J. Kissel, O. Patane, D. Barker
ECR E1700387
IIET Ticket 9392
WP 11743

More details to follow in the comments, but today started the campaign of SUS / SEI watchdog improvements. This required us to restart the
    h1susetmx
    h1susetmy
    h1sustmsx
    h1sustmsy
    h1susetmxpi
    h1susetmypi
    h1isietmx
    h1isietmy
front-end models.

The update makes the following changes:
    (1) Establish a sensible band-limited RMS user watchdog trigger (where it had been some old broken RMS code prior), where the OSEM trigger signals are now calibrated into microns (um_RMS), and the threshold has been adjusted to match it.
    (2) Entirely removed the USER DACKILLs from the SUS, so we no longer AND together the output triggers for any stages' user watchdog and shut down the output of DAC. In other words, we'll no longer kill damping and offsets if we drive the L2 stage too hard with ISC control.
    (3) Entirely removed the connection between the ETM and TMS DACKILLs to the ISI and PI models, and replaced the input to the ISI's PAYLOAD flag with "always good" constants.

As originally promised in ECR and IIET ticket -- this drastically reduces the unnecessary complexity and redundancy of the watchdog system, it now behaves in a much less trigger-happy fashion, and delightfully, we now can answer the question "how much motion trips the watchdog?"

I've accepted the threshold at 25 um_RMS over the 0.1 to 10 Hz band. We can and should adjust this threshold as we gather more experience.

All changes to settings have been accepted in the corresponding SUS safe.snap such that the work in calibrating the OSEM trigger signal won't disappear with the ISC_LOCK guardian's SDF_REVERT state.