We had Lock-loss just after we handed the obs. shift over.
From 17:05, We requested "PRFPMI_LSC_LOCK". The state reached to "Observation" but the lock-loss happened a couple of minutes later.
We also tried "ALS_DARM_LOCK" and "PRFPMI_LSC_LOCKED_RF", however, situation was not recovered.
In these recovery process, we noticed followings;

(1) "ALS-Y_FIB_SEVRO_IN1_DQ" shows large sometime glitches. And its oscillation range in unlock state between 0 and -200. However, for X, it is between +1000 and -1000.
Input power of Y Green laser into interferometer is 0.6, which is also smaller than the optimum value.
We asked to Sugimoto-kun what values are correct and confirmed that behavior of "ALS-Y_FIB_SEVRO_IN1_DQ" is strange.
In "FIB Common Mode Servo Board" for Y Green laser, the "Common offset" value was changed to -16.0 from -5.06.

(2) MCO outputs (H2, H4) looks to be over the limit value of 10,000 in GUARDIAN in the process of "Observation".
This can be a reason of the lock-loss.

We tried lock process step by step with a bit interval in each process, and we successfully started "Observation" at 24:17 at last.
Then MCO outputs of H2 and H4 were about -9000 and +9000.
We changed limit value in GUARDIAN from 10,000 to 15,000.

The SDF status of ALSFIB was updated to new "Common offset".