CST MWS microstrip port size

[elektr0]

microstrip port

Hello,

i still have problems with the size of the waveguide port.
I excite a microstrip structure with four different dielectric materials.
The line impedance, calculated by the waveguide port, depends strongly on the size. The dispersion (frequency dependency of the line impedance) is totally different, if i use a small or large port.

Any reply is good reply.

Thanks

 

[itzikhaim]

cst+2006+advanced+topics+pdf

Hi,

There is known problem about it with CST - try to make it as bigger as possible- You have to ensure to capture the maximun modes (TE/TM) of the EM wave which propagate in microstrip T.L. !!

Look at the manual documents of CST to understand it fully !

 

[elomatic]

higher order modes for microstrip

Hi,
I would recommend to use the entire face of the bounding box as wave guide port. That way it is as big as possible and the lowest mode is excited (I guess that is the one of intrest).

Ciao

 

[elektr0]

cst mws microstrip port size

@itzikhaim

What do you mean with "the maximum modes of (TE/TM)" ?

@elomatic

Using the "FullPlane" option brings problems with higher order modes.


---

When using waveguide ports, a 2D eigenmode solver tries to find out the possible modes. If the port face is too large, there will be higher order modes (hollow wave guide modes) propagatable.
Even if I do not excite these other modes, they are within the calculation (??).

I was not able to use the recommendations from the CST Manual Advanced Topics.pdf in every case, due to stability problems.
In some cases I needed to decrease the port height extensively.

Thanks for reply...
Would be great to discuss it a bit....because I need to use the waveguide port very often.

 

[itzikhaim]

cst size

Hi elektr0 !

Sorry for the late reply.

About your question: I meant with maximum TE/TM modes, that, as you might know, every waveguide have modes that exist and propagate in it. (e.g TE11, TE10 and more)
Every mode has its unique Electric and Magnetic propagation waves.
The size of the waveguide sets the modes to be generated so if you uses the waveguide port (which is rectangular 2D area and may seems to be the edge of the waveguide T.L. - Transmission Line) to be small, you are limting the modes which means the frequency of the EM waves that may not propgate into your model.

You can get some basic knowledge of the modes and how they look like at Pozar - Microwave Engineering excellent book


Hoped I Helped,

Itzik

 

[elektr0]

waveguide port cst stability

@itzikhaim

Hello,

ok may be we can work it out.
You got exactly my problem. Too large size enables higher order modes (hollow waveguide modes) to propagate. Too low size decreases the accuracy of the calculated s-matrix and sometimes results in stability problems, because the energy cannot leave the calc. room.

Anyway, I do only excite mode 1, which is the QTEM mode of the microstrip in my case. DOES THE W.PORT ALREADY BRING THE TE-MODE or DOES THE TE-MODE DEVELOP WITHIN MY CIRCUIT. I would understand the last.

The next problem is to determine, when the higher order mode becomes propagatable. I use the 2D frequency domain eigenmode solver of the waveguide port. If I reduce the size, there is critical size, where all higher order modes are described with an -40dB distance. So I know they are decreasing.
But the -40dB distance is sometimes longer than my circuit dimensions. Means, if these modes develop, they would be there. ???

The next problem is, within coplanar waveguid structures, there are some natural higher order modes. We should consider them. But just considering them and not exciting them, makes it impossible to extract a Touchstone file or even to Renormalize the S-Matrix.
My solution to this problem for now, is to use the MUR-Boundary on the port.
Means special solver setting: Absorb unconsidered modes at port.
So you can only consider mode one, but the other modes are also absorbed.

This topic is very important to me.
So if anybody has any comments, please write it.

Thanks to itzikhaim for reply.

 

[leo333]

te mode tm mode cst

Hi!

What dimensions should have a port for a 50Ohm microstripline (1.1mm width) on a 0.5mm thick Rogers 4003 substrate (Er=3.55, tandelta=0.0027) for CST time dom. simulation for freq 0-40 GHz? It is something wrong width my s21 insertion loss - much more than that of the same line simulated in ADS and in HFSS.
Can somebody help me?
Thanks