Assigning port in CST

[Henry797]

Can someone tell me if there is a minimum dimension of the waveguide port for any transmission line? Is it related to wavelength somehow? And also when I try to use only a small area to define a port in a microstrip line in CST it shows "error during calculation"

 

[volker@muehlhaus]

The size must be large enough to capture the electric and magnetic field in the port region. Large enough so that the field outside the port area does not matter.

 

[Henry797]

Hi volker,
Thanks for your reply. Was wondering why CST shows an error, when I assign a port to a microstrip line at the stripline( dimensions of the port= dimensions of the stripline, except for the height/thickness equal to the substrate. Once I increases the width of the port, it simulates well. Now my question is why do I have to invrease the width? Why is it such a necessity?

 

[volker@muehlhaus]

Was wondering why CST shows an error, when I assign a port to a microstrip line at the stripline( dimensions of the port= dimensions of the stripline, except for the height/thickness equal to the substrate.

Can you show a picture?

Once I increases the width of the port, it simulates well. Now my question is why do I have to invrease the width? Why is it such a necessity?

Can you show a plot of the electric field around your line (near the port), so that we can compare it to the size of your wave port?

 

[Henry797]

Please have a look at the attached image and perhaps suggest a reason for such behavior. I've added the E-filed at the two ports just as you wanted.
Many Thanks!

 

[volker@muehlhaus]

Please have a look at the attached image and perhaps suggest a reason for such behavior.

Ooops ... I thought that you would see it yourself from the pictures ....

It is very obvious that the small port area is much to small. In the left plot, you can see that field is not only under the line, but it goes left and right into the dielectric. This is what we expect from a microstrip. CST complains about your small port area because the size is too small to model the fields of the microstrip (which extend to the left and right of the conductor also). Rule of thumb is port width = 3 * conductor width.

Looking at the E field lines at the top of your large port, we can see that they are quite dense. This indicates that you should increase the port size in z direction. Rule of thumb is port height = 3 * substrate height.

 

[Henry797]

Hi Volker,
Many Thanks for the reply. I'm actually aware of the rule of thumbs which is Width(port)=6*Wf(width of feedline) & Height of port(5*h+t) [h=height of substrate; t=thickness of ground plane]. Now my question was actually, don't know whether you've noticed or not that in Fig 2(previous image) The port end(downwards) didn't reach the end of the substrate or ground plane. Now when I created a port that touches the ground plane [height=1.2mm (substrate)+1mm(above)], it doesn't give any error and the results I've attached here. Hence I guess CST ports tolerate a minimum height of port of 1.8*h and also it must touch or come atleast in level with the ground plane.

 

[volker@muehlhaus]

Now my question was actually, don't know whether you've noticed or not that in Fig 2(previous image) The port end(downwards) didn't reach the end of the substrate or ground plane.

Ok, indeed I missed that.

and also it must touch or come atleast in level with the ground plane.

Sure, the port must be able to excite the correct mode of the transmission lines. This means we need both signal and return (ground).

Hence I guess CST ports tolerate a minimum height of port of 1.8*h

Tolerate is the wrong word ... you won't get an error message, but results for the port excitation will not be accurate.

One more comment: In your picture, there are grounds on the sides. That configuration is grounded CPW, not microstrip, and for accurate results we need to include the additional conductors as well. Then, because of the many different ways to excite the 4 conductors, you will need to help CST with finding the mode that you are interested in. I'm not a CST user myself, so I don't know how CST has implemented that mode setting. In other tools, you typically identify conductor with plus and minus to define the mode of interest.

 

[Henry797]

Hi Volker,
Firstly Thanks for your reply. Now you are absolutely right, the design that I uploaded is a cpw, and was trying to assign waveguide port to observe the field pattern across the transmission line. Can you also help me to give me an idea as to how the surface current pattern would look like in a cpw structure, and do you think the pattern would look to some extent different than a microstrip line?

 

[volker@muehlhaus]

Can you also help me to give me an idea as to how the surface current pattern would look like in a cpw structure, and do you think the pattern would look to some extent different than a microstrip line?

Yes, it will be different.

For your grounded CPW, the strips on both sides of the "signal" conductor have to be connected to the bottom ground. With these three conductors at the same voltage, most of the E field will be between the center conductor and bottom ground, but you will also see relevant field to the side grounds, because that distance (capacitance) isn't that much more than the distance (capacitance) to the bottom ground. For correct line impedance, you need all these conductors in the port excitation.

 

[Henry797]

Hi Volker,
Thanks alot for the reply. Was indeed helpful. Now I was also thinking of connecting the cpw to a edge sma port. If I want to do so, do you think I'll get a fair simulation result, provided that I account for the connection between the upper ground planes and the lower ground plane using the metal holder of the sma connector?

 

[Henry797]

Can anyone also tell me, how can I assign a waveguide port which will have 50 ohm characteristic impedance?
Many Thanks in advance

 

[fiwi]

Henry797,see this...**broken link removed**

 

[Henry797]

Hi Fiwi,
Thanks for the video. But I meant to say only using waveguide port, not using an sma, and also not bottom fed. Can you perhaps suggest me a way to assign a waveguide port(Simulation tab in CST) that results in a 50ohm characterestic impedance. Is it only sma that I can use as a 50ohm port? What if my design resonates above 18GHz? Do you have any suggestion for that?

 

[fiwi]

Hi Henry,


,, if you do not understand something, ask ... !!!



- - - Updated - - -

,,,, or so ...

 

[Henry797]

Hi fiwi,
Many Thanks for the reply. I didn't quite get your first method(looked like a coax mid conductor with discrete port). And your second method, trust me I've tried connecting the stipline with the ground using discrete ports and even simulated changing the direction. The S11 & S21 results didn't match with those simulated using a 50ohm sma at lower Microwave frequencies. However, if you happen to have the dimensions of a 3.5mm connector(smaller in size than sma) used for 18GHz and above, could you share it with me? I can't find any other solution to account for a 50ohm feedline in CST.

 

[fiwi]

Hi fiwi,
Many Thanks for the reply. I didn't quite get your first method(looked like a coax mid conductor with discrete port). And your second method, trust me I've tried connecting the stipline with the ground using discrete ports and even simulated changing the direction. The S11 & S21 results didn't match with those simulated using a 50ohm sma at lower Microwave frequencies. However, if you happen to have the dimensions of a 3.5mm connector(smaller in size than sma) used for 18GHz and above, could you share it with me? I can't find any other solution to account for a 50ohm feedline in CST.

,,,well, but I think the best way of solving the problem is to transmit me,the CST file.

 

[Henry797]

Hi fiwi,
Did you get my CST file? Please check your messages. If you haven't received it please let me know.
Many Thanks!