Following up on the observtion that the SRCL FF is increasing the DARM low frequency (2-3 Hz) motion (73561) and that this motion might be the origin of non-stationary noise (73546), I designed a new high pass filter for SRCLFF and retuned the FF.

The new high pass gives us about 4 times less reinjection between 2 and 4 Hz, at the price of a phase rotation of 20 degrees at 10 Hz. Therefore the old SRCL FF was not performing well anymore. I did a noise inection and refit the FF filter. Fitting this filter is tricky and I don't think I got an optimal solution. But it looks like the new highpass with the new FF performs fine, is at least as effective as the old highpass and the old FF, by looking at the coherence between DARM and SRCL.

DARM RMS is improved a little between 3 and 4 Hz, so maybe we'll see some improvement in the non-stationary noise. We need to figure out what else is making DARM noise at 3.4 Hz. There is also an enormous line at 1.23 Hz that is the cause of most of DARM RMS. I think the bicoherence shows that the 1.23 Hz line is a source of noise modulation too.