I just turned off the length input to MC2 from the ISC.  The shutter is closed during the PSL work, so there is no light, so we have (apparently) just been sending nonsense signals to MC2 for a long while now.  The top mass's integrator has been on, so we've been saturating the DAC for a while also as a result.  I cleared the integrator, so now MC2 can actually damp.

We should probably add this as an option to the guardian - remove MC2 actuation, without misaligning MC2 (which is what is done in the current Offline state, to prevent flashing of the cavity).