A question on one of Lee's paper on LNA

[Ruritania]

Hi,

I have a question on one of the most often cited papers of Thomas H. Lee's: "A 1.5-V, 1.5-GHz CMOS Low Noise Amplifier" (IEEE JSSC, Vol.32, No.5, 1997)

Where does the denominator (1+ωT*Ls/Rs)² come from in the equation (12)?? Isn't that the channel resistance noise (id) itself is located at the output, and contributes itself totally to the output noise power?

Thanks.

Ruri

 

[dsjomo]

Split id to two identical noise sources, one is inserted between output and ground, the other is inserted between source node and ground. Calculate the total output noise current while the output node is shorted to ground.

Then, you'll get the exact equation above. Don't forget to ignore rds

Why the output noise current is not id? It is because that id is injected to the source node, this cause the source node votage to rise, then rise the Vgs in negative direction and so does gmVgs. The total output noise is (gmVgs + id). If Ls = 0, then the injected noise current to source node does not affect the source node voltage. Under this situation, the total output noise current is id.