In HFSS, why ports need to far away from discontinuity?

I know it is for high order modes to die out before they reach ports. If they are real, why not include them in sim? Thanks.

Physically, the excitation should just be the fundamental mode, since your cable is very long and high-order modes die out. However, in the simulation the port is modelled by a finite surface with the field distribution of the fundamental mode. There will be unwanted modes excited, but they are not physical.

Hope this helps.

Do you mean ports can excite or accomodate non-physical modes ? is that why transmissio line to DUT should be long enough? Thanks.

Yes, these can be caused numerically.

lhasayak:

I think it is so that the stray fields from your discontinuity are not picked up in the field port. This would cause numerical error in the S-parameter extraction.

--Max

It is a limitation of Wave ports.
But I think you can use Lumped ports anywhere inside the geometry. More explanation:

Wave ports solve actual field distributions for one or more propagating or non-propagating Modes.

Lumped ports excite a simplified, single-mode field excitation assuming a user-supplied Zo for S-parameter referencing.

Limitation of lumped ports:

Lumped ports excite only one mode, and therefore are not appropriate for excitations where modal superposition is expected.
Since impedance is supplied by the user, not computed, no alternate definitions (Zpi, Zpv, Zvi) are supplied.


More information can be studied in the attached file.

good luck!

Yes, you also have to be careful not to use lumped ports if they could possibly couple to other things near them. Since they are not de-embedded, you should be careful about using them at high frequencies.