ALICE is a general-purpose detector that is designed to study the physics of quark-gluon plasma. The Time Projection Chamber (TPC) is one of the major sub-detectors of ALICE. The TPC electronics consists of 4356 Front-end cards (FECs), which are controlled by 216 Readout Control Units (RCU). In LHC Run1, the Readout Control Unit 1 (RCU1) performed even better than specification. However, in Run2 the energy of colliding beams is increased from 8 TeV to 14 TeV (maximum value) and higher luminosity, which leads to larger event size and higher radiation load on the electronics. As a solution, the Readout Control Unit 2 (RCU2) is designed to provide faster readout speed and improved radiation tolerance with respect to the RCU1.
The RCU2 is conceptually similar to the RCU1 and it reuses the existing readout architecture of the TPC electronics. However, it has improved parallelism in both hardware design and firmware design. This ensures that the readout speed of the RCU2 can be improved by a factor of ~2 with respect to the RCU1. The flash-based Microsemi Smartfusion2 FPGA SOC is used as the main FPGA instead of the SRAM based Xilinx Virtex 2 Pro FPGA that was used on the RCU1. Because its configuration cells are immune to Single Event Effects, the radiation tolerance of the RCU2 was expected to be improved.
The primary objective of this thesis has been to study and improve the radiation tolerance of the RCU2. This is done through several irradiation tests. Actions have been taken against all the radiation related problems that were revealed during the irradiation tests. Running experience shows that radiation tolerance of the readout system based on RCU2 is about 10 times better as compared to the RCU1 for p-Pb collisions at similar energy level. Development of the readout algorithms and system integration of the RCU2 are also important tasks covered in this thesis. Readout performance of the RCU2 has been characterized and it has met the design requirements. The current data acquisition system of the ALICE TPC is measured to record data at factor two higher rates than the readout rates during Run1.