I am tring to simulate a CPW in HFSS, but when I set Delta S as 0.01, 0.001, 0.0001, the RF mode index I get is 2.42, 2.35, 2.33, and the simulation time become longer and longer. I find that in most website they said that the Delta S set to be 0.01 is enough, but in my case it seems not right. Can anyone give me some suggestion about how to simulate a CPW for electro-optic modulator?
of course. you set higher precision, and it takes more memory and time to solve.... and the number of re-meshing steps until the solutions converge grows geometrically. those are the basics of any full-wave simulation.
you should simulate some basic representative structure and see the field distribution. increase precision until you feel that the field distribution is fine enough, and the S-params solution is close to your practical data. then make a researched structure simulation with that setup.
P.S. with CPW you may find that defining a dielectric and air area separately just around, and inside the CPW, and seeding those areas basing on the via space length, or substrate thickness is beneficial, and helps cutting on converging steps.
You see, the FEM solution is always correct with respect to order of field interpolation between the nodes. But the mesh may be to coarse to give you a realistic result. So the software in its large part is an advanced mesher AI, which analyses the coarseness and adds more nodes where needed, until the solution does not change anymore (the Delta S is reached)
Just forgot to say. HFSS of latest versions may lie about the port impedance for complex transmission lines. I.e. on some random occasions it is mixin up Zpv, Zvi, and Zpi. It may lead you to completely wrong conclusions where that is important. It was discovered on simple rectangular waveguides. So I would recommend forgetting about Ansys, and taking CST, Comsol, or ADS
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators