Want help in designing 1-2 wilkinson power divider in HFSS

I designed a 50 ohm 1-2 wilkinson power divider for for 2.4 GHz. I don't know why but my S11 is not correct. It is -40db at 2.8 GHz instead of 2.4 GHz. My S22, S33 and S23 are also not correct please help me. This is my HFSS design. I used Appcad for initial line width measurement and then used HFSS seperately for tuning the lines to requirement. Please help.

I haven't had time to simulate the design. However I don't understand why you have copper layers that are almost half a millimetre thick.That might be the reason for the inaccuracy.

mkonca said:

I haven't had time to simulate the design. However I don't understand why you have copper layers that are almost half a millimetre thick.That might be the reason for the inaccuracy.

Click to expand...

Actually I used AppCad to calculate 100 ohm transmission line and optimized it using HFSS (0.35 mm t-line). I think that's the mistake. When I simulated the same design in ADS and used 100 ohm resistor in place of transmission line, the result was satisfactory but I don't understand how can I design 100 ohm resistor in HFSS. Do I have to use some other material?

---------- Post added at 17:11 ---------- Previous post was at 17:10 ----------

mkonca said:

I haven't had time to simulate the design. However I don't understand why you have copper layers that are almost half a millimetre thick.That might be the reason for the inaccuracy.

Click to expand...

Actually I used AppCad to calculate 100 ohm transmission line and optimized it using HFSS (0.35 mm t-line). I think that's the mistake. When I simulated the same design in ADS and used 100 ohm resistor in place of transmission line, the result was satisfactory but I don't understand how can I design 100 ohm resistor in HFSS. Do I have to use some other material?

you need to right click on an object an assign boundary>impedance OR Lumped RLC OR Finite impedance. I have never used these options but I am sure you can find the best one for yourself with a quick google search.

mkonca said:

you need to right click on an object an assign boundary>impedance OR Lumped RLC OR Finite impedance. I have never used these options but I am sure you can find the best one for yourself with a quick google search.

Click to expand...

Thanks for reply it really helped me a lot. I searched edaboard and found many threads on same issue. Thanks.

sarslanali said:

Actually I used AppCad to calculate 100 ohm transmission line and optimized it using HFSS (0.35 mm t-line). I think that's the mistake. When I simulated the same design in ADS and used 100 ohm resistor in place of transmission line, the result was satisfactory but I don't understand how can I design 100 ohm resistor in HFSS. Do I have to use some other material?

---------- Post added at 17:11 ---------- Previous post was at 17:10 ----------



Actually I used AppCad to calculate 100 ohm transmission line and optimized it using HFSS (0.35 mm t-line). I think that's the mistake. When I simulated the same design in ADS and used 100 ohm resistor in place of transmission line, the result was satisfactory but I don't understand how can I design 100 ohm resistor in HFSS. Do I have to use some other material?

Click to expand...

I'm not familiar w/ HFSS, but in ADS Layout, you can use thin-film resistor layers to make nice loads for circuits like this. You might give that a look. Define the sheet resistance to be 50 ohms/square, then make your resistor have a 1:2 aspect ratio. Each "square" of thin-film material will look like a 50 ohm resistor. If you have a long, skinny strip, they will add up in "series"... so a 1:2 aspect ratio will give you 100 ohms (R = Rs*L/W.... 50*2/1 = 100 ohms). Conversely, a short, wide thin-film resistor will act like "square" 50 ohm resistors in parallel, so a 4:1 aspect ratio would yield 50*1/4 = 12.5 ohms.

In all cases, make sure that your resistor is electrically short (< 5 degrees), or you'll mess up the reflected signal phases, which will reduce the voltage-wave cancellation and hurt your port-to-port isolation.

For these frequencies, you could use a thin-film R that is similar in dimensions to your final packaged part (0201, 0402, 0603, 0805, etc), then simply install the real part when/if you fabricate it.