A Future History of Practical AMS DfT

Stephen Sunter - Siemens Digital Industries Software, Canada

Abstract

The rate of change in analog and mixed-signal (AMS) design-for-test (DfT) has been increasing in the last 30 years and this is likely to continue in the next 5 years. The first papers on systematic AMS DFT approaches and analog fault simulation appeared in the early-1990s. Progress in the next 20 years was mostly in academia and a few start-ups, quite slow compared to changes in digital DfT, and focused on developing built-in self-test (BIST) as a lower-cost equivalent to AMS testers. Ultimately, industry didn’t trust this BIST because it added complexity and an unknown impact on quality.

About 10 years ago, the pace of change quickened as large companies became interested in analog fault/defect simulation to systematically improve quality, as measured by outgoing defective parts per million (DPPM), especially for the automotive market. Just as digital fault simulation has facilitated rapid improvements in scan-based DfT for logic testing, industrial-strength analog defect simulation will undoubtedly facilitate rapid changes in AMS DfT during the next 5 years.

This lecture will highlight significant steps in AMS DfT during ever shorter spans of years, including the near future, and the emerging problems. We’ll also investigate how they were solved, except of course for the problems that will arise in the next 5 years, which attendees are expected to solve.

Syllabus

• The State-of-the-Art in 2013    (1.5 hours)
  • Significant achievements in previous 20 years
  • Emerging problems
• The State-of-the-Art in 2023    (1.5 hours)
  • Significant achievements in previous 10 years
  • Emerging problems
• The State-of-the-Art in 2028    (1 hour)
  • Significant achievements in previous 5 years
  • Emerging problems
• Conclusions