Test Data Truncation for Test Quality Maximization under ATE Memory Depth Constraint

Research output: Contribution to journalArticle

Abstract

Testing is used to ensure production of high quality integrated circuits. High test quality implies the application of high quality test data; however, technology development has led to a need to increase test data volumes to ensure high test quality. The problem is that the high test data volume leads to long test application times and high automatic test equipment memory requirement. For a modular core-based system-on-chip, a test data truncation scheme is proposed, that selects test data for each module in such a way that the system test quality is maximised while the selected test data are guaranteed to overcome constraints on time and memory. For test data selection, a test quality metric is defined based on fault coverage, defect probability and number of applied test vectors, and a scheme that selects the appropriate number of test vectors for each core, based on the test quality metric, defines the test architecture and schedules the transportation of the selected test data volume on the test access mechanism such that the system's test quality is maximised. The proposed technique has been implemented, and the experimental results, produced at reasonable CPU times on several ITC'02 benchmarks, show that high test quality can be achieved by careful selection of test data. The results indicate that the test data volume and test application time can be reduced to about 50% while keeping a high test quality.

Details

Authors
External organisations
  • Linköping University
Research areas and keywords

Subject classification (UKÄ) – MANDATORY

  • Electrical Engineering, Electronic Engineering, Information Engineering

Keywords

  • testing, electronic systems, memory constraints
Original languageEnglish
Pages (from-to)27-37
JournalIET Computers and Digital Techniques
Volume1
Issue number1
Publication statusPublished - 2007
Publication categoryResearch
Peer-reviewedYes
Externally publishedYes