Identifying Component Parasitics Using Equivalent Circuit Modelling

Abstract

Within circuit simulations, only the ideal case is considered. This is insufficient is many ways as it is not possible for a physical component to act ideal outside of a very limited frequency range. In order to account for this equivalent circuit modelling can be applied to identify the parasitic elements introduced by a components imperfections and structure. There are two broad methods which can be applied to obtain an equivalent circuit model. The first is a heuristic approach, which is a brute force form of guess and check in which an approximated topology and values is applied. However, this method is not suitable for high complexity analysis. The second method explored uses partial fraction expansion to separate a transfer function of the input frequency response into a series of fractions. These fractions can then be compared to known circuit topology equations such that an equivalent circuit model can be produced. In doing this there remains a degree of experimentation for the input variables being the number of poles and zeros of the function, determining the order, as well considering both the impedance and admittance of the input frequency response. The use of the PFE method allows for a simplified analysis, compared to that of the heuristic approach. The PFE also produces a higher degree of accuracy when considering the magnitude of either the impedance or the admittance, with the extent of this dependent upon the accuracy used in determining a sufficient equivalent model when using the heuristic approach. In each case, the modelling method produces an equivalent circuit topology, which has an output more closely resembling that of a measured input than the ideal response. This can then be used for the purpose of circuit simulations, allowing for a higher degree of accuracy within the production of component and circuit behaviour.