simulating a trace for S11, S12 and Z0, no ground plane

[lnkaa]

I am simulating for Insertion loss, Return loss and characteristic impedance of a trace on a piece of dielectric.
I would like to know if the simulation results will be correct, if I do not use a ground plane (imagine a microstrip but no GND plane).
Is return plane a necessary criteria for calculating S-parameters?

I am a newbie to EM simulations, pls clarify if you know regarding the same.

Thanks
Ink

 

[FvM]

Your design must not necessarily have a ground plane, but each circuit port must have two poles, either single ended or differentially.

A stripline without ground plane isn't a microstrip. It may be e.g. a coplanar waveguide.

I think you should review the tranmission lines chapters of your text books to understand about meaningful transmission line geometries. The prerequisite for measuring S-paramters S11 and S12 is to have two ports, however they look like.

 

[lnkaa]

Thanks for your response.
I am using 2 waveports at each end of the trace. (port 1 and port 2)
I am simulating space-limited, "trace-only" configuration.
where there is no place to use conventional structures like CPW or stripline or microstrip.

I was apprehensive if the simulation results will be correct when there is no return plane, in HFSS.

 

[FvM]

How do you connect the waveports?

 

[lnkaa]

Waveports are placed exactly at the start and end of the trace
similar to the image added

https://obrazki.elektroda.pl/5246859700_1392892891.png

 

[natnoraa]

Hi,

Pardon my beginner's question but shouldn't the waveport touch the ground plane as well? there's also a specific dimension assigned to a waveport rectangle if i am not mistaken.

check this out : https://hfss13.blogspot.sg/2012/05/wave-port-sizing-guidelines.html (but please double check on the source!)
and a tool: https://www.emtalk.com/waveport_calc.htm

Natnoraa

 

[lnkaa]

Hi natnoraa,
Yes. I checked the links. I followed the same when i simulated for microstrip config.

But I am now trying to simulate for a trace with no ground and hence my doubt for result correctness.
any idea about that?

Ink

 

[natnoraa]

Hi lnkaa,

check this out : https://www.edaboard.com/threads/211362/

I am learning from this too

Natnoraa

 

[lnkaa]

Thanks! Natnoraa. It is very relevant.
I plan to simulate such a trace with a very distant ground plane. (implying no immediate ground, i hope so) and compare the results of both (No Ground Vs Distant Ground)

 

[natnoraa]

Hi inkaa,

earlier on i was trying some nonsense by drawing a ground plane (assigned pec to it) at the bottom of my substrate to act as ground with perf E, which is about 700um+ away. i use wave ports and the inductance i got was unreal. well i am simulating for inductors.

You can try the Si sub as your ground plane as it has conductivity itself if I'm not mistaken. I'll usually draw metal layers so that it will be insensitive to whatever is below the metal layer ground

Natnoraa