I have looked at the few minutes of data with DARM locked using B1p PD1.

Figure 1 shows that DARM is coherent with MICH and SRCL between 20Hz and 50Hz (MICH and SRCL are coherent with each other). MICH and SRCL show prominently the bump at 157Hz which is related to an SIB1 mechanical resonance and laser frequency noise.

Figure 2 shows that MICH between 20Hz and 50MHz is also very coherent with DAQ FMOD 56MHz, which is a pick-off from the main input beam after the IMC. It shows a wide shelf between 10Hz and 60Hz, and the high bump at 157Hz. This spectrum of DAQ FMOD 56MHz was already present during the DARM DC read-out lock of a few weeks ago. And it was also coherent, but there was also demodulation phase on top in MICH a few weeks ago.

Figure 3 shows that these features are also coherent with the IMC length correction.

My guess is that the 56MHz modulation frequency is not well matched to the IMC length. And that an Fmod Err tuning would help the situation.