HFSS 14 Initial Mesh takes very long time

[neoflash]

I have a design of spiral inductor with multi-layer dielectric material substrate.

If I create 20 layers of different SiO2 materials, initial mesh takes forever. (Solving matrix also takes long time.) If I consolidate the dielectric material into 3-4 layers, it becomes really fast.

It seems like consolidation can give me different solution from the original one. I still prefer to use original complex design.

Is there anyway to make the initial mesh generation faster?

 

[volker@muehlhaus]

Yes, that is typical. The FEM method is not efficient for this RFIC work. For planar stackups, Method of Moments solvers (Momentum, Sonnet) are more efficient because that code is specialized for planar EM problems. Both methods solve the full EM problem, so there is no accuracy advantage in using FEM, as long as the stackup is planar.

I've done extensive work on EM simulation of RFIC inductors. FEM can be used if other tools are not available, but most often, the better solution is a planar solver.