Device Modeling using Neural Network Techniques for Solid State Power Amplifier Applications

In this paper, the integration of an advanced device model for solid state power amplifiers is proposed. The model accounts for both the static and dynamic response of HEMT devices very accurately. This large signal model utilizes a feed-forward neural network using the back-propagation method with the Levenberg-Marquardt (LM) algorithm. The model presented was used on a 3MI 0.15 BC;m power pHEMT process developed by Triquint. Excellent agreement is observed between the advance model and measured DC, AC and load pull data.