Design and implementation of on-detector and off-detector electronics for the high-luminosity upgrade of the CMS endcap calorimeter system within LHC at CERN

The CMS experiment requires various technical upgrades to sustain its performance whilst coping with a harsher radiation environment and having to resolve a greater number of proton-proton interactions per beam cross-over, both brought by the increased luminosity projected for the high-luminosity LHC. One key upgrade is replacing the existing endcap calorimeter with a novel device, the High Granularity Calorimeter (HGCAL). This thesis will focus on two mutually related aspects considering the design and implementation of its electronics system. The first aspect is about a test system that prototypes a full vertical slice (start-to-end) of the detector electronics chain. As the final system, this includes both on-detector (front-end) and off-detector (back-end) components, as well as trigger and DAQ sub-systems. Key operational aspects including fast control, timing distribution, slow control, trigger and DAQ were activated and tested. This allowed operating the system in a beam test, measuring a data transfer bit error rate upper limit of O(1E−16), and evaluating the digital agreement between the different devices in the chain. These results mark the first-ever implementation of an HGCAL readout system prototype of such complexity. The second aspect involves the design and FPGA implementation of two back-end algorithms, including simulations to estimate their physics performance. The first algorithm emulates front-end readout chip buffers and occasionally throttles CMS Level-1 trigger decisions to prevent data loss due to buffering limitations. An analysis showing this algorithm results in approximately 1.5E−8 deadtime will be presented. The second algorithm calculates 3D cluster properties (energy, position, shape) from the front-end energy deposits for use in the CMS Level-1 trigger. It will be shown that this algorithm has excellent and very good ability to identify photons and Vector Boson Fusion jets, respectively. Both blocks were implemented in software and firmware, and successfully tested in hardware.