Issue link: https://iconnect007.uberflip.com/i/1295812
12 DESIGN007 MAGAZINE I OCTOBER 2020 Feature Interview by the I-Connect007 Editorial Team This month, we're focusing on technology roadmaps. We wanted to get a feel for the future of high-speed EDA tools, so we spoke with Todd Westerhoff, product marketing man- ager for high-speed tools at Mentor, a Siemens Business. Todd discussed the future of high- speed PCB design tools, as well as some of the many market forces that are driving the devel- opment of the company's tools and the overall EDA segment. He also details a new approach that Mentor is taking: progressive design and analysis. Andy Shaughnessy: Todd, why don't you give us a look ahead at Mentor's high-speed tools and tie it in with what you see going on in the industry in the near future. Todd Westerhoff: Thanks, Andy. When we met at DesignCon this year, we said, "It's the small stuff that gets you." We meant that if you look at what holds up design projects, it's often not the state-of-the-art problems you would expect. It's the stuff that you thought was straightforward that bites you in the end. As a design community, we're pretty good at the complicated stuff because we pay careful attention to it, but we get can be tripped up on what should have been easy. Most companies see signal integrity (SI) as too time-consuming or too complicated for system designers, so analysis gets put into the hands of dedicated specialists to make the pro- cess "more efficient." SI analysis then becomes concentrated at the end of the design cycle. Good post-layout verification will find prob- lems before fab-out and avoid wasting money on prototypes, but you're still finding errors well past the point where they were intro- duced into the design. That means parts of the design have to be ripped up and rerouted, and there's a schedule impact associated with that. We want to give system designers tools where they can find and correct those errors as early as possible. We need a class of analysis tools that will allow system designers to effectively evaluate trade-offs and make design decisions with- out having to get the SI expert involved. The accuracy doesn't have to be as good as what the expert uses—something we call first-order analysis. The goal is to look at things more qualitatively. We don't need to compute sys- tem margin down to millivolts and picoseconds on an absolute basis; we want to know how a With Progressive Design and Analysis Looking Into the Future