It turns out that the bad PDH signal caused by monitor daughter board (D040424) note being properly seated in its plug. It wasn't screwed in and I must have just bumped it out when putting in some test leads.
I have screwed it in properly now.
As pictured in PSL:1994, there was a burn mark in the center of the final lens before the north reference cavity. I've switch out this PLCX-25.4-77.3-UV-1064 for a new lens. Cavity is realigned with a Vmin,Vmax = 0.492V,1.56V, giving a visibility of 52% (about as good as it will get with this MM solution). Polarization is well aligned with the cavity basis.
Something is up with the PDH error signal (see attachment below), last time we checked this it was the textbook shape. Nothing has changed electronically, I need to double check if things are ok further up the path going through EOMs. North has been locked nicely for about 3 weeks now with pretty much 100 % uptime. Power into the cavity is 1.08 mW, so not a saturating thing. It could be AM modulation. I'll check systematically the EOM polarization alignment and that of the AEOM tomorrow. If that fails I'll look at the electronics.
I also blew some dust of optics around the periscope with the ion gun, this may need to be repeated a few times as there is a lot of dust on the table and will take a while to purge out of the tent after each blast.