At the core of every data logger is a sensor with a defined measurement range and stated level of accuracy. Before installation, it is essential to confirm sensor accuracy, calibrate the unit, and document the completed procedure.
Some sensing and measurement devices are subject to drift, i.e., change over time. After one year, a repeat calibration is recommendable to certify that the sensor's accuracy is still within specification.
In-process calibration is a quick and relatively inexpensive method, typically used on-site or in-situ with thermometers. However, it tends not to be as precise as the complete procedure in stabilized conditions.
Essentially, the in-process approach involves comparing the readings of devices under examination with reference devices. By definition, these comparisons take place only at current process temperatures.
Nonetheless, even though the theoretical risk of error is higher than with the fuller procedure, in-process results are usually accurate enough.
As components age, the risk of drift varies according to the sensor type, the measuring principle used, and the degree of physical protection. Nonetheless, suitable quality sensors tend to be stable for many years, whether used in fixed positions or portable applications.
In contrast, the risk for drift in humidity sensors depends on the usage and the sensing technology. As a general rule:
Use capacitive humidity sensors at room temperature and in transport vehicles.
Use electrolytic humidity sensors for high-precision or demanding applications, e.g., in cleanrooms, hot cabinets, and incubators.
If we run a PT100 at its most resistant environmental temperature of +150ºC continually, it’s tested to produce a maximum drift of 0.03% for every 1000 hours. So essentially, at the harshest possible temperature and worst case scenario, the calibration integrity will only ever drift to a maximum of 0.045ºC over 41 days (maximum of 0.4ºC per 365 days). In most cases, this will still be within tolerance at the annual service interval. Considering that typical ambient, refrigerator, and freezer conditions are much less resistant, the accuracy drift reduces substantially.