What is CMC in calibration?

A laboratory's accredited scope lists, for each quantity and range, a Calibration and Measurement Capability (CMC). It is the smallest measurement uncertainty the laboratory can achieve during a routine calibration of a near-ideal device under normal conditions. It is a statement about the laboratory, not about any particular customer instrument.

CMC is a floor, not a promise

The uncertainty reported on your certificate can never be smaller than the CMC, and will usually be larger because the instrument under test contributes its own resolution, repeatability, drift and — for many devices — the inability to reach a truly stable reading. A laboratory with a CMC of 20 mK will not give a 20 mK uncertainty for a 0.1 °C-resolution digital thermometer.

How CMC differs from accuracy

• Accuracy is a qualitative closeness-to-true-value concept; manufacturers often use it for a specified maximum error.
• CMC is a quantified expanded uncertainty (typically k = 2) tied to an accredited scope.
• Accuracy describes the instrument; CMC describes the laboratory's capability.

Using CMC when choosing a laboratory

Compare the required uncertainty for your application with the laboratory's CMC for the right quantity, range and, where relevant, medium and method. Ask how the reported uncertainty will be built up for your device. The CMC alone does not tell you the laboratory can calibrate your specific instrument to the uncertainty you need.

Reading a scope entry

A CMC is meaningless without the row it sits in. Accredited scopes list, for each quantity, the range, the method or medium, and the capability — and all four have to match your job before the number applies to you.

Note the fourth row. A data logger is an indicating instrument calibrated in a chamber, and its CMC here is more than ten times the bath figure in the first row. A supplier quoting 0.015 °C from their scope for a logger job is quoting the wrong row, and the distinction is the medium rather than the laboratory's skill.

Why your number is always bigger

The CMC is defined for a near-ideal artefact: a device that contributes almost nothing itself. Your instrument contributes its resolution, its repeatability, its drift, its response time, and for many devices these dominate. A 0.1 °C logger contributes 0.029 °C from resolution alone, which is already larger than the whole chamber CMC in the table above.

ILAC-P14 states the rule: the reported uncertainty must not be smaller than the CMC, and it normally exceeds it. A laboratory reporting exactly the CMC for every instrument it touches has probably not included the customer's device in the budget at all, and that is worth asking about directly.

Using CMC to choose a laboratory

• Start from your tolerance, not from the laboratory's brochure. Decide what uncertainty you need — a quarter of the tolerance is the conventional target — and then look for a scope that can plausibly deliver it once your instrument is added.
• Match the row: quantity, range, medium and method. A scope covering your range in a bath does not cover your logger in a chamber.
• Ask how the reported uncertainty will be built up for your device. A laboratory that can answer that question in a sentence is one that actually builds a budget per job.
• Check the accreditation is current, on the accreditation body's website rather than on the certificate template. Symbols outlive suspensions.
• Compare like with like. A CMC at 0 °C is not the CMC at 200 °C, and scopes quote the best figure prominently.

CMC is an expanded uncertainty tied to an accredited scope, usually at k = 2. Accuracy on a manufacturer datasheet is a maximum permissible error under stated conditions. They are different kinds of number and comparing them directly produces nonsense in either direction.