I wrote a basic Python code to convert the png screenshots of camera into playable data.
I used it on screenshots of klog #21777 and applied normalization.
The result is in attached figure.

The code is available in /users/matteo/20220815/image_plot.py

UPDATE: I fixed the aspect ratio of the figure so not to distort the beam shape. The code has been updated and the result is in fig2.

Similar to the work on klog #21759.

For this work the green Y shutter was closed since some scattered light was seen on POP and POP_P cameras.
The camera images and value of CAM_SUM and CAM_EXP have been recorded in two configurations: first with Yarm unlocked and ETMY misaligned, second with Yarm locked. The results are in fig1 and fig2 respectively and in the following tables.

Yarm unlocked and ETMY misaligned (fig1)

Yarm locked (fig2)

The results as well as the beam shapes seem consistent with what previously reported.

In addition, fig3 shows the time series of K1:CAM-POP_P_SUM and K1:TMS-Y_IR_PDA1_OUT16 while Yarm is locked. It seems the two signals are anti-correlated.
In this configuration, POP-YarmL.avi and POP_P-YarmL.avi have been recorded. The shape of the beam does not change a lot on POP despite the beam moving around while on POP_P the beam shape changes a lot. 

[Ushiba, Matteo]

Abstract:
We measured the calibrated Gr PDH correction signal for X and Y arms with and without ALS_FIBX/Y while the arm length was locked with IR and green.

Details:
Arm cavity was locked with IR using INITIAL_ALIGNMENT guardian and green was manually locked. Then the spectrum of channel K1:ALS-(X,Y)_PDH_SLOW_DAQ_OUT_DQ was acquired and calibrated to Hz using the VCO: Frequency Actuation Efficiency reported in klog #6544 and applying the arm cavity pole at 33Hz. The spectra for Xarm are in Fig1 while for Yarm are in Fig2. For Xarm a large peak is visible at 0.43Hz while for Yarm a similar peak is present at 0.38Hz.
In the table the RMS for each case is reported.

It could be possible. If no filters are available, we could install dichroic mirrors in front of the RFPD to remove the residual IR. This will likely reduce the amount of green also by few %.

[Matteo, Aritomi]

Summary
We checked the distances between optics after RST2 and the available optics for the WFS installation.

Details
The layout of POP table can be found here.
We checked the distances of optics starting from RST2. The distances are not precisely measured but gauged in terms of " holes" (25mm).

The POP layout will be adjusted accordingly.

We could find all the necessary optics for the installation of WFSs as well as the mechanics. We didn'd found the PICO/PZT mounts. We will continue investigation at a later stage.

[Matteo, Aritomi]

Summary
We realigned POP_P camera and measured the power on POP and POP_P cameras as function of HWP angle. We then took pictures of POP and POP_P in different locking situations

Details
The POP layout can be found here.
First we moved POP_P camera closer to TFP. The new distance between TFP and POP_P camera is 25mm. Note that other optics position in the current layout are not correct and will be reported later. POP camera is also at 25mm from TFP, so both images have the same scale. Then we measured POP and POP_P power as function of HWP_angle. POP power is computed as K1:CAM-POP_SUM/K1:CAM-POP_EXP while POP_P power is computed as K1:CAM-POP_P_SUM/K1:CAM-POP_P_EXP. Those channel are not stored. The result of this measurement is in Fig.1. With the exception of POP_P@150deg, the plot is as expected. We decided to fix HWP_angle = 110deg.

Then we checked the cameras in different configurations: MICH_locked, Xarm_unlocked (but aligned), Xarm_locked, Yarm_unlocked (but aligned), Yarm_locked. In all those configurations, the optimal HWP_angle was found to be around 110deg (for X/Yarm_locked the beam fluctuation was quite large and was difficult to find the optimal point precisely).

Finally we took pictures of POP and POP_P cameras in the mentioned configurations, as well as the power on each camera (note that between POP and POP_P camera images there is a left/right inversion).

For X arm some clipping seems to happen which makes POP image very strange. In both X and Y arm the power on POP_P decreases when the arm is locked and in Yarm POP image the beam shape seems to improve when arm is locked.

The updated POP layout can be found here: https://gwdoc.icrr.u-tokyo.ac.jp/cgi-bin/private/DocDB/ShowDocument?docid=9623

[Matteo, Hirata]

We added label on the QPD cables and on the installed optical components.
Attached some pictures.

[Yamamoto, Matteo]

The ASC model was modified to take into account the two new QPDs on POP for the forward beam.
Here a screenshot of the new channel connected to adc:

which are connected to the k1asc/ASC model and two new DCQPD blocks are added as in the next picture:

The PIT and YAW channels of each QPD are connected to the relative input PIT/YAW matrix.
The new model has not been compiled yet.

Two more DCQPDs plots have been added to the MEDM screen ASC_TRANS_POP_QPD.adl

[Akutsu, Ushiba, Matteo]

Abstract:
We manually offloaded IP2_PIT and adjusted IP1_PIT and IP1_YAW offsets. The mode matching after the work is 95%.

Detail:
Since IMC-PZT2_PIT_OFFSET value was 10 and therefore close to edge of the range (0,150) we manually offloaded it. This was done with IMC locked without ASC and the lock was maintained through all the procedure (mostly). After the manual offload the new value of IMC-PZT2_PIT_OFFSET was 75. Then we addressed IMC-PZT1_YAW_OFFSET which was 142 so also close to the edge of the range. This is less problematic since IP1 is not inside any alignment loop but we addressed it in any case. Before manually offload it we tried to change the value digitally while IMC was locked with ASC engaged and that improved the transmitted power. Given that, manual offload of IMC-PZT1_YAW_OFFSET was not necessary.
We kept adjusting IMC-PZT1_PIT_OFFSET and IMC-PZT1_YAW_OFFSET with IMC locked and ASC engaged while simultanously decreasing the values of IMC-REFL_QPDA1_RF14_I_PIT_OFFSET and IMC-REFL_QPDA1_RF14_I_YAW_OFFSET (original value were 1200 and 4000 respectively) and monitoring DOF4_P and DOF4_Y outputs to avoid railing. In the end we managed to reduce IMC-REFL_QPDA1_RF14_I_PIT_OFFSET and IMC-REFL_QPDA1_RF14_I_YAW_OFFSET to 0 and improve the modematching to 0.9497. The mode matching is computed as the ratio of LAS-POW_IMC_DC_IN1 and LAS-POW_PSLOUT_OUT.
The final values are reported in the table.

The values for IMC-REFL_QPDA1_RF14_I_PIT_OFFSET and IMC-REFL_QPDA1_RF14_I_YAW_OFFSET have already been implemented in imcparam.py.
Futher improvement of modematching can be possible by fine tuning IMC-PZT1_PIT_OFFSET and IMC-PZT1_YAW_OFFSET.

Attached a picture of MC_REFL camera. 

Continuation of klog #20128 and klog #20287.

[Matteo, Akutsu, Ushiba]

This morning we finalized the alignment of all the optics for POP forward beam.
I will update POP optical layout at a later stage.
The beam size is around 300um (radius) for both QPDs but QPD1 is in the far field while QPD2 is on the waist.

QPD1 is connected to ASC0 (ch 20-23).
QPD2 is connected to ASC0 (ch 24-27).

At the moment no interface is present.

Continuation of klog #20128.

I installed L1 (f=500mm) on the POP forward beam and characterized the beam after L1. The results are in fig1.
The beam wais is consistent with simulation and but beam waist position is 5cm far from simulation position. I suspect I forgot to add 5cm shift in the code. In any case this wll not have any impact for next part of the work.

Tomorrow I will continue the installation of L2 and QPDs.

[Ushiba, Matteo]

Abstract:
We adjusted the IMC input beam direction to maximize the IMC transmission.

Detail:
We changed K1:IMC-PZT2_PIT_OFFSET from 21.9 to 10.0 and K1:IMC-PZT2_YAW_OFFSET from 56.6 to 60.0. K1:IMC-DOF4_P_LIMIT was changed from 20.0 to 10.0 to avoid issues after the modification of K1:IMC-PZT2_PIT_OFFSET.
We updated the value of K1:IMC-REFL_QPDA1_RF14_I_PIT_OFFSET and K1:IMC-REFL_QPDA1_RF14_I_YAW_OFFSET to good values: (1200, 4500)
The alignment could be further improved by adjusting PIT_OFFSET but that caused DOF4_P to rail so we abstained.

Necessary work:
K1:IMC-PZT2_PIT_OFFSET value range is from 0 to 150 so at the moment it is very close to the edge. This means we should recenter this value by manually adjusting PZT2_PIT.

[Akutsu, Matteo, Ushiba]

After recovery work of IR Xarm (klog #20265), we went to Xend to center the IR QPD on TMS-X. Note that QPDA1 is currently not aligned so should not be considered. The good one is QPDA2.
The alignment was performed by using the QPD's translation stages.

After alignment of QPD was done, thanks to Ushiba-san, the IR Xarm was better aligned using the dithering system, This allowed for longer lock which allowed a good power spectrum measurement of K1:TMS-X_IR_QPDA2_PIT_OUT_DQ and K1:TMS-X_IR_QPDA2_YAW_OUT_DQ. The spectra can be seen in attached fig1 and are saved in /users/Commissioning/data/AOS/2022/0328/TMS-X_IR_QPDA2.xml

In the spectra are visible several peaks identified as ITMX and ETMX dithering lines and suspension resonances.
In the following table the values of several of them with computed coupling. The PIT2YAW coupling seems to be between 2.3% to 2.5% while the YAW2PIT coupling ranges from 1.8% to 2.7%.

[Yasui, Yokozawa, Matteo, Akutsu]

Abstract:
IR Xarm alignment has been recovered.

Detail:
Alternating the use of PR3 target and checking beam position around IFI and REFL, we recovered some very weak flashes of Xarm by adjusting the position of IMMT2 and PR2. At this stage the IR Xarm transmission could reach at maximum 0.4. Then, we closed all the vacuum chambers and put sinusoidal excitation on IMMT2 and PR2 to further improve the alignment. By selecting good values, the Xarm transmission could reach at maximum 1.4. This was enough to try to lock. The Xarm could lock, but was not very stable due to the still not perfect alignment. The alignment was later improved engaging the dithering system by Ushiba-san.

[Matteo, Hirata, Haoyu]

Continuation of klog #20086.

After IR was back on POP with good alignment, we continued the installation of optics for forward beam QPDs. First we corrected the position of periscope and already installed mirror to minimize overlap with other optics, still the space is very tight. Then we added additional mirror up to the lens position and measured the beam in the position in which the lens should be installed. The beam shape is in the attached picture and the beam size is as follow:
Width x: 4.890 mm
Width y: 4.349 mm

The beam is very well collimated.

[Matteo, Haoyu]

We tried to measure the beam shape of the single bounce reflection from ITMX at POP. The beam shape was ugly and most definetely not gaussian. Initially we suspected some clipping at the level of POP but we found none. We measured the beam before the periscope but it was still ugly. A location where clipping might happen is at the gate valve between ITMX and BS. We might need to check the situation after this gate valve is open (July?).

We also rotated the half waveplate which controls the splitting ratio before s and p pol cameras. The idea was to minimize the power on p pol camera. This is to ensure that the image collected at POP reflects the conditions at ITMX. After rotation (which indeed changed not only the total power but also the shape of the s and p pol images), we increased the exposure of p pol cam (K1:CAM-POP_P_EXP). New value is 34310.

Tomorrow we will try again to measure the beam shape in s and p to compare it with simulations.

I aligned the periscope for the forward beam on POP. This and other optics were roughly placed previously (klog 19226) but not aligned due to the absence of good IR beam.
The attached pictures shows the situation of the periscope and mirror after it. Since forward beam is very close to greenX beam the space between the two periscope is very small.
This lead to the need of installin the mirror after the periscope in the "wrong way" inside the optical mount.

I will continue the alignment work tomorrow.

I added a second cable from the rack, connected to the driver, same length of the previously installed cable (5m) which follows the same route to POP.

Memo: both cables follow a large boundle of RF cables to POP so in the future might be important to check the noise and if better routing is possible.

I measured the OLTF of the tweeter part for FIB_X/Y. The excitation point used is EXC B on each common mode servo board.
The results are attached. In both cases the unity gain is around 3.3kHz.
For FIB_X there is a feature at 59kHz, probably related to tweeter resonant frequency, while for FIB_Y, a similar feature, even if much smaller, is present at 65kHz.

[Matteo, Ushiba (control room)]

Continuation of klog 20010.

Today I investigated the reason why previously the shutter was not responding remotely. I operated it in manual mode and confirmed that the hardware was working properly, then reverted back to trigger mode and "primed" the controller (as requested by controller manual). Prime procedure requires to turn the wheel up when a open request is issued remotely. After this operation the shutter restarted working properly.
Then Ushiba-san aligned the Xarm but since the lock was not stable due to non perfect overall alignment, he misaligned ETMX to have single ITMX bounce. In this condition I aligned the DC of the photodiode. Since the amount of light reaching the shutter was very small I manually rotated the position of HWP2 to an angle of 200deg (nominal position is 214deg) which allowed some more power downstream. Both WFSs are now aligned and the total numer of counts is approx. 350cts for K1:ASC-REFL_QPDA1_DC_SUM_OUTPUT and approx 570cts for K1:ASC-REFL_QPDA2_DC_SUM_OUTPUT (see fig1).
Then I engaged the DC centering which seems to work properly (no deep investigation, just checking by eye that the beam is kept in the center).
Finally I swithed off the DC centering and closed the lateral panel of REFL.

Here I report a 5minutes strech of ALS:FIB_X to investigate its behaviour.
The system unlocked the first time since the fiber transmission dropped below 0.3 (threshold for relock), the second time since the WOOFER output exceeded 4V (threshold for relock are +/-4V) and the third time was manually switched off. In all the streches it is visible that the thermal controll cannot offload the woofer output DC and that the woofers output has several sudden jumps (expecially visible in the third lock).

The UTC time to use to get this stretch is 22 03 08 05 39 13 and length 300secs.

The channels are:
K1:GRD-ALS_FIBX_STATE_N
K1:ALS-X_ARM_INPUT_OUTPUT
K1:ALS-X_FIB_WOOFER_OUTPUT
K1:ALS-X_FIBER_TEMP_SERVO_OUTPUT

[edit]
In fig2 the same channels with the addition of K1:GRD-ALS_PLLX_STATE_N
It seems that also for FIB_X the big jumps are caused by PLLX unlocks. No explanation for small jumps.

I checked a stretch of 25minutes of lock for ALS:FIB_Y. The system remained lock for almost the entirety of the stretch, but the woofer signal had several big jumps. After a bit of investigation, it seems that all the jumps and/or unlocks are caused by unlocks of PLLY, which it seems to be caused by IMC_ALS which is kicking the main laser frequency.

GPS time for the stretch is:　1330825095　and length is 25mins.

Channels are:
K1:GRD-ALS_FIBY_STATE_N
K1:ALS-Y_ARM_INPUT_OUTPUT
K1:ALS-Y_FIB_WOOFER_OUTPUT
K1:ALS-Y_FIBER_TEMP_SERVO_OUTPUT
K1:GRD-ALS_PLLY_STATE_N

[abstract]
I tried to align the reflected beam while Xarm is locked to the IR WFSs on REFL.

[detail]
The reflected beam from Xarm (when arm locked) was very miscentereded at the level of HWP2 (see KARAAGE REFL). A picture of it is fig1. I realigned HWP2 (fig2) and moved TFP2 and BD just after HWP2 (fig3). I aligned the beam to the center of the shutter, but I could not open it from remote, so I started investigating why. The reason was that the power supply of shutter (same as HV driver for piezo and board of picomotors) was disconnected. I connected it to SC-3 12 (fig4) and the electronics restarted to work. Unfortunately the shutter did not responded to remote commands and kept closed. I will continue tomorrow after understanding how to open the shutter.

Today I installed four additional heaters (two on FIB_X and two on FIB_Y) to increase the actuation range of the thermal control.
At the moment there are four heaters for each fiber and they are connected in couples to the CLIO coil driver: channel 1 and 2 are connected to FIB_X and channel 3 and 4 are connected to FIB_Y. Since the old cable for FIB_Y was taped to the ground, I took the opportunity to rerute this as well. Now all the cables follow the cable rack on the side of the cleanroom. I will investigate later the impact of the additional heaters.