Subscribe
Tools Automation

Analog Input Resolution Calculator (bits, step size, ENOB, PLC platforms)

step size = EU span ÷ steps ; steps = 2^bits bits, steps, EU/count, dB

ENOB vs nominal bits. A "16-bit ADC" typically delivers only 12–14 effective bits due to noise, sampling artifacts, and reference-voltage stability. Manufacturers publish ENOB (effective number of bits) in datasheets — always the honest number. For instrumentation-grade converters (sigma-delta, dual-slope), ENOB usually equals the nominal bits within 0,5–1 bit. For fast SAR converters at high sample rates, ENOB can be 2–4 bits below nominal. Toggle the ENOB option to see the practical resolution vs the marketing number.

What ADC resolution actually gets you

Every analog input on a PLC or DAQ card is really an analog-to-digital converter (ADC) that turns a continuous signal into a series of discrete numeric codes. The number of possible codes — and therefore how finely you can distinguish between different signal values — is set by the ADC’s resolution. A “12-bit ADC” gives 2¹² = 4096 discrete codes across its input range. Divide your engineering unit span by that number, and you have the smallest change you can theoretically detect. Everything else in this calculator is a variation on that idea.

How to use this calculator

Three modes cover the three real questions:

  • Bits → step size — you know the resolution (either bit count or a specific PLC platform), you want to know the smallest change you can resolve in your engineering units. Best for spec review and system design.
  • Step size → bits needed — you know how precisely you need to measure, you want to know what resolution the ADC needs. Best for early design and requirement generation.
  • Platform comparison — see the resolution of the same measurement across all common PLC platforms and generic bit depths side-by-side. Best for platform selection decisions.

The formula

The core relationship is elementary:

step size = EU span ÷ number of steps

Where number of steps = 2^bits for a standard binary ADC. Every measurement fits somewhere in this equation:

  • A 12-bit ADC (4096 steps) reading a 0–10 bar pressure transmitter: step size = 10 ÷ 4096 = 0,00244 bar = 2,44 mbar
  • A 16-bit ADC (65536 steps) same range: 10 ÷ 65536 = 0,153 mbar — 16× finer
  • A Siemens S7 analog input (27648 steps, close to 15 bits): 10 ÷ 27648 = 0,362 mbar

The vendor quirks

Different PLC vendors use different count ranges for the “normal” 4-20 mA measurement:

  • Siemens S7-1200 / S7-1500 — 0 to 27648. Not a power of 2. Log₂(27648) ≈ 14,75 effective bits. The odd number preserves diagnostic bandwidth for over/under-range on the module’s full 16-bit signed output.
  • Allen-Bradley Logix (CompactLogix / ControlLogix) — 0 to 32767. Full positive half of signed 16-bit = exactly 15 bits. Slightly better resolution than Siemens for the same input.
  • AB Micro850 — 0 to 4095. Straight 12-bit. About 8× coarser than the bigger AB CPUs.
  • Mitsubishi FX — 0 to 4000. Slightly less than 12 bits effective.
  • Omron NX / NJ — 0 to 6000. About 12,55 effective bits.

These small vendor differences translate to meaningful practical differences. For a 0–10 bar pressure transmitter over 4–20 mA: Siemens gives 0,362 mbar per step; AB Logix gives 0,305 mbar per step; Micro850 gives 2,44 mbar per step. Six times worse on the entry-level AB — that matters for tight control loops.

Effective Number of Bits (ENOB)

Here is the trap: a “16-bit ADC” almost never delivers 16 clean bits of information. Real converters add noise from the reference voltage, from the sampling process, and from thermal effects in the analog front end. The effective number of bits (ENOB) is the honest number, and it appears in the manufacturer’s datasheet — always look for it.

The relationship between noise (SNR in dB) and ENOB, from a theoretical sine-wave input:

SNR (dB) = 6,02 × ENOB + 1,76

Real-world ENOB depends on the ADC architecture:

  • Sigma-delta (industrial process ADCs) — ENOB usually within 0,5–1 bit of nominal. A 24-bit sigma-delta might genuinely deliver 20–22 ENOB.
  • SAR (successive approximation, general purpose) — ENOB typically 2 bits below nominal at low sample rates, 3–4 bits below at high sample rates.
  • Pipeline (high-speed video/comms) — ENOB usually 2–3 bits below nominal.
  • Flash (very fast, low resolution) — ENOB usually equals nominal because there are few bits total.

The calculator’s ENOB toggle lets you input the datasheet’s effective bits and see the practical step size — often noticeably worse than the marketing headline resolution.

Signal types this calculator handles

  • 4-20 mA — the industrial standard for process instrumentation. Span = 16 mA. “Live zero” at 4 mA distinguishes valid measurement from broken wire (see the 4-20 mA scaling calculator for more).
  • 0-20 mA — legacy or specialty. No live zero, so a dead sensor and true zero are indistinguishable.
  • 0-10 V — common on machine tools, HVAC, and older systems. Vulnerable to voltage drop on long cable runs.
  • 0-5 V — legacy analog audio and some sensor outputs. Rare in modern industrial.
  • ±10 V (bipolar) — the full range covers both polarities. On a signed 16-bit ADC, uses all 65536 codes across -10 to +10 V, giving 20 V ÷ 65536 = 305 µV per step.
  • 1-5 V — the voltage equivalent of 4-20 mA, sometimes used when a device outputs voltage but “live zero” is still needed.

Common design decisions this calculator supports

  • “Is my ADC good enough for the measurement I want?” — enter your target step size, get the required bits, compare to your platform.
  • “Can I switch from Siemens to a Micro850 without losing precision?” — use the comparison mode and see the numbers next to each other.
  • “How much better is the S7-1500 than the S7-1200 for my sensor?” — they use the same 27648 count range, so ADC-wise they’re equivalent. The difference is elsewhere.
  • “My PID loop is oscillating — is it a resolution problem?” — if your engineering-unit step size is more than 0,5% of your setpoint, expect noticeable stepping in the process variable that the PID may amplify.

Practical rules of thumb

  • For simple monitoring: ~100 discrete EU values across the range is usually enough (7-bit equivalent).
  • For control loops: aim for at least 1000 EU values across the range (10-bit equivalent) to avoid visible stepping in the control response.
  • For precision measurement: 10,000+ values (14-bit equivalent) if you’re calibrating or trending small deviations.
  • For calibration and metrology: 100,000+ values (17-bit+ ENOB) if you’re building a reference or verifying another instrument.

What this calculator doesn’t cover

  • Sensor accuracy — the transmitter itself has a nominal accuracy (0,1%, 0,5%, etc.) that’s usually larger than the ADC step size. Sensor accuracy typically dominates for good ADCs on decent industrial ranges.
  • Total measurement uncertainty — the practical uncertainty is a combination of sensor accuracy, ADC step, cable noise, calibration drift, and temperature effects. This calculator only computes the ADC step.
  • Sampling rate — resolution is one axis; how often the ADC samples is the other. A 24-bit ADC sampling at 10 Hz is very different from a 16-bit ADC at 100 kHz — even if the ENOB is similar.
  • Anti-aliasing filters — sample rate must be at least 2× the highest frequency of interest (Nyquist). Aliasing artifacts can look like measurement error but aren’t about resolution.
  • Actual field noise — long cables, ground loops, and EMI can dominate over ADC-level noise. Field practice matters more than datasheet numbers for real measurement.

For real system design, treat this calculator’s output as the ADC contribution to overall uncertainty. Then add sensor accuracy, wiring noise, and any post-processing filter effects to get an honest total error budget.

← All Tools