butterworth filter optimization help using ADS

[TerryADS]

Hi all,
I want to translate Butterworth lump component LPF into microstrip filter.I calculted and done.But when i simulated it,I found it's cutoff frequency is not 2.4GHz.I used optimization in ADS2008,spent much time,but can not get EF=0,How can i get a effectively optimization?(Attached my file.)
Thanks,

 

[TerryADS]

Has anyone help me ?

 

[kspalla]

You have 3 different values c=2.15pf, L1=2.05nH and L2 =6.63nH which requires line widths W1, W2 and W3 too. Just with two line widths its not possible.
So increase the variables.

 

[TerryADS]

It is still hard to get the result.Is it something wrong with my calculation or other reason?I set Zh1=110 for 2.05nH,Zh12=130 for 6.63nH,Zl=11R for cap.Anyone share your experience about optimization?

 

[RealAEL]

Some experience when using optimizers:

1. Use Random optimization until the results are getting close. Using Gradient to start will typically be slower.
2. Using a goal for just the difference between S11 and S33 is not sufficient. You should include at least S21 and S43 difference also. Maybe also phase difference as well and magnitide. Weighting will then become important.21
3. Widen the variable limits a bit. L11 and L22 are already at or close to their limits.

Finally look at the layout.

You cannot implement TL3 and TL5 a very wide but short straight transmission lines. These are open circuit stubs and should each be implemented with MLOC or MLEF elements with an MCROS in the middle.

 

[TerryADS]

I fixed it now.Thank kspalla for help to me,especially thank RealAEL for warm-hearted help and detail explain.