Most general-purpose test and measurement instruments are calibrated once every 12 months, but the annual default is a starting point, not a rule. The correct interval is the one that keeps the instrument within tolerance for your application between calibrations. It is set from the instrument's stability, how heavily it is used, the environment it works in, and the consequence of an out-of-tolerance reading.
The factors that decide the interval
- Manufacturer recommendation, the baseline (often 12 months).
- Stability / drift history. If past calibrations show little drift, the interval can often be extended; if it drifts, shorten it.
- Usage intensity. A multimeter used daily on a production line ages faster than one used monthly.
- Environment. Heat, humidity, vibration, dust and mechanical shock accelerate drift.
- Criticality / risk. The higher the cost of a wrong measurement (safety, product release, billing), the shorter the interval.
Typical starting intervals
| Instrument type | Common starting interval |
|---|---|
| Digital multimeters, clamp meters | 12 months |
| Process & multifunction calibrators | 12 months |
| Reference thermometers / dry-wells | 12 months (references often 12–24) |
| Humidity & temperature loggers | 12 months |
| Pressure calibrators / gauges | 12 months |
| New or critical instruments | Shorten initially, then review |
These are typical defaults. Always confirm against the manufacturer's specification and your own quality system.
Let the data set the interval
The most defensible approach is to review the "as-found" results at each calibration. If an instrument repeatedly returns well within tolerance, you have evidence to extend the interval. If it returns near or beyond the limit, shorten it. This is exactly what auditors want to see: an interval justified by history, not guesswork.
What happens if you find an instrument out of tolerance
If a calibration shows the instrument was out of tolerance, you have a problem that reaches backwards: every measurement made since the last good calibration is now suspect. This is why critical instruments get shorter intervals. It limits how much work you would have to re-check. A good calibration certificate gives you the "as-found" values so you can assess the impact.
A worked example: setting an interval from real drift data
Take a digital multimeter with a manufacturer tolerance of ±0.5% on its DC voltage range, calibrated annually for three years running. Year one's as-found reading drifts 0.05% from nominal; year two, 0.08%; year three, 0.11%. That is a small, steady, predictable drift, well inside the ±0.5% tolerance band even after three years, so extending this instrument to an 18 or 24-month interval is a defensible, evidence-based decision, and one an auditor will accept because it is backed by three data points, not a guess. Now compare a second multimeter used daily on a factory floor in a hot, vibration-heavy environment: year one drifts 0.15%, year two 0.35%, close to the tolerance limit. That trajectory says the opposite. Shorten the interval to 6 months before the next reading risks landing outside tolerance altogether. The instrument, the environment and the usage pattern are different variables; the calibration history is what actually tells you which way to move the interval, not a blanket house rule applied to every meter in the toolbox.
Special cases that override the drift-history approach
- New or recently repaired instruments: start with a shorter interval (often 3 to 6 months) regardless of the manufacturer's default, until you have enough calibration history of your own to judge its actual stability.
- Reference standards used to calibrate other instruments: these sit at the top of your traceability chain, so an undetected drift cascades into every instrument they were used to check. Reference-grade equipment typically warrants a shorter interval than the field instruments it calibrates, and should always be sent to a lab with a demonstrably tighter uncertainty than the reference itself.
- Instruments that have been dropped, overloaded, or exposed to conditions outside their rated environment: calibrate immediately, independent of the scheduled due date. A single overload event can shift an instrument's calibration permanently, and continuing to trust the last certificate after such an event is a real audit finding waiting to happen.
- Instruments coming off a long idle period: a meter that has sat in a drawer for a year since its last calibration should be re-verified before returning to service, particularly if it will be used for a compliance-critical measurement immediately.
How this fits ISO 9001 and ISO/IEC 17025 expectations
Neither ISO 9001 nor ISO/IEC 17025 mandates a specific interval number. What both expect is a documented, risk-based rationale that you can produce on request, showing how the interval for each instrument class was set and, ideally, reviewed periodically against actual performance. A blanket "we calibrate everything annually" policy with no supporting data is the single most common audit finding in this area; it isn't wrong exactly, but it can't be defended when an assessor asks why. Keeping the as-found/as-left values from every calibration in a simple register, even a spreadsheet, is usually enough to build that defensible history over time. For guidance on reading those values correctly once the certificate arrives, see our companion guide on how to read a calibration certificate.
Build it into a schedule
Track due dates, keep certificates, and recall instruments before they expire. A lapsed calibration on a critical instrument can fail an audit on its own. Unitest Instruments issues recall reminders and can calibrate most common instruments with a fast turnaround at our SAC-SINGLAS accredited lab, and for instruments needed urgently while yours is away for recalibration, short-term rental keeps work moving without a compliance gap.