Sheila, Peter, Danny, Stefan, Jenne

We have been having intermittent locklosses when transitioning DARM from ETMY to ETMX since I made the changes to the UIM filter described in 46861. This looks like a crossover instability between the PUM and ESD at 4.5-5 Hz.  It is interesting that this only seems to happen early in the lock, either in the transition or in the first 15 minutes of the lock, if we get through that then the interferometer can be stable for many hours. 

Since we only have this problem when actuating on ETMX, not ETMY, we took some measurements to compare the actuators. The first two attachments show the comparison of an excitation at drivealign to CALDAELTAL without the whitening corrected for.  The EX PUM has 20% more gain at 5 Hz than EY, but nothing in these measurements can account for an instability. 

Yesterday Peter F and I looked at the DARM filters from LLO, at LLO they have a larger relative gain between the PUM and the ESD, we have more boosts in the DARM filter bank here which means that we end up with a shallower cross-over between the PUM and ESD.  We tried (unsucsesfully) to transition DARM to a configuration that would get rid of the low passes and high passes that we have in drive-align and simplify our loop design. 

Instead of trying to debug this simpler configuration, we decided to make a small change to the loop which was to add a boost in the PUM lock filter (3rd png attachment shows the filter). The overall loop before and after the change is shown in the first two pdf attachments.

We attempted to measure the PUM crossover after this change, but had difficulty getting  a coherent measurement, even though our excitation was a couple orders of magnitude above the noise in both In1 and In2.  The third attached PDF shows the crossovers before and after this change. 

We also took a Pcal sweep after making this change and updating the actuator model.  It is saved in the usual place.