Figure 1, shows the usual noise budget, with a simple frequency independent coupling of laser phase noise to the B1 DC power fluctuations in pink. In cyan is PRCL noise coupling into DARM, this comes from SSFS noise entering into PRCL sensing and then being reinjected by the PRCL loop.
Note that the measured strain from DARM (blue curve) has a dip at 100Hz, this seems to be the internal noise budget calibration going bad (DARM line hidden by SSFS noise injection?), the official h(t) (black curve) doesn't have this feature.
Figure 2, using the transfer functions from the noise budget, adjusting their gains slightly and adding the PRCL (blue) and SSFS (red) noise coherently allows to have a nice overall projection (yellow) that matches very well the measured h(t) (purple), including the dips at 180Hz and 230 Hz, and the wider bump at 300-400Hz, coming from the interference between noise coupling directly from SSFS to B1_DC and indirectly through the PRCL loop.
This would confirm that the direct optical coupling of SSFS to B1_DC/DARM is frequency independent. I do not remember if that corresponds to finesse or losses asymmetry.
