Front end calibration of MLX90316 three-axis rotary position sensor
“This application note describes the front-end calibration of the MLX90316 three-axis rotary position sensor before the device performs angular position extraction. The document also describes the dynamic offset cancellation mechanism, sensitivity mismatch, orthogonality, signal nonlinearity and the overall accuracy of the system.
This application note describes the front-end calibration of the MLX90316 three-axis rotary position sensor before the device performs angular position extraction. The document also describes the dynamic offset cancellation mechanism, sensitivity mismatch, orthogonality, signal nonlinearity and the overall accuracy of the system.
As shown in Figure 1 and Figure 2, if the magnet (radial magnetization) rotates above the MLX90316, the three-axis Hall plate provides two orthogonal signals Vx and Vy (respectively for the magnetic flux density along the X axis and Y axis) ).
These Hall signals are processed through a fully differential analog chain using classic offset cancellation techniques. The adjusted analog signal is converted by ADC and provided to the DSP module for further processing.
Although the on-chip dynamic offset cancellation mechanism (Hall plate quadrature rotation and chopper stabilized amplifier), the analog signal may show residual offset. The representation of this offset on the sinusoidal signal is shown in Figure 3, where the X0 and Y0 offsets are both amplified (X0 and Y0 are the digital representations of the analog levels V_x,0 and V_y,0). The offset value is usually very small and depends on the temperature.
Although the two Hall signals (V_x and V_y) are both generated by matched Hall plates and amplified by the common multiplexing amplification chain, the two signals may show residual amplitude differences. The two main reasons for this mismatch are the imperfect alignment of the IMC relative to the Hall plate constellation and the difference between the sensitivities of different three-axis Hall plates. An illustration of the amplitude mismatch is shown in Figure 4.
Quadrature error, also called quadrature error, is the phase error between sine and cosine signals. This means that the phase separation of the two signals is not exactly 90 degrees. The sensor’s DSP will continuously adjust the relationship between sine and cosine to obtain a constant phase separation of 90 degrees.
In normal operation, signal nonlinearity is negligible. Its signature is easy to identify; four cycles over 360 degrees. Therefore, it can solve the main source of non-linearity, that is, magnetic saturation (the applied field at the IMC position is greater than 70mT).