Even and Odd-mode Wilkinson power divider

[belma1511]

Hello,

My task is to calculate all the impedances of a 1-4 Wilkinson power divider by using the even and odd-mode method.
I'm not quiet sure how to do that,so could anyone help me with that?

 

[belma1511]

Ok now, I designed a 1-4 W. power divider at 5.5 GHz frequency and substrat FR4 in Ansoft designer. The results for S-parameters are not bad but they have to be better. Any idea how could I optimize my power divider?

 

[jeeudr]

I do not have Ansoft designer, so I am unable to open the file that you have attached. It would be better if you could provide pictures of your design and S-parameter plots.

Is it an equal split 1:4 power divider ? If so, why don't you just arrange three 1:2 Wilkinson power dividers in a tree fashion ?

 

[belma1511]

It is an equal split power divider and I arranged a cascade of three 1:2 W.divider, with Z0=50 Ohm. I used A.Designer just to calculate width and length of the microstrip quarter-wave trans.lines. Furthermore,my main task is to design a 1:4 W.divider in HFSS v.12.
I'm not really familiar with HFSS so I would really appreciate any kind of help and suggestions (even how to start,how to design a microstrip trans.line,etc.)
Here are the images for the S-parameters.
Thanks in advance.

 

[jeeudr]

I have used HFSS (a lot) but never Ansoft Designer.

Here is an example file from HFSS which is supposed to be a 1:8 corporate feed. It will provide you with an idea on how transmission lines / substrates / ports are realized in HFSS.

 

[belma1511]

Thank you a lot! Also if you have more tutorials to become familiar with HFSS don't hesistate to post them!

Please let me know if you already glanced over the S-parameters of my divider and if you have any suggestions how to optimize them.

 

[jeeudr]

Remember the following points

1. The conventional Wilkinson power divider is a narrow band device. So fix your frequency band of operation and try to optimize within those constraints. It would not be fruitful to try for anything over 20% bandwidth of operation.

2. Even if you assume no loss in the resistors, the maximum you can reach for |S12|, |S13|, |S14| or |S15| is going to be -6 dB (because total input = total output power). Due to power loss in the resistors, fabrication tolerances, resistor tolerances, substrate losses etc... you will generally end up with some more energy being lost. If you can get final HFSS simulation values better than -7 dB, that is very good.

3. |S11| < -20 dB is generally enough for most specifications. It is only for rare applications that you need |S11| to be any lower.

4. Ansoft Designer seems to be a circuit simulator and there is only a limited level of accuracy that you can achieve with it. Solve your design with HFSS and then we can get a closer prediction of real-world performance.