The shift was dedicated in recovering the stability of the 480 mW step of the CARM offset reduction and in investigating the unlocks that affect the West arm ALS since a few days.

ALS glitches:
Most of the shift was affected by random unlocks of the West arm ALS. The unlocks were (except for a few notable exceptions noted in the paragraphs below) related to glitches (similar to the ones affecting the North arm ALS in the past weeks) appearing at different rates. During some parts of the shift (mostly at the beginning and after 19.30 UTC) they  appeared every few minutes or even every several tens of minutes (allowing a couple of CARM offset reduction attempts), while during the period between 16.00 and 19.30 UTC the rate of glitches increased up to a glitch every few seconds, resulting in continuous lock and unlocks of the green beam (see figure 1). The reason of these unlocks is still under investigation by the ALS team.

Other ALS unlocks during the lock acquisition:
During the unlock investigations, the reason a relatively common kind of unlock was discovered. This unlock happens in two possible steps of the lock acquisition: either when engaging the boost filter after the handoff of the arm control from the IR beam to the ALS beams, or later while increasing the power just after the handoff of CARM2MC to the IR beam. During these steps, the power of the transmitted arm IR beam (B7 and B8) goes from below to above 1 mW, and this triggers the enabling of the SNEB and SWEB IR tracking loops, causing a transient in the SNEB/SWEB_LC_TX and SNEB/SWEB_LC_TY that sometimes affects the ALS lock stability, as can be seen in figure 2  and 3 . To avoid unlocks, the tracking loops should be disabled during the lock-acquisition steps that rely on the ALS beams.

Oscillations at 480 mW:
During the few successfull attempts of reaching CARM offset reduction, two different differential oscillations affected the interferometer at the 480 mW of transmitted power step, leading to unlocks. One was the same ~55 Hz oscillation encountered during the morning shift  (53247), related to an increase of UGF that affects the DARM loop after a few tens of seconds of lock in this state (see figure 4). Attempts of manually  or automatically (using the UGF servo) tuning the DARM GAIN in order to keep the UGF under control (around 35 Hz) had mixed results, since lowering too much DARM (below 0.02, from the current 0.025) leads to the appearance of low frequency ~1.8 Hz differential oscillations that equally distrupt the lock.

The interferometer has been left with the arms locked on the IR beam and the other nodes in DOWN.