Sheila, Ryan S, Oli
We spent some time yesterday looking at the 05/31 16:45 UTC lockloss, which was caused by a couple of large earthquakes. As the ground started moving more and we got closer to losing lock, DIAG_MAIN was telling us that IMC-REFL_SERVO_SPLITMON was close to saturating (>8V), and we also got some SRM saturations. We wanted to look into what the actual cause of the lockloss was (we weren''t able to confirm that) and see if we could change anything wrt the SRM control loops or IMC SPLITMON in order to make the detector more robust against earthquakes.
We first looked at the length, pitch, and yaw of SRM M3 DRIVEALIGN to see which of the three was contributing the most drive during the earthquake (attachment1). All three have a frequency of around 57mHz. Multiplying the max L/P/Y DRIVEALIGN values reached right before the lockloss by their respective EUL2OSEM values, we find that L was 100,000, P was 52,380, and Y was 7,333, so we will want to offload some of the driving on L.
We also looked into the saturations in IMC-REFL_SERVO_SPLITMON that occured prior to the lockloss (attachment2, attachment3). The IMC-REFL_SERVO_SPLITMON trace shows it saturating multiple times, with the saturations sometimes lasting as long as 3 seconds. We compared SPLITMON to some other relevant channels - I can''t explain why these other channels are relavent, but Sheila said that we might want to talk to Daniel about reducing the load on IMC-REFL_SERVO_SPLITMON.
I looked into how often SPLITMON saturations seem to occur, and they almost exclusively happen when we have high ground motion from earthquakes or wind(attachment4). I found that there have only been two of these ground motion SPLITMON locklosses in O4b (05/14 and 05/31), but before O4b there were many more, usually more than one per month and mostly seen in clumps. I also checked which of these locklosses from the beginning of O4 were accompanied by SRM saturations, and was only able to find 7, with yesterday's being the only one this year.
Ground motion + SPLITMON + SRM saturations during O4 so far:
2023-06-16 11:48 UTC
2023-10-02 00:54 UTC
2023-10-07 08:16 UTC
2023-10-11 01:43 UTC
2023-10-15 04:38 UTC
2023-10-16 03:18 UTC
2024-05-31 16:45 UTC
