S. Dwyer, J. Kissel

The third attachment shows one of the fast glitches that has been causing our large drops in range.  This glitch in particular nearly saturated the ESD DAC, and we think that some of our sudden locklosses could be caused by these glitches when they are slightly larger.  

Since the frequency of this glitch is above 1 kHz, we decided to try to roll of the DARM loop a little faster, to make it more similar to LLO's.  We copied the soft elliptic low-pass from LLO's violin notch filter.

The following filter has been installed into FM9 of the DARM1 bank (replacing an old, unused RG3 filter)

Name: LLO_LP

Design String: zpk([200-900i; 200+900i; 2000], [250-500i; 250+500i; 500+700i; 500-700i], 1, "n").

(see attached BODE plot of the new filter.)

With this new filter, the new DARM OLG phase margin was reduced from ~42ish to 35ish (see attached OLG).

we lost lock right as we were measuring the changed open loop gain transfer function, so it's yet unclear if this improves anything.