source pull
Another question about stability cheking of a PA. I can easily sheck it in small-signal mode with k-factor and SSB and LSB. But how to in large signal mode? Should I use large signal S-parameters?
Source pulling is used to find the optimum match at the input of the DUT. Some pulling measurements (as Power Gain) requires to do source pull after load pull.
If you have access to the large signal S-parameters is ok to find high power oscillations.
Other option is to use load pull setup and check for oscillations using a spectrum analyzer.
Can you advice a measurement system for source/load pull ? Can I use nonlinear VNA to get source/load pull measurement from large signal S-parameters?
* the measurement system from Agilient
Measurement load pull systems makes Maury and Focus:
http://www.maurymw.com/
http://www.focus-microwaves.com/
On their sites you can find a lot of informative applications.
You cannot use a standard VNA to find large signal S parameters
vfone: "You cannot use a standard VNA to find large signal S parameters"
But I meant nonlinear VNA.
did u check the new agilent X paramters and NVNA
khouly
Yes, I mean this one. But I am not shure, is it possible to extract correct Load-pull and source-pull contours from the large signal S-parameters (which I can get with NVNA)?
Load Pull setup is dedicated to the large signal measurements , u need to check the manufacturing companies provided by vfone
khouly
I'm sorry if my question is not clear... What I really want to know, can I
1) Measure X-parameters with NVNA.
2) Export the X-parameters to ADS (or another CAD tools).
3) Simulate load-pull and source-pull contours with the imported X-parameters.
4) Are the load-pull and source-pull contours correct in this case?
If yes, this means that I don't need special load/source-pull measurement instrumentation to get optimum load/source impedance of a PA (if I have NVNA :))...
Acording to Agilent you can do all these measurements using the NVNA
http://cp.literature.agilent.com/lit...989-8575EN.pdf
From the moment you imported the X parameters into ADS (I don?t know if are supported by other CAD tools) you can do a load pull simulation.
Theoretically after whatever simulation you don?t need lab instrumentation for real life measurements. But in reality you need.
"X-parameters enable accurate nonlinear simulation under small
to moderate mismatch conditions." - from 5989-8575EN.pdf page 4.
Does it mean the X-parameters can be used (in simulation) only for small variation of impedance at input and output?
So if I simulate load-pull countors with X-parameters, it is not accurate (if the load pulled far from 50 Ohm).
In contrast direct load-pull measument data allows accuratly determine optimum impedance, even with big variation of impedance at output.
vfone
i think X parameters is not supported yet in other CAD tools
khouly
This make sense, when I think that X-parameters are the ?invention? of Agilent.
Good point. This is a contradiction because on the same page is specified that : ?X-parameters are a fully nonlinear framework?.
Fully nonlinear framework means everything up to the worst conditions.
Anyway, this prove to myself that there is a lack of range using these kind of parameters. Strong nonlinearities in general are hard to be predicted.
There is a quote of Niels Bohr: ?It?s very difficult to make predictions - especially about the future.?
[/quote]
yeah ,i agree with u vfone.
i think in the application note. sure i agree with that. coz when u have a large mismatch the refelected wave will affect the nonlinearities of the device , and this can't be modeled
khouly
Hi
Recently I met a problem of PA unstability under load mismatching (high VSWR). There is low frequency oscillation problem. But I can observe these phenomenon by k factor from small signal s parameter from neither measurement nor simulation result. My colleague also try to use non-linear model (Gummel-Poon I guess) to try to simulate this behavior but it does not work neither.
I saw Agilent provide the new non-linear model. Have anyone ever use Agilent X-parameter to simulate PA stability?
Thanks
MacGyFu
Real stability sometimes shows some divergences from k-Factor theory.There are some recent development on NDF (Normalized Determinant Function) and I have personally not tried seriously this stability criteria on a circuit but it seems to be more analytic and straightforward method and accurate.