Help on Chioce of dielectric height for Ku band mwave design

[syncmaster]

what dielectric height

Currently, I am working on a ku band module design. What dielectric height should I choose for Rogers RT/duroid 5880? 10mil or 20mil?Could anybody explain it.

Thank you!

 

[vitko]

dielectric substrate ku band

Hi, The of choice of microwave substrate thickness is based on mechanical, manufacturing and performance criteria.
I'm using RO4003 20 mil and RO3203 10 mil.
10 mil material is less rigid, it bends more easily, when large and/or multilayer PCBs are being produced the material size changes quite much and this leads to some changes in artwork dimensions.
The 50-O microstrip line is almost twice wider on 20 mil material. Some layouts may become unrealizible on the material with 10 mil thichness, others - on 20 mil material.
My advice is to calculate roughly the more critical dimensions on both materials, to have a talk with your PCB manufacturer and think about mechanical issues. Than you will have your chioce done

 

[lguancho]

Re: Help on Chioce of dielectric height for Ku band mwave de

10 mil 5880 substrate is OK. Usually, you do gold plating on it, reflow it to some metal backing (gold plated aluminium plate).

8)

 

[flyhigh]

Re: Help on Chioce of dielectric height for Ku band mwave de

Hi,

just to add one more detail. If you have to make a lot of vias, especially if you have to ground the sources of the transistors, than the via can be an important issue to adress if it is critical to achieve enough gain or precise oscillating condition. In that case, vias will be more closer to short circuits in 10mils rather then on 20mils substrate.

However, sometimes it is advantagous to have some parasitics coming from via, e.g. it can stabilize the transistor to some extent.

flyhigh

 

[lguancho]

Re: Help on Chioce of dielectric height for Ku band mwave de

Hi, all. I come across the book " Microwave Solid state circuits and application" by Kai Chang (Wiley). The is a section on microstrip line and the optimum thickness determination for particular operating frequency.

To summarise, it boils down to achievable Q factor of 50 ohm lamda/2 resonator against particular thickness. There is an old paper"Maximum Q factor of Microstrip Resonators" IEEE MTTS Vol 29 Feb 1981 by A. Gopinath. There are some graphs which shows measured Q against thickness, for various dielectric constant and frequency.

For ku band (16GHz) application, the optimum thickness is about 0.2mm (10 mil), for Er=2.2


This a more theoretical way of determing a substrate thickness 8)