Looking into the suggestion that the pedestals are affected by the ripple and
hence it would be more sensible to subtract a common pedestal for all timebuckets on a given anodes. This will increase the sampling of a given pedestal by a factor of 128. Below I've looked at the average pedestal as a function on timebucket for barrel 1 ladder 4 for 3 different pedestal files. The 4 plots are for
4 different hybrids on the ladder. You can see that there is a distinct shape to the pedestals as a function of time which is similar for the first 3 plots and inverted for the last plot. I put this down to a remnant of the ripple. The
last hybrid is read out on the other side of the ladder to the first 3 and hence should see the inverted signal. This has been observed before by Dave and is put down to some capacitive coupling along the ladder frame. The fact that the shape of the pedestal varies from run to run tells you that this is not a true
shape driven by the pixel pedestal but is a result of the ripple not being washed out thoroughly by the event sample we take.
This similarity in shape is emphasised in this plot where I've draw the pedestals on top of each other. You can clearly see that hybrids pedestals
oscillate in phase or are inverted.
One small aside, it seems that at least for the recent data things get worse after timebucket 64
Last modified: Thu Sep 27 13:56:17 EDT 2001