Reduction of sensitivities in sampled-data filters

Sampled-data filters can exhibit extreme sensitivity to coefficient inaccuracy. This correspondence introduces design techniques that do not restrict the delay interval in the fiter to being equal to the sampling period, thereby providing flexibility in the reduction of sensitivity. An example illustrates the differences between the performance of a filter designed by using these techniques and that of a filter designed by using the well-known method based on the bilinear z transformation.