Issue link: https://iconnect007.uberflip.com/i/1514189
JANUARY 2024 I DESIGN007 MAGAZINE 13 of different plating processes. Seeing boards travelling down a conveyor in the etching lines and watching copper disappear as the board is slowly sprayed with the etching chemicals gives credence to the necessity to give our fab- ricators margins that are not arbitrarily small. is all made me realize that there was a lot of physics going on inside any PCB. It was shortly aer that I started building my Rolodex of peo- ple whom I would reach out to when I didn't know something. Fab shops are just rife with people who want to help you design boards that will go through their facilities smoothly. Matties: What did you learn in a few facility visits that changed your approach to design? Bell: It changed my perspective, which of course changes your approach. It shied again once I became attached to APCT. As a designer at Plexus, we designed boards and had all kinds of process written around design- ing for manufactur- ability (DFM), design- ing for fabrication (DFF), and designing for assembly (DFA). We called our preferred and qualified fab houses, ask- ing about their capabilities. But on the assembly side, we didn't have 10 different CEMs; we had one: "Here at Plexus, we design for assembly, and these are the rules that we use. If we can't implement these minimums, we won't be able to build it efficiently." We had to take painstaking efforts to make sure that our DFA was in very good shape. at said, there are three main aspects for a successful design: You design for electrical characteristics, for your ability to fabricate it, and for your ability to assemble it. At Plexus, electrical characteristics and assembly tech- niques got prioritized. If we needed to build complexity into the design, we had a matrix to select fabricators' capabilities to fit the design. e reality was that if a smaller shop couldn't build this board, we could go to a shop with greater capabilities. Your perspective changes when you've visited fab shops and talked with the engineering and sales folks at fab shops who really know their processes. ey want to help you. It is okay to ask for grace when getting technical questions from the fabrica- tors. ey are building, in most cases, the most expensive component to any PCBA BOM, so it's good to have a relationship with your fab- ricators. Push when necessary, but give them what they need if it isn't impactful to other aspects of the design. Shortly aer I became the manager of the team, we implemented a training program for junior engineers that talked about those provoc- ative questions in fab. For example, we would ask a junior hardware engineer, "If everything else on the board is the same, and you have a 0.5 mm pitch component and a 1 mm pitch component, which one will you pick?" Most of the newer hardware engineers would pick the smaller pitch, because that means that the designer could have more room to do what they do. But inadvertently, this can drastically increase the cost and complexity of the fab by simply making an uninformed decision early in component selection or concept convergence. Happy Holden: I've seen many PCB design problems that were related to the assembler. I especially dislike assemblers doing printed circuit design, because they favor assembly, and when they favor a certain pitch or pro- cess capability, they push it off to the fab- ricator. The assembler is simply refusing to improve his process. Instead, he wants every board to cost 15–20% more. He saves 5%, which gets distributed over the entire volume, but now every board will cost more. Of course, assemblers don't like me saying this. For instance, at HP, we started using microvias, and we developed an assembly We called our preferred and qualified fab houses, asking about their capabilities.