Faculty:
Bruce Wooley and
Bruce Lusignan
Students: David Shen and Chien-Meen Hwang
The growing emphasis on "personal communications" throughout the telecommunications industry has stimulated the demand for substantial reductions in the cost, size and power dissipation of radio components. Unfortunately, the present generation of radio front-end electronics has not benefited in a significant way from modern VLSI technology. Instead progress has come largely from the advances in the technologies used to package discrete components. The objective of this research is to develop a radio receiver architecture that eliminates the need for external tank circuits and intermediate frequency (IF) filters, thereby allowing most of the receiver to be integrated in a standard silicon process on a single chip. It is expected that when implemented in an aggressive VLSI technology the proposed radio will operate in the 900 MHz portable cellular radio band.
The radio architecture is based on discrete-time signal processing. Integrated analog discrete-time filters are used to avoid the need for mixing and IF filtering. Instead, the input signal is bandpass sampled at a rate that is lower than the highest frequency components the input signal. The Nyquist criterion is satisfied as long as the sampling rate is more than twice the information bandwidth of the signal.
Following sampling, the signal is filtered to select the desired baseband in the presence of large interfering signals. This is accomplished by means of successive bandpass downsampling stages that are implemented using a series of analog discrete-time filters. The filters prevent aliasing due to downsampling and provide a narrow bandpass function to select and amplify the desired signal.
Once the desired signal has been selected, it must be demodulated. Several techniques have been developed for demodulating discrete-time AM and FM signals. Discrete-time AM demodulation can be performed by means of an approach similar to envelope detection, as long as the aliasing of spectral images is well accounted for. A technique for FM demodulation relies on generating in-phase and quadrature components of the desired signal, and then tracking the phase. By sensing and correcting false zero crossings, the thresholding effect in FM demodulation can be substantially reduced.
Integrated Circuits Lab