What should a plant do when an instrument is found out of tolerance?
An out of tolerance instrument is more than a calibration backlog item. It may mean past measurements, alarms, controls or quality decisions were based on inaccurate information. The response must distinguish an actual instrument error from a test setup problem and evaluate the consequences of the as found condition.
Preserve the as found evidence
Before adjustment, record the instrument tag, serial number, range, process connection, test points, indicated values, reference values, tolerance, uncertainty where relevant, ambient conditions and test equipment identity. Confirm the reference equipment's calibration status and suitability. If the as found result may reflect an incorrect hookup, range, scaling or test method, resolve that uncertainty before assigning a failure cause.
Determine the instrument's operational role
Identify whether the instrument is used for indication, control, alarming, protective trips, emissions monitoring, custody transfer or product quality. Map the signal to the PLC or DCS, historian and any downstream calculations. Check redundancy, voting arrangements, bypasses and compensating indications. An error in a local display and an error in a safety critical trip input have different consequences and approval needs.
Assess the lookback period and impact
Use the last known acceptable calibration, drift history, operating trends, maintenance events and independent measurements to estimate when the error may have developed. Do not assume the device became inaccurate only on the calibration date. Review affected operating decisions and records with operations, quality, environmental or engineering personnel as applicable. If the impact cannot be bounded, record the uncertainty and escalate rather than declaring all prior data valid.
Restore the measurement chain
- Diagnose: inspect impulse lines, wiring, grounding, sensor condition, transmitter configuration and scaling as relevant.
- Correct: repair, replace or adjust using an approved procedure and appropriate isolation.
- Verify: record as left results across the required range and test the end to end signal where the risk warrants it.
- Reconcile: update CMMS history, calibration certificate, control system settings and any affected compliance records.
- Prevent: consider interval changes or design improvements only after understanding the failure mechanism.
Practical example: differential pressure transmitter
A differential pressure transmitter reads high at several test points. The team first confirms the calibrator and manifold lineup, then checks for a partially plugged impulse leg and compares with a redundant measurement. Engineering evaluates whether the error could have affected flow control or an alarm. After corrective work, the as left test and loop verification are attached to the asset record, along with the review of earlier operating data.
Distinguish traceability from fitness for use
NIST explains that metrological traceability is a property of a measurement result supported by a documented calibration chain and uncertainty. It does not by itself guarantee that the result is accurate enough for a specific application. A plant should choose tolerances and acceptance criteria based on the instrument's actual function and risk, not merely the availability of a calibration sticker.
Explore KANE consulting
Discuss practical maintenance and reliability improvements for your facility.
Request a consultation →Related answers: instrument calibration records · management of change handover
Technical references
NIST Metrological Traceability Policy · OSHA Process Safety Management, 29 CFR 1910.119