AIEC Wish list

• CDC Make sure that the initialization is scaling the gcf to its mean pressure equivalent, if it is not doing that already. So far all the inits have been done at mean pressure or close to it.
  • CDC - need to scale 2018 gains up (gcf down) by 6% & retrain.
  • CDC - look more closely to find what the correction should be. See if we can match beam current & radiator (ideally amo) & pressure.
  • Separate the EPICS readings & rethink the averaging times.
    • Use the wait and see approach for the HVB current - a few seconds of beam-on time is probably good enough.
    • Use a longer lookback for the temperature, because the readout is coarse and the value is important for the UQ. We would have to use the mean instead of the max - maybe 1 minute?
    • Consider what to use for pressure. Maybe 1 minute is good?
  • Set up a standard series of tests for the raw gcfs from a new model, and also for the UQ corrected data. Start with a book of plots and possibly automate looking through the book.
  • Discuss an option for this is it or get more data, when to use auto-off and should there be an auto-on.
  • Plot predicted std dev from the model vs the difference between predicted and traditional gcfs.
  • Re-enable the HVB current when we have the wait-and-see measurement AND when we have a model trained with the corrected 2018 data.
• FDC - correlation between gcf & pressure was good for 2020, inconsistent for 2018. Torri needs to talk to Lubomir.

• BCAL - baselines and crate temperatures. Need an executive status summary.
• Other detectors - for capstone students. Make a list of useful calibration tables & detectors. Naomi will ask the experts for a list of tables.