In our runup to O3, we have been using the ESD on ETMX for both acquisition and low noise operation.  When we do the transition from acquisition to lownoise, we move all actuation away from ETMX.  In O1 and O2, this meant that we just moved all the actuation (L1, L2, L3) to ETMY, which was already setup for lownoise.  Now, we move L1 and L2 to ETMY, but the L3 actuation over to ITMX.  ETMX's ESD driver gets some analog switching, then we move the L3 actuation back to ETMX.  However, we have been leaving the L1 and L2 actuation on ETMY. 

Part of the reasoning for leaving the L1 and L2 actuation on ETMY was that we had hoped we'd be able to fix and use the ETMY ESD, and just go entirely back to the O2 sequence with all lownoise DARM actuation on ETMY.  This would save some calibration measurements from needing to be redone.  However, we will be using ETMX's ESD for O3, so it seems sensible to have all of the DARM actuation on ETMX. 

To this end, I have looked at the differences in actuation on the L1 and L2 stages between acquisition and lownoise, and determined that we already have all of the necessary filters installed on ETMX, just in a different order than they are installed on ETMY.  I have set up a choice flag in the LOWNOISE_ESD_ETMX state where we can choose whether to move the L1 and L2 actuation back to ETMX, or leave it on ETMY.  If the flag is set to False (which it is currently), then everything will stay as it has been.  If the flag is changed to True, then ETMX's L1 and L2 LOCK_L filter banks will be set up to look like the ETMY banks, and then all DARM actuation will be put back to ETMX at the end of the state (not leaving some on ETMY).  The ALS_DIFF state resets the ETMX L1 and L2 stages for acquisition.

Since I have not yet tested this, the choice flag is left False for now, so that no settings will actually change. Once we're able to relock, we'll try a transition with the final actuation only being ETMX.

The attachment shows the difference in filters for acquisition vs. low noise for the L1 stages.  The L2 stage difference is just engaging an inverse pendulum filter.