Noise Figure of an amplifier - RF Cafe Forums

 Post subject: Noise Figure of an amplifier

Posted: Tue Jun 26, 2007 11:39 am 

Hi all,

I have this amplifier which stated noise figure 10dB at 20dB gain.

I understand that noise figure, NF = SNRi/p - SNRo/p

Does this mean that if i set the gain to 20dB, the SNRo/p will decrease by 10dB?

Thanks.

BT

 Post subject:

Posted: Tue Jul 03, 2007 11:26 am 

Hi BT,

You have the right idea. You are correct in stating that the NF is the difference (in dB) of the i/p SNR to the o/p SNR. What this tells you is that the signal at the output of your circuit will be more noisy than the signal at your input. This should be expected after traveling through a noisy environment of resistors and active devices. Not only does the signal get amplified by the gain of your circuit, but the noise does as well. Then, because of all the added noise from the active devices, the noise level raises even higher.

For example, let's assume your input signal is at -74 dBm and (for the sake of ease) the noise floor is at -174 dBm. Your input SNR is then -74-(-174) = 100 dB.

Then it passes through your amplifier which has a 20 dB gain and a NF of 10 dB. The signal at the output of your amp will be at -74+20 = -54 dBm and the noise floor will now be at -174+20+10 = -144 dBm. Therefore, your output SNR is now -54-(-144) = 90 dB.

So basically what happens is the gain and noise of the circuit raise the overall noise floor and the signal is only affected by the gain of your amplifier. Therefore, your SNR degrades as you move from input to output of your circuit by the amount of your noise figure.

I hope this helps.

J

 Post subject:

Posted: Tue Jul 17, 2007 11:24 am 

Hi j,

Thanks for the info.

By the way, Do you know what is the typical input & output return loss of an
amplifier?

Thanks.


BT

 Post subject:

Posted: Tue Jul 17, 2007 12:17 pm 

Hi BT,

Ideally (for stability purposes) you would like the magnitude of S11 and S22 to be below 1 (the further the better). If the magnitude of either of these parameters goes above 1 you cannot say your amplifier is conditionally stable.

You should also check your K value across PVT. This value should always be above 1 for unconditional stability.

Hope this helps.

-J

 Post subject:

Posted: Tue Jul 17, 2007 12:21 pm 

Sorry for the numerous posts in one section, but I have a correction to my previous post.

In the first section I say,


What I meant to say was:

Ideally (for stability purposes) you would like the magnitude of S11 and S22 to be below 1 (the further below 1 the better). If the magnitude of either of these parameters goes above 1 you cannot say your amplifier is unconditionally stable.


-J