Continued to this post.

1. Demodulation phases were re-tuned.
Just to duble check, the demod phases for each quadrant (such as K1:IMC-REFL_QPDAx_RF14_SEGx_PHASE_R) were re-assessed. MCE L was excited at 5 Hz and checked the peak heights of I and Q signals on diaggui. They are mostly OK and several minor changes were applied. Now most of PHASE_R is 90 degrees both for WFS1 and 2.

2. Closing the loops
With the new demod phases, I manually measured the sensing matrix quickly and decided the input matrices (screenshot attached). DC centering loops are all closed for all the measruements. For P, MCE and MCI+O (comnmon) are degenerate and could not close separately. Probably it's OK as they actually apply the same optical axes change on the IMC. As for the record, gain and filter setting (all p1 plus p0:z1 as an integrator) are written in guardian as a new state "ASC_RFON." However this state is skipped in the guardian sequence not to bother the downstream alignment. We'll implement them permanently once the downstream alingnment is confirmed with the loop closed.

3. Some notes
- Probably because the spot position is very high at MCE, compared with at MCI and MCO, MCE loop needed (DOF1_P and DOF1_Y) needed some offsets.
- DOF3_P, which is MCI-O differential, also needed offsets. Not sure why. IMC TRANS gets slightly better this way. Maybe the entire alignment is not ideal yet.
- Therefore the control configurations may be changed in the future with better alignment (or after the vent!).
- I left the loops running, and when the lock is lost and re-locked, the ASC RF part will be skipped by Guardian.
- Snapshot is saved but not as "IOO burt." just as k1imcasc burt.

4. Still to do
- UGFs have to be measured, and more filter and gain tuning can be done for the faster loops.
- Noise measurement
- WFS calibration
- Compare with the simulation (I feel the sensing matrices are healthy).