LNA noise matching, Imaginary part

[ldynasa]

Hello, everyone.

While doing noise matching for LNA, suppose Zopt = Gopt + j*Bopt. If Gopt shall be 1/Rs, then what about Bopt? Should it zero? Or does it affect if it's not?

Thanks

 

[hosein_rabiee_g]

it affect the noise power. you must match the impedance to conjugate matched Impedance optimized for noise. see constant noise circle in smith chart.(Gonzalez book)

 

[ldynasa]

So what's the common solution to make this imaginary part zero?

 

[BigBoss]

So what's the common solution to make this imaginary part zero?

If your LNA has narrowband, you should use an inductor in series at emitter/source to compansate Cgs/Cbe capacitance.While doing this, you may arrange and pull the imaginary part of Optimum Noise Impedance.Quality factor of this inductor is very important and it should be carfully optimized.For wideband LNAs, it's not possible.

 

[ldynasa]

So you mean that emitter inductor is able to pull this imaginary part of optimum noise impedance to zero?

 

[BigBoss]

So you mean that emitter inductor is able to pull this imaginary part of optimum noise impedance to zero?

Emitter inductance will compansate capacitive part of input impedance of the LNA and you will able to do an optimization by adding emitter inductor and maybe an additional base inductor to implement your wanted input impedance.