Low Cost PCB's Low Cost Components

Author Topic: Microstrip two port network design tips  (Read 557 times)

0 Members and 1 Guest are viewing this topic.

Offline Cacao

  • Contributor
  • Posts: 26
  • Country: ph
Microstrip two port network design tips
« on: August 06, 2017, 09:03:06 PM »
Can anyone give me advice on designing two port network, here is the plan: https://ibb.co/jCHVev
Operating frequency: 2-3GHz

Thank you guys!
« Last Edit: August 06, 2017, 11:43:40 PM by Cacao »
 

Offline Howardlong

  • Super Contributor
  • ***
  • Posts: 3741
  • Country: gb
Re: Microstrip two port network design tips
« Reply #1 on: August 06, 2017, 10:04:51 PM »
That's a directional coupler, which is either a three- or four-port device.

In your case it's configured as a three port, the "fourth port" goes directly to a dummy load.

 
The following users thanked this post: Cacao

Offline Cacao

  • Contributor
  • Posts: 26
  • Country: ph
Re: Microstrip two port network design tips
« Reply #2 on: August 06, 2017, 11:44:49 PM »
That's a directional coupler, which is either a three- or four-port device.

In your case it's configured as a three port, the "fourth port" goes directly to a dummy load.

Can you give me some things to consider when designing it?
 

Offline evb149

  • Super Contributor
  • ***
  • Posts: 1568
  • Country: us
Re: Microstrip two port network design tips
« Reply #3 on: August 07, 2017, 05:37:29 AM »
Well you say you want to use microstrip so that probably means that you're working with 1-10 GHz range signals otherwise microstrip on a PCB is probably not of practically good dimensions / characteristics.

There are some (several related topics of different titles but all relating to directional coupling in some sort) articles on this blog:
http://k6jca.blogspot.com/2015/01/notes-on-directional-couplers-for-hf.html

There are also several good introductory articles with some useful theoretical explanations at the following site among several related linked pages.  There are also possibly some spreadsheets / tools on the site that are related, I'm not sure.
https://www.microwaves101.com/encyclopedias/coupled-line-couplers

You should also look at the free software called ATLC and MMTL and maybe MEEP since they can help you calculate / model the transmission line impedances for even and odd modes and simulate the coupling behavior.

The problems you face will include:
* Microstrip on cheap FR4 types of PCB is lossy so you will have to account for the loss particularly if you're operating much above about 3 GHz.
* Microstrip by definition is assymetric with air covering the line on top and solid dielectric underneath.  If you use solder mask over the microstrip top it will affect the performance in ways that can be hard to predict.
* PCB plating and etching makes microstrips have some kind of trapezoidal cross sectional strip profile rather than the ideal rectangle.  So particularly if you use narrow lines and particularly also if you use edge coupling (which is pretty much implied by a microstrip coupler) then the trace edge angles, solder mask thickness profile vs. edge profile, and trace width / etching / spacing variation due to PCB manufacturing processes will affect both the edge coupling constant and the impedance.
* FR4 types of  low cost PCB material can vary significantly in dielectric constant, and loss vs. frequency.  There is even an effect where the dielectric constant is not uniform depending on the trace angle across the woven fiberglass matting of the PCB since the FR4 is woven in a somewhat coarse X-Y square mesh pattern.  Prepreg layers are used on top of a core for multi-layer PCBs whereas two layer PCBs might use only a double sided copper clad core.  The core will have better dimensional predictability / stability than prepreg since prepreg deforms more during lamination whereas a single core only 2-L PCB doesn't need lamination and doesn't deform much at the core anyway.  So using a well specified type of 2L core only PCB will help you get the dielectric constant, loss, and microstrip height parameter data more accurately known but still there are uncertainties due to etching / plating and process variations.  So for better control / predictability you could specifiy an impedance controlled manufacturing requirement for the PCB.  You could also specify the use of higher RF performance non-FR4 based PCB materials depending if you need a high performance coupler and if you need to use multi-GHz frequencies.
* You would find that using broadside coupled stripline might yield a more predictable design but that would require a controlled stackup and maybe impedance controlled PCB fabrication with at least 4 copper layers so it would be more costly than microstrip.
* There are "off the shelf" directional couplers available for sometimes quite low costs that would be smaller than and less of a difficulty in engineering than using a custom made microstrip based design.  Some DCs are even basically monolithically integrated as thin/thick film ceramic hybrid or similar SMT type PCB component modules.  Minicircuits, Johanson, Anaren and such might be places to look.  As that first web site I mention shows as well as others you can make either lumped element or coupled line based DCs yourself somewhat simply.  The coupled line designs could be made with microstrip or stripline but it is also possible to use coaxial cable, pipe, or other kinds of coupled transmission line structures to achieve the result.  Microstrip of FR4 is not a particularly high performance and compact type of design compared to some other options.

* Your diagram doesn't show anything about the actual operating parameters:
+ Center frequency
+ Bandwidth
+ Coupling ratio between the direct and coupled ports.
+ Return loss vs. frequency requirements
...so to get anywhere with selecting a design you must specify the performance.

« Last Edit: August 07, 2017, 08:43:19 AM by evb149 »
 
The following users thanked this post: Cacao

Offline hendorog

  • Frequent Contributor
  • **
  • Posts: 791
  • Country: nz
Re: Microstrip two port network design tips
« Reply #4 on: August 07, 2017, 06:20:04 AM »
It is probably simpler to get good results with an RF directional bridge instead of a coupler for a beginner.
Assuming you don't need to handle high power.

Good info on both approaches can be found in this impressive VNA project:
Coupler:
http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html

Bridge:
http://hforsten.com/improved-homemade-vna.html
 
The following users thanked this post: Cacao

Offline BigBoss

  • Contributor
  • Posts: 49
  • Country: fr
Re: Microstrip two port network design tips
« Reply #5 on: August 07, 2017, 06:35:07 AM »
Can anyone give me advice on designing two port network, here is the plan: https://ibb.co/jCHVev
Operating frequency: 2-3GHz

Thank you guys!

If you don't have a possibility to reach a EM simulator, I'd recommend you to use commercial products such as
http://www.yantel-corp.com/en/products/datasheel/DC/DC2500P02.pdf

There are also other different manufacturers.The component that you're looking for "Directional Coupler".
It's a 4 port device and forward and reflected power can easily be measured.
 
The following users thanked this post: Cacao

Offline Cacao

  • Contributor
  • Posts: 26
  • Country: ph
Re: Microstrip two port network design tips
« Reply #6 on: August 10, 2017, 12:48:12 AM »
Do you guys know how can I change the operating frequency of a wilkinson power divider such as this into 2.5GHz? https://www.cst.com/academia/examples/wilkinson-power-divider
« Last Edit: August 10, 2017, 01:12:48 AM by Cacao »
 

Offline Cacao

  • Contributor
  • Posts: 26
  • Country: ph
Re: Microstrip two port network design tips
« Reply #7 on: August 10, 2017, 01:52:09 AM »
Here is the function of this network. As I supply a microwave signal (2.5GHz) into one of its ports, that signal will go to the probe that is attached to some material. The reflected signal from the probe then will be detected by the rf detector. Do you think a power divider or a directional coupler can be used?
 

Offline evb149

  • Super Contributor
  • ***
  • Posts: 1568
  • Country: us
Re: Microstrip two port network design tips
« Reply #8 on: August 10, 2017, 01:02:06 PM »
A directional coupler would be applicable, a power divider would not.

A directional coupler is a RF device that lets you direct power from a signal source to a load and direct the reflected signal from that load to one other terminated port on the device.

At DC or AC (which could be low frequencies, not just RF) one could also measure the impedance of a two port network just as an ohm-meter measures the resistance of a two terminal resistor.  Impedance bridge circuits are one way to measure an unknown impedance given a signal source and some known impedances.  So if you want to measure the scalar or vector impedance of a network you can do that with some topologies of scalar or vector network analysis and that will tell you the characteristics of resistance and reactance of your "probe" two port network.  Whether it is most useful to you to measure the probe's scalar / vector impedance or just measure the magnitude of the reflected signal is sort of up to you.  Some of the below publications discuss different topologies for RF measurements and network analysis and signal measurement.  Different measurement configurations are more applicable depending on the nature of the measurement to be made, the impedances involved, the power involved, the frequency involved, the required accuracy and precision, etc.


These may be interesting for you:

Impedance measurement handbook (5950-3000) - Keysight
http://literature.cdn.keysight.com/litweb/pdf/5950-3000.pdf

Agilent AN 1287-1 Understanding the Fundamental Principles of ...
http://care.iitd.ac.in/People/Faculty/bspanwar/crl721/VNA_Agilent%20AN%201287-1.pdf

Agilent AN 1287-2 Exploring the Architectures of Network Analyzers
http://anlage.umd.edu/Microwave%20Measurements%20for%20Personal%20Web%20Site/Architecture%20of%20NWAs%205965-7708E.pdf

Microwave Network Analysis Application Notes
http://web.doe.carleton.ca/~nagui/Appnotes/appnotes.html

RF Basics, RF for Non-RF Engineers - Texas Instruments
http://www.ti.com/lit/ml/slap127/slap127.pdf

Course 1 RF Measurements Techniques Instructions - CERN Indico
https://indico.cern.ch/event/115334/sessions/5217/attachments/50210/72213/C1_Piotr_Ex_all.pdf

RF measurements I: signal receiving techniques - CERN Indico
https://indico.cern.ch/event/216963/session/1/attachments/347577/484628/C1-Measurement1_5.4.11.pdf

RF engineering basic concepts: S-parameters
https://cds.cern.ch/record/1415639/files/p67.pdf


Here is the function of this network. As I supply a microwave signal (2.5GHz) into one of its ports, that signal will go to the probe that is attached to some material. The reflected signal from the probe then will be detected by the rf detector. Do you think a power divider or a directional coupler can be used?
 
The following users thanked this post: Cacao


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf