Prediction of Harmonic Distortion in ADCs using dynamic Integral Non-Linearity model

This paper investigates the application of dynamic, integral non-linearity (INL) model of analog-to-digital converters (ADC) to predict the harmonic distortion in pipelined ADCs. The device used to validate the model is Texas Instruments´ ADS6129; a 12-bit, 250 MSPS¹ pipelined ADC. The performance of the model is measured using single tones in the first Nyquist zone, with an amplitude ranging between -1 dBFS² and -46 dBFS. Total harmonic distortion (THD) and spurious free dynamic range (SFDR) are used as metrics to compare the performance of the model to the actual ADC. The results indicate that a model for high code frequency (HCF) INL is essential for correct prediction of distortion for input signals with small amplitudes. The results also suggest that dynamic INL models offer an effective method to model the distortion in ADCs until Nyquist frequency only. Beyond Nyquist frequency, the non-linear effects of track and hold (T/H) start manifesting and the INL model can no longer account for all the distortion seen at the output of ADCs.