Experimental evaluation of large-signal modeling assumptions based on vector analysis of bias-dependent S-parameter data from MESFETs and HEMTs

The authors presents a systematic experimental examination of the validity of basic large-signal modeling assumptions by subjecting measured S-parameter data versus bias from MESFETs and HEMTs (high electron mobility transistors) to various mathematical operations of vector analysis. Several approaches are used to determine the degree to which pairs of device nonlinearities can be accurately modeled by charge-based nonlinear capacitors, voltage-controlled current sources, and higher-order elements arranged in a standard equivalent circuit topology. Implications are discussed for such circuit modeling concepts as terminal charge conservation and its extension to other state-functions.<>