miniaturized Wilkinson power divider

[Nicholas_UTM]

Hi friends,
I am Nicholas from University Technology Malaysia. I am doing my final year project on a 2-way miniaturized wilkinson power divider.I have a few confusion and need help. I manage to get results as below:
S11=-32dB
S22=-26dB
S33=-26dB
S12,S21,S13,S31=-3.01dB
S23,S32=-30dB

Are we suppose to get the same s-parameter values for S11,S22 and S33?If the valueare not same as above,whatwill be the problem?
By inserting groove to the layout design, i manage toget almost same value for the returnloss.Is it better?because i have seen designs with almost same value and designs with different values.
Another question is, assuming 10% power loss at each output arm,is itstill acceptable in the industry for 40%:40% equal power split?
Thank for the help!

 

[biff44]

Looks fine to me. Due to symmetry, you would expect S22 and S33 to be identical in both phase and magnitude, but S11 could be entirely different.

Try simulating the results with circuit loss, and see if you still like the performance.

 

[Pushhead]

Hi Nicholas,

I agree with biff44. For a lossless model the results are fair enough.
Try introducing dielectric and conduction losses to your model, and then see what happens.
%10 precent losses sounds a little bit high for a "small size" divider, but it depends on the specific design. for example, a low power application can suffer losses without overheating the circuit, whereas high power application will probably melt the circuit.

Try using different substrates and design a minimum loss model.
Assuming you implement the divider by microstrip, you can find a good calculator here:

http://mcalc.sourceforge.net/

Good luck,
P.

 

[daoyuan]

I suggest Genesys simulate ,it supply many
useful thing

 

[DDavid]

HI,

I suggest to you not GENESYS take ADS he is more powerful and more true
lately i design wilkinson splitter 300MHz to 3GHz on microtrip (fr4)...
And after I produced it I got exactly the same result as on simulation...

David

 

[sweetughirish]

S11=-32dB
S22=-26dB
S33=-26dB
S12,S21,S13,S31=-3.01dB
S23,S32=-30dB

Are we suppose to get the same s-parameter values for S11,S22 and S33?
No..you should not get the same value for this 3 parameter, S22 and s33 must be same while s11 must be different

If the value are not same as above,whatwill be the problem?
One arm might be longer then the other or practically the manufacturing can be a problem

By inserting groove to the layout design, i manage toget almost same value for the returnloss.Is it better?
No, you should get almost same as the shematics did cater a good model

because i have seen designs with almost same value and designs with different values.
Layout is a good EM simulator, it will give you a result which might resemble the exact result you will get once fabricated

Another question is, assuming 10% power loss at each output arm,is itstill acceptable in the industry for 40%:40% equal power split?
Err...power divider should not consume much power. If you do have a problem with there must be something wrong with your model[/b]

 

[kevinj]

HOW MUCH IS THE POWER DIVIDER MINIATURIZED, what %?