Active RC Synthesis for Significant Reduction in Magnitudes of Reactive Elements

Abstract

Low impedance levels and/or large system time constants lead to network realizations with large total reactance values in ordinary passive and active RC synthesis. It is shown how active RC networks may, instead, be designed, in which one may freely a priori specify the total capacitance and the input and output resistances. The basic design philosophy is to assign relatively faroff poles to the passive part of the system. Feedback is used to obtain the desired close-in (large time constant) system poles. In one method, employing a single negative-impedance-converter, any number of system poles may be simultaneously realized. However, the output is very sensitive to variations in active and passive parameter values. Also, the negative-impedance-converter must act as such over the bandwidth of the passive part of the network, which part is much larger than that of the system as a whole. A second approach is far more practical. It employs a cascade of isolated canonic second-order structures, each of which realizes a pair of system poles. The resulting design is much less sensitive to active and passive parameter variations, and is much easier to tune. The structure used is especially suitable for narrow-band filtering. Also, the active element bandwidth need only be of the same order of magnitude as that of the system poles. Design relations are developed and optimized with respect to a figure of merit ASA, being the product of amplifier gain required and system sansitivity to amplifier gain variations.