Worked example
Example: Flow loop 0–100 t/h at 60 t/h. The inputs, in the units selected above:
| Input | Value |
|---|---|
| Span of the loop, in its own units | 100.00 - |
| Reading, in the same units | 60.000 - |
| Transmitter reference accuracy (share of span) | 0.075000 % |
| Ambient temperature effect (share of span) | 0.050000 % |
| Static pressure or other process effect (share of span) | 0.050000 % |
| Input card or indicator (share of span) | 0.10000 % |
| Other (calibrator, drift) (share of span) | 0.050000 % |
Applying the formula below gives:
| Result | Value |
|---|---|
| Total probable error, root-sum-square (share of span) | 0.15207 % |
| Worst-case error, straight sum (share of span) | 0.32500 % |
| Total probable error, in span units | 0.15207 |
| Total probable error as a share of reading | 0.25345 % |
Open this example in the calculator and change any input.
Formula
\(u = \sqrt{\sum e_i^2}, \quad u_{wc} = \sum e_i\)
\(u S, \quad u S / x\)
Variables
| Symbol | Meaning | SI unit |
|---|---|---|
| \(S\) | span | units |
| \(x\) | reading | units |
| \(e_1\) | transmitter accuracy | - |
| \(e_2\) | ambient effect | - |
| \(e_3\) | static pressure or process effect | - |
| \(e_4\) | input card | - |
| \(e_5\) | other | - |
| \(e_i\) | error term i (share of span) | - |
| \(u\) | root-sum-square error | - |
| \(u_{wc}\) | worst-case error | - |
| \(u S\) | error in span units | units |
| \(u S / x\) | error as a share of reading | - |
Method and standards
Independent random errors combine by root-sum-square; the straight sum is the worst case.
Where it is used
- Error budgets for custody and efficiency-test measurements.
- Comparing transmitter and card choices.
- Answering audit questions on loop accuracy.
Common mistakes
- Adding per-cent-of-reading and per-cent-of-span terms directly.
- Combining errors at different confidence levels.
- Treating systematic errors as random.
Assumptions and limits
- Independent errors at the same confidence level. Makers' accuracy figures are often limits (rectangular) or 2σ values: convert them to one basis (as in the GUM, with a coverage factor) before combining; this tool does not.
References
- Root-sum-square combination of independent errors
Questions and answers
What is the total error of a loop with 0.075 %, 0.05 %, 0.05 %, 0.1 % and 0.05 % terms?
About 0.15 % of span by root-sum-square, against 0.33 % worst case; at 60 % of span that is 0.25 % of reading.
Why root-sum-square?
Independent errors rarely all peak in the same direction at once, so their combination grows as the square root of the sum of squares.
Why does error grow at low readings?
Span-based errors are fixed in units, so as a share of a small reading they grow.
Method and article reviewed by FireCAD engineering on 12 November 2026. Every result is checked against an independent calculation on each release: how we verify our calculators.