rolling bhabha calibration issues

quick recap, status at beginning of restart of data taking

-one technical problem (farm going into "freaky mode" after deleting temporal db content) was replaced by another (output of number of selected bhabha events wrong)
-solved by some manipulation from Andy Salnikov, problem due to the bugfix of the first problem
-nevertheless: accumulated data is usable
-but temporal db has not been deleted before startup :-(
-first calibration finished (successfully) May, 22nd
-last calibration finished last wed (June, 14th)

rolling bhabha calibration results - some information and statistics

-no problem after deletion of temporal db - problem seems to be solved :-)
-All crates stored, nothing unusual, everything fine
-12 channels with too low number of direct hits
-13 non validated channels
-27 dead channels and lots of next-to-dead channles (as expected)

rolling bhabha calibration results - comparison plots

-comparison difficult, because temporal db not deleted, but nevertheless possible, because:
-first calibration finished on May, 22nd and deleted the temporal db
-constants stored in the test condition, which is not used for next calibration - no change of actual constants
-second calibration calculated constants, relative to unchanged default constants, stored in the test condition again
-used first and second absolute constants to derive relative constants, multiplied to the first calibration
-can be compared with the relative constants of the offline calibration, used similar data sample
-left plot: zoomed absolute difference between the relative constants, right plot: unzoomed relative difference between the constants

-

-very good agreement, difference only 8.3 * 10^-4
-looked at the absolute difference vs. theta and phi:



-very homogene plot, channels with larger difference all in theta=56, except one:
-channel 6:80 failed the validation in second calibration, but passed in offline calib --> constants cannot be the same!
-alltogether it looks very good...
-Next steps:
-load new offline bhabha calibration into both conditions (once it's validated)
-delete the temporal db (this time, I'll double check) ;-)
-Hopefully have a first complete successful iteration...

rolling bhabha calibration is on a very promising way (again)