Today’s fast-paced, high-growth electronics market requires electromagnetic compliance (EMC) test facilities to develop shorter product test cycles while maintaining test accuracy. Commercial test requirements recommend emissions prescan measurements to help reduce overall test times and improve facility throughput. However, these measurements can introduce frequency and amplitude errors into the suspect list.
This article provides an overview of the errors associated with prescan analysis and how to overcome them using spectrum analysis and intermediate-frequency (IF) spectrum monitoring capabilities.
Benefits and Challenges of Prescan Analysis
Making final EMC measurements can be a time consuming process. Making a final measurement at every frequency in the compliance frequency range, whether there is an emission or not, is extremely inefficient. Conducting prescans to create a suspect list of emissions from a device under test (DUT) allows you to focus final measurement efforts only on frequencies with known emissions – saving time and money.
However, prescanning methodologies can introduce frequency and amplitude errors into the suspect list, preventing you from accurately characterizing DUT emissions. Prescan measurements may not accurately represent the characteristics of peak emissions because of the way stepping and sweeping receivers collect and display information as well as the characteristics of the emission itself.
Prescan Errors Caused by Receiver Function
When scanning for suspect emissions using a stepping or sweeping receiver, measurements are made at every resolution bandwidth (RBW) spacing and typically at fractional resolution bandwidth spacing. For CISPR-based emissions scans from 30 MHz to 1 GHz using a 120 kHz CISPR RBW, prescan data is typically collected at every 60 kHz (two points per RBW) or every 30 kHz (4 points per RBW).
Amplitude and frequency errors are introduced when the emission frequency is not exactly at the receiver tune frequency, in the center of the RBW. When using a stepping receiver, the measured amplitude will be less than the maximum emission level due to the filtering effects of the RBW. For example, when using a 120 kHz (-6 dB) RBW, a measurement scan at every 120 kHz will record an emissions signal that falls exactly between two measurement points at an amplitude 6 dB lower than the maximum. The result is a suspect that may fall beneath the limit line, when it is actually over it.
Overall prescan amplitude accuracy can be improved by using a greater number of prescan data points. In the example given above, doubling the number of data points by measuring at every 60 kHz results in a recorded amplitude error of 1.5 dB. However, this approach adds to overall prescan measurement time, as the required dwell time needs to be added for each measurement point, reducing the benefits of prescan. In addition, the stepping receiver reports the amplitude at the current measurement frequency. In the above example, the lower amplitude would be reported as being 60 kHz away from the actual measurement frequency.
Figure 1 illustrates the potential frequency and amplitude errors as a function of prescan step size for a 120 kHz RBW. This example highlights the worst-case where the emission frequency falls exactly between two receiver measurement frequencies, and the reported value is 6 dB lower than the actual emission level.
When using a swept receiver to make this same measurement, the maximum emission amplitude will be captured correctly but the displayed emission frequency will be recorded at the measurement point, offset from the actual frequency. Sweeping receivers record all the amplitude information as the RBW is swept through the frequency range, [±(step size)/2 around the measurement point] but they report the maximum amplitude emission as if it occurred at the selected measurement point. Using the example above of a measurement at every 120 kHz and an emission exactly halfway between two points, a sweeping receiver would report the correct emissions amplitude as occurring at one of the selected measurement points, a 60 kHz offset.
In order to ensure that both the correct emissions amplitudes and frequencies are accurately recorded during final measurement, you must review each signal in the suspect list using a single-frequency measurement and adjust the list frequency as needed prior to making a final measurement.
Prescan Errors Due to Signal Variations
Frequency and amplitude values measured during prescan are dependent on the characteristics of the input signal during the measurement period. When DUT emissions have significant amounts of frequency and/or amplitude modulation, the emission values in the suspect list recorded during prescan can differ significantly from the peak emission values. If the emission being measured for the suspect list does not exhibit its maximum amplitude during the dwell time of the prescan measurement, the amplitude recorded for that frequency will not accurately reflect the true emissions profile of the DUT.
An example of this type of emission is one with a significant amount of frequency modulation, such as a local oscillator, clock frequency, or digital power supply frequency that employs frequency modulation (also known as “dithering”) to spread out the spectral energy in order to reduce its peak emission profile. This is a common design technique used to help products meet compliance limits.
Ensuring the Accuracy of Prescan Data Prior to Final Measurements
Prescan amplitude and frequency errors call for close examination of each signal collected during prescan prior to making a final measurement. For commercial compliance, CISPR requires monitoring of the weighted emission amplitude before final measurement to characterize the amplitude change as a function of frequency. It also requires adjustment of final measurement dwell times based on DUT amplitude variation. Modern EMC receivers offer two capabilities that facilitate this examination—full spectrum analysis and IF spectrum monitoring.
Full Spectrum Analysis
Spectrum analyzers are the most powerful tools available for observing the characteristics of an emission. They offer a full complement of detectors, resolution bandwidths, and video bandwidths to analyze a suspect signal. Multi-trace capability allows you to observe both the instantaneous value and envelope of a modulated signal. Spectrum analyzers also have powerful marker capabilities which allow you to easily identify peak emission frequencies and emission envelope bandwidths.
However, commercially-available spectrum analyzers only have one downconversion signal path, preventing you from both rapidly scanning a suspect signal and measuring the single-frequency weighted detector value of the suspect signal at the same time. Scanning with both peak and weighted detectors simultaneously is possible, but a weighted detector can significantly slow down the marker update rate used to follow the amplitude fluctuations of the peak emission.
IF Spectrum Monitoring
IF spectrum monitoring provides both simultaneous single-frequency weighted measurements and span-limited spectrum displays of suspect emissions. The instrument’s local oscillator (LO) is fixed at the frequency of interest and the receiver measures the single-frequency amplitude at the center of the IF passband, typically on a meter. The instrument also displays the spectral content of the information within the IF bandwidth around the center frequency using a fast Fourier transform (FFT). This makes it is easy for you to manually tune the receiver to peak the emission response while observing the spectral content. In addition, receivers that offer multi-trace capability for the IF spectrum monitoring display provide views of both instantaneous spectrum and signal envelope.
While powerful, IF spectrum monitoring also has limitations. The displayed frequency span is only a portion of the receiver IF bandwidth. In addition, available bandwidths are limited to values less than the RBW used to measure the center frequency amplitude, typically the RBW required by the regulatory agency. For these reasons, IF spectrum monitoring is not as flexible as full spectrum analysis for performing diagnostics on the suspect signal.
Using Spectrum Analysis and IF Spectrum Monitoring to Improve Prescans
The value of these prescan analysis tools is illustrated in the following example. Consider the measurement of a 64 dBuV emission at 500 MHz signal with an FM modulation deviation of 5 MHz and AM modulation of ~25 dB. This signal is injected into a receiver and measured using a 120 kHz CISPR bandwidth, 2 points per RBW, and a 62 µs dwell time (equaling a scan time of 1 s, just over the minimum CISPR scan time of 970 ms for this frequency range).
With a single-scan prescan measurement, the signal is recorded as a 500.6 MHz signal with amplitude of approximately 58 dBuv (Figure 2). Without further analysis, a final measurement at this frequency would not fully capture the maximum value of the emission. The modulation would cause the measured frequency and amplitude pairs to differ on subsequent scans. In the worst case, the signal would not even be considered as a suspect signal if the AM modulation caused the measured prescan amplitude to be below the limit line.
Viewing the emission in more detail with either spectrum analysis or IF monitoring capabilities built into modern EMC receivers clearly indicates the modulation and the correct emission amplitude and frequency. Figure 3 displays the signal in more detail using spectrum analysis. In Figure 3, a max hold (blue) trace is used to display the modulation envelope of the signal and an additional clear write (yellow) trace is used the instantaneous value and frequency of the signal. Marker capability, available in all modern EMC receivers, can then be used on the max hold trace to identify the correct peak-detected amplitude and frequency to be used in the final measurement with a weighted detector. Engineers and technicians can also use the full power of the spectrum analyzer to do additional diagnostic measurements on the emission to better understand the cause of the emission, leading to a repair and a product that passes compliance testing.
Figure 4 displays the same emission in more detail using IF spectrum monitoring. In Figure 4, a max hold (blue) trace is used to display the modulation envelope of the signal and an additional clear write (yellow) trace is used the instantaneous value and frequency of the signal. One advantage of IF spectrum monitoring is that the measurements are made with the local oscillator (LO) fixed at a single frequency. This enables the ability to actively update the weighted detectors while maintaining frequency display update rates. Engineers are able to observe weighted detector measurements prior to the final measurement, all while observing the emissions spectrum. Markers can also be used to characterize the modulation envelope amplitude and frequency values.
Analyzing emission suspect signals collected during prescan prior to final measurement is important to ensure final measurement accuracy and properly characterize the DUT. Powerful tools such as spectrum analyzers and IF spectrum monitoring are available for prescan analysis. These tools increase measurement accuracy while helping to reduce compliance test time by making it easy to view and characterize suspect signals.
Mark Terrien is the EMC Business Manager for Keysight Technologies and is responsible for compliance and precompliance test solutions. He holds an MSEE in Electromagnetic Wave Theory from the University of Wisconsin, Madison and an MBA from Golden Gate University in San Francisco. Terrien can be reached at mark.terrien@ keysight.com.