This Paper details two different contributions related to practical CM/DM EMI measurements. A first part investigates sources and implications of measurement errors that result for CM/DM separators in a practical measurement environment with a particular focus on the recently presented input impedance criterion for CM/DM separators. Furthermore, the realization of an active CM/DM separator, which features competitive separation capabilities (DMTR/CMRR > 51 dB and CMTR/DMRR > 47 dB for frequencies up to 10 MHz), is presented.

This paper discusses the Δ-networks and other circuits designed to separate the conducted electromagnetic interference (EMI) into its common mode (CM) and differential mode (DM) components. The input impedances of CM/DM separators must be 50 Ω resistive in the measurement frequency range, and they must be independent of the values of the noise signals and noise source impedances. The conditions for achieving such input impedances are derived. It is shown that many of the proposed separators, including the Δ-network suggested in the International Special Committee on Radio Interference (CISPR) 16-1-2 standard, do not fulfill the input impedance requirement. This leads to unreliable CM and DM measurements and, consequently, to the oversizing of EMI filters and design by trial and error.Index Terms-Common mode (CM), delta network, differential mode (DM), electromagnetic compatibility (EMC), electromagnetic interference (EMI), International Electrotechnical Commission (IEC), measurement, noise separator, power filters.

No abstract

No abstract

This work investigates sources of measurement errors that result for CM/DM separators in a practical measurement environment, with a particular focus on the recently presented input impedance criterion for CM/DM separators, derives the respective analytical expressions, and employs a detailed analytical model to verify the obtained findings. Furthermore, a method is derived, which determines the worstcase measurement error by reason of cross coupling for given measured DM and CM output voltage components. Based on an example, this work illustrates how the obtained expressions can be advantageously used in a computer program to automatically decide whether a particular spectral measurement component represents a useful measurement result or if it is strongly affected by cross coupling (CM to DM and DM to CM). Finally, the paper presents the realization and accompanying experimental results of an active CM/DM separator, which allows for low realization effort and features competitive separation capabilities (DMTR/CMRR > 50 dB and CMTR/DMRR > 42 dB for frequencies up to 10 MHz). Keywords-Analog processing circuits, electromagnetic compatibility, electromagnetic interference, frequency domain analysis, measurement errors, power electronics, voltage measurement. NOMENCLATUREV rms value of a considered voltage, V = V e jϕ phasor corresponding to a sinusoidal voltage, v(t) = √ 2 V cos(ωt + ϕ), at given angular frequency, ω (alternative time-domain signals are considered by means of Fourier analysis), Z = R + jX complex impedance, V sep,l , V sep,n phasors of the input voltages of the CM/DM separator, V sep,DM , V sep,CM phasors of the DM and CM components corresponding to the separator's input voltages, V out,DM , V out,CM phasors of the output voltages of the separator, DMTR = | V out,DM V sep,DM | Differential Mode Transmission Ratio, CMTR = | V out,CM V sep,CM | Common Mode Transmission Ratio, CMRR = | V out,DM V sep,CM | Common Mode Rejection Ratio, DMRR = | V out,CM V sep,DM | Differential Mode Rejection Ratio, e DM , e CM DM and CM gain errors due to CM/DM separator's input impedances, F. Krismer and J. W. Kolar are with the Power