FPGA Qualification and Failure Rate Estimation Methodology for LHC Environments Using Benchmarks Test Circuits

When studying the behavior of a field programmable gate array (FPGA) under radiation, the most commonly used methodology consists in evaluating the single-event effect (SEE) cross section of its elements individually. However, this method does not allow the estimation of the device failure rate when using a custom design. An alternative approach based on benchmark circuits is presented in this article. It allows standardized application-level testing, which makes the comparison between different FPGAs easier. Moreover, it allows the evaluation of the FPGA failure rate independent of the application that will be implemented. The employed benchmark circuit belongs to the ITC'99 benchmark suite developed at Politecnico di Torino. Using the proposed methodology, the response of four FPGAs-the NG-Medium, the ProASIC3, the SmartFusion2, and the PolarFire-was evaluated under high-energy protons. Radiation tests with thermal neutrons were also conducted on the PolarFire to assess its potential sensitivity to them. Moreover, its performances in terms of total ionizing dose (TID) effects have been evaluated by measuring the degradation of the propagation delay during irradiation.

Automated summary

This study presents an alternative approach based on benchmark circuits to evaluate the failure rate of FPGAs under radiation. The proposed methodology allows standardized application-level testing and facilitates comparison between different FPGAs. The response of several FPGA models to high-energy protons and thermal neutrons, as well as their performance in terms of total ionizing dose effects, were evaluated.