LockLoss :
https://ldas-jobs.ligo-wa.caltech.edu/~lockloss/index.cgi?event=1377686651
I just happend to see the PI's ring up this morning durning my shift and started to trend them back to see just how high PI 24 can go with out us losing lock. While just goofing off I discovered that just last night, literally the last lockloss, PI 24 had run up from 20 to 2213.5 and caused a lockloss in under 12 minutes. screenshot included. The cursor T1 cursor is set at the point were PI 24 reaches 20. T2 cursor is set to GPS Time 1377686651 which is the Lockloss timestamp.
PI FOM plots over the time LockLoss time. Top plot is log scaled.
We think that the PI guardian did not have enough time to evaluate the correct damping phase for mode 24 (10.431 kHz PI) so we updated the PI guardian as we did in alog69800. We increased the phase step from 30 to 60 deg, the average time to calculate the new rms from 1 to 5s, and the timer for each step from 5 to 10s. The PI guardian is not reloaded yet and we asked the operator to reload it when lockloss happens.