Reading time ( words)
At the IEEE International Microwave Symposium (IMS) show in Atlanta, I met with John Coonrod, technical marketing manager with Rogers Corporation and a Design007 columnist. I asked John to discuss the papers he presented at IMS and to give us an update on Rogers’ materials and technology.
Andy Shaughnessy: Good to see you, John. How’s it going?
John Coonrod: Very good. It’s great to see anybody, even at a trade show.
Shaughnessy: Yes, it is. Now, I understand you’re giving a couple of presentations at this show. Tell us about them.
Coonrod: Sure. I gave a presentation yesterday; it was a MicroApp, so these are kind of informal presentations, but hopefully with a good amount of information. Yesterday’s presentation focused on thermal management, and that’s one of those topics that just doesn’t go away. And of course, as the technology advances, there are always more things to be concerned about with thermal management. I explained the basics of heat and how heat flow works, how it works in circuit form, and then some of the things that you want to be mindful of. Everyone knows that you need to have a material that is high in thermal conductivity, of course, but there are some other tricks out there too, like the via form. That problem has been around a long time; the plated through-hole vias from one layer to another act like a heat channel.
And that works well, too, but sometimes if it’s under a chip, then some of these chips are not friendly to that because you’re basically grounding out. You’re having a plated through-hole from the signal plane to the ground plane, and some chips don’t like that. Some chips with the ground pads, no problem. And when you can’t use that, there’s a neat trick that I ran across called a thermal fence. That’s what we’d call it. Instead of a microstrip, it would be a stripline. On both sides of the signal on one plane, you would have grounds with plated through-hole vias. So now you have a lateral path for the heat to flow from the signal to both sides and then down, and that makes a pretty big difference.
Yesterday, I talked about the techniques you can use for managing the heat and the heat flow, and some of the basic ideas behind it. One of them is insertion loss. Of course, people don’t stop and think about that sometimes, but as you have more insertion loss, you will have more heat generated when RF power is applied, of course. So that’s kind of a juggling act. If you get a material that’s very good for insertion loss, but not so good for thermal conductivity, then sometimes it works out pretty well, but you just have to understand the trade-off there.
Shaughnessy: Thermal management has been a problem for a while now, but people didn’t really deal with it before, right?
Coonrod: Right. That’s what I’ve seen, too. In the past, the people who really knew thermal management well, were the engineers working with power amplifiers and these applications were pushing some serious power. But now with millimeter-wave and a lot of the things that they’re doing, even the 77 gigahertz radars that you wouldn’t think have that much power to them, they are generating a fair amount of heat, and they have to do something with that. And besides the heat management, the heat itself can actually affect the circuit performance because all materials have a TCDk, which is how much the Dk will change with the change in temperature. If you can minimize that change in temperature, then you have less of an issue there, too. As technology evolves, it’s one thing after another for thermal management.
Shaughnessy: And speaking of millimeter-wave, you also did a talk on millimeter-wave.
Coonrod: I did. We’ve been working with millimeter-wave a lot over the last few years for automotive radar, 77 gigahertz, 79, or even 24 gigahertz and 60, but I’ve been getting more and more questions from the industry about using stripline and millimeter-wave. Normally they’re using microstrip as part of a multilayer hybrid, so the top layers would be the microstrip RF, and the others would be usually FR-4 or something like that. There’s a lot of interest. SIW is an interesting topic, but stripline is a good one because it offers you a lot of benefits. The big problem with stripline is trying to get the energy from the outside world inside the signal plane that’s buried, and that signal transitions to major headaches.
In my presentation, I explained different methods to make that transition a little quieter for the RF signals, so you have better return loss. I gave some pointers about some of the things like unwanted resonances that can pop up if you’re not careful with the design, spurious wave modes, and how to deal with that. It’s a variety of things related to stripline at very high frequency, and I also gave some examples of it, as well.
Shaughnessy: And the millimeter-wave is important to 5G.
Coonrod: That’s true, and it seems like it’s big now, but I’m seeing more and more of that for future applications, so 5G is evolving still. That’s my perspective anyway. It seems like there’s more and more of the millimeter-wave coming out with 5G. And that’s a good thing. I did a paper some time ago on the transition from microwave to millimeter-wave. A lot of the designers may have been thinking about microwave, and then when they transition to millimeter-wave applications they have to consider some other things that were not problematic at lower frequency. That’s another thing to be concerned about, of course, when you have a lot of designers designing at lower frequencies, and they jump to millimeter-wave, there’s some gotchas in there.
Shaughnessy: I think a lot of people that read your column may not be in microwave, but they’re seeing a lot of microwave-like issues.
Coonrod: That’s true. High-speed digital is another area I’ve been getting more interested in and working with a lot more. It seems like, as high-speed digital is getting faster and faster, they’re having to incorporate more of the millimeter-wave techniques. They’re trying to get cleaner signals. Well, for one thing, the PAM4 format for high-speed digital is actually several layers of digital bits. And that is actually very sensitive to insertion loss and insertion loss differences. At millimeter-wave frequencies, insertion loss can vary for a lot of different reasons. That’s another thing I’ve been looking into, as well.
Shaughnessy: So, how did you all handle the last crazy year? Were you already working remotely?
Coonrod: You know, I work in the lab a lot. I have a special arrangement because we have different groups at Rogers. One group is work-at-home only, another group needs to be at the business unit, so they do go to work and I’m in the group that goes to work as needed. It’s usually lab-related. So if I have customer issues pop up or something going on with research and development, then I’ll go in and do some tests in the lab. I’ve been kind of coming and going, but it’s nice to get out and about again. This is the first time in a year and a half I traveled.
Shaughnessy: Me too.
Coonrod: It is just really nice, actually. I expected the show to be a little slow just because you’re not going to have a lot of international travelers and all that, but it’s really not been a bad show. It’s not back to normal yet, but this is a good step in the right direction.
Shaughnessy: Exactly. Well, great talking to you John. Maybe I’ll see you at DesignCon.
Coonrod: Good talking with you, Andy.