Issue link: https://iconnect007.uberflip.com/i/1116895
14 PCB007 MAGAZINE I MAY 2019 Maybe the glass has to come out, and we could have something else in there. The prob- lem is the infrastructure and vested interests are all geared around making boards the way we make them today, so any change will be a gradual change. But it's a great testament to the FR-4 materials that we've been using for over 50 years. Matties: When you mention the variation of di- rection on the laminate during board manufac- turing, is this something that a designer would be mindful of and call out to ensure that they build in one direction versus another, or is this just random based on their optimal through- put? Morgan: It's random. The designers have no idea at all. And that's why I love talking to designers and explaining to them that these things matter. Designers are aware of the is- sue and that there's a bigger problem now, but they don't tend to do much about it. The only solution that I've seen, which is a very poor one, is that they rotate the circuitry around 15 degrees on the panels. Instead of the risk of traces running vertically above or coinciden- tally with the glass fibers, they always run at an angle across them. You get an average dis- tribution between the two traces of a differen- tial pair. There's no potential for having a very high glass content or very low glass content, which is very wasteful. You waste a huge amount of the real estate on the laminate material because it's square. Some OEMs will call up and specify glass fabrics that have directionality they can live with or square weave, etc. But many of the people design- ing boards don't have a clue about these things. The boards may come out fine one time, but they may not another time. My job for the last few years has been to try to explain to de- signers that these are things they need to be careful of that may impact their design. When they're building toler- ances, they should make sure to allow for this as well. aments. So, you end up with a very dense rope of glass with an area next to it of resin, then a dense area of glass again. Imagine those ropes being round. If you can make them flat, you can make them more oblong in shape or el- liptical and can spread the fibers out and get a more uniform distribution of fibers, so that's another option. The next thing you can do is change the properties of the glass. Most of what we use is E-glass—electrical-grade glass; it's the most widely used glass fiber in the world for every- thing, but it has a certain set of properties. If you change the formulation of the glass, you can tweak some properties and improve things that you want to improve. You can reduce the dielectric constant and loss. The entire supply chain is making further tweaks, and they'll carry on doing so. The question is, "At what point will this not work anymore?" With copper foil, there are now treatments called ANP, which means "almost no profile," so it's almost no profile—submi- cron. And if you talk to companies like Circuit Foils, you'll see that they've been developing these for some years and are coming into the market now. We use lower-loss resin, low-loss glass, and flat copper. The so-called "copper crunch" was supposed to be upon us a few years ago, but we're still using copper. We have to replace copper with something else, though; graphene has been mentioned. Ventec International Group lab.