Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics 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.

How to match broad bandwidth using Smith Chart?

imtiaz369

Member level 2
Member level 2
Joined
Jul 8, 2021
Messages
53
Helped
0
Reputation
0
Reaction score
0
Trophy points
6
Activity points
568
Hello, I have an S-parameter, and I imported it into ADS. Generally, it covers from 3.4-8GHz. But I want to make it from 2-8 GHz, so I need a matching circuit. I did a simple pi matching for 3.00GHz, and it matched at 3.00GHz, but the other frequency went above -10dB. How can I do a broadband impedance-matching circuit? (Attached image for reference - the top picture is before matching, and the second one is after matching)

Any guideline, tutorial or reference! Highly appreciate it!

Screenshot 2025-01-22 110338.png
 
You cannot match an impedance over this frequency band with a simple PI Matching Network.
Wideband Impedance Matching is a serious and troublesome problem in RF. There are some numerical techniques such as Real Frequency.
A practical method is to use few cascaded PI or T matching networks that are resonated at different frequencies with lack of efficiency.
 
You cannot match an impedance over this frequency band with a simple PI Matching Network.
Wideband Impedance Matching is a serious and troublesome problem in RF. There are some numerical techniques such as Real Frequency.
A practical method is to use few cascaded PI or T matching networks that are resonated at different frequencies with lack of efficiency.
Is there any tutorial or guideline with those examples?
 
There are youtube videos using ADS about broadband design that could lead you somewhere, or perhaps nowhere.
 
Last edited:
I am trying to improve the bandwidth. At present, the S-parameter shows a bandwidth of 3.3–8.6 GHz. I am trying is to design a matching circuit to extend the bandwidth from 2.4–8.6 GHz.

To provide some context, the S-Parameter matched for 3.3–8.6 GHz. But I have a few questions regarding the design process:
  1. When designing a matching network, which frequency should I use as the target for matching?
  2. In broadband matching, adjusting the impedance at one frequency often impacts the impedance at other frequencies. How can I effectively handle this?
  3. I have access to ADS for design, but is there a specific Smith chart program available that can help track impedance movement across frequencies?
  4. Could you please share any recommendations or strategies to achieve this bandwidth extension?
Your insights and suggestions would be greatly appreciated. Thank you in advance for your time.

impedance matching.png
 
which frequency should I use as the target for matching?
This is where your approach fails: you have a wide bandwidth and your data circles all around the Smith chart multiple times. The simple matching that you have in mind will not work here.

There is a great commercial tool for wideband matching, targeting antennas: Optennilab.
I had used that tool for automated wideband matching. It will find matching circuits with different numbers of elements, evaluating all alternatives automactically, and give you the best circuit with lowest insertion loss (it doesn't help to have good S11 if your matching network itself is very lossy).

But I am afraid even that best-in-class tool will not be able to solve your matching challenge.
 
This is where your approach fails: you have a wide bandwidth and your data circles all around the Smith chart multiple times. The simple matching that you have in mind will not work here.

There is a great commercial tool for wideband matching, targeting antennas: Optennilab.
I had used that tool for automated wideband matching. It will find matching circuits with different numbers of elements, evaluating all alternatives automactically, and give you the best circuit with lowest insertion loss (it doesn't help to have good S11 if your matching network itself is very lossy).

But I am afraid even that best-in-class tool will not be able to solve your matching challenge.
That's premium software. Do you have any idea how I can start manually? Or, if it is not possible, how can I show that thing?
 
For me, the logic behind Smith chart maching is that you move impedance points from one location to another, but your starting points are all across the chart. I don't see how you could identify any useful "move" (i.e. topology) here by staring at the data.

The commercial software does a lot of trial and error with hundreds of different topologies, in addition to calculating component values. Here is another software package, they have a demo version and maybe you can try that: https://www.antune.net/
 
Your 3.5GHz to 8GHz matching network is not too far from the target 2GHz to 8GHz.
I think just adding another Pi network to the existing network you will get the target.
ADS have its own impedance matching and optimization options:

 
Hello, I have an S-parameter, and I imported it into ADS. Generally, it covers from 3.4-8GHz. But I want to make it from 2-8 GHz, so I need a matching circuit. I did a simple pi matching for 3.00GHz, and it matched at 3.00GHz, but the other frequency went above -10dB. How can I do a broadband impedance-matching circuit? (Attached image for reference - the top picture is before matching, and the second one is after matching)

Any guideline, tutorial or reference! Highly appreciate it!

View attachment 196863
you could try broad band matching synthesis. This involves designing a wide band bandpass or low pass filter with end elements that replicate the impedance of the load you wish to match.

1) plot s11 of you complex load, (no matching network)

2) depending on shape of s11 plot, choose an appropriate rlc network topology that can transverse the smith chart the same way.

3) fit the rlc values so s11 & the rlc s11 overlap. you now have an rlc network that represents the load.

4) you now want to design the filter network. in filter design you start with normalized lowpass prototype and then transform from low pass to bandpass or remain low pass but shift up in frequency. You need to perform the reverse transformation... transform your rlc to the low pass domain.

5) you will now have a normalized low pass equivalent to your high frequency rlc.

6) now synthesize a normalized low pass filter such that the normalized load is the same as your normalized load... maybe your normalized load is a RC in parallel. You can try a chebyshev normalized low pass. usually you start with rs and rl and synthesize the network. it is possible instead to start with the r&c at one end instead and synth the network from there. The equations are not closed form and there is iteration required to solve. To much to replicate here.

7) one you have the normalized filter network, you then transform back up in frequency (low pass or band pass). At this point you will have a network that contains the rlc you fitted to your s11.

8) strip off the rlc... the remaining circuit is your matching network for your complex load.

Although not the full solution for you, this is a procedure for designing robust broadband matching network based on reliable filter theory. It is not trial and error.

Hope this gives some insight
 


Write your reply...

LaTeX Commands Quick-Menu:

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top