WO2007051801A1

WO2007051801A1 - Arrangement for magnetic field measurement

Info

Publication number: WO2007051801A1
Application number: PCT/EP2006/067984
Authority: WO
Inventors: Roland Weiss
Original assignee: Siemens Aktiengesellschaft
Priority date: 2005-11-04
Filing date: 2006-10-31
Publication date: 2007-05-10
Also published as: DE102005052688A1

Abstract

A magnetic field sensor (2) comprises two electrical connections (4a, 4b) for voltage supply, and at least three branches (8a - 8c), which are connected in parallel between the connections (4a, 4b) and form a measurement bridge, wherein each branch (8a - 8c) contains at least two series-connected MR sensors (16ik) with a voltage tap (18a - 18d) located therebetween for a bridge voltage (Uac, Uad, Ubc, Ubd) and forms at least three branch groups (10a - 10c). The individual branch groups (10a - 10c) are arranged in mirror-symmetrical or point-symmetrical fashion to form a single balancing element (SE, SP) within the realms of measurement-engineering tolerances.

Description

description

Arrangement for magnetic field measurement

The invention relates to an arrangement for magnetic field measurement with a measuring bridge of MR sensors.

Magnetoresistive sensors, so-called MR sensors, are based on so-called XMR effects. The electrical properties of such a sensor change depending on the magnetic field applied to the sensor. XMR effects denote a whole class of similar effects, e.g. GMR sensors, in particular, represent an alternative to Hall sensors. MR sensors are always used for measuring the magnetic field and are used in the position and velocity measurement systems. , Speed, field or current sensors. Particularly in the field of position and current sensors, the market has increasingly relied on sensors based on XMR effects in recent years.

Compared to Hall sensors are MR sensors mainly the advantages of a simpler system structure, a greater resistance to interference due to the more reduced foreign ^¬ field sensitivity in the system, and a lower noise on. The higher sensitivity eliminates the need for additional, previously necessary flux concentrators at the sensor, which leads to further advantages of the overall sensor, such as lower weight and fewer saturation problems.

Fully integrated solutions are particularly suitable for MR-based sensors, since the MR elements can be applied as a back-end process on top of an electronics already mounted on a chip substrate, thus requiring no additional chip area.

In many applications, especially in position, speed and current sensors, four MR elements each are connected in a so-called Wheatstone bridge in order to produce a more precise, from temperature fluctuations, external fields, etc. to achieve more independent measurement.

From DE 34 26 784 Al is for example a Magnetfeldsen- sor known as particular in the specific part insbeson ^¬ is shown in FIG. 1 In detail, Figure 1 shows a magnetic field measuring device 100 according to the prior art in the embodiment as a Wheatstone bridge. The magnetic field sensor 100 has two electrical connections 102a and 102b, with which it is connected via connection lines 104a and 104b to a voltage source, not shown. Between the terminals 102a and 102b, two electric branches 106a and 106b are connected in parallel. Each of the branches 106aun 106b consists of the series connection of two MR resistors 108 and a respective voltage tap 110a and 106b lying between the two MR resistors 108 of a branch 106a and 106b. From each voltage tap 110a and 106b leads a measuring line 112 to a separate, not shown in Figure 1 transmitter.

The measuring arrangement 100 as a Wheatstone measuring bridge therefore supplies the bridge voltage U _BE between the voltage taps 110a and 100b, which is picked off via the measuring lines 112 and further processed in the evaluation electronics. If the MR resistors 108 are penetrated by a magnetic field 114, they change their electrical resistance. By appropriate dimensioning of the MR resistors 108 or the entire measuring bridge, it is then achieved that the bridge voltage U _BE changes. The bridge voltage U _BE thus represents a measure of the magnetic field 114.

Furthermore, DE 34 26 784 A1 discloses a magnetoresistive sensor for emitting electrical signals as a function of the position or rotational speed of a ferromagnetic body, in which ferromagnetic measuring strips on a substrate are penetrated by a stationary magnetic constant field. In this case, a bridge may be formed, the conductor tracks are assigned in a specific way. An than Brü ^¬ bridge circuit formed magnetic field sensor is known from DE 10 2004 040 079 B3, wherein a first and a two ^¬ ter bridge arm, each having at least two geschal- preparing resistors in series, at least one magneto of which is each resistive, are present , In this case, the individual resistors are arranged spaced from each other for measuring a field gradient in the direction of a base axis and a third bridge branch is connected, which serves together with the first or second bridge branch for measuring the magnetic field strength. Finally, a two-Wheatstone bridge is known from US Pat. No. 6,100,686 A. To form a magnetic field sensor, a plurality of magneto-resistive elements are formed on a substrate, forming a double Wheatstone bridge, wherein at least one resistive element in each case Element is magnetic field sensitive. The two bridges are constructed identically, the polarity of each corre sponding ^¬ magnetoresistive elements is opposite. This circuit should be able to eliminate an offset signal in the measuring arrangement.

In the known Messandnungen from MR sensors, however, often occur in practice problems such. As an offset and associated offset drift and / or hysteresis effects with non-linearities and temperature drifts. These undesirable phenomena are caused by intrinsic rialeigenschaften Mate ^¬ and process-related variations in the deposition and patterning of the MR layers on a substrate. Above all, the overcurrents required in the application in the field of current measurement, or else generally exceeding the target field strength to be measured, can lead to a temporal change in the offset via the material intrinsic properties of the hysteresis and thus significantly reduce the measurement accuracy of the measuring arrangement.

The cause of the offset, that is the voltage of the MR sensor when the field strength disappears, and partly also the non-linearity, lies in the lacking pair accuracy of the MR bridge. resistors. This is caused by procedural fluctuations, which today lies in the percentage range; The aim here is an accuracy in the per thousand range.

The goal in the development of MR sensors is to compensate for offset and nonlinearity in the MR element downstream electronics, so z. As a purely digital compensation to achieve. For this, however, the offset must not exceed a certain extent.

So far, MR sensors z. B. sorted by quality, which means a high reject rate in production. Also, the MR sensors are trimmed with the laser, which is complicated and expensive. MR sensors can also be used in egg ner so-called closed-loop circuit in which an artificial field for compensation of the measured field it is evidence ^¬ to match the resulting field to zero at the sensor location.

The closed-loop circuit thus has the disadvantage of a very high power consumption and more expensive discrete electronic amplifier components or a very expensive BiCMOS process for the realization of the electronics. The above-mentioned laser trimming and sorting out, ie a greatly reduced yield, cause immense costs in the production.

The object of the present invention is therefore to improve the arrangement of MR sensors known from the prior art and the measuring bridge formed thereby.

The object is achieved by the features of Pa ^¬ tentanspruches 1. Further developments are specified in the subclaims.

The invention relates to a measuring arrangement of magnetic ^¬ field sensors with electrical connections for power supply, with at least three parallel between the terminals, a measuring bridge forming branches, wherein each branch at least two series-connected MR sensors including, with an intermediate voltage tap for a bridge voltage and wherein nigstens two branch groups are formed by the branches in each case we ^¬. The branches of a branch group can be directly next to each other or even nested. Within the scope of technical tolerances, the individual branch groups are advantageously mirror-symmetrical or point-symmetrical with respect to the same symmetry element.

Thus, each two of the at least three rallel connected between the terminals pa ^¬ branches form a measuring bridge in the manner of a Wheatstone measuring bridge. Between each two branches a separate bridge voltage can be removed in each case between their voltage taps. However, since according to the invention at least a third branch is provided with an additional voltage tap, at least three different independent bridge voltages are available between the three voltage taps of the three branches.

By at least three voltage taps not only a single bridge ^¬ ^¬, during egg nes single measurement cycle voltage as before, but at least three bridge voltages are detected. The amount of information obtained per measuring cycle is therefore significantly increased.

By a subsequent further processing of the generally different, independent bridge voltages z. B. statistical errors caused by the processing of the MR sensors are compensated.

Of course, with the two electrical connections for power supply, a power supply of the bridge, z. B. with constant supply current, be accomplished, if they should work in another measurement mode. The conditional by an ohmic resistance noise of the voltage U _r in a Net Zwerk with resistance R i st in well Wei ^¬ ter se def ined al s:

U _r = ^ 4KTRAf, (1) with K = Boltzmann constant T = abs. Temperature R = ohmic resistance Δf = frequency difference

U denotes _RBE the resistance noise in the measuring bridge with 2p = z = 2, ie, two branches of Figure 1, and U _rBP the counter ^¬ was noise in a measurement bridge with several branches 2p = z, which 2p VAS independent values for bridge voltages U _BP provides , then:

By an appropriate communication of the various 2p-l

Bridge voltages can be such. B. a smaller fluctuation of the output signal of the magnetic field sensor due to process ^¬ cal fluctuations, ie the specimen scattering of individual MR sensors, can be achieved.

If AR denotes the standard deviation of the ohmic resistance R of a single MR sensor on the sensor chip, then the bridge voltage at the magnetic field sensor can be calculated taking into account the error propagation. For the voltage error of the bridge voltage, ie the offset voltage ^AU _{BMr, P} ', the following results:

m is a factor for the meandering structure of a single MR sensor consisting of m meanders of partial sensors, as will be explained below.

Of course, in a magnetic field sensor with multiple branch groups, not all of the branch groups need to contain multiple branches. There are also single branches - in other words to single ^{two ¬} gen degenerate branch groups - possible.

At least one of the MR sensors may include a series connection of a plurality of MR sub-sensors. These may in particular be strung together meandering.

Latter measure is already only peo ^¬ within a gen MR sensor instead of a compensation of manufacturing variations in the production of each part MR sensors. The statistical fluctuations, z. B. the resistance of the MR-part sensors in this case within a single ^¬ nen MR sensor by the x-fold series connection partially off, so that a plurality of entire MR sensors against each other have fewer fluctuations.

In this case, the MR sub-sensors can be interconnected meander-shaped to form an MR sensor. The meandering arrangement of MR partial sensors is well known, as well as the resulting advantages, such as the small footprint for producing an MR sensor. The noise voltage with meandering of m MR partial sensors increases according to the formula

approximately wurzelförmig, but this is compensated by the mentioned noise voltage change in parallel connection of 2p branches, as stated above.

In addition, as explained above, the offset voltage is further reduced by meandering with the factor m, but less than by parallel connection. Due to the combination of m meander and 2p parallel elements, ie branches in the construction of a magnetic field sensor, two degrees of freedom are available, in order to best adapt this to existing specifications given given design rules.

The magnetic field sensor may include at least two branch groups each comprising at least two branches. This offers several combi ^¬ nation possibilities for magnetic field sensors with various measuring applications, as explained below.

This is especially true if the branches are constructed identically within a branch group. "Identical" means here at ^¬ not necessarily be completely identical, but also to the branches. B. not only for manufacturing tolerances of the MR sensors, but in the dimensioning and the different acting on them local magnetic field.

For in relation to the local behavior of the magnetic field closely adjacent branches within a branch group prevail then in general the same magnetic fields, so that set within the branch group, again apart from the manufacturing fluctuations, almost identical potenti ies at the voltage tap in each branch.

Exactly these differences can then z. B. be compensated by the above ge ^¬ called error reduction by averaging.

The spacing of individual branches within the branch group may be smaller than the spacing of the branch groups from each other.

In other words, the magnetic field sensor is then formed from at least two, relatively widely spaced branch groups, each branch group consists of at least two individual branches, which in contrast behave ^¬ tively close to each other. In the simplest case, that is with two branch groups of two branches, a structure is formed similar to a Wheatstone bridge, wherein each of the two bridge branches of the bridge has two paralle ^¬ le branches, each with a voltage tap. In comparison to the classical Wheatstone bridge, which supplies only a single bridge voltage, the arrangement ^according to the invention then provides a total of four bridge voltages, namely from the two voltage taps of the one branch group to the two voltage taps of the second branch group.

Of the four bridge voltages, three are independent of one another and, as described above, can advantageously be ^further processed, for example averaged.

As an alternative to the above, the intervals between individual branches within the branch group for various branch groups un ^¬ may be differently. With different branch groups can thus locally different field strengths or fields, namely corresponding to the respective different distance of their branches, to realize the magnetic field measurement. This is especially helpful when measuring gradient fields.

The branch groups can be arranged nested. Between the branches of a first branch group are then z. For example, all show one or more other topic groups. As a result, for example, a current sensor can be constructed according to the principle of a gradient bridge. In this principle, the further inside of a branch group associated MR elements, ie z. B. exposed in the nesting to innermost branch group, in pairs a lower magnetic field strength. These MR elements are thus saturated only at higher currents and thus a larger field gradient. The outside, ie far apart MR elements are then exposed to a larger field gradient and go earlier in saturation. The MR elements of inner and outer branch groups may be geometrically different. At low field strengths, however, the field gradient is low, may no longer be detectable by the innermost branch group because too small and below the working area ei ^¬ nes MR sensor, but then from the outermost branch group, which is relatively far from each other MR Elements in their workspace are captured.

Depending on the height of the shape anisotropy, stray field coupling, etc., which is predetermined by the layout, that is to say the arrangement of the branches, branch groups and the geometry of the MR elements, this effect can be intensified.

The magnetic field sensor thereby has a considerably extended measuring range. Due to an extension of the measuring range, it is z. B. possible to better predict the hysteresis at the offset by a suitable signal post-processing of the bridge voltage. This hysteresis component at the offset is caused by overcurrents or high magnetic fields, wel ^¬ che beyond the required measuring range z. B. may occur in special operating conditions or in the event of a fault.

Such a magnetic field sensor is fundamentally "open-loop" and / or "closed-loop" capable.

The magnetic field sensor can be arranged as an integrated magnetic field sensor together with evaluation electronics on a single chip. In the case of a geometrical arrangement of the measuring branches on the chip, the area requirement for chip area increases greatly in the presence of a plurality of voltage ^taps , since the voltage tap has to be conducted to a bonding pad for subsequent contacting. The bond pads are here in the

Usually much larger than an MR sensor itself or the transmitter.

By integrating the magnetic field sensor together with the evaluation electronics on a single chip, the bonding pads are avoided because the voltage taps can be performed directly on the off ^¬ evaluation electronics. The contacting of voltage taps of the transmitter can, for. B. by Through holes, so-called "vias" are produced .Thus, for example, a lying between the evaluation and MR sensor electrical insulation layer is overcome.

Each voltage tap in a branch may be assigned an amplifier stage. This too is not critical in the case of integration on a chip surface, since an amplifier stage can also be realized in a significantly smaller space than a bond pad or MR element. Because of the high input impedance of the amplifier stage, due to the amplifier stage at each voltage tap, a particularly good voltage measurement between the individual branches takes place with respect to the bridge voltages.

The MR sensors of individual branches may be different. So z. B. be covered by the integration of MR sensors with different sensitivities for magnetic fields large areas to be measured field strength. Corresponding MR sensors therefore have a greatly increased dynamic range compared to those which are constructed from the same MR sensors.

Individual branches may include ^¬ therebetween predetermined angle. This results in so-called MR angle sensors with which field angles in the magnetic field to be measured can be determined. Through the use of branches offset by certain angles within a branch group or between two branch groups, the amount of information per measurement cycle can be increased again. Thus, measuring errors or measurement inaccuracies, as explained above, can also be reduced with respect to the field angles.

The angles can z. B. in particular 30 °, 45 °, 60 ° or 90 °. These are common angles for field determination, in which particularly simple geometric angular relationships apply. The reworking of measured values is thereby considerably simplified. Further details and advantages of the invention will become apparent from the following description of exemplary embodiments with reference to the drawing in conjunction with the claims.

They show, in each case in a schematic outline sketch:

1 shows a magnetic field sensor in the form of a Wheatstone

Figure 2 shows a magnetic field sensor with three branch groups of two branches, Figure 3 is a diagram illustrating the sensitivity and measuring range

FIG. 4 shows an alternative magnetic field sensor with two groups of two branches each,

FIG. 5 shows an alternative embodiment for a branch from FIG

4 with two meander-shaped MR sensors, FIG. 6 shows the layout of a magnetic field sensor according to FIG. 4 and FIG

FIG. 5, FIG. 7, the layout of a magnetic field sensor according to FIG. 2 and FIG

FIG. 4 and FIG.

6 shows the layout of a magnetic field sensor in an alterna tive embodiment for ^¬ field angle measurement.

Figure 1 has already been described in detail in the presentation of the prior art. In the measuring arrangement with MR sensors described there, two branches are present. Branch groups are not formed there.

2 shows a magnetic field sensor 2 with three branch groups 10a, 10b, 10c are arranged each consisting of two branches 8a, 8b and 8c, 8d and 8e, 8f, the symmetrically to a symmetry plane SE ^¬ and in pairs in each case form a measuring bridge.

In FIG. 2, all the branch groups 10a to 10c are executed with branches 8a to 8f extending parallel to one another, the distances a, b, c of the branches within the branch groups 10a to 10c being different for each of these branch groups. they are. The branches 8a, 8b of the branch group 10a have the smallest distance 30a, the branch group 10b a greater distance 30b, and the branch group 10d the largest distance 30c.

In addition, the branch group 10c includes the two branch groups 10a and 10b between their branches 8e and 8f. The branch ^¬ group 10b, in turn, includes between its branches 8c and 8d, the branch group 10a. The branch groups 10a to 10c are thus arranged nested inside each other.

The arrangement in Figure 2 is characterized as MR Gradientenmessbrücke be ^¬, as the component, a gradient magnetic field to be measured ^¬ up 21 has, in or opposite to the direction of the arrow 32, also shown netfeld 21 by the locally different Mag ^¬ This leads to different field strength differences between the individual branches 8a to 8f of the various branch groups 10a to 10c. Thus, the branches 8a, 8b of the branch group 10a experience a significantly smaller one in the case of a given field gradient in the direction of the arrow 32

Field strength difference at its MR resistors 16 as the MR resistors 16 of the show 8e and 8f of the branch group 10c.

Therefore, as in FIG. 4, the voltages between different branch groups 10a to 10c are not tapped via the leads 20, as is the case in FIG. 4, but the three voltages U _ab , Ucd and U _e f occurring within the branch groups 10 a - c, ie the voltages Tensions between the branches 8a, 8b and 8c, 8d and 8e, 8f.

The individual MR sensors in FIG. 2 and the following figures have the reference numerals 16, where (i = 1, 2) stands for the two series-connected sensors of a measuring branch and k stands for the individual measuring branches. Each branch 10a - 10c thus consists of at least two series geschal ^¬ ended MR sensors 16i _D and _D 1œ2 _ac with an intermediate voltage tap 18a -18d for a bridge voltage U, _ad U, Practice or U _b d- Of these four bridge voltages are each independently of three ^¬.

The MR sensors 16 'of the branch groups 10a to 10c in FIG. 2 have pairs of different geometries (width of the strips). The different geometries (width of the strips) or sensor surfaces cause a different measuring range of the individual partial bridges.

The measurement sensitivity of a measurement arrangement according to Figure 2 it is located ^¬ of Figure 3. In Figure 3, the characteristics of the individual measurement branches are shown. Specifically, the abscissa is the measuring current I _M and the ordinate, the measurement voltage U _B ^¬ applied. The graphs 26, 27 and 28 characterize the characteristic curves of the individual bridges from the branch groups 10 ₁ .

In FIG. 3, A represents the working range and Ü the overflow range. The steepness of the graphs of the graphs associated with the individual bridges 10a, 10b and 10c from FIG

26, 27 and 28 shows that the measuring sensitivity with the Ab stands ^¬ a, b and c of the respective bridge measuring forming ^¬ branches increases.

Are 4 shows an alternative embodiment of a magnetic field sensor ^¬ 2 with two electrical terminals 4a and 4b, to which in turn supply lines 6a and 6b are connected, which lead from the magnetic field measuring device 2 to a not shown in the Figure 1 power supply. In the magnetic field sensor 2 are provided between the terminals 4a and 4b are a total of four electrical branches tet 8a to 8d parallel geschal ^¬. The branches 8a and 8b form a first branch group 10a, and the branches 8c and 8d form a second pointing group 10b. Inner ^¬ half of the branch groups 10a and 10b have the branches 8a and 8b and 8c and 8d to each other a distance d of 100 microns, where ^¬, however, the branch groups 10a and 10b have a distance e of 1500μm to each other. The distance d is therefore significant smaller than the distance d of the branch groups (d << e).

Each branch 8a to 8d in turn includes the series connection of two MR resistors 16ik and 162k with a respective voltage tap 18a to 18d arranged between them. All voltage taps 18a to 18d are connected via leads 20 to a transmitter, not shown.

Via the supply lines 20 and the voltage taps 18a to 18d, four bridge voltages Uac, U _ad , Übbe and Ubcu are respectively detected between the branches 8a and 8c, 8a and 8d, 8b and 8c and between the branches 8b and 8d , Instead of a single bridge voltage U _BE at the

Wheatstone measuring bridge according to Figure 1 according to the prior art so four independent voltages U _ac , Uacu Üb and Ubd are detected here.

The magnetic field measuring device 2 is intended to measure a magnetic field 21 "in plane", that is to say in the plane of the magnetic field sensor 2. The magnetic field 21 has a known local characteristic which ^subjects all elements within the branch groups 10a and 10b to the same field strength and field direction These differ only between the branch groups 10a and 10b.

Since the branches 8a and 8b or 8c and 8d are each dimensioned identically to ^¬ each other but by Fertigungsto ^¬ tolerances different from each other, the four measured voltages are dependent just from these manufacturing tolerances. The measurement accuracy of the magnetic field sensor 2 is therefore ge ^¬ increases when processed instead of a single bridge voltage according to the prior art, all four bridge voltages, z. B. be averaged. Of these four bridge voltages, three are linearly independent.

The number of bridge voltages can be achieved by further parallelization of the magnetic field measuring arrangement 2, ie by the insertion of additional branches in the branch groups 10a and / or 10b further increased and thus the measurement accuracy for magnetic field ^¬ measurement further improved.

FIG. 5 shows a specific embodiment of one of the branches 8a to 8d using the branch 8a as an example. The two, the voltage tap 18a enclosing MR resistors 16 _lk are not made in one piece, but each of three MR partial resistors 22i, 22 ₂ and 22 ₃ constructed, which are paral ^¬ lel and spaced from each other on a substrate, not shown. In order to form the actual or entire MR resistor 16 _lk , the MR partial resistors 22 _{k are} connected in series in meandering fashion at their ends via electrically conductive bridges 24 _x . The bridges 24 _k may alternatively be made of MR material.

The two end MR partial resistors 22i and 22 _{3 of} the upper MR resistor 16i _k in FIG. 5 are connected via the leads 26 and 28 to the terminal 4a on the one hand and to the voltage tap 18a on the other hand

FIG. 6 shows the structure according to FIG. 4, in which the MR resistors 16 _lk are designed in a meandering manner according to FIG. 2, in a realization as an integrated circuit. On a carrier material or substrate 40, the above-mentioned evaluation electronics 42 and the leads 20 and additional Sig ^¬ nalleitungen 44 are applied in the form of interconnects. The leads 20 and signal lines 44 are connected to the transmitter 42. In this case, each supply line 20 is assigned an amplifier stage 43, which amplifies the voltage present at the corresponding voltage tap, which here corresponds to the bond pads 52b, with a high input impedance.

The entire substrate 40 is nik together with the Auswerteelektro ^¬ 42 and the aforementioned conductor tracks of an intermediate layer 46 covers, which is shown cut away in Figure 4 in the field of electronics Auswerteelek ^¬ 42nd In turn, the eight MR resistors 16 as well as four bond pads 48a-d and printed conductors 50 are applied to the intermediate layer 46. The MR resistors 16 are ge in Mäandertechnik ^¬ Mäss Figure 2 but formed with five instead of that shown in Figure 2, three meander. Each MR resistor 16 _lk has at its two ends per contact areas 52a and 52b. In this case, the contact areas 52a are connected above the interconnects 46 via the interconnects 50 to the bonding pads 48a and 48b, which thus correspond to the terminals 4a, 4b of FIG.

The contact portions 52b are on the one hand above the layer between ^¬ 46 also through conducting paths 50 each in pairs, ver ^¬ connected so as to form the four branches from figure 1 to 8d 8a. On the other hand, in four of the contact regions 52b, namely one per branch 8a to 8d, four so-called "vias" or through-contacts 54 are arranged, which provide an electrical connection between the corresponding contact regions 52b through the intermediate layer 46 to the underlying ones Leads 20 ago.

As already described above, the evaluation electronics 42 carries out the evaluation of the four voltages U _ac , U _ad , U _bc and U _b d from FIG. The measurement result is output via the signal lines 44 which electrically connect the evaluation electronics 42 to the bonding pads 48c and 48d. For this, different variants are possible: For example, an overall mittelte bridge voltage to the bonding pads 48c and 48d abgege ^¬ be ben analog. Alternatively, a digitally coded voltage or a voltage profile can be output as a digitized measurement result. This can be z. B. the measured magnetic field strength or coded binary values of the four voltages just mentioned or otherwise further processed data enthal ^¬ th.

FIG. 7 shows, according to FIG. 6, the realization of a magnetic field sensor operating on the principle of FIG. In contrast to FIG. 3, however, only two branch groups 10a and 10b are interleaved with each other so as to detect various magnetic field gradients as described above. However, each branch corresponding to the individual branches 8a to 8d from FIG. 5 is again constructed according to FIG. 2 or FIG. 4 from the parallel connection of two individual branches and the MR resistors 16 are meandered. Within each ^¬ of the four branches 8a to 8d according to Figure 3 finds such. B. an averaging of the manufacturing tolerances of the MR resistors 16 _1D according to Figure 2 and Figure 4/5 instead.

In Figure 7 an evaluation electronics ^¬ not shown, is again present below the intermediate layer 46, which delivers to figure at the two bond pads 48c and 48d according to the Description 6 output information.

FIG. 8 shows the layout of a further embodiment of a magnetic field measuring arrangement 2, here with four parallel angle sensor bridges 60a to 60d. By the four pairs of two by 90 ° crosswise arranged branches 62a to 62h are in pairs each offset by 90 ° Feldkompo ^¬ nenten a magnetic field, indicated by the arrow 64, detected. The individual angle sensor bridges 60a to 60d are arranged point-symmetrically to a central point of symmetry SP.

The bond pads 48a to 48d in turn serve both the power supply of the measuring arrangement 2 and the delivery of the measurement information, as has already been described in detail.

The symmetrical design of the branches or branch groups optimum Ge ^¬ obtained with the inventive teaching accuracy of the measuring device - connected with maximum sensitivity and extended measuring range.

Claims

claims

1. Arrangement for magnetic field measurement with a measuring bridge of MR sensors with electrical connections (4a, 4b) for voltage supply, with at least two between the terminals (4a, 4b) connected in parallel, a measuring bridge forming branches (8a - 8c), wherein each branch (8a - 8c) at least two series-connected MR sensors (16) with a voltage tap located Between the seats ^¬ rule (18a -18d) for a bridge voltage (U _ac _ad, U, _bc U, U _b d) characterized in that at least three measuring branches (8a - 8f) are provided, of which at least two individual branches (8a - 8f) we ^¬ nigstens a branch group (10a - 10c), wherein the a ^¬ individual branch groups (10a - 10c ) are arranged in the context of metrological tolerances mirror or point symmetry to a single symmetry element (SE, SP).

2. Arrangement according to claim 1, characterized in that at least one of the MR sensors (16 _lk ) contains a series connection of a plurality of MR-part sensors (22 _1D ).

3. Arrangement r according to claim 2, characterized in that the MR sub-sensors (22 _1D ) meandering to a single MR sensor (16 _1D ) are interconnected.

4. Arrangement according to one of claims 1 to 3, characterized in that at least two branch groups (10a, 10b) each consist of at least two branches (8a - 8d).

5. Arrangement according to claim 4, characterized in that the branches (8a - 8d) are constructed identically within a branch group (10a, 10b).

6. Arrangement according to claim 4 and / or claim 5, characterized in that the distance (d) of individual branches (8a - 8d) within the branch group (10a, 10b) smaller than the Ab ^¬ stand (e) of the branch groups (10a , 10b) to each other.

7. Arrangement according to claim 4 and / or claim 5, characterized in that the distance (a, b, c) of individual branches (8a

- 8f) within the branch group (10a - 10c) is different for different branch groups (10a - 10c).

8. Arrangement according to claim 7, characterized in that the branch groups (10a - 10c) are arranged interleaved.

9. Arrangement according to claim 4, characterized in that the accuracy of the magnetic field measurement is inversely proportional to the resistance noise. (Equation 1, equation 2)

10. Arrangement according to claim 9, characterized in that for the offset error, which is caused by a statistical deviation (ΔR) of the resistor (R), the following applies:

where 2p is the number of measuring branches and m is the number of meanders of an MR sensor (16 _lk ).

11. Arrangement according to claim 9 and 10, characterized in that the accuracy of the number (z = 2p) of the measuring branches (8a

- 8e) on the one hand and the number (m) of the individual magnetic meander (22 _1D ) is dependent.

12. The arrangement according to claim 4, characterized in that the branches (8a - 8d) of different branch group (10a, 10b) are constructed un ^¬ different.

13. Arrangement according to claim 12, characterized in that the measuring range and / or the measuring sensitivity of

Measuring bridges are dependent on the formation of the measuring branches.

14. Arrangement according to one of Ansprüchel2 or 13, characterized in that the measuring range, which is determined by the geometry (width / area) of the MR sensors (16 _lk ), is vorgeb ^¬ bar.

15. An arrangement according to claim 14, characterized in that the width of the MR sensors (16 ^) in the measuring groups formed by the Zweiggrup ^¬ pen (10a - 10c) with the distance (a, b, c) of the measuring branches (8a - 8e) increases.

16. Arrangement according to one of the preceding claims, characterized in that the measurement sensitivity depends on the distance (a, b, c) of a measuring bridge (10a - 10c) form ^¬ the measuring branches (8a - 8Of).

17. Arrangement according to one of the preceding claims, characterized in that the integrated magnetic field ^¬ sensor together with a transmitter (42) on a single chip (40) is arranged.

18. Arrangement according to claim 17, characterized in that each voltage tap (18a-d) in a branch (10a - 10c) ei ^¬ ne amplifier stage (43) is associated.

19. Arrangement according to one of the preceding claims, characterized in that the MR sensors (16) of individual branches (8a - 8d) are different.

20. Arrangement according to one of the preceding claims, character- ized in that individual branches (8a - 8d) include between them predetermined angle.

21. Arrangement according to claim 20, characterized in that the angles are 30 °, 45 °, 60 ° or 90 °.