Undersampling and SNR Degradation in Practical Direct RF Sampling Systems

Abstract

Advances in semiconductor technology are leading to ever-higher sampling rates for analog-to-digital converters (ADCs). In traditional systems, (cellular) radio signals had to be downconverted in the analog domain from radio frequency (RF) to baseband (BB) using direct conversion receivers (DCRs) in which the ADCs could operate at low sampling rates. Due to advances in ADC technology, modern direct RF (DRF) sam-pling ADCs are able to sample directly in the RF domain, opening up new application possibilities. In order to reduce the required sampling rate, it is particularly interesting to use undersampling techniques in these DRF systems. However, there are some SNR-decreasing problems associated with this approach, such as the noise folding effect, which need to be carefully considered. This paper provides a detailed investigation of practical DRF undersampling systems in theory and practice. It is shown analytically and with experimental verification that noise enhancement due to noise folding is a negligible effect in real-world DRF implementations. Furthermore, the results of this paper motivate reduced complexity and reduced power consumption DRF undersampling ADC architectures with high analog bandwidth but relatively low sampling rate.