An SEU fault injection platform for radiation-harden design debugging in the FPGA

Author:

Zhang F.,Zhou Z.,Wang Y.,Wang D.,Wu M.,Zhu L.

Abstract

Abstract An SEU fault injection platform was designed to ease the debugging of radiation-harden design in FPGA. The platform includes the FPGA being tested, the functional firmware being debugged, the essential-bit extraction algorithm based on Python, and the random fault injection algorithm. The platform can inject single-bit SEU failures, adjacent two-bit SEU failures, and three-bit SEU failures. It works in two modes: normal injection mode and essential-bit injection mode. Functional failure rate is the performance metric which used to evaluate the development. It is the probability of triggering a DUT function failure. In this experiment, the SEU fault injection platform is verified by measuring whether TCP/IP communication links are disconnected due to SEU faults. Experimental results show that the probability of TCP/IP link disconnection in normal injection mode is 0.13%, 0.23% and 0.25% respectively when random injection of single-bit SEU failure, adjacent two-bit SEU failure and multi-bit SEU failure occurs 10,000 times. In the essential-bit injection mode, the probability of TCP/IP link disconnection caused by single-bit SEUs failure, adjacent two-bit SEUs failure and three-bit SEUs failure is 0.87%, 6.97% and 11.76% respectively, indicating a significant increase in the functional failure rates. This shows that the essential-bit injection mode can be used to expose problems more quickly and accelerate the debugging process of FPGA radiation-harden design.

Publisher

IOP Publishing

Subject

Mathematical Physics,Instrumentation

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3