EMI emissions testing: peak, quasi-peak, and average measurements

Article By : Dorine Gurney

Learn the difference between these three EMI emissions measurements and when to use them. It's not just for EMI engineers.

EMI emissions testing is critical to bringing a product to market. Most electrical and electronic products require regulatory compliance testing, which must be performed in an approved test lab. Such testing can be expensive, especially if the product fails. A failed test can result in lost time to market and cost overruns.

To minimize unwanted trips to the test lab prior to compliance tests, you can measure EMI emissions from your product. Not only will this increase your chances of passing emission compliance testing the first time, it can help you identify potential issues early on and reduce the need for last-minute product redesigns.

There are several measurements you can do using a spectrum analyzer or EMI receiver.

Quasi-peak, peak and average
In a precompliance test, you scan your product for noise and compare it to limit lines described in various standards based on region and class of equipment under test. In the U.S., that’s FCC Part 15 for commercial devices. There are many other standards that are industry specific such as for automotive and military products. The standards also specify how to measure the noise. Most commercial standards use three major detection methods defined by CISPR 16-1: peak, quasi-peak and average. Bench and USB spectrum analyzers and EMI receivers from numerous suppliers have this measurement capability.

Peak detection, as its name implies, retains the peak value of each harmonic in an emitted signal, indicating the worst-case scenario. Average detection provides the average amplitude of each signal component across its period. Quasi-peak detection weighs each component based on its repetition rate: the faster repetition rate, the higher the weight given to that component.

Figure 1 shows the three detectors output response to two similar pulsed signals, with the top one having a higher repetition rate. Notice that the quasi-peak detector has a higher voltage output.

Figure 1. The quasi-peak detector generates a higher voltage output when the event occurs more frequently.

To quickly perform EMI scans, you should start with peak detection for precompliance testing. Compare the results against the published quasi-peak limits. If the design meets the quasi-peak limits, with lots of margin to spare (say 6 dB or more), the product is likely to pass compliance testing. Thus, and quasi-peak testing may not be necessary. If, however, the product marginally passes peak detection, quasi-peak testing is recommended. This is the same approach used in compliance testing and will assist in identifying potential areas of concern.

Precompliance testing
Reproducing the exact same test conditions found in a compliance lab at your own location is nearly impossible unless you’re willing to spend heavily on facilities and equipment. Fortunately, you can still increase the probability of passing EMI compliance testing by investing in lower cost equipment and facilities.

The ideal location for testing is in a radio free environment, far away from cellular, base stations, broadcast stations, or any other ambient signals. One suggestion is to set up radiated emission testing at an external site such as a large empty warehouse. The ideal setup avoids walls with hidden metal harnesses or objects that can create reflections and may null some of the faulty frequencies and/or amplify some “passing” ones. Reducing those reflections can be expensive because you need to absorb emissions to prevent reflections. Many engineers will build a shielded room using chicken wire to at least reduce the effects of outside signals. It depends on your budget. You will also need to invest in antennas, a tripod, and an amplifier in addition to a spectrum analyzer. Many antennas are available, but if you’re on a tight budget, you can opt for using a DTV antenna or make them from PCBs. You will need a good isolated table, an isolation transformer, and a line impedance stabilization network (LISN). Figure 2 shows a typical radiated test setup using a USB-based real-time spectrum analyzer.

Figure 2. A test setup for radiated emissions conssits of an antenna, a spectrum analyzer, an isolated table, a LISN, and possible a preamp.

Start by scanning the ambient environment with your device in position but turned off. The first scan for radiated emission should be using peak detection because it’s fast and gives a quick overall perspective of how far the noise amplitude is from the limit and which frequencies are a concern.

Figure 3 shows an ambient trace and a first peak scan compared to the limit line of FCC Part 15 Class A up to 960 MHz. The ambient noise is the teal trace and the peak scan is the yellow trace. The table shows the results of peak scan versus the limit line. As you can see, the results show a large number of spurs in the A region as well as a failure.

Figure 3. These results show a large number of spurs in the A region and a failure.

The scan in Figure 4 uses quasi-peak detection across region A (zoomed up to 59 MHz) to get more precise results and is more in line with the official testing done at an accredited test lab.

Figure 4. The quasi-peak test shows that the devices barely passes and most likely will need to be modified.

Fig. 4 shows the results for peak (green trace) and quasi peak (yellow trace) versus the limit line. In this example, the device is now passing with a margin of 1.5 dB. This is probably not enough to ensure passing at the test house. To account for inaccuracies in the pre-compliance test setup, it is best to have more margin. Ideally, you’ll want at least 6 dB of margin, if achievable.

In this case, the device will likely need to be modified to achieve more margin, before it will be ready for testing in the regulatory compliant environment of an EMC/EMI test lab.

A word about spectrum analyzers
For EMC/EMI pre-compliance testing, it is important to select a spectrum analyzer with software supporting quasi-peak, peak and average detection. Many instruments support this function. When designing and testing for compliance, you may need to use all three of these tests to get an accurate picture of your product’s ability for it to pass.

There’s simply no getting around the need for radiated emission testing. Pre-compliance peak and quasi-peak testing help ensure that excess EMI emissions don’t leave you stranded at the compliance test house.

—Dorine Gurney has over 15 years of experience as a product planner at Tektronix. She is now retired.

Related articles: