Analog Adaptive Filtering Techniques for High-Speed Data Communications

An area receiving significant attention both from academia and industry is data communications. Great effort is being placed at pushing data transmission rates over copper to near the Shannon limit which requires support electronics such as filters operating in the VHF range. Thus, tuning mechanisms are essential to accommodate fabrication and channel variations. One technology to accommodate these variations is that of analog adaptive filtering and is the main topic of this thesis. Specifically, this thesis is an investigation to determine and demonstrate the capability of analog adaptive filtering in practical high-speed applications. The issues that are addressed are the implementation of the tuning algorithm, simple and effective building blocks and the effects of DC offset. Herein, a curve fitting approach is proposed for the adaptation algorithm and focus is placed on simple algorithms and simple filter building blocks that can accommodate a wide tuning range and high speed. Experimental results of a 100Mb/s transmit pulse-shaping filter verify the proposed adaptive techniques and highlight practical problems. This prototype is the first demonstration of an analog adaptive filter operating in the VHF range. Attaining a maximum processing frequency of 230MHz, it is also one of the fastest integrated filters reported in the technical literature. DC offset, being a major hindrance to optimal performance, is considered and its effect on four different possible variations of the basic