I am simulating an UWB power divider and I am getting entirely different results if I change the excitation from "Wave Port" to "Lumped Port". I would like to notice that this happens the same way in HFSS and CST i.e. in HFSS if I use waveport then change to lumpedport, results before and after change entirely the same way it does in CST when I use waveport then change to discreteport.
Hi Ahmed,
If you mean a microstrip circuit, then the discrete port may introduce some additional discontinuity at the connetion point. It can result in additional effective reactive load. In contrast a wave port means a continuous transition to a semi-infinite line of the same cross-secion. Normally the difference should be weak, however in some cases (e.g. in resonant circuits) such a change of port type can result into phase shifts and some changes in resonance position.
Since both methods predict the same, the port type is important to be chosen correclty for this certain geometry. In your situation it is better to model feed poits as close to reality as possible. For example if a coax-to-microstrip feed is supposed, it shold me modeled as well - then a wave port can be assigned to the coax input plane.
Thank you Stanislav,
I wonder if this can be predicted (I mean this harmful phase shift)? Your words mean I can only detect such variation only if I tried both types of ports for each design. Otherwise, I should simulate an SMA connector. It's a buried error if I can understand you correctly :)
Ahmed,
for my opinion, wave port is more suitable for theoretical study of wave propagation in uniform transmission lines (when effects of line terminations must be eliminated). For example, to calculate a propagation constant. Or if one can ensure that in the considered model there is no transition to another type of transmission line in the port's plane.
But if in practice the transition or termination exists, then only detailed model of the feed can give relaible results. If coax-to-microstrip transition is used it can be somehow modeled as a discrete port. In this case the results should be closer to the experiment, compared to ones calculated wuth a wave port.
But again, it is better to simulate SMA connetcor.
Got it, Stanislav. Thank you :)
Zinal, Thank you :) But how can you explain they do not substantially vary, actually they almost vary at all, in the majority of cases ??
Yes,surely you will see a different response. If you simulate a planar structure, it is better to use waveguide port...
best regards
Thank you Hadi, but what are you suggesting if the spacing between the network ports don't allow for near "Waveports" to be placed ??
If you simulate in CST studio, you could use from multiple pin port, this kind of port will solves your problem.
best regards
Do the results of HFSS lumped port and CST discrete port are similar. Actually, I want to convert my design to CST.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence