ADC low-complexity modelling identification and mitigation of baseband intermodulation distortion with instantaneous frequency dependence

Abstract

The linearization of active electronic components such as power amplifier or analog-to-digital-converter (ADC), is a vast subject. Many issues come into play, including behavioural modelling with the selection of a relevant model in terms of accuracy, complexity, identification, and compensation. In this paper we consider an ADC linearization problem, we propose an original low-complexity modelling identification and mitigation of baseband intermodulation distortion with instantaneous frequency dependence. Thus, we introduce a refinement of a power series model by adding a dependence on the variation of the instantaneous frequency of the signal. Based on this, we describe a method suitable for the model calibration using a two-tone signal. We finally present a measurement bench adapted to the calibration of the coefficients of the model, followed by some results of linearization. All the methods presented are intended to be operated in real time in embedded systems and are therefore studied to present very low computational complexities.