Different versions of IE3D give different simulation results

[mstripman]

IE3D Different versions

Hi All

Well first of all let me mention that I have simulated gap coupled microstrip antennas in which 11 patches were coupled to driven element. the simulation was done using IE3D v 9.21 and the first fabricated model itself matched the simulated results. But for other gap coupled configurations I found that different versions of IE3D gave different results. For eg. for gap coupled MSA consisting of 2 elements, v9.21 gives 68 MHz BW, while 10.33 gives 75 MHz and v 11.15 gives 88 MHz. Why there is so much disparity between different versions.

Sincerely
mstripman

 

[jian]

Re: IE3D Different versions

Hi, mstripman: In our tests, different versions of IE3D versions are quite consistent. However, we did change the Automatic Edge Cells (AEC) feature of IE3D. the parameters are different in IE3D 11. It may not covert the AEC parameters correctly from IE3D 10.X or earlier. If the AEC is not in the right range, it may affect accuracy when you are modeling edge coupled structures. Please send me the files and I will help you jian@zeland.com.

 

[mstripman]

Re: IE3D Different versions

Hi Jian,

I did not use the geo file of v 9.21, instead I constructed a seperate file for v 11.15. What I tried to simulate was non-radiating edge gap coupled MSA consisting of two elements. One was fed using co-axial probe while other was gap coupled to one of it's non-radiating edge. What I was interested in was the maximum bandiwidth obtained using this configuration. So accordingly I optimized the parameters i.e. gap width and feed point location. what I observed was that both these version gave different values for optimized value of these two parameters. And particularly for maximum bandwidth, using v 9.21 I could only get a bandwidth of 68 MHz, whereas v 11.15 gave me 88 MHz. That is what I wanted to know as to why there is so much differnce between these two values.

Is it only for this particular case of non-radiating edge gap coupled MSA's, as some other papers also have reported disparity between theoretical and measured results.

Sincerely
mstripman

 

[jian]

Re: IE3D Different versions

Hi, mstripman: Again, for edge coupled feeding structure, you need to be careful on the edge cells. If you do not model it correctly, you may not be able to get high accuracy results. IE3D has automatic ways to choose AEC for you. However, you still need to give it some guideline. You may want to check the meshing to make sure the edge cell size is appropriate. If the edge size is obvious larger than the gap size, you may under estimate the coupling. If it is very small compared to the gap size and the regular cell size, it may also affect the accuracy. You need to define the correct AEC ratio in IE3D 11. In earlier version, AEC is defined by absolute value. They may differ if you do not make sure the AEC cell size is about the same because edge cell size is critical to the accuracy for edge coupled fed structures. On IE3D 11, you also have the option to use multiple layers of AEC for extreme accuracy using the contemporary meshing scheme. This scheme was implemented to meet the requirement of some customers for single pass design of -40 dB coupler. In MTT 2004, I met a customer who was doing coupler design. He told me he could use IE3D for single pass design of -20 dB coupler. Just define the gap as variable and run the optimizer. He could get the optimized geomtry in a some time and the optimized geometry would meet his requirement. However, when he tried to design -40 dB coupler, he had to manually create multiple edge cells in order to achieve the goal with more iterations. For this reason, we have implemented the multiple layers of AEC in the IE3D 11 for high accuracy requirements. In case you do not get very accurate results using different single layer AEC, you may consider using 2 layers of AEC for it. You can suppport me the files at jian@zeland.com and I can help you.