Robust Extraction of Cardiff Model Parameters from Appropriately Tailored Measured Load-Pull Data

In the design of power amplifiers the ideal methodology would be a CAD based design cycle using a physics-based state-function (I-V, Q-V) non-linear model. In practice state-function models are often more measurement-based than physics-based and by design very general, which does limit their accuracy. Behavioral models extracted directly from measured data provide a complementary methodology, that should address the accuracy limitations of state function models. However, their success in achieving this is highly dependent on the user's behavioral model complexity selection and their measurement dataset choice. In this paper, it will be shown that this issue for the Cardiff Model can be resolved via the design of appropriately tailored measurement datasets.