Issue link: https://iconnect007.uberflip.com/i/1337116
FEBRUARY 2021 I DESIGN007 MAGAZINE 15 failures." I was shocked, but once I learned more details, I found that over 30% of the boards were failing to get through fabrication because of problems with footprints or things like that. Now, we go back to cause and effect. "Okay, what's the cause?" It was our multiple libraries. As a matter of fact, we had 1,123 libraries that we were working from in Legrand, and you cannot manage that many libraries. You just can't do it. You must have a singular library that can be managed. One Friday evening, aer everybody went home, I deleted every single one of their libraries and their backups, and I purposely came in late on Monday morning. I knew what was about to go down. As soon as I walked in, the whispers began: "He's here." Everybody was in the conference room, and I said, "As you probably have now realized, all your libraries are gone. We're starting over, and we're starting with a singu- lar library which we can manage and control." I knew it was the only way you could fix the effect. e effect was bad boards, and we had to go back to the root cause. Matties: One thing you did was to require everybody to stop doing their busy work and take that time. ey're under time pressure, and a lot of companies may not be willing to take that time, I would think. Watson: Yes, you've got to look at the whole picture, though. "Okay, we're losing time rede- signing a board." Matties: People don't stop to delete thousands of libraries because the thinking is they don't have the time to do it. Watson: Exactly. Why is it that we have time to do something again, but we don't have the time to do it right the first time? We kind of compartmentalize our processes and we don't look at the big picture. We always talk about saving money. How much money was wasted in an engineering drawer of boards that were just wasted? e other comment I wanted to make about libraries and footprints is you've got to have a single library, period, because that's what's managed and what's controlled. Now, there are changes that happen all the time on libraries. IPC, for example, just came out with 7351C. ere were some real major changes they did with footprints from revision B. at's where they brought in rounding off the pads and dif- ferent things like that. How do you make those changes in your library when you've got 1,000 libraries? Shaughnessy: Nobody likes dealing with that stuff. How did they get to that point? How would you know where to look and which library to look in? I mean, I guess that's a whole other string of nomenclature or what- ever. What a nightmare. Watson: It was actually a living nightmare. Usu- ally, every design that went through had its own library, and every designer had their own library, what they liked, what they preferred, and what their standards of a component were. I've been doing design for 20 years. When we first began, we used to take a resistor that con- sisted of a symbol and a footprint. We would lay down 0603 resistors of a certain value, we would just copy that same resistor, lay it down, change the value—just copy and paste. at's the way we did our schematics. Now, compo- nents have so much in them, which is phenom- enal. is is really nice. When you have a component, what are we looking at? We're looking at the symbol and at the footprint, but now we're also adding in parametric information in that component. Now we have sourcing, we have the manu- facturers who will build that component for you who will be your vendors, who will sell it to you. Now we've even tied in component dynamic information: What are your vendors' quantities right now and what is their pricing?