Digital filtering performance in the ATLAS Level-1 Calorimeter Trigger

The ATLAS Level-1 Calorimeter Trigger is a hardware-based system designed to identify high-p_T jets, electron/photon and tau candidates, and to measure total and missing E_T in the ATLAS Liquid Argon and Tile calorimeters. It is a pipelined processor system, with a new set of inputs being evaluated every 25 ns. The overall trigger decision has a latency budget of ~2 μs, including all transmission delays. The calorimeter trigger uses about 7200 reduced-granularity analogue signals, which are first digitized at the 40 MHz LHC bunch-crossing frequency, before being passed to a digital Finite Impulse Response (FIR) filter. Due to latency and chip real-estate constraints, only a simple 5-element filter with limited precision can be used. Nevertheless, this filter achieves a significant reduction in noise, along with improving the bunch-crossing assignment and energy resolution for small signals. The context in which digital filters are used for the ATLAS Level-1 Calorimeter Trigger is presented, before describing the methods used to determine the best filter coefficients for each detector element. The performance of these filters is investigated with commissioning data and cross-checks of the calibration with initial beam data from ATLAS are shown.