Studying High-Speed I/O Failures: Tales from the Intel I/O Test Road Map

After successful deployment of AC I/O loopback for single-ended interfaces, we rapidly had to prepare for High Speed I/O (HSIO) circuitry. The HSIO derived from Serializer/Deserializer (aka SerDes) interfaces commonly used in telecommunications. Computer systems’ thirst for higher data rates drove the adoption of this interface architecture. These I/O circuits had significant differences from single-ended interfaces, which AC I/O loopback supported.  The feature list included:

  • Unidirectional signaling
  • Point-to-point interconnect on the system board
  • Differential signaling
  • 8/10-bit encoding
  • Protocol aware interfaces to move from sleep mode to active transmission
  • Training as part of the wake-up procedure to compensate circuitry to strobe and impedance
  • Significantly higher data rates than any of Intel’s testers

In summary, these circuits were a heck of a lot more complicated.

We started the pathfinding work in 2001 to prepare for an introduction of products in 2005. I led a small team of engineers for this work. In graduate school I learned that the best way to approach testing is to understand how things fail. Hence, we began a study of defect’s impact on circuit performance. A test chip design team gave us access to their 3GIO circuits (aka Third Generation I/O.) Inserting defects into a good circuit and simulating the resulting performance required multiple fault simulations. We used a heuristic to insert defects into the circuit, based upon the schematic connections. We had developed a set of scripts to perform this function.

To simplify the analysis, we split the 3GIO prototype circuit into sub-circuits for each team member. I took on the data clock recovery circuitry, Pat took on the driver circuit, and Bob took on the receiver circuitry. The Data Clock recovery circuit had an implementation based upon a clock mixer, which was different than another over-sampling technique. Stay tuned for what I found.

Have a productive day,

Anne Meixner

Dear Reader, What memory or question does this piece spark in you? What has been the essential engineering technique to your current engineering practice? Please share your comments or stories below. You, too, can write for the Engineers’ Daughter–See Contribute for more information.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.