Philips Stud Sensor Magnetoresistive Sensor User Manual

DISCRETE SEMICONDUCTORS  
General  
Magnetoresistive sensors for  
magnetic field measurement  
2000 Sep 06  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
The KMZ range of magnetoresistive sensors is  
characterized by high sensitivity in the detection of  
magnetic fields, a wide operating temperature range, a low  
and stable offset and low sensitivity to mechanical stress.  
They therefore provide an excellent means of measuring  
both linear and angular displacement under extreme  
environmental conditions, because their very high  
sensitivity means that a fairly small movement of actuating  
components in, for example, cars or machinery (gear  
wheels, metal rods, cogs, cams, etc.) can create  
measurable changes in magnetic field. Other applications  
for magnetoresistive sensors include rotational speed  
measurement and current measurement.  
2
handbook, halfpage  
R = R  
R cos  
α
0
0
Permalloy  
H
α
Current  
I
MLC127  
Examples where their properties can be put to good effect  
can be found in automotive applications, such as wheel  
speed sensors for ABS and motor management systems  
and position sensors for chassis position, throttle and  
pedal position measurement. Other examples include  
instrumentation and control equipment, which often  
require position sensors capable of detecting  
Fig.2 The magnetoresistive effect in permalloy.  
displacements in the region of tenths of a millimetre (or  
even less), and in electronic ignition systems, which must  
be able to determine the angular position of an internal  
combustion engine with great accuracy.  
Figure 2 shows a strip of ferromagnetic material, called  
permalloy (19% Fe, 81% Ni). Assume that, when no  
external magnetic field is present, the permalloy has an  
internal magnetization vector parallel to the current flow  
(shown to flow through the permalloy from left to right).  
If an external magnetic field H is applied, parallel to the  
plane of the permalloy but perpendicular to the current  
flow, the internal magnetization vector of the permalloy will  
rotate around an angle α. As a result, the resistance of R  
of the permalloy will change as a function of the rotation  
angle α, as given by:  
Finally, because of their high sensitivity, magnetoresistive  
sensors can measure very weak magnetic fields and are  
thus ideal for application in electronic compasses, earth  
field correction and traffic detection.  
If the KMZ sensors are to be used to maximum advantage,  
however, it is important to have a clear understanding of  
their operating principles and characteristics, and how  
their behaviour may be affected by external influences and  
by their magnetic history.  
R = RO + ROcos2α  
(1)  
Operating principles  
Ro and Ro are material parameters and to achieve  
Magnetoresistive (MR) sensors make use of the  
magnetoresistive effect, the property of a current-carrying  
magnetic material to change its resistivity in the presence  
of an external magnetic field (the common units used for  
magnetic fields are given in Table 1).  
optimum sensor characteristics Philips use Fe19Ni81,  
which has a high Ro value and low magnetostriction. With  
this material, Ro is of the order of 3%. For more  
information on materials, see Appendix 1.  
It is obvious from this quadratic equation, that the  
resistance/magnetic field characteristic is non-linear and in  
addition, each value of R is not necessarily associated  
with a unique value of H (see Fig.3). For more details on  
the essentials of the magnetoresistive effect, please refer  
to the Section “Further information for advanced users”  
later in this chapter or Appendix 1, which examines the MR  
effect in detail.  
Table 1 Common magnetic units  
1 kA/m = 1.25 mTesla (in air)  
1 mT = 10 Gauss  
The basic operating principle of an MR sensor is shown in  
Fig.2.  
2000 Sep 06  
3
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
In this basic form, the MR effect can be used effectively for  
angular measurement and some rotational speed  
measurements, which do not require linearization of the  
sensor characteristic.  
R
handbook, halfpage  
In the KMZ series of sensors, four permalloy strips are  
arranged in a meander fashion on the silicon (Fig.4 shows  
one example, of the pattern on a KMZ10). They are  
connected in a Wheatstone bridge configuration, which  
has a number of advantages:  
H
MLC128  
Reduction of temperature drift  
Fig.3 The resistance of the permalloy as a  
function of the external field.  
Doubling of the signal output  
The sensor can be aligned at the factory.  
MBC930  
Fig.4 KMZ10 chip structure.  
2000 Sep 06  
4
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
Two further resistors, RT, are included, as shown in Fig.5.  
These are for trimming sensor offset down to (almost) zero  
during the production process.  
For some applications however, the MR effect can be used  
to its best advantage when the sensor output  
characteristic has been linearized. These applications  
include:  
Weak field measurements, such as compass  
applications and traffic detection;  
Current measurement; and  
MLC129  
Rotational speed measurement.  
handbook, halfpage  
For an explanation of how the characteristic is linearized,  
please refer to the Section “Further information for  
advanced users” later in this chapter.  
Philips magnetoresistive sensors  
R
R
T
T
Based on the principles described, Philips has a family of  
basic magnetoresistive sensors. The main characteristics  
of the KMZ sensors are given in Table 2.  
4
3
2
1
V
V
V
O
GND  
CC  
O
Fig.5 Bridge configuration with offset trimmed to  
zero, by resistors RT.  
Table 2 Main characteristics of Philips sensors  
SENSITIVITY  
FIELD  
RANGE  
(kA/m)(1)  
LINEARIZE  
MR  
EFFECT  
SENSOR  
TYPE  
VCC  
(V)  
Rbridge  
(k)  
APPLICATION  
EXAMPLES  
PACKAGE  
(mV V)  
---------------------  
(kA m)  
KMZ10A  
KMZ10A1(2)  
KMZ10B  
KMZ10C  
SOT195  
SOT195  
SOT195  
SOT195  
0.5 to +0.5  
0.05 to +0.05 9  
2.0 to +2.0  
7.5 to +7.5  
9  
16.0  
22.0  
4.0  
1.2  
1.3  
2.1  
1.4  
Yes  
Yes  
Yes  
Yes  
compass, navigation, metal  
detection, traffic control  
12  
10  
current measurement,  
angular and linear position,  
reference mark detection,  
wheel speed  
1.5  
KMZ51  
KMZ52  
SO8  
0.2 to +0.2  
0.2 to +0.2  
8  
8  
16.0  
16.0  
2.0  
2.0  
Yes  
Yes  
compass, navigation, metal  
detection, traffic control  
SO16  
Notes  
1. In air, 1 kA/m corresponds to 1.25 mT.  
2. Data given for operation with switched auxiliary field.  
2000 Sep 06  
5
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
Flipping  
The field (e.g. ‘Hx’) needed to flip the sensor  
magnetization, and hence the characteristic, depends on  
the magnitude of the transverse field ‘Hy’: the greater the  
field ‘Hy’, the smaller the field ‘Hx’. This follows naturally,  
since the greater the field ‘Hy’, the closer the  
magnetization's rotation approaches 90°, and hence the  
easier it will be to flip it into a corresponding stable position  
in the ‘x’ direction.  
The internal magnetization of the sensor strips has two  
stable positions. So, if for any reason the sensor is  
influenced by a powerful magnetic field opposing the  
internal aligning field, the magnetization may flip from one  
position to the other, and the strips become magnetized in  
the opposite direction (from, for example, the ‘+x’ to the  
x’ direction). As demonstrated in Fig.6, this can lead to  
drastic changes in sensor characteristics.  
Looking at the curve in Fig.7 where Hy = 0.5 kA/m, for  
such a low transverse field the sensor characteristic is  
stable for all positive values of Hx and a reverse field of  
1 kA/m is required before flipping occurs. At Hy = 2 kA/m  
however, the sensor will flip even at smaller values of ‘Hx’  
(at approximately 0.5 kA/m).  
MLC130  
handbook, halfpage  
V
O
(mV)  
10  
0
4
2
2
4
H
(kA/m)  
y
10  
reversal  
of sensor  
characteristics  
Fig.6 Sensor characteristics.  
MLC131  
V
O
(mV)  
100  
H
=
y
2 kA/m  
50  
0.5 kA/m  
3
0
3
2
1
1
2
H
(kA/m)  
x
50  
100  
Fig.7 Sensor output ‘Vo’ as a function of the auxiliary field ‘Hx’ for several values of transverse field ‘Hy’.  
2000 Sep 06  
6
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
Figure 7 also shows that the flipping itself is not  
instantaneous, because not all the permalloy strips flip at  
the same rate. In addition, it illustrates the hysteresis effect  
exhibited by the sensor. For more information on sensor  
flipping, see Appendix 2 of this chapter.  
MBB897  
3
handbook, halfpage  
R
bridge  
(k)  
Effect of temperature on behaviour  
Figure 8 shows that the bridge resistance increases  
linearly with temperature, due to the bridge resistors’  
temperature dependency (i.e. the permalloy) for a typical  
KMZ10B sensor. The data sheets show also the spread in  
this variation due to manufacturing tolerances and this  
should be taken into account when incorporating the  
sensors into practical circuits.  
2
In addition to the bridge resistance, the sensitivity also  
varies with temperature. This can be seen from Fig.9,  
which plots output voltage against transverse field ‘Hy’ for  
various temperatures. Figure 9 shows that sensitivity falls  
with increasing temperature (actual values for given for  
every sensor in the datasheets). The reason for this is  
rather complex and is related to the energy-band structure  
of the permalloy strips.  
1
40  
0
40  
80  
120  
T
160  
o
( C)  
amb  
Fig.8 Bridge resistance of a KMZ10B sensor as  
a function of ambient temperature.  
2000 Sep 06  
7
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
MLC134  
15  
V
O
o
T
=
25 C  
amb  
(mV/V)  
o
25 C  
10  
o
75 C  
5
0
o
125 C  
5
10  
15  
operating range  
3
2
1
0
1
2
3
H
(kA/m)  
y
Fig.9 Output voltage ‘Vo’ as a fraction of the supply voltage of a KMZ10B sensor as a function of transverse field  
‘Hy’ for several temperatures.  
2000 Sep 06  
8
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
Figure 10 is similar to Fig.9, but with the sensor powered  
by a constant current supply. Figure 10 shows that, in this  
case, the temperature dependency of sensitivity is  
significantly reduced. This is a direct result of the increase  
in bridge resistance with temperature (see Fig.8), which  
partly compensates the fall in sensitivity by increasing the  
voltage across the bridge and hence the output voltage.  
Figure 8 demonstrates therefore the advantage of  
operating with constant current.  
MLC135  
75  
o
T
=
25 C  
amb  
V
O
o
25 C  
(mV/V)  
50  
o
75 C  
o
125 C  
25  
0
25  
50  
75  
operating range  
4
2
0
2
4
H
(kA/m)  
y
Fig.10 Output voltage ‘Vo’ of a KMZ10B sensor as a function of transverse field ‘Hy’ for several temperatures.  
2000 Sep 06  
9
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
Using magnetoresistive sensors  
linear characteristics is required for compensation. Philips  
KTY series are well suited for this purpose, as their  
positive Temperature Coefficient (TC) matches well with  
the negative TC of the MR sensor. The degree of  
compensation can be controlled with the two resistors R7  
and R8 and special op-amps, with very low offset and  
temperature drift, should be used to ensure compensation  
is constant over large temperature ranges.  
The excellent properties of the KMZ magnetoresistive  
sensors, including their high sensitivity, low and stable  
offset, wide operating temperature and frequency ranges  
and ruggedness, make them highly suitable for use in a  
wide range of automotive, industrial and other  
applications. These are looked at in more detail in other  
chapters in this book; some general practical points about  
using MR sensors are briefly described below.  
Please refer to part 2 of this book for more information on  
the KTY temperature sensors; see also the Section  
“Further information for advanced users” later in this  
chapter for a more detailed description of temperature  
compensation using these sensors.  
ANALOG APPLICATION CIRCUITRY  
In many magnetoresistive sensor applications where  
analog signals are measured (in measuring angular  
position, linear position or current measurement, for  
example), a good application circuit should allow for  
sensor offset and sensitivity adjustment. Also, as the  
sensitivity of many magnetic field sensors has a drift with  
temperature, this also needs compensation. A basic circuit  
is shown in Fig.11.  
USING MAGNETORESISTIVE SENSORS WITH A COMPENSATION  
COIL  
For general magnetic field or current measurements it is  
useful to apply the ‘null-field’ method, in which a magnetic  
field (generated by a current carrying coil), equal in  
magnitude but opposite in direction, is applied to the  
sensor. Using this ‘feedback’ method, the current through  
the coil is a direct measure of the unknown magnetic field  
amplitude and it has the advantage that the sensor is being  
operated at its zero point, where inaccuracies as result of  
tolerances, temperature drift and slight non-linearities in  
the sensor characteristics are insignificant. A detailed  
discussion of this method is covered in Chapter “Weak  
field measurement”.  
In the first stage, the sensor signal is pre-amplified and  
offset is adjusted. After temperature effects are  
compensated, final amplification and sensitivity  
adjustment takes place in the last stage. This basic circuit  
can be extended with additional components to meet  
specific EMC requirements or can be modified to obtain  
customized output characteristics (e.g. a different output  
voltage range or a current output signal).  
Philips magnetoresistive sensors have a linear sensitivity  
drift with temperature and so a temperature sensor with  
V
= 5 V  
S
offset  
sensitivity  
adjustment  
adjustment  
R7  
2.4 kΩ  
R9  
33 kΩ  
R2  
500 kΩ  
R5  
140 kΩ  
R1  
100 kΩ  
R12  
150 kΩ  
R3  
22 kΩ  
KMZ10B  
R6  
KTY82-210  
1
R11  
22 kΩ  
8
op-amp  
4
IC1  
op-amp  
3
2
2
3
6
5
7
V
= 0.2 V to 4.8 V  
4
O
TLC2272  
1
(with resistive load  
R4  
greater than 10 k)  
14 kΩ  
R10  
C1  
R8  
33 kΩ  
10 nF  
2.4 kΩ  
MBH687  
Fig.11 Basic application circuit with temperature compensation and offset adjustment.  
10  
2000 Sep 06  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
Further information for advanced users  
THE MR EFFECT  
In sensors employing the MR effect, the resistance of the  
sensor under the influence of a magnetic field changes as  
it is moved through an angle α as given by:  
Barber pole  
handbook, halfpage  
R = RO + ROcos2α  
(2)  
I
I
It can be shown that  
H2  
sin2α =  
for H H  
O
(3)  
-------  
HO2  
MLC125  
Permalloy  
Magnetization  
and  
sin2α = 1 for H > HO  
(4)  
where Ho can be regarded as a material constant  
comprising the so called demagnetizing and anisotropic  
fields.  
Fig.12 Linearization of the magnetoresistive effect.  
Applying equations (3) and (4) to equation (2) leads to:  
H2  
R = RO + RO 1 –  
for H H0  
(5)  
(6)  
-------  
HO2  
A Wheatstone bridge configuration is also used for  
linearized applications. In one pair of diagonally opposed  
elements, the Barber poles are at +45° to the strip axis,  
while in another pair they are at 45°. A resistance  
increase in one pair of elements due to an external  
magnetic field is thus ‘matched’ by a decrease in  
resistance of equal magnitude in the other pair.  
The resulting bridge imbalance is then a linear function of  
the amplitude of the external magnetic field in the plane of  
the permalloy strips, normal to the strip axis.  
R = RO for H > HO  
which clearly shows the non-linear nature of the MR effect.  
More detailed information on the derivation of the formulae  
for the MR effect can be found in Appendix 1.  
LINEARIZATION  
The magnetoresistive effect can be linearized by  
depositing aluminium stripes (Barber poles), on top of the  
permalloy strip at an angle of 45° to the strip axis (see  
Fig.12). As aluminium has a much higher conductivity than  
permalloy, the effect of the Barber poles is to rotate the  
current direction through 45° (the current flow assumes a  
‘saw-tooth’ shape), effectively changing the rotation angle  
of the magnetization relative to the current from α to  
α 45°.  
2000 Sep 06  
11  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
H2  
HO2  
R  
2
H
R = RO  
+
O + R  
-----------  
1 –  
(7)  
-------  
H O  
-------  
O
The equation is linear where H/Ho = 0, as shown in Fig.7.  
Likewise, for sensors using Barber poles arranged at an  
angle of 45°, the equation derives to:  
R
handbook, halfpage  
H2  
H02  
R  
2
H
R = RO  
+
O R  
-----------  
1 –  
(8)  
-------  
H O  
------  
O
This is the mirror image of the characteristic in Fig.7.  
H
Hence using a Wheatstone bridge configuration ensures  
the any bridge imbalance is a linear function of the  
amplitude of the external magnetic field.  
MLC126  
FLIPPING  
As described in the body of the chapter, Fig.7 shows that  
flipping is not instantaneous and it also illustrates the  
hysteresis effect exhibited by the sensor. This figure and  
Fig.14 also shows that the sensitivity of the sensor falls  
with increasing ‘Hx’. Again, this is to be expected since the  
moment imposed on the magnetization by ‘Hx’ directly  
opposes that imposed by ‘Hy’, thereby reducing the degree  
of bridge imbalance and hence the output signal for a  
given value of ‘Hy’.  
Fig.13 The resistance of the permalloy as a  
function of the external field H after  
linearization (compare with Fig.6).  
For sensors using Barber poles arranged at an angle of  
+45° to the strip axis, the following expression for the  
sensor characteristic can be derived (see Appendix 1 on  
the MR effect):  
MLC132  
150  
V
O
H
=
x
(mV)  
4 kA/m  
100  
2 kA/m  
1 kA/m  
50  
0
0
2
4
6
8
10  
H
12  
(kA/m)  
y
Fig.14 Sensor output ‘Vo’ as a function of the transverse field ‘Hy’ for several values of auxiliary field ‘Hx’.  
2000 Sep 06  
12  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
The following general recommendations for operating the  
KMZ10 can be applied:  
The greater the auxiliary field, the greater the disturbing  
field that can be tolerated before flipping occurs.  
For auxiliary fields above 3 kA/m, the SOAR graph shows  
that the sensor is completely stable, regardless of the  
magnitude of the disturbing field. It can also be seen from  
this graph that the SOAR can be extended for low values  
of ‘Hy’. In Fig.15, (for the KMZ10B sensor), the extension  
for Hy < 1 kA/m is shown.  
To ensure stable operation, avoid operating the sensor  
in an environment where it is likely to be subjected to  
negative external fields (‘Hx’). Preferably, apply a  
positive auxiliary field (‘Hx’) of sufficient magnitude to  
prevent any likelihood of flipping within he intended  
operating range (i.e. the range of ‘Hy’).  
Before using the sensor for the first time, apply a positive  
auxiliary field of at least 3 kA/m; this will effectively erase  
the sensor’s magnetic ‘history’ and will ensure that no  
residual hysteresis remains (refer to Fig.6).  
TEMPERATURE COMPENSATION  
With magnetoresistive sensors, temperature drift is  
negative. Two circuits manufactured in SMD-technology  
which include temperature compensation are briefly  
described below.  
Use the minimum auxiliary field that will ensure stable  
operation, because the larger the auxiliary field, the  
lower the sensitivity, but the actual value will depend on  
the value of Hd. For the KMZ10B sensor, a minimum  
auxiliary field of approximately 1 kA/m is recommended;  
to guarantee stable operation for all values of Hd, the  
sensor should be operated in an auxiliary field of 3 kA/m.  
The first circuit is the basic application circuit already given  
(see Fig.11). It provides average (sensor-to-sensor)  
compensation of sensitivity drift with temperature using the  
KTY82-210 silicon temperature sensor. It also includes  
offset adjustment (via R1); gain adjustment is performed  
with a second op-amp stage. The temperature sensor is  
part of the amplifier’s feedback loop and thus increases the  
amplification with increasing temperature.  
These recommendations (particularly the first one) define  
a kind of Safe Operating ARea (SOAR) for the sensors.  
This is illustrated in Fig.15, which is an example (for the  
KMZ10B sensor) of the SOAR graphs to be found in our  
data sheets.  
The temperature dependant amplification A and the  
temperature coefficient TCA of the first op-amp stage are  
approximately:  
R5  
2RT  
A =  
1 +  
for R8 = R7  
------  
R3  
----------  
R7  
MLC133  
12  
handbook, halfpage  
TCKTY  
H
d
(kA/m)  
TCA  
=
for R = R  
7
--------------------  
R7  
8
1 +  
----------  
2RT  
8
RT is the temperature dependent resistance of the KTY82.  
The values are taken for a certain reference temperature.  
This is usually 25 °C, but in other applications a different  
reference temperature may be more suitable.  
SOAR  
4
I
Figure 16 shows an example with a commonly-used  
instrumentation amplifier. The circuit can be divided into  
two stages: a differential amplifier stage that produces a  
symmetrical output signal derived from the  
magnetoresistive sensor, and an output stage that also  
provides a reference to ground for the amplification stage.  
II  
0
0
1
2
3
4
H
(kA/m)  
x
To compensate for the negative sensor drift, as with the  
above circuit the amplification is again given an equal but  
positive temperature coefficient, by means of a  
KTY81-110 silicon temperature sensor in the feedback  
loop of the differential amplifier.  
Fig.15 SOAR of a KMZ10B sensor as a function of  
auxiliary field ‘Hx’ and disturbing field ‘Hd’  
opposing ‘Hx’ (area I).  
2000 Sep 06  
13  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
V
S
OP1  
R10  
R
T
R12  
KTY82-110  
R14  
R5  
R6  
R1  
R
A
V
OP3  
O
R4  
R13  
offset  
R
B
R7  
R3  
R9  
OP2  
R11  
KMZ10B  
R2  
MLC145  
Fig.16 KMZ10B application circuit with instrumentation amplifier.  
The amplification of the input stage (‘OP1’ and ‘OP2’) is  
given by:  
RT × TCKTY  
RA + RB + RT  
TCA  
=
(11)  
----------------------------------  
RT + RB  
For the given negative ‘TC’ of the magnetoresistive sensor  
and the required amplification of the input stage ‘A1’, the  
resistance ‘RA’ and ‘RB’ can be calculated by:  
A1 = 1 +  
(9)  
--------------------  
RA  
where RT is the temperature dependent resistance of the  
KTY82 sensor and RB is the bridge resistance of the  
magnetoresistive sensor.  
TC  
1
R B = R T ×  
KTY × 1 –  
-----------------  
1  
(12)  
-------  
A1  
TCA  
The amplification of the complete amplifier can be  
calculated by:  
RT + RB  
RA  
=
(13)  
--------------------  
A1 1  
R14  
A = A1 ×  
(10)  
--------  
where TCKTY is the temperature coefficient of the KTY  
sensor and TCA is the temperature coefficient of the  
amplifier. This circuit also provides for adjustment of gain  
and offset voltage of the magnetic-field sensor.  
R10  
The positive temperature coefficient (TC) of the  
amplification is:  
2000 Sep 06  
14  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
APPENDIX 1: THE MAGNETORESISTIVE EFFECT  
Figure 17 shows the geometry of a simple sensor where  
the thickness (t) is much smaller than the width (w) which  
is in turn, less than the length (l) (i.e. t « w ‹ l). With the  
current (I) flowing in the x-direction (i.e. q = 0 or Q = f) then  
the following equation can be obtained from equation 1:  
R = R0 + DR cos2f(2)  
Magnetoresistive sensors make use of the fact that the  
electrical resistance ρ of certain ferromagnetic alloys is  
influenced by external fields. This solid-state  
magnetoresistive effect, or anisotropic magnetoresistance,  
can be easily realized using thin film technology, so lends  
itself to sensor applications.  
and with a constant current Ι, the voltage drop in the  
x-direction Ux becomes:  
Resistance- field relation  
The specific resistance ρ of anisotropic ferromagnetic  
L
∆ρ  
------  
ρ
Ux = ρΙ  
1 +  
cos2φ  
(3)  
-----  
wt  
metals depends on the angle Θ between the internal  
magnetization M and the current I, according to:  
ρ(Θ) = ρ+(ρρ||) cos2 Θ  
Besides this voltage, which is directly allied to the  
resistance variation, there is a voltage in the y-direction,  
Uy, given by:  
(1)  
where ρand ρ|| are the resistivities perpendicular and  
parallel to M. The quotient ρ||)/ρ= ρ/ρ  
is called the magnetoresistive effect and may amount to  
several percent.  
1
∆ρ  
ρ
Uy =ρΙ  
sinφcosφ  
(4)  
-- ------  
t
This is called the planar or pseudo Hall effect; it  
resembles the normal or transverse Hall effect but has a  
physically different origin.  
Sensors are always made from ferromagnetic thin films as  
this has two major advantages over bulk material: the  
resistance is high and the anisotropy can be made  
uniaxial. The ferromagnetic layer behaves like a single  
domain and has one distinguished direction of  
magnetization in its plane called the easy axis (e.a.),  
which is the direction of magnetization without external  
field influence.  
All sensor signals are determined by the angle φ between  
the magnetization M and the ‘length’ axis and, as M  
rotates under the influence of external fields, these  
external fields thus directly determine sensor signals. We  
can assume that the sensor is manufactured such that the  
e.a. is in the x-direction so that without the influence of  
external fields, M only has an x-component  
(φ = 0˚ or 180˚).  
Two energies have to be introduced when M is rotated by  
external magnetic fields: the anisotropy energy and the  
demagnetizing energy. The anisotropy energy Ek, is given  
by the crystal anisotropy field Hk, which depends on the  
material and processes used in manufacture. The  
demagnetizing energy Ed or form anisotropy depends on  
the geometry and this is generally a rather complex  
relationship, apart from ellipsoids where a uniform  
demagnetizing field Hd may be introduced. In this case, for  
the sensor set-up in Fig.17.  
handbook, halfpage  
L
M
y
ϕ
W
ϑ
Ι
MBH616  
x
M s  
t
H ≈  
(5)  
--- ------  
w µ 0  
d
where the demagnetizing factor N t/w, the saturation  
magnetization Ms 1 T and the induction constant  
-7  
µ0 = 4π Vs/Am.  
Fig.17 Geometry of a simple sensor.  
The field H0 Hk + t/w(M0/m0) determines the measuring  
range of a magnetoresistive sensor, as f is given by:  
2000 Sep 06  
15  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
resistance-field (R-H) dependence, so a simple  
Hy  
sinφ=  
(6)  
--------------------------  
magnetoresistive element cannot be used directly for  
linear field measurements. A magnetic biasing field can  
be used to solve this problem, but a better solution is  
linearization using barber-poles (described later).  
Nevertheless plain elements are useful for applications  
using strong magnetic fields which saturate the sensor,  
where the actual value of the field is not being measured,  
such as for angle measurement. In this case, the direction  
of the magnetization is parallel to the field and the sensor  
signal can be described by a cos2α function.  
Hx  
H +  
------------  
o
cosφ  
where |Hy| |H0 + Hx| and Hx and Hy are the components  
of the external field. In the simplest case Hx = 0, the volt-  
ages Ux and Uy become:  
2
Hy  
L
∆ρ  
------  
ρ
Ux =ρl  
1 +  
1 –  
(7)  
-----  
wt  
------  
H0  
Hy  
-- ------ ------  
H0  
1
t
∆ρ  
ρ
1 (Hy H0)2  
(8)  
Sensors with inclined elements  
Uy =ρl  
Sensors can also be linearized by rotating the current path,  
by using resistive elements inclined at an angle θ, as  
shown in Fig.18. An actual device uses four inclined  
resistive elements, two pairs each with opposite  
inclinations, in a bridge.  
(Note: if Hx = 0, then H0 must be replaced by  
H0 + Hx/cos φ).  
Neglecting the constant part in Ux, there are two main  
differences between Ux and Uy:  
1. The magnetoresistive signal Ux depends on the  
The magnetic behaviour of such is pattern is more  
complicated as Mo is determined by the angle of inclination  
θ, anisotropy, demagnetization and bias field (if present).  
Linearity is at its maximum for φ + θ 45˚, which can be  
achieved through proper selection of θ.  
square of Hy/H0, whereas the Hall voltage Uy is linear  
for Hy « H0.  
2. The ratio of their maximum values is L/w; the Hall  
voltage is much smaller as in most cases L » w.  
A stabilization field (Hst) in the x-direction may be  
necessary for some applications, as this arrangement only  
works properly in one magnetization state.  
Magnetization of the thin layer  
The magnetic field is in reality slightly more complicated  
than given in equation (6). There are two solutions for  
angle φ:  
φ1 < 90˚ and φ2 > 90˚ (with φ1 + φ2 = 180˚ for Hx = 0).  
Replacing φ by 180˚ - φ has no influence on Ux except to  
change the sign of the Hall voltage and also that of most  
linearized magnetoresistive sensors.  
Therefore, to avoid ambiguity either a short pulse of a  
proper field in the x-axis (|Hx| > Hk) with the correct sign  
must be applied, which will switch the magnetization into  
the desired state, or a stabilizing field Hst in the  
x-direction can be used. With the exception of Hy « H0, it  
is advisable to use a stabilizing field as in this case, Hx  
values are not affected by the non-ideal behaviour of the  
layer or restricted by the so-called ‘blocking curve’.  
M
handbook, halfpage  
0
ϕ
ϑ
Ι
ϑ
ϕ
Ι
M
0
MBH613  
The minimum value of Hst depends on the structure of the  
sensitive layer and has to be of the order of Hk, as an  
insufficient value will produce an open characteristic  
(hysteresis) of the sensor. An easy axis in the y-direction  
leads to a sensor of higher sensitivity, as then  
Ho = Hk Hd.  
Fig.18 Current rotation by inclined elements  
(current and magnetization shown in  
quiescent state).  
Linearization  
As shown, the basic magnetoresistor has a square  
2000 Sep 06  
16  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
BARBER-POLE SENSORS  
equal widths. The characteristic is plotted in Fig 20 and it  
can be seen that for small values of Hy relative to H0, the  
R-H dependence is linear. In fact this equation gives the  
same linear R-H dependence as the planar Hall-effect  
sensor, but it has the magnitude of the magnetoresistive  
sensor.  
A number of Philips’ magnetoresistive sensors use a  
‘barber-pole’ construction to linearize the R-H relationship,  
incorporating slanted strips of a good conductor to rotate  
the current. This type of sensor has the widest range of  
linearity, smaller resistance and the least associated  
distortion than any other form of linearization, and is well  
suited to medium and high fields.  
MBH615  
handbook, halfpage  
R
handbook, halfpage  
Permalloy  
Barber pole  
Ι
R  
+
Ι
y
Ι
Magnetization  
ϑ
R
0
MBH614  
x
0
1  
0.5  
0
0.5  
1
H
Y
H
0
Fig.19 Linearization of the magnetoresistive effect  
with barber-poles (current and  
Fig.20 Calculated R-H characteristic of a  
barber-pole sensor.  
magnetization shown in quiescent state).  
Barber-pole sensors require a certain magnetization  
state. A bias field of several hundred A/m can be  
generated by the sensing current alone, but this is not  
sufficient for sensor stabilization, so can be neglected. In  
most applications, an external field is applied for this  
purpose.  
The current takes the shortest route in the high-resistivity  
gaps which, as shown in Fig 19, is perpendicular to the  
barber-poles. Barber-poles inclined in the opposite  
direction will result in the opposite sign for the R-H  
characteristic, making it extremely simple to realize a  
Wheatstone bridge set-up.  
Sensitivity  
Due to the high demagnetization, in most applications  
field components in the z-direction (perpendicular to the  
layer plane) can be ignored. Nearly all sensors are most  
sensitive to fields in the y-direction, with Hx only having a  
limited or even negligible influence.  
The signal voltage of a Barber-pole sensor may be  
calculated from the basic equation (1) with Θ = φ + 45˚  
(θ = + 45˚):  
2
H
H y  
Definition of the sensitivity S contains the signal and field  
variations (DU and DH), as well as the operating voltage  
U0 (as DU is proportional to U0):  
L
1 ρ ∆ρ  
y
UBP =ρl  
α 1 +  
±
1 –  
(9)  
-----  
wt  
-- ------ ------ ------  
------  
H 0  
2
ρ
ρ H 0  
U 1  
------- ------  
H U 0  
U  
So =  
=
(10)  
---------------  
U0H  
where a is a constant arising from the partial shorting of the  
resistor, amounting to 0.25 if barber-poles and gaps have  
2000 Sep 06  
17  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
requires a high sensor resistance R with a large area A,  
since there are limits for power dissipation and current  
density. The current density in permalloy may be very high  
(j > 106 A/cm2 in passivation layers), but there are weak  
points at the current reversal in the meander (see section  
on sensor layout) and in the barber-pole material, with  
five-fold increased current density.  
This definition relates DU to a unit operating voltage.  
The highest (HG) and lowest (Hmin) fields detectable by  
the sensor are also of significance. The measuring range  
HG is restricted by non-linearity - if this is assumed at 5%,  
an approximate value for barber-pole sensors is given by:  
HG 0.5(H0 + Hx)  
(11)  
A high resistance sensor with U0 = 25 V and a maximum  
S0 results in a value of 2.5 x 10-3  
(A/m)-1 for Su or, converted to flux density, ST = 2000 V/T.  
This value is several orders of magnitude higher than for a  
normal Hall effect sensor, but is valid only for a much  
narrower measuring range.  
From this and equation (9) for signal voltage (UBP) for a  
barber-pole sensor, the following simple relationship can  
∆ρ  
be obtained: HGS0 0.5  
(12)  
------  
ρ
Other sensor types have a narrower range of linearity and  
therefore a smaller useful signal.  
Materials  
The lowest detectable field Hmin is limited by offset, drift  
and noise. The offset is nearly cancelled in a bridge circuit  
and the remaining imbalance is minimized by symmetrical  
design and offset trimming, with thermal noise negligible in  
most applications (see section on sensor layout). Proper  
film deposition and, if necessary, the introduction of a  
stabilization field will eliminate magnetization switching  
due to domain splitting and the introduction of ‘Barkhausen  
noise’.  
Sensitivity S0 is essentially determined by the sum of the  
anisotropy (Hk), demagnetization (Hd) and bias (Hx) fields.  
The highest sensitivity is achievable with Hx = 0 and  
Hd « Hk, although in this case S0 depends purely on Hk  
which is less stable than Hd. For a permalloy with a  
thickness greater than or equal to 20 µm, a width in  
excess of 60 µm is required which, although possible, has  
the drawback of producing a very low resistance per unit  
area.  
There are five major criteria for a magnetoresistive  
material:  
Large magnetoresistive effect Dr/r (resulting in a high  
signal to operating voltage ratio)  
Large specific resistance r (to achieve high resistance  
value over a small area)  
Low anisotropy  
Zero magnetostriction (to avoid influence of mechanical  
stress)  
Long-term stability.  
Appropriate materials are binary and ternary alloys of Ni,  
Fe and Co, of which NiFe (81/19) is probably the most  
common.  
Table 1 gives a comparison between some of the more  
common materials, although the majority of the figures are  
only approximations as the exact values depend on a  
number of variables such as thickness, deposition and  
post-processing.  
The maximum theoretical S0 with this permalloy (at  
Hk = 250 A/m and ρ= 2.5%) is approximately:  
mV  
--------  
1
Table 3 Comparison of magnetoresistive sensor  
V
A
----  
m
S0(max) = 104  
= 100  
(13)  
--------------  
materials  
kA  
------  
m
Materials  
ρ (108m)  
22  
ρ/ρ(%)  
2.2  
ΙΙk(/m)  
250  
NiFe 81:19  
NiFe 86:14  
NiCo 50:50  
NiCo 70:30  
For the same reasons, sensors with reduced sensitivity  
should be realized with increased Hd, which can be esti-  
mated at a maximum for a barber-pole sensor at 40 kA/m.  
A further reduction in sensitivity and a corresponding  
growth in the linearity range is attained using a biasing  
field. A magnetic shunt parallel to the magnetoresistor or  
only having a small field component in the sensitive direc-  
tion can also be employed with very high field strengths.  
15  
24  
26  
3
200  
2.2  
3.7  
0.07  
2500  
2500  
2000  
CoFeB 72:8:20 86  
∆ρ is nearly independent of these factors, but r itself  
increases with thickness (t 40 nm) and will decrease  
during annealing. Permalloys have a low Hk and zero  
magnetostriction; the addition of Co will increase ρ, but  
A high signal voltage Ux can only be produced with a  
sensor that can tolerate a high supply voltage Uo. This  
2000 Sep 06  
18  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
this also considerably enlarges Hk. If a small temperature  
coefficient of ρ is required, NiCo alloys are preferable.  
The amorphous alloy CoFeB has a low ρ, high Hk and  
slightly worse thermal stability but due to the absence of  
grain boundaries within the amorphous structure, exhibits  
excellent magnetic behaviour.  
the more the magnetization rotates towards 90˚ and  
therefore it becomes easier to flip the sensor into the  
corresponding stable position in the ‘-x’ direction. This  
means that a smaller -Hx field is sufficient to cause the  
flipping action  
As can be seen in Fig 22, for low transverse field strengths  
(0.5 kA/m) the sensor characteristic is stable for all positive  
values of Hx, and a reverse field of approximately 1 kA/m  
is required to flip the sensor. However at higher values of  
Hy (2 kA/m), the sensor will also flip for smaller values of  
Hx (at 0.5 kA/m). Also illustrated in this figure is a  
noticeable hysteresis effect; it also shows that as the  
permalloy strips do not flip at the same rate, the flipping  
action is not instantaneous.  
APPENDIX 2: SENSOR FLIPPING  
During deposition of the permalloy strip, a strong external  
magnetic field is applied parallel to the strip axis. This  
accentuates the inherent magnetic anisotropy of the strip  
and gives them a preferred magnetization direction, so that  
even in the absence of an external magnetic field, the  
magnetization will always tend to align with the strips.  
Providing a high level of premagnetization within the  
crystal structure of the permalloy allows for two stable  
premagnetization directions. When the sensor is placed in  
a controlled external magnetic field opposing the internal  
aligning field, the polarity of the premagnetization of the  
strips can be switched or ‘flipped’ between positive and  
negative magnetization directions, resulting in two stable  
output characteristics.  
MLC131  
V
O
(mV)  
H
=
100  
y
2 kA/m  
50  
0.5 kA/m  
3
0
3
2
1
1
2
H
(kA/m)  
x
MLC130  
50  
handbook, halfpage  
V
O
(mV)  
100  
10  
0
4
2
2
4
Fig.22 Sensor output ‘Vo’ as a function of the  
auxiliary field Hx.  
H
(kA/m)  
y
10  
reversal  
of sensor  
characteristics  
The sensitivity of the sensor reduces as the auxiliary field  
Hx increases, which can be seen in Fig 22 and more  
clearly in Fig 23. This is because the moment imposed on  
the magnetization by Hx directly opposes that of Hy,  
resulting in a reduction in the degree of bridge imbalance  
and hence the output signal for a given value of Hy.  
Fig.21 Sensor characteristics.  
The field required to flip the sensor magnetization (and  
hence the output characteristic) depends on the  
magnitude of the transverse field Hy. The greater this field,  
2000 Sep 06  
19  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
MLC132  
150  
V
O
H
=
x
(mV)  
4 kA/m  
100  
2 kA/m  
1 kA/m  
50  
0
0
2
4
6
8
10  
H
12  
(kA/m)  
y
Fig.23 Sensor output ‘Vo’ as a function of the transverse field Hy.  
A Safe Operating ARea (SOAR) can be determined for  
magnetoresistive sensors, within which the sensor will not  
flip, depending on a number of factors. The higher the  
auxiliary field, the more tolerant the sensor becomes to  
external disturbing fields (Hd) and with an Hx of 3 kA/m or  
greater, the sensor is stabilized for all disturbing fields as  
long as it does not irreversibly demagnetize the sensor. If  
Hd is negative and much larger than the stabilising field Hx,  
the sensor will flip. This effect is reversible, with the sensor  
returning to the normal operating mode if Hd again  
becomes negligible (see Fig 24). However the higher Hx,  
the greater the reduction in sensor sensitivity and so it is  
generally recommended to have a minimum auxiliary field  
that ensures stable operation, generally around 1 kA/m.  
The SOAR can also be extended for low values of Hx as  
long as the transverse field is less than 1 kA/m. It is also  
recommended to apply a large positive auxiliary field  
before first using the sensor, which erases any residual  
hysteresis  
MLC133  
12  
handbook, halfpage  
H
d
(kA/m)  
8
SOAR  
4
I
II  
0
0
1
2
3
4
H
(kA/m)  
x
Fig.24 SOAR of a KMZ10B sensor as a function of  
auxiliary field ‘Hx’ (MLC133).  
2000 Sep 06  
20  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
APPENDIX 3: SENSOR LAYOUT  
different for these three families of sensors in every case,  
the elements are linked in the same fashion to form the four  
arms of a Wheatstone bridge. The meander pattern used  
in the KMZ51 is more sophisticated and also includes  
integrated compensation and flipping coils (see chapter on  
weak fields); the KMZ41 is described in more detail in the  
chapter on angle measurement.  
In Philips’ magnetoresistive sensors, the permalloy strips  
are formed into a meander pattern on the silicon substrate.  
With the KMZ10 (see Fig 25) and KMZ51 series, four  
barber-pole permalloy strips are used while the KMZ41  
series has simple elements. The patterns used are  
MBC930  
Fig.25 KMZ10 chip structure.  
2000 Sep 06  
21  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
In one pair of diagonally opposed elements the  
barber-poles are at +45˚ to the strip axis, with the second  
pair at 45˚. A resistance increase in one pair of elements  
due to an external magnetic field is matched by an equal  
decrease in resistance of the second pair. The resulting  
bridge imbalance is then a linear function of the amplitude  
of the external magnetic field in the plane of the permalloy  
strips normal to the strip axis.  
MLC129  
handbook, halfpage  
This layout largely eliminates the effects of ambient  
variations (e.g. temperature) on the individual elements  
and also magnifies the degree of bridge imbalance,  
increasing sensitivity.  
R
R
T
T
4
3
2
1
V
V
V
O
GND  
CC  
O
Fig 26 indicates two further trimming resistors (RT) which  
allow the sensors electrical offset to be trimmed down to  
zero during the production process.  
Fig.26 KMZ10 and KMZ11 bridge configuration.  
2000 Sep 06  
22  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
WEAK FIELD MEASUREMENT  
Contents  
static offset, an offset drift due to temperature variations of  
about 6 (µV/V)K1 can be expected and assuming an  
ambient temperature up to 100 °C, the resulting offset can  
be of the order of 2 mV/V.  
Fundamental measurement techniques  
Taking these factors into account, with no external field a  
sensor with a typical sensitivity of 15 mV/V (kA/m)1 can  
have an offset equivalent to a field of 130 A/m, which is  
itself about four times the strength of a typical weak field  
such as the earth’s geomagnetic field. Clearly, measures  
to compensate for the sensor offset value have to be  
implemented in weak field applications.  
Application note AN00022: Electronic compass design  
using KMZ51 and KMZ52  
Application circuit: signal conditioning unit for compass  
Example 1: Earth geomagnetic field compensation in  
CRT’s  
Example 2: Traffic detection  
Example 3: Measurement of current.  
A technique called ‘flipping’ (patented by Philips) can be  
used to control the sensor. Comparable to the ‘chopping’  
technique used in the amplification of small electrical  
signals, it not only stabilizes the sensor but also eliminates  
the described offset effects.  
Fundamental measurement techniques  
Measurement of weak magnetic fields such as the earth’s  
geomagnetic field (which has a typical strength of between  
approximately 30 A/m and 50 A/m), or fields resulting from  
very small currents, requires a sensor with very high  
sensitivity. With their inherent high sensitivity,  
magnetoresistive sensors are extremely well suited to  
sensing very small fields.  
When the bi-stable sensor is placed in a controlled,  
reversible external magnetic field, the polarity of the  
premagnetization (Mx) of the sensor strips can be switched  
or flipped between the two output characteristics (see  
Fig.27).  
Philips’ magnetoresistive sensors are by nature bi-stable  
(refer to Appendix 2). ‘Standard’ techniques used to  
stabilize such sensors, including the application of a strong  
field in the x-direction (Hx) from a permanent stabilization  
magnet, are unsuitable as they reduce the sensor’s  
sensitivity to fields in the measurement, or y-direction (Hy).  
(Refer to Appendix 2, Fig. A2.2).  
V
O
M
x
To avoid this loss in sensitivity, magnetoresistive sensors  
can instead be stabilized by applying brief, strong  
non-permanent field pulses of very short duration (a few  
µs). This magnetic field, which can be easily generated by  
simply winding a coil around the sensor, has the same  
stabilizing effect as a permanent magnet, but as it is only  
present for a very short duration, after the pulse there is no  
loss of sensitivity. Modern magnetoresistive sensors  
specifically designed for weak field applications  
offset  
H
y
M
x
MLC764  
incorporate this coil on the silicon.  
However, when measuring weak fields, second order  
effects such as sensor offset and temperature effects can  
greatly reduce both the sensitivity and accuracy of MR  
sensors. Compensation techniques are required to  
suppress these effects.  
Fig.27 Butterfly curve including offset.  
This reversible external magnetic field can be easily  
achieved with a coil wound around the sensor, consisting  
of current carrying wires, as described above. Depending  
on the direction of current pulses through this coil, positive  
and negative flipping fields in the x-direction (+Hx and Hx)  
are generated (see Fig.28).  
OFFSET COMPENSATION BY FLIPPING’  
Despite electrical trimming, MR sensors may have a  
maximum offset voltage of ±1.5 mV/V. In addition to this  
2000 Sep 06  
23  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
Flipping causes a change in the polarity of the sensor  
output signal and this can be used to separate the offset  
signal from the measured signal. Essentially, the unknown  
field in the ‘normal’ positive direction (plus the offset) is  
measured in one half of the cycle, while the unknown field  
in the ‘inverted’ negative direction (plus the offset) is  
measured in the second half. This results in two different  
outputs symmetrically positioned around the offset value.  
After high pass filtering and rectification a single,  
continuous value free of offset is output, smoothed by low  
pass filtering. See Figs 29 and 30.  
current  
pulses  
coil  
H
sensor  
y
V
O
H
x
Offset compensation using flipping requires additional  
external circuitry to recover the measured signal.  
MLC762  
Fig.28 Flipping coil.  
CLOCK  
T
I
F
PHASE  
SENSITIVE  
DEMODULATOR  
FLIPPING  
L
PRE-  
AMPLIFIER  
OFFSET  
FILTER  
V
out  
F
SOURCE  
MBH617  
Fig.29 Block diagram of flipping circuit.  
2000 Sep 06  
24  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
T
T
T
flipping current I  
F
time  
internal  
magnetization  
V
O
V
O
offset  
time  
H
y
(a)  
(b)  
(c)  
V
O
time  
V
O
time  
MBH618  
Fig.30 Timing diagram for flipping circuit (a) output voltage; (b) filtered output voltage; (c) output voltage filtered  
and demodulated.  
2000 Sep 06  
25  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
SENSOR TEMPERATURE DRIFT  
not negligible, as it can produce a difference of a factor of  
three within a 25 °C to +125 °C temperature range, for  
fields up to 0.5 kA/m. This effect is not compensated for by  
the flipping action described in the last section.  
The sensitivity of MR sensors is also temperature  
dependent, with sensitivity decreasing as temperature  
increases (Fig.31).The effect on sensor output is certainly  
MLC134  
15  
V
O
o
T
=
25 C  
amb  
(mV/V)  
o
25 C  
10  
o
75 C  
5
0
o
125 C  
5
10  
operating range  
15  
3
2
1
0
1
2
3
H
(kA/m)  
y
Fig.31 Output voltage ‘Vo’ as a fraction of the supply voltage for a KMZ10B sensor, as a function of transverse  
field ‘Hy’, at several temperatures.  
2000 Sep 06  
26  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
The simplest form of temperature compensation is to use  
a current source to supply to the sensor instead of a  
voltage source. In this case, the resulting reduction in  
sensitivity due to temperature is partially compensated by  
a corresponding increase in bridge resistance.  
output voltage ‘Vo’, and reduces the variation in sensitivity  
to a factor of approximately 1.5 (compared to a factor of  
three using the voltage source). However, this method  
requires a higher supply voltage, due to the voltage drop  
of the current source.  
Thus a current source not only improves the stability of the  
MLC135  
75  
o
T
=
25 C  
amb  
V
O
o
25 C  
(mV/V)  
50  
o
75 C  
o
125 C  
25  
0
25  
50  
75  
operating range  
4
2
0
2
4
H
(kA/m)  
y
Fig.32 Output voltage ‘Vo’ of a KMZ10B sensor as a function of transverse field ‘Hy’ using a current source, for  
several temperatures.  
2000 Sep 06  
27  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
The optimal method of compensating for temperature  
dependent sensitivity differences in MR measurements of  
weak fields uses electro-magnetic feedback. As can be  
seen from the sensor characteristics in Figs 31 and 32,  
sensor output is completely independent of temperature  
changes at the point where no external field is applied  
(the null-point). By using an electro-magnetic feedback  
set-up, it is possible to ensure the sensor is always  
operated at this point.  
The magnetic field produced by the compensation coil is in  
the opposite direction to the measured field, so when it is  
added to the measured field, it compensates exactly for  
the change in the output signal, regardless of its actual,  
temperature-dependent value. This principle is called  
current compensation and because the sensor is always  
used at its ‘zero’ point, compensation current is  
independent of the actual sensitivity of the sensor or  
sensitivity drift with temperature.  
To achieve this, a second compensation coil is wrapped  
around the sensor perpendicular to the flipping coil, so that  
the magnetic field produced by this coil is in the same  
plane as the field being measured.  
Information on the measured magnetic signal is effectively  
given by the current fed to the compensating coil. If the  
field factor of the compensation coil is known, this  
simplifies calculation of the compensating field from the  
compensating current and therefore the calculation of the  
measured magnetic field. If this field factor is not precisely  
known, then the resistor performing the current/voltage  
conversion must be trimmed. Figure 34 shows a block  
diagram of a compensated sensor set-up including the  
flipping circuit.  
Should the measured magnetic field vary, the sensor’s  
output voltage will change, but the change will be different  
at different ambient temperatures. This voltage change is  
converted into a current by an integral controller and  
supplied to the compensation coil, which then itself  
produces a magnetic field proportional to the output  
voltage change caused by the change in measured field.  
compensation coil  
compensation field  
flipping field  
earth's field  
MLC757  
flipping coil  
sensor KMZ10A1  
Fig.33 Magnetic field directions and the flipping and compensation coils.  
2000 Sep 06  
28  
 
Philips Semiconductors  
Magnetoresistive sensors for  
magnetic field measurement  
General  
The influence of other disturbing fields can also be  
eliminated provided they are well known, by adding a  
second current source to the compensating coil. Such  
fields might be those arising from the set-up housing,  
ferromagnetic components placed close to the sensor or  
magnetic fields from electrical motors.  
The brief summary in Table 3 compares the types of  
compensation and their effects, so they can be assessed  
for their suitability in a given application. Because these  
options encompass a range of costs, the individual  
requirements of an application should be carefully  
analysed in terms of the performance gains versus relative  
costs.  
CLOCK  
PRE-AMPLIFIER  
WITH  
SUPRESSION  
OF OFFSET  
PHASE-  
SENSITIVE  
DEMODULATOR  
L
FLIPPING  
SOURCE  
F
L
C
CURRENT  
REGULATOR  
VOLTAGE & CURRENT  
OUTPUT  
MBH619  
Fig.34 Block diagram of compensation circuit.  
Table 4 Summery of compensation techniques  
TECHNIQUE  
EFFECT  
Setting  
avoids reduction in sensitivity due to constant stabilization field  
Flipping  
avoids reduction in sensitivity due to constant stabilization field, as well as  
compensating for sensor offset and offset drift due to temperature  
Current supply  
reduction of sensitivity drift with temperature by a factor of two  
Electro-magnetic feedback accurate compensation of sensitivity drift with temperature  
2000 Sep 06  
29  
 

Peavey Music Mixer AMD 1220 User Manual
Peavey Portable Speaker DTH 4215f User Manual
Peavey Stereo Amplifier PV1600 User Manual
Philips Car Amplifier TDA1563Q User Manual
Philips Flat Panel Television 12 User Manual
Philips Paper Shredder 7FF3FPB User Manual
Philips Stereo System FW V220 21 User Manual
Plantronics Headphones 47349 01 User Manual
Porter Cable Nail Gun 893885 007 User Manual
Proxim Network Router AP 700 User Manual