Signal Processing for the ATLAS Liquid Argon Calorimeter : studies and implementation

The ATLAS detector has been operating successfully in the LHC\nbeam line studying the proton-proton collisions produced with an\nenergy of up to 8 TeV in the center of mass. In the peak luminosity,\nmore than 36 independent collisions happened every 50 ns. Given the\nrarity of some of the processes a larger rate of\ncollisions is being envisaged.\n\nThe ATLAS detector has a complex set of calorimeters measuring the\nenergy of different types of particles. The Liquid Argon calorimeters\nwork by ionization of their active material and the free electrons are\ncollected by a high voltage applied to electrodes within the liquid argon.\nThe generated current signal has a well known pulse shape. Multiple\nsamples of the analog signal are collected and digital signal\nprocessing techniques are used to extract the pulse amplitude\ndirectly correlated to the energy deposited in the detector. The\nanalog pulse is relatively long (up to 400 ns) and the probability of\nminimum bias events happening in the same detector area during that\ninterval is quite high. This pile-up effect can jeopardize the\nquality of the results and a technique called optimal filter is\nused to minimize its effects.\n\nIn the next data taking periods, however, the increase of luminosity will\nbe much more important, leading to a discussion on new techniques\nthat could mitigate the impact of such pile-up induced noise.\nSome of these techniques work in free-running mode, which can be applied,\nas well, for trigger purposes.\n\nThe paper will discuss some of these new techniques\nas well as the implementation of these in new ATCA standard technology.\nA prototype using the processing power of FPGAs is being prepared for studies\nand may be used when the LHC return to operation in 2015.