Bandpass filter design approach

Hi all,

It has been a constructive session for the last hour after browsing all the threads on this topic. It provided me a surface understanding of the availability of methods around.

I wish to design a 1.5GHz BPF. I have been constructing a 7th order BPF based on AADE filter design software. I used the values provided at 1.5GHz and begin soldering the LC on a FR4 board. After 2wks, im still unable to produce one that barely make it near 1.5GHz. From these, i concluded that the filter software does not take parasitics and board substrate into considerations.

Is it possible to produce a 1.5GHz BPF? The LC values im dealing with are in terms of pico and nano. Its very difficult for me to achieve the exact values as per simulated. Is there any better filter software that accounts for parasitics and board details?

Im keen in exploring a microstrip BPF design and i have just started looking around. I have a few terms which im unsure of "tapped line", "center stubbed resonator".
Thank you all!

Hi,
1.5 GHz is a frequency that is high for lumped components and low for microstrips.
A microwave filter will works but it will be no so compact as the lumped one (that will have probably worse insertion loss performances).
However it's hard to design such a filter (the 7th order is quite high) without help of a simulator that's able to take into account how the board affect the performances. You can try "QUCS": a freeware simulator able to simulate transmission lines such as microstrip and stripline.
By means of QUCS you'll can try both microwave and lumped components network to compare them. I strongly suggest you to make use of s-parameters model for L and C instead of the ideal parts. these last can be used only at the beginning, just to see what about the frequency shift caused by the board tracks.
Regards.

Accurate filters are not constructed with LC elements at that frequency due to their tolerances,standard values,temperature effects and non-ideal models..
Also, board and its prasitic effects will shift the filter characteristics.

You can build a microstrip filters on a PCB boad by EM simulating and then doing optimization.

All is not lost. Here are some things to try:

1. Redesign your frequency scheme to use commercially made filter modules.

2. If that cannot be done, use the top coupled resonator topology. This will absorb the parasitic capacitances which will change the center frequency but not change the shape..

3. Use coplanar waveguide on your PC card. This is where the signal traces and ground area are on thee same side of the board and there is no metal under the filter. This will reduce the parasitic capacitances and also the inductance of vias (which will not be used.)

4. Keep the coils far apart so that there is no mutual inductance between them. This can be done by alternating them on either side of the straight through signal trace.

Thanks guys!

By the way, im using ADS. Attached is the paper i got from ieee on a kind of BPF. Im thinking of selecting an existing microstrip BPF design then tunning it my desired 1.5GHz by changing its length and width.

I wonder how many types of BPF approach are there?

I had altered the BPF design from 4.4GHz to 1.5GHz.
The schematic and layout is attached, can anyone point out some practical considerations beforei send this design for fabrication? Thanks!

Any responses? Thank you!