Increasing band width or band by using saw filter in parallel

[adnan012]

hi,

Suppose that there is a saw filter A with center frequency 105MHz (bandwidth 10 MHz) , second saw filter B with center frequency 110 MHz
(bandwidth 10 MHz). Is it possible to use both saw filter in parallel configuration to use the bandwidth of both filter. Is there any ADS model for saw filter available.

 

[biff44]

at the edge where the two filters overlap you will have a lot of amplitude and phase distortion, so if your modulation stradles that region it will not work. If you have lots of narrowband channels, and just throw away the few that are in the overlap, it could work.

 

[adnan012]

thanks for reply.

Do i need additional impedance matching for parallel configuration.?

 

[biff44]

why bother? It is going to have horrible vswr. Put a 3 dB pad on either side of the two filters and you can probably live with it.

 

[adnan012]

why 3 db pad (attenuator) is required and how i can implement it for saw filter.

 

[biff44]

in general, if you are making a system, you want to have components that are all matched to the same impedance (usually 50 ohms). That way you can add in components without creating mismatches. Impedance missmatches cause gain and phase ripples that can make a system not work. So ways you can add a poor quality component into a system without causing these ripples are:
1) Pad down both sides. a 3db pad on both sides guarantees better than a 6 dB return loss.
2) Add ferrite isolators on both sides
3) Do an impedance match network (IF the impedances are well behaived--i.e. SAW filters are not well behaved)
4) live with the ripples if your system is tolerant of such things

 

[adnan012]

i am still confused about 3db pad , is it a resistive pi or t network?
What is a ferrite isolator? Does it come in SMD package.

 

[E Kafeman]

If possible, select two SAW filters that not is overlapping.
A SAW filter have often rather low impedance outside of its pass band. Paralleling two filters with slightly different frequency range and they will shortcut each other. But we can use that as an advantage.
If both these filters have low real and low imaginary impedance in each others pass band is it possible to get a quite good 50 Ohm match and with low loss by a impedance matching network over whole frequency range covered by two SAW filters.
From each SAW, connect it to a network resulting in 45 degree delay at 50 Ohm. Each network can be done with just a inductor and a cap.
Connect these both network with each other, they will be almost blind for each other due to the total of delay is 90 degree.
A additional network is connected to same spot as the both other networks are interconnected, for impedance correction to a single 50 Ohm output.
A alternative method is to use coaxial cables instead of L/C network.

Two SAW's with overlapping bands, phasing part can maybe be solved for a part of the frequency range they have in common.
Worse is that one SAW have a step impedance/phase variation in the middle of pass band for the other SAW.
Cheapest way to reduce that effect is as biff44 says, two 3 dB resistor network. You will still get a very nonlinear frequency and impedance response, but variations will be limited so response not becomes absolute zero in parts of the total pass band.
About the resistor network design, think about it as a 3-way power divider.