作者: Sisir Kumar Jena , Santosh Biswas , Jatindra Kumar Deka
DOI: 10.1109/VDAT50263.2020.9190571
关键词:
摘要: A gigantic amount of test pattern covering all faults is required to attain a high quality in modern VLSI testing. With this, the Test Data Volume (TDV) rises, which leads an increase Application Time (TAT) as well power requirement. This paper proposes Fault-based Approximate Testing (FAT) technique that generates patterns only for distinguished faults. The basic idea identifying set (tolerable faults) can be ignored and left untested. remaining (intolerable are tested by generating those We examine impact overlooking some injecting stuck-at appropriate place netlist. Deviation output analyzed respect application-specific error threshold. experimental result shows irrespective erroneous produced, these circuits used kinds error-tolerant applications like image/video processing, speech digital communications. Along with above benefit number fault-sites reduces 40–50%, reduction patterns.