SensorCatalog

Sensor accuracy spec converter

Put a datasheet's "±(% of reading + % FS)", tempco and stability into one number at your reading, in units, % of reading and % FS.

Range and reading

Full scale means

Datasheet specification

Relative terms in

Result

±0.02256bar

Probable error: 1.128 % of reading, 0.2256 % FS. Worst case ±0.033 bar.

Basic accuracy
±0.003 bar0.15 % of reading · 0.03 % FS
Reading term
±0.001 bar0.01 % FS
Full-scale term
±0.002 bar0.1 % of reading
Temperature effect
±0.02 bar20 K from the reference temperature
Long-term stability
±0.01 bar1 yr since calibration
Worst case
±0.033 bar1.65 % of reading · 0.33 % FS
Probable error in ppm
2256 ppm FSof full scale
Full scale
10 barspan of the range
Full-scale term dominates below
4 barwhere both relative terms are equal

Over the range

10 % · 1 bar
±0.0225 bar2.25 % of reading · worst ±0.0325 bar
25 % · 2.5 bar
±0.0226 bar0.904 % of reading · worst ±0.03325 bar
50 % · 5 bar
±0.02281 bar0.456 % of reading · worst ±0.0345 bar
75 % · 7.5 bar
±0.02309 bar0.308 % of reading · worst ±0.03575 bar
100 % · 10 bar
±0.02343 bar0.234 % of reading · worst ±0.037 bar
Formulas and standards
Ebasic = a × |x| + b × FS + c + n × d
Eprobable = √(Ebasic² + Etemp² + Estab²)   Eworst = Ebasic + Etemp + Estab

The basic accuracy written as ±(a % of reading + b % of full scale + c digits) is a sum by definition, so its terms add. Temperature effect (tempco × distance from the reference temperature) and long-term stability are separate effects with their own causes: the probable error combines them with the basic accuracy as a root sum of squares, the worst case adds everything.

Full scale usually means the span of the range; some datasheets for bipolar ranges (±10 V) mean the largest value instead, so choose what yours states. A % of reading term shrinks with the reading while a full-scale term does not, which is why the error in % of reading rises steeply in the lower part of the range. This converter uses only the terms you enter; linearity, hysteresis and repeatability are normally inside the basic accuracy, but check the datasheet's notes.

Need a sensor whose accuracy holds where you measure?

Get pricing from SensorSpan

How the accuracy spec converter works

A datasheet states accuracy in whatever form suits the product: ±0.1 % of full scale for a pressure transmitter, ±(0.05 % of reading + 0.02 % FS) for a precision transmitter, ±(0.5 % + 3 digits) for a meter, ppm for a reference. To compare two sensors, or to put one into an error budget, all of these must become one number at the reading you actually measure.

The converter adds the basic terms, which is how a “±(a + b + c)” specification is defined, then combines the result with the temperature effect and the long-term stability. It shows the error in your unit, in % of reading, in % of full scale and in ppm, both as the probable error and as the worst case.

Why % of full scale hurts at low readings

A full-scale term is the same number of units everywhere in the range. A 10 bar transmitter with ±0.02 % FS has ±0.002 bar at 10 bar and at 1 bar alike, which is 0.02 % of reading at the top and 0.2 % at a tenth of the range. A % of reading term shrinks with the reading. The table under the result shows the total at 10, 25, 50, 75 and 100 % of the range, so you can see where a sensor stops being good enough.

The converter also gives the reading below which the full-scale term is the larger one. For the example ±(0.05 % of reading + 0.02 % FS) on 0 … 10 bar this is 4 bar.

Temperature effect and stability

The temperature effect is a coefficient per kelvin multiplied by how far the sensor is from its reference temperature, usually 20 or 23 °C. A sensor in an outdoor enclosure can be 30 K away for much of the year. Stability is the drift per year since calibration. Both are separate effects with their own causes, which is why the probable error combines them with the basic accuracy as a root sum of squares; the worst case adds everything and is what a guaranteed limit needs.

Full scale: span or largest value

Full scale usually means the span of the range: 10 bar for 0 … 10 bar, 20 V for −10 … 10 V. Some datasheets for bipolar ranges mean the largest value instead, 10 V in that example, which halves every full-scale term. Choose what the datasheet states; when it does not say, the span is the cautious reading.

What the specification may leave out

Read the notes under the accuracy table. Linearity, hysteresis and repeatability are normally part of the basic accuracy; the effect of supply voltage, mounting position and static pressure often is not. A turndown, where a transmitter is set to part of its maximum range, can scale the full-scale term up; the pressure transmitter accuracy calculator handles that case. For a temperature chain, the temperature measurement error calculator combines sensor class, wiring and transmitter.

Frequently asked questions

How do I convert ppm to percent?

1 % is 10 000 ppm, so 50 ppm is 0.005 %. Switch the relative terms to ppm and enter the datasheet’s figures directly.

What does “± 3 digits” mean?

A digit, or count, is one step of the last displayed digit. On a meter showing 0.000 … 9.999 V one digit is 0.001 V, so 3 digits are ±0.003 V. Enter the number of digits and the value of one digit.

Is the probable error a confidence interval?

Only if every term is stated as the same coverage, which datasheets rarely say. Treat it as an engineering estimate of the typical error and the worst case as the limit.

↑ ↓ choose · Enter open · Esc close