Matt, Keita

As noted in entry 5279 we had some trouble with changing MC2 acuation after the MC2 model change/restart.  It is not clear if the state before the change was the "right" one, and no explanation for the change was found (all MC2 coil driver bits were flipped, with no apparent change).  This should be looked at again with a model of the system in hand so that the measurements can be compared with expectations.  Since the "new" MC2_M2 TFs make the L1 filters stable, I am inclined to guess that we have returned to the "right" state.

In this state I changed the M2 stability filter to be very simple (zero at 2Hz, pole at 8Hz).  This gives the M2 cross-over resonable phase margin for all UGFs below 10Hz.  The attached images show the MC2 TFs with the M2 gain set to zero (no M2 actuation), M2 gain set to 0.03 (generally enough, and good for locking), and M2 gain of 0.1 (good for ISI work or high seismic times).  The maximum "good" value of the M2 gain is about 0.2, more than 0.3 is unstable.

Note that in the TFs involving MC2_M3, this system is working with heirachical control, so the measurement includes M3 and M2 (e.g., the whole MC2 actuation, or combine MC2 M2 and M3 actuator).  These TFs show that the M3 gain has about 8dB of gain margin, so gains more than about -2000 are likely to cause problems (-1000 is the nominal value).