Imperfect integration and DC offset effects in the convergence performance of gradient search based analog adaptive filters

The advance of the analog VLSI technology has produced a growing interest in analog gradient-search-based adaptive filtering, rendering it a field of active research during the last several years. This is because analog adaptive filters have potentially, faster convergence rates, a much smaller size and lower power consumption than their digital counterparts. However, much effort is required yet to solve some problems inherent to analog implementation of adaptive filters, not present in digital adaptive filters, such as imperfect integration circuits and DC offset, which greatly degrade the system performance. This paper presents theoretical and experimental results which show that DC offset and the pole position of the integrator circuits, are two of the most important distortion sources in gradient-search-based analog adaptive systems.