Table of Contents
Advertisement
Philips Semiconductors
Electronic Compass Design using
KMZ51 and KMZ52
5. SIGNAL CONDITIONING UNIT (SCU)
5.1 Requirements
The SCU consists of two separate "channels" fulfilling the basic task of amplifying the x- and y-field sensor
output voltages (refer to Figure 3). Considering a minimum earth field strength in the sensor plane of
approximately 15 A/m and a sensor sensitivity of typically 80 mV/(kA/m) (at Vcc = 5V, refer to Table 1), an MR
sensor will deliver an amplitude of approximately 1.2 mV, when rotated in that field. Therefore, significant
amplification is required in order to provide reasonable voltages for the following direction determination stage.
Depending on the desired system accuracy, the SCU has to fulfil up to three further requirements. These are
the elimination of the following error sources:
•
Offset voltages Vox, Voy at the SCU output:
These are caused by the offsets of the sensor elements and the connected amplifier. Offset of a sensor
element arises due to the tolerances and temperature drift of the four magnetoresistive elements, which are
arranged as a Wheatstone bridge. Thus, the sensor output voltage deviates from zero, if no magnetic field is
applied.
•
Sensitivity difference ∆S between x- and y- channel of the SCU:
This is due to tolerance and temperature drift of sensor sensitivities and the following amplification.
•
Non-orthogonality β of sensors:
Due to mounting tolerances, the real angular displacement between the sensors deviates from the desired
90° by an angle β.
Equation (3) indicates the effect of these error sources on the azimuth reading. Here it is assumed that the
direction determination unit carries out equation (1) by replacing Hey and Hex with the respective SCU output
voltages Vy and Vx. As each output voltage equals the measured earth´s field component times channel
sensitivity plus offset, the azimuth reading is:
Vy
α
=
=
arctan
Vx
Equation (3) becomes identical to equation (1), i.e. the real azimuth is derived, if Vox, Voy, ∆S and β are
eliminated. The azimuth errors caused by offsets, ∆S and β are periodic functions of α. The amplitudes of these
functions, i.e. the maximum azimuth errors, can be assumed as proportional to the magnitude of the respective
error source. Table 2 in section 10 states the equations for these azimuth errors together with an indication of
their maximum value depending on the respective error source:
•
Offset error: E
offset
i.e. an offset to amplitude ratio of 1% causes a max. azimuth error of 0.8 °.
•
Sensitivity difference error: E
i.e. a ∆S/S ratio of 1% causes an error of 0.3 °.
•
Non-orthogonality error: E
i.e. a 1° deviation from orthogonality between the sensors causes a max. azimuth error of 1°.
The KMZ52 is specified to have an offset voltage of max. ±1.5mV/V and an offset drift of max. ±3µV/(V/K) (refer
to Table 1). Thus, at the recommended supply of 5V, the max. offset voltage at 25°C is ±7.5mV and the max.
offset drift, e.g. over a temperature range of 100°C, is 1.5mV. Comparing these values with the sensor voltage
amplitude of 1.2mV as derived above, it becomes evident, that offset level and even offset temperature drift can
be significantly higher than the desired signal. As a consequence, an efficient in-circuit offset compensation is
crucial for every compass system, even for "low end" products. Section 5.2 shows techniques to realize this.
∆
⋅
+
⋅
He
(
S
S
)
sin(
arctan
⋅
⋅
He
S
cos
≈ 0.8 °/%,
≈ 0.3 °/%,
∆S
≈ 1 °/°,
β
α
β
+
+
)
Voy
α
+
Vox
16
Application Note
(3)
AN00022
Advertisement
Table of Contents
