Issue link: https://iconnect007.uberflip.com/i/1116895
34 PCB007 MAGAZINE I MAY 2019 Senese: Right, and that's exactly where we sit. One particular high-visibility change over the last few years is the use of low Dk glass to re- duce the Dk and loss in high-end materials. You can only get so much out of the resin. This is a fairly easy change to make for the lami- nators. It's expensive because the glass costs more than E-glass. But when designs go from 6 Gb to 12, 25, and 100 over the next 10–15 years, the loss has to go down. Right now, most of the low-loss material in the market is E-glass like MEGTRON 6. The next generation is moving to a low Dk glass, which would've been the same that was used by NELCO-13SI 20 years ago. That glass has ex- isted in some volume for a long time, and it's only now beginning to be needed for certain speeds. It's still a difficult glass to get; It's on- ly about 1–2% of the glass made for our business. Certain types of equipment in the server and switch business needed to have that extra 20–30% lower loss, so they've started building it. We build it every day: our MEGTRON 7N uses a low Dk glass that has not yet transitioned to mass produc- tion like E-glass. Even for us–who make quite a bit of low-loss material–it's still nowhere close to the volume of MEGTRON 6, and it won't be for some time. We have two partners that we buy that ma- terial from. Until very recently, we didn't have supply issues. Last year for the first time, be- cause several other OEMs adopted this at pret- ty much the same time without doing much research into the supply chain, there was a shortage of this glass. The shortage affected everyone differently, but during that time, I heard all kinds of stories from people. Most of them were made up, but some of them are true. People make promises based on what you say minus some guesstimate. As a supplier, if I believed what every end user or fabricator tells me, I would predict hundreds of points per year of expansion in a particular market, but we know that doesn't happen. You have to plan your business around something that makes sense, so you're a little more conservative. Sometimes, that conserva- tism is wrong. As a group, we haven't learned to manage the risk of developing technologies in the supply chain very well. We are work- ing very hard with suppliers, as opposed to 10 years ago, to give them up-to-the-minute and up-to-the-month information about where things are going. What's happening right now is that the second gen- eration of low Dk glass is being developed by two different sources. However, it's being looked at by anybody who makes glass for the same reason that we're looking at mak- ing higher volumes of low-loss material and lower volumes of FR- 4—because it's how you survive and make money in the box that we've ped ourselves into. Johnson: Tony, how do you see the industry getting out of this box? Senese: The industry may not need to elimi- nate this box, but the industry has to work to- gether to do a better job communicating needs at a very basic level– monthly, daily, and week- ly forecasting from the whole supply chain as well as a more honest and rigorous discussion of megatrends. One of those is the advent of 5G. There are devices within this realm requir- ing low-loss material and maybe even high-fre- quency materials like PTFE, but there will be a constant pressure to drive down the cost. In the past, that has created opportunities for companies that can make materials and boards that have a low-cost constituent in the digital portion—a hybrid circuit board with FR-4 and PTFE, for instance. Or in the case of high-speed circuits, for a long time, hybrids with FR-4 in the ground power pairs have been designed in