DVCS data processing 10-Oct-2007

Sergey,

I've looking through some of the P0 events and have questions concerning how to define what a cluster is.

For events 67,104,197, and 279 there is a region where there are towers with a large charge, but there isn't anything that seems obvious to identify 2 individual hits (there is no dip then increase in charge as you travel away from the tower with the most charge). Are these events some examples of what should be identified as having 2 or more clusters? If so, why?

For event 133, there is a single tower hit (with a large charge), but has no adjacent hit towers. I guess that this is the other half of the cluster pair. If this is true, this means that clusters must be allowed to have only a single hit tower to be counted as a cluster (in attempt to not loose any trigger). Accepting this would mean a lot of false triggering for setting high discriminator thresholds, or very poor trigger efficiency for keep discriminiator thresholds low.

There are many additional similar events in the file, though most show 2 individual clusters clearly. Having to accept these types of events with the current cluster finding logic makes it hard to reject other events in the dvcs_047213_10000event file, which helps to explain why there was little reduction in the events for that file. This small reduction also held true in the 2 cluster case (see the attached excel file - there are 2 sheets, 1st for 1 cluster min, 2nd for 2 cluster min).

Please let me know what you think about these type of events.

Thanks, Ben