Fortunately, we don't need to move the PR2/PR3. This zig-zag shape came from the demodulation phase, and after the phase tuning, the error signal got fine.
Here, I wanna emphasize that this phase tuning was performed with a great new tool to simulate the IFO response developed by Niigata Univ students.
I guess they will make a report about that tool's detail soon, so I just show a couple of time series plots. I attached the error signal of the PRCL with several demodulation phases. The Upper is the measured one, and the lower is the simulated one with that tool.
Basically, this was calculated by finesse, so the result itself is not really new. However, I could get this result extremely easily with this new tool. Let's wait for the release report!!