Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

how to using ADS or other software to design a 3-dB coupler?

Status
Not open for further replies.

Needbias

Advanced Member level 4
Joined
Mar 3, 2002
Messages
103
Helped
2
Reputation
4
Reaction score
2
Trophy points
1,298
Activity points
756
In this forum,I think there have many expert on coupler design.but how to just use software, such as ads2002 or microwave office to design a 3-dB stripline coupler?I use mutilayer model of ads2002,but the result seems not very good.wish expert give me some advice.
Thx.
Best regards,
Needbias
 

cwcwecan

Junior Member level 3
Joined
Apr 14, 2002
Messages
27
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
126
What is the frequency of your stripline 3dB coupler?
And What is the Substrate you use?
3dB is a Strong Coupling
So you should design it carefully
 

Element_115

Advanced Member level 2
Joined
Mar 23, 2002
Messages
509
Helped
72
Reputation
144
Reaction score
43
Trophy points
1,308
Activity points
3,767
Are you making a Power spliter? /\ D S has a wilkinson power spliter
example in their design kit. I built a (3dB)Branch line coupler in /\DS, using Monentum but the results where not very accurate. A few things to remember
are the wave length in the substrate you are using and the trace width for
50 Ohms at your center frequency. You can find many app notes on the web.

Hope this helps
 

flatulent

Advanced Member level 5
Joined
Jul 19, 2002
Messages
4,629
Helped
489
Reputation
980
Reaction score
150
Trophy points
1,343
Location
Middle Earth
Activity points
46,689
paper and pencil

The wilkins and branch line couplers are so well understood that you should be able to design them with paper and pencil and use the software to verify the design. You can do a frequency sweep to see where the design works best and then scale the lenghts to move the best performance part to the frequency you want.

The main errors you will probably have to look out for as mentioned above is the impedance of the lines and the wavelength. Rogers corp has a free calculator for determining the impedance and velocity at a given frequeny.

Here is a copy of their program.
 

Needbias

Advanced Member level 4
Joined
Mar 3, 2002
Messages
103
Helped
2
Reputation
4
Reaction score
2
Trophy points
1,298
Activity points
756
Thx so much for your all help.
For my design,the bandwidth is from 2GHz~6GHz,I want use two 8.34-dB coupler to form a tandem 3-dB coupler,so the branch-line coupler sames not appropriate for this design.Do you have any advice?
Thx.
bgz

Needbias
 

flatulent

Advanced Member level 5
Joined
Jul 19, 2002
Messages
4,629
Helped
489
Reputation
980
Reaction score
150
Trophy points
1,343
Location
Middle Earth
Activity points
46,689
Lange coupler

The Lange coupler is the type which will give you these bandwdiths.
 

toonafishy

Full Member level 6
Joined
Jun 18, 2002
Messages
369
Helped
38
Reputation
76
Reaction score
12
Trophy points
1,298
Location
Earth
Activity points
3,187
You might also be able to use a two or three section wilkinson in series using stepped impedance. This will give you better bandwidth. I did a two section one that covered 1-2GHz. That's 100% bandwidth. If you have @DS, look in @DS\examples\training\*omentum\mom_class_prj and you will see a nice single section splitter.
 

pll

Junior Member level 2
Joined
Jul 15, 2002
Messages
22
Helped
2
Reputation
4
Reaction score
2
Trophy points
1,283
Activity points
107
Hi

I am seeing some printed balun with ADS (schematic and momentum) and I see that the simulation on microstrip balun (rat race or Marchand) are very good (I measured the s-parameter of these baluns in some boards and the results are very closed to the simulation). Now I am trying to explore new tipology in CPW-slotline structure in ADS in order to reduce the dimensions of these balun, but I have a lot of problem because the results are differents from the article where I take this tipology (I think to simulate the CPW-slotline structure I need a three dimension simulator). Do you know other tecnique (balun over two or three layers for istance) to reduce the balun dimension?

Regards
 

toonafishy

Full Member level 6
Joined
Jun 18, 2002
Messages
369
Helped
38
Reputation
76
Reaction score
12
Trophy points
1,298
Location
Earth
Activity points
3,187
If your design is based on edge couplers, then you can create multiple fingers. This will make a very small balun. But if I remember correctly, the output ports are 90degrees, not 180 so if you need 180, you will need an additional 90 degree external phase shift. What are you tryng to do?
 

Needbias

Advanced Member level 4
Joined
Mar 3, 2002
Messages
103
Helped
2
Reputation
4
Reaction score
2
Trophy points
1,298
Activity points
756
Thx,my design is based on broside couple, and I want the phase error must less than 90+/- 2 degree,amplitude error less than 0.5 dB at 2GHz~6GHz,so I think lange is not a good choice.
 

pll

Junior Member level 2
Joined
Jul 15, 2002
Messages
22
Helped
2
Reputation
4
Reaction score
2
Trophy points
1,283
Activity points
107
HI toonafishy

I am trying to design a balun for the W-LAN 802.11a standard (5 to 6 GHz). I designed a rat race for this application with very good results, but the dimensions are good for research and not for an application board. So I am exploring new structure to design a wideband balun but with little dimensions. I am trying with CPW-slotline structure but I can't simulate it correctly with Momentum of ADS: is it for my mistake? Do you know other structures very low in dimension but with a large bandwidth?

Thanks
 

Lupin

Full Member level 6
Joined
Jan 22, 2003
Messages
342
Helped
12
Reputation
24
Reaction score
9
Trophy points
1,298
Location
Italy
Activity points
3,235
Just one question for coupler expert people,

A branch-line coupler is a 4 port component. The output ports (O1 and O2) are loaded with 2 equal loads at -15 dB each one...
The question is: how much is the return loss at the input and at the other port?
I think that theorically I can have -
 

Lupin

Full Member level 6
Joined
Jan 22, 2003
Messages
342
Helped
12
Reputation
24
Reaction score
9
Trophy points
1,298
Location
Italy
Activity points
3,235
(sorry.. some problem occured..) I was saying:

-infinite dB at input port and -9 dB at the other port.
What do you think?

Regards
Lupin
 

jamez

Member level 3
Joined
Jan 15, 2002
Messages
56
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Activity points
370
Hi Needbias,

I have make a 4-8Ghz 3dB coupler with tow five step coupler,
in my view, the step uncontinues is very important. The ads and mw not consider their step uncontinues and coupling in same time.
So, you can try it in ads momental or ansoft ensemble.
finally,I get -2.5 to -3.5 dB coupler in full band.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top