A Universal Behavioral Model for COTS ADCs with Predictive Cold Temperature Performance Capability

Abstract

To further enable exploration of Cold Worlds, this paper seeks to develop a temperature-predictive universal model for CMOS Nyquist-rate commercial-off-the-shelf (COTS) analog to digital converters (ADCs). The proposed modeling methodology is designed to enable rapid model construction from readily available data sheet information without incurring extensive and costly overhead associated with data collection in order to produce a behavioral model with predictive cold temperature performance estimation. Thorough temperature characterization of two different popular architecture COTS CMOS ADCs from different manufacturers is used to add predictive temperature performance capability to this universal behavioral model. Parametric trends in the ADC transfer curve, which have been measured and established across temperature, are applied to the initial parameters of an ADC's datasheet transfer curve, allowing the initial transfer curve to evolve across temperature in an analogous fashion to what has been experimentally observed. From this simulated transfer curve, the expected output and performance of the device may be computed. The paper concludes with a comparison of part to part variations for several devices along with simulated results. It may be possible to predict a rough estimate of a part's performance beyond datasheet temperature ranges.