Parent Category: 2014 HFE
By NI AWR
Overview
This application note illustrates the use of the Analyst 3D finite element method (FEM) electromagnetic (EM) simulator, fully integrated within the NI AWR Design Environment platform, for optimizing (matching of) a coaxial connector microstrip transition. Analyst simplifies design by keeping all design data within one project file, including connectors, bonded packages, housing, or any other arbitrary 3D objects. Various transition models from different connectors to different board substrates can likewise be stored as a library and conveniently used in any subsequent circuit design.
The Design
In the schematic (parent document), a printed circuit board has been drawn with the signal trace. The 3D connector that is designed for 20 mil substrates is, technically speaking, a sub model (or child) of the parent document, making the design hierarchical. The connector model allows for the placement of a port on the coaxial cable end of the connector (input port) and for the other port to be defined as a regular wave port at the end of the microstrip (output port). The reference plane of the output port is moved just after the connector.
It is important to note that when comparing measurement results to simulated ones, the transition from the measurement hardware’s coaxial cable to the signal trace of the device under test (DUT), as shown in Figure 1, is commonly assumed to be ideal. In reality, however, this ideal condition can result in data mismatch (simulation data =/ measurements) at higher frequencies.
Figure 1 • The transition from the measurement hardware’s coaxial cable to the signal trace of the DUT.
This ideal assumption is implicitly made as soon as the simulation input and output ports are placed on the trace. In this case it is the microstrip or coplanar waveguide, as shown in Figure 2. Here the simulation port is directly placed on the microstrip, thereby neglecting the discontinuity represented by the connector in Figure 1 and thus systematically distorting the simulated versus measured results.
Figure 2 • The “ideal” assumption is implicitly made as soon as the simulation input/output ports are placed on the trace, in this case the microstrip or coplanar waveguide.
While there are multiple methods that can be used to ensure that the transition has as little reflection as possible, using Analyst for the method described in this application note enables designers to treat the transition parasitics in an analytic fashion, moving away from the ideal assumption towards a more realistic approach.
In Figure 3, the S11 of the transition reveals that the inherent matching is satisfactory but only up to about 2 GHz. Given that the target design frequency is 10 GHz, the reflection is unacceptable (as high as -10 dB in this case) and thus optimization of the transition is warranted, not only to address the loss of energy, but also to account for mismatch.
Figure 3 • The S11 in of the transition in this coax-to-microstrip graph reveals that the inherent matching is good only up to about 2 GHz.
Optimization Strategy
The transition model can be optimized in a schematic within NI AWR Design Environment by using ‘EM document’ as a regular sub model. By inserting a series-L parallel-C matching circuit, the transition can be readily optimized at 10 GHz. In microstrip, a series-L can be realized by a narrow segment of strip, while a parallel-C can be realized by a wide strip segment. Within NI AWR Design Environment software, it is straightforward to optimize the required strip dimensions, as shown in Figures 4 and 5.
Figure 4 • It is a straightforward process to optimize the required strip dimensions in this transition model schematic using the EM document as a regular sub model.
Figure 5 • Reflection coefficient of the transition at coax port using the closed-form models of the matching circuit (green curve).
The next step for optimization is to put the matching circuit dimensions into the 3D connector model and run another verification simulation (shown in Figure 6).
Figure 6 • Reflection coefficient of the transition at coax port including a full 3D model of the matching circuit geometry (blue curve).
As revealed in Figure 6, this first matching effort results in excellent performance. It is also insightful to view and animate the surface currents at 10 GHz (see Figure 7).
Figure 7 • Surface current annotation of the optimized transition at 10 GHz.
Conclusion
To summarize, this simple matching circuit/optimization exercise provides signal transmission from the coax cable into the microstrip with less than -20 dB reflection at 10 GHz (the target design frequency). The matching bandwidth is now about 2 GHz and is easily and reliably optimized with the resulting EM modeled connector using conventional circuit models. Finally, a full 3D EM verified solution is readily available with the optimized geometry as well.
The use of Analyst for the EM simulation provided a more realistic, reliable approach to achieving design success in this application. Analyst 3D FEM technology within NI AWR Design Environment enables designers to move smoothly from circuit concept to full 3D EM verification with a single mouse click.