Cheryl and I are doing a quick re-initial alignment starting with XARM IR, since the DRMI ASC keeps pulling us out of lock. (I'm hopeful that we'll get away with just doing this ASC, and not the green arms). While watching, I was curious why we seemed to be *railing* the PRM during PRX lock. The PRCL gain was set to -5250000, which seems more like a typo than a gain that usually goes in our LSC dofs. Looking through the svn archive, this change was made between the April 5th 2019 checkin and the April 23rd checkin of lscparams. There are no notes either in the inline comments or the svn log as to why this would be, since the previous value was -3200.
I changed the params value to be -5200, but by then we were moved on to other parts of init align. If Ed finds a problem with this state when doing initial alignment tomorrow, we can think about taking a quick open loop gain measurement.
UPDATE: We did an initial alignment starting from green arms, since we weren't able to get past TR CARM. PRX was unable to lock with the gain value of -5200, but was able to lock with -5200000 (-5.2e6), but this value railed PRM. So, I tried -5.2e5, and PRX was locked and PRM was not railed. So, we'll try this value for a while, but again if there are problems with the next initial alignment, someone could revert this value in lscparams.py.
Since this looks like a possible long night, want to make a note of where the value for this gain is (since I'm a novice with editing guardian scripts):
File:
/opt/rtcds/userapps/release/isc/h1/guardian/lscparams.py
Value in question:
'PRXY' (line #81 as of tonight) with current giganto value of -520000 (-5.2e5), BUT previously it was a larger -5250000 (-5.25e6).
My Note For Night:
Will be going through an alignment soon. And will try to do so WITHOUT the new INIT_ALIGN guardian node trying to step through everything automatically. But if I have issues, I will revert to the LARGER issue noted above.