Issue link: https://iconnect007.uberflip.com/i/1529118
14 DESIGN007 MAGAZINE I NOVEMBER 2024 Limitations of Rules of Thumb LaRont: Where do rules of thumb stop being effective? Dack: I think it's when you start getting into space or diving deep into the ocean as I previ- ously mentioned. As Kris mentioned, there are other environments as well. We might say that this is a good rule of thumb for not using rules of thumb: When human life is at stake or the end product is mission-critical, Class 3 opera- tion will be required. Moyer: Correct. So, then we have manufactur- ability and documentation. Dack: Right. Documentation consists of graph- ics, material, and process notes and dimen- sions which include tolerances. Moyer: Also, graphics, data sheets, and file formats for data output. Are you doing every- thing on paper or PDF, or are you producing IPC-2581 files? Another bullet point would be the manufacturability challenges for the new packaging technologies. You have difficulty in soldering, placement, and rework and repair. You're no longer a technician with a soldering iron desoldering a part. You're using a hot air station to carefully desolder the BGA or QFN apart. Now you have to carefully get your sol- der in there; you're no longer optically inspect- ing, but X-ray inspecting, and that leads us to test. Dack: Exactly. Testability is another point. With test, a common rule of thumb says that we need one test point per net, plus a bunch of other extra test point coverage on the power and ground portions. Moyer: How do you get to the test point in an old part? You just take your probe and touch the pin. It's coming off the side of the body or the bottom. If I'm doing microvias from layer 1 to layer 2 to do my escape routing, unless I come over and drop another blind via all the way to the bottom or top side of my board, where do I get my test point? I don't have any convenient features to test with. Access and density are both problems for testability of these reduced package sizes. Dack: We had a customer come in with a really small IC package, and they wanted testability. We told them, "Yes, it's good to have 100% test- ability." Well, from all those high-speed con- tacts, they pulled test points out around the IC, which made a bunch of stubs and degraded the performance. It's a challenge that design- ers have to understand. Moyer: I'll give you another one. If you're doing high-volume production with full ICT and not just flying probe test, if you put that many pogo pins in one area, you've got so much spring force pushing up. If your board's not thick enough, you actually have enough force that you can actually crack and fracture your board if they don't have enough down- ward force pins. But now you've got to have enough open board space so the downward force pin isn't crushing your resistors and capacitors and other things. " When human life is at stake or the end product is mission-critical, Class 3 operation will be required. "