sheila.dwyer@LIGO.ORG - posted 16:38, Sunday 11 November 2018 - last comment - 18:44, Sunday 11 November 2018(45205)
LOCKLOSS_SHUTTER_CHECK guardian can be confusing when manually adjusting powers
Georgia mentioned yesterday that the fast shutter lockloss guardian was failing it's check. We saw this again today, after we powered up and then reduced the power with the interferometer still locked.
The problem here was that the shutter lockloss check guardian watches the transmitted arm powers, and if they go above a threshold it watches for the shutter to fire when we loose lock. We went below this threshold when we reduced the power again to do OMC scans, which caused the LOCKLOSS_SHUTTER_CHECK guardian to go into the failed state, which it stays in until a human intervenes by taking it to init.
The shutter has passed it's test a few times today.
This happened again, but even when we didn't decrease the power, this is probably more similar to what Georgia mentioned.
Attached is a screenshot of the guardian log, and the channels used to check the arm circulating power at the time.
This happened at Nov 12 2018 00:05:03 UTC. You would expect the guardian to jump from state 10 (HIGH_ARM_POWER) to state 20 (CHECK_SHUTTER) if the circulating power drops below it's lower threshold in either arm while the in the HIGH_ARM_POWER state. The lower threshold is 41000 counts on the TR_B_NSUM units plotted in the attached dataviewer, and neither channel drops below that threshold.