The major conference on semiconductor reliability is the International Reliability and Physics Symposium . Although a lot of the focus these days is on semiconductors, the symposium been going on for 55 years and presumably covered a different beat when it started. This year it takes place April 2nd to 6th, in Monterey. Recently the conference has been focused on very low-level reliability issues, heating effects in transistors, for example. However, increasingly, a more holistic system view is needed. One big driver, as with so much in semiconductor, is autonomous driving (and its predecessor ADAS). The reliability requirements are high for obvious reasons. Importantly, they are higher than the reliability of the underlying silicon (or GaN or whatever is being used) and so just focusing on low-level effects, while important, is not enough. Keynotes Perhaps this is best illustrated by the first keynote, which is by Nancy Currie-Gregg, who works at NASA and is both an astronaut and a principal engineer of NASA's engineering and safety center. Her talk is titled Reflections on the Risk of Human Space Exploration—Lessons Learned from Past Failures . She will talk about lessons learned from the three biggest failures in the last 50 years of the space program: The Apollo 1 launchpad fire The shuttle Challenger's explosion during ascent The shuttle Columbia's breaking up during re-entry She will draw out lessons any engineering organization applicable to all organizations managing extremely complex systems operating in high-risk environments. Continuing with the theme of the reliability of systems as a whole, the second keynote is by Ronald Newhart of IBM on System Level Reliability Challenges with Technology Scaling . He is a distinguished engineer in the Systems and Technology Group focusing on (surprise!) reliability. He has been the lead technology interface for the POWER and Z-Systems microprocessors for nine product generations. The keynotes open the conference proper on Tuesday morning. But there are more the previous two days. Sunday and Monday Sunday and Monday are a couple of days of tutorials. Again, these cover a wide range, from system-level issues to low-level process issues. The tutorials run in parallel, so you can't go to all of them. Some are aimed at true experts, and some are high-level aimed at beginners. Unfortunately, it is not always obvious which are which. I remember reading a sentence in a book on artificial intelligence---something I've always remembered: "You can only learn something if you almost know it already." So you need to go to tutorials where you have enough knowledge to understand the material, but not so much that you learn nothing. On Sunday there are two full-day sessions on: Silicon modes Integration On Monday there are four most-of-a-day sessions on: Advanced silicon and package Automotive and IoT (I intend to attend this one) Integration Circuits and systems Then, at 2:30pm, there is a Year in Review, looking at the most important developments in the reliability area over the last year. Tuesday to Thursday The regular sessions take place in three parallel tracks. There are far too many papers for me to even list the invited papers. But let me highlight three that look especially interesting. Ken LaBel of NASA on NASA Past, Present, and Future: The Use of COTS in Space . COTS stands for Commercial Off-The-Shelf, so in this context, it refers to semiconductors that were never designed to go into space (see table below for a flavor). Tom Sharpe of SMT Corporation on Today's Functional Electronic Clones. Advanced counterfeiters are producing "functional die emulation" manufacturing processes, which can pass a lot of the testing but are made on the cheap without the same reliability. R. Kwasnick of Intel on Telemetry for Reliability . He will be talking about acquiring use condition data from users in the field. There are special sessions on wide bandgap devices, primarily GaN (gallium nitride) and SiC (silicon carbide). There is a special session on FinFET oxide reliability and FinFET BTI (bias temperature instability). In another session, a series of field tests of photovoltaic reliability will be reported on. On Tuesday evening there are six workshops on topics from COTS, through middle-of-line, to standards-based qualification. Details and Registration If you are not from the Bay Area, Monterey is an interesting place to visit with a long history. It was the first capital of California. John Steinbeck wrote Cannery Row about it, and Cannery Row is a real place, although the canneries are long gone and it is mostly trendy restaurants. The most famous place there is the Monterey Bay Aquarium, which will almost certainly have long lines but is definitely somewhere that you have to visit. This is all walking distance from the conference hotel (quite a long walk all the way to the aquarium). If you have a car, Carmel is not that far away. Big Sur is not that far away as the crow files, but for us non-crows the road is currently closed, although it may have re-opened by the time of the conference. Sections got washed away in the recent rains (and a bridge collapsed, although I think that is coming in from the south, which will take months to replace). Full details are on the IRPS website . Early registration ends on March 16th, details here .![]()
![]()
↧