R&S MXO3 Oscilloscope for EMC measurements: part 2

Rohde & Schwarz recently announced their MXO3 1 GHz bandwidth 12-bit oscilloscope [1], and I managed to get one to review. The R&S MXO38 is ideal for EMC troubleshooting and characterizing design issues early. The 1 mV low noise vertical sensitivity, 12-bits, 500 Mpts memory depth, and 21 ns trigger re-arm allow a terrific FFT spectrum display. The waveform capture is an amazing 4.5 million waveforms per second, providing real-time capture of up to 99%. From an EMC point of view, this provides a nearly real-time spectrum capture.

In Part 1 of this series, I discussed how to use near-field probes to help identify high-frequency harmonic energy sources on PC boards. Characterizing each energy source, such as processors, memory, and power conversion, helps identify the source of coupled EMI being radiated from your product. This is “Step 1” of my three-step process for troubleshooting radiated emissions issues and has been pretty well documented in other articles and industry application notes.

What has not been as well covered is my “Step 2”, how to use RF current probes to further characterize attached cable harmonic currents and reveal “red flags” in your designs. The ability to identify potential coupling mechanisms from observing harmonic energy on attached cables is an important part of the troubleshooting process. In Part 2 of this series, I’ll show how I use one of my most important tools for troubleshooting radiated emissions issues.

RF current probes

I suspect most product designers are familiar with the smaller current probes designed for oscilloscopes or digital multimeters (DMMs). These typically have smaller apertures that fit a wire or small cable and generally extend from DC to 100 MHz, at best. There are also current probes for electrical measurements with larger apertures that range up to only a few MHz and are really designed for mains frequencies.

RF current probes, on the other hand, are designed to measure microamps of RF current from kHz to hundreds of MHz flowing on I/O and power cables. They usually have a hinged aperture that can accept everything from a single wire to large-diameter cables (Figure 1). When their 50Ω port is connected to a spectrum analyzer, you’ll observe an RF spectrum similar to that when using a near-field probe. Many manufacturers make these probes, but for this article, we’ll use the affordable Tekbox Model TBCP2-750 ($879). See [2].

Cable radiation is usually the dominant reason for radiated emission failures. The problem is that the various harmonic energy sources measured on your circuit boards or system can couple to any attached cables, resulting in high-frequency harmonic currents that now create an antenna effect, causing radiated emissions.

We’ll use the RF current probe to characterize and reduce these coupled RF currents by clamping it around each I/O and power cable in turn and recording the spectral characteristics for each. The typical RF current probe is sensitive enough to measure µA of RF current, and only 6 to 8 µA of harmonic current can fail the FCC class B limit.

If cables are the dominant cause of emission failure, then I often leave the probe connected and fixed in place while I’m trying various mitigations back in the PC board or system. Observing changes in real time is quite efficient!

RF current probe measurements

The RF current probe is merely a current transformer that measures RF currents in the primary (wire or cable to be measured) and couples that to the secondary, which is loaded by the 50Ω input impedance of the spectrum analyzer (Figure 2). This produces a voltage across 50Ω that is usually in terms of dBµV. Switching the MXO38 spectrum display from the default dBm to dBµV allows us to calculate and directly compare to the emissions limits, as you’ll see further on.

I usually insert a bit of “bubble wrap” within the probe aperture to keep the wire or cable centered and away from the metal probe case in order to minimize measurement errors.

Because of resonances on cables, it’s best to slide the RF current probe back and forth on the cable or wire in order to maximize the dominant harmonic or harmonics. Once the larger harmonics are maximized, I tape the probe down to the table to minimize variables while trying different mitigations to reduce cable coupling from the board.

Mitigations could include rerouting internal cables, improved bonding of cable shields to chassis or digital return plane, adding or improving common mode filtering at the I/O or power connectors, shielding energy sources using local shields, etc.

Caution

RF current probes are so sensitive that when clamped around a wire or cable, that cable tends to act as a receiving antenna. Ambient transmissions, like AM/FM broadcast stations, two-way communications, digital TV, and cellular phones, can confuse measurements from the equipment under test (EUT). It is important to be aware of certain bands in the RF spectrum where these services reside, so they can be differentiated from product emissions.

To ease this confusion, I usually clamp the current probe around the cable under test with the EUT turned off. You’ll like to see at least the FM broadcast stations in the region 88 to 108 MHz. Placing these captured signals in Max Hold mode provides an initial plot of current ambient signals. On the MXO3, this can be saved as a “reference” trace, as you’ll see in the next section.

Making a measurement

We’ll be measuring the common-mode harmonic currents on the power input cable of a 1 MHz GaN buck converter. This converter has a strong 230 MHz ring frequency that I use to demonstrate how ring frequency can manifest as a broad peak in the frequency spectrum. We’ll use a Fair-Rite 43 material snap-on choke to demonstrate a reduction in harmonic currents in the cable. Figure 3 shows the test setup.

As mentioned above, when using the RF current probe to test common mode cable emissions outside of a shielded chamber, I first perform an ambient measurement with the EUT off. This will record and save the specific RF environment so you can better differentiate EUT emissions from ambient signals.

First, connect the current probe to channel 1 (C1) and set the Termination to 50Ω. Leave all other settings to their defaults. Turn on the EUT in order to set up the oscilloscope for an acceptable time domain signal. I usually just press Auto-set at first.

Then turn on the Spectrum mode (Menu > Spectrum), turn On the Display, set the Source to C1, touch Start/Stop and set the frequency limits to Start = zero, Stop = 1 GHz, Normally, the EMC standards have you set the resolution bandwidth to either 100 kHz (military) or 120 kHz (commercial), but I find 1 MHz may have a cleaner averaged signal for troubleshooting purposes. Turn off Auto BW and set RBW for 1 MHz. Press Max Hold to save the ambient plot.

Now, switch the Menu to Apps > Reference and select R1 (reference waveform 1), turn Show = On, and select Source = S1MaxH. Select Create/Update and Save As “Ambient”. This will save the ambient Max Hold plot as reference 1 (R1).

Note that every time you reopen the Reference screen, the Source keeps defaulting to C1, so you’ll need to keep selecting S1MaxH, or it will save a time domain reference for C1. This messed me up several times when trying to add reference waveforms to the spectrum display.

Next, turn on the EUT and select Max Hold, then save this as R2 [2]. This will become the saved “before” spectrum case. You can change the default color by going to Settings > Appearance, Category = Reference, and Color Source = R2. Go ahead and set the color as desired. For this article, I set it to light blue. In a similar fashion, save this as reference 2 (R2) and name it “EUT_On”.

You should end up with a screen capture as in Figure 4 with the ambient in default grey, EUT-On (“before case”) in the color of your choice, and the real-time waveform in the default yellow. From here, you can try various mitigations and compare the resulting spectrum with the saved reference 2 (R2) waveform. The ambient spectrum (R1) may be used to differentiate the ambient RF signals from the “live” signals.

The blue trace is the measurement with the EUT turned on and shows the broad peaks at the ring frequency of 233 MHz and board resonance at 558 MHz. Now, with the saved reference plots, you can commence troubleshooting and applying various mitigations while observing immediate results!

Estimating pass/fail

One important use for the RF current probe is to provide an estimate of passing or failing specific emission test limits. By knowing the dominant currents in an I/O or power cable, we can calculate the E-field at the test distance per the standard used (usually 3m or 10m for commercial products). While this won’t necessarily be precise, it still gives us a “ballpark” estimate to which we can compare with the test limit at that frequency.

Commercial RF current probes come with a calibration chart of transfer impedance versus frequency (Figure 5). Using Ohm’s Law, we can use this chart to calculate the measured common mode current in the wire with respect to the voltage measured at the probe output port, assuming a 50Ω system. This is based on work by Dr. Clayton Paul [3] and further refined by Henry Ott [4]. There are example calculations in [5] and [6].

Let’s assume we measure one of the dominant harmonics in a cable as 28 dBµV at 120 MHz at the spectrum analyzer. We can also read of a transfer impedance of 20 dBΩ at 120 MHz from the calibration chart in Figure 5.

Using Ohm’s Law, we can calculate the common mode current (Icm) in the cable:

Icm (A) = E (V) / R (Ω), or, in converting to terms using log identities,

Icm (dBµA) = Vprobe (dBµV) – 20 dBΩ = 28 – 20 = 8 dBµA

Now using the E-field equation from Paul and Ott:

where

Ec is the calculated E-field in V/m due to common-mode current flowing on the cable,

Ic is the current through the wire or cable (A),

f is the harmonic frequency being measured (Hz),

L is the length of the cable in meters and

d is the measured distance during the compliance testing (usually 3 or 10m).

Converting the measured values to basic units and plugging into the E-field equation, we get 1.26E-4 (V/m). Converting this back to log units, we get 42.03 dBµV/m. Comparing this with the FCC class B limit at 120 MHz (43.5 dBµV/m) indicates we may be just under the limit by only 1.47 dB. Typically, we’d want at least a 6 dB margin, so this is probably not good enough.

To streamline all these calculations, I developed a simple Excel spreadsheet, which may be downloaded from my website [7]. Figure 6 shows an example calculation. By entering the specific probe transfer impedance, the frequency of concern, the cable length, and test distance (typically 3 or 10m), the E-field in dBµV/m is calculated and may be compared to the appropriate test limit.

Summary

The RF current probe is not only a useful tool for general troubleshooting, but may be used to determine potential passing or failing due to harmonic currents on a radiating cable. While they may be a bit pricy, I find the RF current probe is one of my most-used tools for troubleshooting emissions.

Most importantly, if you know that cables are the dominant radiating source, all the troubleshooting can be done in-house with an RF current probe. No need to run back and forth between your facility and the 3rd-party test lab to perform this troubleshooting, saving cost and time!

I also have a short video showing how to use these RF current probes in [8].

With its fast acquisition update rate, the MXO38 spectrum feature works extremely well for general bench-top EMC troubleshooting and debugging, and it’s become one of my favorite bench-top tools. The ease of setting up the analyzer is a step ahead of other leading manufacturers. Once saved, the reference waves stay in memory along with the instrument setup, so if you have to move locations, everything is preserved.

References

[1] Rohde & Schwarz
[2] Tekbox current probes
[3] Paul, Introduction to Electromagnetic Compatibility (2nd Edition), Wiley Interscience, 2006, pages 518-532.
[4] Ott, Electromagnetic Compatibility Engineering, Wiley, 2009, pages 690-693.
[5] Wyatt, Workbench Troubleshooting EMC Emissions (Volume 2), Amazon.
[6] Wyatt, The RF Current Probe: Theory and Application, Interference Technology
[7] Wyatt, E-Field Calculator (scroll down to find it)
[8] Wyatt, Current Probe Demo, Current Probe Demo