Browsing by Author "Zhou, Jun"

Now showing 1 - 1 of 1

Open Access
Software based self test under memory, time and power constraints
(2009) Zhou, Jun; Wunderlich, Hans-Joachim (Prof. Dr. habil.)
The use of embedded system is ubiquitous in modern life, from automotive industry, robot techniques to the household appliances. As the core component of embedded systems, microprocessors play the central role in information processing, and fulfill several specialized tasks. The involvment in safty-critical applications requires high reliability of these systems. Test is known to be important for product quality and used to detect defects both during manufacturing or in-field scenario. In particular, the following aspects are crucial for microprocessor test. First, a low-cost test is preferrable in terms of hardware overhead, test time for example. To detect timing faults, testing at the system frequency, namely at-speed testing, is required. Design modification to incorporate hardware dedicated for test is not desirable due to potential performance degradation. Overtesting accounts for possible yield loss in that chips may fail the structural test, although those faults can never be sensitized in normal functional operations. The advent of software-based self-test solve these problems and therefore becomes the target of this work. However, constrained test generation is still of need. The dissertation looks into this aspect and proposes a novel SBST scheme, which optimizes memory, test application and power consumption at the same time without penalty of fault coverage. In addition, this work also presents a method to improve test quality of SBST through adding new instructions under the ASIP (application-specific instruction-set processor) framework.