Issue link: https://iconnect007.uberflip.com/i/1479191
SEPTEMBER 2022 I PCB007 MAGAZINE 43 at's not only good for your costs and better for your yields, but it's better for the capacity of the industry, because now half the materials you once required are available for other designs. Where's that confluence of what designers can do to change how they do their job and materials come together? Strubbe: It really hasn't changed much over the years, it's just the requirements which have changed. Fundamentally, addi- tive is nothing new; it was built in Kollmorgen (PCK) in the '70s. e process has improved and it's better. You can start knocking layers out if you can use 2-mil lines and spaces and run them 30 inches. Not all designs can. eir sig- nal will survive that tiny little trace, but if you don't have to worry about current, which is with a lot of applications, they all work just fine. If you look at the technology and materials from a designer standpoint, the drivers are primar- ily signal integrity (SI) and power/heat. We're seeing more applications of coins and thicker planes. Because there are more electronics being packed in the same area, the heat goes up. Korf: e SI engineer says, "I need to lower my losses and I can only get so much out of this material, so I must keep my copper traces wider, maybe make them a little bit thicker. I can only go down so far in layer thickness to increase density, but if I go too far, my imped- ance drops down. at creates another prob- lem." A lot of the work is trying to control some constraints from timing, from a loss bud- get standpoint, from the trace with materials. When materials run out, now you start play- ing with the trace and that's where you get difficulty because you can only bring a trace width down so far and that creates another set of problems. You have plated through-holes to transition signals between layers. ese holes require pads which consume a significant amount of real estate. e SI folks are continu- ously looking at copper: copper roughness, the grain structure of copper, and the roughness on all four sides. e challenge is to make traces wide enough to reduce losses, but that causes all these density questions. e challenge from the designer standpoint is to keep the traces wider because of SI reasons but they've got to keep their 50 ohms, which constrains two of the five con- straints. I can get you a lower-loss copper foil, lower-loss glass, and lower-loss resins. It's just a matter of how big your checkbook balance is. Signal integrity drives most of these changes. Strubbe: On the automotive side, it's a matter of asking, "I've got a lot of current, so what do we do with all this current we need to get from the batteries back to motors?" at's their chal- lenge. e sensors and everything else, exist- ing materials, could be tuned from a cost stand- point for the various frequencies required— you see a lot more people in the microwave space now because it's getting so huge due to automotive. It's a matter of the signal integrity standpoint from the designer, "I need to keep my traces wide, but my BGA pitch is going down." It's a natural conflict. We're seeing some 3D technologies coming out now, and it's true 3D. ey give designers a little different option, but it has its constraints too. Johnson: Is TUC working in the 3D space? Korf: ere may be development that can't be talked about yet. Most of the materials are based on inks. Laminators are starting to pro- vide solutions in the 3D space. For a designer, it's the traditional challenge: they must get all the traces in. Signal integ- John Strubbe