4.1 Calibration Metrology, Standards & NIST Traceability
Key Takeaways
Calibration establishes the relationship between reference values and indications with their uncertainties. Record a separate adjustment and subsequent checks if performed.
A percent-of-span limit gives a fixed error magnitude across the specified range. Percent of reading depends on the magnitude of the reading.
Traceability belongs to a measurement result and requires a documented chain to a stated reference with uncertainty at each step.
For the stated two-sided TUR model, divide the tolerance width by twice the expanded uncertainty of the measurement process.
Use the applicable reporting, environmental and decision-rule requirements for the work.
Separate calibration, adjustment and verification
Calibration uses a reference with assigned values and uncertainties to establish a relationship with the measured indications. The relationship can then be used to obtain a result from an indication. See VIM 2.39. An adjustment changes how the instrument responds. A verification evaluates whether it meets a stated requirement. In field work, record the initial observations before adjustment when an as-found result is required. If the device is adjusted, repeat the prescribed checks and keep the as-left results.
Interpret the tolerance given for the device and condition
Accuracy describes how close a measurement is to the quantity value of interest. The International Vocabulary of Metrology treats accuracy as qualitative. In field data sheets, an “accuracy” rating commonly gives error limits under defined conditions. Read the basis of the limit and any additional terms for range, temperature or other influences.
| Stated error limit | How to calculate the magnitude |
|---|---|
| Percent of calibrated span | Multiply the percentage by the difference between the upper and lower range values. |
| Percent of reading | Multiply the percentage by the magnitude of the specified reading. |
| Mixed or multiple-term limit | Use the combination and conditions given in the specification. |
A pressure transmitter with a 0–400 psig range and a symmetric limit of ±0.25% of span has a ±1.0 psi error limit under the stated conditions. At reference values of 400, 40 and 10 psig, that same magnitude is 0.25%, 2.5% and 10% of the respective reference values. For a meter with only a ±0.5% of reading term, the limit is ±1.0 lb/min at 200 lb/min and ±0.1 lb/min at 20 lb/min. A specification may add an offset or a term related to zero stability. Select the range for required performance at the actual operating points and excursions.
- Precision describes the spread of repeated results under stated conditions. A cluster can be tight and still displaced from a reference value.
- Repeatability is precision under a specified common set of conditions over a short time. Reproducibility considers changed conditions such as locations, operators or systems. See the VIM measurement precision entries.
- Linearity describes agreement with a stated straight-line model. An end-point line and a best-fit line can give different deviations.
- Hysteresis is direction-dependent response at a common input after approaching from opposite directions. If a pressure test at 50 psi gives an indication of 50.2 psi on the rise and 49.8 psi on the fall, the separation there is 0.4 psi. Deadband is an input interval over which the expected output change is absent, such as after reversal with mechanical play.
Use a suitable standard and document the reference chain
A primary standard can be established by a primary reference procedure. A secondary standard is established by calibration against a primary standard for the same kind of quantity. A working standard is used routinely for measurements or calibrations. These labels describe the role of the standard in the measurement system. See VIM 5.4, 5.5 and 5.7.
- Pressure. A piston gauge balances pressure against force. In the ideal relation, . Working gauges can have their effective area calibrated by comparison. Use the characterized area and masses, local gravity and applicable corrections. The NIST piston gauge analysis includes buoyancy, pressure distortion, temperature and head terms.
- Temperature. A triple-point-of-water cell realizes an ITS-90 fixed point assigned 273.16 K on that scale. The modern SI kelvin is defined using the fixed Boltzmann constant. Real cells and the best estimate of the thermodynamic triple-point temperature have uncertainty. See BIPM guidance on realizing the kelvin.
A result has metrological traceability when a documented, unbroken calibration chain relates it to a specified reference and each link contributes uncertainty. As an example, a field pressure result may refer through a shop calibrator and a laboratory reference to a realization of the SI pascal. The provider of the result supports the claim. A certificate number or laboratory accreditation alone cannot establish how the instrument was used or whether the uncertainty is suitable. NIST’s policy and FAQ distinguish traces to NIST reference standards from the general concept. Determine the recalibration interval and intermediate checks using equipment use, stability history and risk. ILAC G24 provides guidance on intervals.
Calculate TUR for the actual measurement process
The NASA measurement quality handbook describes the two-sided Z540.3-style test uncertainty ratio. Divide the width of the permitted error band by twice the approximately 95% expanded uncertainty of the calibration measurement process. For a symmetric tolerance of , this simplifies to . Include the relevant uncertainty from the reference and how it is used, such as resolution, drift, repeatability and environment. A data-sheet error limit for a test gauge alone does not supply the entire denominator. When a report states , use the given coverage basis. The relationship between coverage factor and coverage probability depends on the distribution and supporting information. See NIST’s uncertainty guidance.
For a 0–200 psig transmitter with ±0.50% of span as the test limit, psi. A job requiring TUR ≥ 4:1 calls for psi. If the evaluated process uncertainty is 0.75 psi, TUR is about 1.33:1. With a different setup giving 0.15 psi, TUR is about 6.67:1. This answers the capability question. Use the applicable decision rule to determine acceptance. For example, a rule requiring the entire uncertainty interval to fit inside ±1.00 psi gives acceptance limits of ±0.75 psi when psi. False acceptance and false rejection remain possible. The applicable rule controls how that risk is handled. ILAC G8 describes decision rules for statements of conformity.
Record conditions that affect the result
Control and record environmental conditions that matter for the reference, instrument and method. A piston gauge may need a temperature correction. A temperature comparison depends on immersion and stability of the source. Check vibration, drafts, humidity or electrical interference when relevant to a valid result.
Use the applicable site and laboratory reporting rules. Document the item and reference IDs, date, operator, method, necessary conditions, results, any changes, and the applicable tolerance or decision rule for a conformity statement. Record uncertainty and coverage as required for the work. ISO/IEC 17025:2017 addresses the competence and consistent operation of testing and calibration laboratories. Covered U.S. processes under OSHA 1910.119(j)(4)(iv) must document equipment inspections and tests, including the test date, performer, equipment identifier, work description and results.
A DP transmitter with a 250 inH₂O span has a symmetric limit of ±0.5% of span. At an applied reference value of 25 inH₂O, what are the error limit and its size as a percentage of that reference?
±0.125 inH₂O and ±0.5%.
±1.25 inH₂O and ±5.0%.
±1.25 inH₂O and ±0.5%.
±2.50 inH₂O and ±10.0%.
A 0–400 °F transmitter is tested to a symmetric limit of ±0.20% of span. A job specifies a two-sided TUR of at least 4:1. What is the maximum expanded uncertainty U₉₅ for the whole calibration measurement process?
0.80 °F.
0.40 °F.
0.20 °F.
0.05 °F.
Which evidence supports a traceable working piston-gauge measurement?
Calibrated effective area and masses, a documented model and corrections, and a reference chain with uncertainties.
The piston’s nominal diameter alone, treated as exact.
A supervisor’s signature in place of calibration evidence.
A portable enclosure and readable gauge face.
Sections you finish are checked off in the contents.