Resolver/digital converter and control apparatus using the same

0001The present invention relates to a resolver/digital converter, and particularly to a resolver/digital converter having a failure detecting function or to the failure detecting function of the resolver/digital converter.

0002The servo control system needs a rotation-angle sensor in order to detect the rotation angle and make the feedback control. The brushless motor control also requires the rotation-angle sensor in order to cause current to flow in the coil of the motor in accordance with the rotation angle of the motor as well as the servo control system does. The resolver has so far been widely used as a rotation-angle sensor since it is rigid and resistant to the environment because of its simple structure.

0003In addition, the servo control system that is used for motor-driven power steering, or x-by-wire control system, particularly steer-by-wire or fly-by-wire control system is required to have a failure detecting function because it needs safety and reliability.

0004Moreover, a resolver/digital converter has been developed for converting the signal from the resolver to a rotation angle and supplying it as digital data to a microcomputer or the like (see the catalog of Smartcoder AU6802, Tamagawa Seiki Co., T12-159N1 (Apr. 20, 2002), http://www.tamagawaseiki.co.jp/ct1/1591n3j.pdf (nonpatent literature 1)). The resolver/digital converter according to this document forms a feedback loop that the estimate φ of a rotation angle is modified by the residual error ε resulting from the computation of sin(φ), cos(φ) generated from the estimate φ of rotation angle and the input signal, thereby making the estimate φ of rotation angle be converged to the actual rotation angle θ. This system also has the function to notify the microcomputer of the occurrence of a failure when the residual error ε exceeds a certain value.

0005In addition, according to <patcit id="pcit0001" dnum="JP9280890A"><text>JP-A-9-280890</text></patcit> (patent document 1), the failure of the resolver is detected by using the property of trigonometric function that sin²(θ) + cos²(θ) = 1 as the output from the resolver.

0006Furthermore, according to <patcit id="pcit0002" dnum="JP9072758A"><text>JP-A-9-72758</text></patcit> (patent document 2), an abnormality-detecting signal processor for finding the angle θ is provided independently of an angle-detecting signal processor, and the failure is detected from the difference between the angles θ produced from both processors.

0007<patcit id="pcit0003" dnum="US5347277A"><text>US 5347277</text></patcit> shows a dual phase resolver to digital converter having two separate channels for generating rotor position data. The data from the two channels are selectively combined in the output of the converter in order to provide greater accuracy.

0008According to the prior art described in the nonpatent document 1, if the signal from or to the resolver is disconnected, the signal from the resolver becomes abnormal, and thus it does not satisfy the relation between sin(θ) and cos(θ) that is the condition of the feedback loop convergence. As a result, the residual error ε becomes large, thus this fact being sent to the microcomputer as a failure. Also, if the feedback loop for finding the estimate φ of rotation angle does not work, the residual error ε does not converge, but diverges, thus this fact being sent to the microcomputer as a failure. In this case, if the function for judging that the residual error ε exceeds a certain value and the function to notify the microcomputer of the failure break down (fixed failure of the output, open or short circuit failure), the event of this failure occurrence cannot be transmitted to the microcomputer. Thus, this fact must be further considered.

0009According to the prior art described in the patent document 1, if the signal from or to the resolver is disconnected, the signal from the resolver becomes abnormal, and thus it does not satisfy the relation between sin (θ) and cos (θ). Thus, since the value of sin² (θ) + cos² (θ) is deviated from 1, a failure is detected, and hence the microcomputer can be informed of the failure. However, the failure that can be detected in this system is only the failure of the abnormal signal waveform due to the disconnection of the signal from or to the resolver. That is, if a failure occurs in the resolver/digital converter function itself that performs the process for finding the angle θ, this failure cannot be detected.

0010Moreover, according to the method in patent document 2, since both angles θ redundantly found are compared, it is possible that, even if a failure (3) occurs in the resolver/digital converter function, and if another failure occurs in the function to notify the microcomputer of the failure (3), the failure (3) and this other failure can be both detected.

0011Therefore, the kinds of failure that can be detected by the microcomputer with the techniques mentioned in the above documents combined are as follows. <ol id="ol0001" compact="compact"><li>(1) Failure of abnormality of signal waveform caused when the signal from or to the resolver is disconnected.</li><li>(2) Failure in the function to detect failure (1) and notify the microcomputer of it.</li><li>(3) Failure in the resolver/digital converter function itself that performs the process for finding θ on the basis of signal from the resolver.</li><li>(4) Failure in the function to detect the failure (3) and notify the microcomputer of it.</li></ol>

0012However, the method in patent document 2 still needs consideration about the following points.

(i) Simplicity in computation

0013The computation load is large in finding θ from values of trigonometric functions (sin, cos), and judgment to be made over the boundary between the regions of θ is required to reconsider. For example, when one θ is 1°, the other 359°, the difference between both angles must be judged to be 2°.

(ii) Design diversity

0014The processes for redundantly finding θ are extremely similar since both values of θ are found from sin and cos component signals. Thus, these processes are easy to cause error in the design common to both, and accompany a systematic error. Particularly for the application in which safety is emphasized, the idea of design diversity and n-version design is desired to introduce.

SUMMARY OF THE INVENTION

0015It is the objective of the invention to provide a resolver/digital converter having an improved error detection structure.

0016This object is accomplished by the features of claim 1.

BRIEF DESCRIPTION OF THE DRAWINGS

0017<ul id="ul0001" list-style="none" compact="compact"><li><figref idref="f0001">FIG. 1</figref> is a block diagram showing a basic embodiment of the invention.</li><li><figref idref="f0001">FIG. 2</figref> is a block diagram showing an embodiment of the invention in which a conversion trigger generator is used as a voltage comparator.</li><li><figref idref="f0001">FIG. 3</figref> is a block diagram showing an embodiment of the invention in which the conversion trigger generator is used as a zero cross detector and delay circuit.</li><li><figref idref="f0002">FIG. 4</figref> is a block diagram showing an embodiment of the invention in which the conversion trigger generator is used as a peak detector.</li><li><figref idref="f0002">FIG. 5</figref> is a block diagram showing an embodiment of the invention in which the conversion trigger generator is used as a 90°-phase shifter and zero cross detector.</li><li><figref idref="f0002">FIG. 6</figref> is a waveform diagram showing an example of the operation of the conversion trigger generator.</li><li><figref idref="f0003">FIG. 7</figref> is a block diagram showing an example of an error detecting function.</li><li><figref idref="f0003">FIG. 8</figref> is a block diagram showing another example of the error detecting function.</li><li><figref idref="f0004">FIG. 9</figref> is a block diagram showing still another example of the error detecting function.</li><li><figref idref="f0004">FIG. 10</figref> is a block diagram showing an embodiment of the invention in which common mode input means and normal mode input means are provided.</li><li><figref idref="f0005">FIG. 11</figref> is a block diagram showing an embodiment of the invention in which differential input means and single-end input means are provided.</li><li><figref idref="f0005">FIG. 12</figref> is a diagram showing the waveform of resolver signals that is produced when an inter-phase short circuit occurs.</li><li><figref idref="f0006">FIG. 13</figref> is a block diagram of a detailed example of the embodiment shown in <figref idref="f0004">FIGS. 10</figref> and <figref idref="f0005">11</figref>.</li><li><figref idref="f0007">FIG. 14</figref> is a block diagram showing an embodiment of the invention in which a common mode component detector is provided.</li><li><figref idref="f0007">FIG. 15</figref> is a block diagram showing an embodiment of the invention in which the resolver signals are supplied as differential signals to the resolver/digital converting portion and fed as single-end signals to the A/D converter of the microcomputer.</li><li><figref idref="f0008">FIG. 16</figref> is a block diagram of an embodiment of the invention in which the differential input signals and single-end input signals are switched and fed to the A/D converter of the microcomputer.</li><li><figref idref="f0008">FIG. 17</figref> is a block diagram showing a detailed example of the embodiment shown in <figref idref="f0008">FIG. 16</figref>.</li><li><figref idref="f0009">FIG. 18</figref> is a block diagram showing an embodiment of the invention in which the differential input means and single-end input means are operated in a frequency division manner.</li><li><figref idref="f0009">FIG. 19</figref> is a diagram showing an example of the frequency characteristic for the frequency division in the embodiment shown in <figref idref="f0009">FIG. 18</figref>.</li><li><figref idref="f0009">FIG. 20</figref> is a block diagram showing an example for implementing the frequency division shown in <figref idref="f0009">FIG. 19</figref>.</li><li><figref idref="f0010">FIG. 21</figref> is a block diagram showing another example for implementing the frequency division shown in <figref idref="f0009">FIG. 19</figref>.</li><li><figref idref="f0010">FIG. 22</figref> is a block diagram showing an embodiment of the invention in which a bias correcting function is provided.</li><li><figref idref="f0011">FIG. 23</figref> is a diagram showing another example of the frequency characteristic for the frequency division in the embodiment shown in <figref idref="f0009">FIG. 18</figref>.</li><li><figref idref="f0011">FIG. 24</figref> is a block diagram showing an example for implementing the frequency division shown in <figref idref="f0011">FIG. 23</figref>.</li><li><figref idref="f0012">FIG. 25</figref> is a table of failures that can be detected by this invention.</li><li><figref idref="f0012">FIG. 26</figref> is a block diagram showing an embodiment of a fault tolerant resolver and fault resolver/digital converter according to the invention in which each of the fault tolerant resolver and resolver/digital converter is doubled to be fault tolerant.</li><li><figref idref="f0013">FIG. 27</figref> is a block diagram showing an embodiment of a motor control system according to the invention.</li><li><figref idref="f0014">FIG. 28</figref> is a block diagram showing an embodiment of a power steering apparatus according to the invention.</li><li><figref idref="f0015">FIG. 29</figref> is a diagram of an example showing the threshold for detecting abnormality is changed with the change of car speed.</li><li><figref idref="f0015">FIG. 30</figref> is a block diagram showing an embodiment of the invention in which reliability order judging portions are provided.</li><li><figref idref="f0016">FIG. 31</figref> is a table showing an example of the judgment in the reliability-order judging portion.</li><li><figref idref="f0016">FIG. 32</figref> is a table showing an example of the classification of failures into minor, middle and serious defects.</li><li><figref idref="f0017">FIG. 33</figref> is a block diagram showing an example of the actuator.</li><li><figref idref="f0017">FIG. 34</figref> is a block diagram showing another example of the actuator.</li><li><figref idref="f0018">FIG. 35</figref> is a block diagram showing another example of the actuator.</li><li><figref idref="f0018">FIG. 36</figref> is a block diagram showing still another example of the actuator.</li><li><figref idref="f0019">FIG. 37</figref> is a block diagram showing an embodiment of the invention in which this invention is applied to brakes.</li><li><figref idref="f0020">FIG. 38</figref> is a block diagram showing an embodiment of the invention in which a selector is provided.</li><li><figref idref="f0020">FIG. 39</figref> is a block diagram showing an embodiment of the invention in which selectors and actuators are redundantly provided.</li></ul>

DESCRIPTION OF THE EMBODIMENTS

0018Embodiments of the invention will be described with reference to the accompanying drawings.

0019<figref idref="f0001">FIG. 1</figref> is a block diagram of a basic embodiment of a controller 100 having a failure detecting function for detecting the failure in the resolver circuits.

0020An exciting current f(t) generated from an exciting signal generator 3 is supplied to a resolver 5. The exciting signal f(t) is generally often assigned a trigonometric function such as A · sing (ωt), where A is the amplitude, and ω is the angular velocity that is expressed by ω=2πf of which f is frequency. The resolver 5 produces resolver signals Y_sin and Y_cos. The resolver signals Y_sin and Y_cos, respectively, can be expressed by<maths id="math0001"><math display="block"><msub><mi>Y</mi><mi>sin</mi></msub><mo>=</mo><mi>k</mi><mo>⋅</mo><mi>sin</mi><mfenced><mi>θ</mi></mfenced><mo>⋅</mo><mi>f</mi><mfenced><mi>t</mi></mfenced></math><img file="EP1560085B1_D0001.tif" /></maths><maths id="math0002"><math display="block"><msub><mi>Y</mi><mi>cos</mi></msub><mo>=</mo><mi>k</mi><mo>⋅</mo><mi>cos</mi><mfenced><mi>θ</mi></mfenced><mo>⋅</mo><mi>f</mi><mfenced><mi>t</mi></mfenced></math><img file="EP1560085B1_D0002.tif" /></maths> where θ is the rotation angle, and k is the gain. A resolver/digital converting portion 2 produces an estimated angle φ on the basis of the resolver signals Y_sin and Y_cos The resolver/digital converting-portion 2 also produces an error detection signal and supplies it together with the estimated angle φ to a microcomputer 1. The resolver/digital converting portion 2 can be constructed differently, and the typical one can be considered to use the method proposed in the nonpatent document 1. The resolver/digital converting portion 2 described in the nonpatent document 1 is the unitary body formed of the resolver/digital converting portion 2 and exciting signal generator 3 built in the same chip.

0021Although the above-given resolver/digital converting portion and exciting signal generator are the same as in the prior art, an embodiment of the invention further uses a conversion trigger generator 4 to generate a conversion trigger on the basis of the exciting signal f(t), an A/D converter 11 to convert the resolver signals Y_sin and Y_cos to digital signals by using the conversion trigger, and an error detecting function 12 to detect error due to a failure.

0022The conversion trigger generator 4 can also be constructed differently. <figref idref="f0001">FIG. 2</figref> shows an example of the conversion trigger generator 4 in which a voltage comparator 41 is used to compare the exciting signal f(t) and a reference voltage (V_ref) 42. In this embodiment, since the resolver signals Y_sin and Y_cos are converted to digital signals when the exciting signal f (t) reaches V_ref, the resolver signals Y_sin and Y_cos, when f (t)=V_ref, or at the time of digital conversion, are respectively given by<maths id="math0003"><math display="block"><msub><mi>Y</mi><mi>sin</mi></msub><mo>=</mo><mi>k</mi><mo>⋅</mo><mi>sin</mi><mfenced><mi>θ</mi></mfenced><mo>⋅</mo><msub><mi>V</mi><mi mathvariant="italic">ref</mi></msub></math><img file="EP1560085B1_D0003.tif" /></maths><maths id="math0004"><math display="block"><msub><mi>Y</mi><mi>cos</mi></msub><mo>=</mo><mi>k</mi><mo>⋅</mo><mi>cos</mi><mfenced><mi>θ</mi></mfenced><mo>⋅</mo><msub><mi>V</mi><mi mathvariant="italic">ref</mi></msub></math><img file="EP1560085B1_D0004.tif" /></maths>

0023<figref idref="f0001">FIG. 3</figref> shows another example of the conversion trigger generator 4 in which the voltage comparator 41 is used to compare the exciting signal f(t) and 0 volt and detect zero cross points, and a delay circuit 43 is used to delay it by t_delay. If the exciting signal satisfies f(t)=A · sin (ωt_delay) , the resolver signals Y_sin and Y_cos at the time of digital conversion are respectively expressed by<maths id="math0005"><math display="block"><msub><mi>Y</mi><mi>sin</mi></msub><mo>=</mo><mi>k</mi><mo>⋅</mo><mi>sin</mi><mfenced><mi>θ</mi></mfenced><mo>⋅</mo><mi>A</mi><mo>⋅</mo><mi>sin</mi><mfenced><msub><mi mathvariant="italic">ωt</mi><mi mathvariant="italic">delay</mi></msub></mfenced></math><img file="EP1560085B1_D0005.tif" /></maths><maths id="math0006"><math display="block"><msub><mi>Y</mi><mi>cos</mi></msub><mo>=</mo><mi>k</mi><mo>⋅</mo><mi>cos</mi><mfenced><mi>θ</mi></mfenced><mo>⋅</mo><mi>A</mi><mo>⋅</mo><mi>sin</mi><mfenced><msub><mi mathvariant="italic">ωt</mi><mi mathvariant="italic">delay</mi></msub></mfenced></math><img file="EP1560085B1_D0006.tif" /></maths> Thus, sin(θ), cos(θ) and tan(θ) can be found in the same way as those in the example of <figref idref="f0001">FIG. 2</figref>. If the differences between those values and sin (φ), cos (φ) and tan (φ) based on the estimate φ of angle are within allowable ranges, the result is normal. If they are out of the ranges, the result can be judged to be a failure. In order for the resolver signals Y_sin and Y_cos to be maximized to improve the S/N ratio, it is preferable to satisfy ωt_delay=π/2, or t_delay=π/2ω. In this embodiment, the delay circuit 43 for delaying by t_delay can be constructed by a timer provided within the microprocessor, and thus the number of discrete components to be used can be greatly reduced.

0024<figref idref="f0002">Fig. 4</figref> shows another example of the conversion trigger generator 4 as a peak detector 44. The peak detector 44 can be achieved by a method for differentiating the exciting signal f(t) and finding the time when it becomes zero, a method for using a phase shifter to delay the exciting signal f(t) by 90 degrees so that the zero cross can be obtained, or a method for holding peaks in turn while they are being updated, and finding as a time of peak the time when the current value becomes below the held peak value.

0025According to this embodiment, the resolver signals Y_sin and Y_cos at the time of conversion are respectively given by<maths id="math0007"><math display="block"><msub><mi>Y</mi><mi>sin</mi></msub><mo>=</mo><mi>k</mi><mo>⋅</mo><mi>sin</mi><mfenced><mi>θ</mi></mfenced><mo>⋅</mo><mi>A</mi></math><img file="EP1560085B1_D0007.tif" /></maths><maths id="math0008"><math display="block"><msub><mi>Y</mi><mi>cos</mi></msub><mo>=</mo><mi>k</mi><mo>⋅</mo><mi>cos</mi><mfenced><mi>θ</mi></mfenced><mo>⋅</mo><mi>A</mi></math><img file="EP1560085B1_D0008.tif" /></maths> The values of sin (θ), cos (θ) and tan (θ) are found in the same way as those in the embodiments shown in <figref idref="f0001">FIGS. 2 and 3</figref>. If the differences of those values to the values of sin (φ), cos (φ) and tan (φ) are within the allowable ranges, the result is regarded as normal. If the differences are out of the allowable ranges, the result can be judged to be a failure.

0026<figref idref="f0002">FIG. 5</figref> is still another detailed embodiment in which a phase shifter 45 is used to delay the phase of the exciting signal f(t) by 90 degrees and the comparator 41 is used to take zero cross points. The exciting signal f(t) is fed to the 90°-phase shifter 45 so that it is delayed in phase by 90 degrees. The delayed signal is compared to a reference value by the comparator 41 so that the zero cross point can be detected. The output from the comparator 41 is used as the conversion trigger. If the 90°-phase shifter 45 is formed of primary delay elements of a resistor and a capacitor, it is able to delay the phase of the exciting signal by about 90 degrees over a range from the cutoff frequency to a sufficiently high frequency.

0027<figref idref="f0002">FIG. 6</figref> is a waveform diagram showing the conversion timings for the resolver signals Y_sin and Y_cos in each embodiment shown in <figref idref="f0001 f0002">FIGS. 2 through 4</figref>. The abscissa is time, and the ordinate is the amplitudes of the resolver signals Y_sin and Y_cos, and the exciting signal f(t). As illustrated in <figref idref="f0005">FIG. 12</figref>, the resolver signals Y_sin and Y_cos are converted at the time when the exciting signal f(t) reaches V_ref, or at point A. In the example of <figref idref="f0001">FIG. 3</figref>, the resolver signals are converted at the time when the exciting signal f(t) has progressed t_delay from the zero cross point, or at time B. If t_delay-π/(4ω), the conversion is made at the peak of the signal, and thus it is optimum because the S/N ratio can be increased. According to the embodiment shown in <figref idref="f0002">FIG. 4</figref>, the resolver signals Y_sin and Y_cos are converted when the exciting signal f(t) reaches the peak.

0028<figref idref="f0003 f0004">FIGS. 7 through 9</figref> show embodiments of an abnormality detecting function 12 that detects from those obtained signals Y_sin, Y_cos the failure occurring in the resolver and resolver/digital converting portion. <figref idref="f0003">FIG. 7</figref> shows a first embodiment of the abnormality detecting function 12.

0029First, a computing portion 122 finds sin(θ) and cos(θ) according to the following expressions. <maths id="math0009"><math display="block"><mi>sin</mi><mfenced><mi>θ</mi></mfenced><mo>=</mo><msub><mi>Y</mi><mi>sin</mi></msub><mo>/</mo><mfenced open="[" close="]"><mi mathvariant="italic">SQRT</mi><mo>⁢</mo><mfenced><msup><msub><mi>Y</mi><mi>sin</mi></msub><mn>2</mn></msup><mo>+</mo><msup><msub><mi>Y</mi><mi>cos</mi></msub><mn>2</mn></msup></mfenced><mo>⋅</mo><msub><mi>V</mi><mi mathvariant="italic">ref</mi></msub></mfenced></math><img file="EP1560085B1_D0009.tif" /></maths><maths id="math0010"><math display="block"><mi>cos</mi><mfenced><mi>θ</mi></mfenced><mo>=</mo><msub><mi>Y</mi><mi>cos</mi></msub><mo>/</mo><mfenced open="[" close="]"><mi mathvariant="italic">SQRT</mi><mo>⁢</mo><mfenced><msup><msub><mi>Y</mi><mi>sin</mi></msub><mn>2</mn></msup><mo>+</mo><msup><msub><mi>Y</mi><mi>cos</mi></msub><mn>2</mn></msup></mfenced><mo>⋅</mo><msub><mi>V</mi><mi mathvariant="italic">ref</mi></msub></mfenced></math><img file="EP1560085B1_D0010.tif" /></maths>

0030A computing portion 121 finds sin (φ) and cosφ on the basis of the estimated angle φ. A comparing function 123 compares the values of sin(θ) and cos(θ) with the values of sin(φ) and cos (φ). If the differences are within the allowable ranges, the result is regarded as normal. If the differences are out of the allowable ranges, the result can be judged to be a failure. The failure of the resolver/digital converting portion can be detected by this method.

0031<figref idref="f0003">FIG. 8</figref> shows another embodiment of the abnormality detecting function 12.

0032A computing portion 124 finds tan(θ) from the following expression.<maths id="math0011"><math display="block"><mi>tan</mi><mfenced><mi>θ</mi></mfenced><mo>=</mo><msub><mi>Y</mi><mi>sin</mi></msub><mo>/</mo><msub><mi>Y</mi><mi>cos</mi></msub></math><img file="EP1560085B1_D0011.tif" /></maths>

0033A computing portion 125 finds tan (φ) on the basis of the estimated angle φ. The comparing function 123 compares the tan(θ)and the tan (φ). If the difference is within the allowable range, the result is regarded as normal. If the difference is out of the range, the result can be judged to be a failure. The failure of the resolver/digital converting portion can be detected by this method.

0034<figref idref="f0004">FIG. 9</figref> is still another embodiment of the abnormality detecting function 12.

0035The computing portion 122 finds sin(θ) and cos (θ), and the computing portion 124 finds tan (θ). Computing portions 126 through 128 respectively find θ₁, θ₂ and θ₃ by substituting the obtained values of sin (θ), cos (θ) and tan (θ) into the following expressions.<maths id="math0012"><math display="block"><msub><mi>θ</mi><mn>1</mn></msub><mo>=</mo><mi>a</mi><mo>⁢</mo><mi>sin</mi><mfenced><mi>sin</mi><mfenced><mi>θ</mi></mfenced></mfenced></math><img file="EP1560085B1_D0012.tif" /></maths><maths id="math0013"><math display="block"><msub><mi>θ</mi><mn>2</mn></msub><mo>=</mo><mi>a</mi><mo>⁢</mo><mi>cos</mi><mfenced><mi>cos</mi><mfenced><mi>θ</mi></mfenced></mfenced></math><img file="EP1560085B1_D0013.tif" /></maths><maths id="math0014"><math display="block"><msub><mi>θ</mi><mn>3</mn></msub><mo>=</mo><mi>a</mi><mo>⁢</mo><mi>tan</mi><mfenced><mi>tan</mi><mfenced><mi>θ</mi></mfenced></mfenced></math><img file="EP1560085B1_D0014.tif" /></maths> The comparing function 123 compares the obtained values of θ₁, θ₂ and θ₃ with each other. If the differences are within the allowable ranges, the result is regarded as normal. If the differences are out of the ranges, the result can be judged to be a failure.

0036This method is able to check the reasonableness of the signal from the resolver and detect the failure of the resolver itself.

0037<figref idref="f0004">FIG. 10</figref> shows an embodiment of the controller 100 that has both normal mode input means 101 that supplies the normal components of resolver signals Y_sin and Y_cos to the next stage, and common mode input means 102 that supplies the common mode components of the resolver signals to the next stage. <figref idref="f0005">FIG. 11</figref> shows another embodiment of the controller 100 that has both differential input means 103 that supplies the resolver signals Y_sin and Y_cos as differential signals to the next stage and single-end input means 104 that supplies the resolver signals as single end signals to the next stage.

0038If a short circuit occurs between the winding phases of the resolver (between the windings of exciting signal, sine signal and cosine signal), the normal resolver signals (broken line) cannot be obtained when a single-end configuration is used as shown in <figref idref="f0005">FIG. 12</figref>. The resolver signals are as indicated by the solid line because they are affected by the inter-phase short circuit. When a differential configuration is used, however, the component affected by the inter-phase short circuit is superimposed as a common mode component upon the original signal (normal mode component: broken line) to produce the waveform indicated by the solid line. Thus, if only a differential amplifier takes out the normal mode component, the original resolver signals (solid line) can be produced. Therefore, if the resolver signals Y_sin and Y_cos are received by differential-configuration input means, the inter-phase short circuit effect can be avoided. However, the occurrence of the inter-phase short circuit cannot be detected under this condition. Thus, if the common mode input means 102 or single end input means 104 is provided as in the examples of <figref idref="f0004">FIGS. 10</figref> and <figref idref="f0005">11</figref> to detect the common mode component that is to be superimposed due to the inter-phase short circuit, a control system can be realized that is not affected by the inter-phase short circuit and that can detect the inter-phase short.

0039<figref idref="f0006">FIG. 13</figref> shows a specific example of <figref idref="f0004">FIGS. 10</figref> and <figref idref="f0005">11</figref>.

0040The resolver signals Y_sin and Y_cos produced from the resolver 5 are supplied through the normal mode input means 101 or differential input means 103 to the resolver/digital converting portion 2, and also fed through the common mode input means 102 or single end input means 104 to the A/D converter 11. Even if an inter-phase short circuit occurs, the resolver/digital converting portion 2 is not affected by the inter-phase short circuit, but can produce the estimated angle φ on the basis of the resolver signals Y_sin and Y_cos. On the other hand, the input signals to the A/D converter 11 are affected by the inter-phase short circuit. If the error detecting function 12 monitors the input to the A/D converter 11, the inter-phase short circuit occurrence can be detected. When the short circuit occurs, the signal waveform of the signals supplied to the resolver/digital converting portion 2 that is not affected by the inter-phase short circuit is different from that fed to the A/D converter 11 that is affected by the inter-phase short circuit, and thus the inter-phase short circuit can be detected from the difference between the values of the trigonometric functions based on the signal waveforms or the estimated angles. The detection method in the error detecting function 12 may be the method shown in <figref idref="f0003 f0004">FIGS. 7 through 9</figref>.

0041Contrary to <figref idref="f0006">FIG. 13</figref>, the resolver signals Y_sin and Y_cos can be supplied through the common mode input means 102 or single end input means 104 to the resolver/digital converting portion 2, and also fed through the normal mode input means 101 or differential input means 103 to the A/D converter 11 of microcomputer, in which case the same effect can also be obtained.

0042<figref idref="f0007">FIG. 14</figref> shows a more specific example of <figref idref="f0004">FIGS. 10</figref> and <figref idref="f0006">13</figref>. The solver signals Y_sin and Y_cos are supplied to the resolver/digital converting portion 2 and to the A/D converter 11 of the microcomputer 1, and the common mode input means 102 is also provided. In <figref idref="f0006">FIG. 13</figref>, the exciting signal generator 3, resolver 5 and conversion trigger generator 4 are not shown for the sake of simplicity. The resolver/digital converting portion 2 also serves as the normal mode input means 101 that supplies the normal mode signals to the next stage or as the differential input means 103 that supplies the differential signals to the next stage. The resolver/digital converting portion 2 has, provided at the first stage, differential amplifiers 21 and 22 through which the resolver signals Y_sin and Y_cos can be received as differential signals. Differential amplifiers 51 and 52 are also provided before the A/D converter 11 of the microcomputer 1 in order to receive the resolver signals Y_sin and Y_cos as differential signals. In addition, resistors R₁, R₂ and resistors R₃, R₄ that constitute the common mode input means 102 can be used to add the common mode components of the resolver signals Y_sin and Y_cos. As a result, operational amplifiers 61, 62 produce common mode components of resolver signals Y_sin and Y_cos with the normal mode components cancelled out.

0043<figref idref="f0007">FIG. 15</figref> shows another specific example of <figref idref="f0005">FIGS. 11</figref> and <figref idref="f0006">13</figref>. As illustrated in <figref idref="f0007">FIG. 15</figref>, if the resolver signals Y_sin and Y_cos are supplied as differential signals to the resolver/digital converter portion 2, and fed as single-end signals to the A/D converting portion 11 of microcomputer 1, the resolver/digital converting portion 2 is not affected by the inter-phase short circuit, but the signals affected by the inter-phase short circuit are fed to the A/D converter 11 of microcomputer 1. As a result, the inter-phase short circuit occurrence can be detected from the abnormality of signal waveforms. In this example, the resolver/digital converting portion 2 also serves as the differential input means 103 of <figref idref="f0010">FIG. 21</figref>, and the A/D converter 11 of microcomputer 1 also serves as the single-end input means 104.

0044Thus, as described above, the inter-phase short circuit occurrence can be detected from the abnormality of signal waveforms by use of the common mode input means 102 or single end input means 104. However, since the electric noise often comes as common mode components, use of single-end input means 104 for one input side as in the example of <figref idref="f0006">FIG. 13</figref> results in the reduction of noise resistance so that the redundancy effect of the input circuits 103, 104 is reduced. Thus, examples for detecting the inter-phase short circuit occurrence with the noise resistance kept are shown in <figref idref="f0008 f0009 f0010 f0011 f0012">FIGS. 16 through 25</figref>.

0045<figref idref="f0008">FIG. 16</figref> shows an embodiment of the invention in which a single-end/differential input means 105 and the differential input means 103 are provided that are supplied with the resolver signals Y_sin and Y_cos. This single-end/differential input means 105 switches the single end input means and differential input means and receives the resolver signals through the selected input means. As illustrated, the single-end/differential input means 105 is used instead of the single end input means 104 in order to normally supply the redundant differential input signals to the next stage, and when a diagnosis is made in booting the system just after the power supply is turned on, to supply the single end signals to the next stage so that the inter-phase short circuit can be detected.

0046<figref idref="f0008">FIG. 17</figref> shows a specific example of the single-end/differential input means 105. The single-end/differential input means 105, as illustrated, has an operational amplifier OpAmp, a switch SW, and resistors. The output Vo of the operational amplifier OpAmp is supplied to and converted by the A/D converter 11 not shown to digital data. In order for the controller 100 to reduce the size and weight, it is desired to incorporate the A/D converter 11 in the microcomputer 1 not shown, but it may be discrete as long as the function can be kept. The switch SW is desired to control by the microcomputer 1, but it is not always necessary to do so.

0047The resistors R_b3, R_b4 are used to divide the power supply voltage V_cc and supply a bias voltage (center voltage) to the signals fed from the resolver. If R_b1=R_b2, the bias voltage (center voltage) equals to V_cc/2. The resistors R_b3, R_b4 are used to divide the source voltage V_cc to produce a voltage V_b2. If R_b3=R_b4, the voltage V_b2 equals to V_cc/2. The signal Ysin or Ycos from the resolver is differentially supplied as an input voltage V_inp, V_inn in this circuit, or each signal Y_sin or Y_cos is fed to each of two circuits that constitute the single-end/differential input means 105.

0048The relation between the input voltage V_inp, V_inn and output voltage V_o is as follows.

0049When the switch SW is opened, the equation of<maths id="math0015"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>R</mi><mi>f</mi></msub><mo>/</mo><msub><mi>R</mi><mi>i</mi></msub><mo>⁢</mo><mfenced><msub><mi>V</mi><mi mathvariant="italic">inp</mi></msub><mo>-</mo><msub><mi>V</mi><mi mathvariant="italic">inn</mi></msub></mfenced><mo>+</mo><msub><mi>V</mi><mrow><mi>b</mi><mo>⁢</mo><mn>1</mn></mrow></msub></math><img file="EP1560085B1_D0015.tif" /></maths> is satisfied.

0050When the switch SW is closed, the equation of<maths id="math0016"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>R</mi><mi>f</mi></msub><mo>/</mo><msub><mi>R</mi><mi>i</mi></msub><mo>⁢</mo><mfenced><msub><mi>V</mi><mrow><mi>b</mi><mo>⁢</mo><mn>1</mn></mrow></msub><mo>-</mo><msub><mi>V</mi><mi mathvariant="italic">inn</mi></msub></mfenced><mo>+</mo><msub><mi>V</mi><mrow><mi>b</mi><mo>⁢</mo><mn>1</mn></mrow></msub></math><img file="EP1560085B1_D0016.tif" /></maths> is satisfied.

0051Therefore, when the switch SW is opened, the single-end/differential input means operates as the differential input means, and when the switch SW is closed, it operates as the single end input means. In other words, the differential input means and single-end input means can be switched by opening and closing the switch SW.

0052In addition, even if the input means to which the signals are fed from the resolver is not made to be redundant, the input means can be constructed as the single-end/differential input means 105. According to this embodiment, the input means can be operated as a differential input means in the normal mode, and operated as a single-end input means in the mode in which the system makes a diagnosis while it is being booted immediately after the power supply is turned on, so that the inter-phase short circuit can be detected.

0053<figref idref="f0009">FIG. 18</figref> shows an embodiment in which the differential input means 103 is provided to receive the resolver signals Y_sin, Y_cos and supply differential input signals to the next stage, and frequency division input means 106 is provided to receive the resolver signals and supply the single-end input and differential input signals in a frequency division manner to the next stage. According to this embodiment, the differential input means that is resistant to noise is used for the frequency region including noise components, and the single-end input means is used for the frequency band that is little affected by noise components but much affected by the inter-phase short circuit, so that both inter-phase short detecting function and noise resistance can be made compatible. The frequency division input means 106 can be realized by inserting filters (high-pass filter, low-pass filter, band-pass filter and band-stop filter) having pass characteristics depending upon frequency in one of the two input lines of the differential input circuit, or inserting filters having different pass characteristics depending upon frequency in the two input lines. <figref idref="f0009 f0010 f0011 f0012">FIGS. 19 through 25</figref> show detailed examples of the input means.

0054<figref idref="f0009">FIG. 19</figref> shows an example of the characteristics of input means 106 designed to serve as the single-end input means for the lower frequency region and as the differential input means for the higher frequency region including the resolver signal frequency. This example is effective particularly when the resolver signal frequency and higher frequency region are affected by noise and when the inter-phase short effect appears in the lower frequency region. For example, when the switching frequency of the PWM inverter for driving a motor is near the resolver signal frequency, the resolver frequency and higher frequency region are affected by noise.

0055In addition, if a different center voltage (bias voltage) is used for each phase, the lower frequency region is affected by the inter-phase short circuit. For example, since the input voltage range of the resolver signals Y_sin, Y_cos to the resolver/digital converting portion 2 and A/D converter 11 is normally from 0 to 5 volts, the center voltage (bias voltage) is generally fixed to about 2.5 V. In addition, since the exciting signal produced from the exciting signal generator 3 is desired to be as large as possible in order for the S/N ratio to be raised, the center voltage (bias voltage) is usually set at about 1/2 the battery voltage. If the battery voltage is 12 V, the center voltage is fixed to 6 V. When considering that the battery voltage is gradually reduced while it is being operated, the center voltage (bias voltage) is generally fixed to about 4 ∼ 5 V. Therefore, when the inter-phase short circuit occurs between the resolver signal Y_sin, Y_cos and the exciting signal, the center voltage (bias voltage) of the resolver signal Y_sin, Y_cos is pulled to rise from about 2.5 V up to the center voltage (bias voltage) of the exciting signal, and thus the short circuit can be detected from this change. In addition, if the center voltage (bias voltage) of the resolver signal Y_sin, Y_cos is slightly varied around 2.5 V, the inter-phase short circuit between the resolver signals Y_sin, Y_cos can be similarly detected from this change.

0056In order to achieve the frequency characteristic shown in <figref idref="f0009">FIG. 19</figref>, it is necessary that a high-pass filter be inserted in one of the two input lines of the differential input circuit. <figref idref="f0009">FIG. 20</figref> shows a specific example of this case. This circuit arrangement has a capacitor C inserted in series with one input line of the two input lines of the normal differential input circuit. This capacitor makes the input circuit be a single-end input circuit in terms of DC and makes it be a differential input circuit in terms of AC. The signal Y_sin or Y_cos from the resolver is differentially supplied as the input voltages V_inp and Vinn to this input circuit. Thus, the frequency-division input means 106 is constructed by a circuit for each signal Y_sin or Y_cos, or by two circuits for both signals.

0057If the capacitor C in <figref idref="f0009">FIG. 20</figref> is imaginarily opened, an equivalent circuit for DC region is obtained. The input/output relation in this equivalent circuit can be expressed by the following equation.<maths id="math0017"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>R</mi><mi>f</mi></msub><mo>⋅</mo><msub><mi>V</mi><mi mathvariant="italic">inp</mi></msub><mo>/</mo><mfenced><msub><mi>R</mi><mi>i</mi></msub><mo>+</mo><msub><mi>R</mi><mi>f</mi></msub></mfenced></math><img file="EP1560085B1_D0017.tif" /></maths>

0058If the capacitor C in <figref idref="f0009">FIG. 20</figref> is imaginarily shorted, an equivalent circuit for AC region is obtained. The input/output relation of this equivalent circuit is given by<maths id="math0018"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>R</mi><mi>f</mi></msub><mo>/</mo><msub><mi>R</mi><mi mathvariant="italic">i</mi></msub><mo>⁢</mo><mfenced><msub><mi mathvariant="italic">V</mi><mi mathvariant="italic">inp</mi></msub><mo mathvariant="italic">-</mo><msub><mi mathvariant="italic">V</mi><mi mathvariant="italic">inn</mi></msub></mfenced></math><img file="EP1560085B1_D0018.tif" /></maths>

0059Thus, for AC region this circuit arrangement operates as a differential input configuration, and for DC region it operates as a single-end input configuration. Therefore, the frequency characteristic shown in <figref idref="f0009">FIG. 19</figref> can be achieved by this example, and thus the inter-phase short circuit detection function and the resistance to noise can be made compatible.

0060<figref idref="f0010">FIG. 21</figref> shows another embodiment of the frequency-division input means 106 in which the gain for DC region is fixed at 1.

0061The input/output relation for AC region is the same as in <figref idref="f0009">FIG. 20</figref>, or given by<maths id="math0019"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>R</mi><mi>f</mi></msub><mo>/</mo><msub><mi>R</mi><mi>i</mi></msub><mo>⁢</mo><mfenced><msub><mi mathvariant="italic">V</mi><mi mathvariant="italic">inp</mi></msub><mo mathvariant="italic">-</mo><msub><mi mathvariant="italic">V</mi><mi mathvariant="italic">inn</mi></msub></mfenced></math><img file="EP1560085B1_D0019.tif" /></maths>

0062The input/output relation for DC region is as follows. The input signal V_inp is not divided, but directly supplied to the + input terminal of the operational amplifier, and the output voltage V_o is controlled so that the - input terminal of the operational amplifier can be maintained at V_inp. That is,<maths id="math0020"><math display="block"><msub><mi>V</mi><mi mathvariant="italic">inp</mi></msub><mo>=</mo><mfenced><msub><mi>R</mi><mi>f</mi></msub><mo>⋅</mo><msub><mi>V</mi><mi mathvariant="italic">inn</mi></msub><mo>+</mo><msub><mi>R</mi><mi>i</mi></msub><mo>⋅</mo><msub><mi>V</mi><mi>o</mi></msub></mfenced><mo>/</mo><mfenced><msub><mi>R</mi><mi>i</mi></msub><mo>+</mo><msub><mi>R</mi><mi>f</mi></msub></mfenced></math><img file="EP1560085B1_D0020.tif" /></maths> For DC region, V_inp=V_inn is satisfied, and thus<maths id="math0021"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>V</mi><mi mathvariant="italic">inp</mi></msub></math><img file="EP1560085B1_D0021.tif" /></maths>

0063From the above, it will be understood that for AC region the input means operates as a differential input configuration and for DC region the input means operates as a single-end input configuration of gain 1. According to this example, since the gain for DC region can be fixed at 1, the center values of the signals in the differential input means 103, frequency-division input means 106 and the following-stage A/D converter 11 can be set to be the optimum points for treating the maximum amplitude, or 1/2 the operating source voltage to the circuits.

0064<figref idref="f0010">FIG. 22</figref> shows an embodiment of the invention in which it is prevented that, if the center voltages (bias voltages) of the signals Y_sin, Y_cos are made different in order to detect the inter-phase short circuit between the signals Y_sin, Y_cos, the center voltage (bias voltage) of the input signal to the resolver/digital converter 2 may be deviated from an ideal value (for example, 2.5 V) to cause the waveforms of the upper and lower ends of the signal to be clipped due to the saturation, thus limiting the maximum amplitude. In this embodiment, a bias applying circuit 107 is provided to apply different center voltages (bias voltages) to the signals Y_sin, Y_cos. Then, a bias correcting circuit 108 adds a correction bias to correct the center voltages (bias voltages) into an ideal value (for example, 2.5 V), and supplies the corrected signals to the normal mode input means 101 or differential input means 103 that are placed at the input stage of the resolver/digital converter 2. Therefore, when the inter-phase short circuit does not occur, the signals Y_sin, Y_cos fed to the resolver/digital converter 2 have an ideal value (for example, 2.5 V) as their center voltage (bias voltage) since they are corrected after predetermined different center voltages (bias voltages) are added to the resolver signals, and thus those signals have the ideal value (for example, 2.5 V) as the center voltage (bias voltage). When the inter-phase short circuit occurs, the center voltages applied to the signals Y_sin, Y_cos are deviated from a certain value, and thus the signals Y_sin, Y_cos fed to the resolver/digital converting portion 2 also have the center voltages (bias voltages) deviated from the ideal value (for example, 2.5 V) so that the inter-phase short circuit occurrence can be detected.

0065Also, even if the resolver signals are received by the differential input means but not the single-end input means, the waveforms of the upper and lower ends of the signal of a narrow input voltage range are clipped due to saturation, and thus the inter-phase short circuit can be detected as abnormal waveforms.

0066This embodiment functions effectively when the input voltage range to the resolver/digital converting portion 2 is narrow. However, when the input voltage range to the A/D converter 11 or common mode input means 102 or single-end input means 104 is narrow, the bias correcting circuit 108 is similarly needed to place before the common mode input means 102 or single-end input means 104.

0067<figref idref="f0011">FIG. 23</figref> shows an example of the frequency characteristic of the frequency division input means 106 that takes the single-end input configuration for the frequency region around the resolver signal frequency. This example is effective particularly when the region around the resolver signal frequency is little affected by noise, but when the other frequency regions are much affected by noise. For example, this corresponds to the case when the switching frequency of the PWM inverter for driving the motor that is the main noise source is much lower or much higher than the resolver signal frequency.

0068The following three methods can be considered in order to achieve the frequency characteristic shown in <figref idref="f0011">FIG. 23</figref>.

0069The first method is that a filter for eliminating the frequency region around the resolver signal frequency is inserted in one of the two input lines of the differential input circuit. According to this method, the differential input means functions for the other frequency regions.

0070The second method is that a high-pass filter with a cutoff frequency sufficiently lower than the resolver signal frequency is inserted in one of the two input lines of the differential input circuit and that another high-pass filter with a cutoff frequency sufficiently higher than the resolver signal frequency is inserted in the other line. According to this method, the differential input means functions for the frequency band higher than the resolver signal frequency region, but the lower frequency band than the resolver signal frequency region is cut off.

0071The third method is that a low-pass filter with a cutoff frequency sufficiently lower than the resolver signal frequency is inserted in one of the two input lines of the differential input circuit and that a low-pass filter with a cutoff frequency sufficiently higher than the resolver signal frequency is inserted in the other line. According to this method, the differential input means operates for the lower frequency band than the resolver signal frequency region, but the higher frequency band than the resolver signal frequency region is cut off.

0072<figref idref="f0011">FIG. 24</figref> shows an example of the circuits for implementing the second method. The cutoff frequency of the high-pass filter can be substantially determined by the time constants of C₁ · Ri, C₂ . R_i.

0073Here, under the condition of<maths id="math0022"><math display="block"><mn>1</mn><mo>/</mo><mfenced><mn>2</mn><mo>⁢</mo><mi>π</mi><mo>⁢</mo><msub><mi>C</mi><mn>1</mn></msub><mo>⋅</mo><msub><mi>R</mi><mi>i</mi></msub></mfenced><mo>⁢</mo><mfenced><msub><mi>f</mi><mi mathvariant="italic">rez</mi></msub><mo>〈</mo><mn>1</mn><mo>/</mo><mfenced><mn>2</mn><mo>⁢</mo><mi>π</mi><mo>⁢</mo><msub><mi>C</mi><mn>2</mn></msub><mo>⋅</mo><msub><mi>R</mi><mi>i</mi></msub></mfenced><mo>〉</mo></mfenced></math><img file="EP1560085B1_D0022.tif" /></maths> where f_rez is the frequency of the resolver signals, the very high frequency region of f&lt;&lt;1/ (2πC₁· R_i) to the + input end and - input end of the operational amplifier is cut off, resulting in<maths id="math0023"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>V</mi><mrow><mi>b</mi><mo>⁢</mo><mn>2</mn></mrow></msub></math><img file="EP1560085B1_D0023.tif" /></maths>

0074The frequency region of<maths id="math0024"><math display="block"><mn>1</mn><mo>/</mo><mfenced><mn>2</mn><mo>⁢</mo><mi>π</mi><mo>⁢</mo><msub><mi>C</mi><mn>1</mn></msub><mo>⋅</mo><msub><mi>R</mi><mi>i</mi></msub></mfenced><mo>&lt;</mo><mo>&lt;</mo><mi>f</mi><mo>&lt;</mo><mo>&lt;</mo><mn>1</mn><mo>/</mo><mfenced><mn>2</mn><mo>⁢</mo><mi>π</mi><mo>⁢</mo><msub><mi>C</mi><mn>1</mn></msub><mo>⋅</mo><msub><mi>R</mi><mi>i</mi></msub></mfenced></math><img file="EP1560085B1_D0024.tif" /></maths> to the + input end of the operational amplifier is cut off, resulting in<maths id="math0025"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>R</mi><mi>f</mi></msub><mo>⋅</mo><mfenced><msub><mi>V</mi><mrow><mi>b</mi><mo>⁢</mo><mn>2</mn></mrow></msub><mo>-</mo><msub><mi>V</mi><mi mathvariant="italic">inn</mi></msub></mfenced><mo>/</mo><msub><mi>R</mi><mi>i</mi></msub><mo>+</mo><msub><mi>V</mi><mrow><mi>b</mi><mo>⁢</mo><mn>2</mn></mrow></msub></math><img file="EP1560085B1_D0025.tif" /></maths> and thus the single-end operation.

0075The very high frequency region of 1/(2πC₂· R_i)&lt;&lt;f is passed through the + input end and - input end of the operational amplifier, resulting in<maths id="math0026"><math display="block"><msub><mi>V</mi><mi>o</mi></msub><mo>=</mo><msub><mi>R</mi><mi>f</mi></msub><mo>⋅</mo><mfenced><msub><mi>V</mi><mi mathvariant="italic">inp</mi></msub><mo>-</mo><msub><mi>V</mi><mi mathvariant="italic">inn</mi></msub></mfenced><mo>/</mo><msub><mi>R</mi><mi>i</mi></msub><mo>+</mo><msub><mi>V</mi><mrow><mi>b</mi><mo>⁢</mo><mn>2</mn></mrow></msub></math><img file="EP1560085B1_D0026.tif" /></maths> and thus the differential input operation.

0076According to the examples shown in <figref idref="f0007 f0008 f0009 f0010 f0011">FIGS. 15 through 24</figref> mentioned above, a control system can be achieved that is not affected by the inter-phase short but that can detect the inter-phase short.

0077Thus, according to the above examples, as listed in <figref idref="f0012">FIG. 25</figref>, the failure (1) of the abnormality of waveforms due to the disconnection of the signal from or to the resolver can be detected by the failure detecting function provided in the resolver/digital converter 2 and the error detecting function 12 of the invention.

0078In addition, since the function to detect the failure (1) and notify the microcomputer is redundantly provided in the resolver/digital converting portion 2 and detecting function 12, the failure (2) of the function to detect the failure (1) and notify the microcomputer can be detected by the other function even if one function breaks down.

0079The failure (3) of the resolver/digital converter function itself for finding θ on the basis of the signal from the resolver can be detected by the failure detecting function provided in the resolver/digital converting portion 2 and the error detecting function 12 of the invention.

0080In addition, since the function to detect the failure (3) and notify the microcomputer is redundantly provided in the resolver/digital converting portion 2 and the detecting function 12, the failure (4) of the function to detect the failure (3) and notify the microcomputer can also be detected by the other function even if one function goes out of order.

0081Thus, according to the above examples, the failures of the resolver and resolver/digital converter can be detected. If a failure is detected in the application to the motor-driven power steering apparatus or the like, the motor-driven assist mechanism is stopped, thus the fail-safe operation being assured. In addition, since the system operation is not allowed to stop in x-by-write, particularly steer-by-wire or fly-by-wire, the fault tolerance is demanded that can continue the operation even when a failure occurs. Thus, as illustrated in <figref idref="f0012">FIG. 26</figref>, if resolvers 5-1, 5-2 and resolver/digital converting portions 2-1, 2-2 are doubly provided, and combined with the error detecting function 12, a triple system can be substantially achieved to form a fault tolerance system. If one of the double resolvers 5-1, 5-2 and resolver/digital converting portions 2-1, 2-2 breaks down, the failure can be detected by the failure detecting function provided in the resolver/digital converting portions 2-1, 2-2 and the error detecting function 12, and the defective portion can be located. Then, the failure-free ones of the double resolvers 5-1, 5-2, resolver/digital converting portions 2-1, 2-2 can be used to maintain the operations. It would be further desirable to make the microcomputer 1 be a multiple structure.

0082<figref idref="f0013">FIG. 27</figref> shows an embodiment of a motor control system according to the invention. The microcomputer 1 is responsive to the estimated angle φ from the resolver/digital converting portion 2 to supply a motor-drive command to a timer 6 that generates a PWM signal. The timer 6 responds to the command from the microcomputer 1 to generate the PWM signal of a predetermined duty cycle. A motor driver 7 drives a motor 8 according to the waveform of the PWM signal. The output shaft of motor 8 is connected to a controlled object 9 and thus moves the controlled object 9. The output shaft of the motor is also connected to the resolver 5 by which the rotation angle is measured. The rotation angle is supplied to the microcomputer 1.

0083In the motor-driven power steering control apparatus, the controlled object 9 is the whole steering system. In the steer-by-wire control apparatus, the controlled object 9 is the steering column and steering mechanism (steerage apparatus).

0084Also, in the power steering control apparatus, as illustrated in <figref idref="f0014">FIG. 28</figref>, a torque monitoring function 14 is used to monitor the torque signal from a torque sensor 13, thus further improving the safety. The function of power steering apparatus is to assist the steerage by motive energy and to thereby alleviate the driver's steering force. In case that the power steering apparatus operates contrary to the driver's will due to a failure, the driver should exert an unusual steering force (torque) against this operation. Thus, the torque monitoring function 14 detects the driver's unusual steering force (torque) when the power steering apparatus breaks down, and stops the assist mechanism, thus assuring the failsafe operation.

0085When the steering force is aided by the power steering, the steering force (torque) is substantially 2 to 3 (Nm), but when the driver widely turns the steering wheel, it is about 15 (Nm). Therefore, when the driver exerts a steering force (torque) that exceeds 20 to 30 (Nm) including a margin, the power steering apparatus is regarded as being broken down, and the assist mechanism is stopped so that the failsafe operation can be assured. In addition, since the widely turning of the steering wheel is limited to the low speed of car, various thresholds to the steering force (torque) are provided for the detection of the abnormality in different speeds of car as shown in <figref idref="f0015">FIG. 29</figref>. The threshold is stepwise changed in accordance with the speed range of car. That is, for high speed the threshold is selected to be 5 to 6 (Nm) and for low speed it is fixed to 20 to 30 (Nm). Alternatively, by continuously changing the threshold according to the car speed, it is possible to precisely detect the abnormality. Moreover, in order that the assist mechanism can be prevented from being stopped when the steering torque is transient and thus normal, it can be considered that the assist mechanism is stopped when the steering force continues to exceed the above threshold during a constant time or above.

0086<figref idref="f0015">FIG. 30</figref> shows an embodiment of a redundant-structure control apparatus. Redundant multiple controllers 100-1 to 100-n supply control signals 110-1 to 110-n as control commands to actuators 200-1 to 200-n. The controllers 100-1 to 100-n also produce the results 111-1 to 111-n of their own diagnosis. Reliability order judging portions 400-1 to 400-n produce ON/OFF signals 410-1 to 400-n on the basis of the diagnosis results 111-1 to 111-n supplied from the controllers 100-1 to 100-n.

0087The actuator 200-i, when the input ON/OFF signal 410-i is ON, controls a controlled object 300 to act by the amount based on the inputted control signal 110-i. When the input ON/OFF signal 410-i is OFF, it does not control the controlled object 300.

0088<figref idref="f0016">FIG. 31</figref> shows a table of the judgment in the reliability order judging portions 400-1 to 400-n, or the relation between the diagnosis result 111-1 to 111-n and the ON/OFF signal 410-1 to 410-n. In this table, if the reliability order judging portion 400-1 is now taken as an object of interest, its number is i=1, and the numbers of the other related objects are j= 2 ∼ n. Also, if the reliability order judging portion 400-2 is now taken as an object of interest, its number is i=2, and the numbers of the other related objects are j=1, 3 ∼ n. As shown in this table, when the failure grade of an object of interest based on the diagnosis result 111-i from this object is lower than those of the other objects based on the diagnosis results 111-j from the other objects, the ON/OFF signal 410-i of interest is turned ON. On the contrary, when the failure grade of an object of interest based on the diagnosis 111-i is higher than those of the other objects based on the diagnosis results 111-j from the other objects, the ON/OFF signal 410-i of interest is turned OFF.

0089If the case 1 is taken as an example, the diagnosis result 111-i of interest has no failure, and thus the ON/OFF signal 410-i of interest is turned ON. If the case 2 is taken, the diagnosis result 111-i of interest has a minor failure, but the diagnosis results 111-j of the other objects also have a minor failure, and thus the ON/OFF signal 410-i of interest is turned ON. If the case 3 is taken, the diagnosis results 111-j of the other objects have no failure, but the diagnosis result 111-i of interest has a minor failure, thus the ON/OFF signal 410-i of interest being turned OFF. If the case 4 is taken, both the diagnosis result 111-i of interest and the diagnosis results 111-j of the other objects have middle failures, and thus the ON/OFF signal 410-i of interest is turned ON. If the case 5 is taken, the diagnosis results 111-j of the other objects have middle failure, but the diagnosis result 111-i of interest has a serious failure, thus the ON/OFF signal 410-i of interest being turned OFF.

0090In addition, it is also possible to add the judgment that, when the diagnosis result 111-i of interest has a failure more serious than a certain degree, the ON/OFF signal 410-i of interest is turned OFF irrespective of what degree the diagnosis results of the other objects show. If the case 6 is taken, the diagnosis result 111-i of interest shows a serious failure, and thus the ON/OFF signal 410-i of interest is turned OFF independently of the diagnosis results 111-j of the other objects.

0091<figref idref="f0016">FIG. 32</figref> shows an example of the definition of failure grades. The grades of failure can be classified according to the scale of the function lost due to the failure or to the probability of the occurrence of failure indicated by the diagnosis result. In the embodiments of the invention, since the failure of sensor is not detected by diagnosis until it becomes obvious, the degrees of failure in the sensor are classified according to the scale of the function lost due to the failure of the sensor in the former case. Since the memory can be recovered from failure by the error correction code (ECC), the degrees of failure in the memory can be classified according to the possibility of recovery, or to the probability of the occurrence of failure indicated by the diagnosis result in the latter case. The degrees of failure in the processor can be classified as minor or middle according to the probability of failure occurrence indicated by the diagnosis result, and as serious when the processor function is seriously damaged.

0092<figref idref="f0017 f0018">FIGS. 33 through 36</figref> are block diagrams showing embodiments of the invention in which the operation of actuator 200-i is controlled by using ON/OFF signal 410-i. <figref idref="f0017">FIG. 33</figref> shows an embodiment in which the ON/OFF signal 410-i controls a main switch 211 for the power supply to the actuator 200-i that drives a motor 210 by use of an inverter formed of semiconductor devices 201 to 206. When the ON/OFF signal 410-i is ON, the main switch 211 is closed, and the inverter formed of semiconductor devices 201 to 206 in the actuator 200-i is actuated according to the control signal 110-i to drive the motor 210. When the ON/OFF signal 410-i is OFF, the main switch 211 is opened, and the inverter of semiconductor devices 201 to 206 is stopped from producing the output, thus ceasing to drive the motor 20.

0093<figref idref="f0017">FIG. 34</figref> shows an embodiment in which a switch 212 is opened or closed by the ON/OFF signal 410-i to allow the inverter to supply or not to supply the output to the motor 210. When the ON/OFF signal 410-i is ON, the switch 212 is closed, and the inverter of semiconductor devices 201 to 206 in the actuator 200-i is actuated by the control signal 110-i to drive the motor 210. When the ON/OFF signal 410-i is OFF, the switch 212 is opened, and the inverter of semiconductor devices 201 to 206 is stopped from producing the output, thus ceasing to drive the motor 210.

0094<figref idref="f0018">FIG. 35</figref> shows an embodiment in which a logical product (AND) 213 is used to control the gate signals to the semiconductor devices 201 to 206 on the basis of the ON/OFF signal 410-i. When the ON/OFF signal 410-i is ON, the inverter of semiconductor devices 201 to 206 is actuated by the control signal 110-i to drive the motor 210. When the ON/OFF signal 410-i is OFF, the gate signals to the semiconductors 201 to 206 are always OFF, and thus the inverter is stopped from producing the output, thus ceasing to drive the motor 210.

0095<figref idref="f0018">FIG. 36</figref> shows an embodiment in which a clutch 214 is provided on the output shaft of motor 210 and intermitted according to the ON/OFF signal 410-i. When the ON/OFF signal 410-i is ON, the clutch 214 is fastened, and thus the power of the motor 210 is, when the motor is driven, transmitted through its output shaft to the outside. When the ON/OFF signal 410-i is OFF, the clutch 214 is opened, and thus the output power of the motor 210 is not transmitted through the output shaft.

0096According to the above embodiments of the invention, it is possible to assure the reliability and safety of the high reliability control system, particularly of the steer-by-wire system, x-by-wire system, and the power steering system for large-size cars.

0097<figref idref="f0019">FIG. 37</figref> shows an embodiment in which this invention is applied to brakes. The right front brake and right back brake of the controlled object (automobile) 300 are controlled by a doubled control apparatus formed of a set of controllers 100-1, 100-2, actuators 200-1, 200-2, and reliability order judging portions 400-1, 400-2. The left front brake and left back brake of the controlled object (automobile) 300 are controlled by another doubled control apparatus formed of a set of controllers 100-3, 100-4, actuators 200-3, 200-4, and reliability order judging portions 400-3, 400-4. According to this embodiment, if a serious failure occurs in any one of the controller 100-1, actuator 200-1 and reliability order judging portion 400-1 to disable the right front brake, the right back brake is continuously controlled by the controller 100-2, actuator 200-2 and reliability order judging portion 400-2 even if a minor failure occurs in any one of the devices 100-2, 200-2 and 400-2 so that the right brakes of the controlled object (automobile) 300 can be maintained to operate, exerting a certain braking force without causing yaw moment. The same action is taken even if a serious failure occurs in the controller 100-2 ∼ 100-4, actuator 200-1 ∼ 200-4, or reliability order judging portion 400-1 ∼ 400-4.

0098<figref idref="f0020">FIG. 38</figref> shows an embodiment in which the control signals 110-1, 110-2 are switched by a selector 500 and supplied to a single actuator 200. The selector 500 may be Modified Votor (MV) described in Patent 2085879 that the inventor already filed. In addition, selectors 500-1 ∼ 500-n, and actuators 200-1 ∼ 200-n can be redundantly provided as illustrated in <figref idref="f0020">FIG. 39</figref>.

0099Thus, according to one embodiment of the invention, it is possible to provide a resolver/digital converter capable of detecting failures by simple computation.

0100In addition, according to another embodiment of the invention, it is possible to provide a resolver/digital converter that is not immediately affected by the inter-phase short circuit, and that can immediately detect the short circuit.

0101Moreover, according to still another embodiment of the invention, it is possible to provide a fault tolerance system capable of continuously operating as the whole system even if a failure occurs in both subsystems.