ADC ENOB vs Resolution

Hi harianalog ,

Both calculations you mentioned are useful depending on where we are going to use the LSB. Resolution is used when we are in architectural or designing stage, while ENOB is used when evaluating real-world accuracy or measurement performance.

Datasheets prominently feature the resolution since this will help what the ADC can represent in terms of digital interface width and memory size, as well as its possible maximum precision. It is not a marketing spec but you are correct that is a theoretical spec. ENOB on the other hand tells the real usable precision of the ADC.

Example: For a 24-bit resolution with 20-bit ENOB.

Resolution = size of each code (24-bits)
ENOB = size of each meaningful code (20-bits, 4-bits noise)

Please let us know if the information above is insufficient to answer your question.
Thank you.

Regards,
Francis

The following links are good background reading: 

• MT-001 Taking the Mystery out of the Infamous Formula, "SNR = 6.02N + 1.76dB," and Why You Should Care https://www.analog.com/media/en/training-seminars/tutorials/MT-001.pdf
• MT-003 Understand SINAD, ENOB, SNR, THD, THD + N, and SFDR so You Don't Get Lost in the Noise Floor https://www.analog.com/media/en/training-seminars/tutorials/MT-003.pdf
• MT-229 Quantization Noise: An Expanded Derivation of the Equation, SNR = 6.02 N + 1.76 dB https://www.analog.com/media/en/training-seminars/tutorials/MT-229.pdf
• MT-230 Noise Considerations in High Speed Converter Signal Chains https://www.analog.com/media/en/training-seminars/tutorials/MT-230.pdf
• Analog Dialogue Issue #90 Resolution vs. ENOB – Still Hazy After All These Years https://www.analog.com/en/resources/analog-dialogue/raqs/raq-issue-90.html
• Noise Spectral Density: A New ADC Metric? https://www.analog.com/en/resources/technical-articles/noise-spectral-density.html

There are also some online calculators: 

• SNR/THD/SINAD Calculator (Computes RMS noise and ENOB from SNR, THD and SINAD) https://www.analog.com/en/resources/interactive-design-tools/snr-thd-sinad-calculator.html
• Data Conversion Calculator https://www.analog.com/en/resources/interactive-design-tools/data-conversion-calculator.html
• Effective Number of Bits Calculator Tutorial https://www.analog.com/en/resources/design-notes/effective-number-of-bits-calculator-tutorial.html

Higher resolution ADCs are limited by their intrinsic noise (use the measured SNR to derive the ENOB) whereas in lower resolution ADCs, quantization noise dominates (SNR is closer to that of an 'ideal' n-bit converter).

Thank you, I willgo through all the materials provided. Below is mycurrent understanding. May I know is this correct or not.

My current understanding:

Resolution defines the theoretical quantization - a 16-bit ADC has 2^16 discrete output codes, so with a 1V reference:

1 LSB = 1V / 2^16 = 15.26 µV (the step size between codes)

ENOB represents the actual usable performance accounting for real-world noise and distortion. If the same 16-bit ADC has ENOB = 14 bits, then:

Effective noise floor ≈ 1V / 2^14 = 61 µV RMS. Signals smaller than ~61 µV will be buried in the noise (hard to distinguish from random fluctuations)

This means that while the ADC outputs 16-bit codes with 15.26 µV steps, the inherent noise causes the readings to fluctuate by approximately ±2 LSB, making the effective resolution equivalent to a cleaner 14-bit converter.

Thank you, I posted my initial understanding about this topic. Could you please check

Hi harianalog ,

Yes, your understanding on resolution and step size is correct.

ENOB is a guide to how much useful resolution you can get, taking into account noise and distortion.  Dynamic range, SNR, THD, SINAD and ENOB are all mathematically related.  If you have the SNR and test condition (e.g. -1 dBFS) you can use that to work out the RMS noise of the converter.

As the RMS noise is essentially Gaussian in nature, there is a subtlety that you can reduce the noise floor and increase signal to noise ratio by oversampling and decimating.  Take the AD4080, for example.  It is a 20-bit converter with 22-bit linearity (DNL 0.25 ppm), a measured SNR of 93.6 dB at fin = 1 kHz, and a dynamic range of 94.6 dB.  The theoretical dynamic range of a 20-bit converter, based on quantization noise alone) is 120 dB.  If you use it to oversample an input signal, you can get processing gain (+3 dB every time you double the sample rate and average) that takes you closer to the theoretical dynamic range of an n-bit converter.  The example in the datasheet shows an increase in signal to noise ratio to 102.5 dB using the Sinc5+ compensation filter and decimating by 8, taking you from an ENOB of 15.4 to an ENOB of 16.7.

So resolution, linearity and noise are all important and will have a bearing on your signal chain design.