Reading time ( words)
If your company is having trouble designing PCBs for profitability, where does the blame fall? Is it management’s fault for having inefficient processes, or is it the fault of the designers and design engineers for not keeping up to date with their training? During DesignCon, I asked Al Neves of Wild River Technology to weigh in on this question. As he explains, you and your manager might both be to blame for inefficiencies in the design process.
Andy Shaughnessy: Al, we were just talking about designing for profitability. You mentioned that you had a few pithy thoughts on the subject.
Al Neves: Yes, and I see a trend. The work that we focus on is high-risk, extremely demanding signal integrity projects. About half of our business is service, and the other half is standard products. On the service side, somebody will come out with silicon and want a test platform. Many of the test platforms are 50–70 gigahertz right now, with very advanced embedding, complicated breakouts from their BGA, and high-density routing. The signal integrity has to be really good. One specification—IEEE P370—defines what good signal integrity is at 50 gigahertz.
A customer will say, “We have something that we’d like you to help on.” Once we’re at the technical credibility part of the meeting, they will say, “We need it done by next month,” and we will say, “We don’t do things like this by next month. It takes a long time.” One of the issues with profitability is that if you want to save money, you need the project manager or an executive in that group to do their job. Your job is to manage the team in a way that you reduce risk.
If you’re going to want to do 112 gigs in 2020, you should have been working on the test fixtures—the compliance, firmware, and all of that—three years ago. You can’t just pull a stack-up that’s suitable; you don’t know what the routing is. You have to throw a team together ahead of time. Then, you have to do some test vehicles because each stack-up for that level is very demanding. I’m talking about 112-gig test vehicles, and the test boards are difficult. You have to do things early, manage the project carefully, and engage in systematic design.
Shaughnessy: Planning ahead is a big part of it, especially for cutting-edge designs.
Neves: Absolutely. Let’s say you wanted to climb Mount Everest. That’s a daunting endeavor. You may go to the general area of Nepal and hang out at 7,000 feet. Next, you may want to go to base camp, which is around 14,000 feet, and you’ll stay there for weeks or a month. Then, you set up a series of camps up the mountain to help mitigate your risk. You’re climbing the mountain in a very systematic way, which reduces your risk, and you can acclimate to the problem. That’s how advanced signal integrity is; you have to acclimate to the problem.
That’s my idea of designing for profitability. If you embark on a project and don’t know what you’re doing, you delay the silicon in the marketplace, have a lot of people running around like crazy, and have to spend a lot. There’s nothing more costly.
To read this entire interview, which appeared in the March 2020 issue of Design007 Magazine, click here