broadband microstrip filter

I want to design a broadband(50%) bandpass filter with a 1mm thick FR4 board at the center frequency of 600MHz. First,I used the interdigital model.But the S11 was very poor.I can only lay out the microstrip filter on a Maximum area of 10cm X 10cm.So I can't use 1/4 wavelength stub model.the hairpin and combline model can be only applied to narrow band filter.Can anybody help me?TKS very much!

Use "stripline" as shown in the figure.this structure can achieve high coupling coef.
an then can achieve wide band.

_______________________ top

air
____ _____ strip
----------------------------
---------------------------- FR4

air
_________________________ bottom

at 600MHz, why not use lumped elements. or at least, use lumped capacitors for coupling.

Hi,

most Rf designs on FR4 suffer from inacurate input of dielectric constant and that can be very critical for filter application. Also, depending on a particular manufacturer, dielectric losses can be limiting factor for achieving good insertion loss of a filter. As stated above, you might be better off using lumped elements.

flyhigh

TKS very much,you are so kind.
But this structure is so hard for general manufacture .
I see many books and papers saying interdigital model can achieve a broadband function.But I can't get the result with microstrip.whether it's impossible achieving a broadband function with interdigital microstrip? ?

The bandwitdh you need is very high.
Normally filters based in coupled structures achieve up to 20%.
I don't know if an optimization process would help you to go to higher bandwidths.

For a high bandwidth I would propose you to make 2 separate filters, a high pass (using lumped capacitors and inductors or distributed lines) and a low-pass (distributed element plus adjustable capacitors).

After the 2 separate designs are made, you can then use an optimizer to do a fine adjust and get the desired response (correction needed after 2 filters will be cascaded).

S.

it is a so called 'suspended microstrip'.you should look up to text books.
in fact, coupled microstrip can not achieve high coupling because the the thickness
is limited,but as to the suspended microstrip,the air under the substrate add to the thickness, and high coupling can be achieved.
here the FR4 is also used to support the conducting strips.

Hello Amicloud.
Would you happen to have a paper about coupled lines in suspended microstrip technology?
I'm surching for a paper where I could find closed or approximate formulas to get coupling coeficient, effective dielectric constants (even and odd) and even and odd characteristic impedances,for given W (strip sidth), S (gap between strips), h1 (substrate thickness) and h2 (air gap).

Thanks.
S.

If i were u....i will go for SMD lumped element as such a low frequencies device.....bear in mind....when using lump element.....there are equivalent circuit for C and L..................