Re: UWB Antenna using IE3D
Hi:
I would like to make a few comments on this structure:
The gnd is very small compared to the patch. I am afraid there is a significant inbalance between the patch and the gnd at the feeding point. The inbalance may cause much uncertainly in the results, no matter whether measurement or simulation is concerned. This is a similar situation when you try to meausre a dipole antenna with the 2 arms not the same length or you are using a coaxial to feed a dipole without any balun. In the reality, the current will go onto the outer surface of the coaxial feed line and become part of the radiator. If you measure the impedance, depending how you feed the antenna, you will get different results because the feed line is part of the antenna. I think it should be one of the most significant source of error involved.
Some users discussed with me on matching the measured results and simulated results for CPW fed antennas. It is the same problem. They didn't make sure the feed is balanced between the positive arm and the gnd of the antenna and it introduces all kinds of uncertainly into the measured or simualtion results. If the structure is symmetrical such as symmetrical dipole antennas, or the positive arm is significantly smaller than the gnd (monopole on a big piece of gnd), the feed will become well balanced and you will get more consistency between measured and simulation results. In case you have a unbalanced structure, you should need to take some measure to make sure what you are simulating and what you are measuring a consistent. If you don't get good match, it is quite likely due to the setup of your measurement and the setup of the simulation are not consistent. Regards.
Added after 8 minutes:
More comments on the simulations:
1. You don't need to choose so few frequency points for such a simulation. You can define the frequency range from 1 to 15 GHz with couple hundred frequency points for smooth results. It will slow down much.
2. If you choose the Contemporary Meshing, it will reduce the meshing and you will get faster results.
3. If y ou can divide it manually or use the Adv Edit->Rectanglization to divide into more rectangular shapes, you will also be able to speed up the simulaton much. Anyway, as it is, the speed is quite good already. It take a few minutes on a good computer to run. Some simple revision may make it better.
Added after 10 minutes:
Sorry. I had a typo in the additional comments.
Putting couple hundred frequency points will allow you for smooth results without slowing down the speed. You will accurate results in about the same time as long as you use Adaptive Intelli-Fit (AIF). Regards.