Best procedure for designing stripline for fabrication?

How many commercial fabrication iterations is typical before you get a good stripline result? Also, has anyone had good success designing stripline with Rogers' transmission line application? I am working in Ku-Band.

Thanks.

It depends on a few things, but 3 spins should be good for anyone.
One spin if you're lucky ( easy circuit or common design).

One good practice is to do a "Pizza Run" on your first layout.
That's when you make a circuit from simulation, then repeat it but make everything
a little fatter and repeat it again but make everything a little thinner(~10%). This is
so you can gauge/correlate your sim tool to the layout house.

And always make a 50 Ohm tracew/ length = 1 λ and measure it. Some times a
board house will give you the wrong dielectric or material with the wrong thickness.

Cheers

hi edf
i'm also working on the KU band filter,12Ghz. The stripline filter confused me for a long time. i designed the filter use the RO4003 in HFSS.
i test it use the via. The result was so bad ,the center frequence shifted to the left about 400M .The top of the curve is not flat so much.Test results and simulation results are quite different.
in my opinion, there are perhaps two reasons :
(1)the RO4003 has two versions of the dielectric constant ,3.38 and 3.55. I used 3.38 ,maybe i should use 3.55 to design. i don't konw which one should i use. does anybody know that? please help me, thx.
(2)the bad test result maybe brought by via. dose any one has the experence of using via in such band ? The frequence is so high, the via maybe bring bad S11 and S21. dose anyone known that, please help me,thx.