Today we wanted to keep working on the issue related to BS AA TX and the 400 mHz oscillation arising when we switch to full bandwidth control mode.

Unfortunately the activity was a bit delayed due to the issue of the Neovan chiller (https://logbook.virgo-gw.eu/virgo/?r=65298). When the problem was solved we started again the lock acquisition. During the first period of the shift we unlocked here and there, probably due to the bad weather activity. After a while the robustness of the lock acquisition was recovered.

We wanted to go to LN3 with the BS TX in drift control, and eventually perform several test on the alignment loops both for BS and for the DET bench.

However, we didn't have the chance to perform any of the test we planned due to several unlocks while we were going towards Low-Noise 3.

One unlock occurred after the arise of a 250 mHz oscillation visible in the B1 PD1 audio channel, in correspondence of its saturation.

We decided to perform the subsequent lock by putting the engagement of the BS TX in fullbandwidth by default in the automation. To be noted that since yesterday we postponed its engagement at the beginning of LN3, so at the same time in which SR TY is misaligned to adjust the DCP.

The lock after if failed exactly after the engagement of FBW, due to the overcoming the safety threshold on B1s DC.

We noticed that in the initial transient of LN3 the fluctuation of B1s are usually higher, so we thought to post-pone even further the BS TX FBW engagement, after 90 second of SRTY misalignment, in a much calmer situation.

To do so we put in the automation a self.timer condition of 90 second starting from the beginning of SR TY misalignment. The modification worked and the FBW  engagement was succesfull. It unlocked shortly after anyhow.

OPEN POINTS:

• As far as we remember, in the past we were used to put BS AA in FBW before going to Low-Noise 2, before the BS MAR reallocation. But for some reason this was problematic so that's why it has been decided to put it after LN2. It should be worth to better investigate the optimal point of engagement of full bandwidth and see if there's the possibility to put it back in some earlier step (e.g. CARM NULL 1f ?)
• There's a second possibility which is also worth to be explored: there's the chance that the 400 mHz oscillation is somehow a loop instability due to B4 quadrants. Maybe the sensor is measuring some other cross coupling with other DoFs. So, when we blend B1p and B4, there's the chance that the plant we think we are using to build the loop is not the same for the two quadrants. In this case, we should dedicate proper measurement to better understand if this is the case.