Calibration lines for arbitrary (up to 4) UIM, PUM and up to 2 TST stages, handling in both front end and pyDARM epics generation

Currently LLO is using both ETMX and ETMY UIM and PUM stages to feed back DARM (along with just ETMY TST).

The front end calibration model can handle calibrating all UIMs, PUMs and TSTs, but can't actually handle the necessary TDCFs that would be needed to measure them all those stages individually. In fact, it's setup to handle exactly 1 line per stage (UIM, PUM, TST).

In talking with Valera, he wouldn't rule out the possibility of running with all 4 PUM stages if that helps gets better low noise performance, for example. So instead of adding a bunch of configuration to handle 1 to 4 lines for each "stage" that might (or might not) happen, would it be better to go all in on the combined actuator concept?

Currently, we assume our compensation of the individual coils for an individual PUM is good enough that we can just treat the drive signal as going out to a single actuator which doesn't have individual coils drifting in a frequency dependent way. We can extend that concept to all 4 PUM stages, as they are supposed to have very similar frequency dependent transfer functions, and if they don't, we would want to craft the digital actuation filters such that they do.

Taking that view, we should be able to take a single oscillator and drive 1,2,3 or 4 PUM stages, while also applying correction in GDS for a single overall measured "super actuator" PUM stage. This also simplifies GDS and DCS pipelines, as then they don't need to potentially sum 10 different actuators instead of the 3 we do now (and which I've been faking given we have 5 right now).

So I propose changing suspension models and CALCS to drive a signal from CALCS, that is then sent via Dolphin Reflected memory to all 4 quad suspensions and is controlled by a matrix or local switch whether it's injected or not. Local demodulation back into CALCs would need to take into account the extra delay, but makes the setup generic to arbitrary numbers of actuators at each stage.

Thoughts?