Why matching impedance of LNA since we only concern about signal voltage?

[tendo R]

Hi all,
I am wondering why we do impedance matching for LNA. Because in most sensor circuits where opamps are used it is the input voltage to amplifiers are maximized.
From this post https://www.edaboard.com/threads/67820/ I only understand that when maximum power transfered to system the SNR would be better. But for MOSFET LNA which is voltage controlled amplifier the input voltage should decide the output signal quality.
Please let me know where is my mistake and explain why it is different from sensor circuits which use instrumentation amplifiers, since their SNR are also important.
Thanks for your help!

 

[mtwieg]

The impedance matching will affect not only the voltage from the source, but also the input-referred noise of the amplifier. It turns out that often times maximizing the source voltage and the power gain does not give the best noise performance, because the noise will be increased even more. There is a condition called a noise match (which is different from the power match) which gives maximum SNR, and that's what we usually aim for.

 

[tendo R]

The impedance matching will affect not only the voltage from the source, but also the input-referred noise of the amplifier. It turns out that often times maximizing the source voltage and the power gain does not give the best noise performance, because the noise will be increased even more. There is a condition called a noise match (which is different from the power match) which gives maximum SNR, and that's what we usually aim for.

Thanks for your reply.
But when we talk about impadance matching for LNA, it is Zl=Zi*. Is it happen to be that power matching is the same situation as noise matching?

 

[FvM]

Power and noise matching points are generally not the same, as mtwieg already stated. For the simple case of a purely real amplifier impedance, noise matching is achieved when en and in contribution is balanced, Ropt = en/in.