We had some problems engaging the SRC ASC after maintenance day today.

• In ENGAGE_ASC_FOR_FULL_IFO (which is at 2W input power) while the SRC1_P loop was converging the buildups and alignment got terrible until we lost lock. It seemed like a phasing problem, and indeed when we added and offset (-5, comparatively large) to the SRC1 pit loop so that the error signal was zeroed at a nice alignment, we no longer had this problem.
• Before powering up I tried opening the loop, removing the offset, and re-phasing AS_A_RF72, but the phase difference needed to minimise the offset in the Q-quadrature when the alignment was good was huge (~70 degrees), and when I tried to close the loop again we lost lock, so I undid my rephasing.
• Next time we tried to lock, we acquired with the SRC1 offset, but when we powered up to 5W the alignment ran away and we lost lock again
• Next time we locked at 2W with the offset in, and as we powered up to 5W I ramped the offset off as the power was ramped up. This seems to have done the trick and we made it to NLN.

I don't understand why we needed this offset at 2W and not at 5W. But as a temporary stopgap I will put it in the guardian, switching it on at PREP_ASC_FOR_FULL_IFO (line 2475), and switching it off at INCREASE_POWER (lines 3391 and 3392, hopefully no typos this time, sorry Jim!)

Patrick and Travis also noted that ALS_COMM required human intervention a few times while locking, the offset needed to be tuned up by hand and at least once the x-arm green alingment required tweaking in order to complete the FIND_IR state.