Calibration Log Template: What to Record

A calibration log records what happened each time an instrument was calibrated: when, by whom, against what reference, what was found, what was left, and who decided the instrument could go back to work. One entry per calibration event, accumulating over the instrument's life.

It is a different thing from the certificate a laboratory sends back, and a different thing again from the calibration register, which is the list of what you own. The certificate says what the laboratory measured. The log says what you did about it.

The Fields

Identity and timing

• Instrument identifier — your own asset number, matching the register and the label on the instrument.
• Date of calibration.
• Performed by — the laboratory, or the named person if calibrated in-house.

Traceability

• Reference standard used, by its own identifier, and its calibration status on the day. This is the field that makes an in-house calibration mean anything: a reference whose own calibration had lapsed does not confer traceability, and without recording which reference was used you cannot demonstrate afterwards that it had not. For work sent to an accredited laboratory this is carried by the laboratory's accreditation. The chain this sits in is covered in metrological traceability.
• Environmental conditions where they materially affect the measurement — temperature and humidity for dimensional or electrical work. Not every calibration needs this; the ones where it matters, it matters a lot.

The measurement itself

• As-found readings. What the instrument was doing when it arrived, before any adjustment.
• As-left readings. What it was doing afterwards. Identical to as-found where nothing was adjusted, and it is worth writing "no adjustment" rather than leaving the field blank so that a blank always means "not recorded".
• Measurement uncertainty for the calibration.
• Tolerance or permissible error the instrument is being judged against.
• Verdict — in tolerance, out of tolerance, or in tolerance with a decision rule applied.

The decision

• Certificate number — the pointer to the laboratory's evidence.
• Accepted by, and the date. Someone decides the instrument is fit to go back into service. Recording who turns a filing exercise into a control.
• Next due date.
• Outcome — returned to service, adjusted and returned, quarantined, withdrawn.

As-Found Is the Field That Earns Its Place

If a log omits one thing, it is as-found, and it is the wrong thing to omit.

As-left tells you the instrument is fine now. As-found is the only record of what it was doing during the months it was actually being used. When an instrument comes back out of tolerance, the question is not whether it is fixed — it is which measurements taken since the last calibration are now in doubt, and how badly. As-found is the input to that assessment. Without it, the honest answer to "how far out was it?" is "we don't know", and the impact assessment has to assume the worst case across the whole interval.

This is also the field that feeds interval review. An instrument found comfortably inside tolerance at four consecutive calibrations is telling you its interval could be longer; one found near its limits each time is telling you the opposite. That signal only exists if as-found was recorded as a value rather than a tick. The methods for reviewing intervals all depend on having it.

Where an instrument does come back out of tolerance, the out-of-tolerance procedure covers the sequence that follows.

A Worked Entry

Abstract field lists are easy to agree with and hard to apply, so here is one entry filled in — a digital micrometer returned from an accredited laboratory, found slightly high, adjusted, and accepted back into service.

The entry that matters here is the as-found row. Without it, this reads as a routine calibration — instrument went out, came back in tolerance, carry on. With it, the record says the micrometer was reading 0.005 mm high when it arrived, against a tolerance of 0.004 mm, and every measurement it took since October 2025 needs considering against that. That is the difference between a log and a filing system, and it is one row.

Two details worth copying. The uncertainty is quoted with its coverage factor rather than as a bare number, because ± 0.002 mm means different things at k = 1 and k = 2. And the acceptance is a named person with a date two days after the calibration — the gap is realistic, and recording it is what makes acceptance a decision rather than an assumption.

The Due Date Is Yours, Not the Laboratory's

Worth stating plainly, because logs are frequently populated from whatever the certificate says. UKAS LAB 5, Edition 5 at section 3.7:

"Any recommendation on the calibration interval must have been agreed with the customer, therefore it shall be clear from the certificate that if a calibration due date is reported it is at the customer's request, e.g. 'Customer requested calibration due date'."

A due date on a returned certificate is there because you asked for it, or it is a recommendation you have not agreed to. Either way the decision is yours and the log is where you record having made it. Copying due dates from certificates without that step is how organisations end up with intervals nobody set.

The same document, at section 4.4, lists what belongs on the calibration label — the UKAS accreditation symbol (with the laboratory's accreditation number), instrument identification, date of calibration, certificate number, and space for the recalibration due date. The log, the label and the certificate should agree; when they do not, the label is usually the one that was never updated.

Paper, Spreadsheet, or System

A calibration log works on paper. Plenty of small workshops run one perfectly well in a ring binder, and there is nothing wrong with that at twenty instruments.

What determines whether it keeps working is not the number of instruments but how often you need to read it across entries. Recording is easy in any format. Answering "show me this instrument's last four as-found results" is easy on paper for one instrument and slow for thirty — and that query is exactly the one interval review depends on. A log that is easy to write into and hard to read out of will be written into faithfully and never read, which is the state most calibration logs are in.

The other pressure is retention: entries accumulate for years and have to remain findable, which how long to keep calibration records covers.

If you are checking whether the certificates coming back to you carry everything they should before you log them, the calibration certificate completeness checker runs through the expected elements, and what a calibration certificate must contain covers them in full.

How CalProof Fits

CalProof keeps each calibration as an entry against the instrument, so the log and the register are the same data rather than two spreadsheets that disagree. Each entry carries the date, provider, measurement uncertainty, the verdict against tolerance and the certificate, with the instrument's full history on one screen. As-found and as-left readings — the values this post argues you should keep, not just the tick — aren't yet a field CalProof captures; until they are, log them alongside the certificate in whatever form you already use.

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

If you are setting this up in a spreadsheet first, the calibration schedule template covers the register and due-date side. To see the shape of the output, the sample audit pack is ungated and downloads as a single PDF.

Sources

• UKAS LAB 5, Edition 5 (January 2025) — Reporting calibration results (sections 3.7, 4.4)

This guide applies to UK manufacturers and laboratories working to ISO 9001 or ISO/IEC 17025. It is general guidance based on published UKAS policy; what your own records need to show depends on your scope, your equipment and your assessor's interpretation. Verify against your certification body or UKAS assessor. This is not legal or compliance advice.