Dieletric resonators filters

top c coupled

Any good resources about designing filters using Dielectric resonators

You can get some general guide lines on
http://www.trans-techinc.com/

But, let me know if you have specific questions.

NandoPG

Hi everyone,

I was wondering how do you insert a dielectric resonator between 2 50Ohm stages without shifting the frequency? As I understand, not to kill the resonator Q (and thus the filter Q), you would need very small capacitors (<0.25pF at 2Ghz for example) which do not exist or have poor spec?

Hi Ghost,
You are right, the caps should be small and they can be simulated easily.
As to caps ranging from 0.1pF to 100pF you have the expellant caps ATC series 600S. I have been characterized this caps and they are really outstanding. Take a look on the ATC web site.

NandoPG

Models are avilable for coaxial resonators in MWoffice.
With proper pcb layout we can achive results same as the simulated.
We have done that.
thanks

Ansoft CAD, both Serenade8.7 and Designer1.0 have a built-in model for DR coupled with one or two microstrip lines, including spacer. I think it is quite accurate. You can design your filters directly i.e. by inputing physical dimensions, not any conversions are needed from lumped equivalents etc.

flyhigh

First of all, any RF program is able to simulate these filters, not only MWO. For filters above 3GHz lots of considerations like grounding of the resonators and box effect start to came into the scene.

One question to Gladiator: for what frequency range would you be interested to get your filters?

NandoPG

thanx everyone,

nandopg, what i do not like about this solution is that the capacitors cause a shift of the filters resonant frequency. Did you already realize such a filter?

Hi

I have the impression that you are trying to bult a DR filter and have the desired response without tuning, so you are worried about frequency shift of resonant frequency when DR is coupled to external circuit. Well, regardless the caps present or not, there is no way that you can make DR filter (I am talking about cilindrical DR coupled to microstrip) without having some means of tuning it. DRs have large Q, and are very selective, chances to have desired response straight away are minimal! Beside any modeling and CAD simulation, mounting and tolearance issue allways require you to have some tuning available. Even if you have only one DR to couple, you will need tunning, not to mention filters having more DRs. A screw (metal, to increase frequency when screwing it in, or dielectric to decrease the resonant frequency) above the DR is the most popular and convenient way to achieve a tuning possibility.

I made some DR filters in the pass myself, have some experience on this.

flyhigh

Dear Ghost Tweaker
Maybe I have misunderstood this post:
What I am talking about is bandpass filters made up by rectangular ceramic resonators operating below 1/4 wavelength. If this is the case, yes I already built several of these filters and if you have specific questions just post them here.

If you are talking about filters made up by cylindrical DR, some time ago I characterized some cylinders but I didn't get the proper thermal stability (-40 to+85C) with them. Because of this I quit this approach and didn't build filters using cylindrical DRs.

NandoPG

dear Nandopg,

you were right, I'm talking about 1/4wavelength rectangle dielectric resonators. My question is now how do u proceed when the resonator is already at the correct frequency (passband of your filter) and you still need capacitors to keep a correct Q when the filter is inserted between two 50Ohm loads. In my case I have a resonator at 1400MHz with a Q of 400. My filter has to have 30Mhz bandwidth around 1400MHz (Q~50).

Dear Ghost Tweaker,
First of all the self parallel resonance of a resonator to be used in a bandpass filter should be higher than the frequency where they are going to be used. Doing this way the resonators will act like a high Q inductor. As a rule of thumb, to start you should choose resonators with the self parallel resonance 15% higher than the center frequency of your filter.
When you couple two resonators through a capacitor, what you are doing is to reflect a small portion of the coupling cap in parallel with the resonator in addition to create a mutual impedance between the resonators. This fact displace the resonance points defined by each parallel circuit creating thereby a passband.
During the filter calculation and simulation it is true that you will have to correct a little bit the self resonance of the resonator to get flatness and insertion loss, but you have total freedom to order (from the vendor) the resonator cut the length you need, not to say that the ceramic resonator allow you in the real life a small amount of tuning, so that the filter can be optimized in production.
For all of this a coupling cap doesn't introduce any kind of prejudice to a ceramic resonator filter by the contrary, they create the proper bandwidth and decrease the insertion loss.

NandoPG

The filter topology is top coupled resonator filter. This was done in the past with LC networks as the resonators. Several low cost filter design programs have this topology. The problem with using them is finding the effective L and C values of the ceramic loaded transmission lines.

If you have an analysis program that has an optimizer function, you can enter your schematic and let the program find the right coupling capacitor values. At your frequency they will be 1 pF or smaller.

Commercially made filters have these capacitors made with copper on opposite sides of the printed circuit board.

www.aade.com has this topology in their filter program. You can use the program for a few days for free.

Just to point out that my previous posts in this forum are a bit out of focus as it was not clear to me that topics are about DRs of rectangle dielectric resonators, so apologies to everybody, I didn't understand it from the beginning.

flyhigh

The synthesis method of bandpass filter implemented with ceramic resonators defers from the method used to synthesize the same topology with lumped coils in the sense that the resonators work as an inductor that is a function of the frequency.
Be aware that the outcome from programs supporting top C coupled resonators bandpass topology is just a rough first guess to get the filter values to be implemented using ceramic resonators.

NandoPG

What is the range of Dieletric resonators filters?
Can it work as a BPF to meet these requirments:
passband 8-12G IL<-3dB,ripple<2dB
stopband 16-24G attenuation>40dBc

Thanks for your response!

Dear STD

The useful range for ceramic resonators filters is from about 300MHz to less than 6GHz. There is no theoretical constraints about bandwidth although for narrowband filters above 1GHz a very low coupling capacitor is required what makes the implementation of the filter sometimes impossible.

NandoPG