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.

loadpull in power amplifier

Status
Not open for further replies.

SA1234

Newbie level 3
Joined
Jun 3, 2019
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
61
hello everyone,

I’m a bit stuck in the concept of load pull…



I’m using ADS to model and tune a amplifier, my questions are:

(regardless of stability)

1- by loadpull, we simulate the effect of different loads on the performance of the system. Imagine I picked the load ZL which leads the amp to have the best performance.

As I assume I should put this as the output load, but what I observe in the examples is that starting from this point, a matching network to 50ohm is put in the output…

I agree to match to 50ohm, but the unclear part for me is that when we match to 50ohm, the amp sees 50ohm as the output so the performance is not what I expected??



2- why normaly left half of smith chart is used to simulate arbitrary load? why I have not seen the right portion in any example?



3- when I find a desirable load for fundamental harmonic (ZL)

I think in the ideal case for 2nd and 3rd harmonic, would be to get the max reflection coefficient (=1) (or Zout=inf???)

and most probably it would not be the case with that very ZL that I picked.

so what would be the approach here?
 

1. Loadpull determines the optimal load impedance for your PA. You then design a matching network that transforms this impedance to the actual load impedance, e.g. 50 ohms.

2. Left halve of Smith charts means |ZL| <= 50 ohms. Typically semiconductor PA are working with low voltage and high current, thus left half plane.

3. Complete reflection of harmonic power is usually achieved by an additional harmonic filter. A matching network with low pass characteristic can work as its first stage, but is typically not sufficient for required harmonic suppression.
 
  • Like
Reactions: SA1234

    SA1234

    Points: 2
    Helpful Answer Positive Rating
1. Loadpull determines the optimal load impedance for your PA. You then design a matching network that transforms this impedance to the actual load impedance, e.g. 50 ohms.

2. Left halve of Smith charts means |ZL| <= 50 ohms. Typically semiconductor PA are working with low voltage and high current, thus left half plane.

3. Complete reflection of harmonic power is usually achieved by an additional harmonic filter. A matching network with low pass characteristic can work as its first stage, but is typically not sufficient for required harmonic suppression.


Dear FvM thank you very much for responding me,

The last unclear point for me is that, is there any method to perform a matching network design that positively effects all the harmonics? I got your point regarding 3rd comment that you made, but considering many template in ADS that also simulate the fundamental & 2nd and 3rd harmonc freq, I wonder how to use them simultaneously to design the matching network. I mean how to proceed to optimise it.
I do appreciate If you guide me through this point.

Regards,
 

is there any method to perform a matching network design that positively effects all the harmonics?
Depends on which positive effects you want to achieve. For most applications, surely for radio transmitters, you want maximal attenuation of harmonics. That's basically regarded by selecting a low pass topology. But specific power stages, e.g. class C can have also requirements for the load impedance at harmonic frequency. Ultimately you place LC series circuits resonating at the harmonics.
 
  • Like
Reactions: SA1234

    SA1234

    Points: 2
    Helpful Answer Positive Rating
Depends on which positive effects you want to achieve. For most applications, surely for radio transmitters, you want maximal attenuation of harmonics. That's basically regarded by selecting a low pass topology. But specific power stages, e.g. class C can have also requirements for the load impedance at harmonic frequency. Ultimately you place LC series circuits resonating at the harmonics.

dear FvM, again thank you very much.

I have a class A amplifier, and the effect I want is simultaneously having best performance for fundamental freq and offcourse maximal attenuation for the harmonics. I perfectly get that with lowpass design of the matching network it is possible.

But what arises question in my mind is more related to ADS design flow and guides that I have found.
The flow that I have seen (unfortunetly without clear explaination) is that, starting from fundamental freq, for example 'ZL' is the specific value that I find for optimum load, so far OK. As I precieve you responce, with a lowpass matching circuit I can finish the design right here.

But what I see instead, is that after finding 'ZL', the performance of system is again evaluated for both 2nd and 3rd harmonics, with sweeping the gamma phase (ofcourse around the edge of smith chart so considering the magnitude of reflection coefficient =1 ) and finding the optimum values for the load again!

There are 2 things that I absolutely don't get:
1- by this approach we will find 3 different load impedances for fundamental, 2nd and 3rd harmonic. now how to make a matching network compatible with these values in ofcourse these three different frequencies?

2- while performing the simulation for the 2nd, there is an option to put the optimum value of load impedance for fundamental frequency, I don't get how it could possibly effect the simulation result for the 2nd harmonic?? ( I did put different values and I observed different results).

I'm sorry for the long questions that I made, but unfortunetly I barely could have find convincing material on the web, and it is kindda driving me crazy.
I really appreciate for the time you spend to answer me.

Thank you.
 

A Matching Circuit which has been found for max. Delivered Power or max. Efficiency or max. Gain ( simultaneous matching is not possible for these three metrics ) will also suppress 2nd and 3rd Harmonics due to its nature.You can check how much it will attenuate these harmonics and then if it's necessary you may add harmonic traps either series or parallel resonance circuits.There are also some matching circuits that perform these two functions e.g. Matching and Suppressing Harmonics.
I recommend you to read Grebennikov's textbook to get more information.But academical papers have also very successful works.
 
  • Like
Reactions: SA1234

    SA1234

    Points: 2
    Helpful Answer Positive Rating
I got the idea, I will find and search through the book.
Thank you very much.

Regards,
Sara
 

2- while performing the simulation for the 2nd, there is an option to put the optimum value of load impedance for fundamental frequency, I don't get how it could possibly effect the simulation result for the 2nd harmonic?? ( I did put different values and I observed different results).

I am also on the same boat as you and working on Power amplifier at the moment. I am also not quite and I would also want to find out the impact of Zopt(f) at fundamental on Zopt(2f) at second harmonic. May be first, you can find out the Zopt(2f) without considering Zopt(f) and see the effect and then compare it with the case considering Zopt(f).
 

I don't understand what Zopt(2f) is meant to be. The matching network will be optimized for 1. maximum power transfer of the fundamental and 2. (as far as provided by the network topology) for suppression of harmonics, mostly 2nd and 3rd.

Load impedance at the fundamental affects the generation of harmonics in so far as it modifies the power stage operation. Additional parameters of the matching network related to harmonic suppression are a general low pass characteristic and optionally dedicated resonant traps.
 

I don't understand what Zopt(2f) is meant to be.
I meant to say optimum impedance at second harmonic. The reason I said is that in some works they try to combine fundamental and 3rd harmonic power(without supressing it) to get more output power. So in those cases, we would have an optimum load for fundamental and 3rd harmonic.
 

Thanks for explanation. I think it's a very uncommon application.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top