How to calculate ADC resolution
Divide the input range by the number of codes: 1 LSB = Vref ÷ 2^N for a 0 to Vref input, or 2·Vref ÷ 2^N for ±Vref. A 12-bit ADC with a 3.3 V reference resolves 805.7 µV per code. The code for an input is V ÷ LSB rounded to a whole number, and a code converts back to code × LSB. The ideal SNR is 6.02N + 1.76 dB, 74.01 dB at 12 bits.
ADC resolution formula, SNR and ENOB
An N-bit converter divides its full-scale range into 2^N equal steps. ADI MT-010 places the first code transition ½ LSB above the bottom of the range and every LSB after that, so each code is centered on a whole number of LSBs.
q = FSR ÷ 2^N code = round(Vin ÷ q) V = code × q
- q
- 1 LSB in volts
- FSR
- full-scale range: Vref (0 to Vref) or 2·Vref (±Vref)
- N
- resolution in bits
The code is clipped to 0 … 2^N − 1 (or −2^(N−1) … 2^(N−1) − 1 for bipolar), so the top code represents Vref − 1 LSB. Quantization adds noise of q ÷ √12 rms, which sets the best SNR an ideal converter can reach with a full-scale sine (ADI MT-001):
SNR = 6.02N + 1.76 dB SNR(BW) = 6.02N + 1.76 + 10·log10(fs ÷ 2BW)
The ENOB formula converts a measured SINAD to effective bits the same way, corrected for a test tone below full scale (ADI MT-003):
ENOB = (SINAD − 1.76 dB + 20·log10(full scale ÷ tone amplitude)) ÷ 6.02
ADC resolution calculation examples
Example 1: a 1.234 V sensor on a 10-bit, 5 V ADC
1 LSB = 5 V ÷ 1024 = 4.883 mV. 1.234 ÷ 4.883 mV = 252.72, which rounds to code 253. That code’s center is 1.2354 V, so reading the code back gives the input to within ½ LSB (2.44 mV), before the converter’s own offset and gain errors.
Example 2: −0.5 V on an ADS1115 set to ±2.048 V
The span is 4.096 V, so 1 LSB = 4.096 V ÷ 65,536 = 62.5 µV, as in the ADS1115 datasheet. −0.5 V ÷ 62.5 µV = −8000, which the part outputs in two’s complement as 0xE0C0. In offset binary the same code would be 0x60C0.
Example 3: ENOB from a datasheet SINAD, and what oversampling buys
A 12-bit ADC is specified at 70 dB SINAD with a −1 dBFS tone. ENOB = (70 − 1.76 + 1) ÷ 6.02 = 11.50 bits, against an ideal 74.01 dB. Sampling at 1 MS/s and filtering digitally to a 10 kHz band adds 10·log10(1 MHz ÷ 20 kHz) = 16.99 dB of process gain.
ADC resolution table: LSB size by bits and reference voltage
1 LSB = Vref ÷ 2^N for a 0 to Vref input. For a ±Vref bipolar input, double these values. For any other reference, use the ADC calculator at the top of the page.
| Bits | Codes | Vref 2.5 V | Vref 3.3 V | Vref 4.096 V | Vref 5 V | Ideal SNR |
|---|
| 8 | 256 | 9.766 mV | 12.89 mV | 16 mV | 19.53 mV | 49.93 dB |
| 10 | 1,024 | 2.441 mV | 3.223 mV | 4 mV | 4.883 mV | 61.97 dB |
| 12 | 4,096 | 610.4 µV | 805.7 µV | 1 mV | 1.221 mV | 74.01 dB |
| 14 | 16,384 | 152.6 µV | 201.4 µV | 250 µV | 305.2 µV | 86.05 dB |
| 16 | 65,536 | 38.15 µV | 50.35 µV | 62.5 µV | 76.29 µV | 98.09 dB |
| 18 | 262,144 | 9.537 µV | 12.59 µV | 15.63 µV | 19.07 µV | 110.13 dB |
| 20 | 1,048,576 | 2.384 µV | 3.147 µV | 3.906 µV | 4.768 µV | 122.17 dB |
| 24 | 16,777,216 | 149 nV | 196.7 nV | 244.1 nV | 298 nV | 146.26 dB |
SINAD to ENOB table
Effective bits for a SINAD measured with a full-scale tone. Each 6.02 dB is one bit. For a tone below full scale, use the ENOB calculator on the SNR and ENOB tab.
| SINAD | ENOB |
|---|
| 50 dB | 8.01 bits |
| 56 dB | 9.01 bits |
| 62 dB | 10.01 bits |
| 68 dB | 11.00 bits |
| 74 dB | 12.00 bits |
| 80 dB | 13.00 bits |
| 86 dB | 13.99 bits |
| 92 dB | 14.99 bits |
| 98 dB | 15.98 bits |
| 110 dB | 17.98 bits |
What limits a real ADC
Resolution is not accuracy. The reference sets the scale of every code, so a reference that is 0.1 % off moves a 12-bit full-scale reading by 4.1 LSB; use a dedicated reference rather than a noisy supply when the absolute value matters. Offset, gain error and INL add more, and thermal noise, clock jitter and distortion keep every real converter’s ENOB below its nominal resolution, so read the SINAD or ENOB line of the datasheet. Keep the input inside the range, scale it with a divider or amplifier so the signal uses most of the codes, and put an anti-alias filter in front: anything above fs/2 folds back into the band and no amount of resolution removes it.
Glossary
- LSB (least significant bit)
- The input change that moves the code by one: full-scale range ÷ 2^N. Also called the quantization step q.
- Full-scale range (FSR)
- The analog span the codes cover: Vref for a 0 to Vref input, 2·Vref for ±Vref.
- Code center
- The input voltage a code represents exactly, code × LSB; inputs within ½ LSB of it give that code.
- Dynamic range
- Full scale over one LSB, 20·log10(2^N): about 6.02 dB per bit, or 72.2 dB at 12 bits.
- SINAD
- Signal to noise and distortion ratio: the rms signal over everything else in the spectrum, harmonics included.
- ENOB
- Effective number of bits: the resolution of an ideal ADC with the same SINAD.
- Two’s complement
- Signed binary where negative codes have the top bit set; the usual output format of bipolar ADCs.
Frequently asked questions
What is ADC resolution?
Analog-to-digital converter resolution is the number of bits in the output code, and the voltage step that follows from it: 1 LSB = full-scale range ÷ 2^N. A 12-bit ADC with a 3.3 V reference has 4,096 codes and resolves 805.7 µV per code.
What is the resolution of a 12-bit ADC with a 3.3 V reference?
One LSB is 3.3 V ÷ 4096 = 805.7 µV. The top code, 4095, represents 3.2992 V, one LSB below the reference.
How do I convert an ADC value to voltage?
Multiply the code by 1 LSB: V = code × Vref ÷ 2^N for a 0 to Vref input. Code 2048 on a 12-bit ADC with a 3.3 V reference reads 1.65 V. The Code to voltage tab is an ADC value to voltage calculator for unipolar and bipolar codes.
Should I divide by 4095 or 4096 to convert ADC counts to volts?
Divide by 4096 (2^N) for a 12-bit converter: that is the LSB size, and the all-ones code sits one LSB below Vref (ADI MT-010). Dividing by 4095 instead makes the top code read exactly Vref; the difference is under 1 LSB. TI SLAA013 measures full scale between the first and last code centers, which is 2^N − 1 steps of the same LSB.
Why is the ideal SNR of an ADC 6.02N + 1.76 dB?
Quantization error is close to a sawtooth one LSB high, so its rms value is q/√12. A full-scale sine has an rms value of q·2^N/(2√2). Their ratio is 2^N·√1.5, which is 6.02N + 1.76 dB: 74.01 dB for 12 bits and 98.09 dB for 16 bits (ADI MT-001).
What is ENOB and how do I calculate it?
The effective number of bits is the resolution of an ideal converter with the same SINAD: ENOB = (SINAD − 1.76) ÷ 6.02 for a full-scale tone. If the test tone was below full scale, add the difference in dB first (ADI MT-003). A 12-bit ADC with 70 dB SINAD at −1 dBFS has an ENOB of 11.50 bits.
What is the quantization noise of an ADC?
q ÷ √12 rms, where q is one LSB, spread over dc to fs/2. A 12-bit converter with a 3.3 V reference has 233 µV rms of quantization noise.
Does oversampling increase ADC resolution?
It can increase SNR in a narrower band. If noise outside the signal bandwidth BW is removed with a digital filter, SNR rises by 10·log10(fs ÷ 2BW) (ADI MT-001): about 6 dB, or one bit, for every 4× of oversampling. Sampling at 1 MS/s for a 10 kHz signal gives 16.99 dB.
How do I read a bipolar ADC code?
Many bipolar ADCs, such as the TI ADS1115, output two’s complement: positive full scale is 0x7FFF, zero is 0x0000 and −1 LSB is 0xFFFF for 16 bits (ADS1115 Table 7-3). Others use offset binary, which is the signed code plus 2^(N−1), so zero is 0x8000. Pick the format in the calculator, and enter Vref as the positive limit of a ±Vref range.
More free PCB tools
See all free electronics tools →
- Nyquist frequency formula and aliasing calculatorWhere a signal aliases for a given sample rate, the Nyquist frequency, and the anti-alias filter attenuation you need.
- Stepper motor steps per mm calculatorSteps per millimeter for belt and lead screw drives from motor step angle, microstepping, pulley teeth and pitch, and the top speed.
- R2R DAC calculator (R-2R ladder)Output voltage for any code of an R-2R resistor ladder DAC, the step size, output resistance and resistor-tolerance error.
- IEEE 754 floating-point converterSee the sign, exponent and mantissa bits of a float32 or float64, convert hex to float and back, and the rounding error.
- 555 timer calculatorFrequency, period and duty cycle of a 555 in astable mode, or the pulse width in monostable, and the parts for a target.
- Fixed-point calculator (Q format)Convert numbers to and from Qm.n fixed point: integer, hex, resolution, range and rounding error.
Sources
- Analog Devices MT-001, Taking the Mystery out of the Infamous Formula “SNR = 6.02N + 1.76 dB”
- Analog Devices MT-003, Understand SINAD, ENOB, SNR, THD, THD + N, and SFDR
- Analog Devices MT-010, The Importance of Data Converter Static Specifications (LSB size, ideal transfer function)
- Texas Instruments ADS1115 datasheet (FSR and LSB size, ideal output codes)
- Texas Instruments SLAA013, Understanding Data Converters
Design the whole board, not one number
ProtoFlow is free AI PCB design software: describe a circuit and get a schematic with real parts, PCB layout with an autorouter, SPICE simulation and Gerber files. Native KiCad, Altium and EAGLE import and export. Free on macOS, Windows and Linux.