0
\$\begingroup\$

Suppose I have an inverting op-amp (part of a trans-impedance amplifier circuit) that gives me a slightly-positive-but-small signal, and the output goes into a ADC. I have two choices:

  1. Increase the Rf (feedback resistor) value which should amplify the output.

  2. After recording from the ADC, use software to scale the value to "look" like I amplified by Rf.

Is there any difference between these two scenarios?

\$\endgroup\$
0
3
\$\begingroup\$

You ADC doesn't have infinitesmal resolution. Measure in a small signal = less resolution for that signal. Once its digitized, that's it. The detail is gone.

It's the audio equivalent of "ENHANCE!" when TV police investigate video footage.

What I described above is called quantization noise, but doing so will also makes the noise that your ADC will introduce more significant relative to your signal.

\$\endgroup\$
1
  • \$\begingroup\$ lol makes sense - thanks! \$\endgroup\$ – Cam.Davidson.Pilon Feb 12 at 19:30
2
\$\begingroup\$

Multiplying your ADC measurements in software is like digital zoom. If you had enough resolution in the original digitized data, then sure, it will work great. If your opamp only provides a gain of say two, it only provides one extra bit of resolution, and most likely you can remove it and a 10-12 bit microcontroller ADC will work just fine with one less bit.

But if the opamp has a gain of 1000, and you remove it, your 12 bit ADC becomes a 2 bit ADC.

Sometimes, you need more pixels.

enter image description here

\$\endgroup\$
2
  • 2
    \$\begingroup\$ Zoom! Enhance! More! \$\endgroup\$ – DKNguyen Feb 12 at 20:40
  • \$\begingroup\$ Great analogy "Multiplying your ADC measurements in software is like digital zoom." thanks \$\endgroup\$ – Cam.Davidson.Pilon Feb 13 at 1:49
0
\$\begingroup\$

Generally the optimal solution (if it meets requirements, of course) is to amplify the signal enough that it (just) remains within the full range of the ADC under all conditions. If it stays within 20% to 80% of range, that's most of the benefit.

As well as the quantization issue, most ADCs (especially those on MCU chips) have quite a bit of noise, so you may only get 9 or 10 bits from a 12 bit ADC. At higher sampling speeds, the noise tends to increase. You can often increase the signal to noise ratio by amplifying the signal. Of course the noise is mixed in with the signal at the input, then you amplify that portion of the noise as well, so there is no improvement, but the effect of the ADC noise will be reduced.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.