Discussion Forum

I'm not an expert on this subcategory of ports, but I suggest to you that the key insight here follows from the physics: Lumped-ports are only approximations to reality if/when used over any extended space, so the key is to use them in such a way that the approximation involved is not too severe. E.g., if you really need to model the "thickness" in the lumped-port region to get accurate results, then you should probably be modeling most/all of that particular region instead as a regular part of your FE problem, rather than as a lumped port. Again, remember that a "lumped-port" is not a truly physical thing, in contrast to including an actual transmission-line (TL) and a physical junction. The "real-world" part of your problem of interest presumably includes an actual TL junction, with a distribution of real-world currents on/in it and fields on/around it. If that real-world junction to your antenna is sufficiently small and simple, and if the TL that feeds it does not otherwise interfere/interact with the antenna's currents and fields, then you may be able to drop the details of the TL from the problem and represent its presence via a simple "lumped-port" instead. But if not, then you really need to model the junction itself in some greater level of detail.

So, I suggest you take another look at the physical problem you are trying to represent with your model. There may still exist a better-possible representation which still takes advantage of using a lumped-port, but one in which the lumped-port's specification can be made to be independent of your dipole's thickness.

Thanks for your help. I will try to do what you told me.