After realigning the IFO, I lost lock a couple of times during ASC_PART1, due to PRC2_Y engaging.

Going slow the next time, it first looked like the IFO was more happy with the PRC2_Y input matrix of RF9I=0, RF45I=-1 .

But after about 15min the the input matrix from the guardian script - RF9I=0.7, RF45I=-1 , seemed to work just fine.

In the attached plot, notice that the RMS in all signals shrank significantly over about 15min.