Modeling Scattering Parameters

[Humberto Riveiro]

Hello Everyone

This is my first post.

I wanted to know what are the real applications of modeling (or extracting) s-parameters from a transmission line using the telegraph equations and its characteristic impedance.

I saw a few papers, all of them talk about measure the s-parameters using reflectometers and then use that parameters for describe the device under test (Transmission line for example).

My question is, if people know the material and the geometry of that device, why they don´t model mathematically the s-parameters instead of measure them?

I understand that some devices are dificult to model because of its geometry. This is the only reason?

Thanks in advance

Humberto

 

[volker@muehlhaus]

My question is, if people know the material and the geometry of that device, why they don´t model mathematically the s-parameters instead of measure them?

I understand that some devices are dificult to model because of its geometry. This is the only reason?

Yes. You just have to decide what is easier. For standard cross sections, there are good scalable models available. And if you have an EM simulator ("field solver") you can get S-parameters from simulation, based on dimensions and materials, even for exotic geometries. Some users have more trust in measurements, so they might prefer that even if simulation is possible/available.

 

[Humberto Riveiro]

Thanks for your response!

But if I want to make my own model, for example, of a microstrip TL at different frequencies and calculate de return loss for every frequency, how can I achieve that model (and the corresponding S matrix) using Telegraph Equations such I can make an accurate model and get parameters similar to that obtained by measurement?

 

[volker@muehlhaus]

how can I achieve that model (and the corresponding S matrix) using Telegraph Equations such I can make an accurate model and get parameters similar to that obtained by measurement?

You can use Telegraph Equations to calculate S-parameters if you know the R', L', C' and G' values.

You can use a field solver to calculate these values. Usually they are frequency dependent due to skin effect etc.

 

[Humberto Riveiro]

Thank you
I just want to start with a simple model of a single ended microstrip whic I know the characteristic impedance. I will try with an uniform impedance along the whole distance and see what happends.

I think for that I don't have to know RLCG, just how to calculate Zo which is function of the geometry and the permittivity of substrate. Is this correct?

Thanks in advance.