The Counter-Based Output Selection Method with Unknown-Preventing Capability

• Main Authors: Yu-Hua Wu, 吳育樺
• Other Authors: Kuen-Jong Lee
• Format: Others
• Language: en_US
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/93596304302728094425

碩士 === 國立成功大學 === 電機工程學系碩博士班 === 96 === In this thesis, we propose a novel output response compaction technique, called the Counter-Based Output Selection Method, to address the problems of large test volume, aliasing, unknown value (X) and difficult-to-diagnosis that commonly exist in traditional space/time compactors. The basic idea is to directly select only a subset of output bits for observation so as to achieve aliasing-free and X-tolerant capabilities. An ATPG- independent selection algorithm called the Dynamic Selection Procedure is developed to efficiently and effectively select the necessary output bits to observe based on the control of a simple counter. Experimental results show that on the average over 10x compaction ratio can be achieved even with highly compressed test sets. More importantly, we find that almost the same diagnosability achievable using the original test sets can be preserved. To make this technique more practical, the influence of unknown value is also considered. An output selection algorithm that can efficiently avoid the selection of unknown bits is developed. Experimental results show that for all ISCAS89 and ITC99 benchmark circuits, when 2% unknown bits are randomly inserted to the output response, the algorithm always can identify a control sequence for the counter to select a set of deterministic output bits to detect all detectable faults. Even for the cases where 5% or even higher (10%) unknowns present in the circuits, the developed algorithm can still avoid almost all unknown values by using a scan chain reordering procedure.