Probes let you home in on the source of emissions that can couple into functional circuits. EMI can cause your product to fail compliance tests or cause circuits to malfunction.
EMI from on-board DC-DC converters is a common problem with IoT products. These small circuits generally switch between 1 MHz and 3 MHz and use very fast sub-nanosecond edge speeds. The result is broadband EMI often extending above 2 GHz. This EMI can affect the sensitivity of sensitive receiver circuits, especially cellular and Global Navigation Satellite System (GNSS).
One useful way to characterize the EMI performance of a DC-DC converter is to measure the rise time and ringing using a small H-field probe in the time domain. This may be done non-invasively by coupling an H-field probe to the converter output inductor (Figure 1).
It is important to characterize ringing on the switched waveform, because this ring frequency can translate to broad peaking in the emission characteristics. H-field probes are quick and safe because they don’t require direct connection to the circuitry – just couple it to the DC-DC converter output inductor.
The Rohde & Schwarz HZ-15 near-field probe kit (review) includes several H-field probes (loops). Because we want to couple to currents in traces and components, that’s the type to use. The largest one may be too sensitive and thus has less resolution that you may need to isolate the source of emissions. The next smaller one (model RS H 50-1), which is about 1 cm in diameter, is about right to identify and characterize EMI at the board level. Simply connect the probe to a 50-Ω oscilloscope input and adjust for an adequate displayed waveform.
[Editor’s note: EMI probe kits are also available from Beehive, Com-Power, ETS-Lindgren, Keysight Technologies, Langer EMV, TekBox, Tektronix, and others.]
To prove this characterization measurement, let’s examine the math (Figure 2). There will be some unknown mutual coupling factor (M in the equations below) between the inductor and H-field probe. Because we don’t know the mutual coupling factor, the amplitude won’t compare with actually measuring with an oscilloscope probe. For EMI purposes, we’re mainly interested in the rise time, general switched wave shape, and ringing frequency, if any.
A DC-DC converter usually has a near square wave signal (VL) from the converter switch node (SW) and output inductor (L) input to ground return and this is what we’d measure with an oscilloscope probe. The current through the inductor is related to that voltage as:
Assuming the H-field probe is held close to the inductor, you get some mutual coupling, M (unknown), and the output of the probe is:
Combining the first two equations results in:
Therefore, factoring out the constant, M/L, we see VOUT ∝ VL.
Because VOUT is proportional to VL, the most important characteristics for EMI are now easily and quickly measured without the risk of shorting connections with oscilloscope probe tips. By using the H-field probe held close to each DC-DC converter inductor, you can measure the rise time (indicates the upper range of harmonic frequencies), pulse width and period (also factors into harmonic frequencies), and ringing frequency (which can cause broad resonant peaking in the broad band spectrum.
Figure 3 and Figure 4 compare the switched waveform characteristics from a RT-ZS20 1.5 GHz bandwidth oscilloscope probe (with short probe tips) and the RS H 50-1 H-field probe. The measured results are comparable, except for amplitude.
With the same H-field probe connected to a Siglent SSA 3032X spectrum analyzer with start and stop frequencies at 1 and 500 MHz, respectively, and with a 120 kHz resolution bandwidth, you can the 8 MHz resonance within the resulting broadband spectrum (Figure 5).
In many cases I’ve seen, this ring frequency can easily be in the 100s of MHz and cause major broad peaking in the emission spectrum, which can cause EMI failures if coupled to an efficient antenna-like structure (cables, typically).