ADS, Not to Consider Parallel MS Lines

[moreho]

Hi All
when i design a microstrip circuit in the software Agilent ADS, i may insert multiple elements such as a number of parallel ms lines.
in theory and simulation, as a true concept, these lines may have effects on each other, but when i make the circuit, i may have some walls between these lines, therefore they can't effect on each other.
can i consider this walls in the software (Momentum Environment) or even can i use special settings in momentum that can remember ADS not to simulate the effect of these lines?

thanks in advance.

 

[volker@muehlhaus]

I am not sure if I understand your question/problem.

If you want to model lines as shielded/uncoupled, you can include the shield in your Momentum model. This simulates the real layout, including the reduced coupling effect.

Or you can move the lines into separate EM models, where they can't couple.

 

[moreho]

the second issue is that i knew it but did not use because of a little difficulty and complexity of creating and managing multiple individual layouts...
but the first issue is one that i am seeking it.
so can you explain how can i include the shield in my momentum model?

thank very much, dear sir.

 

[volker@muehlhaus]

so can you explain how can i include the shield in my momentum model?

Include the shield metal and vias in the EM simulation.

 

[moreho]

do you mean that i add metal traces between these lines as those metal walls that i have in my circuit box?

 

[volker@muehlhaus]

do you mean that i add metal traces between these lines as those metal walls that i have in my circuit box?

Yes, you could add vias from top infinite ground to bottom infinite ground. The vertical length of each meshed via segment should not be larger than 1/10 lambda, so you might need to split the substrate and air layer into multiple pieces.

Sounds complicated? Yes, I would not want to do this. Instead, simulating the layout in multiple separate EM blocks is much easier, if you use subcircuit hierarchy where each "box" has a separate layout block. It also make more sense for the simulation method (simulation time and memory) because the MoM memory requirement scales with N^2 where N is the number of mesh cells. So splitting into multiple smaller EM models is a good idea, if they don't couple anyway.