RPC and DTBX Trigger Sensitivity and Efficiency

Of great importance is a test of the sensitivity of the trigger to predefined threshold values. As we need regions of similar geometrical to compare the two trigger systems we restrict our analysis to .

  
Figure 9: Trigger response : probability for (reconstructed) lower than nominal threshold (upper), and for 80% reduced threshold (lower) for RPC and DTBX.

In fig. 9 we show, how often the reconstructed will be less than the generated one - which could cause trigger losses. For RPC this occurs only in 10% of all events, whereas for DTBX we observe a dependence up to 20 GeV/c rising from to and than staying constant. Therefore if we set our trigger thresholds to only 80% of the predefined nominal trigger threshold value, we will find only a few % of muons in the RPC where the generated will be higher than the reconstructed . For the DTBX this is the case in 5 to 15%.

To understand this we investigated, how the generated will be seen and reconstructed by the trigger.

  
Figure 10: Trigger response : (reconstructed) versus (generated) for RPC and DTBX.

  
Figure 11: Comparison of trigger output in per event for DTBX and RPC.

In fig. 10 versus is displayed. Quite clearly there is a difference in the assignment between the two systems. Obviously the DTBX trigger makes a more isomorphic transformation from the generated to the assigned , whereas the RPC trigger tries to avoid any loss especially of high muons and consequently - also due to the rather coarse binning above 40 GeV/c - has a very high proportion of muons with  GeV/c. Therefore RPC will consequently produce a higher trigger rate than DTBX, whereas DTBX will be able to make a finer discrimination of transverse momenta, but needs a better tuning in order to get lower trigger losses.

If we require exactly one triggered muon in the RPC and DTBX systems (from one generated muon), we can compare different outputs of the two trigger systems on an event to event basis. In fig. 11 we plot the difference vs and observe it to be mostly less than 25 GeV/c, with the exclusion of a region of 40 - 70 GeV/c in . This will be just the effect of different binnings in of the two trigger systems in this intermediate region.

In fig. 12 and fig. 13 we give the efficiency curves for different thresholds fitted with an integral of a Gaussian. The fit parameters are a measure for the needed threshold values and the width (window) around the setting of the nominal trigger thresholds. The DTBX and RPC curves are shifted with respect to each other, as the nominal value of the trigger threshold should correspond with 90% probability to the required trigger value of the RPC, whereas this is designed to be 50% for the DTBX resolution.

  
Figure 12: Efficiency of the DT trigger for various thresholds.

  
Figure 13: Efficiency of the RPC trigger for various thresholds.

We also note, that RPC can reach a higher efficiency than DTBX. In the first place the reason is the rather high rate of not reconstructed tracks, in the second place the high proportion of tracks reconstructed in DTBX with a lower value than generated and therefore always lost for the corresponding threshold.