2021 Virtual Undergraduate Research Symposium
2021 Virtual Undergraduate Research Symposium
Improving Microwave Device Efficiency Through the Quantitative Visualization of Electromagnetic Fields: Automating Load-Pull Simulations
Improving Microwave Device Efficiency Through the Quantitative Visualization of Electromagnetic Fields: Automating Load-Pull Simulations
PROJECT NUMBER: 47 | AUTHOR: Kristen Ung, Electrical Engineering
MENTOR: Peter Aaen, Electrical Engineering
ABSTRACT
The purpose of this project is to develop new high-throughput simulation capabilities to allow us to understand the design trade-offs that are made within very high-power GaN power transistors. In these transistors, the output power, efficiency, and linearity are impacted by the layout of the transistor and to explore a wide range of layouts we need to have a flexible and automated method to simulate these devices on a high-performance computing platform. In this project, a simulation procedure was developed in MATLAB to automate load-pull simulations and to automatically process the resulting simulation data. The script runs the simulations and processes the data to plot surfaces to convey the efficiency, gain, and gain compression of these transistors. The automation allows for the testing of different configurations of a single transistor to be run without supervision. The developed script uses Keysight Advanced Design System (ADS) to perform the simulation. With this data, insights into the layout of the transistors and the influence of electrical performance are examined. This will allow designers to make more informed decision on how to develop these transistors.
PRESENTATION
AUTHOR BIOGRAPHY
Kristen Ung is a Junior in Electrical Engineering with an emphasis in Integrated Circuits and Electronics and a minor in Computer Science. Kristen has been working with the Electrical Engineering Department to automate high-throughput load-pull simulations of power transistors to build a library of data for design reference and experimental research. She hopes to continue this research to add more analysis in the data processing of these simulations, as well as making the simulations more efficient to minimize the time required to run the simulations.
Great presentation! I have one question:
– In your “Swept Cells” experiment were the different configurations with respect to “Cells” and “Gate Width” run in-parallel or in serial?
Thank you! In the current experiment, all of the tests were run in serial. The next step of this project would not only be to filter out any outliers, but also to make the script compatible with a software that allows for parallel testing.
Thank you! The current experiment ran all of the tests in serial. The next step of the project would not only be to filter out any outliers, but also to make the script compatible with a software that allows for parallel testing.