3
\$\begingroup\$

This is from Analog-to-Digital Converison by Marcel Pelgrom

He defines an approximate relationship between Signal to Quantization Noise Ratio (SNQR or just SNR) and number of bits as: enter image description here

The author then uses this later on, enter image description here

How does he replace SNR with SINAD in 5.13 to get the ENOB equation? Aren't they different?

\$\endgroup\$

1 Answer 1

4
\$\begingroup\$

SNR and SINAD perform a similar function.

SNR is Signal to Noise ratio. When considering a perfect quantiser, it's a good measure of how much noise is introduced by the quantisation steps.

SINAD is Signal to Noise and Distortion. Distortion is caused by a non-linear converter. Its effect on the signal is to introduce other tones that should not be there, in other words, energy that is not part of the wanted signal. Noise and distortion are lumped together as being energy that should not be there.

ENOB, or Effective Number of Bits, is generally used to get a measure of how imperfect a signal converter is, so generally uses SINAD rather than plain SNR.

However, these are approximate figures, and there's little point getting picky about them if the actual performance of a converter matters to you. If it matters, you'll specify it and measure it for exactly the performance you need.

The only two classes of people who would want to get picky about exactly how ENOB is calculated are (1) Picky people who like a fight and (2) Marketting people from converter brand A who want their ENOB to be 0.1 bits better than the converter from brand B.

\$\endgroup\$
1
  • \$\begingroup\$ Brilliant response. Thank you very much. That cleared up everything. Also one more thing, SNR and SNQR are the same thing right? I've seen some books where they are used interchangably \$\endgroup\$ Commented Dec 26, 2019 at 17:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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