What is a good unloaded Q factor in an RF filter?

[Hi-Tone]

Hi everyone.

I have a 4 pole combline cavity filter in the band 1710-1755 MHz.

As I have an insertion loss of -0.16 dB and calculated my loaded QL = 38.5 and my unloaded Q0 = 2109.

1) A Q is calculated for each resonator (each pole), but can I say that the filter has a Q as a whole unit?

2) Is there a general rule to say that a Q above a 1000 is considered good?

Thank you so much,
Steve

 

[Phytech]

I doubt you can obtain insertion loss of 0.16 dB on a 4-pole comline filter in the frequency you specified, it would require unloaded Q of more than 8000 to meet your insertion loss.

The loaded Q you calculated is just the revers of the percentage bandwidth of the filter. When people say a filter, they normally say the center frequency and bandwidth (or percentage bandwidth). If people say high Q resonator filter, it usually refers to the resonator unloaded Q.

There is no general rule to say Q0 above 1000 is good or not good. It all depends on your application and technology. For LC design, you will never get a Q0 of 1000. Actually if you can get 180, it is very good. For ceramic filters, 1000 Q is extremely good. For combline filters, 1000 Q is just below average. And for dielectric resonator loaded cavity filters, 1000 Q is very bad as you can easily get Q of more than 10000.

Good luck!

 

[Hi-Tone]

My filter (measured from -3dB) goes from 1685 - 1792. Between 1710 - 1755 i have IL = 0.16dB. This is measured on several network analyzers and a noise figure meter.

Is there something im doing wrong in my calculations?

 

[Phytech]

My filter (measured from -3dB) goes from 1685 - 1792. Between 1710 - 1755 i have IL = 0.16dB. This is measured on several network analyzers and a noise figure meter.

Is there something im doing wrong in my calculations?

This sounds reasonable, the unloaded Qs for the resonators are still very high (more than 3000), but it is obtainable for combline filters. Now you cannot claim the loaded Q is 38.5, you may claim the unloaded Q as the center frequency divided by the 3-dB bandwidth, which some people call this the loaded Q of the filter.

Actually, the important Qs for a coupled resonator filter are: 1. unloaded Q of the resonator, 2. the coupling Q of the input and output, which measure how strongly you couple the first resonator to the input port, and the last resonator to the output port. In order to make the filter work, you need to make the coupling Q right. The coupling Q is not only a function of the percentage bandwidth, but also a function of the passband return loss.