In systems wherein information is represented and processed digitally, the level of overall system performance that can be achieved is often limited by the speed and dynamic range of the interfaces between analog and digital representations of that information. Since digitally encoded signals can be processed with virtually arbitrary precision, continued progress in the scaling of integrated circuit technology is accomplished by corresponding demand to improve the precision and speed of so-called data conversion interfaces. This pressure is, in turn, accompanied by two constraining factors: the desire to embed data conversion interfaces within large digital signal processing circuits and explosive growth in the demand for portable electronic systems. The integration of data conversion circuits in VLSI technologies imposes severe constraints on the dynamic range available to implement those circuits, while the concern for battery life in portable systems often dictates substantial reductions in power dissipation in comparison to traditional applications. The principal objective of this research is to discover circuit design techniques for digitizing signals with bandwidths of several MHz using standard digital CMOS technologies. The research focuses on the use of oversampling methods to encode MHz-bandwidth radio signals centered at IF frequencies in the range of a few tens of MHz and on methods of increasing the resolution of Nyquist-rate converters for digitizing baseband signals with bandwidths on the order of 10 MHz. A complementary theme of the research is operation from a low supply voltage with minimum power dissipation.
