Investigating some Run 10 DCA anomalies

Following up on a report of some unusual <DCA> measurements reported by Toshihiro in Run 10 (see his blog page, or ShinIchi's 2020-05-13 PWG presentation slide 2). One of the things to verify is whether the Booster Main Magnet (BMM) operation had anything to do with these observations.

One note to make from the outset: we could disable use of the TPC (via the TPC status database table) during these periods if we want to prevent them being processed in any reproductions.

Run 11086085:
There was an unexpected beam abort at the end of the run. The scalers dropped to zero, and what little data is there may have improper distortion corrections (SpaceCharge for Run 10 AuAu62 used the ZDC west singles rate). Here's Toshihiro's <DCA>, the BBC coincidence rate, and BMM currents during that run:

Run 11085031:
Something odd happened in the middle of the fill, but it's unclear what. Here's Toshihiro's <DCA>, the BBC coindidence rate, and BMM currents during that run, or fill (the run was from 8:34 - 9:00):

Whatever it was also changed the east-going and west-going particles differently, which means that the impact on the TPC was unlikely to be symmetric, and probably cannot be corrected without significant work (if it's even possible).

Run 11126010:
It appears that this run and runs 011 and 012 that immediately followed it were started before injection of RHIC completed. In the plots of various BBC scalers I show here, the black curve is all the data over the time of the 5 fills containing these 5 runs, and the red points and curve highlight the data during the 5 runs. The east/west singles rates ratio appears to be a good indicator that things are notably different during injection.

BMM currents were constantly on this whole time, even after injection, so they cannot be correlated with the problem:

We don't actually apply any SpaceCharge distortion correction for these runs (during Run 10 AuAu7 running). But I've shown on other occasions that data acquired during RHIC injection can be problematic, so it is probably not worth spending the effort needed to correct such data (if it's even possible).