What is Calibration Uncertainty

How to calculate calibration uncertainty?

Calculate calibration uncertainty per the JCGM 100:2008 and ILAC P14 following the instructions below.

1. Find the CMC Uncertainty from the Scope of Accreditation.
2. Find the UUT resolution (i.e. smallest incremental change).
3. Determine the UUT repeatability (i.e. standard deviation of repeated measurements).
4. Convert each uncertainty contributor to a standard uncertainty.
   1. CMC Uncertainty: Divide uncertainty by k=2.
   2. UUT Resolution: Divide uncertainty by 2√3.
   3. UUT Repeatability: Divide uncertainty by √n (i.e. observations in final result).
5. Calculate the combined standard uncertainty.
   1. Square each uncertainty (i.e. convert to variance).
   2. Add together the squared uncertainties (i.e. combined variance).
   3. Calculate the square root of the combined variance.
6. Determine the coverage factor (k) for a 95.45 % confidence interval.
   1. Use k=2 from JCGM 100:2008 Table G.1, or
   2. Use effective degrees of freedom and JCGM 100:2008 Table G.2.
7. Calculate the expanded uncertainty – Multiple the coverage factor and the combined standard uncertainty.

How to round uncertainty to two significant digits?

Round uncertainty to two significant digits per JCGM 100:2008 and ILAC P14 following the instructions below.

1. Review the numerical value of measurement uncertainty.
2. Find the first significant digit (i.e. first non-zero number).
3. Find the third significant digit (i.e. two numbers immediately to the right).
4. Round the second significant digit using an appropriate method (JCGM 100:2008 section 7).
   1. Conventional rounding (i.e. Round up n ≥ 5, Round down n < 5)
   2. Round up to the next larger number.
   3. Round down to the next smaller number.
   4. Round to nearest even digit.

What is difference between CMC Uncertainty and Calibration Uncertainty?

CMC Uncertainty is reported in the scope of accreditation and represents the laboratory’s measurement capability offered to customers under typical conditions while calibration uncertainty is reported in the calibration report and represents the expanded measurement uncertainty associated with the reported result.

The CMC Uncertainty in the scope of accreditation can represent the calibration uncertainty for a best existing device per ILAC P14 section 5.3. However, the actual UUT resolution and reproducibility must replace the best existing device at the time of calibration and be reported in the calibration certificate. Otherwise, contributions from the best existing device can be omitted from the CMC uncertainty so long as it is disclosed in the scope of accreditation (typically with a footnote).

What is CMC Uncertainty?

According to the proceedings from the 96^th meeting of the CIPM (2007), CMC uncertainty is a calibration and measurement capability available to customers under normal conditions:

1. As published in the BIPM key comparison database (KCDB) of the CIPM MRA; or
2. As described in the laboratory’s scope of accreditation granted by an ILAC-MRA signatory.

 
The term was approved by the ILAC General Assembly in October 2007 according to the proceedings from the 96^th meeting of the CIPM and is considered identical to the previously used term “Best Measurement Capability” or BMC.

For more information, refer to ILAC P14 policy, appendix A.

What is Test Uncertainty Ratio for Calibration?

Test uncertainty ratio is ratio of the tolerance or specification interval in relation to the expanded uncertainty interval associated with a measurement process or value.

It is defined in the ANSI Z540.3 Handbook. Additionally, an identical term “measurement capability index” is defined in the JCGM 100 106:2012.

It is commonly used to evaluate the measurement risk or suitability of a measurement process. However, it does not take into account the specific risk (e.g. PFA or PFR) associated with a measurement result.

Common test uncertainty ratios recommended by many standard methods and guidance documents are between 3:1 and 5:1. The most commonly accepted or referenced test uncertainty ratio is a 4:1 ratio.

It should be noted that calibration labs are not allowed to report accredited results with a TUR ratio less than 1:1.