Amplifiers

[LezBurns]

Hello. Can anybody help me with a problem that im having. I have designed a system that has a very low signal to noise ratio. The signal that i am trying to measure is changing very slightly and it is hard to measure due to all the noise that i am getting. The noise is completly random but is there a system (instrumentation amplifier) that I can use to filter out some of the noise and if so how can I achieve it

 

[walker1]

You say your noise is 'random'. I take that as more or less white noise.
Have you tried doing an FFT on the noise ?
A few destinct frequencies in a noise signal could make it look random in time domain.

How about your signal, is it also wide band or are we talking about a single frequency, a few specific frequencies or only a few kHz wide ?

If you narrow your amplifiers bandwidth you will reduce
the overall noise level by a factor that is the sqareroot of the bandwidth reduction. (for white noise)

What signal level are we talking about ?
At low signal levels you could gain something by using an amplifier with lower noise data, but if we are above 0.2-1V, the noise reduction will most likely not be significant.

 

[IRstuff]

There are some other possibilities:

> Low pass filter the noise or, as suggested above, reduce the amplifier bandwidth. Some of this can also be accomplished on the collected digital data.

> Average the signal, particularly if it's repetitive

> Use a lock-in amplifier, works great if signal is single frequency

>

TTFN