Design and characterization of downconversion mixers and the on-chip calibration techniques for monolithic direct conversion radio receivers

Abstract

This thesis consists of eight publications and an overview of the research topic, which is also a summary of the work. The research described in this thesis is focused on the design of downconversion mixers and direct conversion radio receivers for UTRA/FDD WCDMA and GSM standards. The main interest of the work is in the 1-3 GHz frequency range and in the Silicon and Silicon-Germanium BiCMOS technologies. The RF front-end, and especially the mixer, limits the performance of direct conversion architecture. The most stringent problems are involved in the second-order distortion in mixers to which special attention has been given. The work introduces calibration techniques to overcome these problems. Some design considerations for front-end radio receivers are also given through a mixer-centric approach.

The work summarizes the design of several downconversion mixers. Three of the implemented mixers are integrated as the downconversion stages of larger direct conversion receiver chips. One is realized together with the LNA as an RF front-end. Also, some stand-alone structures have been characterized. Two of the mixers that are integrated together with whole analog receivers include calibration structures to improve the second-order intermodulation rejection. A theoretical mismatch analysis of the second-order distortion in the mixers is also presented in this thesis. It gives a comprehensive illustration of the second-order distortion in mixers. It also gives the relationships between the dc-offsets and high IIP2. In addition, circuit and layout techniques to improve the LO-to-RF isolation are discussed.

The presented work provides insight into how the mixer immunity against the second-order distortion can be improved. The implemented calibration structures show promising performance. On the basis of these results, several methods of detecting the distortion on-chip and the possibilities of integrating the automatic on-chip calibration procedures to produce a repeatable and well-predictable receiver IIP2 are presented.