Injected a MICH or SRCL line at 45 Hz, tuned the MICH or SRCL feedforward gains to minimize the line in DARM.

Small improvement by changing MICH_FF_GAIN from -1 to -0.955. No improvement by chaning SRCL_FF_GAIN. Those gain changes have been absorbed into the Oct10 filter, so the nominal gains in the filter banks are still -1 and -1

The attached plot shows the effect of switching on the MICH and PRCL feedforward (beware that the reduction of the ~50 Hz / ~ 100 Hz bumps is due to closing the beam diverters)