According to the spectrum measurement results in klog30699, we might improve the RMS around 0.45 Hz by increasing the gain of damping loop.
So, I tried to increase the gain.

To do that, I first measure the TF from BFL to TML with 1mHz BW to confirm the resonant frequency precisely (fig1).
Then, I designed NB filters with 50% pass-band width of the target resonant frequency.
I also set a notch filter to avoid oscillation at resnant frequency at 0.5-0.7 Hz.

Figure 2 shows the OLTF before and after modifying the filters.
Since entire gain was increased by a factor of 2, gain peaking around UGF becomes larger but it seems accetable.

Figure 3 shows the spectrum of TML OpLev of PR2.
Red and blue shows the spectra with new filters and old filters, respectively.
Since overall spectrum becomes better, it is difficult to say the reduction of RMS is due to the filter modification but at least the performance is not bad.
So, I keep this filter for LOCK_ACQUISITION state.