is calculated on all trigger levels and in the offline event selection programs. At the first level trigger the calculation is based on calorimeter trigger towers which each contain several calorimeter cells. The result is therefore less accurate than the calculation of at higher trigger levels where the full readout information is available. Moreover the calculation is based on uncalibrated detector information. The cut imposed is . Figure 5.1 shows the distribution of for the Monte Carlo data sample with and without the FLT-cut as well as the ratio of both distributions, the turn on curve for the FLT: The FLT trigger turn on is quite slow, only at about is it fully efficient for charged current events.
The second and third level trigger calculation of is more accurate because here the full readout information and granularity of the UCAL are available.
The DST selection programs have the calibrated detector information available. Furthermore is calculated using the reconstructed vertex which yields the most accurate results possible.
Figure 5.2 shows the distribution of for random events triggered in the unpaired and open bunches before and after the first level trigger cut. There are many events at close to zero, which are due to empty triggers by the FLT. These events are rejected at the SLT.