Design for Test and Fault Tolerance for Nanoscale Circuits

Lecture

May 23 14:00
ETS Track B

Sybille Hellebrand
University of Padeborn

Nanoscale integration comes along with an increased variability of circuit parameters as well as with growing soft error rates. To achieve acceptable yield and to ensure a reliable system operation in the field, a "robust" design must compensate both permanent and transient faults to a certain extent. Testing becomes particularly difficult in this context, because different instances of a circuit may need different test sets and a robust design style makes it hard to distinguish between critical and non-critical failures during test. Moreover, a pass/fail-test is no longer sufficient, but the remaining robustness for system operation must be determined ("quality binning").

This tutorial shows how classical fault tolerant architectures can be used for yield and reliability improvement and also introduces emerging self-calibrating and adaptive architectures. Furthermore, the challenges of testing robust systems are discussed and specific DfT solutions are presented. Finally, techniques for analyzing the robustness properties of a circuit are explained.

Prerequisites

  • Basic knowledge about test and DfT, in particular built-in and embedded test
  • Basic knowledge about fault tolerance

Suggested preliminary readings

  • Michael L. Bushnell, Vishwani D. Agrawal: Essentials of Electronic Testing for Digital, Memory, and Mixed-Signal VLSI Circuits; Kluwer Academic Publishers (Springer), 2000
  • I. Koren and C. M. Krishna: Fault-Tolerant Systems; Morgan-Kaufman Publishers, San Francisco, CA, USA, 2007
  • IEEE Design and Test: Special Issue on Design for Yield and Reliability, Vol. 21, No. 3, May/June 2004

Learning outcomes

  • Understanding the challenges stemming from parameter variations and increased soft error rates
  • Basic knowledge about design and DfT for robust architectures

Syllabus

  1. Introduction: New challenges for DfT and fault tolerance due to the increasing impact of parameter variations and soft errors.
  2. Fault tolerance for yield and reliability improvement
  3. New approaches for robust design (variation-tolerant, adaptive, self-calibrating, ...): basic ideas and architectures
  4. Prerequisites for robust design:
    • Online monitoring, online error detection and correction
    • Monitoring of aging problems
    • Infrastructure for self-calibration
  5. Test and DfT for robust design
    • Minimizing overtesting and yield loss
    • Test calibration
    • Design for "Quality Binning"
  6. Robustness analysis