Design of 13.7-14.55GHz Band Pass filter

[srlraghuls]

I want to design a Microstrip BPF with Pass Band 13.7-14.55Ghz. I have selected selected RT/Duroid 5880 Laminate with dielectric constant 2.2, Which Dielectric Height I should go 10mils or 62mils?
In Nuhertz Filter I have generated the Interdigital ground coupled Layout of BPF and while simulating in ADS the result i got is not even 50% matching with AWR Nuhertz Simulation.
Can you please suggest some document or way to design a 13.7-14.55Ghz Microstrip BPF.

 

[vicky001.iith]

Read this
**broken link removed**

Hope it will help you

 

[Darktrax]

The 5880 substrate is PTFE with a thin glass matrix. It is soft, and to have grounding vias, requires a special sodium etch before plating.

The loss tangent is 0.0009 at 10 GHz. There are many other RT Duroids that use part ceramic loading, still have low loss, are dimensionally stable, and can be etched and plated with conventional PCB techniques. There is also Taconic RF-35A. The strip losses for filters are often more affected by the suface roughness under the strip than simply going for the lowest value tan(delta) loss. Rolled versions are better.

For an edge coupled micristrip filter for which there are handy examples, the gaps can get very thin. Choose a substrate thickness that is more forgiving for this parameter. The very thinnest are not suitable for filters. Lines get very thin.

It is hard to tell why the simulated filter does not agree well with the first speculated design, but it will have something to do with differences in assumptions about the substrate properties. Taking an example, and scaling it for frequency may be a better approach. Keep in mind this means scale the resonant lengths. To retain the line impedances, you have to leave the widths, and the substrate alone, unless you know what you want, and alter those deliberately.

 

[biff44]

that is a pretty high frequency for a duroid substrate. Due the the mix of Teflon and fiberglass weave, it is hard to know the exact dielectric constant. Alumina would be a better choice. but if u do use the duroid, leave a guard band on either side for variations in er/dimensions

 

[tony_lth]

Hi, I used to do a lot of microstrip filter with RO4350B 0.2oz 10mil substrate, even to 26GHz, your filter bandwidth is not very wide, it should be easy to do.
You should use edge coupled line filter.

 

[srlraghuls]

Thank You all for suggestions.

While simulation in Layout momentum in ADS the mesh density default is 20 and arc resolution in 45 default. Which mesh density and arc resolution i should keep to get PCB tested result and designed result to be matched

 

[tony_lth]

I suggest you use AWR to simulate it, or ADS SCH to do so, without considering the mesh issue.
Then you can simulate it in ADS for layout.