Zero-Acceleration Output
Transducer outputs measured under conditions of zero physical acceleration define the baseline offset of the measurement system. The value of the zero-g bias offset must be subtracted from the raw sensor readings to obtain the true acceleration. This baseline error arises from internal mechanical stresses in the sensing element and electronic offsets in the amplifier.
The offset value is determined during factory testing and stored in the sensor’s non-volatile memory.
Thermal Drifting
Changes in the operating temperature of the sensor can cause the baseline offset to shift during operation. For an accelerometer affected by zero-g bias offset, the temperature-dependent drift is modeled and compensated using a polynomial correction curve. The sensor holds a temperature sensor close to the sensing element to provide real-time thermal data.
This compensation maintains the stability of the sensor output across the full operating range.
Calibration Rectification
Precise identification of the offset is achieved by measuring the sensor output in opposing orientations and averaging the results. To calculate the zero-g bias offset, the sensor is rotated 180 degrees in the gravity field, which reverses the sign of the gravity vector while keeping the bias constant. Adding the two measurements together cancels out the gravity signal, leaving twice the bias offset.
This method separates the bias from the scale factor and alignment errors. The calibration must be verified periodically to account for the long-term aging of the sensor materials. The aging rate is measured during high-temperature burn-in tests.
System Startup
Startup initialization routines can also perform a quick calibration if the system can be held stationary for a short period. In applications requiring a fresh zero-g bias offset measurement, the sensor is sampled while the vehicle is at rest to establish the current offset. This on-the-fly correction compensates for any turn-on instability of the electronics.