Performance of the TFU network of the First Level Trigger of HERA-B and determination of efficincies for J/ψ→μ+ μ-

The experiment HERA-B at the deutsches elektronen synchrotron (DESY) is designed to measure the violation of CP symmetry in the neutral B meson system. The relatively small cross section for bb production for proton collisions at 920 GeV together with the small branching fractions require a fast and efficient trigger system. The trigger system comprises four levels and is optimized to select decays of J/ψ → l+l-. The first level trigger (FLT) performs a fast online tracking which is implemented in a custom-designed electronics board, the track finding unit (TFU). The FLT has to cope with an input rate of 10.4 MHz and has to reach a suppression factor of 200. In total, the network consists of 55 boards. The performance of the TFU network is studied with respect to deadtime and trigger rate as a function of the interaction rate. The upper limit of the deadtime is 4.4% for an interaction rate of 5 MHz and 10.1% for a rate of 10 MHz. Under the trigger condition to require a single track the trigger rate is 20.7 kHz for 5 MHz and 52.9 kHz for 10 MHz proton interaction rate. The limit for the input rate of the Second Level Trigger is 30 kHz which can be achieved by operating a multiplicity veto at a loss of 5% of J/ψ events. The trigger rate at a higher interaction rate can only be reduced sufficiently by requiring instead of one track two tracks with opposite sign and an invariant mass of at least 2.0 GeV/c2. The FLT efficiency to find a muon track originating from a J/ψ → μ+μ- event is determined from data taken during the year 2000 and from a perfect Monte Carlo (MC) sample and a realistic MC sample that describes the detector performance of the year 2000. The single-track efficiency is 59% for perfect MC, 16% for realistic MC and 18% for the data. The efficiency to trigger a J/ψ event is 36% for perfect MC and 4% for realistic MC as well as for the data. These efficiencies are also determined with respect to the transverse momentum pTtrack of the track as well as the pTJ/ψ of the J/ψ and with respect to the Feynman variable xF. The analysis of the data shows a decrease of efficiency of the order of 36% in the first muon layer MU 1. A major fraction, (25-28)% (abs.), are due to a bad efficiency of the tracking layer. The inefficiency due to the algorithm is determined by perfect MC and results in 9.2%. (orig.)