• DocumentCode
    3387302
  • Title

    Data retention fault in SRAM memories: analysis and detection procedures

  • Author

    Dilillo, Luigi ; Girard, Patrick ; Pravossoudovitch, Serge ; Virazel, Arnaud ; Hage-Hassan, Magali Bastian

  • Author_Institution
    Lab. d´´Informatique, de Robotique et de Microelectronique de Montpellier, Univ. de Montpellier II, France
  • fYear
    2005
  • fDate
    1-5 May 2005
  • Firstpage
    183
  • Lastpage
    188
  • Abstract
    In this paper, we present a novel study on data retention faults (DRFs) in SRAM memories. We analyze in detail the electrical origins of these faults, starting from the most common till those that lead to what we have called hard to detect DRFs. In general, DRFs are supposed to be produced by very high resistive-open defects that affect the refreshment loop of the core-cell. We demonstrate that lower values of resistance may produce hard to detect DRFs. Moreover, each resistive-open defect produces a particular faulty behavior of the core-cell that changes for different ranges of the resistive value. We analyze different cases and we propose for each one an efficient test procedure based on March tests. In particular, we propose to stimulate the defective cells in some cases by indirect accesses and in some other cases by emphasizing natural noise phenomenon of SRAM memories (such as the ground bounce).
  • Keywords
    SRAM chips; electric resistance measurement; failure analysis; fault diagnosis; integrated circuit testing; system-on-chip; DRF; March tests; SRAM memories; core-cell faulty behavior; data retention fault; electrical origins; fault analysis; fault detection; ground bounce; refreshment loop; resistance values; resistive-open defects; Circuit faults; Circuit testing; Electric resistance; Electrical fault detection; Fault detection; Integrated circuit noise; Random access memory; Robots; System-on-a-chip; Uniform resource locators;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    VLSI Test Symposium, 2005. Proceedings. 23rd IEEE
  • ISSN
    1093-0167
  • Print_ISBN
    0-7695-2314-5
  • Type

    conf

  • DOI
    10.1109/VTS.2005.37
  • Filename
    1443421