JP2002303555A - Magnetostrictive torque sensor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce hysteresis more than in a conventional sensor. SOLUTION: A magnetostrictive torque sensor 10 includes a sensor shaft 12 of which magnetic characteristics change depending on applied torque, an exciting means (an excitation circuit 14 and excitation coils 161, 162) for applying an AC magnetic field to the sensor shaft 12, and a detecting means (detecting coils 181, 182, and a signal processing part 20) for detecting the torque based on the change in the magnetic characteristics of the sensor shaft 12. An excitation frequency lies between a first frequency and a second frequency. the first frequency is a value where the hysteresis is zero at a first torque value not less than a rated torque value. The second frequency is a value where the hysteresis is zero at a second torque value not more than the first torque value and having a maximum hysteresis when the AC magnetic field of the first frequency is applied.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetostrictive strain sensor (stress sensor or load sensor) utilizing an inverse magnetostrictive effect, and more particularly to a magnetostrictive torque sensor. ) Can be reduced.

[0002]

2. Description of the Related Art A conventional magnetostrictive torque sensor is disclosed, for example, in Japanese Patent Laid-Open Publication No.
One disclosed in Japanese Patent Publication No. 36 is known. This type of sensor includes a sensor shaft having a magnetostrictive layer whose magnetic properties change according to stress (strain), and excitation means for applying an AC magnetic field thereto. It has a magnetic head or a solenoid coil to detect a target change.

[0003] Some magnetic materials exhibit a magnetostrictive phenomenon in which distortion occurs when magnetism is applied. In such a material, when a strain is applied, an inverse magnetostriction phenomenon in which a magnetic property (magnetic permeability) changes is caused. The magnetostrictive strain sensor has a feature that very sensitive, resistive strain gauges distortion detecting means such as 10 ^{- 3}
While an order distortion is detected, a distortion of the order of 10 ⁻⁶ can also be detected.

[0004] By utilizing the property of being able to detect very small strains, it is possible to construct a highly sensitive and highly rigid torque sensor. This is a magnetostrictive torque sensor.
The magnetostrictive torque sensor detects a stress (strain) in a tensile direction or a compressive direction on the sensor shaft surface. The detection method includes:
A method in which a groove inclined at 45 ° to the central axis is formed so that the tensile stress and the compressive stress on the sensor shaft surface become maximum against a torsional load, and a change in this direction is detected by shape magnetic anisotropy. There is a method of arranging the magnetic head so as to detect a change in the 45 ° direction of the shaft surface. In either detection method, the sensor axis can be made thicker because the sensitivity is significantly increased as compared with the method of obtaining torque from the angular displacement between the input side and the output side of the torsion bar, and the method using a strain gauge. A torque sensor can be realized.

As an applicable field of the magnetostrictive torque sensor, there is, for example, an electric power steering for an automobile. This is a power steering having a configuration in which an electric motor is controlled in accordance with a torque input to a steering wheel to generate an assist force. Magnetostrictive torque sensors, which are also directly operated by humans and have the characteristics of high sensitivity and high rigidity in view of sensory performance such as rigidity and natural feeling, are suitable for this application. However, on the other hand, strict performance is required such that the sensor characteristics must not change even if an excessively high torque is input with respect to the rated torque.

It is known that the magnetostrictive torque sensor has extremely high sensitivity, but when a large load (torque) is applied far from the rated load (rated torque), hysteresis occurs. ing. This indicates that the midpoint output value of the sensor changes. Therefore, in a power steering having a configuration in which torque assist is performed in accordance with a sensor output, this midpoint output fluctuation, that is, hysteresis, is a very serious problem.

[0007] In the first place, this problem is caused by a large difference between the rated torque and the guaranteed maximum torque value. This will be described with an example. The torque sensor has both a function as a transmission component for transmitting torque and a function as a sensor for measuring transmitted torque. From the aspect of the transmission component, if the shaft diameter that is not broken at the guaranteed maximum torque value is selected, the function is satisfied. On the other hand, from the viewpoint of the sensor, it is sufficient that the rated torque can be measured with the required performance with this shaft diameter. However, in reality, the rated torque and the guaranteed maximum torque value are separated by about 10 times, so that when an excessive torque is applied, a large hysteresis occurs.

For example, the guaranteed maximum torque is 1 of the rated torque.
Consider the case of 0 times. In this case, it is assumed that the magnetostrictive torque sensor has 1% hysteresis with respect to the output at the guaranteed maximum torque. Then, the value of the hysteresis becomes 10% of the rated torque.

In the present specification, the sign of the torque is positive for clockwise (CW) and negative for counterclockwise (CCW). The magnitude of the hysteresis is the amount of output change due to hysteresis, and the amount of output change at rated torque (one side).
Use a numerical value that has been rendered dimensionless by the value obtained by dividing by (% FS). In this specification, the term "hysteresis" is positive when a clockwise torque is input and unloaded, and the sensor output at that time is larger than the original output value. Also, in this specification, the magnitude relationship between torque and hysteresis will be discussed in terms of absolute values unless otherwise specified.

[0010]

A method for solving such a problem is disclosed in Japanese Patent Application Laid-Open No. 2000-9558. In this method, two layers having a positive hysteresis and a negative hysteresis with respect to the torque are arranged in the magnetostrictive portion so that the hysteresis with respect to the excessive torque is made to be apparently zero. Specifically, using a manufacturing method such as shot peening, two layers of a positive hysteresis area and a negative hysteresis area are arranged in an extremely thin area on the sensor shaft surface, and the excitation frequency at which the hysteresis is zero at the guaranteed maximum torque value This is a technique for apparently reducing the hysteresis to zero by using.

However, the inventor of the present invention has found that this method is effective for excessive torque in which optimal conditions are set,
It has been found that hysteresis still occurs at other torques (that is, torques lower than that).
That is, an intermediate torque range that is equal to or higher than the rated torque value and equal to or lower than the guaranteed maximum torque value (hereinafter, referred to as “excess torque range”).
Then, hysteresis still exists.

[0012]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a magnetostrictive torque sensor which can suppress the hysteresis more than before.

[0013]

A magnetostrictive torque sensor according to the present invention comprises: a sensor shaft whose magnetic properties change according to an applied torque; an exciting unit for applying an AC magnetic field to the sensor shaft; Detecting means for detecting a torque based on a change in the magnetic property of the sensor shaft. The excitation frequency of the AC magnetic field is between the first frequency and the second frequency. Here, assuming that the torque value and the hysteresis are absolute values, the first frequency is a value at which the hysteresis becomes zero at the first torque value equal to or higher than the rated torque value. The second frequency is a value at which the hysteresis becomes zero at a second torque value that is equal to or less than the first torque value and has the maximum hysteresis when an AC magnetic field of the first frequency is applied.

At the first frequency, the hysteresis at the first torque value becomes zero, but the hysteresis at the second torque value becomes maximum. On the other hand, at the second frequency, the hysteresis at the second torque value is zero, but the hysteresis at the first torque value is not zero (it becomes considerably large). Therefore, in the present invention, by setting the excitation frequency between the first frequency and the second frequency, the hysteresis of the first torque value is smaller than that in the case of the second frequency, and the second torque value Is made smaller than in the case of the first frequency. Therefore, the hysteresis is reduced as a whole torque value.

[0015] The first torque value may be a maximum guaranteed torque value. In general, when the maximum guaranteed torque value is applied, the hysteresis is the largest, so that the effect is most remarkable.

The exciting means has a function of changing the exciting frequency, and when the detecting means detects an excessive torque value equal to or more than a predetermined value, the excessive torque value is set as a new first torque value. 2). Since the maximum guaranteed torque value is not applied so often, the present invention is applied to a possible excessive torque. That is, when an excessive torque value is detected, the first and second frequencies are determined using the excessive torque value as a first torque value, and an excitation frequency is set between them.

For example, when an excessive torque value is detected, a first frequency at which the hysteresis becomes zero at the excessive torque value is determined. Subsequently, when the torque is below the excessive torque value,
A second torque value having a maximum hysteresis when an AC magnetic field of a first frequency is applied is determined. Subsequently, at the second torque value, a second frequency at which the hysteresis becomes zero is determined. Subsequently, for example, the excitation frequency is set such that the hysteresis at the excessive torque value and the hysteresis at the second torque value are opposite in polarity and equal in absolute value. The first and second frequencies, the second torque value, and the final excitation frequency are determined, for example, by previously measuring the relationship between torque and hysteresis for each excitation frequency and creating a three-dimensional map. This is realized by searching this three-dimensional map.

The excitation frequency may be fixed at all times. In this case, the exciting means can be simplified.

The exciting frequency may be a value at which the hysteresis at the first torque value and the hysteresis at the second torque value are opposite in polarity and equal in absolute value (claim 5). In this case, the hysteresis is minimized as a whole torque value.

[0020]

FIG. 1 is a block diagram showing an embodiment of a magnetostrictive torque sensor according to the present invention. Hereinafter, description will be made based on this drawing.

The magnetostrictive torque sensor 10 of the present embodiment is
A sensor shaft 12 whose magnetic properties change in accordance with the applied torque, and exciting means (exciting circuit 14 and exciting coils 161, 162) for applying an alternating magnetic field to the sensor shaft 12
And detection means (detection coils 181 and 182 and a signal processing unit 20) for detecting the torque based on a change in the magnetic property of the sensor shaft 12.

The sensor shaft 12 is made of a ferromagnetic material such as iron, nickel, or an alloy thereof. When a torque is applied to the sensor shaft 12, ± 45 with respect to the center line of the sensor shaft 12.
Tensile stress and compressive stress are applied in the direction of degree. There is a difference in the magnetic permeability between the direction in which the tensile stress is applied and the direction in which the compressive stress is applied due to the magnetostriction phenomenon. That is, torque can be detected by detecting the difference in the magnetic permeability. In other words,
Magnetic anisotropic portions 121 and 122 are formed on the surface of the sensor shaft 12 by forming a groove having an angle with the central axis, and the magnetic anisotropic portions 121 and 122 are subjected to shot peening. The magnetic permeability of the isotropic portions 121 and 122 changes according to the torque.

Therefore, the excitation coils 161, 162 and the detection coils 181, 182 are arranged on the surface of the sensor shaft 12,
The exciting coils 161 and 162 are excited by the AC voltage of the exciting circuit 14. Then, AC voltage values corresponding to the magnetic permeability in the tension direction and the magnetic permeability in the compression direction are respectively detected by the detection coils 16.
1,162 separately. Therefore, if these AC voltages are rectified and smoothed to DC, the difference between them corresponds to the torque value applied to the sensor shaft 12. In addition,
Although not shown, the sensor shaft 12 is connected between, for example, an input shaft attached to a steering wheel and an output shaft attached to a steering mechanism such as a pinion and a rack.

Although not shown, the signal processing unit 20 includes a rectifier circuit for rectifying the AC voltage output from the detection coils 181 and 182, a comparison circuit for obtaining a difference between the output voltages of the rectifier circuits, and a DC output for the comparison circuit. , A temperature sensor for temperature correction, a microcomputer, and the like.

Further, the excitation circuit 14 may have a function of changing the excitation frequency according to the excessive torque value when an excessive torque value exceeding the rated torque is output from the signal processing unit 20. Such a function can be realized by, for example, a microcomputer, a memory, a VCO, and the like.

Hereinafter, the magnetostrictive torque sensor 1 according to the present embodiment will be described.
The operation of 0 will be described.

As a result of investigating the phenomenon described in "Problems to be Solved by the Invention" through detailed experiments, the inventor has found that an excitation frequency at which the hysteresis becomes zero at a certain torque value (hereinafter referred to as "optimal excitation frequency"). Is present at one or two points, and it has been found that the optimum excitation frequency changes depending on the torque value.

In the magnetostrictive torque sensor having the excitation frequency fTm at which the hysteresis becomes zero at the guaranteed maximum torque Tm, the relationship between the torque in the excessive torque region and the hysteresis is plotted as shown in FIG. Guaranteed maximum torque Tm
The hysteresis is increased by the following excessive torque, and the hysteresis is maximized at the torque THmax.

In order to examine this in detail, FIG. 3 plots the relationship between the excitation frequency and the amount of hysteresis.
From this, it can be seen that the optimum excitation frequency fTHmax at the torque THmax has shifted to a higher frequency side than the case of the guaranteed maximum torque Tm. That is, at the torque THmax, the amount of hysteresis indicated by HTHmax in FIG. 3 occurs at the excitation frequency fTm.

Let us look at this phenomenon in more detail. A positive hysteresis region and a negative hysteresis region immediately inside the sensor shaft surface are formed on the sensor shaft surface. here,
Under the excitation frequency fTm at which the apparent hysteresis becomes zero at the guaranteed maximum torque Tm, the respective hysteresis cancels each other due to the same amount and opposite sign. However, a smaller torque THma
In x, since this relationship is broken, a positive hysteresis remains.

This is due to the correlation between the stress and the hysteresis and the deviation of the stress distribution in the round bar. The stress τ in the round bar has the following relationship with the diameter d. ^{τ = 16T / (πd 3)} T: torsional torque i.e., stress round bar surface, as well as proportional to the torque value is inversely proportional to the cube of diameter. Therefore, when the input torque value is small, the stress is also reduced in proportion thereto. At this time, if the stress-hysteresis relationship does not have a linear relationship, the hysteresis of each of the positive and negative regions has a deviation in magnitude when the torque value changes, and cannot be canceled.

Further, since the stress distribution is proportional to the cube of the diameter, the change in the stress value with respect to the change in the diameter is large. Therefore, it is necessary to generate a large amount of hysteresis in the inner region with smaller stress than in the outer region.
Generally, the stress-hysteresis relationship is constituted by a non-linear region and a linear approximation region. However, due to the large change in stress value due to the diameter, the difference in stress value between the outermost surface and immediately below it is large. It happens. As a result, each hysteresis changes in magnitude and cannot be canceled each other, so that the hysteresis remains.

For the above reasons, hysteresis still occurs in the excessive torque range. However, at a sufficiently small level near the rated torque, the surface stress is small and the hysteresis itself is small, so that the magnitude of the hysteresis does not matter.

To solve the above problem, the excitation frequency may be varied according to the excessive torque value input in the excessive torque region. Specifically, a survey as shown in Fig. 3 was conducted over the entire excessive torque range, the excitation frequency at which the hysteresis became zero was measured, and a map of the excessive torque value and the excitation frequency was created, and the excitation frequency was changed. It should be done.
It should be noted here that the excitation frequency needs to be changed only when excessive torque is input.

That is, the excitation frequency may be fixed as long as it is used only in the rated torque range (not more than a set torque value). Then, the excitation frequency is changed only when an excessive torque is input, and thereafter, the excitation frequency is not returned to the original excitation frequency but is used until the next excessive torque is input again. This point is crucially different from the case where temperature correction or the like is performed by changing the excitation frequency or the excitation current, which will be described later.

Focusing on this point, a slightly simpler method can be considered. That is, the torque THma in FIG.
This is a method in which the excitation frequency is varied only when an excessive torque between x and the guaranteed maximum torque Tm is applied.
In this case, in order to reduce the hysteresis on the low torque side, the optimum excitation frequency fTHm at the torque THmax is set.
ax, and only when a torque equal to or greater than the torque THmax is input, the excitation frequency may be changed according to the input stress. In general, the frequency of inputting excessive torque is not so high, and thus, sufficient performance is exhibited in practical use.

It is ideal that the excitation frequency in the region from the torque THmax to the guaranteed maximum torque Tm is varied based on a map created based on the optimal excitation frequency obtained experimentally. In order to simplify this, it is also possible to cope with a method in which the excitation frequency map between the excessive torque THmax and the guaranteed maximum torque Tm is linearized with respect to the torque value. With this method, the map itself can be omitted, which is advantageous in terms of cost.

In addition, when the variation range of the optimum excitation frequency due to the excessive torque value is small and the hysteresis in the excessive torque region is relatively small, as shown in FIG.
A method in which the excitation frequency is fixedly used at an intermediate excitation frequency from Hmax is also conceivable. In this case, if it is desired to reduce the total width of the excessive torque-hysteresis relationship, as shown in FIG. 3, an intermediate excitation frequency fa such that the magnitude of the hysteresis becomes equal is obtained.
You just have to select In this case, the relationship between the excessive torque and the hysteresis initially shown in FIG. 2 becomes as shown in FIG. Then, since the torque output in the excessive torque range changes across zero, the total width becomes small when considered in the excessive torque range of right and left torsion. Further, when the excitation frequency is fixed, many parts such as a part for changing the excitation frequency, a part for detecting and measuring excessive torque, and a part for correcting the sensitivity according to the changed excitation frequency can be omitted, so that cost and cost can be reduced. There are many advantages in terms of reliability and the like.

FIG. 5 is a three-dimensional graph showing the relationship between FIGS. Hereinafter, an example of the operation of the magnetostrictive torque sensor 10 will be described with reference to FIGS.

The relationship between the torque and the hysteresis is measured in advance for each excitation frequency, and this is stored in a memory of the excitation circuit 14, for example, as a three-dimensional map shown in FIG. The initial value of the excitation frequency is fTHmax.
Therefore, the excitation circuit 14 determines the excitation frequency fTHmax
Excites the exciting coils 161 and 162.

Here, when the guaranteed maximum torque Tm is detected by the detection coils 181 and 182, the detected value is applied to the signal processing unit 2
0 is output to the excitation circuit 14. Then, the excitation circuit 1
The microcomputer 4 operates as follows. First, the excitation frequency fTm at which the hysteresis becomes zero at the guaranteed maximum torque Tm is searched from the three-dimensional map of FIG. Subsequently, a torque value THmax that is equal to or lower than the guaranteed maximum torque Tm and has the maximum hysteresis HTHmax when an AC magnetic field having the excitation frequency fTm is applied is searched. Subsequently, an excitation frequency fTHmax at which the hysteresis becomes zero is searched for in the torque value THmax. Next, the excitation frequency fa is searched such that the hysteresis H2 at the guaranteed maximum torque Tm and the hysteresis H1 at the torque value THmax are opposite in polarity and equal in absolute value. Finally, the excitation circuit 14 sets the excitation frequency to fT
The excitation coils 161 and 162 are excited at the excitation frequency fa by changing from Hmax to fa. As a result, the hysteresis is minimized as a whole torque value.

Subsequently, although not shown in FIG. 5, it is assumed that an excessive torque T (THmax <T <Tm) is detected. Then, the microcomputer of the excitation circuit 14 operates as follows. First, from the three-dimensional map of FIG. 5, the excitation frequency fT at which the hysteresis becomes zero at the excessive torque T is shown.
(Not shown). Subsequently, a torque value TH (not shown) that is equal to or less than the excessive torque T and has the maximum hysteresis when an AC magnetic field having the excitation frequency fT is applied is searched. Subsequently, an excitation frequency fTH (not shown) at which the hysteresis becomes zero is searched for in the torque value TH. Subsequently, an excitation frequency fb (not shown) is searched such that the hysteresis at the excessive torque T and the hysteresis at the torque value TH are opposite in polarity and equal in absolute value. Lastly, the excitation circuit 14 changes the excitation frequency from fa to fb, and sets the excitation coils 161 and 161 at the excitation frequency fb.
162 is excited. As a result, the hysteresis is minimized as a whole torque value.

Japanese Patent Application Laid-Open No. 10-153502 discloses a technique (hereinafter, referred to as "conventional example 1") in which two excitation frequencies are switched and used in connection with a technique for changing the excitation frequency. . In Conventional Example 1, measurement is performed alternately at an excitation frequency for measuring torque and an excitation frequency that is insensitive to torque in order to detect temperature, and the torque value measured at the excitation frequency for torque detection is used for temperature detection. The correction is performed based on the output obtained at the excitation frequency. The differences between Conventional Example 1 and the present invention are as follows. In the present invention, the excitation frequency is changed only when excessive torque is detected, and thereafter, the excitation frequency is fixed until excessive torque is input again. On the other hand, in the conventional example 1, since the main purpose is to correct the temperature, the excitation frequency is alternately changed at regular time intervals.

Japanese Patent Application Laid-Open No. 9-159553 discloses a technique for improving the temperature drift characteristic by adjusting the excitation frequency (hereinafter referred to as "conventional example 2"). In the second conventional example, a module in which the excitation frequency is variable and a map in which the excitation frequency is determined with respect to the temperature are provided, and the temperature drift is corrected by adjusting the excitation frequency according to the temperature. The differences between Conventional Example 2 and the present invention are as follows. In the second conventional example, the main purpose is placed on the temperature correction, and therefore, the excitation frequency always changes following the temperature change (when the temperature returns to the original value, the excitation frequency returns to the original value). On the other hand, in the present invention, the hysteresis improvement at the time of excessive torque input is made daily, and in the case of the rated torque range, the excitation frequency is fixedly used at a specific value, and only when excessive torque is input. The excitation frequency is varied, and thereafter, the excitation frequency is fixedly used until the next excessive torque value is input.

Next, the present invention will be described more specifically with reference to examples.

[0046]

Example 1 JLS SNCM815 alloy steel round bar (N
Material) from the specified bar material, rolled it,
Grooves (magnetically anisotropic portions) inclined by 45 ° with respect to the central axis were provided on the surface. This part was subjected to induction hardening and then shot peening. Shot peening was performed twice under different conditions. The first shot peening had an arc height of 0.20 mm (A piece), and the second shot peening had an arc height of 0.16 mm (A piece). The shot media used was steel and had a grain size of 0.25.
mm. A solenoid coil and an electric circuit were attached to this sensor shaft, and the torque-output characteristics were evaluated. FIG. 1 shows a schematic configuration of the torque sensor.

The measurement of the hysteresis was performed by measuring the sensor output before and after the application of the torque. When the sensor output was measured, the sensor was removed from the torsion tester and completely mechanically free in order to ensure zero torque.

First, since the maximum guaranteed torque value of this torque sensor is ± 150 N · m, the hysteresis was measured by changing the excitation frequency at this torque. As a result, a relationship between the excitation frequency and the hysteresis as shown in FIG. 7 was obtained, and it was found that the hysteresis became zero at the excitation frequency of 37 kHz.

Next, the relationship between the input torque value and the hysteresis was investigated in a torque range exceeding the rated torque at this excitation frequency. Then, as shown in FIG.
At 50 N · m, the hysteresis becomes zero, but at other torque values, the hysteresis exists and the torque 9
It was found that the hysteresis was maximized at 0 N · m.
Also, in this case, the sign of the right-hand twist (CW: torque value plus) and the left-hand twist (CCW: minus torque value) are reversed, so that the overall width has twice the width of the hysteresis value in CW. You can see that.

Next, the torque 90 at which the hysteresis is maximized
At N · m, the relationship between the excitation frequency and the hysteresis was investigated in the same manner as described above. Then, the excitation frequency is 39.3k
It was found that the hysteresis became zero at Hz.

FIG. 8 shows the result of selecting 38.2 kHz as the excitation frequency based on these results and again examining the relationship between excessive torque and hysteresis. In this case,
Hysteresis at 150 N · m and 90 N · m is
It can be seen that the magnitudes are the same and the signs are inverted. As a result, the peak value of the change is suppressed, and the hysteresis over the entire excessive torque region is averaged.

By the way, the rated torque of this sensor (10
The hysteresis in (N · m) is 0.1 before the excessive torque load.
It was 5% FS and 0.8% FS after loading, which was at a level that was practically no problem.

In this embodiment, the excitation frequency is 90 N
The frequency at which the amount of hysteresis between m and 150 N · m is equal is used, but in reality, 37 kHz and 39.3 k
It is possible to select any frequency between Hz.
As a result, the hysteresis characteristic (curve shape) in the excessive torque range is controlled. For example, if the hysteresis after a load of 150 N · m has a margin, the excitation frequency is set to 39 kHz so as to specialize in the characteristic up to 90 N · m. For example, optimization can be performed according to the day.

[0054]

Embodiment 2 A sensor shaft was manufactured in the same manner as in Embodiment 1, and a similar coil was assembled. To this, an electronic circuit capable of detecting a torque value up to an excessive torque range and having a variable excitation frequency was attached. This circuit has a torque of 90 Nm
The excitation frequency is fixed at 39.3 kHz until 90 to 1
The electronic circuit is configured such that when a load of 50 N · m is applied, the excitation frequency changes linearly in the range of 39.3 to 37 kHz according to the torque value. The hysteresis of the torque sensor in the rated torque range is 0.5% FS.

When an excessive torque is applied to this sensor,
FIG. 9 shows the relationship between excessive torque and hysteresis. It can be seen that the hysteresis is improved over the entire excessive torque range, and the amount of hysteresis combined with CW and CCW is significantly improved as compared with the case of FIG. Incidentally, the hysteresis in the rated torque range after the experiment was 0.7% FS, which was a level having no problem in practical performance.

[0056]

According to the magnetostrictive torque sensor of the present invention, the first frequency at which the hysteresis becomes zero at the first torque value equal to or higher than the rated torque value, and the first frequency at which the hysteresis is equal to or lower than the first torque value, The hysteresis of the first torque value is set by setting the excitation frequency between the second frequency at which the hysteresis becomes zero at the second torque value having the maximum hysteresis when an AC magnetic field of the frequency is applied. Can be made smaller than in the case of the second frequency, and the hysteresis of the second torque value can be made smaller than that in the case of the first frequency, so that the hysteresis can be suppressed as a whole torque value.

According to the magnetostrictive torque sensor according to the second aspect, since the first torque value is the maximum guaranteed torque value, it is possible to prevent the hysteresis from being maximized when the maximum guaranteed torque value is applied. .

According to the magnetostrictive torque sensor of the third aspect, when an excessive torque value is detected, the first and second frequencies are determined using the excessive torque value as a first torque value.
By setting the excitation frequency between these, even if excessive torque is frequently applied, it is possible to finely suppress the hysteresis.

According to the magnetostrictive torque sensor of the fourth aspect, since the excitation frequency is always fixed, the excitation means can be simplified.

According to the magnetostrictive torque sensor of the fifth aspect, the excitation frequency is set so that the hysteresis at the first torque value and the hysteresis at the second torque value are opposite in polarity and equal in absolute value. By doing so, it is possible to minimize the hysteresis as a whole of the torque value.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing an embodiment of a magnetostrictive torque sensor according to the present invention.

FIG. 2 is a graph showing a relationship between torque and hysteresis at an excitation frequency at which hysteresis becomes zero when a maximum guaranteed torque value is applied.

FIG. 3 is a graph showing a relationship between an excitation frequency and a hysteresis for each torque value.

FIG. 4 is a graph showing a relationship between torque and hysteresis at an excitation frequency at which hysteresis is minimized as a whole torque value.

FIG. 5 is a three-dimensional graph showing the relationship between FIGS. 2 to 4;

FIG. 6 is a graph showing a relationship between torque and hysteresis in the first embodiment.

FIG. 7 is a graph showing a relationship between an excitation frequency and a hysteresis for each torque value in the first embodiment.

FIG. 8 is a graph showing a relationship between torque and hysteresis in the first embodiment.

FIG. 9 is a graph showing a relationship between torque and hysteresis in the second embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Magnetostrictive torque sensor 12 Sensor axis 14 Excitation circuit (excitation means) 161 and 162 Excitation coil (excitation means) 181 and 182 Detection coil (detection means) 20 Signal processing unit (detection means)

Claims (5)

[Claims] 1. A sensor shaft whose magnetic property changes according to an applied torque, an exciting unit for applying an AC magnetic field to the sensor shaft, and the torque based on a change in magnetic property of the sensor shaft. Wherein the exciting frequency of the AC magnetic field is between a first frequency and a second frequency, and a torque value and a hysteresis are absolute values. The frequency is a value at which the hysteresis becomes zero at a first torque value equal to or higher than a rated torque value, and the second frequency is equal to or lower than the first torque value and an AC magnetic field of the first frequency is applied. A second torque value having a maximum hysteresis when the hysteresis is zero. 2. The magnetostrictive torque sensor according to claim 1, wherein the first torque value is a maximum guaranteed torque value. 3. The exciting means has a function of changing the exciting frequency. When the detecting means detects an excessive torque value equal to or more than a predetermined value, the exciting means converts the excessive torque value into a new first torque. The magnetostrictive torque sensor according to claim 1, wherein the value is a value. 4. The magnetostrictive torque sensor according to claim 2, wherein said excitation frequency is always fixed. 5. The excitation frequency is a value in which the hysteresis at the first torque value and the hysteresis at the second torque value are opposite in sign and have the same absolute value. Or the magnetostrictive torque sensor according to 4.