Issue link: https://iconnect007.uberflip.com/i/978458
22 PCB007 MAGAZINE I MAY 2018 Down at sub-six gigahertz, there's not so much of a dramatic change in the electrical requirements of the materials, but one of the things that is happening is, perhaps not nec- essarily in the very first iterations, but certain- ly in newer designs that we see coming out on the horizon, there's a lot more integration between the antennas and the power compo- nents and the transceivers, and going forward even more integration with the high-speed dig- ital parts. So, you see much more complex multilayer print- ed circuit boards, and a lot more integration. You see these modular stacks starting to look a bit more like high- speed digital boards. And so, the manufacturability of the PCB becomes more crit- ical, and the ability to make much more complex multi- layer PCBs becomes more critical than it was from 2, 3 and 4G. I think that's how I would sum it up. Shaughnessy: Ben, why don't you tell us a lit- tle bit about yourself, then just give us a few thoughts on 5G. Ben Jordan: Sure. My background is more dig- ital than anything. I was an FPGA program- mer before getting into the EDA space. For me, RF has always been this scary black box thing. And I know that 5G is going to affect many people, especially at the PCB design lev- el. That's because those people don't necessar- ily have any knowledge or experience with do- ing RF PCB design or RF system-level design. But the whole point of moving to 5G is to enable many thousands of additional devices to join the network, where the whole push for this from my point of view is so the Internet of Things could go to its next stage of evolution. There are going to be many thousands of de- vices needing reasonable bandwidth, a lot of them are going to be thirsty, and a lot of devic- es will have very, very small bits of informa- tion. So, if you read up on 5G, you read that there could be many hundreds of thousands of simultaneous connections from sensor arrays, or it will be used as a sensor network. May- be the flagship consumer will still be mobile phones, but the same networks in our homes, in our streets, in our cities, are going to be 4G or 3G, will be completely saturated if all those different IoT sensors and services that we're dreaming up were to have to use those older technologies. That's really what's driv- ing this and what's pushing bandwidth to be increased to accommodate this. What do you do to increase band- width? You make much more complex modulation schemes, and you also must increase carrier frequency. So now with 5G we're going to be seeing in the millimeter wave, bandwidths up to 60 gigahertz, for example. What does that mean for PCB de- signers and people like me who don't know a lot about RF? It's kind of scary, if you think about it. I think, to a large extent, we have been rely- ing on our chip manufacturers, companies like NXP or Broadcom and Qualcomm, to come up with the goods in a very easy-to-consume sort of package, so that a lot of those materials and a lot of the RF signal pathways are contained on a chip, and the antennas are there. One of the good things about high frequency carriers is the sizes of the antennas can shrink. One of the other things about 5G is its ability to go peer to peer. Then various nodes on the net- work will be able to quickly negotiate direct links and communicate peer to peer and not require a base station all the time. Some of those bottlenecks will disappear and it will become more of a true mesh. And in do- ing so, we can have lower power requirements because there's less signal strength required in many instances. So that can help with the hundreds or thousands of sensor arrays or oth- er smart devices. In terms of PCB design, re- John Hendricks