Yesterday we measured again the MC2/MC3 crossover frequency. Again, tecnical problems prevented us from going all the way to the actual crossover frequency, but we were close enough to estimate confidently it to be close to 100 mHz.  See attached plot.

Because of the peak right above 1 Hz and the observed instability if we try to increase the MC2 path gain, Matt designed a new filter to replace the elliptical one and roll off the gain before the offending peak (basically a LP filter at 1 Hz with at attenuation of at least 20 dB at 1.2 Hz, where the peak is). We then increased the MC2 loop gain by a factor to (0.005->0.01)