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.

[SOLVED] One Simple Question on Load Pull for PA

Status
Not open for further replies.

wyattmengy

Newbie level 5
Joined
May 8, 2012
Messages
9
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Location
Ontario, Canada
Activity points
1,394
Hi everyone, I suddenly had a brain cramp and confused myself on this question I thought I understood:

After obtaining the load pull result and the corresponding impedance value from ADS (shown in figure attached), is this impedance value (8.779+j3.15) Zopt, or conjugate of Zopt? In other words, for designing the output matching of a PA using load pull, give that the final load impedance is 50 ohms, do I need to design the output matching network to transform 50 ohms to Zopt (8.779+j3.15), or Zopt* (8.779-j3.15), so that the transistor will output maximum power or gain?

Thanks!

load pull.PNG
 

"Load Impedance" should mean the impedance of the load (as if you were standing at the PA, looking into the load termiantion). Output impedance is the impedance that the PA would look like if you disconnected the load and "looked into the PA". That would require you to conjugate match.

Since you are told what the LOAD impedance is, use it directly.
 
I guess I was a bit confused between conjugate matching and matching according to Zopt provided by load pull test. According to Pozar (see figure below), the output impedance is defined as looking into the drain of the transistor, and load impedance is defined as looking into the output matching network from the transistor. So I think in the case of nonlinear PA and load pull, the Zopt returned by load pull is actually the optimum Zl, which is whatever value the output matching network converts the 50 ohm load into, and this has nothing to do with Zs and neither conjugate matching for small signal amps at the output.

Am I right?





Pozar.PNG
 

In fact the load pull tuner is actually your matching network, having 50 ohms in one side (Zo), and on the other side having the required impedance that you have to provide to get optimized parameters of the PA (ΓL)
 
In fact the load pull tuner is actually your matching network, having 50 ohms in one side (Zo), and on the other side having the required impedance that you have to provide to get optimized parameters of the PA (ΓL)

Does this mean that if the load pull test tells me a Zopt value which can return the maximum power output, then I can design the output matching network to convert 50 ohms load into this Zopt appears to the FET, the dynamic load line will be in maximum swing length on the I-V curve?

Thanks!
 

If the simulation is telling you Zopt = Zload for the case of max power out, then yes... design your matching network to transform 50 ohms into Zopt. As vfone said, you are simply replacing the tuner object with your own implementation to perform the impedance transformation.

Two caveats:
1. If this is for a school project/homework, or a narrow band design, then a single-point match should be sufficient.

2. Rarely is your output impedance a nice 50 ohms (say, if your next stage is a filter or an antenna). You quite often have to tweak the output match once you marry up the PA to the next stage because of a non-50 ohm termination impedance seen by your matching network. If this is just a paper exercise, it might be a good thing to mention in your write-up, but you won't need to incorporate it into your design steps. If you are doing this to build and implement, be aware that retuning the match could provide some benefit in maximizing the circuit's performance.
 
Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top