Efficient computation and minimisation of root sensitivities of analogue and digital networks

An efficient computational method is described, to find roots and their sensitivities with respect to parameters of analogue and digital circuits. It is indicated how the root sensitivities can be minimised, while ensuring that the roots remain at specified locations. The computations are organised so that the roots of intermediate designs need not be found, and the roots of the final successful design are the desired roots, having minimum sensitivities. It is shown through illustrative examples that sensitivity minimisation can lead to considerable improvements, even in the case of well designed networks.