Between two calibrations, the instrument keeps working and its components change. Some changes are slow and predictable; others are abrupt. In both cases the instrument can leave its tolerance without giving any sign, since the indication still shows a believable value. That is the main reason for calibrating at intervals and for making intermediate checks.
Slow causes
- Component aging: electronic references, resistors, sensors and elastic elements change their properties over time.
- Creep and material relaxation: in mechanical elements under stress.
- Corrosion and contamination: these alter surfaces, orifices and contacts.
- Thermal cycling: repeated expansion and contraction produce cumulative changes.
Abrupt causes
- Impacts and drops, which can displace mechanisms or damage sensors.
- Overpressure, mechanical overload or overvoltage.
- Exposure to temperature, humidity or vibration outside specification.
- Ingress of liquids or dust.
Causes related to people
Unauthorized adjustments, misuse, improper storage, cleaning with incompatible products or incorrect connections. These cases are reduced with clear procedures and training.
How the risk is detected and limited
- Intermediate checks with a check standard or a second instrument.
- Recording the error found at each calibration to see trends.
- Intervals adapted to the observed behavior.
- Post-event protocols: if there was an impact or overload, verify before continuing.
When it is found out of tolerance
The impact must be assessed: how long it may have been off, which measurements depended on it and whether the affected products or decisions need review. That is why it is so important to have the as-found state, not only the state after adjustment.
What the engineer sees in practice
Usually nobody notices the problem until a calibration reveals an error larger than allowed, or until a product is rejected downstream by a customer or by another measurement. An instrument that gives stable, plausible readings can be biased, and stable readings are not proof of accuracy. Checking from time to time against an independent reference is the cheapest way to find out in time.