Click here to learn about Henrik Forstén's on-line 2D EM field solver tool
Click here to go to our main page on coupled line couplers
Click here to see the same GaAs microstrip coupler in this page analyzed with Microwave Office
Click here to see PolyStrata 10dB coupler design using Henrik Forstén's on-line EM tool (much better result!)
New for February 2026. There's nothing like being unemployed for making you look around for free EDA tools!
Here we are using Henrik Forstén's on-line 2D EM tool to follow up on a microstrip coupler design we first posted on this page using Microwave Office's linear model, and we will compare the results. Linear models are a great way to start a design but before you fabricate anything you should take a look at the artwork with an electromagnetic solver. There are 3D simulators (HFSS, CST) and 2.5D simulators (Momentum), but how about a 2D solver? We took a look at a 10 dB coupler on GaAs using Henrik Forstén's totally free on-line 2D EM field solver tool.
2D EM simulations almost perfectly valid for TEM transmission lines of uniform cross-section, but microstrip is not a pure-TEM media. Henrik's solver produces dispersion-free results for microstrip, which we know is not the case. Is it as good as a linear solver, or better? We don't yet know the answer, but sometimes a free tool is all you have at your disposal.
One of the features of the Henrik's EM tool is that you can store all of your parameters (geometry, material properties, frequency points, etc.) by copying a link to your work. Here is a link with all the coupler's parameters loaded, just click solve and we will all be on the same page!
Click here to see the 10 dB coupler project
Here is an image of the coupler's geometry that we downloaded from the project.
Even and odd mode impedances for 10 dB coupler in 50 ohm system (see this page for equations) should be 69.4 ohms, and 36.0 ohms respectively. The resulting structure has Zeven=67.7ohms and Zodd=35.2 at 10 GHz; it's pretty close.
And finallly, here are the S-parameter magnitude results. Note that the tool does NOT allow you to assign port numbers.
Now let's compare the previous Microwave Office coupler analysis with this EM analysis. Port 1 is in input, port 2 is isolated, port 3 is direct and port 4 is coupled. Why did we pick these part numbers? We can't recall but it must have been important!
Here is the Microwave office result we generated long ago.
We no longer have access to MWO, otherwise we would have plotted comparison data on one plot for each parameter. Below we downloaded the EM result data and re-plotted it on the same scale and with the same port numbering as the MWO project so you can compare the magnitudes of the S-parameters between the two analyses. The coupling (S41) is very similar, and so is the isolation (S21). However, there is a distinct difference in return loss (S11). Have you ever tried plotting an S4P file in Excel? It was no fun but we came up with a solution, if anyone is interested you could try it.
Which analysis is more accurate? We are going to place a bet on the linear analysis (Microwave Office) as it takes into account frequency dispersion, and about a million man-hours went into developing the model.
If anyone has time on their hands, please analyze the circuit in a 2.5D or 3D solver and send us the results, which should be the "most real" of any prediction.