Methods to Increase the Bandwidth of Broad-Band Time-Domain Electromagnetic Interference Measurement Systems

Standards have been defined to specify the limits of electromagnetic radiation power levels any electric or electronic component may emit or be exposed to in specific environments. These limits are defined to ensure that an electric or electronic device is not disturbing the operation of other devices in its vicinity or even poses a risk to health and wellbeing of humans or animals. On the other hand, exposed to radiation within these specifications, the device itself should operate unimpaired. EMI receivers are built to measure and evaluate the electromagnetic emissions from a device under test (DUT). Based on the test results obtained by the receiver a DUT can be considered compliant to the standards may be made available to the market. The primary objective in this work is to enhance the time-domain EMI measurement receiver beyond the available technology. In addition, new methods for measuring and evaluating EMI emissions are introduced to the time-domain EMI measurement system.

To increase the bandwidth of broad-band time-domain electromagnetic interference (TD-EMI) measurement systems a method with interleaved sampling of the measurement signals by multiple parallel mutually time-delayed samplers has been investigated. By this method the effective sampling rate and the system bandwidth could be increased by a factor equal to the number of channels. Such an increase in the baseband to a level beyond the available technology introduces high level spurious signals in the final spectrum due to the mismatch between different analog to digital converters. Automatic calibration algorithms for minimization of the mismatch are thoroughly explained.

To increase the dynamic range of the systems multifrequency sampling has been introduced. This approach introduces a multifrequency sampling to digitize the input signal in two branches simultaneously. Such an approach will generate two spectra, where the signal contributions representing the measured EMI signal are the same, and the spurious signal contributions are different in both channels. The comparison of the two spectra requires the introduction of a multirate filter in the second channel that changes the rate of the data stream to the sampling rate of the first channel. The description of the implementation of the possible multirate structures at such high sampling rates is discussed.

The traditional EMI test procedures test the effect of the emissions of any DUT at predefined frequencies. A measurement of the envelope amplitude, using the well-known detectors like max-peak, average, RMS, and others, was sufficient to evaluate the effect of emissions of the DUT. Based only on the measured amplitude levels, a DUT was considered passing or failing the EMI test. New communication systems use more sophisticated transmission methods with comparison with the traditional systems. The new methods use, for a single system, several frequency bands. In addition, coding algorithms are introduced in the transmission of information. Such new systems show robust behavior against some interfering signals. Not to disturb the transmission of information, an interfering signal should have special properties regarding its rate of occurrence and amplitude level. An interfering signal could then have relatively high amplitude with a low occurrence rate while not disturbing the frequency band under test. So measuring the rate of occurrence of an electric, magnetic, or electromagnetic emission of a DUT would relax the requirement of an EMI test.
For this purpose, an interesting approach to measure the amplitude and rate of an EMI stochastic signal is to use the Amplitude Probability Density function (APD). The parallel implementation of the APD measurement function on thousands of frequencies is explained. The relation between APD and the traditional detectors is presented. In addition, a new definition of the limitlines is explained, where the DUT test would not depend only, in the future, on the amplitude limit but also on its probability distribution.