Port Placement CPW

[khadkasushan]

Hi,

I am trying to simulate CPW lines, but I am quite confused about the port placement. The purple and red boxes are lumped ports. Is the way I have placed the ports correct?

Also, considering the purple box as Port 1 and the red box as Port 2, shouldn’t the p_dir be y and x, respectively?

I tried going through all the previous discussions, but I could not find a clear answer to this.

Thanks.

[LubomirJagos42]

about field direction if it's gnd plane around traces then yes, port direction must be same as direction source and target layer so if gnd plane is not below trace but around then definitely port direction must be for one port X for second Y

[khadkasushan]

Hi, thanks for the clarification.

I have an additional question related to my setup. My structure is actually a CPWG (Coplanar Waveguide with Ground), meaning I have:
- Coplanar ground planes on the same layer as the traces
- An additional ground plane on the second PCB layer below the traces

You mentioned that the port direction should point from ground to center conductor in the coplanar direction. However, since I also have a bottom ground plane, the E-field exists both horizontally (coplanar ground) and vertically (bottom ground plane).
My question is:

• Does the presence of the bottom ground plane affect the port direction (p_dir) definition
• Should I still use p_dir='y' and p_dir='x' for my two ports, or does the CPWG configuration require a different port setup?

[LubomirJagos42]

hi, there is example in openEMS example for coplanar port, but since I started looking on it and trying to make some simple example for it I got that I never tested coplanars port and need while to make it work :D

[LubomirJagos42]

so I rerecreated example from openEMS octave scripts in freecad and it's as I wrote at beginning, p_dir is wave propagation direction and evec is excitation direction so p_dir must point direction where your stripline is heading and excitation should be orthogonal to it as it's vector for excitation E field between stripline and ground, in case of normal coplanar stripline it's clear it should be 'y' about grounded coplanar stripline I'm not sure but if it you generate excitation between stripline and top gnd then 'y' if you generate excitation between stripline and bottom then 'z' that one I would have to try to be sure

[LubomirJagos42]

when I wrote about this grounded CPW with Gemini AI it really depends how thick is substrate, if substrate is thin enough and prevail fields between bottom ground and stripline input port should be microstrip line, if it's thicker and both top ground and bottom ground have impact on field then lumped port which covers gap and bottom ground with e field direction in 'z' should be used

[LubomirJagos42]

another thing when I made simulation of CPW you must used AddCPW... port method since it adds lumped port on both side of stripline in your picture you have lumped port just on one side of that coplanar stripline what is wrong

[khadkasushan]

Hi, thanks a lot for the input. I simulated using the lumped port earlier and I got the results, which I have attached. Anyway, I have also attached the CAD file of my design if you are interested in checking the CPW port for the simulation, as I have compared the actual VNA measurements of the PCB with the openEMS simulation using the lumped port :D I will also try myself simulating with CPW port and let's see how it compares with the simulated result.

freecad_file+results.zip

[VolkerMuehlhaus]

Unrelated to simulation: the large pad at coax center connector is not good RF routing, it will add massive shunt capacitance. If you need to design for controlled impedance lines, this "design with Z=sqrt(L'/C') in mind" also applies to the connector transition region.

[khadkasushan]

Thanks a lot. I was not considering it.