Signal to noise ratio for amplifier

[stanislavb]

Hi,
Some guru told me that for amplifier increasing gain we reducing bandwidth (and I agree with him). However he claimed that it leads to increasing Signal/noise ratio. And this is not so clear. Increasing gain we also increase noise amplitude. Although we decrease noise bandwidth we increase noise amplitude with the same factor. So I am feeling that ratio should stay the same.
What is yours opinion?
Thank you

 

[keith1200rs]

It depends where the noise is coming from, but it is quite likely that higher gain improves signal to noise ratio. That will not always be the case though. If you had some more details it would be possible to be specific.

Keith

 

[Audioguru]

A lousy old opamp like a 741 is extremely noisy when it has high gain even when the high gain reduces its bandwidth. Some of the noise is "popcorn" noise at very low frequencies.
There are many better low noise opamps available today.

 

[stanislavb]

Hi,
We are speaking about noise with incoming signal. If you are thinking that for such noise it improves SNR, what is the reason. I repeat that enlarging gain reduce bandwidth in the same factor(according theory).
Thank you

 

[Audioguru]

Noise on the incoming signal has nothing to do with the signal to noise ratio of the amplifier.

If the noise is hiss on the incoming signal then you can chop off the high frequencies.
If the noise is hum on the incoming signal then you can chop off low frequencies or notch it out.

Then you are left with a telephone type of sound.

 

[keith1200rs]

It you are just talking about noise contained within the source signal then yes, on your example the signal to noise ratio will increase with increased gain. It isn't a helpful increase though and one that could be achieved simply by reducing the bandwidth at a lower gain.

A 20dB increase in gain would increase both the signal and noise by 20dB but if the noise bandwidth reduces by a decade as well you will have less noise. You could achieve that bandwidth reduction in other ways.

The scenarios I was thinking of were transimpedance amplifiers where noise is often dominated by the feedback resistor and a single transistor amplifier where you can plot constant noise figure curves for a given input impedance and choose the optimum gain and operating conditions.

In general, increasing gain just to reduce bandwidth is not the best way of reducing bandwidth.

Keith