Sensor characterisation is done based on the n-point(usually 9) characterisation of the sensor. Characterisation is performed at a specific temperature (25°C) and excitation.
In 9 point characterisation method, humidity levels are swept the through the RH values and measuring the corresponding dc output voltage for the individual sensor: Values are taken at humidity levels of 0%, 25%, 53.2%, 75.3%, 93.8%, 75.3%, 53.2%, 25% and 0%. Based on the characterisation results, Best Fit Straight Line (BFSL) is plotted and sensor characteristics are specified in the datasheets.
Accuracy is specified based on the specific calibration curves for any individual sensor. It is specified using the linear Best Fit Straight Line (BFSL) and the non-linear 2nd order curve.
As an example let us consider a sensor with an accuracy of ±2% RH (BFSL). If the sensor has an output voltage of 0.689 V at 0%RH, an average slope(BFSL) of 0.036 V/%RH and offset of 0.662, then its BFSL accuracy error is given by (0.689 - 0.662)/0.036 = 0.75% RH. As sensors accuracy is ±2% RH (BFSL), i.e. 0.072V, the sensor should always output 0.662 ±0.072 V or a value in the range of 0.59 V to 0.734 V.
· Hysteresis is the difference between the two voltages to %RH conversions (using average BFSL slope) at each of the four duplicated points in the nine point characterization. Hysteresis is recorded in absolute %RH terms.
The value taken is the largest %RH figure for an individual sensor over each of the four characterization points.
Interchangeability defines the range of voltages for any population of sensors at this RH point.
As an example let us consider a sensor from a particular company with an interchangeabilty of ±5% at 0% RH. With an average slope (BFSL) of 0.036 V/%RH and offset of 0.662 V, ±5% RH is equal to ±0.18 V. This means that the output voltage for this device is 0.662 V ±0.18 V, or a range of 0.482V to 0.842 V. When exposed to an RH of 0%, the output of the entire population of sensors will fall within this range.
Linearity indicates the voltage deviation from the BFSL value and the measured output voltage value, converted to RH.
Sensors are subjected to accelerated stress tests. If the tests causes the sensor to drift and report RH outside prescribed specifications, the sensor is considered a failed sensor. Based on such tests, reliability figures like MTTF(Mean-time-to-failure) and FIT (Number of Failures per billion operating hours) are specified.
Repeatability is the maximum variation between sensor outputs for repeated sweeps of humidity levels across the sensors’ measurement range under identical conditions.
For example, if the point value is 0.013 V using the 31 mV/RH slope this is 0.42% RH.
· Response Time
Response Time is measured in “slow moving air” (less than 5 m/s). Typically, maximum time required for the output voltage of the sensor to rise to 63% of its final value or to fall to 37% of its final value when exposed to a step rise or fall in humidity is specified as response time
· Temperature Compensation
Voltage output for an individual sensor at a given excitation and RH is affected by temperature. In many sensors, the temperature is measured and the effect of temperature of humidity measurement is reduced and this is referred to as temperature compensation.
Output voltage stability is the output voltage drift in time at the specified RH level converted to a %RH value.
This figure is also generated through accelerated stress tests and is typically taken as the change in mean output voltage from a large batch of sensors in specific environmental conditions.