The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> Relevance of S22 of integrated PA
https://designers-guide.org/forum/YaBB.pl?num=1187355418

Message started by Prabhu on Aug 17^th, 2007, 5:56am

Title: Relevance of S22 of integrated PA
Post by Prabhu on Aug 17^th, 2007, 5:56am

Is S22_PA a specification?

As I understand we pump in current into a resitive impedance (R), looking into the matching network + antenna, determined by available voltage swing limits. The matching network transfers the power dissipated in R to antenna.
In this picture I dont see any relevance of specifying S22 for the PA.

Is my view on the irrelevance of S22_PA correct?

Thanks,
Raja

Title: Re: Relevance of S22 of integrated PA
Post by didac on Aug 17^th, 2007, 8:19am

Hi,
I think the answer to this question is "it depends", I mean what are you designing?
A stand-alone PA for the semi-custom market?
A transceiver?
A complete system(transceiver+PCB+antenna)?
If it is a complete system you have the freedom to choose the output impedance for you PA, if you are designing a transceiver usually input and output RF impedances are referenced to 50 ohms, and for a a stand-alone chip I think it's compulsory to achieve a good input and output matching.
Just my thoughts,
hope it helps.

Title: Re: Relevance of S22 of integrated PA
Post by aaron_do on Aug 17^th, 2007, 5:54pm

Hi Raja,

If i understand your question correctly, R represents the resistance seen by the RF power transistor looking into the matching network connected to the antenna, right? The power transistor itself has some output resistance, R2. For maximum power transfer, R2 must equal R. This is equavalent to saying S22 = 0 from the antenna side.

cheers,
Aaron

Title: Re: Relevance of S22 of integrated PA
Post by RFICDUDE on Aug 17^th, 2007, 8:47pm

There is relevance; however, it is not in the same sense as conjugate power matching.
For a PA, efficiency, linearity, and stability are the main considerations. To maximize efficiency, we want to produce the largest possible amount of power delivered to the load given the power supply constraints available. Note this is not power transfer since we want to maximize the power delivered AND minimize the power dissipated by the source. If the transistor is an ideal current source then this condition occurs when the current source delivers a peak value which causes the peak load voltage to be Vdd for a common source amp with an ideal choke connected to the drain and AC coupling cap from the drain to the load.

Of course in a real device there is capacitance and little bit of resistance. In this case we look for a load impedance which permits the required amount of peak current swing and voltage swing at the drain of the power device. The somewhat empirical technique for characterizing or finding the optimal load is called "load pull."

There is a somewhat infamous paper by Steve Cripps on the concept.
Cripps, S.C., "A Theory for the Prediction of GaAs FET Load-Pull Power Contours," Microwave Symposium Digest, MTT-S International , vol.83, no.1, pp. 221-223, May 1983

He has published a few books on PA design too.

Title: Re: Relevance of S22 of integrated PA
Post by WillN on Aug 20^th, 2007, 5:13pm

Agree with your statement Probhu and RFICDUDE. Basically, output matching network is used to transform antenna/duplexer impedance to your required load impedance presented to the output device base on your efficiency-linearity trade-off. So, this output network can only serve this only purpose and you get whatever output impedance you can get.

If you really need good (close to 50Ohm) impedance, say for matching to your duplexer, you will need to have balanced (actually quadrature) PA which essentially cancel out the reflection from the load and matching networks are located between the output devices and quadrature combiner.

Title: Re: Relevance of S22 of integrated PA
Post by Prabhu on Aug 21^st, 2007, 6:34am

Thanks for all your valuable replies.

I conclude that once the resistive impedance looking into the matching n/w (for the given antenna impedance ) is fixed, for efficent use of the available voltage supply, s22 becomes irrelevant or ceases to be a design specification.

Raja

Title: Re: Relevance of S22 of integrated PA
Post by RFICDUDE on Aug 21^st, 2007, 9:25pm

Basically you are right, but there is a big assumption that the load is 50 ohms. Usually S22 for an optimized PA is not specified; however, the efficiency and linearity measurements/specifications are valide when the load is 50 ohms. Stability is also stated with a range of VSWR mismatch implying the amplifier is stable for a wide range of none 50 ohm impedances; although, none of the other performance specifications are speficied under the mismatch conditions.