[Aso, Ushiba]

Abstract:

We performed P2L decoupling of PR3 by usign PRCL signals to reduce the offset of PR3 ADS.
Also, we reduced the ADS frequency at 7.1 Hz, which is below the UGF of PRCL loop at OBSERVATION state.
After that, we teted severa configuration of PR3 ASC but do not fully succeed yet.

Detail:

<P2L decoupling of PR3>

We excited PR3 in L and P, and compared the peak hight of PRCL error signals.
During the measurement, normalization by POP90 signals is hold not to induce fluctuation from POP90 signals.
Frequency and amplitude of excitation are 7.1 Hz and 4 cnts for both DoF, respectively.
Figure 1 shows the results (See details in the legend).
P2L coupling reduced by a factor of 4.

<ADS phasing and test at observation state>

After decoupling of PR3 actuators by using DRIVEALIGN matrix, we tested PR3 ADS.
We reduced dither frequency to 7.1 Hz so that L motion can be well supressed by PRCL loop, which has a UGF of 20 Hz at OBSERVATION state.
After changing dither frequency, demodulation phase was remeasured (fig2).

First, we measured the spectra and transfer function from LO to POP90 signals with 0.125 Hz BW but failed due to 0.1 Hz sidebands in the signals.
We are not so sure what this sideband is, but probably it causes the low coherence when exciting YAW for PR3 ADS (klog24766).
So, we measured the spectra and TF with 0.01 Hz BW, and then the coherence at 7.1 Hz is high enough for measuring the relative phase.
According to the measurement, demodulation phase was decided as 21.1 degs.

Then, we tested ADS at OBSERVATION state but still PRFPMI alignmet became bad after engaging PR3 ADS and WFS somehow.

<PR3 self-BPC>

Since the purpose of PR3 ASC is just fixing PR3 angle (reducing the drift of PR3), it is not necessarily ADS and WFS.
So, we tested BPC for fixing PR3 angle.
I used 50.1 Hz and 45.1 Hz with 1 cnt CLK_GAIN for pitch and yaw, which are used for PR2 BPCs (klog25180), respectively.
Figure 3 shows the phasing results of PR3 BPCs.
Obtained demodulation phasees are -129 and 48 degrees for pitch and yaw, respectively.

After deciding demodulation phases, we closed PR3 BPC by using PR3 itself.
Figure 4 shows the time series data of transmission powers, BPC error and feedback signals, and several OpLev signals.
Loop itself seems stable but gradually transmission power was reduced.
Since the earthquake happened and lock was lost, I couldn't check more long-term stability today.