The ePIC-dRICH streaming readout system

The investigation of the nucleon structure is one of the primary goals of the ePIC experiment at the future Electron–Ion Collider. Considering both inclusive and exclusive processes, ePIC will access unprecedented ranges in Q$^2$ and Bjorken x. The dual-radiator RICH (dRICH) detector is a key component of the ePIC Particle Identification (PID) system and is essential for the reconstruction of hadrons produced in high-Q$^2$ scattering events. Covering the pseudorapidity range $1.5< \eta <3.5$, the dRICH ensures $\pi$/K/p separation from $\sim3$ to $\sim 50$ GeV/c thanks to two Cherenkov radiators. More than 300 thousand 3x3 mm$^2$ SiPMs are employed as photosensors, each of them representing one readout channel. Given an expected radiation dose of $\sim6\cdot10^{10}$ 1-MeV neq/cm$^2$, radiation damage to the SiPMs leads to an increase in the dark count rate (DCR), which constitutes the main limitation to single-photon resolution. Mitigation strategies based on low-temperature operation and in-situ annealing allow the DCR to be contained below 300 kHz after $\sim200$ fb$^{-1}$ of integrated luminosity. 

The highest data throughput in ePIC of $\approx 7$ Tbit/s is expected from the dRICH readout system, primarily due to the DCR contribution. To address this challenge, efficient data-reduction techniques have been developed together with a streaming readout architecture segmented into 1248 Photon Detection Units (PDUs). Each PDU integrates four matrices of 64 SiPMs each, along with front-end electronics based on 4 ALCOR ASICs and an FPGA-based Readout (RDO) card. The ALCOR ASICs generate precise timestamps data, which are aggregated by the RDO card and streamed to the ePIC DAQ system via a 10 Gb/s optical link.

The data-push readout architecture of the dRICH detector will be presented, focusing mainly on the front-end and readout boards. A data throughput of $\approx 1.4$ Tb/s is achievable using different data reduction methods that will be shown, highlighting the main role of the DAQ architecture design. Finally, results of prototype PDU readout will be discussed, in terms of data throughput, together with future studies exploiting the full bandwidth available of the selected optical link.