RLC circuit extraction with the differential evolution algorithm for conducted electromagnetic emission model of integrated circuits

Abstract

This paper examines the modeling of conducted electromagnetic emissions of integrated circuits. In this study, test circuits were designed and printed circuit boards were prepared to measure the input impedance at the power supply pin of the microcontroller. After S-parameter measurements of the test circuit were performed in a frequency range of 10 MHz to 2 GHz, Z-parameters were obtained via S-parameters. Next, the impedance--frequency curve of the test circuit's power supply pin, which consists of a microcontroller core, printed circuit board, conductive trace, and SMA connector global impedances, was obtained. Therefore, all the traces, printed circuit board, and SMA connector impedance effects were eliminated with the deembedding technique. After all other impedance effects were deleted with the deembedding technique, the input impedance at the power supply pin of the microcontroller was obtained and shown with impedance--frequency characteristics. The passive (RLC) circuit was modeled from this curve and the results were compared to the measurements. The differential evolution algorithm was used to extract the optimal RLC passive elements. Finally, sensitivity analyses were successfully performed to ensure the accuracy of the extracted circuit.