Influences of hardware implementation on a high speed digital adaptive filter using the Residue Number System

A study of the performance and certain characteristics of an eight-weight adaptive FIR digital filter implementing the LMS algorithm using Residue Number System (SRNS) arithmetic hardware has been done in order to draw conclusions as to the optimum hardware configuration. Affects of the hardware and unknown system on the rate of convergence and adaptive algorithm step size were found by running filter simulations for different values of the systems quantities. It was found that characteristics of the unknown system (or plant) do not affect the choice of filter hardware.