ASC-MICH loop works after ASC-AS_B rephasing (it looked nonsensical before). I have no idea why they changed. AS_A phase was OK though not great, no adjustment was made.
Before making any adjustment:
I refined PRMI alignment manually, locked PRMI and measured the response of ASC-AS_B_RF45 used for MICH. I also measured AS_A just because.
In the first attachment (AS_B_RF45_VS_AS_A_before_phase_adjust.png), I'm misaligning BS in PIT both ways without enabling MICH nor PRC1 ASC feedback (but did enable DC3 and DC4 to center the beam on AS_A and AS_B) to test how AS_A and AS_B RF45 output respond.
MICH ASC error signal (AS_B_RF45_Q_PIT, yellow) is useless as it has a very small response to BS misalignment and it never crosses zero. This is what Elenna, Louis and I observed yesterday, too. In comparison AS_A_RF45_Q_PIT looks more promising (but should not).
In ASC-AS_B_RF45_phase.png, I'm injecting PRCL (not angle) dither at 80Hz to PRM M3 stage TEST_L and see the line in on each quadrant of each WFS. The idea is:
- PRCL signal must appear in I phase (dashed lines) not Q (solid) for AS_A and AS_B quadrants (left panel, top is AS_B, bottom is AS_A) as well as the LSC sensor (H1:LSC-REFLAIR_A_RF45, the right panel).
- For each WFS, all quadrant signals in I phase MUST have the equal phase relative to the PRCL error signal at the dither frequency (middle panel).
As you can see, AS_B doesn't make sense. For all AS_B quadrants, PRCL line shows stronger in Q than in I, and the phase of the third quadrant is flipped.
ASC-MICH works after rephasing.
I made the following changes to roughly minimize PRCL line in Q AND to make the phase relation of all I signals the same relative to PRCL error:
| before | after | |
| H1:ASC-AS_B_RF45_SEG1_PHASE_R | -59.698 | 10 |
| H1:ASC-AS_B_RF45_SEG2_PHASE_R | -31.5 | 40 |
| H1:ASC-AS_B_RF45_SEG3_PHASE_R | 146 | 20 |
| H1:ASC-AS_B_RF45_SEG4_PHASE_R | -80 | 10 |
The phase numbers themselves look more similar to each other now, PRCL line in Q phase is much smaller in all quadrants (ASC-AS_B_RF45_phase_after_adjustment.png), and most importantly, ASC-AS_B_RF45_Q_ERR responds to the MICH misalignment sensibly with good amplitude and it crosses zero (AS_B_RF45_VS_AS_A_after_B_phase_adjust.png).
Dark offset adjustment might be needed, though.
Anyway, after this adjustment ASC MICH loop worked fine with ASC-MICH_P_GAIN and Y_GAIN of 0.5. See ASC-MICH_works.png.
Demod phase values were accepted in sdf.
ASC PRC1 is not working yet.
According to Elenna, she's having a hard time finding a good PRC1 sensor in PRMI these days. She's been trying to use POP_X (because REFL sensors stopped working after the last vent, and because we haven't phased REFL WFSs for which we want to use CARM for phasing).
Since the POPX signal level is fairly small in PRMI, I set the dark offset (and accepted in SDF), locked PRMI, enabled DC6 loop to keep the POP_X centered, and moved PRM in PIT and YAW manually. Turns out that PIT never crosses zero, it comes close but then stops changing there. YAW looked OK in that at least it crosses zero.
When turning the servo ON with an integrator, YAW doesn't run away but PIT does, see POPX_P_not_working.png. In the same screen shot, you can see that the 80Hz PRCL dither appears in Q phase not I for all quadrants. I temporarily changed the phase so the dither shows up in I not Q, but PIT still didn't cross zero so I reverted them back.
| Original (current value) | Minimize PRCL in Q (reverted to original value) | |
| H1:ASC-POP_X_RF_SEG1_PHASE_R | -21 | -70 |
| H1:ASC-POP_X_RF_SEG2_PHASE_R | -15 | -90 |
| H1:ASC-POP_X_RF_SEG3_PHASE_R | 18 | -70 |
| H1:ASC-POP_X_RF_SEG4_PHASE_R | 10 | -90 |
