| The electromagnetic spectrum from dc to 2--3 GHz has been recently saturated with wireless communications. The largest market share of this technology is primarily for voice and data through cellular networks. Soon, larger bandwidths will be needed to accommodate the transfer of multimedia such as live video and the Internet to roaming users. The carrier frequencies on which this information is modulated will push into the tens of gigahertz, generating a high-volume global market for monolithic microwave integrated circuits (MMIC). Millimeter-wave technology has been used for a few decades, mostly for military applications. The components were often bulky and costly, but only needed in small numbers. Once the demand for millimeter-wave systems has reached the consumer level, simplicity and cost will drive the technology. The advent of the MMICs has made this possible; however, it remains to be seen which technology will dominate the evolution of millimeter-wave systems. The voltage-controlled oscillator (VCO) is the critical component in most of these systems due to carrier phase noise requirements. The goal of this work is to identify challenges in the design of fully monolithic oscillators and link performance limitations to technology parameters.;The choice of foundry process is most often the critical factor when attempting to achieve a given performance at a given cost. It must be recognized that once the minimum performance specifications are met, the solution with the lowest cost will be selected. In current MMIC oscillator research, there is little prediction of the phase noise in the design process. Therefore, the results of past studies have been largely empirical. An empirical study only sets a performance benchmark for a particular device and circuit design. Empirical results in this dissertation show that the phase noise adequate for many proposed systems can be achieved with a low-cost technology such as GaAs metal-semiconductor field effect transistor (MESFET). The results highlight the inextricable linkage of phase noise to both device parameters and circuit design methodology. The prediction of phase noise is accomplished through modeling of low-frequency noise of devices and the use of commercial simulation tools. |
