Issue link: https://iconnect007.uberflip.com/i/1348195
38 DESIGN007 MAGAZINE I MARCH 2021 world, especially in the design of printed cir- cuit boards. ere were so many times where the inaccuracies of the data that I was given to work with caused problems, that I eventually learned to double check and ask questions of everything first. Here are some of the prob- lems that I've encountered. Design Database When I first started laying out boards many years ago, some designers found the automated features of the tools too restrictive and would disable them. When one of these designs later required an ECO, it was a nightmare to work on because it had been routed on a 1-mil grid without any online DRCs. Aer having to completely re-route a few of these designs to make a simple change, I learned to work with these designers before they started to ensure that database was set up correctly. Another lesson I learned was to make sure that the design rules were correctly set up before I started working on an existing design. Many designers wouldn't take the time to set up their constraints and would alter trace widths manually for nets such as power and ground. If you weren't paying attention to what you were editing, those traces would snap back to their default skinny width, and the DRCs would be none the wiser. I also learned to look at those designs for areas of metal that were altered using simple graphical shapes instead of intelligent cop- per pours. Although those planes and thermal pads looked okay, there were oen vias or pins going through them, causing a short to a dif- ferent plane on another layer that the DRCs wouldn't catch. Design Libraries Laying out boards at a service bureau back in the day oen meant re-creating library parts for each new customer. In some cases, the designers would cut corners and use the incor- rect method for creating a pad or shape in their footprint that either wouldn't photoplot well, or it would register a DRC incorrectly. Check- ing the integrity of new library parts became part of my regular routine. On the flip side, it also paid to make sure that a designer was using the correct part rather than merely copying an existing footprint without checking it first. We oen found prob- lems with polarity or pin-outs that could have resulted in an unexpected re-design of the board and an angry customer. I also learned to keep an eye out for the odd kind of library mistakes, like a surface mount pin being created with a through-hole drill attribute. Although this would eventually light up a design's DRCs like a Christmas tree, it would still take time to correct and was always better to catch it before it got into layout. Design Information A good PCB layout depends on getting com- plete and correct data to work with. In the service bureau we would work from a netlist that the customer gave us, and we got used to having to add bypass caps and test points, as well as to double check net names. ankfully, with the integrated design systems we work with now, you typically don't see these kinds of problems anymore. I also learned to check the components in the netlist against the parts listed in the bill of materials. It was amazing how many times this didn't match, and we had to go back to When I first started laying out boards many years ago, some designers found the automated features of the tools too restrictive and would disable them.