Derui chamberCalibration & Accuracy
How Often Should a Testing Machine Be Calibrated?
There is no reliable one-size-fits-all interval. Set the initial interval from measurement risk and use, then shorten or extend it using actual drift and verification history.
How should a calibration interval be set?
Begin with the intended use of each measurement channel, not the machine as one undivided object. A fatigue tester may contain force, displacement, speed and cycle-count channels. A dynamometer may add torque, rotational speed, voltage, current and temperature. Each channel can have a different risk and calibration need.
When a standard, customer contract, accreditation requirement or regulator specifies an interval, follow that requirement. Otherwise, establish a documented initial interval from the manufacturer’s information, expected use and the consequence of a wrong result. Review it after enough history is available.
| Factor | Reason to shorten the interval | Evidence that may support extension |
|---|---|---|
| Usage | High cycles, frequent overloads, multiple shifts or continual movement | Light, stable use with controlled handling |
| Environment | Temperature variation, vibration, contamination or humidity | Monitored and stable laboratory conditions |
| History | Drift approaching tolerance or repeated adjustment | Several consecutive as-found results comfortably within limits |
| Measurement risk | Small decision margin or severe consequence of false acceptance | Large margin and low consequence, supported by uncertainty review |
| Events | Repair, relocation, collision, overload, sensor replacement or software change | No disruptive event and satisfactory intermediate checks |
Accuracy, repeatability and uncertainty are not the same
Accuracy
Accuracy describes closeness of agreement between a measured value and the value being measured. In equipment specifications it is often represented through limits such as maximum permissible error, but the exact definition and stated conditions must be read carefully.
Repeatability
Repeatability describes how closely repeated results agree under defined repeatability conditions. A machine can repeat the same biased result very consistently; good repeatability therefore does not by itself establish accuracy.
Resolution
Resolution is the smallest change indicated or displayed by the system. More digits do not automatically mean lower error. Resolution, noise, calibration, alignment and the complete measurement chain all influence usable performance.
Measurement uncertainty
A measurement result is interpreted together with its uncertainty. Relevant contributors can include the reference standard, repeatability, resolution, environmental influence, interpolation, alignment, fixtures and the test method. The acceptable uncertainty depends on the decision the result must support.
What parts of a testing machine need calibration?
Calibrate or verify every measurement or control function that materially affects the reported result or test condition. The exact list depends on the machine and method.
| Channel | Typical reference | Points to consider |
|---|---|---|
| Force | Calibrated force-proving device or reference load cell | Direction, range, loading sequence, alignment and multiple working points |
| Torque | Reference torque transducer or calibrated torque system | Clockwise/counter-clockwise operation, adapters and parasitic loads |
| Displacement | Traceable length standard, indicator or laser system | Range, mounting, cosine error and actual point of measurement |
| Speed/frequency | Reference tachometer, timer or frequency standard | Steady and dynamic operation across normal settings |
| Temperature | Calibrated thermometer or simulator | Sensor position, chamber uniformity and stability |
| Electrical quantities | Calibrated voltage, current or power reference | AC/DC mode, bandwidth, range and complete acquisition chain |
| Time/cycle count | Traceable time reference or independent counter | Long-duration accumulation, stop logic and missed pulses |
Calibrating the sensor alone may not characterize the complete installed channel. Cabling, signal conditioning, controller scaling, software, fixtures and alignment can all change the result. Define whether calibration is performed on a component, an installed channel or the full system.
What should be checked between calibrations?
Intermediate checks are planned comparisons that provide evidence the equipment remains under control. They do not automatically replace calibration. Choose a check that is sensitive to likely failure modes and can be performed consistently by trained staff.
- Zero and span checks with a stable check standard
- Reference specimen or control sample trend
- Cross-check against an independent instrument
- Fixture, alignment, fastener and cable inspection
- Control-chart review for drift or step changes
- Software version and parameter verification
- Environmental-condition review
Set warning and action limits before reviewing results. Record the standard used, environmental conditions, operator, raw result and disposition. A check with no predetermined response cannot reliably protect the measurement process.
How do you review a calibration certificate?
Do not file the certificate based only on the presence of a logo or the word “traceable.” Confirm that it identifies the correct instrument and channel, calibration date, method, environmental conditions where relevant, reference standards, results at useful points and associated measurement uncertainty.
- Does the identification match the equipment, sensor and serial number?
- Are both as-found and as-left results reported when adjustment occurred?
- Do calibration points cover the range actually used?
- Are errors and uncertainty stated clearly with units?
- Is conformity stated, and which specification and decision rule were used?
- Is metrological traceability supported by a documented calibration chain?
- Is the provider competent for the required quantity, range and uncertainty?
NIST emphasizes that metrological traceability is a property of a measurement result and requires a documented unbroken calibration chain, with each link contributing to uncertainty. Traceability alone also does not guarantee fitness for purpose; the uncertainty must be suitable for the measurement need.
What if equipment is out of tolerance?
Quarantine the affected channel and stop using it for reportable results until disposition is complete. Record the as-found condition before adjustment where possible. Then evaluate the potential effect on results produced since the last acceptable evidence—this may be the previous calibration, a satisfactory intermediate check or another justified point.
- Identify affected equipment, ranges, methods and dates.
- Compare the as-found error and uncertainty with actual decision limits.
- Review intermediate checks, control samples and product margins.
- Determine whether reports, customers or product decisions are affected.
- Repair or adjust, recalibrate and document authorization to return to service.
- Review the interval, handling, environment and preventive controls.
The impact review should be proportionate and evidence based. An out-of-tolerance point does not automatically invalidate every previous test, but ignoring it can leave false acceptance or false rejection unaddressed.
Frequently asked questions
Is annual calibration mandatory for every test machine?
No universal rule makes twelve months correct for every machine. It is a common initial interval, but method, contractual, regulatory or accreditation requirements take priority. Otherwise use a documented risk- and history-based interval.
Does calibration include adjustment?
Not necessarily. Calibration establishes the relationship between indications and reference values under specified conditions. Adjustment changes the system. When adjustment is made, retain as-found data and perform the necessary as-left calibration.
Can our laboratory perform calibration internally?
It may, if it has competent personnel, suitable procedures, reference standards, environmental control, uncertainty evaluation, traceability and records appropriate to the intended result. Customer, accreditation or regulatory requirements may still require an external accredited service.
When should calibration be performed after installation?
Confirm measurement performance after installation and before relying on results, especially when transport, assembly, alignment or site conditions can affect the system. Factory calibration alone may not cover the installed measurement chain.
Define calibration before your machine is accepted
Include required quantities, working ranges, tolerances, calibration points, certificate content and factory-acceptance checks in the equipment specification. Send Derui your test method and measurement requirements so the sensing and verification plan can be reviewed with the machine configuration.
Authoritative references
- ISO — ISO/IEC 17025:2017, requirements for competence, impartiality and consistent operation of testing and calibration laboratories.
- NIST Policy on Metrological Traceability, traceability, measurement uncertainty and fitness-for-purpose principles.
- ISO — ISO 10012:2026, measurement management systems and confidence in measurement results.
Always use applicable contractual requirements, current standards and your laboratory’s approved quality procedures. This article provides general planning guidance, not a calibration procedure for a specific machine.
