Figure 1 shows the evolution of the Pstab and SSFS line couplings to DARM over ~1 week. The PSTAB1 (at 236Hz) and PSTAB2 (at 1501Hz) show the same trend while the LF PSTAB0 line (43Hz) has a different trend. The Pstab coupling at LF (PSTAB0) seem to follow the 56MHz LSB amplitude on B1p (see bottom right plot) or the sidebands imbalance, the correlation is not always clear. The HF lines (PSTAB1&2) seem to follow the 6MHz B1p power. This is clearer on Fig 2 which zooms on March 31st and one can see that the PSTAB1&2 couplings drops at the beginning of the lock following the 6MHz SB power.
Figure 3 shows data from Apr 7 when the 6MHz modulation depth was increased from 12 to to 15dBm (see 59687 and 59697). B5_DC clearly sees an increase of noise by a factor ~1.5 at HF while there is no impact on B1 (and no coherence with B5). So it seems that B5_DC sees the 6MHz amplitude noise (at least when the modulation depth is at 15dBm, while it might be limited by sensing noise at 12dBm). I tried to see if B5 could be used to do a noise projection by rescaling it using the Pstab2 line (see Fig 4) but it gives a projection above B1 spectrum so the mechanism is not straightforward (maybe due to the fact that the carrier and sidebands relative weights are different in B1 & B5).
Figure 5 shows also data from April 7th when the 56MHz was decreased from 13dBm (blue) to 10 dBm (purple). In that case there is no increase of noise on B5 nor on B1 (note that I didn't use the data taken when the modulation depth was further decreased as there was some clear induced misalignment (B1 OG and B1p power decreasing, see Fig. 6) ). The fact that B5 does not see extra noise could mean that the 56MHz RAM noise is lower than the 6MHz and/or due to the lower modulation index of the 56MHz.
Fig 1 also shows the B1 noise in the bucket (bottom 2nd plot) monitored by the BRMS from 120-130 Hz (region without lines/structures). It seems very stable over time.