Optimal Design of Passive Devices for Verifying On-Wafer Noise Parameter Measurement Systems

Abstract

We propose the optimal design of passive devices that can be used to verify on-wafer noise parameter measurement systems. The design principles result from obtaining the minimum relative uncertainties of four noise parameters: $F_{\mathrm {min}}$ , $R_{\mathbf {n}}$ , $\vert \Gamma _{\mathrm {opt}} \vert $ , and $\angle \Gamma _{\mathrm {opt}}$ for a wide range of S-parameters of a passive two-port network. A Monte Carlo method has been used for the investigation, and simulation results show that $\vert S_{11} \vert $ plays a primary role in deciding the optimal design and must be within 0.5– $0.6.~\vert S_{21} \vert $ plays a secondary role in the design and, ideally, it should be as small as possible. Based on these findings, we designed and fabricated three planar attenuators on a semi-insulating GaAs substrate. The test results (at up to 40 GHz) show excellent agreement with the simulation. This is the first time that the effect of different designs of passive verification devices on the system noise measurement has been analyzed, and the design principles of optimal passive devices are given.