Reliable and redundant FPGA based read-out design in the ATLAS TileCal Demonstrator

TL;DR

This paper presents a highly reliable, redundant FPGA-based read-out system for the ATLAS TileCal detector, enabling full data read-out and improved radiation tolerance, with a demonstrator to evaluate long-term performance before full deployment.

Contribution

Introduction of a new FPGA-based, redundant read-out system with radiation-tolerant design for the ATLAS TileCal, including a demonstrator for thorough reliability testing.

Findings

Full data read-out capability achieved with FPGA and optical links

Redundancy and radiation-tolerant design improve reliability

Demonstrator validates long-term performance and fault protection

Abstract

The Tile Calorimeter at ATLAS is a hadron calorimeter based on steel plates and scintillating tiles read out by PMTs. The current read-out system uses standard ADCs and custom ASICs to digitize and temporarily store the data on the detector. However, only a subset of the data is actually read out to the counting room. The on-detector electronics will be replaced around 2023. To achieve the required reliability the upgraded system will be highly redundant. Here the ASICs will be replaced with Kintex-7 FPGAs from Xilinx. This, in addition to the use of multiple 10 Gbps optical read-out links, will allow a full read-out of all detector data. Due to the higher radiation levels expected when the beam luminosity is increased, opportunities for repairs will be less frequent. The circuitry and firmware must therefore be designed for sufficiently high reliability using redundancy and radiation tolerant components. Within a year, a hybrid demonstrator including the new read-out system will be installed in one slice of the ATLAS Tile Calorimeter. This will allow the proposed upgrade to be thoroughly evaluated well before the planned 2023 deployment in all slices, especially with regard to long term reliability. Different firmware strategies alongside with their integration in the demonstrator are presented in the context of high reliability protection against hardware malfunction and radiation induced errors.