WAC Notes July 13 2021

From PREX Wiki
Jump to: navigation, search
RCDB HALOG HAPLOG DocDB Online Prompt BlueJeans Link Runlist spreadsheet, respin 1 spreadsheet

PREX Main << Weekly Analysis Coordinator << WAC Notes CREX << WAC Post-Run_Meeting Notes

WAC Notes July 6 2021 << >> WAC Notes July 20 2021


Date and time: July 13, 2021, 11am
Bluejeans ID: 564945377
Meeting link: https://bluejeans.com/564945377
Runlist spreadsheet

July 13th

  • Ryan gives an update about AT - Slides
  • Next week Cameron will talk about beam corrections and respin2 results

July 6th

No update, but Ryan will talk next week about AT stuff

June 22nd

Decided to turn these meetings into a series of thesis-work updates and long-paper technote presentations.

May 18th 2021

Comparing respin2 to 1, changes, etc.

  • Conclusions decided during the meeting:
    • Respin2 run list looks pretty good. Documented on the HAPLOG
    • Dithering vs. regression outputs from respin2 multipletwise plots look good - see HAPLOG
    • Aq wien state dependence actually increased a bit, likely due to the inclusion of more data in respin2 dataset
    • The known Aq wien state dependent large Aq immediately after a trip (obtained from cutting SinceLastTripEnd < 250) got much larger, which I suspect is due to Fewer fake-trips being included which had diluted the post-trip large Aq problem in the respin1 results (respin 1 had a bunch of mislabeled beam trips due to the bcm_an_diff burp cut being considered a beam trip)
    • There are no noteworthy main detector asymmetry OR diff_bpm dependences on wien state or SinceLastTripEnd (which is comforting)
      • There is a 1 sigma effect in the BPM 12X data, which could really be some sort of Aq pickup... our BPM pedestals are not better than 2% precise
    • The data-inclusion comparison between respin1 and 2 looks good. A lot of new data in CREX part 3 is found, partly due to fixing the bcm_an_diff burp cut, and also due to improving cuts everywhere
    • A pull plot of reg_asym_us_avg looks really nice
      • The only non-gaussian portions can be attributed to lower current running and/or to a few outlier events
      • It would be nice to do a weighted histogram filling, where the weight comes from the minirun-wise mean-error, which doesn't exist in the mul-plot data, but does exist in the aggregator, and would be feasible to add into these trees and see what kind of result we get that way


Paul, Cameron, Victoria, AJZec, Weibin, Eric, Ryan