symmetry within SonnetEM

[elektr0]

Hallo,

is there a possibility to use decrease the ammount of memory if my structure has one or more symmetry axes.

Thx.

greets elektr0

 

[maxwellian]

elektr0:

If your structure has symmetry and all of your ports are located on and centered on the axis of symmetry, you can use the symmetry feature in Sonnet. It should save you a factor of 4 for memory and around a factor of 8 for simulation time.

I think they can handle one axis of symmetry, which is essentially a magnetic wall. If your circuit has an electric plane symmetry as well, you might be able to use a wall of the PEC bounding box to create an electric symmetry plane. But that is only if you don't have ports on that side of the circuit, I think.

--Max

 

[elektr0]

i am talking about a symmetrical differential planar inductor, looking like this

................|------------|
--port1-----................|
................................|
--port2-----................|
................|------------|

When looking at my drawing, IGNORE ALL DOTS.


This means there is no possibility to use the symmetric configuration for better simulation speed.
RIGHT ?

 

[maxwellian]

No, you can't use symmetry for this case, unless you're interested in common-mode S-parameters alone (which is what you would get if you used the box symmetry/H-plane symmetry feature in Sonnet).

For full general S-parameters, you need to simulate the full structure.

If you have Sonnet Professional, you could use Conformal Mesh, which really speeds up inductor simulations and other simulations with transmission lines that have high memory requirements. It can be combined with Sonnet's thick metal simulation to create pretty efficient inductor analysis.

--Max

 

[elektr0]

Hallo maxwellian,

i've got a full version, but the coil has 8 layers and is too complicated for sonnet

 

[maxwellian]

elektr0:

Have you applied conformal mesh to all of the coil windings? That should help the simulation requirements quite a lot.

--Max

 

[elektr0]

conformal meshing did not work.

But I will use ansoft HFSS for it.

Added after 19 minutes:

thx to all

 

[Joe_User]

elektr0,

Have you tried sending your file to Sonnet Tech Support? I had a pretty complex multi-layer coil which I thought was too complicated for Sonnet too, but then I contacted their support guys and they taught me how to be more efficient. So my suggestion would be to send your Sonent project file to support. Sonnet is more accurate than HFSS for planar coils, so it should be worth it.