View source for 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 [http://clonweb/wiki/clondocs/DVCS_Trigger/P0_event67.png 67],[http://clonweb/wiki/clondocs/DVCS_Trigger/P0_event104.png 104],[http://clonweb/wiki/clondocs/DVCS_Trigger/P0_event197.png 197], and [http://clonweb/wiki/clondocs/DVCS_Trigger/P0_event279.png 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 [http://clonweb/wiki/clondocs/DVCS_Trigger/P0_event133.png 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 [http://clonweb/wiki/clondocs/DVCS_Trigger/dvcs_047213_trigger_efficiency_oct10.xls 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