Enhanced Millimeter-wave Filter Design under Stochastic Uncertainties

A fully adjustable and tolerant technique for the design of millimetre-wave filters in the presence of stochastic uncertainties, and in particular of fabrication imperfections, is presented in this paper. In particular, a novel frequency-independent algorithm with a versatile implementation, on the basis of Chebyshev polynomial analysis, is developed. Essentially, the featured method leads to a compressed polynomial, under the restriction that the maximum local peaks of the original distorted polynomial are below the unit. As a consequence, the equivalent filter design can efficiently and reliably manipulate any arbitrarily-generated imperfection, without the need of additional conventions. The proposed formulation is successfully verified by means of three real-world microstrip filter configurations, namely a 5th-order Chebyshev step-impedance lowpass filter at 2.5 GHz, a 5th-order Chebyshev hairpin-line bandpass filter at 10 GHz, and a 4th order Chebyshev open-loop bandpass filter at 50 GHz.