We were able to close the MICH/SRC1 loops (for BS/SRM alignment) using the new AS72 signal today. Other ASC loops engaged at the mean time were INP1 and PRC2 (for IM4 and PR2).
The loops were first engaged at 00:15:00 Jul 09 2018 UTC in the DRMI config. The offset in SRCL for avoiding SRC mode hopping was also zeroed.
We then intentionally brought the IFO to DOWN and tried to engage the DRMI ASC (by hand) directly using the new AS72. The time stamp for this re-engagement was 00:50:00 Jul 09 2018 UTC. The guardian input matrix in the PREP_DRMI_ASC stage was modified to enable AS72.
=================================================================================
Details:
We start from the GRD state 'PREP_DRMI_ASC'. We first optimized the optics' alignment by hand to maximize POP18 and minimize POP90. We further tuned the BS and SRM alignment by zeroing the old AS36 error signals. Once this process was done, we phased the AS72 WFS such that for each quadrant the DC signal showed in I-phase. The Q-phase signal should then be free from error due to spot centering but contain only the true wave-front distortion.
We then measured the step response. For YAW the values were
| YAW response | A_Q | B_Q |
| BS | 0.29 | 1 |
| SRM | 1 | 0.84 |
This was further inverted to yield the following sensing matrix:
-250 A_Q + 300 B_Q ==> BS,
-500 A_Q + 150 B_Q ==> SRM,
where the overall scale was set to match the original AS36 response.
For PITCH, the step-response matrix was
| PIT response | A_I | A_Q | B_I | B_Q |
| BS | -0.24 | -0.74 | 1 | -0.09 |
| SRM | 0.82 | 1 | 0.02 | 0.40 |
The SNR in the B_Q to SRM response didn't seem to be great so we chose the following sensing scheme
500 A_Q + 2000 B_Q ==> BS,
-200 A_Q + 150 B_I ==> SRM.
Including some I-phase signal in the SRM sensing made it not immune to spot centering. The lock in the DRMI (cold IFO) appeared to be quite stable, and we could decouple SRM/BS sensing better than the original AS36 scheme. Yet its robustness against thermal effects would require further tests.