With increasing requirements for 5G applications, the demand for RF filter design and production has never been greater. RF/microwave filter design engineering is an iterative and experienced-based process that can be very time consuming for even the most experienced filter design engineers. The product development process requires experience-based knowledge, continuous tweaking and repeated, iterative trials, followed by physically building samples to measure and verify the design’s RF performance. For inexperienced designers or more complicated design requirements, additional build cycles are required before reaching an optimized design, extending the R&D cycle time and increasing engineering development costs.
SynMatrix is a cloud-based, intelligent, passive component design platform that can accelerate the RF/microwave filter design process by introducing an accurate and efficient step-by-step method to achieve RF performance goals. The design platform combines proprietary advanced algorithms and machine learning technologies with decades of experience in RF/microwave component R&D. Users will dramatically improve the component design simulation and tuning process, armed with a comprehensive set of tools and automation features to help produce designs more accurately and more quickly. A complete line of design structures is available on the platform, including cavity filters, ceramic filters, waveguide, coaxial, LTCC and coplanar microstrip. The design platform was launched in 2019; the latest feature, an intelligent AI-tuning system, was added in February 2020.
The SynMatrix design platform addresses three main elements of the design process:
- Advanced coupling synthesis and comprehensive design analysis (see Figure 1).
- Advanced computer-aided tuning (see Figure 2).
- An AI-optimized tuning system that runs automatically.
The RF/microwave filter design process begins with the product requirements and specification analysis, which can be managed using SynMatrix’s advanced matrix synthesis tool. The comprehensive synthesis module includes arbitrary topology synthesis, dispersive effect control, matrix sensitivity analysis, thermal drift margin analysis, peak power handling analysis and matrix optimization. Once the “golden matrix” is established, designers can use advanced algorithms in the computer-aided tuning module, which includes two-port and multi-port tuning techniques and generic optimization, to incorporate dispersive effects and spurious prediction. These features help eliminate repetitive processes to improve development speed and the design accuracy of the end product.
SynMatrix recently released an advanced intelligent AI-tuning feature enabling Ansys HFSS users to access one of the world’s first AI-based optimization systems (see Figure 3). SynMatrix directly integrates with HFSS and provides the following benefits:
An advanced intelligent tuning and space mapping design platform in one project workflow - Designers can import Ansys HFSS design parameters directly and automatically into SynMatrix. Once the data is imported, HFSS users can use the user-friendly platform to auto tune and edit the parameters to meet the desired RF design output.
An intelligent AI-tuning tool - The tool uses a machine learning, nonlinear curve-fitting technique, accompanied by an anti-error mechanism with fast convergence. The original and aggressive space mapping functions are also available.
Dispersive simulator - SynMatrix is one of few tools to offer dispersive simulation in the design process. It handles complex structures such as ceramic and waveguide filters and multi-modes with high dispersive or close spurious effects (see Figure 4).
Cloud-based or desktop - Users have the flexibility and convenience of either secure, cloud-based or desktop deployment options.
Compared to other design platforms, few can match the technical array of features available on SynMatrix. Many of the advanced component design functions offered by SynMatrix - arbitrary topology synthesis, RF performance optimization, thermal shift analysis and Monte-Carlo analysis - are not included in many other synthesis tools in the market. SynMatrix’s intelligent AI-based optimizer uses reinforcement learning backed by years of practical human experience. This differentiator feature enables filter designers to “auto tune” a design with a click of the mouse, then multi-task while the design is being optimized.
The SynMatrix design platform goes beyond basic filter design: with the Ansys HFSS integration, the combination offers an integrated and comprehensive workflow that can improve efficiency and flexibility to design advanced microwave filters. The platform offers tools to tweak and tune the outputs automatically, greatly reducing the engineering development time to produce designs. With the increasing demand for 5G systems, the corresponding demand for complex RF/microwave filters in these frequency bands will increase substantially. With a design platform like SynMatrix, users have a formidable tool to meet these growing demands.
Richmond Hill, Ontario, Canada