Issue link: https://iconnect007.uberflip.com/i/1400005
20 DESIGN007 MAGAZINE I AUGUST 2021 quency signals we are working with just by using today's parts. To me, it's very important that you learn to discern which information is right for the par- ticular board the designer is working on. For example, datasheets will tell you you're sup- posed to do A, B, and C to design the part onto the board. Many engineers have told me, "e datasheets say A, B, and C, so it must be right." In reality, the information might not be right on this board. When the designer is just rely- ing on the datasheet or the application note they are not getting the whole picture. ey're getting the part of the picture that the manu- facturer made work in one particular setting, on one particular board, with one particular stackup. e designer's board, stackup, and everything else may be very different. It reminds me of when you see a soware demonstration; the soware always works beautifully when they demonstrate, because they have it set up perfectly. But when you sit down and try to do the same thing, it doesn't always work well. Shaughnessy: Right. What do you think about specific CAD tool training from the EDA com- panies? Webb: Some of them have conferences and classes, and they try to teach things above and beyond the basics of their tool, and that's a very good thing. Everybody needs to understand the basic information. But if you can go to training where they teach about their tool and some more advanced information as well, then you've gotten two good things in one because you understand why to do things a certain way. As an example, we all know that soware companies push their autorouters. I have a friend who has worked with four or five differ- ent companies as part of their autorouter devel- opment teams. I recently asked him what he thought about autorouters in general, and he said, "ey're probably not a good way to design a board." He said they are not able to keep up with current high-speed design technology, so they are always doing a catch-up kind of work. I have always felt that way, but I was sur- prised someone who was in the development of those tools said that as well. I put this infor- mation into one of my classes, because I know that there are people who want to plug the board into an autorouter and come back the next day and it's finished. Routers don't under- stand about return current and they don't know about where the energy of the signal really flows from layer to layer in the stackup. ose things need to be controlled by the designer. In my opinion, autorouters can't do nearly as good a job on complicated boards as the expe- rienced designers can. Shaughnessy: Right. Webb: You asked about four-year colleges, and their education curriculum for design. ere is currently very little about PCB design in most four-year engineering degree programs. e PCEA is working with some colleges to develop a semester-long course about PCB design to help engineers understand what is needed aer they graduate. It is a work in progress, so definitely watch for that. In the meantime, there is the one-week class the PCEA group of designers is doing with EPTAC. ere are also people who teach one-, two-, and three-day classes about PCB design as well. We are not trying to be all things to all people, certainly, but we're trying to be the mentor that isn't there anymore. To me, it's very important that you learn to discern which information is right for the particular board the designer is working on.