CN100414255C - Method of calculating compensation value for angle detecting sensor and angle detecting sensor using the method - Google Patents

Abstract

To provide a compensation value calculation method of an angle detection sensor, capable of obtaining accurate angle output, even when that error signals are contained in the output signal from the sensor, and to provide an angle detection sensor that uses this. A calculation part 14 calculates the rotation angle [phi], before the compensation of an object to be measured, including total error signal [Delta][phi] from SIN signal and COS signal output from a sensor means 1 and a signal converter 12. A control part 11 extracts a signal that minimizes the remaining energy E from the first candidate signal S1, stored to a memory part 16 as a phase compensation value S<SB>[alpha]</SB>. In the same manner, a strain compensation value S<SB>[beta]</SB>is extracted from the second candidate signal S2, and gain compensation value S<SB>[gamma]</SB>is extracted from the third candidate signal S3. By removing S<SB>[alpha]</SB>, S<SB>[beta]</SB>, S<SB>[gamma]</SB>from the total error signal [Delta][phi], an accurate angle output [phi]<SB>OUT</SB>can be detected.

Description

The compensation value calculation method of angle detecting sensor and the angle detecting sensor that uses it

Technical field

The present invention relates to use by the magneto-resistance effect element of GMR element representative the compensation method of the detected anglec of rotation and the angular transducer that possesses this method, particularly, even relate under the situation of the error signal that contains phase error and strain error etc. between signal, also can improve the compensation value calculation method of angle detecting sensor of accuracy of detection of the output angle that is detected and the angle detecting sensor that uses it by magneto-resistance effect element output.

Background technology

The detection of the output angle of automobile steering wheel etc. is to utilize with the wheel disc of rotating members rotation synchronously such as steering axle and angle detecting sensor etc. to carry out.In the sensor part of described angle detecting sensor, adopt perception magnetic and the magneto-resistance effect element of output signal output, as the prior art document that uses such magneto-resistance effect element, for example following patent documentation 1,2,3 and 4 etc. are arranged.

Figure 16 is the planimetric map of structure of expression angle detecting sensor 100, angle detecting sensor 100 be provided with respect to the rotating disk 102 of described rotation center O rotation with and inner assembly 101.

In the inside of described assembly 101 with respect to described rotation center O and the position of symmetry (in 90 ° the position of circumferentially staggering mutually of rotation center O) is respectively equipped with 4 chip substrates (wafer) K1, K2, K3, K4.On a chip substrate, two conducts respectively are set (are expressed as G1～G8) respectively with the GMR element that is configured to basic magneto-resistance effect element that is laminated by exchange-biased layer (antiferromagnetism body layer), fixed bed (pin fixed bed), nonmagnetic layer, free layer (free magnetic layer).

That is, on described chip substrate K1, GMR element G1 and G2 are set, GMR element G3 and G4 are set on chip substrate K2, GMR element G5 and G6 are set on chip substrate K3, GMR element G7 and G8 are set on chip substrate K4.Mounting is in each GMR element G1～G8 of each chip substrate K1～K4, under GMR element G1 and GMR element G4 is connected in series and GMR element G3 and GMR element G2 are connected in series state, both are connected in parallel and constitute first bridge circuit.Equally, under GMR element G5 and GMR element G8 is connected in series and GMR element G7 and GMR element G6 are connected in series state, both are connected in parallel and constitute second bridge circuit (with reference to Figure 16).

Described magnet M1 and M2 are fixed in the inner face of described rotating disk 102.With its N utmost point of a magnet M1 towards rotation center O and its S utmost point of another magnet M2 towards rotation center O state and be fixed, between described magnet M1 and magnet M2, produce certain external magnetic field H.

Rotating member as detected material is rotated, and makes described rotating disk 102 rotations, and then described magnet M1, M2 rotate in a circumferential direction assembly 101.At this moment, the direction of magnetization of the free layer of each GMR element G1～G8 changes along with described external magnetic field H.Thus, because the resistance value of described each GMR element G1～G8, change according to the direction of magnetization of described free layer and the direction of magnetization angulation of described fixed bed, so from the described first bridge circuit output+sin signal and-the sin signal, simultaneously from the output of described second bridge circuit from from first bridge circuit ± the sin signal phase stagger 90 °+the cos signal and-the cos signal.

Control part, in these four signals, by differential amplification described+the sin signal and-the sin signal generates SIN signal (sine wave signal), and by differential amplification described+the cos signal and-the cos signal generates COS signal (cosine wave signal).Next, described control part is calculated tangent value (tan) by described SIN signal (sine wave signal) and COS signal (cosine wave signal), and calculates arc-tangent value (arctan) by asking, and can detect the output angle of described rotating member.

(patent documentation 1) spy opens the 2002-303536 communique.

(patent documentation 2) spy opens the 2000-35470 communique.

(patent documentation 3) spy opens the 2003-106866 communique.

(patent documentation 4) spy opens the 2003-66127 communique.

In described angle detecting sensor 100,, need keep 90 ° of phase differential between described sine wave signal and the cosine wave signal accurately in order to detect the rotation angle of rotating member accurately.So, direction of magnetization (magnetized towards) h that for this reason is located at the described fixed bed of two GMR elements on the same chip substrate is with same direction manufacturing, therefore for example the direction of magnetization h of chip substrate K1 is made as+the Y direction, then need be between adjacent chip substrate, described direction of magnetization h staggers with high-precision 90 ° of intervals mutually and installs, making the direction of magnetization h of chip substrate K2 is-the Y direction that the direction of magnetization h of chip substrate K3 is+directions X that the direction of magnetization h of chip substrate K4 is-directions X.

; because the direction of magnetization h of the fixed bed of GMR element G1～G8 can not confirm by visual; so be difficult to each chip substrate K1～K4 is accurately staggered 90 ° and be installed on the assembly with described direction of magnetization; can not correctly set under the described 90 ° situation; exist phase error to produce, and can not detect the problem of the rotation angle (output angle) of described rotating member accurately as 90 ° ± α.

In addition, in each chip substrate quilt cutting accurately, one side and under the direction of magnetization h of the GMR element G1～G8 situation about forming with respect to the parallel accurately of chip substrate, though can be by utilizing for example device of the compensation setting angle of pattern recognition device etc., and on assembly with chip substrate each other with 90 ° of installations, thereby described direction of magnetization h is correctly staggered 90 °, but exist the manufacturing cost of described chip substrate to rise easily in this case, it is complicated that assembling procedure during installation becomes, the problem that built-up time and assembly cost increase.

On the other hand, in described angle detecting sensor 100, input angle (the magnet anglec of rotation) θ that comparatively it is desirable to 102 rotations of output angle φ and described rotating disk is accurately proportional and export, but the strain error of sinuous signal overlap appears in actual output angle φ on the straight line that changes with linear function, and described output angle φ does not accurately form ratio (with reference to Fig. 6) with described input angle θ.

The generation reason of this strain error, be to cause known for everyone by the distinctive resistance value strain of described GMR element, if in described four signal waveforms, produce this strain error respectively as the output of described angle detecting sensor 100, then in described SIN signal and COS signal, also can produce strain error, and, can not improve from the problem of the precision of the detected output angle φ of angle detecting sensor so exist because in the calculating of described tangent value (tan) and arc-tangent value (arctan), also be subjected to the influence of described strain error.

If described phase error and strain error β can be with given approximations to function, and can utilize this approximate function that the output angle φ from described angle detecting sensor 100 outputs is compensated with pursuing angle, then can improve the precision of described angle output significantly.But, can not easily calculate the penalty coefficient that constitutes given function, in above-mentioned patent documentation 1,2,3 and 4, there is not record about the method that is compensated coefficient yet.

The present invention is intended to solve above-mentioned existing problem, and purpose is to provide the computing method of the offset of the phase error that comprised in a kind of output of the detecting sensor of offset angle in advance and/or strain error equal error signal.

In addition, the object of the present invention is to provide a kind ofly, utilize offset that described compensation value calculation method calculates and the compensating error signal by use, thereby improve the angle detecting sensor of the accuracy of detection of output angle.

Summary of the invention

The present invention is a kind of compensation value calculation method of angle detecting sensor, this angle detecting sensor has: sensor assembly, it produces rotating magnetic field according to the rotation that puts on determinand, and the described rotating magnetic field of perception and export a plurality of output signals corresponding to the input angle of described rotation; Signal transformation portion, it is transformed into two kinds of signals with given phase differential from described a plurality of output signals; The signal adjustment part, it adjusts the skew and the gain of described two kinds of signals; Operational part, it is according to described adjusted two kinds of signals, with the anglec of rotation of described determinand before by way of compensation output angle and calculate; Storer, it stores the offset of the total error signal that is comprised in the preceding output angle of the described compensation of compensation in advance; Compensation section, it is by removing the output angle that described offset calculates described determinand in the total error signal that is comprised from the output angle before the described compensation,

Have:

First step generates two kinds of signals with given phase differential from described a plurality of output signals;

Second step is calculated the anglec of rotation of the described determinand output angle before by way of compensation according to described two kinds of signals;

Third step, in the total error signal that from the output angle before the described compensation, is comprised, when removing the different a plurality of candidate signal of cycle same-amplitude one by one, will be that the candidate signal of minimum is extracted out as described offset with the residual energy that is comprised in the signal after removing.

In the present invention, can calculate the The optimal compensation value that is used for compensating the preceding total error signal that output angle comprised of described compensation.

In above-mentioned, described output angle is being made as θ, the total error signal that is comprised in the output angle before the described compensation is made as Δ φ, described offset is made as S _n(n is α, β or γ), and will from described total error signal Δ φ, remove described offset S _nAfter signal in the one-tenth that comprised be divided into e (=Δ φ-S _n) time, described residual energy E is calculated by following formula 4.

(formula 4)

$E={\&Integral;}_0^{2\mathrm{\&pi;}}{e}^2\mathrm{d\&theta;}$

Said method finds to be similar to the offset (function) of error signal by adopting vague generalization harmonic analysis (GHA:Generalized HarmonicAnalysis), can obtain best offset thus.

For example, with the phase differential that is produced between described two kinds of signals, the offset when becoming 90 ° ± α add phase error ± α on 90 ° of given phase differential after is made as S _αThe time, described offset S _αApproximate by shown in the following formula 5.

(formula 5)

$S_{\mathrm{\&alpha;}}=\frac{\mathrm{\&alpha;}}{2}\{1+\mathrm{<figure-callout\; id="2"\; label="COS"\; filenames="C20051013144400261.png,C20051013144400262.png"\; state="\{\{state\}\}">COS</figure-callout>}2{\mathrm{\&phi;}}_{\mathrm{\&alpha;}}\}$

Wherein, ${\mathrm{\&phi;}}_{\mathrm{\&alpha;}}=\arctan \left\{\frac{\sin (\mathrm{\&theta;}+\mathrm{\&alpha;})}{\cos \mathrm{\&theta;}}\right\}$

In said method, can precompute the caused detected phase error delta of the phase error φ that removes by being produced between two kinds of signals _αPhase compensation value S _α

Perhaps, will be made as S for the offset of the strain error β that produces because of resistance value strain that described sensor part had _βThe time, described offset S _βCan be similar to by shown in the following formula 6.

(formula 6)

S _β＝-β·sin4φ _β

Wherein, ${\mathrm{\&phi;}}_{\mathrm{\&beta;}}=\arctan \left\{\frac{\sin \mathrm{\&theta;}}{\cos \mathrm{\&theta;}}\right\}$

In said method, can precompute and remove the caused detection strain error of the inherent strain error β φ that is had by the GMR element _βStrain compensation value S _β

Perhaps, will be made as S for the offset of gain error with amplitude γ _γThe time, described offset S _γCan be similar to by S _γ=γ sin2 θ represents.

In said method, can precompute and remove by the caused gain error signal Δ of fault in enlargement γ φ _γGain compensation value S _γ

In addition, angle detecting sensor of the present invention is characterised in that, the described offset that utilizes the described compensation value calculation method that each is recorded and narrated to be calculated is deposited into described memory section, and compensates with the offset that reads from the described storer output angle to described determinand.

In the present invention, by from the output angle before the compensation that calculates by operational part, removing the offset that precomputes, and can remove the error signal that forms by the phase error that comprises described in the output angle that has before the described compensation and strain error etc. substantially.For this reason, can improve the precision of the output angle that detects from angle detecting sensor.

In above-mentioned, preferably in described storer, deposit described offset S in _α, S _βAnd S _γIn at least more than one offset.

In said method, because can remove the detected phase error delta φ that is comprised in the total error signal that is comprised in the preceding output angle of described compensation _α, detect strain error Δ φ _β, and gain error signal Δ φ _γPart or all, so can further improve the accuracy of detection of angle detecting sensor.

In above-mentioned, described sensor assembly is preferably two groups the bridge circuit that is formed by magneto-resistance effect element.

In said method, by using magneto-resistance effect element, can miniaturization.And, can provide the sensor of high reliability by using bridge circuit.

Described two kinds of signals are SIN signal and COS signal, be preferably after described operational part goes out tangent value from described SIN signal and described COS calculated signals, by obtaining arc-tangent value, with the anglec of rotation of described determinand before by way of compensation output angle and calculate.

In above-mentioned steps, can positively obtain the anglec of rotation of described determinand.

Described in addition signal adjustment part, described operational part and described compensation section are preferably by a calculation process module and form.

In said apparatus, can under the management of described control part, concentrate processing, and structure that can small-sized intensification angular transducer.

In the computing method of the offset of angle detecting sensor of the present invention, can precompute the offset that the error signal to phase error contained in the output angle before compensating and strain error etc. compensates.

Because use the error signal of calculating in advance, can remove the error signal of phase error contained in the preceding output angle of compensation and strain error etc., so can improve the precision of the angle output of angle detecting sensor.

Description of drawings

Fig. 1 is the structural drawing of the structure of expression angle detecting sensor of the present invention.

Fig. 2 is from the oscillogram of the relation of four signals of two groups bridge circuit output under the rational situation of expression.

Fig. 3 is the oscillogram that is illustrated in when comprising phase error between the SIN signal of the output of bridge circuit and the COS signal.

Fig. 4 is with the output angle φ before the compensation after the ATAN processing _αThe oscillogram of representing as continuous functions.

Fig. 5 is the output angle φ before expression will compensate _αWith the difference of input angle θ as detected phase error delta φ _α(=φ _α-θ) and the expression oscillogram.

Fig. 6 is the angle φ before the compensation of expression when comprising the strain error that produces based on the resistance value strain _βOscillogram.

Fig. 7 is that expression detects strain error Δ φ _β(=φ _β-θ) oscillogram.

Fig. 8 is expression total error signal Δ φ (=Δ φ _α+ Δ φ _β) the oscillogram of an example.

Fig. 9 is another routine oscillogram that expression compensates the total error signal Δ φ that is comprised among the preceding output angle φ.

Figure 10 is expression phase compensation value S _αOscillogram with its candidate signal.

Figure 11 is the error signal Δ φ after the phase compensation _-α(=Δ φ-S _α) oscillogram.

Figure 12 is expression strain compensation value S _βAnd the oscillogram of candidate signal.

Figure 13 is the error signal Δ φ behind phase compensation and the strain compensation _-alpha-beta(=Δ φ _-α-S _β) oscillogram.

Figure 14 is expression gain compensation value S _γAnd the oscillogram of candidate signal.

Figure 15 is the error signal Δ φ behind expression phase place, strain and the gain compensation _{-alpha-beta-γ}Oscillogram.

Figure 16 is the planimetric map of the structure of expression angle detecting sensor.

Among the figure: the 1-sensor assembly, the 10-signal processing module, the 11-control part, 12-signal transformation portion, the 12A-first signal transformation portion, 12B-secondary signal transformation component, 13-signal adjustment part, the 14-operational part, 15-compensation section, 16-memory section, the 101-assembly, the 102-rotating disk, e, e1, e2, e3-residual error composition, the direction of magnetization of h-fixed bed (magnetized towards), G1～G8-GMR element (magneto-resistance effect element), the H-external magnetic field, K1, K2, K3, the K4-chip substrate, M1, M2-magnet, the S1-first candidate signal (being used to extract out the aggregate of the signal of phase compensation value), S2-second compensating signal (being used to extract out the aggregate of the signal of strain compensation value), S3-the 3rd compensating signal (being used to extract out the aggregate of the signal of gain compensation value), S _α-phase compensation value, S _β-strain compensation value, S _γ-gain compensation value, WB1-first bridge circuit, WB2-second bridge circuit, α-phase error, β-strain error, the amplitude of γ-gain error, θ-input angle, the output angle (output of operational part) before φ-compensation, φ _αOutput angle (output of operational part) before-the compensation when comprising the detected phase error, φ _β-comprise the output angle (output of operational part) before the compensation when detecting strain error, Δ φ-total error signal, Δ φ _α-detected phase error, Δ φ _β-strain detecting error, Δ φ _γ-gain error, Δ φ _-αError signal after the-phase compensation, Δ φ _-alpha-betaError signal behind-phase compensation and the strain compensation, Δ φ _{-alpha-beta-γ}Error signal behind-phase place, strain and the gain compensation, Δ φ _OUT-angle output (output of angle detecting sensor).

Embodiment

Fig. 1 is the frame assumption diagram of the structure of expression angle detecting sensor of the present invention.Fig. 2 is relation in ideal conditions the oscillogram of expression from four signals of two groups bridge circuit output.Below illustrated angle detecting sensor, be the device of output angle of the rotating members such as steering axle of test example such as automobile.

Angle detecting sensor shown in Figure 1 has sensor assembly 1 and the signal processing module 10 to handling from the output signal of described sensor assembly 1 output.

The structure of described sensor assembly 1 is identical with device illustrated in above-mentioned " background technology " hurdle.That is to say that as shown in figure 16, described sensor assembly 1 has: can rotate freely and the rotating disk 102 established with respect to rotation center O; With four chip substrates of mounting (wafer) K1, K2, K3 and K4, and be fixed in the assembly 101 of the inside of described rotating disk 102.Described four chip substrate K1, K2, K3 and K4 are set at respectively in the described assembly 101 with respect to described rotation center O and the position of symmetry, promptly in the position of circumferentially staggering mutually with 90 ° of intervals of rotation center.

On a chip substrate, be provided with each two and (be expressed as G1～G8) respectively for the GMR element of basic magneto-resistance effect element with the structure (not shown) that is laminated by exchange-biased layer (antiferromagnetism body layer), fixed bed (pin stop layer), nonmagnetic layer, free layer (free magnetic layer).

Because described chip substrate, hang up the external magnetic field with the state of a plurality of GMR element film forming on a large substrate, described fixed bed magnetized after (direction of magnetization) is unified into a direction and be chip substrate K1～K4 by cutting respectively, so be located at two GMR elements on the chip substrate the direction of magnetization of fixed bed identical.So each chip substrate K1～K4 is fixed in the described assembly 101, make described direction of magnetization between adjacent substrate, have about 90 ° of relations.In addition, described 90 ° of relations are accurately preferred, but because the phase error that causes thus can compensate by mode described later, so also do not need to have high-precision 90 ° of relations.

Mounting constitutes two groups the bridge circuit that is made of the first bridge circuit WB1 and the second bridge circuit WB2 in each GMR element G1～G8 of described chip substrate K1～K4.As shown in Figure 1, the first bridge circuit WB1 is by constituting at GMR element G1, the G2 and G3, the G4 that are located at relative rotation center O and become institute's mounting on the chip substrate K1 of axisymmetric position (with reference to Figure 16) and the chip substrate K2.That is to say that the first bridge circuit WB1 is the circuit be connected in series GMR element G1 and GMR element G4,, be connected in parallel and form with the circuit of be connected in series GMR element G3 and GMR element G2.Equally, the second bridge circuit WB2 is by constituting at GMR element G5, the G6 and G7, the G8 that are set at relative rotation center O and become institute's mounting on the chip substrate K3 of axisymmetric position and the chip substrate K4.The second bridge circuit WB2 is the circuit by be connected in series GMR element G5 and GMR element G8, with the circuit of be connected in series GMR element G7 and GMR element G6, is connected in parallel and forms.

And, the end of the described first bridge circuit WB1 that is connected in parallel and the second bridge circuit WB2 and power supply V _CcBe connected, the other end is connected with ground connection GND.

Described rotating disk 102 be connected by for example gear etc. as the rotating member (steering axle etc.) of determinand, and constitute with rotation and make described rotating disk 102 rotations according to rotating member.Therefore, if make described rotating member rotation, then described rotating disk also rotates, thereby described magnet M1, M2 can rotate around described assembly 101.

At this moment,, each the GMR element G1～G8 in the assembly 101 are paid rotating magnetic field, so the magnetization of each free layer that forms each GMR element G1～G8 is towards changing because result from external magnetic field H between described magnet M1, the M2.Thus, the resistance value of described each GMR element G1～G8, along with the magnetization of described free layer towards changing towards the angle that is become with the magnetization of described fixed bed.Therefore, from the GMR element G3 that constitutes the described first bridge circuit WB1 and the connecting portion of GMR element G2,, export the different sine wave signal of 180 ° of two mutual phase phasic differences with the connecting portion of GMR element G1 and GMR element G4.Simultaneously, from the GMR element G7 that constitutes the described second bridge circuit WB2 and the connecting portion of GMR element G6,, export the different sine wave signal of 180 ° of two mutual phase phasic differences with the connecting portion of GMR element G5 and GMR element G8.

But, because with the chip substrate K1, the K2 that dispose with rotation center O rotational symmetry, with the chip substrate K3, the K4 that dispose with identical rotation center rotational symmetry, also be configured in respect to described rotation center O to become roughly on 90 ° the diverse location, so establish from shown in first bridge circuit circuit WB1 output two signals be+sin signal,-sin signal, then the signal from the output of described first bridge circuit be+the cos signal and-cos signal (with reference to Fig. 2).

Shown in this embodiment, when for example described rotating disk 102 rotates in a clockwise direction, be made as+the sin signal from the sine wave signal of the connecting portion output of the GMR element G3 of the described first bridge circuit WB1 and GMR element G2, then from the connecting portion output-sin signal of described GMR element G1 and GMR element G4.If this moment is from the described GMR element G7 of the described second bridge circuit WB2 and the connecting portion output+cos signal of GMR element G6, then from the connecting portion output-cos signal of described GMR element G5 and GMR element G8.

Described signal processing module 10 mainly has control part 11, signal transformation portion 12, signal adjustment part 13, operational part 14, compensation section 15, memory section 16 etc.

Described control part 11 constitutes main body by CPU, and has the function of a series of signal Processing that comprises described signal adjustment part 13, operational part 14 and compensation section 15 etc.

Described signal transformation portion 12 has first 12A of signal transformation portion and secondary signal transformation component 12B, and is respectively equipped with first, second differential amplifier 12a, 12a and A/D transducer 12b, 12b.In described first 12A of signal transformation portion, described differential amplifier 12a will from described two kinds of first bridge circuit WB1 output+the sin signal and-the sin signal carries out differential amplification, formation is by the sin signal that the amplitude of twice constitutes, follow after described A/D transducer 12b will amplify signal with fixed sampling period be transformed to after the A/D conversion SIN signal (digital signal).

Equally, in the differential amplifier 12a of described secondary signal transformation component 12B, described differential amplifier 12a will from described two kinds of second bridge circuit WB2 output+the cos signal and-the cos signal carries out differential amplification, formation is by the cos signal that the amplitude of twice constitutes, follow signal transformation after described A/D transducer 12b will amplify and be after the A/D conversion COS signal (digital signal) (first step).

At this if will be for example A1, A2, B1 and B2 as amplitude coefficient, with a1, a2, b1 and b2 as the skew (offset) coefficient, with described+sin signal as+A1sin θ+a1, with described-sin signal as-A2sin θ-a2, with described+cos signal as+B1cos θ+b1, with described-cos signal as-B2cos θ-b2 and represent, the described SIN signal that generates by described first 12A of signal transformation portion then, become (+A1sin θ+a1)-(the sin θ of A2sin θ-a2)=(A1+A2)+(a1+a2).The same described COS signal that generates by described secondary signal transformation component 12B, become (+B1cos θ+b1)-(the cos θ of B2cos θ-b2)=(B1+B2)+(b1+b2).

Described signal adjustment part 13 is carried out the skew adjustment and the gain of described SIN signal and described COS signal and is adjusted (amplify and adjust), and has the benchmark (0 point) and amount (amplitude amount) consistent function of the amplitude direction that makes two signals.That is to say, mention above-mentioned example, the amplitude coefficient (B1+B2) that expression makes amplitude coefficient (A1+A2) unanimity of SIN signal or is similar to the COS signal is adjusted in so-called gain, expression is adjusted in so-called skew, as a1+a2=0 and b1+b2=0, benchmark with amplitude is positioned at the mode of origin position (0 point) and adjusts the dislocation that causes thereby elimination is overlapping by bias voltage.

In addition, this adjustment, the amplification coefficient g that sets in (A1+A2)=g (B1+B2) is for example further undertaken by read the deviation ratio that is made as (a1+a2)-gs=0 and (b1+b2)-gc=0 from memory section 16.Therefore, the SIN signal of this time point becomes (A1+A2) sin θ, and the COS signal is roughly consistent (wherein, with (B1+B2) cos θ

).

Described operational part 14 is equipped with and carries out sin, cos, tan, tan ^-1=arctan, sinh, cosh, exp, the software of function calculation such as log and other calculating, for example utilize the numerical evaluation software of known CORDIC (Coordinate Rotation Digital Computer) algorithm, here, have by described SIN signal is handled divided by the TAN that described COS signal calculates tangent value (tan=SIN signal/COS signal=sin θ/cos θ), and from handle by described TAN calculate arc-tangent value (arctan (sin θ/cos θ)) the value of being tried to achieve and obtain output angle Φ before the compensation of determinand (°) ATAN handle.

The compensation section 15 of this embodiment is made of the functional operation module of using the cordic algorithm identical with above-mentioned operational part 14, and carries out calculation process as described later.

In addition, be preferably by a calculation process module and constitute described signal adjustment part 13, described operational part 14, described compensation section 15, described each calculating can be concentrated under the management of described control part 11 by using a described calculation process module be located in reason, and can small-sizedly intensify.

In described memory section 16, according to the data of measuring by described inspection operation before dispatching from the factory and predetermined described amplification coefficient g, described deviation ratio gs, gc, and phase compensation value S described later _α, strain compensation value S _βWith gain compensation value S _γDeng offset etc., in each angular transducer by in the write store 16 in advance.

Next, the computing method of above-mentioned offset are described.

Fig. 3 is the oscillogram when comprising phase error between the SIN signal COS signal of expression as the output of bridge circuit.

Be located in formation between the adjacent substrate of the direction of magnetization of described fixed bed of GMR element G1 to G8 of chip substrate K1～K4, fail to keep under 90 ° the situation of relation, respectively from the described 12A of first, second signal transformation portion, between the SIN signal and COS signal of 12B output, except original phase differential (90 °), also contain phase error based on the dislocation of described direction of magnetization.

The anglec of rotation of described rotating disk 102 will be paid, the input angle of promptly paying described angle detecting sensor 100 is made as θ, with this moment phase error (from the amount of staggering of 90 ° of phase differential) be made as α (°), and will from described second bridge circuit WB2 output+the cos signal is made as+cos θ, general-cos signal is made as-cos θ.In addition, the convenience in order to illustrate is made as A1=A2=B1=B2=1 with described amplitude coefficient A1, A2, B1 and B2 here, and described deviation ratio a1, a2, b1 and b2 are made as a1=a2=b1=b2=0.Even so, the result is handled by described signal adjustment part 13 and TAN to have produced same effect, promptly because described each coefficient by cancellation, so there is not special problem.

＜first step 〉

At first, in first step, calculate SIN signal and COS signal.That is to say, if with from described second bridge circuit WB2 output+cos θ and-cos θ is a benchmark, then can with from described first bridge circuit WB1 output+the sin signal indication is+sin (θ+α), general-sin signal indication be-sin (θ+α).Therefore, become+sin (θ+α)-(sin (θ+α))=2sin (θ+α) from the SIN signal of described first 12A of signal transformation portion output, become+cos θ-(cos θ)=2cos θ from the COS signal of described secondary signal transformation component 12B output, if to illustrate it, (first step) then as shown in Figure 3.In addition, in Fig. 3 to Fig. 8,, represented the situation of α=+ 5 ° as the example of described phase error.

From the described 12A of first, second signal transformation portion, the described SIN signal and the COS signal of 12B output are offset in described signal adjustment part 13 and adjust and gain (amplification) adjustment.

＜the second step 〉

In second step, calculate the preceding output angle φ of compensation _αThat is to say that the TAN by described operational part 14 handles, ((θ+α)/cos θ calculates θ+α)/2cos θ=sin as 2sin with tangent value (tan=SIN signal/COS signal).In addition, though handle to calculate arc-tangent value (arctan) by ATAN, the output of the described operational part 14 when comprising phase differential α is with its output angle φ before as the compensation of determinand _αOutput angle φ before then compensating _αAs φ _α=arctan (sin (θ+α)/cos θ) and calculate (second step).

Fig. 4 is comprising under the situation of phase error, and ATAN is handled output angle φ before the post-compensation _αThe oscillogram of representing as continuous functions.Fig. 5 is with the output angle φ before the compensation _αAnd the difference between the input angle θ is as detected phase error delta φ _α(=φ _α-θ) and the expression oscillogram.

In Fig. 4, the output angle φ before the output angle θ of each rotation and the compensation _αCorresponding one by one., the output angle φ before the compensation after ATAN handles _αBecome in that (φ=θ) goes up the equitant shape of trigonometric function as a straight line of desirable output angle.

Here, if will be by the output angle φ before described compensation _αDeduct input angle θ and detected phase difference _α-θ is as detected phase error delta φ _αAnd obtain, then be detected as sinuous periodic function (trigonometric function) shown in Figure 5.Promptly as can be known, the output angle φ before the compensation after ATAN handles _αIn, include sinusoidal wave shape detected phase error delta φ shown in Figure 5 _α-(=φ _α-θ).

If to described detected phase error af _αAnalyze, then described as can be known four signals (+sin signal ,-the sin signal ,+the cos signal ,-the cos signal) or described SIN signal and COS signal with one-period as 360 ° (with reference to Fig. 2 and Fig. 3), to this, described detected phase error delta φ _αWith one-period as 180 ° (with reference to Fig. 5).Promptly as can be known, under with described four signals or described SIN signal and the situation of COS signal as baseband signal, described detected phase error delta φ _α, become the 2 frequencys multiplication trigonometric function in (1/2 cycle) with respect to this baseband signal.

Therefore, detected phase error delta φ _αCan be similar to by with the represented trigonometric function of following formula 7.Represent detected phase error delta φ so approx _αSignal be called phase compensation value S _α

(formula 7)

$S_{\mathrm{\&alpha;}}=\frac{\mathrm{\&alpha;}}{2}\{1+\mathrm{<figure-callout\; id="2"\; label="COS"\; filenames="C20051013144400261.png,C20051013144400262.png"\; state="\{\{state\}\}">COS</figure-callout>}2{\mathrm{\&phi;}}_{\mathrm{\&alpha;}}\}\&ap;\mathrm{\&Delta;}{\mathrm{\&phi;}}_{\mathrm{\&alpha;}}$

Wherein, ${\mathrm{\&phi;}}_{\mathrm{\&alpha;}}=\arctan \left\{\frac{\sin (\mathrm{\&theta;}+\mathrm{\&alpha;})}{\cos \mathrm{\&theta;}}\right\}$

In addition, described phase compensation value S _αThe amplitude amount of full width (peak to peak), be equivalent to the α of described phase error.

Next, to based on resistance value strain that each GMR element had and the strain compensation value S that produces _βComputing method describe.

Fig. 6 is the output angle φ before the compensation of expression when containing the strain error that produces based on resistance value _βOscillogram, Fig. 7 is that expression detects strain error Δ φ _β(=φ _β-θ) oscillogram.In addition, in Fig. 7, as detecting strain error Δ φ _βExample, show strain error β and be the situation of β=5 degree.

Constitute each GMR element G1～G8 of the described first bridge circuit WB1 and the second bridge circuit WB2, have fixing resistance value strain separately.For this reason, from described four signals of described first, second bridge circuit WB1, WB2 output, include the detection strain error Δ φ that produces based on described resistance value strain respectively _βSo, use to comprise this detection error delta φ _βFour signals, by method same as described above, carry out promptly that TAN handles and ATAN handles, the output angle φ before the compensation that is calculated _β, be represented as desirable output angle by a straight line shown in the dot-and-dash line on (on the φ=θ), the shape (with reference to Fig. 6) that sinuous error signal (strain error) is superimposed.

Here, identical with the situation of above-mentioned phase error, if try by the output angle φ before described compensation _βIn described linear function φ=θ is deducted (φ _β-θ), and obtain the detection strain error Δ φ that is equivalent to both angular misalignments _β, then can represent as sinuous periodic function (trigonometric function) shown in Figure 7.Promptly as can be known, the output angle φ before the compensation after ATAN handles _β-in, include sinuous detection strain error Δ φ shown in Figure 7 _β(=φ _β-θ).

Equally, if to described detection strain error Δ φ _βAnalyze, then with respect to the baseband signal (four signal or described SIN signal and COS signal) that with one-period T is 360 °, described detection strain error φ _βWith one-period T as 90 ° (with reference to Fig. 7).Promptly as can be known, described detection strain error Δ φ _β, be the trigonometric function that forms by frequency (1/4 cycle) with respect to 4 times of described baseband signals.Therefore, detect strain error Δ φ _βCan be similar to by with the trigonometric function shown in the following formula 8, and so and approx expression detects strain error Δ φ _βSignal be called strain compensation value S _β

(formula 8)

Wherein, φ _β=arctan (sin θ/cos θ).

In addition, described strain compensation value S _βThe amplitude amount with respect to described strain error β.

Fig. 8 is expression total error signal Δ φ=(Δ φ _α+ Δ φ _β) the oscillogram of an example.

In output angle (output angle before the compensation) φ of the described operational part 14 of the angle detecting sensor of reality, above-mentioned detected phase error delta φ _αWith detection strain error Δ φ _βTwo errors are mixed with the synthetic state of signal and are existed.If will comprise described detected phase error delta φ _αWith detection strain error Δ φ _βThe comprehensive error signal of the two is as integrated error signal Δ φ=(Δ φ _α+ Δ φ _β) and represent, then described integrated error signal Δ φ, be by as shown in Figure 8 one-period T is made as the signal that 180 ° periodic function forms.Among the output angle φ before the compensation after ATAN handles, include this integrated error signal Δ φ.

Therefore, after the ATAN as above-mentioned second step handles, if can from the output angle φ before the described compensation, remove the detected phase error delta φ that constitutes total error signal Δ φ _αAnd detection strain error Δ φ _β, then should be able to obtain high-precision angle output φ _OUT

But, though distinguished aforesaid detected phase error signal Δ φ _α, can be similar to trigonometric function, and detect strain error signal delta φ by the above-mentioned formula 4 that forms with respect to the baseband signal double frequency _βCan be similar to trigonometric function, but as shown in Figure 8, constitute the detected phase error delta φ of total error signal Δ φ by the above-mentioned formula 5 that forms with respect to baseband signal quadruple rate _αWith detection strain error Δ φ _β, exist with synthetic state, be difficult to directly obtain detected phase error delta φ from described total error signal Δ φ _αWith detection strain error Δ φ _βAmplitude (α/2 and β).

Therefore, in the following to directly not obtaining above-mentioned detected phase error delta φ _αWith detection strain error Δ φ _βEach amplitude, and obtain the described phase compensation value S that forms by suitable amplitude _αWith strain compensation value S _βThird step describe.In addition, third step has been used so-called vague generalization mediation and has been resolved (GHA:Generalized Harmonic Analysis).

＜third step 〉

Fig. 9 is another routine oscillogram that expression compensates the total error signal Δ φ that is comprised among the preceding output angle φ.Figure 10 is expression phase error compensation value S _αAnd the oscillogram of candidate signal, Figure 11 is the error signal Δ φ after the expression phase compensation _-α=(Δ φ-S _α) oscillogram, Figure 12 is expression strain compensation value S _βAnd the oscillogram of candidate signal, Figure 13 is the error signal Δ φ behind expression phase compensation and the strain compensation _-alpha-beta=(Δ φ _-α-S _β) oscillogram, Fig. 14 is expression gain compensation value S _γAnd the oscillogram of candidate signal.Figure 15 is the error signal Δ φ behind expression phase place, strain and the gain compensation _{-alpha-beta-γ}Oscillogram.

Below, describe be contained in the situation of handling the output angle φ before the compensation of calculating by the ATAN in the above-mentioned operational part 14 at as shown in Figure 9 total error signal Δ φ.

(1) compensation detected phase error delta φ _αPhase compensation value S _αComputing method.

At first, preparation a plurality of first candidate signal S1 as shown in figure 10.A plurality of first candidate signal S1 of this moment are used to extract phase compensation value S _αThe aggregate of signal, each first candidate signal S1 is the signal by above-mentioned formula 5 defineds, promptly with respect to the baseband signal of described SIN signal and COS signal etc., and the signal that forms by 2 overtones bands, but the amplitude amount (phase error) of full width is because of each candidate signal difference.In addition, the data volume of each first candidate signal S1 is equivalent to the amount of the one-period (360 °) of described total error signal Δ φ.

Described control part 11 extracts an only signal from these a plurality of first candidate signal S1 as phase compensation value.That is to say that control part 11 sequentially reads one the first candidate signal S1 portion 15 that recompenses from a plurality of first candidate signal S1.

Here, establishing the residual error composition that deducts the described first candidate signal S1 that reads from described total error signal Δ φ is e1, and then residual error ingredient e 1 is represented by formula 9.

(formula 9)

e1＝Δφ-S1

So described compensation section 15 is based on calculate residual energy E1 with following formula 10.

(formula 10)

$E1={\&Integral;}_0^{2\mathrm{\&pi;}}{e1}^2\mathrm{d\&theta;}$

Control part 11 will be based on described formula 9 and formula 10 and residual energy E1 and each described first candidate signal S1 of output compare, and will be the phase place candidate signal S of the first candidate signal S1 of minimum as the best with described residual energy E1 therefrom _αAnd extract out.

In addition, from described total error signal Δ φ, deduct so and the phase compensation value S that extracts _αThe back forms the error signal Δ φ after the phase compensation shown in Figure 11 _-α(=Δ φ-S _α).So, by from described total error signal Δ φ (with reference to Fig. 9), deducting described phase compensation value S _α(with reference to Figure 10) can remove the described detected phase error signal Δ φ that is contained among the described total error signal Δ φ _α(with reference to Fig. 5).

(2) compensation detects strain error Δ φ _βStrain compensation value S _βComputing method.

Next, preparation a plurality of second candidate signal S2 as shown in figure 12.A plurality of second candidate signal S2 of this moment are used to extract strain compensation value S _βThe aggregate of signal, each second candidate signal S2 is the signal by above-mentioned formula 5 defineds, the i.e. signal that is formed by 4 overtones bands with respect to baseband signals such as described SIN signal and COS signals.Just, the amplitude amount (phase error β) of full width is different because of each second candidate signal S2.In addition, the data volume of each second candidate signal S2 is equivalent to the amount of the one-period (360 °) of described total error signal Δ φ.

Described control part 11 sequentially reads one the second candidate signal S2 portion 15 that recompenses from a plurality of second candidate signal S2.

Here, establish error signal Δ φ after the described phase compensation _-αThe residual error composition that deducts the described second candidate signal S2 that reads is e2, and then residual error ingredient e 2 is represented by formula 11.

(formula 11)

e2＝Δφ _-α-S2＝Δφ-S1-S2

So described compensation section 15 is calculated residual energy E2 based on following formula 12.

(formula 12)

$\mathrm{<figure-callout\; id="2"\; label="E"\; filenames="C20051013144400261.png,C20051013144400262.png"\; state="\{\{state\}\}">E</figure-callout>}2={\&Integral;}_0^{2\mathrm{\&pi;}}{e2}^2\mathrm{d\&theta;}$

Control part 11 compares described residual energy E2 and each described second candidate signal S2, and will be the phase place candidate signal S of the first candidate signal S2 of minimum as the best with described residual energy E2 therefrom _βAnd extract out.

In addition, the error signal Δ φ after the described phase compensation _-αIn deduct so and the strain compensation value S that extracts _βAfter become as shown in figure 13 phase compensation and the error signal Δ φ behind the strain compensation _-alpha-beta(=Δ φ _-α-S _β=Δ φ-S _α-S _β).So, by the error signal Δ φ after described phase compensation _-αDeduct described strain compensation value S in (with reference to Figure 11) _β(with reference to Figure 12) can remove the error signal Δ φ that is contained in after the described phase compensation _-αIn described detection strain error signal delta φ _β(with reference to Fig. 7) can confirm to reduce to be contained in the error signal Δ φ behind phase compensation and the strain compensation as shown in figure 13 _-alpha-betaIn error.

But, at phase compensation shown in Figure 13 and the error signal Δ φ behind the strain compensation _-alpha-betaIn, include the error signal Δ φ of SIN shape _γIf at length to described error signal Δ φ _γInvestigate, then this error signal Δ φ as can be known _γBe in the gain adjustment of being undertaken by above-mentioned signal adjustment part 13 (amplify and adjust), not adjust and remaining gain error signal Δ φ _γAnd described gain error signal Δ φ _γCan be by representing with following formula 13.

(formula 13)

Δφ _γ＝γsin2θ

Wherein, γ is the amplitude coefficient as gain error.Therefore, in curve shown in Figure 13

About.

(3) compensating gain error signal Δ φ _γGain compensation value S _γComputing method.

Described gain error signal Δ φ _γAlso can by with above-mentioned detected phase error delta φ _αWith detection strain error Δ φ _βThe identical method of situation and remove.

That is to say, prepare a plurality of the 3rd candidate signals shown in Figure 14 and (be used to extract gain compensation value S _γThe set of signal) S3, and utilize each the 3rd candidate signal S3, obtain the error signal Δ φ behind above-mentioned phase compensation and strain compensation _-alpha-betaIn deduct residual error ingredient e 3 (=Δ φ behind described the 3rd candidate signal S3 _-alpha-beta-S3) carry out e3 after the involution ², in the scope of one-period (0～2 π [rad]), carry out the residual energy E3 of integration gained.So, from will being that the 3rd candidate signal S3 of minimum is as the optimum gain offset wherein with described residual energy E3

And extract out.

In addition, the error signal Δ φ behind described phase compensation and the strain compensation _-alpha-betaIn deduct so and the gain compensation value S that extracts out _γError signal Δ φ behind back formation phase place, strain and the gain compensation as shown in figure 15 _{-alpha-beta-γ}(=Δ φ _-alpha-beta-S _γ).Like this, can remove the error that is comprised among the output angle φ before the described compensation substantially, perhaps approach 0.

So, in the present application,, can from described a plurality of first candidate signal S1, will with residual energy E1 the described by way of compensation detected phase error delta of the first candidate signal S1 φ of minimum by using the aforementioned calculation method _αPhase compensation value S _αAnd calculate, and can from described a plurality of second candidate signal S2, will with residual energy E2 the described by way of compensation detection strain error of the second candidate signal S1 Δ φ of minimum _βStrain compensation value S _βAnd calculate, and then can from described a plurality of the 3rd candidate signal S3, will with residual energy E3 the described by way of compensation gain error signal Δ of the 3rd candidate signal S3 φ of minimum _γGain compensation value S _γAnd calculate.

So, in the storage part 16 of angle detecting sensor, write in advance at least more than one the phase compensation value S that calculates by said method _α, strain compensation value S _βPerhaps gain compensation value S _γTherefore, whenever exporting φ from described angle detecting sensor output angle _OUTThe time, all read phase compensation value S from described storer 16 _α, strain compensation value S _βPerhaps gain compensation value S _γAt least more than one, and the angular error φ before described compensation removes detected phase offset Δ φ _α, detect strain compensation value Δ φ _βPerhaps gain compensation value Δ φ _γ

So, even in four signals from magneto-resistance effect element, exporting, contain under the situation of various error signals, also can reduce its influence and improve the angle output φ that exports as angle detecting sensor _OUTAccuracy of detection.

And by carrying out above-mentioned third step repeatedly, can be with described detected phase error delta φ _α, detect strain error Δ φ _βAnd gain error signal Δ φ _γAll remove, can further improve the φ of the angle output of angle detecting sensor _OUTAccuracy of detection.

In addition, though in the computing method of the offset of above-mentioned embodiment, be illustrated in the following order, promptly be to calculate phase compensation value S at first _α, then be to calculate strain compensation value S _β, be calculated gains offset S at last _γ, but the invention is not restricted to this, with which kind of order computation offset all can.

In addition, in the above-described embodiment, though to remove detected phase error delta φ at first _α, then remove and detect strain error Δ φ _β, remove gain error signal Δ φ at last _γBe order, but the invention is not restricted to this, which kind of removes error in proper order with all can.In addition, can understand,, then can remove above-mentioned any situation error in addition if identical periodic certain error is arranged.

Claims (10)

1. the compensation value calculation method of an angle detecting sensor, this angle detecting sensor has: sensor assembly, it produces rotating magnetic field according to the rotation of granting determinand, and the described rotating magnetic field of perception and export a plurality of output signals corresponding to the input angle of described rotation; Signal transformation portion, it is transformed into two kinds signal with given phase differential from described a plurality of output signals; The signal adjustment part, it adjusts the skew and the gain of described two kinds of signals; Operational part, it is according to described adjusted two kinds of signals, with the anglec of rotation of described determinand before by way of compensation output angle and calculate; Memory section, it stores the offset that the total error signal that is comprised in the output angle before the described compensation is compensated in advance; Compensation section, it is by removing described offset in the total error signal the output angle before being contained in described compensation, and calculates the output angle of described determinand,

It is characterized in that described compensation value calculation method has:

First step generates two kinds signal with given phase differential from described a plurality of output signals;

Second step is calculated the anglec of rotation of the described determinand output angle before by way of compensation according to described two kinds signal;

Third step, in the total error signal the output angle before being contained in described compensation, when removing the different a plurality of candidate signal of cycle same-amplitude one by one, will be that the candidate signal of minimum is extracted out as described offset with the residual energy that is comprised in the signal after removing.

2. the compensation value calculation method of angle detecting sensor according to claim 1 is characterized in that, described output angle is being made as θ, and the total error signal that is comprised in the output angle before the described compensation is made as Δ φ, and described offset is made as S _n, wherein n is α, β or γ, and will remove described offset S from described total error signal Δ φ _nAfter signal in the one-tenth that comprised be divided into e=Δ φ-S _nThe time, described residual energy E is calculated by following formula 1,

(formula 1)

$E={\&Integral;}_0^{2\mathrm{\&pi;}}{e}^2\mathrm{d\&theta;}.$

3. the compensation value calculation method of angle detecting sensor according to claim 2, it is characterized in that, with the phase differential that results between described two kinds signal, the offset when becoming 90 ° ± α add phase error ± α on 90 ° of given phase differential after is made as S _αThe time, described offset S _αIt is approximate by following formula 2 expressions,

(formula 2)

$S_{\mathrm{\&alpha;}}=\frac{\mathrm{\&alpha;}}{2}\{1+\cos 2{\mathrm{\&phi;}}_{\mathrm{\&alpha;}}\}$

Wherein, ${\mathrm{\&phi;}}_{\mathrm{\&alpha;}}=\arctan \left\{\frac{\sin (\mathrm{\&theta;}+\mathrm{\&alpha;})}{\cos \mathrm{\&theta;}}\right\}.$

4. the compensation value calculation method of angle detecting sensor according to claim 2 is characterized in that, will be with respect to based on resistance value strain that described sensor assembly had and the offset of the strain error β that produces is made as S _βThe time, described offset S _βIt is approximate by following formula 3 expressions,

(formula 3)

S _β＝-β·sin4φ _β

Wherein, ${\mathrm{\&phi;}}_{\mathrm{\&beta;}}=\arctan \left\{\frac{\sin \mathrm{\&theta;}}{\cos \mathrm{\&theta;}}\right\}.$

5. the compensation value calculation method of angle detecting sensor according to claim 2 is characterized in that, will be made as S with respect to the offset of the gain error with amplitude γ _γThe time, described offset S _γApproximate by S _γ=γ sin2 θ represents.

6. angle detecting sensor, it is characterized in that, the described offset that utilizes the compensation value calculation method described in the described claim 1 to calculate is deposited into described memory section, and compensates with the offset that reads from the described memory section output angle to described determinand.

7. angle detecting sensor according to claim 6 is characterized in that, in described memory section, deposits described offset S in _α, S _βAnd S _γIn at least more than one offset.

8. angle detecting sensor according to claim 6 is characterized in that, described sensor assembly is two groups the bridge circuit that is formed by magneto-resistance effect element.

9. angle detecting sensor according to claim 6, it is characterized in that, described two kinds signal is SIN signal and COS signal, after described operational part goes out tangent value from described SIN signal and described COS calculated signals, by obtaining arc-tangent value, the anglec of rotation that makes described determinand before by way of compensation output angle and calculate.

10. angle detecting sensor according to claim 6 is characterized in that, described signal adjustment part, described operational part and described compensation section are formed by a calculation process module.

Images