Pre-Compliance Testing for Conducted Emissions

expert insights

EMC Bench Notes

Equipment Needs (AC Mains)

By Kenneth Wyatt

re-compliance testing aims to duplicate the test setup used by your third-party test lab. Fortunately, setting up a conducted emissions test in-house is relatively simple and can be performed on the benchtop (Figure 1).

Figure 1: The commercial conducted emission measurement setup as specified by CISPR 11 or 32.

Ideally, you should procure a copy of the appropriate EMC test standard used, depending on the product type. For example, for military testing, you’ll need a copy of MIL-STD-461. For commercial, industrial or medical products, you’d use one of the IEC standards, such as IEC/EN 61326, IEC/EN 60601 or the generic IEC/EN 61000-6-3, which will refer back to CISPR 11 or CISPR 32. For automotive modules, you’ll need a copy of CISPR 25. These will describe the equipment and setups and test limits required. I also have more detailed information in my EMC Troubleshooting Trilogy, Volumes 1 and 2 [Reference 1].

The purpose is to measure the noise voltages (as referenced to ground or chassis) for the line or neutral circuits. The return path for the common mode currents is shown, and we’ll discuss this in a later article.

First, let’s discuss the various noise voltages produced by common line-operated switch-mode power supplies. The conducted emissions test measures noise voltage EMI, that is, the emission voltages measured from line to ground (or chassis) and neutral to ground (or chassis). This is depicted in Figure 1 and is measured using a line impedance stabilization network, or LISN.

A typical conducted emission plot of a line-operated switch-mode power supply is shown in Figure 2. This was measured using an EMZER EMScope EMI receiver with built-in LISN.

Typical conducted emission test plots for a line-operated switch-mode power supply. In this case, the line to ground noise voltages are compared to the CISPR 32 test limit. Peak (blue), quasi-peak (green), and average (red) plots are shown.

Figure 2: Typical conducted emission test plots for a line-operated switch-mode power supply. In this case, the line to ground noise voltages are compared to the CISPR 32 test limit. Peak (blue), quasi-peak (green), and average (red) plots are shown.

Equipment Required

Let’s start off with the basic equipment you’ll need. This will include a good spectrum analyzer or EMI receiver, a line stabilization network (LISN) and a metallic ground plane over the top of a test bench or table for the equipment under test (EUT).

There’s one caution regarding the more affordable analyzers. Because they rely on a common superheterodyne topology, they tend to have local oscillator leakage or feed through, which manifests as a large peak at the “zero” frequency and an associated “skirt” adjacent to it.

Ideally, the test should be conducted inside a shielded room in case of a strong nearby AM radio station transmitter or other nearby ambient transmission sources within the measurement sweep range. However, for troubleshooting or pre-compliance purposes, this is not required. Just be aware of any local transmitters that could appear in your plots.

Spectrum Analyzer: Figure 3 shows an example of an affordable bench top spectrum analyzer. You’ll want to specify an analyzer with the required test frequency range as specified in the appropriate standards your product requires. Most of the conducted emissions tests will start from 10 kHz (military) or 150 kHz (Commercial) and stop to at least 30 MHz (10 MHz for military).

Figure 3: An example of an affordable spectrum analyzer usable for pre-compliance testing of conducted emissions.

There’s one caution regarding the more affordable analyzers. Because they rely on a common superheterodyne topology, they tend to have local oscillator (LO) leakage or feed through, which manifests as a large peak at the “zero” frequency and an associated “skirt” adjacent to it. This skirt is caused by phase noise from the PLL local oscillator. The width of the peak and skirt depends on the resolution bandwidth (RBW). If you’re testing to the military MIL‑STD-461 starting at 10 kHz, this LO leakage may be an issue in resolving the lower emissions when using one of the affordable analyzers. Fortunately, a 9 or 10 kHz RBW tends to minimize this issue.

For commercial testing, this is not nearly as much an issue, but you’ll still see part of the peak (or associated skirt) at the 150 kHz start frequency. The lab-quality spectrum analyzers won’t generally have this issue due to better internal shielding. In the case of conducted emissions, you definitely get what you pay for.

The example in Figure 4 shows an expanded plot from 9 kHz to 1 MHz showing the zero-frequency peak with associated phase noise “skirt.” A Siglent SSA 3032X analyzer was used for the measurement. The marker at 150 kHz indicates the commercial starting frequency. Notice it’s riding on top of significant phase noise, which appears as a rise in the noise floor of the measurement. Typically, the measurement of switch mode power supplies will be relatively large at this low frequency, so this rise in noise floor may not really be an issue.

Expanding the lower frequency range for conducted emissions shows the “zero” frequency of the LO as well as the associated skirt that raises the noise floor by 30 dB at the 150 kHz commercial lower frequency limit.

Figure 4: Expanding the lower frequency range for conducted emissions shows the “zero” frequency of the LO as well as the associated skirt that raises the noise floor by 30 dB at the 150 kHz commercial lower frequency limit.

Figure 5 shows the same frequency range but with a much-reduced phase noise. A lab-quality Signal Hound model BB60D was used. The 150 kHz marker is about 30 dB lower than that shown in Figure 4.

The same measurement setup as Figure 4, but using a lab-quality Signal Hound model BB60D analyzer. While you can still see the zero-frequency peak, the associated skirt is much less perceptible.

Figure 5: The same measurement setup as Figure 4, but using a lab-quality Signal Hound model BB60D analyzer. While you can still see the zero-frequency peak, the associated skirt is much less perceptible.

Line Impedance Stabilization Network (LISN): Line impedance stabilization networks (LISNs) are used to help match the power line impedance to 50 Ω for the purpose of measuring the conducted emissions emanating from DC or line-operated products. They come in AC 50 µH (90 to 270 VAC) and DC 5 µH (at up to 100 V or more). The µH reference refers to the inductance used in the LISN. Most are designed to measure from 9 kHz to 30 (or 100) MHz.

Figure 6 shows the LISN (Tekbox TBCL08) I like to use for AC line-operated products. It has a switch for line and neutral measurements and can handle up to 8 amps of line current.

An affordable LISN. It has a switch that measures either line or neutral noise voltages and also includes a built-in 10 dB attenuator and transient protector.

Figure 6: An affordable LISN. It has a switch that measures either line or neutral noise voltages and also includes a built-in 10 dB attenuator and transient protector.

Ground Plane: For accurate measurements, you’ll also need a ground plane under the test setup. In the past, I’ve simply used heavy-duty aluminum foil taped down to the benchtop. The LISN and spectrum analyzer should be bonded to this ground plane to allow the noise currents a return path back to the source.

Currently, I’m using a Tekbox model TBGP “roll up” ground plane over the top of my plastic table top. This comes in a 250 x 140 cm roll, and the 250 cm dimension nearly fits across my 6-foot table. I keep the excess rolled up behind the test setup. See Figure 7.

Emissions setup for measuring a common LED light bulb. The equipment is sitting on top of the Tekbox ground plane, and both analyzer and LISN are bonded to it.

Figure 7: Here is my conducted emissions setup for measuring a common LED light bulb. The equipment is sitting on top of the Tekbox ground plane, and both analyzer and LISN are bonded to it.

Isolation Transformer: Some standards require an isolation transformer be connected between mains power and the LISN. It is used to isolate the conducted emission measurement from local power line noise. While this is specified by some EMC standards, like MIL-STD-461, CISPR 11 or 32 does not specify one. I find it is optional for pre-compliance or troubleshooting tests on the benchtop. I’m using a Solar model 7032-3 for any serious military testing.

Transient Protector: A transient protector at the spectrum analyzer input is highly recommended to protect the sensitive RF front end from transients caused by applying or removing main power or when switching from line to neutral measurement. Many of these also include a 10-dB attenuator as extra protection. I like to use the Tekbox TBFL1.

Summary

This summarizes the most basic equipment needed for setting up your own commercial conducted emissions test. Use the References section to find out more about the products mentioned. See my previous article on how to use spectrum analyzers for EMC measurements (Reference 5).

References

Wyatt, EMC Troubleshooting Trilogy (Volumes 1, 2 and 3)
Saelig Electronics (U.S. distributor for Tekbox and Siglent products)
Siglent Technologies
Tekbox Digital Solutions
Wyatt, “How to use spectrum analyzers for EMC,” In Compliance Magazine website, March 7, 2024.

Share this story:

Back to Issue