Update on tracking residuals

From GlueXWiki
Jump to: navigation, search

Improvements to the Kalman Filter tracking for the central region have been checked in:

  1. Treatment of the stereo wires has been simplified, thereby fixing (for as yet unknown reasons) small round-off errors in the calculation of the doca.
  2. Maximum step size in the time-based phase was reduced from 1 cm to 2 mm. I also modified the code doing the extrapolation to the position of closest approach to the beam line. These changes slow down the code but eliminate the bias in phi: left plot shows the "normalized residual (blue)" (formally known as "pull" in my previous presentation") and the sigmas for this distribution (in red) before the changes, right shows the results after changes. Both plots are for 0.1-3.1 GeV/c pions thrown from the center of the target.

Old residual.gif New residual.gif

These changes seem to marginally improve the sigmas for q/pT as well:

Old residual pt.gif New residual pt.gif