Issue link: https://iconnect007.uberflip.com/i/1307491
NOVEMBER 2020 I DESIGN007 MAGAZINE 29 Hofer: That is absolutely correct. Holden: In fact, the military is kind of trailing everybody else. Shaughnessy: You do a lot of the stuff in PTFE, and that offers its own challenges with it being hard to register, and it's kind of squishy. How does that affect the way you create vias? Hofer: As we get into PTFE, dimensional stabil- ity becomes an issue, and not necessarily Z-axis dimensional stability, though there is a bit of that, but very much so in X/Y stability. And when we're doing landless vias or microvias, that makes it even more challenging because with the technology today, every methodology for plating and/or filling those vias requires a mechanical planarization methodology, and that just doesn't work well with PTFE. We have to walk a fine line between being able to planarize the surface, making it flat, without disrupting the X/Y stability of the PTFE. I look forward to a technology that will allow us to fill those vias without the need for planariza- tion. There have been a couple of tries at that technology, but they've not quite gotten it. Holden: Strangely enough, that was the first microvia board we built in 1982: an eight-layer PTFE multilayer with a metal core, cavities, and selectively plated pure gold with micro- vias to do wire bonding, attaching the chips in the cavities down to the metal core. If I cir- culated that board today, nobody would quote on it. They would say it's impossible to build. But that's vintage 1982. Hofer: Yes. I find that to be a little more manu- facturable with the metal core because it helps hold the stability down in the processing of the subsequent microvias. But we do a lot of cavity boards with wire bonding outside of the cavity because with many RF components, the shorter the leads to the pads, the better. Hav- ing the heat absorption of the core that the chip can sit right on is extremely beneficial. Holden: There was always a nickel barrier. Hofer: We, too, prefer a nickel barrier, but we find a lot of our RF folks are not liking the nickel because of the added loss, so we're doing a lot more EPIG and EPAG now, where they can still wire bond without the use of the nickel underplate. Feinberg: One of the things that I've always tried to push with our people as a supplier back in the day was communication with the designers. The designers never communicated much with the suppliers. You really don't have much communication with the designers? Hofer: Once we have an established customer, our dialogue with the layout and the electri- cal engineer and even the mechanical engineer increases dramatically. It's the referrals and the new folks. With their first order, we get their files and say, "Wow, I wish you would have talked to us a few weeks ago." But with our current relationships, the majority of my day is spent on the phone with either the lay- out person, the EEs, or the MEs, or trying to get the project managers on the line. I like that because that's a huge benefit to both the cus- tomer and us. Let's face it, even though board shops all basically do the same thing, there are different nuances between shops. Getting to talk with a customer before the data is over the fence helps them understand what your processes are. I always make it a point to help them understand what other folks' processes are as well so that they design not only so that I can build but also so that other folks can build, because I realize that I'm not going to be everything to everybody all the time. Let's face it, even though board shops all basically do the same thing, there are different nuances between shops.