Issue link: https://iconnect007.uberflip.com/i/808383
26 The PCB Magazine • April 2017 FABRICATORS SPEAK OUT ON HIGH-SPEED MATERIALS an RF board is like playing craps. The difference between FR-4 boards and RF boards is that the FR-4 boards are like playing craps Saturday night in Laughlin, Nevada, which is a $5 min- imum bet, where RF is like playing the Bellagio at 9 o'clock on Saturday night where the minimum bet is $25. Same rules, same game, but the buy- in is huge because of the registration and the other fun stuff that goes with it. You can lose fast or you can win a lot. That's one of the greatest challenges—the material itself, its softness, its registration, as well as processing. One of the things is hold- ing registration and the material. The other parts that are a challenge for the materials are the ability to deposit electroless and maintain- ing peel strength. A lot of them have low peel strength. Many RF materials can have smear. We call it flat, but it's smear when you drill and you cover the inner face of an innerlayer pad; it violates the IPC-6012 rigid PCB performance spec. It's allowed somewhat in IPC-6018, the performance spec for high-frequency/micro- wave PCBs—but smear is always a concern on a lot of high-speed materials. Some of the other challenges with materials are Teflon or PTFE surfaces, where you have to do special treatment to get the soldermask or legend to stick to it. Matties: When you talk about the registration, how do you mitigate the problem? Partida: What we do is we use software that collects data on every panel we build, every resin system, and every lamination cycle. This information is a feedback system to predict fu- ture builds to be on target the first time, to re- duce the risk of losing material. Sometimes, if we look at a structure we've never seen before, we'll do a pilot run to dial in the scale factors, or we work with a customer saying, "You know, you want a Class 3 board but this is an engi- neering board. Can we have Class 2 as we learn the scale factors for this construction and this stack-up?" Sometimes they'll work with us on that. The RF community really works very well with the printed circuit board manufacturers. We have to work together in order to be success- ful, so we use those tools. The other challenge, too, with a lot of the high- speed RF materials or parts is that there are unbalanced constructions. For example, one of the first boards we built was a 1 to 2, 1 to 5, 1 to 7, 1 to 9, 1 to 11, 12 to 14 and then final 1 to 14 lamination structure. With our predictor, we got on target the first time. Matties: Wow. Is this over-the-counter software or is this proprietary software that you've developed? Partida: It's a software that I believe won an award at the APEX show, three or four years ago, for the best product of the show. The company's name is XACT PCB. When you're working with high-speed RF with high frequencies, there are so many different combinations of materials that the engineers use that you're always chal- lenged, whether you have it in your database or you're learning for the first time, in which case the predictor tells you, "I've never seen this be- fore. Good luck." Matties: What about from the flex side, any chal- lenges for you, Joe? Joe Menning: On the flex side of things I would echo a lot of the same concerns. Even though we don't do a whole lot in the high-speed ma- terials area, I'd echo the concerns on the pro- cessing. Obviously in the flex world we're used to materials moving and changing shape, but I would reinforce the fact that a lot of the high- speed materials that we've had experience with are even more challenging than typical flex ma- terials. Regarding the comment about PTFE, it's a great high-speed material but it doesn't bond well to anything, really, unless you have some sort of pre-treatment, so it's got a lot of chal- lenges on that front. Matties: In the flex world, you've been dealing with material movement for a long time. How do you deal with the registration issue? Gerry Partida