Meander simulation in Sonnet vs HFSS

[danial314]

Hi everyone,
we are looking for inductance and capacitance of a meander with 400 legs. ( the exact parameters of the meander is bellow in the figure)
we placed the meander on the 500 um Si substrate and 500 um air on top. we run the simulation for S-parameters (S11) on sonnet and got resonance frequency 7.8GHz. Meander was defined to be lossless material in sonnet.
we tried to reproduce the s-parameter and inductance respectively in the HFSS and Maxwell both for the same Meander with perfect electric conductor but unfortunately our result are quite different. resonance is around 1GHz.
in SONNET: the inductance and capacitance that we extracted from the meander by fitting the data is around L=20 nH and C=22fF
in ANSYS Maxwell: the inductance and capacitance that we extracted from the meander by fitting the data is around L=170 nH and C=150fF.
What is the reason for this discrepancy?

 

[volker@muehlhaus]

Is your Sonnet cell size too large and creates a short between some legs?

 

[BigBoss]

Could you arrange the Boundary Conditions in HFSS similar to Sonnet ? The discrepancies may come from Boundary Conditions due to Box Resonance. Because the Meander Line is too long and comparable to Substrate Height.

 

[volker@muehlhaus]

Yes, good point as well. The box walls are always metal walls in Sonnet, only top and bottom can be switched to absorbing.

 

[danial314]

Is your Sonnet cell size too large and creates a short between some legs?

The cell size is same length as Meander conductor width and conductor spacing (I showed them in the figure)
is that make sense?

Could you arrange the Boundary Conditions in HFSS similar to Sonnet ? The discrepancies may come from Boundary Conditions due to Box Resonance. Because the Meander Line is too long and comparable to Substrate Height.

I made the boundary condition exactly same. can you please explain a bit more about your last point? you meant that I have to make the substrate surface area larger or decrease the thickness of the substrate? (sorry, I did not get what height here means)
thanks

 

[BigBoss]

I made the boundary condition exactly same. can you please explain a bit more about your last point? you meant that I have to make the substrate surface area larger or decrease the thickness of the substrate? (sorry, I did not get what height here means)
thanks

If the Height of the Substrate is Comparable to Structures' Dimensions, it can be cause of some computation errors. The best way to obtain a right solution is to use the Right Tool.
HFSS is a true 3D FEM simulator and it's essentially has been intended/developed to simulate larger systems like Antennas, EMI/EMC Problems, Airplanes, Waveguides etc. Therefore sometimes it can be troublesome.
I recommend you to use Momentum or SONNET for these kind of structures.
It doesn't mean that HFSS cannot never be used .. NO, of course it can be used but it needs more experience and attention because its target is different.
I'm not expert on HFSS but I observed every time that HFSS was a little bit troublesome in terms of Settings due to reasons which I have counted above.