Shrinking geometries and efficient design techniques are helping to reduce die sizes, which lowers the cost of semiconductor devices. Despite these improvements, increased competition and smaller gross margins are forcing semiconductor companies to reduce the overall cost of IC production even more. One of the major contributors to total device cost is the cost of testing.
For the digital portion of ICs, DFT (design for test) techniques have significantly reduced test complexity and test times. Unfortunately, testing the analog portion of an IC is much more complex. As a result, most engineers still perform analog measurements using conventional methods, such as bringing the analog output to a package-level pin and performing the measurement using external instruments. This approach has its disadvantages, especially in terms of time and cost.
Fortunately, there is a way to reduce the test time spent on clock frequency measurements. By performing on-chip frequency measurements, device manufacturers can reduce their dependency on external instruments and can perform concurrent, parallel, and faster frequency measurements without adding any significant silicon area. In fact, one study has shown a reduction in test time of more than 50%. For devices with many clock sources, this can lead to significant test cost savings.
Hi agk, thanks for the comment. Obviously if the target clock is the same as the timer clock, it will not work. The article addresses this and mentions that the timer must be running at twice the frequency of the clock being measured.
On a system there can be multiple clock sources and if all of them are measured using on-chip timers, test time savings will be much higher.
Please continue to post comments, we appreciate it.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.