Hi!

I made an AFE for a gas sensor, and I would like to know how big is my actual noise. The circuit is quite simple, just an current to voltage amplifier, low pass filter (2th order sallen-key), second amplifier with adjustable gain via digital potentiometer and 18-bit ADC.
If I measure, with oscilloscope, the noise present at the ADC's input (when the input of the AFE is not connected and the gain setup is about 70%) I get about 18mVp-p (the noise measured with the probe shorted is about 3mVp-p), this noise has two major components, one of 160kHz (the smaller) and the other of about 6kHz (the bigger).
But the bandwidth of interest is below 10Hz and the ADC (MCP3424) specifies a response of -70dB about 5kHz, when I measure the voltage with the ADC (same conditions: input of the AFE is not connected and gain setup is about 70%) I get about 0.3mV.
So my question is: it's safe to say the resultant noise is below 1mV because I have enough filtering, or is there any possibilities for the higher frequencies to interfere in some way I can't think of in the final measurement?

 

Please post your schematic as 800x600 gif file and the noise as 800x600 jpg file, so our members could give you the suggestion.

 

I was heavily involved in a recent, similar, discussion.

This thread might have some useful info for you.