I looked at signals from the suspension noise monitors in the H1 PUM coil drivers, to see if they typically saturate during lock (result: yes they do), and to see if Duo Tao's monitor board redesign would substantially improve the situation (result: yes it would).

Attachment 1: signals from the existing noisemons, recorded during a Jan 18 lock.  Not all, but about half of them are exhibiting saturation (which sets in at ~25000 ct).

Attachment 2: projected output of the new board with the same drive signals.  The projection was made using the recorded MASTER_OUT data, filtered with the known transfer function through the new board (with the coil driver in state 1), as well as the coil driver's "LP" filter (necessary because the drivers actually run in state 3, which is the most aggressive dewhitening).

Attachment 3: since the noisemons on ETMY and ITMY were not continuously saturating, I ran a quick check of their coherence with the drives.  It revealed excess noise in the ETMY L2 LL channel.  This is unlikely to be a problem for the H1 DARM sensitivity, due to the state 3 dewhitening. But it lets us improve the estimated fault rate to 2 out of 12 (counting also the L1 ETMY PUM channels previously measured).