With the arm cavity locked I tried to measure again the WFS sensing matrix. It still looks almost degenerate in both PIT and YAW.

These are the measurements:

sensP =
   1.0e-03 *
   -0.2374    0.2929
   -0.1506    0.1611
sensY =
   1.0e-03 *
    0.0133   -0.1897
    0.0177   -0.2023

And these are inverted matrices:

iP =
   1.0e+04 *
    2.7373   -4.9782
    2.5596   -4.0341
iY =
   1.0e+05 *
   -2.9772    2.7915
   -0.2610    0.1953

The measuremetns are at about 4 Hz.

They look quite bad. I won't even try to impelment them.