PA's linearization using diode

[jeff0283]

Hi I am doing a PA's linearization using a series Schottky diode connected at PA's gate terminal.
I choose the Avago HSMS280F,and here is the datasheet:

I am trying to tune the diode to achieve gain expansion and decreasing phase shifting, to compensate AM-AM and AM-PM distortion of my PA.
My schematic is as the attachment.

After sweep the series R(1k~16k Ohm) and the parallel C(1pF~20pF),
I find when R=15k, C=2pF ,the result can achieve gain expansion and decreasing phase shifting, but the difference is too small.
(For gain expansion,dB(S21)=-0.13dB@Pin=-20 dBm, dB(S21)=-0.11dB@Pin=20dBm delta=0.02 dB)
(For gain expansion,phase(S21)=4.11 degree@Pin=-20 dBm, phase(S21)=3.97 degree@Pin=20dBm delta=-1.4 degree)
How can I adjust my schematic to have a bigger defference?
I have attach my ADS workspace.
Thanks~~

 

[vfone]

I don't think a simple Schottcky diode can compensate the AM-AM and AM-PM distortion of a PA, even is about a single stage PA.
The gain expansion and phase shift of a diode is lower than a transistor distortion. I think you need a more complex network to do this compensation, or just live with the compensation that you get.
A higher phase shift you can get using a varactor diode, but this is not as stable and predictable as a Schotcky diode.

 

[chuckey]

I think you have discovered why semiconductor manufacturers do not recommend this as a way to get more linear power out of their devices! If it was possible it would have been done 40 years ago.
FWIW the BBC tried this out in 1970 to linearise the amplitude response of their klystron amplifiers (sine transfer characteristic).
Frank

 

[bhatt_190487]

Do the Bias voltage sweep at low voltage values. At that time keep bias resistance in ohms . Yow will
get different expansions.

 

[biff44]

back in the day I did a two diode (antiparallel diodes) predistorter that gave us a good 5 dB reduction in backoff for 64 QAM modulation. we had to screw with the dc bias to tweak it up though.

I do not see how a single diode would work...it seems to be the wrong intermods being generated. You want 3rd order intermod suppression

 

[jeff0283]

ok,I know, and thank you all for replying~