A Statistic-Based Scan Chain Reordering for Energy-Quality Scalable Scan Test

As a rapid progress in technology processes, the design integration of high-performance system-on-chip (SoC) is on the rise rapidly. To incorporate hundreds of IP cores into a single chip, a modern SoC exceeds ten million gates with a large number of scan cells, so that it leads excessive energy consumption. In this paper, we present an energy-quality (EQ) scalable scan test method using new scan chain reordering. The method conducts three stages, which are a new scan partitioning, a scan partition-based X-filling, and a statistic-based scan stitching to reduce test energy consumption without quality degradation. The proposed scan partitioning method prevents excessive routing overhead. Then, the proposed scan chain reordering is performed by a statistical analysis considering EQ scalability. It also covers two frequently-used fault models: 1) stuck-at and 2) transition delay. The experimental results show that the proposed scan chain reordering method achieved lower energy consumption and relieve the routing overhead on ISCAS'89, ITC'99, and IWLS'05 OpenCores benchmark circuits in most cases compared with previously existing methods without excessive runtime overhead.