An lpoptimisation procedure is suggested for the design of activeURCfilters, approximating specified magnitude and linear-phase characteristics. The performance measure is a function of the magnitude error, phase error and the magnitude sensitivity with respect to the coefficients of the transfer function. The proposed technique has the advantage of controlling the sensitivity of a realisable transfer function without limiting the design to a specific circuit configuration. It does not require the assumption that a lumped solution to the problem is available as is the case in most previous design procedures. The technique is illustrated by a design example.