Systematic Adjustment
Software-based correction of sensor output relies on comparing raw measured values against known standard values stored in non-volatile memory. During this procedure, which is termed digital calibration, correction coefficients are calculated to nullify zero-point offsets and scale errors. This method replaces manual potentiometers with digital registers.
Compensation Algorithm
Polynomial equations and multi-point look-up tables correct non-linear response curves across varying temperatures. Execution of digital calibration involves applying these algorithms in real time to the output of an analog-to-digital converter. This action ensures that the digital output corresponds to the physical parameter.
Reference Medium
Accurate calibration relies on the traceability of the reference sources used during the initial characterization phase. Although digital calibration can correct repeatable errors, it cannot fix random noise or drift that happens after the coefficients are set. Verification against a primary standard must therefore be performed periodically to confirm the integrity of the correction coefficients.
Accuracy Limit
Quantization noise and resolution boundaries of the digital-to-analog converter define the ultimate performance floor of the system. Even with digital calibration, the precision is bounded by the word length of the registers and the noise floor of the analog front end. Fine-grain adjustments require high-resolution processors to avoid introducing rounding errors.
