Keywords

FPGA, Reliability, SEU

Abstract

Sponsorship: Los Alamos National Laboratory. Field Programmable Gate Arrays (FPGAs) are an attractive solution for space system electronics. Unfortunately, FPGAs are susceptible to radiation-induced single-event upsets (SEU). As such, the FPGA Reliability Studies research group (http://reliability.ee.byu.edu) at Brigham Young University has studied ways to effectively measure the static, dynamic and persistent cross sections of an FPGA desgin; each of which are characterized in some way by how the part reacts to an SEU. One such method is to actually radiate an FPGA and monitor how it reacts to SEUs. A cheaper, more efficient solution is to use fault-injection to emulate SEUs. In order to validate the use of fault-injection to measure the persistent cross section of an FPGA design, we measured the persistent cross section of several designs using proton irradiation at Crocker Nuclear Laboratory in Davis, CA. Our goal was to show a high correlation between accelerator and simulation data to prove that fault-injection is a reliable alternative to proton irradation. This document is a detailed description of the correlation process.