Issue link: https://iconnect007.uberflip.com/i/1361971
APRIL 2021 I PCB007 MAGAZINE 13 Thompson: Exactly. Matties: When you were in fabrication, how did you look at process life cycle? Because if you want to change a chemical supplier, that's a large undertaking, a lot of qualifying. It's not an easy or light decision that you would make. Thompson: Or a material manufacturer. In fact, frequently we would have issues with materi- al manufacturers. Even with the same prepreg ply from the same manufacturer I would have erratic results in signal integrity and controlled impedance. As an example (and this materi- al manufacturer will remain nameless), I went to a trade show and I said, "Look, I'm getting a 1-mil difference in dielectrics of prepregs straight across the board." Styles 106, 1080, 2116, 2113, 7628 all had a linear 1-mil difference. And they said, "Oh, that's because one of our products is made in Guangdong, and the next one is made in Shenzhen. One of them has a dry curing process and the next one has this other kind of a curing process." I said, "at's the problem. You need to give me a sole source. Make your mind up and tell me where you're going to give me the product from and keep it coming from only that particular facility." You can't mix them because I'd never be able to shoot and hit my controlled impedances trying to go with two different materials from two different facilities with two different cure cycles. It never worked. Even with the same prepreg ply from the same manufacturer I would have erratic results in signal integrity and controlled impedance. or am I just going to be building a four- or a six-layer consumable home product and some- thing simple like that? It really depends great- ly on what the product is that I would be look- ing to build. In fact, I've heard the story many times over the years where a designer will liter- ally camp at a fabricator and nurse a job all the way through the process. And by saying that, what they mean is if there need to be changes—CAD changes, where you have to move a part here or you have to reduce the size here, or you have to move a trace here or there—then they'll do it on the fly at the fabricator. I've heard that many, many times. My brother-in-law did that for many years. He would go to a large company like Compaq and literally camp out for a week in a hotel, just go- ing back and forth through the shop and mak- ing sure they were still proceeding forward with the board. Matties: When we look at fabrication, we're really breaking this down into multiple areas: investment and equipment, and technology. Happy says that there are particular process life cycles. Happy, maybe you want to explain a little bit about what a process life cycle is from your point of view? Happy Holden: In terms of developing a prod- uct, it helps in evaluating or putting in a new process if you actually have a process with milestones from doing beaker level or qualifi- cation tests all the way up through a pilot line and then expansion to medium or higher vol- ume production. But there are specific deliv- erables and specific milestones that are useful to have, including the process of obsolescence. As technology changes, how do we ramp out of this and into the next one? Like you said, Barry, everything should be defined down to a process. New chemical processes and new plating or new materials had a very defined process lifecycle to meet those milestones and deliverables rather than everybody just invent- ing their own way to do process engineering.