There are a few things about the way that the suspensions are controlled or the user interface that could be easier for users, based on observations over the last few days of alignment and PRMI work..
- When we misalign or align ITMX the analog watchdog trips, MN H1 coil, and perhaps also H2 and H3 (I don't exactly remember which coils but it seems to be the same coil on MN each time).
- I'd suggest adding indicators for watchdog trips on the VIS overview screen. For type A suspensions there is a verbal annoucement in the control room, but some visual cues could also be helpfull. For other suspension types, it would be nice to add a verbal alarm as well.
- Also, it is confusing that the medm screen for the beam splitter looks like the nalog watchdog is tripped when it is actually not connected.
We had difficulty aligning the beamsplitter:
- The beamsplitter optical lever loop seems to have a ugf around 10mHz, so when one attempts to align it by moving the set point of this loop we either wait a long time between steps or we get confused.
- We also attempted to turn of the OL DC loop and align by changing the opticalign setpoint, however the guardian seems to set the optic align gain to 0 and not restore it, so that we end up clicking a button that isn't actually movig the optic for longer that I would like to admit :) I don't unerstand why the guardian would be setting this to zero.
- Tonight, even after the DC oplev servo was disengaged, the optic resonded very slowly (40-50 seconds response times) to changes in the opticalign set point. This isn't much faster than using the opticaling set point.
- Kieko and I tried to align the beamsplitter by turning off the test mass optical lever damping, but it seems that the IM loops don't do a very good job of damping things. Masayuki told us that he manually turns off the boosts in BS TM DAMP FM5 when aligning, then engage it while locking.
- Adam has been working on an medm screen to clearly show all of the things which are responsible for alignment (that would show when optic align has gain, when the DC loop is on or off, ect) This kind of screen with a nice visual summary of important information for all the optics can be very helpful.
- It seems like it would be possible to simplify the management of alignment setpoints. If one opens the DC loops and align by hand, then there are two places where the alignment needs to be updated to avoid having it overwritten by the guardian, the DC optical lever set point as well as the "good" DC optical lever set point.
A final suggestion is to check the tuning of the damping loops with simple step responses. It would be easier to align if when we move the optics, they move and stop swinging within a 1-2 seconds.