To investigate the detuning stability, we fixed the homodyne angle to squeezing quadrature and measured CCFC error signal and FDS three times with half an hour interval (attached figures). As you can see, the CCFC error signal and FDS are stable and the detuning fluctuation is ~2Hz.
The strange thing is that although the CC detuning is almost optimal, the CCFC error signal with demodulation phase of 23deg seems good detuning for FDS.
A similar measurement was done in elog2350, I and Michael got quite similar result even with a time scale of 1 hour. However, if I remember and understand well, Aso-san had a comment that different homodyne angle may affect our measurement, which is what we have seen. It seems important to understand why detuning becomes not so stable if we measure with different homodyne angle.
However, the same measurement may be interesting to be made for without CCFC but with AA/pointing.