As responce of klog34102, I investigated the related channels of PSL.

Segment 1 and 4 of PSL ALIGNMON QPD are saturated while PSL output is increasing. But, the behaivor at the moment of saturation seems to be somehow strange. Figure 1 shows the time-series of each segment of PSL-ALIGNMON_OUTPUT_QPDA1, PSL output (K1:LAS-POW_PSL2_OUTPUT), IP PZT feedbacks of IMCASC (K1:IMC-DOF{4,5}_{PIT,YAW}_OUTPUT). As PSL output power is increasing, powers on SEG1 and SEG4 are increasing, and then are saturated at last. On the other hands, the ones on SEG2 and SEG3 are decreasing. This indicates that beam position moves in Yaw direction according to their segmant location while the power is increasing. However, the IP PZT feedbacks seems not to move so much. IP1 PIT PZT feedback moves largely rather than YAW PZTs. On the other hands, before increasing the power, IP1 PIT PZT moved much larger than the amount in increasing the power but any segment amounts seem to be not changed. Therefore, the PZT seems to be not cause of the alignment change in the Yaw direction.

Anyway, the easy one of ways to solve the saturation is to switch the QPD gain from the higher gain to the lower one. According to fig.3 in klog20236, the gain switch on ALIGNMON QPD is toggled in the left direction from the view of the QPD front. In this state, the gain is selected to the higher one according to klog17268. So we can reduce the signal to 1/5 by changing the QPD gain.