Overview of the data acquisition system architecture for the DarkSide-20k experiment

TL;DR

The paper describes the architecture of the data acquisition system for the DarkSide-20k dark matter detector, focusing on its triggerless operation, high data throughput, and multi-stage processing pipeline.

Contribution

It presents a detailed design of a triggerless, high-throughput DAQ system tailored for the DarkSide-20k experiment, including data flow and processing architecture.

Findings

Handles 3 GB/s data flow from detector

Utilizes 48 high-speed waveform digitizers

Implements multi-stage data reduction and processing

Abstract

The DarkSide-20k (DS-20k) detector is designed to directly search for dark matter by detecting weakly interacting massive particles (WIMPs) scattering off nuclei in a 20 tonnes target of low radioactivity Argon from underground sources, in a dual-phase time projection chamber (TPC). The scintillation light produced by these interactions is recorded by custom silicon photomultiplier (SiPMs) assemblies comprising the detector's 2720 readout channels. The data acquisition system (DAQ) for the DS-20k experiment operates in a triggerless mode, handling the estimated 3 GB/s flow of digitized data from MC simulation from the entire detector, dominated by the TPC. Signals are digitized by 48 commercial VX2745 CAEN 16-bit, 125 MS/s, high channel density waveform digitizers. This digitized data is then transferred to 24 Front-End Processor (FEP) machines for an initial data reduction stage. Subsequently, the processed data is sent to a separate set of Time Slice Processor (TSP) computers, where it is assembled into fixed-length time series for further analysis and storage for offline processing.