I am designing an on-chip inductor without metal shield and uses momentum to model the inductor. but then I wonder if Agilent momentum is able to calculate the effect of the eddy current in substrate. I heard from someone that because it is 2.5D simulator, the eddy current in the substrate is not taken care of. Also, there seems to be no meshing for the substrate. However, there is a resistance defined for the substrate, and also green function seems to be used, so I think it should be giving a 3D result.

Actually I do not understand what is so called 2.5D simulation. I think that the method of moment should really be giving the accurate 3D result, since I also used to be using it in antenna simulation. Or is Agilent momentum not based on the method of moment or only a simplified version of it?

Thanks!

Actually I do not understand what is so called 2.5D simulation. I think that the method of moment should really be giving the accurate 3D result, since I also used to be using it in antenna simulation. Or is Agilent momentum not based on the method of moment or only a simplified version of it?

Thanks!

> I am designing an on-chip inductor without metal shield and uses momentum to model the inductor. but then I wonder if Agilent momentum is able to calculate the effect of the eddy current in substrate.

Yes. Substrate eddy currents are included.

> I heard from someone that because it is 2.5D simulator, the eddy current in the substrate is not taken care of.

There are some 3D solver marketing guys who try promote their tools by such wrong statements. Eddy current is included in the Momentum simulations, and Momentum is well suited for conductors over silicon substrate.

> Also, there seems to be no meshing for the substrate.

Correct. The currents in the lossy substrate are included in the results, even though the substrate doesn't need to be meshed. Actually, by including the substrate currents through the Green's function, with no need for meshing the substrate, this approach is more accurate then meshing the substrate (which is always approximate due to finite mesh size)

> However, there is a resistance defined for the substrate, and also green function seems to be used, so I think it should be giving a 3D result.

Yes.

> Actually I do not understand what is so called 2.5D simulation.

The limitation is that conductors can only be on the dieelectric layers (horizontal) or vertical for vias. We can't have arbitrary shapes with arbitrary orientation. That is why some call it "2.5D" or "planar 3D".

> Or is Agilent momentum not based on the method of moment or only a simplified version of it?

Yes, Momentum uses MoM.