P@tch antenna in IE3D
I am simulating a p@tch (Sorry for @ but p@tch is automatically getting translated to fix! ) antenna in IE3D.I defined the p@tch to resonate at 26 GHz and AEC is enabled. When I simulated it was resonating at 26. 275 GHz ( minimum of S11 is my reference for resonance).So I added one more small metal rectangle ( so the two rectangles now share a common edge) to the p@tch to bring the resonance down to 26 GHZ.It worked well.Now when I selected the two rectangles and merged them to form one contiguous p@tch ( Adv Edit->Adv Boolean Oper -> Union/Merge ) and then I simulated , the min of S11 shifts to about 26.3 GHz. though the p@tch dimensions are same Why is this ? My guess is that it is due to meshing as more vertices are present in the first case . Which one is more accurate.Should I disable AEC and define the edge cells manually ?
Also Jian , if you are reading this , can you please throw some more light on " Calculating radiation from portion of a structure ".Basically I want to see whether the feed radiation is contaminating the pattern due to the p@tch. I understand that there is no way of separating out the feed radiation and the p@tch radiation as the fields due to the two will be mutually coupled.Hence how should I use this feature of "pattern for selected cells only". Should I then look at the axial ratio as a better estimate for the disturbance due to the feed radiation ?
Thanks for reading this long post .
-Arun
How is the feed modeled in your project?
Rather than worrying about the radiation from the feed, you should pay attention to the accuracy of the input impedance. If the design is for resonance at 26GHz, and the IE3D give you 27.5G, then something is wrong with your setting. I wouldn't start to fine tune it (add small patch etc.) with such a big discrepency.
For the difference of 26.3GHz vs 26GHz: 1) merging the two rectangle should not affect the result if the mesh remains the same; 2) Edge Cell should definitely be used; 3) you should start from a coarse mesh and refine the mesh sytematically to see the convergence of the resonance frequency, my guess is that neither 26.3 nor 26GHz is the final answer.
In brief, radiation from the feed is not the cause of your problem.
Hi Loucy,
Thanks for your reply.Firstly sorry for a mistake in the post.A 26 has dropped out somewhere.It was resonating at 26.275 GHz and not at 27.5 GHz. I have corrected it in my post now.I am using the microstrip feed with the port as "Advanced extension scheme " with 5 cells on the de-embedding arm.I am getting an S11 of -33 dB which I think is a good match.The design formulae as you know are either emperical or based on many approximations. ( Balanis and other books ). They can give you only an indication regarding the dimensions.Hence the fine tuning.Infact even the IE3D manual also gives a fine tuning example ( Sec 5.12 , Electromagnetic Tuning Simulations ).Hence I am using tuning.I will look at Z11 as you sugested for resonance.
-Arun
well, connecting two rectangle changes the resonance of the whole structure (you may use perturbation methods for this), be happy with this little shift!
and, the resonanceis accured when the imaginery part of the Zin is zero. t is the definition. but, if the imaginary part could be affect by feed structure (as in say, L-probe feed), then the best way to know the resonance is the freq. when real part of the Zin is in a maximum, and the imainary part has a second derivative of zero.
rgrds, marti