Help needed for Patch Antenna With Via Ports in HFSS

Hi all, I happen to have so little experience with HFSS as I am a newbie in antenna modeling. I am trying to design a patch antenna with 3 ports for 802.11n MIMO applications and would like to feed it through 3 via ports. Right now I have created rectangles as lumped ports and placed them just in the substrate, having their integration lines directing upwards from ground to the patch. I am absolutely unsure whether this is a proper port design, hence I got really good results and hope that those are correct. My attempts to insert a coaxial structure into the project failed in many ways. At first I tried to create a small portion of an SMA coaxial (4mm outer radius, 0.6mm inner, teflon in between) in a separate project but I could not manage to get any better than -9dB for the return loss. Though I just pasted it into my project, but this time the analysis was extremely slow, creating tens of thousands of tetrahedra even in the first adaptive pass and eventually gave me the "out of memory" error.

Therefore I would appreciate any feedback upon my current way feeding with lumped ports. I just hope this structure I made does not violate something very serious in terms of electromagnetics. It is even fine by me if these ports may reveal the real results to some degree, although they are not like any conventional port designs.


Here is my project there;
Thanks in advance.

I would appreciate if you could have a look at my project and comment on my port assignment. Even your response would be "This design is no way near an acceptable lumped port", please say so. Thanks.

Your ports are fine, as long as you recognize that they are ideal feeds. The return loss that you get using a coax feed will be more representative of the true structure performance, but for quick sims and tuning these lumped ports will be good. It is of note that you will likely want to use PMLs rather than Radiation boundaries due to the fact that you have more than one antenna in your model. The radiation boundaries may not work as well. But the benefit is that you can bring your airbox down to .1 lambda from your radiating elements.

have fun :)

It's been great help, thank you.