An overall hitfinding efficiency is useful, and mapping its effect on the particle-finding efficiency as a function of, e.g. and p, would provide a useful single particle efficiency. However, of great importance to HBT studies, kinkfinding studies, and the like, is the two-track resolution, which will be affected by the two-hit merging distance.
In Figure 18 are 2-dimensional histograms of the yield of hits on a padrow as a function of the relative seperation in the pad and tdc direction (although these units have been converted to centimeters). The left panel of Figure 18 is the relative seperation yield of all Monte Carlo padrow crossings, while the right panel shows the relative seperation yield of reconstructed hits. It is clear that padrow crossings that are very close together are not being resolved, and only one hit is found.
Figures 19 and 20 show the ratio of the reconstructed and generated relative seperation yields for padrows in the inner and outer sector, when deconvolution is turned off (TCLPAR.DECON=0) and when standard tcl settings are used. Although the efficiency is not a step function as a funciton of relative distance, the merging distance goes from about 3-4 cm to about 1.5 cm when deconvolution is turned on. Therefore, we would like to emphasize that, although the overall hitfinding efficiency only increases about 20% when deconvolution is turned on (going from 71% to 91% for inner sector), those hits that are recovered with deconvolution are exactly those that are important for, say, HBT studies.
Figure 18: Relative seperation yield in the time and pad directions for Monte Carlo padrow crossings (left) and for reconstructed hits (right).
Figure 19: The efficiency of finding a hit as a function of its spatial distance from another hit on the same padrow, when deconvolution is turned off in the hitfinder.
Figure 20: The efficiency of finding a hit as a function of its spatial distance from another hit on the same padrow, for default settings in the cluster/hitfinder.