Yesterday I got a chance to look at the phasing of REFL 45 WFS again, and to measure the sensing matrix for the REFL wfs with a new phasing. In the end, these phasings were reverted so we could go back to running without recommisioning the loops.
I drove a line in CHARD pit, and phased the 45 WFS quadrant by quadrant to minimize the CHARD signal in I.
| current phaseR | phase R to minimize CHARD in I | difference | |
| A1 | -165 | 140 | -55 |
| A2 | -175 | 155 | -30 |
| A3 | -45 | -25 | +20 |
| A4 | -50 | -20 | +30 |
| B1 | 115 | 85 | -30 |
| B2 | 125 | 75 | -50 |
| B3 | -115 | -105 | +10 |
| B4 | -110 | -100 | +10 |
The phases of the CHARD pit signals have the appropriate phases for PIT with both the current settings and the ones that minimize the signal in I. The motivation for doing this was to get a sensing matrix that would allow us to distinguish between CHARD and PR3 motion. After rephasing I drove a line at 8 Hz in CHARD, IM4, and PR3 and measured the transfer function from the excitation to REFL WFS. For reference, I drove 3 Counts in CHARD P EXC, 100 counts inPR3 M3 drivealing P2P, and 1 count in IM4 M1 drivealign P2P. The amplitudes of the lines were choosing to give good coherence, and I have not calibrated this measurement in any units, so only comparisons down the colmuns in this table are meaningful:
| CHARD | PR3 | IM4 | |
| A 45 I | -43dB,161° | -38dB, -75° | -4dB, -1° |
| A 45 Q | -34dB, 151° | -37dB, -170° | 14dB, 1° |
| A 9 I | -11.5dB, 173° | -30dB, -3° | 18dB, 0° |
| A 9 Q | -24dB, 170° | -49dB, 7° | -11dB, 173° |
| B 45 I | -52dB, 96° | -36.4dB, -80° | -10dB, 3° |
| B 45 Q | -37dB, 131° | -32dB, -170° | 14.6dB, 2° |
| B 9 I | -12dB, 173° | -28dB, -2° | 16dB, -180° |
| B 9 Q | -33dB, 174° | -40dB, -2° | 8dB, -180° |
With these phasings, we should be able to use something in our input matrix like:
A9I+B9I-2*(A45I+B45I) to CHARD which would reduce the coupling of PR3 motion into CHARD which could be causing us some grief (alog 20523.) Currently we use only A9I+B9I.
A45I+B45I for PR3 should be a better option than it used to be for the PR3 loop, although we may still need some 9I terms to cancel CHARD.
We can try closing loops around these new phases and adjusting our input matrix next time there is a chance.