• DocumentCode
    2754992
  • Title

    Why would an ASIC foundry accept anything less than full scan?

  • Author

    Oakland, Steven F.

  • Author_Institution
    Microelectron. Div., IBM Corp., Essex Junction, VT, USA
  • fYear
    1997
  • fDate
    1-6 Nov 1997
  • Firstpage
    1031
  • Abstract
    A key force behind IBM´s growth in the application-specific integrated circuit (ASIC) market is the ability to sign off on multi-million-gate designs without requiring test vectors, presenting a savings in both time and money to customers. Once a customer ensures (via formal verification and/or functional simulation) that the design functions as required, static tinting analysis (STA) ensures that the design achieves the required performance targets. Extensive model-to-hardware correlation assures correctness of the timing analysis models, enabling IBM to assure that the design can be manufactured to the required performance targets. Through a combination of full-scan and boundary-scan design-for-test (DFT) structures, the IBM ASIC methodology ensures that automatically generated test patterns will run correctly on test equipment; typically achieving 99+% stuck-fault coverage. In the case of a repeatable manufacturing defect, full-scan-based diagnostic software isolates the problem without customer involvement
  • Keywords
    application specific integrated circuits; automatic testing; boundary scan testing; design for testability; economics; integrated circuit manufacture; integrated circuit testing; logic testing; production testing; ASIC; ASIC foundry; IBM; application-specific integrated circuit; boundary-scan DFT structures; design functions; full-scan DFT structures; model-to-hardware correlation; repeatable manufacturing defect; static tinting analysis; stuck-fault coverage; Analytical models; Application specific integrated circuits; Circuit simulation; Circuit testing; Design for testability; Formal verification; Foundries; Integrated circuit testing; Performance analysis; Timing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Test Conference, 1997. Proceedings., International
  • Conference_Location
    Washington, DC
  • ISSN
    1089-3539
  • Print_ISBN
    0-7803-4209-7
  • Type

    conf

  • DOI
    10.1109/TEST.1997.639721
  • Filename
    639721