i made a braodband ( 1-3 Ghz) wilkinson power divider with FRF4 substrate and it works very well.
but when i try to make another one (2Ghz - 18Ghz) with FRF4 Er = 4.5 or Duroid Er= 2.57 i've got problem with S21 parameters .
between 6Ghz and 18Ghz i've got a S21 who decrease between -8dB an -30 dB?
can somebody tell me why? it's the substrate is the cause of my problem
which topology is the best ? straight lines or curves?
Re: what is the best substrate for a microstrip power divide
Definitely the reason of poor S21 at those frequencies is the FR4 material.
For higher frequencies you can use one of these materials from Rogers: http://www.rogerscorporation.com/mwu/pdf/prodselect.pdf
Curves lines usually are used in broadband splitters to minimize the total physical length.
Re: what is the best substrate for a microstrip power divide
a good material at high freq is RT 5880 or materials from Taconic like RF-35. RT 5880 has low tand=0.00019 n is very stable at high freq works very well in Ku band.
it is not only the loss tangent. there are more effects associated to the substrate material which are important at high frequencies: surface waves, which are generated by leakage and are linked to the substrate thickness and epsilon.
I suppose you use SMA connectors soldered to the MIC traces. These connections are also important.
You also mention that it is a wilkinson splitter: do you know the parasitics of the resistor?