Sheila D, TJ S

There has been a bout of SRM M3 watch dog trips around the SRY locking step of initial alignment over the last few weeks (alog81670 for example). Today Sheila and I discussed locking without the FE triggering at all, and just having the ALIGN_IFO node see good flashes and turn on the necessary filters and banks. I tried this out today but I wasn't able to get it to lock any more reliably than what we have now. No matter the method, it would seem to catch with low AS_A values, ~2500 on NSUM vs the normal 5000. From here, SRM would start to be driven before it would realize that it wasn't quite locked. I tried adding some code in ACQUIRE_SRY to try to turn on and off the SRM M1 LOCK L input and the SRCL FM4 filters and even clear the history if necessary, but this didn't work without long settling periods between attempts.

I ended up keeping the ISC_library.is_locked('SRY') looking as AS_A_DC_NSUM_OUTPUT > 4000. This value is a bit higher but safer. It will let the node run through Down and reset drives, integrators, and let SRM settle a bit. I don't think this is a fix, barely even a band aid. It will need some more thought on how to only catch on the correct mode.