Identification of a power-frequency meta-model for a power amplifier


Currently power amplifiers are modelled either in a component based approach or by linear models, that are parameterized in the operating point of the device. Both these approaches have their drawbacks, however. The component based model requires one to expose the intellectual property of the component, while the model collection leads often to error prone results.  This calls for compact black box models, that combine the shielding of the internal structure with the ease of use of the model. This mixed model can become a stepping stone between a less flexible linear and a complete but more complicated nonlinear model.
In the past, there have been attempts to model a power amplifier by such a compact model. However, the extraction proved to be complex while the model error was not always very well controlled.


In this work, we would like to combine the ideas of Best Linear approximation and the local interpolation to obtain a model that is can be scaled with respect to the power and impedance levels of its operating environment. As this will require the extraction of a model that depends on more than one independent variable, the number of measurements that are needed to extract the model can potentially become too high to be practically usable.  In this work, we will also try to make a contribution to this challenging problem.


Simulations are conducted using matlab and/or ADS. Model extraction and experiment design will be handled in Matlab. The final goal is to verify the model on a real measurement setup, that uses that PNA-X based network analyzer.
Interested? Want to know more? Please feel free to contact the promoter. By asking you dont commit to anything else than getting more information.


Yves Rolain
Back to top