Abstract:
I checked the beam shape when PRMI was locked with 1F and 3F by using sensor card at POP table.
Then, I checked the RF PDs (PDA1 and PDA2) and found that beam on PDA2 was miscentered and hitting at the edge of the effective area.
After centering the PDA2, I tried ADS for PRMI by feedbacking to IMMT1, IMMT2, and PRM.
This time, POP 90 bexomes over 3, which is the highest value at this moment.

Detail:
I engaged ADSs in the same was as in klog23042, tried to lock PRMI_3F but again AS becomes bright.
I checked the signs of control signals but MICH_3F and PRX_3F can be locked without any trouble, so the sign of the control should be fine (If my understanding is correct, sign of MICH and PRCL for PRMI_3F are the same for MICH_3F and PRX_3F, respectively).
So, I suspected that the alignment of PRMI.

Since it is difficult to see the beam shape with the POP SPOL gigE camera when locking PRC because of maybe birefringence, I directly check the beam with sensor card at POP table.
Figure 1 and 2 show the beam at POP forward path whe locking PRMI_3F and PRMI_1F, respectivey.
Both beam seems fine and seems to be locked with TEM00.

After that, I checked the RF PDs (POP PDA1 and PDA2) and found that the beam on POP PDA2 hit the edge of the PD.
So, I moved the mirror (RBS1, see optical layout) to center the beam (although it is put at the common path for PDA1 and PDA2, not other choice for centering PDA2...).
Then, I also moved RST3 to recenter the beam on PDA1.

After beam centering, I tried ADS for PRMI.
I used IMMT1, IMMT2, and PRM as mentioned in the initial alignment procedure review (JGW-G2214532-v1).
After locking PRMI 1F and engaged ADSs, POP 90I signals becomes 3 (fig3), which was about 1 in the last time (klog23042).

ADSs used this time were implemented in the ASC_LOCK guardian and can be engaged by requesting PRMI_ADS_ENGAGED when locking PRMI with 1F signals.

Note:
Figure 4 shows the good oplev values for PRMI today.
Even though the POP 90 signals becomes large, fluctuation of POP 90, POP DC, and AS DC was not so small.
It may come from the MICH contrast, and it might be better to use BS instead of IMMT1 for PRMI alginment in that case.
We need to investigate more.