Is it possible to design a band-pass filter filter (on a substrate with dielectric constant of 4.5) of passband below 1 GHz using distributed components. If yes, are there any special procedures to do it?
Now depends how below 1GHz you want to go.
Between 500MHz and 1GHz the dimensions of the distributed elements on FR4 are large, but not impossible to make. You can try a Hairpin topology.
Now depends how below 1GHz you want to go.
Between 500MHz and 1GHz the dimensions of the distributed elements on FR4 are large, but not impossible to make. You can try a Hairpin topology.
Hairpin filters are generally good choice, but from what I see they are narrow-band below 1 GHz.
What I am designing is a band-pass filter from 300 to 800 MHz. I've succeeded to design it using different impedance lines and stubs, but at the cost of the relatively large size, as you have mentioned.
I wonder if there is any other approach to minimize the size.
You can do it, if you use very high dielectric constant substrates (er>38 )
You can do a quasi distributed, where the inductors are made up of high impedance tranmission lines, or spiral coils. The easier to procure capacitors are either lumped element chips, or interdigital fingers.
Typical coupled line filters, lumped or distributed, would need very high inter resonator coupling values for such large bandwidths which would make interdigital fingers hard to realize. You might have to switch to lumped surface mount components then.
Another approach to this problem is to look at your wide bandpass filter as a combination of a low-pass and high pass filter connected in cascade. This might actually be easier to implement.