Upsilon L2 triggered events analysis and the BEMC Preshower

This is a brief status report on inculding the preshower in the Upsilon analysis

We have used online pedestals for mimicking the triggers, today we will briefly study the L0 trigger and how much remians after cuts

We worked with the L0 trigger for a small sample of the Upsilon data in this study and wanted to investigate where the triggered events "get lost".

Using primarly Vertex Only we mimicked the L0 trigger from the ADC values by bitshifting the ADC by 4: ADC>>4 > 18.

• We scanned: 2190 L2 triggered events

We then made track assocaitions to the towers. The tracks were extrapolated to the center of the tower to improve the resolution

We investigated to see if there were any obvious hot tower cluster, this was not found to be the case, the towers look pretty well distributed

• The mimic L0 trigger condition with no additional cuts left: 724 tracks

Double counting

We then investigated the number of times a track pointed to the same tower after passing the L0 mimic

• 2 or more tracks pointing to a tower after L0 mimic: 195

this leaves (724-195)/2190= 24% of tracks passing the mimic

It was found that the additon of a momentum cut, of at least p>1, reduced the amount of towers with double tracks passing the mimc trigger

• Two track pointing to same tower with L0 mimic and p>1 : 2
• Tracks passing mimc with p>1: 430

This gives (450-2)/2190= 20%

Are there still false associations or tracks pointing to photon clusters? Making the PID cuts as well as tighter cuts on P should help eliminate these

dEdx and the preshower

The Dedx cut leaves of 3e-6 to 5e-6 leaves 212/2190= 10%

We are using a lose cut for now, as the L0 mimc was very sensative to the width of this selection

With the preshower and all other cuts we are left with 79/2190= 4%

No preshower, dedx, p>1 and mimc

Preshower, no dedx, all PID, p>1 and mimc

Preshower, dedx, all PID, p>1 and mimc

Conclusions

• L0 trigger did not seem to be affected by a "hot tower cluster"
• A lot of L0 triggers seem to be "lost" during tracking reconstruction
• Adding the condition that the L0 trigger tower must be in the event did not decrease the L0 mimc tiggers found, suggesting that the mimc is the same (or lower) than the real L0 trigger was
• We ran with a wider dEdx distribution of 3-5 as this seems to effect the L0 mimc trigger most, maybe the L0 was being contaminated with signal with dEdx
• In the L2 triggered events, approx. 5-10% of the samlple is good electrons with the L0!

Next

Next time we will show some results from the study of the preshower on the paired events