Issue link: https://iconnect007.uberflip.com/i/1210212
FEBRUARY 2020 I DESIGN007 MAGAZINE 23 emissions are. You can look for trends, but knowing if it's going to pass or not is really hard to do. You look for the known problems and elimi- nate those known problems. Sometimes, you can look at the near-field radiation, which is one of the things I show in my classes with a board that has good returns and bad returns. And with just a little sniffer probe near the sur- face, you can see the bad returns radiate a fac- tor of 10 or 20 times more near-field than the region of the layout that has good returns. Just using a near-field probe will give you an idea of how bad things are. It comes down to look- ing for the root cause and identifying it, and then fixing the problem. Shaughnessy: I hear some people say that they don't see EMI issues below 50 or 75 kilohertz. Is there a definite cutoff for EMI? Bogatin: The FCC started at 30 megahertz, I be- lieve, but there are some low-frequency emis- sions tests that are a lot lower in frequency. I'm not as familiar with these tests. I've heard that there's a specification for very low-frequency magnetic field radiation or near-field radiation in the hundreds of hertz to kilohertz, but again I'm not sure in the specification. However, if you have the same antenna, the same length of wire, the same current, and you just jiggle the frequency, the far-field radiated emissions drop with frequency. Going to a lower frequen- cy means you're going to have a lower far-field electric field. Dan Feinberg: Eric, is it a combination of fre- quency and power, or is it just specifically the frequency? In other words, if the power that the signals are being generated at is much low- er, would the frequency be less important? Or is it different specifications, different limits? Bogatin: Yes, I think so. When I talked about power before, it's the radiated energy, the pow- er in the radiated energy, not in the voltage or current. Feinberg: Exactly. Bogatin: So, the current doing the radiating is oscillating back and forth, and the FCC require- ment says in this frequency there are five or six different frequency ranges. Within each range, there's a set limit to the maximum far-field electric field strength. If you just take that same current and slosh it around the far-field electric field strength, it's going to be proportional to the current multiplied by the frequency. In the past few years, there's one area in PI and radiated emissions that I think is really exciting. Everybody uses switch-mode power supplies because they're really efficient. But the problem is you're switching lots of cur- rents around, and current switching is a reci- pe for radiated emissions. That's why switch- mode power supplies are notorious for radiat- ing and failing FCC tests. I think it was Linear Tech who was bought by Analog Devices, and they came out with a line of this switch-mode power supplies. They're called Silent Supplies, and they have done some clever engineering on the inside of these modules to dramatically reduce the opportunity for radiated emissions. Shaughnessy: What would you say are some of the biggest mistakes that engineers and design- ers make as far as EMI? Or some things that de- signers should do to preclude EMI? Bogatin: This is where Mentor came up with their design rule checker with a list of design rules. Todd Hubing also has a list of dos and don'ts of how to reduce these problems for reducing radiated emissions by how you do the layout. They may be four or five differ- ent things. Number one is don't have signals cross gaps in planes, so if you have a split pow- er plane, for example, then you don't want to have a signal use that split as a return. Another one is when you have a multilayer board, and a signal goes from one layer to another layer, you want to have a return via in proximity to the signal via, so that says you want to use ground as your return and drop the ground via between the two ground planes when the sig- nal switches. With connectors for coax cables, you want to use a 360-degree connector and make sure