I did a little work on the squeezer ASC today, although I haven't gotten to close any loops because we have been having lots of locklosses.
With anti squeezing injected, I phased the signals into I, which doesn't matter since we take the quadrature sum of I and Q, but it makes it easier to watch what is happening. The settings for the WFS matrices were wrong, now that I fixed them the signals respond as we would expect. It looks like the output matrix that Haocun found for the DC loops will work well for the 42 signals as well, with the pos loop controlling AS A and the ANG loop on AS B.
The squeezer automation has worked well today, in most locks the squeezing is injected and locked automatically. One time the LO loop was not locking, and I had to reduce the LO IN1 gain from it's usual 7dB to -3dB, the loop locked, and I was able to step the gain back to 7dB.
FYI: The Inject Squeeze guardian on the ISC_LOCK will move on without closing the loop if RF3 level is below -25 dBm (if I remember correctly). When this happens it's an indication that we should check our alignment (or lock by hand if we have enough 3MHz signal).