Daniel, Vicky
In FC2 suspension's M1 stage DRIVEALIGN matrix, we decoupled L2P (length-to-pitch) for long-term FC2 lock stability, because over an hours-long lock, we see big FC-ASC pitch offsets build up over a 4+ hour lock as the FC-IR loop feeds back to FC2 Length. At DC, most of this feedback goes to FC2's M1_L stage, so we decoupled L2A in FC2 M1's drivealign matrix.
After ~midnight last night (3/12-3/13 evening), when FC-IR unlocked, the filter cavity struggled to re-lock because FC-GREEN did not catch TEM00 lock again for >30 min. FCGS kept locking on TEM01 (pitch offset). When I cleared the FC-ASC pitch outputs, it relocked fine on TEM00, so likely related to the FC-ASC alignment offsets that built up over several hours of lock. At ~DC, the FC2-IR suspension feedback goes primarily to FC2_M1_L, which we see directly correlated to the pitch offsets that build up (yaw also to a much lesser extent). From the correlation between FC2_L sus feedback (1884) and FC_ASC pitch offsets (12) (1884/12~6.4) which is nice and linear over the ~hours lock, around noon today we set FC2 M1's L2P drivealign gain at +6.4 for decoupling. We also decoupled L2Y at +0.7, TBD if yaw is optimal but leaving that here.
It looks like this L2A (mostly L2P) decoupling possibly helps more than I realized. With L2P, our FC-GREEN locking keeps the cavity length much more stable now (right side of attachment) than on Sunday (left side), in the same GR_LOCKED state (green suspension+vco locked). Before (left side), the FC-IR signal would swing around cavity co-resonance by almost ~ir linewidth in ~seconds; fc length now seems held stably w/in an ir-linewidth.
With the long FDS lock stretches over the weekend 3/17-3/20, we can see the tidal drift in FC length over its ~300m baseline! Very cool! Tidal drift seems of order 10-20-microns, if we look at the teal 1st column, 2nd row, plot of FC2's M1_DAMP_L_INMON.
Fun screenshots of a long 18-hour lock w/FDS from 3/19/23 (here w/o blrms), and we can see long term tidal drifts of FC length!
