Calibration Standards and Reference Equipment

"Calibration standards" means two quite different things depending on who is saying it, which makes searching for guidance frustrating. To a quality auditor it usually means the documentary standards — ISO 9001, ISO/IEC 17025. To anyone actually performing calibrations it means the physical references: the gauge blocks, reference weights, standard resistors and reference thermometers that a calibration is performed against.

This guide is about the second kind. Specifically: how to choose them, how their uncertainty flows into yours, and how to keep them traceable over years rather than months. If you want the documentary side, ISO 9001 Clause 7.1.5 and the ISO/IEC 17025 requirements cover it, and metrological traceability covers the chain these references sit in.

Reference equipment is worth treating as its own management problem because it fails differently from working equipment. A worn working gauge produces one bad measurement. A drifted reference quietly invalidates everything calibrated against it since its last certificate.

The Hierarchy, and Why It Exists

Most organisations end up with a two- or three-level arrangement, whether or not they name it:

• Reference standards — the most accurate items you hold for a quantity. Calibrated externally by an accredited laboratory or a national measurement institute. Used as little as possible.
• Working standards — calibrated in-house against the reference. Used for routine checks on production equipment.
• Process and production equipment — what people actually pick up: the shop-floor micrometers, gauges and indicators.

The split exists to protect the reference. Every level below inherits its traceability, so wear, handling damage or an out-of-tolerance event on the reference propagates downward across everything and every date since its last calibration. Keeping it in a controlled environment, used by named people, is the cheapest insurance available.

Worth being clear that "reference" and "working" describe a role, not a grade. The same gauge block set is a reference standard in a small workshop and a working standard in a calibration laboratory. What makes something a reference is that you have chosen to treat it as the top of your chain and calibrate it externally.

If you calibrate working standards against your reference yourself, that is in-house calibration and the conditions in in-house vs outsourced calibration apply to it.

Your Reference's Uncertainty Becomes the Floor of Yours

This is the part that changes purchasing decisions. The uncertainty on your reference's certificate does not stay on that certificate — it enters every uncertainty budget built on it, as an imported Type B contribution. M3003 section 5.2(a) lists exactly this: "For measuring instruments — the imported uncertainties associated with their calibration and any drift or instability in their values or readings."

Because contributions combine in root-sum-square, the largest term dominates. A worked budget in how to build an uncertainty budget shows the reference's own calibration contributing roughly 39% of the total on its own — the single largest line, contributing over five times repeatability's share of variance (roughly 2.4 times on a standard-uncertainty basis).

The practical consequence: you cannot measure your way below your reference's uncertainty. If your reference carries U = 0.06 °C and you need to report results at 0.05 °C, no amount of repeat readings or environmental control will get you there. You need a better reference. Teams routinely try to solve a too-large uncertainty by taking more readings, which shrinks the one contribution that was already negligible.

This also feeds the conformity question. A reference with a large uncertainty widens the guard band you need, so marginal parts start failing — see decision rules and guard banding for how that arithmetic works. Reference quality is a commercial decision, not just a metrological one.

Two things to check on a reference's certificate before relying on it:

• The coverage factor. An uncertainty with no stated k cannot be converted to a standard uncertainty, so it cannot enter a budget properly. The calibration certificate guide covers what else must be present.
• The range and points covered. A certificate calibrated at three points does not give you traceability across the whole span if you use the reference across the whole span.

Choosing Where to Get It Calibrated

UKAS policy TPS 41 section 2.1 gives two acceptable primary sources for reference standards: a national measurement institute (or designated institute) whose service is covered by the CIPM MRA, or a calibration laboratory accredited for those services by an accreditation body in the ILAC Arrangement. In the UK that means NPL or a UKAS-accredited laboratory — though as covered in the in-house guide, an equivalently accredited laboratory in another ILAC signatory country also qualifies.

Where neither route is available for the quantity you need, TPS 41 allows other sources, but with two hard constraints that are worth quoting because they get ignored. Section 2.2.3:

"The choices in 2.2.1 and 2.2.2 shall not be made on purely economic grounds (such as the cost of using a particular organisation) or on logistical grounds (such as the need to employ an overseas organisation to obtain traceability). It is a last resort if other routes are unavailable."

So "the accredited lab was more expensive" is explicitly not a reason. And where such a source is used, section 2.2.4 requires the organisation to "ensure that appropriate evidence for claimed traceability and measurement uncertainty is available", listing what that evidence may include:

Section 2.2.5 adds a scope limit worth knowing: evidence of traceability "will only be accepted for the specific procedures and quantities audited", because traceability for the organisation's other services has not been assessed. A supplier audit for one quantity does not bless their whole catalogue.

Reading that list, the accredited route is usually cheaper once your own time is counted. Assembling and maintaining that evidence on a supplier is more work than the price difference on a certificate.

Keeping It Traceable Over Time

A reference standard is not calibrated once. TPS 41 section 3.1 requires an ongoing programme: "In order to maintain traceability on a continuous basis, reference standards and measuring equipment shall be subject to further calibrations on an ongoing basis, hence the establishment of a calibration programme is necessary."

It then lists the factors the interval depends on — "including but not limited to":

Factor (d) is the one that earns you longer intervals. If you cross-check the reference against another item — or run intermediate checks between full calibrations — you have evidence of stability between certificates, which supports extending the interval. Without any intermediate checking, a longer interval is a longer period of no information.

Factor (f) mentioning transportation is not incidental either. A reference sent away for calibration travels both ways; a check on return is how you find out whether the courier was careful.

TPS 41 section 3.3 points to ILAC-G24 for interval detail, which is the same guidance behind determining calibration intervals. The free Calibration Interval Recommender gives a starting point per equipment type.

The Review Almost Everyone Skips

When a reference comes back with a fresh certificate, the temptation is to file it and put the item straight back on the shelf. TPS 41 section 3.2 asks for something more:

"When a fresh calibration has been obtained the data provided should be reviewed in order to confirm that the declared performance is still met. It may be necessary to reconsider the calibration interval or the suitability of the equipment in accordance with the outcome of such a review."

Two decisions are being asked for. Does the equipment still meet its declared performance — and if the answer is marginal, is the interval still right, or is the item still suitable at all? A reference that has drifted a little at each of the last three calibrations is telling you something, but only if someone compares the three certificates.

For accredited calibration laboratories, the same section adds a consequence with teeth: "in the case of a calibration laboratory, the CMCs may be affected; in such cases UKAS shall be informed." If your reference has degraded, the capability you publish may no longer be supportable.

There is also the backward-looking question, which TPS 41 does not raise here but which follows: if a reference is found outside its expected performance, what about everything calibrated against it since its last certificate? That is the same impact assessment as any out-of-tolerance event, just with a much wider blast radius.

Register Discipline That Makes This Workable

None of the above is difficult; it is just easy to lose track of. Four things worth having in place:

1. References are in the same register as everything else, with their own due dates. A reference tracked separately — or on someone's calendar — is a reference that goes overdue unnoticed.
2. The link from each in-house calibration to the reference used is recorded. This is the single most valuable field, because it is what lets you answer "what was affected?" when a reference turns out to have drifted. Reconstructing it later from dates and memory is close to impossible.
3. Certificates are kept for the whole history, not just the current one. Interval review under §3.1(b) and performance review under §3.2 both need the run of past results, not the latest number.
4. Intermediate checks are recorded as events, so the evidence supporting a longer interval actually exists when you claim it.

The Short Version

• "Calibration standards" means documentary standards or physical references — this guide is the physical sense.
• Keep a hierarchy: reference (externally calibrated, used sparingly) → working standards → production equipment. "Reference" is a role, not a grade.
• Your reference's uncertainty is an imported Type B contribution (M3003 §5.2(a)) and usually the largest line in the budget. You cannot measure below it — more readings will not help.
• Primary acceptable sources: an NMI covered by the CIPM MRA or an ILAC-Arrangement-accredited laboratory (TPS 41 §2.1).
• Non-accredited sources are a last resort and explicitly not a cost or logistics decision (§2.2.3), requiring appropriate evidence of traceability and uncertainty — §2.2.4 gives seven examples, expressly "not limited to" them — accepted only for the specific procedures and quantities audited (§2.2.5).
• Intervals depend on the factors at §3.1, which UKAS does not present as a closed list — intermediate checks (d) are what justify a longer one.
• Review the data when a reference returns (§3.2): performance still met? interval still right? item still suitable? Accredited labs must tell UKAS if CMCs are affected.
• Record which reference each in-house calibration used, or you cannot scope the damage when one drifts.

How CalProof Fits

Field spreadsheets almost never carry a reference standard's own calibration history in a form you can check quickly — and that history is what an assessor asks for, and what you need in a hurry when a reference comes back out of tolerance.

CalProof holds reference standards as instruments in the same register as everything else, with their own intervals, due dates and full certificate history, so a reference's own record is never buried in a separate spreadsheet or folder. A reference approaching its due date shows up in the same recall view as everything else, before it lapses.

GBP pricing from £29 a month. No per-user fees on Pro and above. UK data hosting. No long-term contract.

To see the shape of the output, the sample audit pack is ungated and downloads as a single PDF.

Sources

• UKAS TPS 41 Edition 6, December 2022 — UKAS policy on metrological traceability
• UKAS M3003 Edition 6, March 2024 — The expression of uncertainty and confidence in measurement
• BIPM — CIPM Mutual Recognition Arrangement (CIPM MRA)
• NPL — National Physical Laboratory

This guide summarises UKAS TPS 41 Edition 6 (December 2022) and UKAS M3003 Edition 6 (March 2024) as published; section references are to those editions. TPS 41 is UKAS policy addressed to accredited organisations — if you hold no accreditation its requirements do not bind you, though its criteria remain a useful benchmark for managing reference equipment. Verify against the current published editions and your UKAS assessor or certification body for requirements specific to your scope. This is not legal or compliance advice.