Architectural Category
Elimination of internal offset voltages is achieved through a specific signal conditioning architecture. An auto-zeroing front-end operates by periodically sampling the input offset of an amplifier and subtracting it from the active signal path. This technique maintains accuracy over time.
Correction Mechanism
Internal switching circuitry disconnects the input and shorts the amplifier terminals during the nulling phase. The resulting error voltage is stored on a capacitor and used to cancel the offset when the auto-zeroing front-end returns to the signal amplification phase. This process repeats at a high frequency to track changes caused by temperature or aging.
By removing the dc component of the noise, the system achieves sub microvolt precision.
Aliasing Behavior
Periodic sampling folds high frequency noise into the baseband of the instrument. Because the auto-zeroing front-end behaves as a sampled data system, wideband noise above the Nyquist frequency appears as increased white noise at the output. Designers mitigate this by using low pass filters before the sampling stage.
System Constraint
Recovery time after a large signal transient limits the usable bandwidth of the sensor. Switching noise or charge injection from the internal gates can introduce small spikes into the output of the auto-zeroing front-end.