Toward Wide-Band High-Resolution Analog-to-Digital Converters Using Hybrid Filter Bank Architecture

Abstract

The paper presents the problem of design and simulation of a high-speed wide-band high-resolution analog-to-digital (ADC) converter working in a bandpass scenario. Such converters play a crucial role in software-defined radio and in cognitive radio technology. One way to circumvent the limits of today’s ADC technologies is to split the analog input signal into multiple components and then sample them with ADCs in parallel. The two main split approaches, time interleaved and frequency splitting, can be modeled using a filter bank paradigm, where each of these two architectures requires a specific analysis for its design. In this research, the frequency splitting approach was implemented with the use of a hybrid filter bank ADC, which requires an output digital filter bank perfectly matched to the input analog filter bank. To achieve this end, an analog transfer function, together with an assumption of strictly band-limited input signal, has been used to design the digital filter bank so far. In contrast, the author proposes dropping the band-limit assumption and shows that the out-of-band input signal has to be taken into account when designing a hybrid filter bank.