There’s a lot of EMC testing that focuses on the 30 MHz to 1 GHz range, specifically radiated emissions tests like FCC/ANSI C63.4 or CISPR 25. There are plenty of requirements that go up from there (up to 18 GHz for MIL-STD-461 RE102 testing or 40 GHz if you’re really unlucky), and some that go down from there, such as the conducted emissions 150 kHz – 30 MHz test for FCC/ANSI C63.4. But then there are some tests that look at the 9 kHz – 30 MHz range specifically in radiated terms, like CISPR 36 taking measurements from 150 kHz – 30 MHz using loop antennas. In these standards, limits are written in A/m for magnetic fields instead of the V/m electric fields we’re all more used to. Why?
It all comes down to the difference between the “near field” and “far field”. (For a lot more detail on this topic, see Ken Javor’s four-part series on near field measurements in this very magazine, starting here.) We think of electromagnetic waves as being very consistent over space. If we measure only the electric field (E-field) coming from a device under test (DUT), then we’ll get good information about both the electric and magnetic fields (H-field) associated with the unit. This works because the H-field and E-field are related to each other by the wave impedance number, which is a constant (120π or 377Ω). But this is only true if you are taking your measurement in the far field distance from the DUT.
As you can see in Figure 1, if you’re very close to a particular source, you’ll get very different information if you measure the E-field only or the H-field only. The relationship between them is not a constant number until you get a certain fraction of a wavelength away, generally considered to be roughly a sixth of a wavelength (λ/2π). Any closer than that, and the wave impedance number will still be varying with distance. In Figure 1, that looks like a well-behaved function, but from the kind of accidental emitters found in EMC testing, it’s a bit more chaotic. You certainly would not want to measure the E-field at a distance of λ/10 and make a prediction about what the H-field is.
For the frequency range most common in EMC tests, the lowest frequency is 30 MHz with a wavelength of 10 m, which makes the λ/2π distance about 1.6m. The 3 m test distance of ANSI C63.4 is well outside this (and the fact that the 1 m test distance of RE102 and CISPR 25 is inside it is one reason the lower frequency ranges of those tests are so frustrating to deal with). But in general, most EMC radiated emissions tests are done in the far field, measuring more-or-less well-behaved plane waves.
However, when you’re dealing with a source that draws a lot of current at a moderately low voltage, and you’re worried about victims that may be very close, you need to do some magnetic field-specific testing. The wave impedance is defined like any other impedance, Z = V/I. Something like an electric vehicle (EV), drawing say 20 A at 300 V, could have a wave impedance of ~15 Ω, much lower than a 377 Ω plane wave. (Compare this to a well-designed aerospace antenna that might drive 28 V into the RF feed but only draw 50 mA—that would have a wave impedance of ~560 Ω, much higher than a plane wave.)
In order to properly characterize the threat from a low-wave-impedance source, like an EV, you need to take measurements with antennas optimized for the magnetic field (loop antennas), at low frequencies (9 kHz – 30 MHz), at close range (less than 1.5 m)—exactly the set up for a CISPR 36 test.
Measuring E-fields in the far field works for the majority of consumer and military hardware developed in the 20th and early 21st centuries. But with EVs becoming more common (for aircraft as well as terrestrial cars), measuring the H-field in the near field is more appropriate.
