Reports until 15:18, Thursday 19 December 2019
H1 GRD (OpsInfo)
camilla.compton@LIGO.ORG - posted 15:18, Thursday 19 December 2019 - last comment - 19:31, Thursday 19 December 2019(53994)
Changed location of LASER_PWR = 2W in ISC_LOCK Guardian

Jenne, Camilla

In my 2019-11-26 lock acquisition (alog 53492), after quitting initial alignment I unintentionally started (and struggled) locking with the laser power at 10W.
We found that this was because after ISC_LOCK requested laser power of 2W, the ALIGN_IFO guardian was still in MICH_BRIGHT state requesting 10W before it was pushed to DOWN. Image shows at 05:29:44 UTC, ISC_LOCK in DOWN (0) state then laser power reducing to 2W before returning to 10W.
To stop this happening in the future, the LASER_PWR = 2W request has been moved in ISC_LOCK from DOWN to PREP_FOR_LOCKING, after nodes have been but to DOWN.

Operator: When next out of observe. Please reload ISC_LOCK Guardian.

Images attached to this report
Comments related to this report
thomas.shaffer@LIGO.ORG - 19:31, Thursday 19 December 2019 (53999)
Link

I worry about this during a lock loss. With this change, we will have have to get through our DOWN state with swinging optics and 38W injected.

What happened for Camilla of the 26th was unfortunate timing. ALIGN_IFO had bailed on a MICH_BRIGHT attempt and started to run DOWN, then ISC_LOCK got the request to go to DOWN setting the power to 2W and doing some other stuff before then requesting ALIGN_IFO to DOWN. In the space between ISC_LOCK requesting 2W and requesting ALIGN_IFO to go to DOWN, ALIGN_IFO made a request to go to 10W. The LASER_PWR node had to finish its first request before it could execute the second. ISC_LOCK had a user message saying:

"2019-11-26_05:50:48.578006Z ISC_LOCK [DOWN.main] USERMSG 1: LASER_PWR: REQUEST CHANGED (was: POWER_2W, now: GOTO_POWER_10W_FAST)"

I think this is a rare case that has a message already associated with it and does not warrent a change.

I've reverted ISC_LOCK.py from 20765 to 20763, and copied the change to ISC_LOCK_Camilla.py if others feel like the change is necessary.