Elliptic Bandpass Filter - Lumped to Distributed Conversion
I have designed a simple 5th order elliptic bandpass filters. I'm using lumped components, however, I would like to convert these lumped elements into distributed elements. I'm having trouble doing that. I've tried using kuroda's identities without success. I included a picture of the design using lumped components in ADS. This is giving me a hard time as I need to eventually create on this using microstrip lines.
To design my elliptic bandpass filter, I converted the lowpass design into bandpass design. My filter is working for fl=1ghz, fh=2ghz.
Any help is much appreciated.
Distributed Filters are not designed in this way.Converting Lumped Elements onto Distributed ones wouldn't work as you expected.
First, 1..2GHz frequency range is not very suitable for distributed filters unless you're able to use high di-electric substrate.Because large design dimensions due to large wavelength will result radiation and metallic losses and this will deteriorate the filter's response.In additional to, in case of coupled structures, insertion loss will be pretty high.
However, different distributed type filter such as coupled resonators.edge coupled filter,comb-line filter etc. can be used but I repeat here, filter response may be a bit disappointment.
I see. Lets say I change my design to a BP with Fc at 12.5ghz. I understand that I will need to use LC resonators for the design. I'm trying to use O.C stubs, however, the circuit is giving me problems because I can't convert the parallel branches using koruda's identities. Is there an easy way to breakdown the parallel branches for conversion?
I'm not too worried about filter response, as I'm doing this exercise for self learning.
If you change the center frequency, everything will also be changed.There are many types of filter for that frequency.I cannot explain how to design them here but you'd find a well reference for Distributed Filters.