After discussion with the stakeholders and deciding that using the Welch was too elaborate for the ESS (the issue is supporting sufficient configuration and then picking suitable values). For now, at least, we have decided to go with plain FFT.
This update supports an arbitrary number of accelerometers per data client and also writes the raw-ish acceleration data, after applying offset and scaling using the ESS accelerometer telemetry topic. If you feel strongly that the data should be truly raw (no offset or scale), then please do note that here and/or in the review. I do see that the XML description really says "raw"; I just feel uneasy ignoring the given calibrations (especially scale; offset can easily be determined from the raw data).
I am asking Petr Kubanek to review the ts_ess_labjack PR and Craig Lage to review the new ESS configuration parameters in ts_config_ocs. Please feel free to comment on the other PR as well. Patrick Ingraham: feel free to weigh in as well.