UPDATE(5/15)

All 100 segments totaling 1M events have been looked at, and the following is the correlation plot between the mean ADC and the number of triggers.

 The cluster trigger

The first 500,000 events from this run are analyzed. Configuration was "production2009_200Gev_Single / Physics"

For 829 cells in Eleanor's table, and total of 23729 triggers came from those cells. (based on Qtdata)

Hottest 25 cells produced 48% of the triggers. Module break downs are; NL:10, SL:8, NS:3, SS:4

Total 170 cells produced 90% of the triggers.

The non-LED FMS fast trigger rate has gone up to ~10% from ~5% before the access.

Another 21450 triggers came from the perimeter cells that are on the ignore list. The portion of the triggers that are "unofficial" seems to be growing as the hottest cells that have been cooled down lately (I think) tended to be part of the official trigger. My understanding is that the "unofficial" cluster sums are formed as if the high tower was in the middle of the qt8 strip, so they are mostly very close to what they should be.

The clusters are generated from Qtdata. These clusters differ from the "appended/internal" L1 DSM input clusters in two aspects. 1. Qt8 sums are NOT truncated. 2. The perimeter high towers are ignored according to Eleanor's table. 

Total of 40 clusters, one for each Qt32 board, are compared to the threshold.

LED suppression is done by looking at the number of channels with ADC>10, and cutting at 800.

For the following plots, the origin is South Up.

Cluster sum distribution. (after low 5 bit suppression)

Trigger multiplicity

Attached spreadsheet shows number of triggers, pseudo-rapidity, and phi for each cell. Sheet 1 is for all the cells that have cluster definition in Eleanor's table. Sheet 2 is for those that do not, including both the physical holes and the trigger ignores. For this, the number of trigger is based on appended L1 input.