CST: Waveguide port and Smith chart question

[andi_]

Hello, I'm a beginner, designing a diamond-shaped patch antenna fed with quarter-wave feedline for 50Ω matching exited by a waveguide port in CST MWS. I got Smith chart result as in the attachment with Z=95.69Ω for S1,1.

My problem is, I'm not sure this impedance stands for the impedance of the antenna or the impedance of the waveguide port. And since to design a ¼-wave transmission line, antenna impedance needs to be known, [Zin = sqrt(Z0*ZA)], how do I get it's value?

I tried with feeding the antenna with a 50Ω µstrip, and add a marker on the Smith chart for my freq., 5.7GHz, get abs(Z) and do the Zin calculation. Is this the correct way?

And how to choose waveport size? The result in the attachment is obtained with waveport width = width_of_50Ω_line and height = height_of_track + height_of_substrate + height_of_ground.

Thanks in advance.

 

[mamali]

First, to see the impedance, go to "Results->S parameter Calculations->Calculate Z and Y matrixes", then go to 1D results tree, choose Z Matrix. Then you can see the Real, Image, Mag and phase of the impedance in the whole frequency range. It is calculated from the S parameters, so it is actually the impedance that you SEE from the port.

So, this is the impedance that you see from the port, including all the components, it's simply something like a Thevenin equivalence of the whole structure.

The impedance of the patch antenna depends on where you excite it. You know that this is a resonant structure with a dominant mode. In that mode, there is a certain arrangement of E and H fields. So, in different points of the patch you see different impedances, and that the way to match the antenna to feeds with different impedances. For example, if you have a prob-fed patch, you can move the location of the probe, and if you have a microstrip-fed patch, you make notches.

I coudln't be sure how did you feed the patch, if it is by a microstrip and if the microstrip line is a 50 Ohm, you can simply assume that the port is at the end of the feedline, which means what you calculated is more or less the impedance of the antenna. To be nore precise (mostly for phase), you can change the reference plane in port definition, and put it next to the antenna.

If you also put your cst file (a zipped project including the file and the associated directory), we can be more helpful.

Matt

 

[andi_]

@mamali, thanks for the reply, I've attached the cst file to this post. The resonant frequency I'm trying to achieve is 5.7GHz and I use εr=4.7

 

[mamali]

Your port definition is incorrect. The width of the port should be between 6 to 10 times the width of the microstrip, and the height should be ~4 times the thickness of the substrate, from lower side of the ground to the top.
However, you have a very unusual thick ground plane, why is it so thick? I thonk you can reduce it dramatically with not much difference, especially with the boundary conditions you have here. The only effect it has is to reduce the back propagation resulting from the finite ground plane, but in order to simulate it, you have to change the boundary conditions: all of them should be "open (add space)".
I have attached a version with thin ground plane and corrected port, with the results. As you see, it is now somehow matched to the feed at the central frequncy, but you can make the matching better by e.g. changing the dimensions of the narrow strip. Here, if you check the port fields, you will see that it's correctly defined and is 50 ohms; so you can assume that the impedance you get is the impedance at the end of the feed, which means at the beginning of the narrow strip. If you need the thick ground plane, let me know.

 

[andi_]

Eventhough this is late, thanks for you help mamali! The problem is actually the design itself. Quarter wave transmission line as feeding technique is not suitable for Fr-4 due to its ε_r, which caused the transmission line, if I remember correctly, needs to be extremely thin.