CERN Accelerating science

ALICE is undergoing upgrades during the Long Shutdown (LS) 2 of the LHC to improve its performance and capabilities, and to prepare the experiment for the increases in luminosity provided by the LHC in Run 3 and Run 4. One of the most extensive upgrades of the experiment (and the topic of this thesis) is the replacement of the Inner Tracking System (ITS) in its entirety with a new and upgraded system. The new ITS consists exclusively of pixel sensors organized in seven cylindrical layers, and offers significantly improved tracking capabilities at higher interaction rates. And in contrast to the previous system, which would only trigger on a subset of the available events that were deemed “interesting”, the upgraded ITS will capture all events; either in a triggered mode using minimum-bias triggers, or in a “trigger-less” continuous mode where event data is continuously read out. The key component of the upgrade is a novel pixel sensor chip, the ALPIDE, which was developed at CERN specifically for the ALICE ITS upgrade. The seven layers of the ITS is assembled from sub-assemblies of sensor chips referred to as staves, and the entire detector consists of 24 120 chips in total. The staves come in three different configurations; they range from 9 chips per stave for the innermost layers, and up to 196 chips per stave in the outer layers. The number of control and data links, as well as the bit-rate of the data links, differs widely between the staves as well. Data readout from the high-speed copper links of the detector requires dedicated readout electronics in the vicinity of the detector. The core component of this system is the FPGA-based Readout Unit (RU). It facilitates the readout of the data links and transfer data to the experiment’s server farms via optical links; provides control, configuration and monitoring of the sensor chips using the same optical links, as well as over CAN-bus for redundancy; distributes trigger signals to the sensor, either by forwarding the minimum-bias triggers of the experiment, or by local generation of trigger pulses for the continuous mode. And the field-programmable devices of the RU allows for future updates and changes of functionality, which can be performed remotely via several redundant paths to the RUs. This is an important feature, since the RUs are not easily accessible when they are installed in the cavern of the experiment and will be exposed to radiation when the LHC is in operation. Radiation tolerance has been an important concern during the development of the FPGA designs, as well as the RU hardware itself, since radiation-induced errors in the RUs are expected during operation. Techniques such as Triple Modular Redundancy (TMR) were used in the FPGA designs to mitigate these effects. One example is the radiation tolerant CAN controller design which is introduced in this thesis. A different challenge, which is also addressed in this thesis, is the monitoring of internal status and quantities such as temperature and voltage in the ALPIDE chips. This is performed over the ALPIDE’s control bus, but must be carefully coordinated as the control bus is also used for triggers. The detector and readout electronics are designed to operate under a wide set of conditions. Considering events from Pb–Pb collisions, which may have thousands of pixel hits in the detector, a typical pp event has comparatively few pixel hits, but the collision rate is significantly higher for pp runs than it is for Pb–Pb runs. And the detector can be used with two triggering modes, where the continuous trigger mode has additional parameters for trigger period. A simulation model of the ALPIDE and ITS, presented in this thesis, was developed to simulate the readout performance and efficiency of the detector under a wide set of circumstances. The simulated results show that the detector should perform with a high efficiency at the collision rates that are planned for Run 3. Initial plans for a dedicated hardware, to handle and coordinate busy status for the detector, was deemed superfluous and the plans were canceled based on these results. Collision rates higher than those planned for Run 3 were also simulated to yield parameters for optimal performance at those rates. For the RU, which was designed to interface to three widely different stave designs, the simulations quantified the amount of data the readout electronics will have to handle depending on the detector layer and operating conditions. Furthermore, the simulation model was adapted for simulations of two other ALPIDE-based detector projects; the Proton CT (pCT) project at University of Bergen (UiB), a Digital Tracking Calorimeter (DTC) used for dose planning of particle therapy in cancer treatment; and the planned Forward Calorimeter (FoCal) for ALICE, where there will be two layers of pixel sensors among the 18 layers of Si-W calorimeter pads in the electromagnetic part of the detector (FoCal-E). Since the size of a calorimeter pad is relatively large, around 1 cm², the fine grained pixels of the ALPIDE (29.24μm × 26.88 μm) will help distinguish between multiple showers and improve the overall spatial resolution of the detector. The simulations helped prove the feasibility of the ALPIDE for this detector, from a readout perspective, and FoCal was later approved by the LHCC committee at CERN.