(awesome compared to what we had in the past, but still only for 3 DOFs).

I started with H1:ALS-X_TR_A_LF_OUT of about 1.1, centered WFSs, engaged the servo, and it didn't work.

Then I manually aligned the arm well such that the transmission becomes 1.2-1.25, centered WFS, and it still didn't work. Since I wasn't sure if the cavity was well aligned when I measured everything before, I started from scratch.

I measured the demod phase, and the phase itself was not a disaster, but I've found that segments were totally imbalanced (I balanced them before). Fixed it such that the seg1 signal strength against length injection is matched to seg3, same thing for seg2/seg4.

Then I measured the sensing matrix again, inverted them, put them in the input matrix, and PIT DOFs started working (but YAW not).

After several rounds of enable WFSs, center WFSs, remeasure sensing matrix, I converged to a state where 2 PIT DOFs plus one YAW DOF that corresponds to PZT2 works wonderfully. Enabling 3DOFs would immediately bring the transmission up to 1.36 or so.

Lisa misaligned ETMX such that the transmission drops to 1.0-ish, we enabled the 3DOF WFS, and immediately the transmission recovered.

Don't know why the 4th DOF doesn't work but I'm leaving it as is. For now WFS is disabled, QPD centering is enabled. If you want to enable WFS, you first need to disable QPD centering by disabling the input or output of H1:ALS-X_QPD_A_PIT and such, as the WFS feedback is also added into the same PZT path and there's no convenient way to disable QPD while allowing WFS signal to be routed to PZT.