Issue link: https://iconnect007.uberflip.com/i/978458
26 PCB007 MAGAZINE I MAY 2018 Jordan: I think the issues with assembly re- main the same for any kind of wireless tech- nology. There's going to be DFT; it's going to create new problems and new opportunities, again because it's new physical connectivity and modulation schemes, and new rules from the FCC and CE and Europe and others, for making sure products don't cause problems. And all of that impacts assembly, in my mind anyway, because assembly is the first task of that final stage of production. The second half of that stage, which is done typically in the assembly process, at the end, is the test. And designing for test is going to have new chal- lenges with 5G. I mean, how are you going to test that your devices are too chatty when it's supposed to be functioning and not using too much bandwidth on a network in a cer- tain geographical region where there could be 1,000 other devices? How do we test for that? And how do we design the PCB at those millimeter wave fre- quencies; the probe effect is going to be a huge problem. You know, the probe effect is that by measuring something you affect how it oper- ates. So how do you measure? And that hap- pens at those kinds of frequencies. So, the as- sembly houses are going to have to develop and buy new test equipment that's going to en- sure these products pass muster. Not just func- tionally, but that they pass the rules. Maybe I have a naïve idea of that, but that's just my two cents on that question. Las Marias: I think you're right, Ben. Thank you for that. Shaughnessy: Happy, is there anything that we're missing? What do you think about the 5G? Happy Holden: At Hewlett Packard, we typically would do testing for up to 800 gigahertz. When you're designing and building boards, and we have to measure something that's in 800 giga- hertz, you have to be 10 times better. And so, RF design and RF materials were always a big headache, but that's where HP made a lot of money, because not too many people could build this test equipment. Now this kind of specialty, rather than being with the OEMs, is going to come down to common board fabrica- tors and board assemblers that may not have the manpower, the equipment, the training or the knowledge. For a while there's going to be an elite few that will have mastered all these needs. And then because of the opportunity and the prices, more people will jump in with more processes and more materials. But for the short term, this is a tough area. You know, I took a degree in electrical engineering, but because of the mathematics and field theory, I stayed away from Maxwell's equations. Unfortunate- ly, it's coming that even digital is RF design, and you can't escape. The only way to escape is to go optical; then you don't have to worry about it at all, because there are no magnetic fields. So, one of the al- ternatives may be that a lot of 5G may jump into the optical area, simply because of the in- herent problems and shortages in the digital electronic area. I think that's what makes it an interesting subject. Going forward, there are an awful lot of challenges. Shaughnessy: Right. It seems like you're just now getting to the point with millimeter wave where they can commercialize it for main- stream. I keep reading about people that have had a hell of a time with millimeter wave, but it seems like they've got it under control now. Holden: UCLA demonstrated for us a millime- ter wave chip that had only three connections, and it had replacement mechanical connec- tors. Because they were so directional that the transmit and receive had to be surface mount- ed on the edge of the PC board and you didn't need the wire or mechanical connectors be- cause of the millimeter wavelength. A millime- ter chip with its own antenna right off the die itself—and they're so small you can put quite a few on a wafer. Designing and the architecture of products, they change. The follow up, we're not using mechanical connectors with sockets or anything like that anymore.