I repeated the scans here:

drupal.star.bnl.gov/STAR/blog-entry/aliceb/2009/feb/16/timing-scans-barrel-preshower-shower-max-update

with tighter QA cuts on the pedestals.  There are significantly fewer bumps in the spectra.  While some remain I worry about losing more statistics with anymore cuts.

It still seems there is no significant dependence of either slopes or sums or "slope deviations" on the timing.