ON THE SURFACE IMPEDANCE USED TO MODEL THE CONDUCTOR LOSSES OF MICROSTRIP STRUCTURES

In the numerical analysis of microstrip structures their finite conduc tivity is often modelled by using an equivalent surface impedance to r elate the nonzero tangential electric field on the conductor surface t o an equivalent electric surface current. The fields on top and below the patch are not necessarily the same and to account for this discont inuous tangential electric field a strict equivalence formulation shou ld include magnetic surface currents. However, it is common practice t o neglect them, presumably because they considerably complicate the su rface integral equation and its numerical solution. Here it is shown t hat this practice leads to the use of an incorrect, although conservat ive, surface impedance. A basic assumption is that delta much-less-tha n DELTA much-less-than lambda, with delta the skin depth, DELTA the co nductor thickness, and lambda the free space wavelength.