LNA Predistorsion

[Leiser]

Hi,

I´m designing a LNA for an FM receiver. I need 60dB of IMR3 due to customer requirement. In order to achieve that high req. I used a diode in the input of the amplifier as a predistorsion method. It really works for intermodulation, but I have doubts about noise. Does anyone know if this is a good solution? What is the impact on noise of this diode in the input of the amplifier?

Thanks!

 

[VisiGoth]

Are you saying that you can count on the distortion caused by the diode will be undone by the LNA first transistor? This sounds like a schoole problem, not something an engineer would do. Or am I out of touch with the state of the art in high linearity LNA's?

I have hear of using the same transistor as a diod and running the same bias through it and thermally connecting the devices to get some semblance of match. But a match to cancle below 60 dB level sounds like an uberEngineer job.

Good luck. I will monitor this thread for reports of your success and advice of others. This should be interesting.

Merry Christmas.



jsolar

 

[biff44]

I am surprised that it works at all. Being an LNA, I assume you are receiving a very small signal. That small signal would not swing the diode current very far, and therefore you will have a pretty linear I-V curve--i.e. NOT what you want to generate predistortion. You want I = V^3.

I guess the answer to your question revolves around what the diode dynamic resistance is. If you are biasing it up so it only has an ohm or two of resistance, then it will probably not screw up the NF. If it is acting like a 200 ohm resistor, forget about calling it a LNA!