Table of Contents
Advertisement
Philips Semiconductors
Electronic Compass Design using
KMZ51 and KMZ52
5.5.3 Pre-amplifier (block 2)
The flipped sensor signal is amplified here by a factor of R4/Rbridge, where Rbridge is the resistance of the
sensor bridge. Due to the electro-magnetic feedback used in this circuit, the pre-amp´s output will be virtually
zero, when the closed loop control has settled. However, it should be provided that OP2 is operated in its linear
range during transient processes of the closed loop control, e.g. after turn-on of the system. Thus, the pre-
amplification should be set, such that OP2´s output is in its linear range, when the control loop is interrupted,
even at maximum sensor signal plus max. offsets. An amplification of approximately 100 is a recommendable
value.
5.5.4 Offset compensation (block 3)
The low pass filter around OP3 extracts the offset, which is the dc component of the flipped output signal (refer
to Figure 12) and feeds it as negative feedback to OP2. The sensor offset is compensated this way. Principally,
this could also be done by means of a capacitor between sensor and pre-amp, however the method shown here
is preferable, as it compensates the offsets of both the sensor and OP2 simultaneously. This allows to use a low
cost op-amp instead of a special low-offset type, thus reducing system cost. The determination of the filter cut-
off frequency is a trade-off between smoothing of the output signal and response time.
5.5.5 Synchronous rectifier (block 4)
This block recovers the desired dc signal from the flipped ac signal. Provided that R7=R8=R9, this block
performs an alternating +1 and –1 amplification, depending on the state of switch S1, which is controlled by the
flipping generator. Thus, each time the output of OP2 changes polarity due to the flipping, this block causes an
additional change of polarity, resulting in a rectification.
5.5.6 Integral controller (block 5)
This block forms the integral part of the PI-controller, built together with block 6 to drive the compensation coil
for electro-magnetic feedback. An integral characteristic is required in the control loop, to force the remaining
error - which is the sensor output signal - to zero. Dimensioning the time constant R10·C2 is a trade-off between
response time and smoothness of output signal.
5.5.7 Compensation coil driver (block 6)
Forms the proportional part of the PI controller for electro-magnetic feedback including current source for the
compensation coil. When the control loop has settled, i.e. when the compensation coil generates a field with
equal magnitude and opposite sign to the respective earth´s field component, (Hex or Hey), the output voltage
Vout is:
=
V
H
out
(
, x
) y
( e
, x
) y
where Acomp is the field factor of the compensation coil (refer to the data sheet of KMZ5x). Equation (9) shows
the desired effect of the electro-magnetic feedback: Vout is independent of sensor sensitivity and its
temperature drift (KMZ52: typ. 0.31%/K). The temperature drifts of R12 and Acomp, which now affect the output
voltage are significantly lower. Typical values are 0.02%/K down to 0.005% for standard or precision smd
resistors and 0.01%/K for Acomp of the KMZ52. For ∆S compensation, Vout can be made adjustable by using a
potentiometer for R12. Note, that due to the asymmetrical OPAMPs supply, Vout in (9) is the output voltage
swing relative to Vref.
R
12
⋅
Acomp
(9),
23
Application Note
AN00022
Advertisement
Table of Contents
