely swinging SRY.
From these data, the carrier buildup factor (ratio of resonant to non-resonant power) for each transmission port was obtained.
The measured values are:

• Carrier buildup factor:
• REFL: 5.0
• POP: 1.3
• POS: 1.6
• AS: 5.0
• Design value: 7.3 (considering R_{BS}=0.5, R_{IY}=0.996, R_{SRM}=0.85)

Check of PD offsets

Fig. 2 shows the PD outputs with the laser shutter closed and no laser injected into the interferometer.
The offsets of all PDs are nearly zero compared to the on-resonance and off-resonance signals, indicating that they do not contribute significantly to the underestimation of the carrier buildup factor.

Measurement of carrier reduction

Fig. 3 shows the time-series data when the SRM was misaligned and the carrier entered each transmission port (REFL, POP, POS, AS) through a single path.
The cursors in the figure indicate the on-resonance and off-resonance levels.
From these data, the carrier reduction factor (ratio of off-resonance to single-path power) for each transmission port was obtained.

• Measured carrier reduction factor:
• REFL: 0.55
• POS: ~1
• POP: ~1
• AS: 0.57
• Design value: 0.47

Discussion

First, the carrier buildup/reduction factors observed at AS and REFL are generally consistent with each other. However, they still show discrepancies from the design values.

One possible explanation is additional losses inside the SRY.
Additional loss decreases the carrier buildup factor while increasing the reduction factor, which is consistent with the observed trend.
Additional loss can be estimated from the measurement results as follows. The buildup factor b and reduction factor q are given by:

b = ((1+r)/(1-r))^2
q = 1/(1+r)^2

where r is the product of all amplitude reflectivities inside the cavity.
Using the measured values and these equations, the additional loss can be estimated.
However, for the AS results, the additional loss estimated from the buildup factor is 31%, and that estimated from the reduction factor is 50%, which seem too large and also are inconsistent with each other.
Therefore, the discrepancy from the design value cannot be explained solely by additional cavity loss.

Another possible explanation is some offset in the PD outputs.
From the present measurements, we confirmed that there is no significant PD offset when no laser is injected into the interferometer.
However, if there exists an offset that appears only when the laser is injected, it would lead to underestimation of the buildup factor and overestimation of the reduction factor.
As a possible source of such an offset, Tanaka-san pointed out p-pol. light generated by birefringence in the ITMY substrate.
The p-pol. light generated inside the SRY does not interfere with the main s-pol. carrier and can therefore appear as a PD offset.
Moreover, this p-pol. light directly enters the POS and POP ports, while it is rejected at AS and REFL by the OFI and IFI, respectively.
This could also explain why the mismatches observed at POS and POP are particularly large.

This hypothesis can be tested by placing PBSs at POS and POP to remove the p-polarized light and checking whether the buildup/reduction factors improve.
A similar effect should also appear in the PRY, so investigating the buildup/reduction factors at the PRY POS and POP ports may also be useful.

For additional information, I attached the photograph of the current SRM OPLEV QPD layout below.

I modified the ALIGNED state's script of the SRM guardian to engage this hierarchical control. I confirmed that this control was engaged by the guardian and that the guardian could transit between ALIGNED and DAMPED states.

Note:

I have not implemented "LOCK_ACQUISITION" state for SRM. So please do not request SRM guardian to go to "LOCK_ACQUISITION" state. Also, MISALIGNED state has not been implemented.

[Tanaka, Takano, Dan]

After the work reported at klog 36886, we started initial alignment procedure.

Summary

We redid the initial alignment following klog36759. This log summarizes the work up to the X-arm and Y-arm alignment.

What we did

We started the initial alignment again following klog36759.

X-arm alignment

At first, the beam position on ETMX was largely displaced, so we manually adjusted ITMX and ETMX while keeping TR_GRX.

• ITMX: P, Y = 4.5, -12.1 -> -5.4, -15.0
• ETMX: P, Y = -2.7, -5.6 -> -9.3, -5.5

The transmitted power did not increase easily. The PR3 dither seemed to have been disengaged because its signal exceeded the threshold(?).

We also found that K1:LSC-PSL_PWR_SCALE_OFFSET had changed from 1 to 2.9 around the DAQ restart, so we set it back to 1. After repeating the X-arm alignment, it worked well.

Y-arm alignment

The beam position on ETMY was also largely displaced, so we adjusted the BS alignment.

• BS: P, Y = 3.4, -40.4 -> 8.0, -44.7

After this adjustment, the Y-arm alignment looked good.

Request from Tanaka-san

We have added two DAQ channels to the k1lsc model.
SRCL_IN1 16384
SRCL_OUT 16384
 

[Tanaka, Fujimoto, Saito, Dan]

Summary

We checked the SRM OPLEV QPD signals after yesterday's centering work. The OPLEV beam was incident on both the TILT and LEN QPDs, but the environmental light and IR contribution were relatively large. We installed an IR filter also in front of the LEN QPD, centered both QPDs, and confirmed that the LEN OPLEV mainly responds to SRM longitudinal motion.

What we did

We first checked whether the OPLEV beam was incident on the QPDs. The environmental light level was large, so we reduced it as much as possible during the measurements.

For the TILT QPD, the sum signal was about 90 counts with the OPLEV beam and about 10 counts when the beam was blocked. Removing the IR filter increased the sum from about 90 to 230 counts, corresponding to an attenuation factor of about 2.6.

For the LEN QPD, no IR filter was installed at first. The sum signal was about 317 counts with the OPLEV beam and about 307 counts when the beam shutter was closed. After inserting an IR filter, the sum decreased to about 130 counts, corresponding to an attenuation factor of about 2.4. Opening the green shutter did not significantly change either QPD signal.

These checks showed that the OPLEV beam was already incident on both QPDs, but the IR contribution was non-negligible. Therefore, we decided to install/keep IR filters in front of both the TILT and LEN QPDs.

IR filter installation and QPD centering

We installed an IR filter in front of the LEN QPD using a right-angle clamp because of the limited space around the QPD.

After reducing the environmental light, with the OPLEV beam on, the IR shutter closed, (and the green shutter open,) the final sum values were:

• TILT sum: 90.5 counts
• LEN sum: 130 counts

We then centered both QPDs using the micrometer stages while monitoring the QPD signals.

LEN OPLEV response check

Finally, we excited the SRM in longitudinal and yaw at 10 Hz and checked the LEN OPLEV response. (Attached)

When the SRM was excited in the longitudinal direction, the response appeared in the LEN OPLEV signal. When the SRM was excited in yaw, the response clearly appeared in the TILT OPLEV signal. Therefore, we judged that the present oplev position is acceptable.

Files: `/users/VIS/TypeB/SRM/LogNotes/260513_oplev/`

[Kimura, Yasui and Uchiyama]
 On the afternoon of May 12, we shut down the Q-mass unit that had been temporarily installed on the OMMT's exhaust line.
 Data on residual gas distribution up to May 12 is stored in a folder within the KAGRA Dropbox.
Going forward, we will consider temporarily operating the Q-mass unit to investigate the exhaust malfunction that occurred this time.
  Details on the operational method will be reported separately.
  

I modified the servo filter "int" (FM10) in the OLDAMP FILTERS in the IM loop. The cross-frequency between the TM and IM loops was set to 0.1Hz for both pitch and yaw directions. I offloaded the IM pitch with the picomotor.

Summary

We found that the power budget with SRY on resonance is not so weird after all. We misunderstood what is the 'reflection light' from SRY.

Detail

As reported shortly in this post, we observed an unfamiliar power change when SRY flashed. Fig.1 shows the DC power signal in several ports: REFL, AS, POP, POP s-pol, POP p-pol, and POS (newly installed on Monday). Surprisingly, the power at both REFL and AS ports increased. Another concern was that the buildup evaluated at POS was smaller than expected from the mirrors' reflectivity: it should be ~7, but it was just 1.6.

Hiroki pointed out that the REFL port in SRY (also SRX) is actually not the cavity reflection in the usual term. Fig. 2 shows the cavity mirror layout, and thinking of it carefully, the 'normal' cavity reflection for SRY is the beam going to ITMX (for SRX it's ITMX), and the REFL port is one of the cavity transmissions. In that sense, REFL and AS are equivalent for SRX/Y, and we should see the built-up power at both ports.

The buildup ratio at REFL and AS is 5.2 and 12, respectively. It seems consistent (but a bit low) at REFL, but at AS it is overestimated for some reason. On the other hand, the other ports show smaller buildup: 1.6 at POS (as mentioned above), 1.7 at POP, 1.5 at POP s-pol, and 2.3 at POP p-pol. We don't have any clear answer for them now.

[Dan, Fujimoto, Tanaka, Takano]

Summary

After adjusting SRM yaw, we successfully implemented the SRCL control with POP f1 signal. Currently the UGF is ~ 18 Hz. Also, SRM ADS worked with AS DC power.

Detail

After the PRM alignment (and the QPD centring work, which was found to be a failure later, though), we tried to close the SRCL control loop. Looking at several RF signals, we decided to use POP f1 I signal, which seemed to have a decent SNR. We copied all the filter banks from PRCL1 and PRCL2 and implemented them into SRCL1 and SRCL2. After tweaking the gain in SRCL2, SRY was stably locked with the setting below:

• SRCL1: FM9 (pole: 0 Hz, zero: 8 Hz)
• SRCL2: FM6 (an elliptic low pass with a cutoff at 300 Hz)
• Gain: -800

With these filter bank settings, the open loop gain was measured as shown in Fig. 1. The UGF was ~ 18 Hz, and the phase margin was 43°.

Following the success of the SRCL control, we implemented a dithering control loop to improve SRM alignment. The dithered frequencies were 4.3 Hz for SRM pitch and 6.3 Hz for SRM yaw, and we demodulated the AS_PDA1_DC signal. The demodulation phase was set as shown in Fig. 2; I set it to 8° for pitch and 13° for yaw. The loops were closed with the gain of -1000 for both pitch and yaw control.

Even though the alignment was improved by SRM ADS, the intracavity buildup was only 1.5, which should be ~ 7, considering the BS and SRM transmission. It was also found that the REFL power increased when SRY was locked, which should be decreased. It might be that we locked SRY to higher-order modes, but judging from the camera images at AS and POP Spol, the fundamental mode appeared to be on resonance. Yesterday, we confirmed that the buildup factor for PRY was ~7, close to the expected value, so the birefringence loss does not seem to limit amplification. We need more information to investigate the reason.

[Dan, Takano, Tanaka, Fujimoto]

Summary

From the previous SRC flash searches, it was found that a large yaw offset of the SRM is required to observe SRC flashes. Under such conditions, the SRM OPLEV beam went out of range of the QPD.
To address this issue, we performed centering of the OPLEV QPDs (TILT QPD and LEN QPD) for the aligned SRM.
Although the centering onto the QPDs were confirmed while we were in the mine, after returning to the control room we found that the beam was no longer incident on the TILT QPD. We plan to check the situation and recover the alignment tomorrow.

What we did

After obtaining a reasonable SRM alignment during the SRC flash search in the morning (klog #36876), we carried out the SRM OPLEV centering work in the mine in the afternoon.
Inspection around the QPD revealed that the beam was clipping on the holder of the first steering mirror. Therefore, we decided that rearrangement of all the optics, including the steering mirror, BS, TILT QPD, lens, folding mirror, and LEN QPD, was necessary.

First, we adjusted the position of the first steering mirror and aligned the beam onto the TILT QPD. After coarse alignment, fine adjustment was performed using the micrometer stage of the QPD while monitoring the QPD signals.

Next, alignment onto the LEN QPD was performed. For the LEN QPD, the distance relationships between the SRM and lens, and between the lens and LEN QPD, are important. Therefore, the downstream optics (BS, lens, folding mirror, and LEN QPD) were repositioned and aligned so that these distances matched the previous setup (klog #7952).

Fig. 1 shows a photograph of the final layout. The distances between optics are as follows:

• TM to upper viewport: 990 mm (assumed to be the same as the previous setup)

• Viewport to first steering mirror: 305 mm

• First steering mirror to BS: 40 mm

• BS to lens: 45 mm

• Lens to folding mirror: 110 mm

• Folding mirror to LEN QPD: 273 mm

After repositioning, final centering onto the QPD was performed using the micrometer stages of the QPD.

After the adjustment, the optical table was covered with aluminum foil, and we returned to the control room. At that point, we noticed that the beam was no longer incident on the TILT QPD. It is possible that the aluminum foil blocked the beam path or disturbed the alignment during covering. We plan to check the situation tomorrow and recover the alignment.

I checked the behavior of the OSEM gains when the Type-B/Bp chambers were evacuated in June, 2024. The gains of some OSEMs in any IMs (PRM, PR2, PR3, BS, SRM, SR2, SR3) were reduced in vacuum.

Takano, Fujimoto, Tanaka, Dan

Summary

We performed the initial alignment of the IFO by following the procedure in klog36759. After that, we tried to find the SRC flash by shaking SRM and moving SRM IP in yaw direction. Finally, we observed clear SRC flashes.

What we did

Initial alignment of IFO

First, we performed the initial alignment of the IFO. Since the transmissivity of SRM was changed, the usual initial alignment procedure cannot be used for the AS side alignment. Therefore, we adjusted the alignment by referring to klog36759.

For X-arm and Y-arm, we used the usual initial alignment procedure. After that, we continued the alignment according to the procedure described in klog36759. As written in the klog, we moved OMMT1 while monitoring OMMT1_TRAS_QPD1. When the beam position on OMMT1_TRAS_QPD1 became almost centered, we regarded the AS side alignment as good.

Search for SRC flash

Next, we tried to find the SRC flash by adjusting the alignment of SRM. We set only ITMY to LOCK_ACQUISITION and misaligned the other Type-A suspensions. We also restored the SRM IM OPTICALIGN pitch value to yesterday’s value of 3825. In this condition, we largely moved only the yaw alignment of SRM. We used the stepping motor and moved SRM IP F0_Y from about -12073 to -27073. Then the flash could be seen. After that, we continued the alignment while monitoring the flash signal. Finally, we found good SRC flashes with the following SRM IM OPTICALIGN values:

• P: 3900
• Y: 480

The flash signal was monitored with K1:LSC-POS_PPOL_DC_OUT_DQ. This is the signal from the temporary PD installed yesterday on the POS table, and the flash appeared clearly in this channel.

However, since the suspension was still oscillating, it was difficult to further optimize the alignment at this moment.

Next, we entered the mine, center the QPDs, and try to turn on the local controls.

This foton error occurred because a filter bank block with one or more filter modules defined was removed from the model.
It was fixed by removing filter module definition of the removed filter bank block manually.

-----
In such case, "File errors" is displayed on a red background, and the error details pop up in a separate window when foton is opened as shown in Fig.1. In many cases, this type of error occurs when a filter bank is removed from the model while its filter design remains. This is because, when the model is rebuilt after deleting a filter bank block, the module name definitions at the top of the foton file are generated based on the new filter bank list, but the filter coefficient definitions are not cleaned up. Normally, this specification should not become a problem, because filter bank blocks are not repeatedly added or removed without sufficient consideration and reviews. But it often becomes an issue within KAGRA’s management and operation framework.

Anyway, VIS-SRM_TM_SENSCORR_L was removed from the k1vissrmp model probably in klog#36457 though some filter module definition was remained on that filter bank. An only safe method is to clean up the filter modules on the filter bank that will be deleted before installing the updated model. But in this case, nobody probably noticed fact that some filter banks would be removed because it's not mentioned in the model update plan (klog#36374). The other method is to edit the foton file manually after installing the updated model. Though foton files are in plain text format, their formatting is so strict and they can very easily become unreadable from foton or models by a manual edit. So it's never recommended. But this is only way if we couldn't notice before installing updated models and I did it in this time. Removed filter modules are shown in Fig.2.

After manual fixing, all errors were cleared on foton. But a warning still remains related to the too low (but non-zero) zero/pole as shown in Fig.3. Though it's not a problem on reading filters by foton and models, numerical errors often occurs on that filter bank. Please remove it if it's unnecessary, or please change corner frequency as a higher value with a careful design of the control if it's used.

Dan, Takano, Fujimoto, Saito, Tanaka

## Abstract

We tried to find the SRC flash by moving alignment SRM. Yaw alignment seems to be larger than the range of IM actuator. We observed SRC flash after moving SRM and SR3.

## What we did

This morning, we tried find the resonancre in SRC by moving the alignment of SRM largely in both P and Y directions. Unfortunately, SRM DC control seems not work well. So we stopped the any DC controls of payload parts, and we decided to use OPTICALIGN to move alignments. This time, we aligned ITMY and misaligned other Type-As, that is, we made up SRY cavity. Moreover, we increased IMC power up to 10 W. We used awggui and set the excitation amplitude ~1000 cnts and the frequnecy 0.05 Hz in Yaw and 0.04 Hz in Pit, respectively. In this parameters, SRM alignments moved largely enough TM oplev gots out of ranges. However, any flash could not be observed in AS RF34 IQ signals and DC power in AS PDs. Therefore, we put the additional PD at the POS table to observe the flash.

This afternoon, we entered the mine and installed the additional PD just after the 2 inch lens in main IR beam path on POS table. Also, we used the POS PPOL PD channel (K1:LSC-POS_PPOL_DC) for this PD tempolarly. After that, we tried again. However, any flash could not be observed. We used TCam to find the beam spot reflected SRM on ITMY. 

Then, we increased Yaw scan amlitude to ~5000 cnts, which is the almost maximum. Then, Saito-kun found that the scattering at the edge of the TCam image.So we adjusted the actuator output when we observed the scattered light. Then, we oscllated the SRM only in Yaw direction. During the oscillation, we adjusted PItch by applying the DC offset with monitoring TCam image. After adjusting, the scattered light could be observed in the edge of both ITMY and ITMX TCam images once per 20 seconds. This duration seems to be consistent with the Yaw excitaition frequency. Therefore, the beam reflected from SRM seems to be shifted largely in Yaw direction. The shift amounts seems to be larger than the IM actuator range. So, it is necessary to use IP to move SRM more largely.

This time, we tried to moved SR3 in Yaw direction with oscllating SRM in Yaw direction. When SR3 moved to -100 urad from initial good setpoint, we can see the beam spot refleted from SRM in POP-SPOL camera. Moreover, the flash signals were observed in the additional PD and REFL PD when the beam spot reflected from SRM was overlaped on the beam spot reflected from ITM. Then, we stopped the excitation and set the actuator output when flash signal is observed in PDs (fig.1). 

However, the beam went away from AS PDs and camera when SRC flashes.