Biasing circuit for Power Amplifier

[haitaoz]

It is known that biasing circuit is critical for power amplifier performance.Can somebody talk about the trade-off between them?Such as from linearity,efficiency,temperature stablity,biasing stability points of view?
Thanks!

 

[electronics_kumar]

i)once if you are travelling from classA,B,C . the efficiency increases from 50%,78,100%.
ii)also its linearity decreases from class A to C.ie class A is more linear than B which is more linear than Class C
iii) CLASS A wll have operating point at Qpoint but for class B it goes near to cutoff.class C goes into deeper cutoff..
iv)class B is having crossover distortion to mitigate Class AB was proposed.
v)cLASS A will have zero signal high power dissipation ie no signal at its input high power dissipation...whereas class B will power dissipation only when it is having signal in its input


still to be explored

 

[haitaoz]

Thanks for your answering!
For the typical handset PA,most of them are biasing around class AB from linearity and efficiency point of view.However,there are some biasing circuits for each input port(like base of the transistors).I wonder what is the typical schematics of them and the trade-off between them.

Thanks again!

 

[boy]

Most PAs use active bias due to its temperatur stability. There are some active bias chips from infineon.

 

[electronics_kumar]

class A,B are used as a audio frequency amplifier whereas class C are used as radio frequency amplifier

 

[boy]

A very linear amplifier for digital modulation uses class A. Also class AB is very popular for linearity and efficiency.

 

[haitaoz]

Is there any typical biasing circuit for CDMA or GAM PA?It should not be just voltage biasing,am I right?Hope you can share your valuable experience.thanks

 

[vfone]

Biasing a PA is a challenge especially when is using digital modulation.
In GSM handset the bias of the PA it depend for what topology is used for Power Amplifier: Base/Gate voltage control (usually for closed loop approach) and Collector/Drain voltage control (open loop approach).
In CDMA/WCDMA things are different. The PA should be very linear, having in the same time a good efficiency. Here one approach is to use Dynamic Bias (different bias for different output powers).
I have to underline one mistake that is made by people very often today. They believe that in GSM/GMSK because is constant envelope can be used a pure Class C amplifier, and in CDMA/WCDMA because is nonconstant envelope the PA class shall be a pure Class A. In reality for both cases the designers use Class AB. If pure Class C is used in GSM/GMSK the Spectrum due to Switching it will go crazy, and for CDMA/WCDMA is never used a pure Class A because the PA efficiency will be less than 25%.
Main challenges of a PA bias circuit are:
-Voltage/Current thermal runaway
-Stability vs Frequency vs Output Power vs Temperature vs Voltage
-Thermal Noise
Here are some articles about PA biasing:
**broken link removed**
http://amsacta.cib.unibo.it/archive/00000573/01/GM_03_03.pdf
http://europe.nokia.com/downloads/aboutnokia/research/library/electronics/ELE5.pdf

 

[haitaoz]

vfone,
Thanks for your reply and it is really helpful

 

[lynnheflex]

Does the biasing circuit influence the matching circuit?

 

[vfone]

Does the biasing circuit influence the matching circuit?

Yes. Changing the bias it change the input impedance of the active device. For this reason the bias of the first stage should be carefully designed.
The output of the PA shall be load-pulled for different bias/Ic/Id, to find the right output match.