Model driven design of high performance microwave filters

Microwave filters are steadily evolving towards the use of more complex transmission line structures as building blocks. This is needed to cope with the highly demanding specifications of modern telecommunication equipment.

The design procedures that are used to realize these filters however have not evolved at the same pace. They still rely on the use of lossless, analytical models that are borrowed from standard line structures to approximate more complex, lossy structures. The lack of accuracy that results from this crude approximation is then filled using massive numerical optimization based on EM simulations. This process is both time-consuming and blocks the intuitive insight in the operation of the device. The loss of physical insight makes it harder for the designer to make proper design choices and can lead to designs that are too sensitive.

To circumvent these disadvantages, we propose to introduce model driven design. This is done by using 2 different approaches:

Introducing an accurate equivalent circuit model into the design procedure

An accurate equivalent circuit model is extracted for an elementary section of the considered microstripstructure. The discrete parameters of the model are related to the geometrical parameters of the structure. We will maintain the analytical models from the classical design procedure and extend them to make the procedure sufficiently accurate, thereby eliminating the need for optimisation.

Introducing metamodeling into the design procedure

The geometrical parameters of the complete microstripstructure are linked to the filter characteristics or the filter specification parameters through metamodels. The metamodels are then used in the optimization process instead of the computationally expensive EM simulations, resulting in an accurate and efficient optimization directly on the structure of interest. Using a metamodeling-based design a designer needs more than 5000 times less CPU time compared to an EM-optimization-based design.

After designing a modulated RF device, the nonlinear behavior of the system needs to be characterized. Several widely used Figures of Merit (FOM) exist for this purpose. To improve the time and effort necessary to extract the traditional FOMs, we are working at an alternative that is based on the Best Linear Approximation.

Reference

  1. Yves Rolain, Maral Zyari, Evi Van Nechel, Gerd Vandersteen. A measurement-based error-vector-magnitude model to assess non linearity at the system level. IEEE MTT-S International Microwave Symposium (IMS), Honolulu, Hawaii, 4-9 June 2017, pp. 1429-1432
  2. Van Nechel, E., Ferranti, F., Lataire, J., & Rolain, Y. (2017). Efficient Design Optimization and Variability Analysis of Defected Ground Structure Filters Using Metamodels. In IEEE MTT-S International Microwave and RF Conference (IMaRC) .
  3. Van Nechel, E., Rolain, Y., & Lataire, J. (2018). Extracting Improved Figures of Merit for Characterizing Nonlinear Devices using Multisine Excitation Signals. In 91st ARFTG Microwave Measurement Conference (ARFTG), IEEE.
  4. Van Nechel, E., Ferranti, F., Rolain, Y., & Lataire, J. (2018). Model-Driven Design of Microwave Filters Based on Scalable Circuit Models. IEEE Transactions on Microwave Theory and Techniques, 66(10), 4390-4396.
Back to top