Issue link: https://iconnect007.uberflip.com/i/860275
10 The PCB Design Magazine • August 2017 Andy Shaughnessy is managing editor of The PCB Design Magazine. He has been covering PCB design for 18 years. He can be reached by clicking here. • The component manufacturer's. • We put our boards through a lot of testing before they get to the field. It is usually an un- foreseen stress or other phenomenon that insti- gates the failure. • This is twofold: The last person to handle the assembly and the final failure analysis out- come results. • As far as the designer's responsibility, in most cases difficult technologies are discovered before the order is accepted, if not shortly there- after. This affords an opportunity to improve upon the original design in concert with the designer, resulting in a potential savings in time to manufacture and cost to the customer. • It could be anyone's fault. It can be the designer plus the fabricator, the fabricator plus assembly, or it could even be everyone's fault. • The designer. By designer, I mean the board architect (the guy who chooses the devices), the design engineer who does the schematics, a lack of SI/PI engineer, and the CAD engineer (who is doing a high-speed board without adequate and relevant knowledge). We asked, "If you, the designer, are truly at fault, how do you avoid making the same mistake again?" Answers included: • Keep a log of past and present faults, en- sure your checkers and design rules are properly set-up, and always check your work. It doesn't hurt to have a peer review. • We put in place and modify processes to cover the cause and verify that we don't do it again. We call it DIVE: Define the issue, inves- tigate why, verify the fix/correction, and ensure it's corrected by monitoring designs that follow. • By changing the company design proce- dures and sharing the information with my colleagues and network. • We designers expect our manufacturing counterparts to be able to inspect and verify, qualify and certify every phase of the PCB man- ufacturing process so that if a problem arises, we can issue a corrective action report. We expect the supplier to be able to comb through their processes to find out where and how a defect occurred and document how they are going to fix the process so that it won't re-occur. Aren't we designers playing the hypocrite if a layout problem is found and we cannot point to a design process? As you can see, there are no easy answers. Everyone plays a part in the board's develop- ment, and everyone is a suspect. So, this month we asked our expert contributors to discuss methods for identifying the cause of board fail- ure, and steps to take to avoid repeating this mistake in the future. (There's also a little bit of finger-pointing, of course.) First, Gaudentiu Varzaru of the Politehnica University of Bucharest discusses all the ways for problems to creep into a PCB design, includ- ing overly automated EDA tools. Scott Decker of UTC Aerospace Systems explains why no one has the right to point fingers—there is plenty of blame to go around when a board fails. EPTAC's Kelly Dack focuses on the use of root cause analysis, and how RCA can help technologists quickly find the cause of failure and make sure it never happens again. And consultant Tim Haag explains how tough it is to find the root cause. He shares a few horror stories about failures that were not his fault, despite what everyone else thought at first. And we have columns from our regular contributors Barry Olney and John Coonrod. We are also introducing Jade Bridges of Electrol- ube, whose first column focuses on coatings for thermal management. It's almost the end of summer, and trade show season is around the corner. We'll be at PCB West and SMTA International in September. I hope to see some of you on the road! PCBDESIGN MISTAKES WERE MADE