Calibration factors of the length QPD: 0.002471 mm/counts (vertical), 0.002451 mm/counts (horizontal), see attachment for the calibration graph.
Other parameters can be found in 8481, 8417 and 8430.
After a few rounds of trial and error, I got:
rotational angle for tilt QPD: 2.5 °
rotational angle for tilt QPD: 3.5 °
beam off-center: 0.02 cm (+V) and 0.083 cm (+T)
With these correction implemented, the diagonalization matrix is:
| TILT_PIT | TILT_YAW | LEN_PIT | LEN_YAW | |
| L | -0.02692811 | 0.16327656 | -6.39818178 | 0.38816218 |
| P | 1.74084729 | -0.06811129 | 2.92579829 | -0.17750109 |
| Y | 0.09504626 | 1.95077516 | 0 | 0 |
I also fixed a subtle mistake (the effect of this mistake is that it will introduce some Y2P coupling) in my code that generates these matrice. So, please see other entries for the corrected diagonalization matrice for SR2 and SRM (didn't have time for correcting that of SR3).
I turned the previous L2P cancellation filters off and changed all TM coil output filter gains back to unity. They are likely to be invalid.
All cross-couplings are likely to be well below 1%. I didn't have time to confirm it with more averages. See attachments for the TM transfer functions.
I have also implemented damping filters for the TM.
