The problem reported by Cheryl from the owl shift (alog 54008) is the same thing as we experienced before (alog 53965), 3MHz beat note degrades as SQZ ASC tilts ZMs, eventually ASC runs away, SQZ guardian resets ASC, beat note goes back to normal, and it repeats it many, many times (1st attachment). We had the same problem in the afternoon.

We first tried to relieve the ASC output by moving ZM2 YAW manually, but the runaway behavior only got faster.

Then we checked the dark offset of ASC-AS_A_RF42 and AS_B for all I and Q channels after closing the beam diverter and some of them were 20 to 30, which is significant as the sum channel is only 150-160 counts when the beam diverter is open. Usually an offset of 20 to 30 counts is not huge, but in this case the signal is very small, therefore we want to monitor this periodically.

After adjusting this, the running away behavior was gone. 2nd and 3rd attachment shows the sdf diff (except for ZM1 P offset because it was unmonitored, we changed the setting to monitor it).

Daniel also found that the NLG was somewhat lower than before.  H1:SQZ-SHG_LAUNCH_DC_POWERMON keeps creeping up (6.9 today VS 5.2 on Dec/17) though OPO TRANS is still 1.0, and CLF REFL 6MHz monitor was ~ -30.5 instead of -29.5-ish. Daniel changed the OPO temperature (H1:SQZ-OPO_TEC_SETTEMP) from 33.246 to 33.214 degree Celsius to gain 1dB or so for H1:SQZ-CLF_REFL_RF6_DEMOD_RFMON. That doesn't mean that degradation is OK, we have to keep monitoring.