Patent Application
20060278022
The present invention relates to a rotation speed detecting apparatus and a load measuring apparatus of a rolling bearing unit. For example, the invention relates to an improvement in a rolling bearing unit for supporting a wheel of a moving member of an automobile, a railroad vehicle, various carrying vehicles or the like utilized for ensuring stability in moving the moving member by measuring a rotation speed of a rotating member constituting the moving bearing unit and measuring a load (one or both of radial load and axial load) loaded on the rolling bearing unit.
For example, a wheel of an automobile is rotatably supported by a suspension by a rolling bearing unit of a double row angular type. Further, in order to ensure a running stability of an automobile, there is used a vehicular running stabilizing apparatus of an antilock braking system (ABS), a traction control system (TCS), a vehicle stability control system (VSC) or the like. In order to control the various vehicular running stabilizing apparatus, a signal indicating a rotation speed of a vehicle, accelerations in respective directions applied on a vehicle body or the like becomes necessary. Further, in order to carry out a higher control, there is a case in which it is preferable to know a magnitude of a load (one or both of radial load and axial load) applied on the rolling bearing unit by way of the wheel.
In view of such a situation, JP-A-2001-21577 (hereinafter, described as “Patent Reference 1”) describes a rolling bearing unit having a load measuring apparatus capable of measuring a radial load. The rolling bearing unit having the load measuring apparatus according to a first example of the background art is for measuring the radial load and is constituted as shown by
An attaching hole 10 penetrating the outer ring 1 in a diameter direction is formed at a middle portion in an axial direction of the outer ring 1 and a portion thereof between double rows of the outer ring tracks 7, 7 at an upper end portion of the outer ring 1 substantially in a vertical direction. Further, inside of the attaching hole 10 is mounted with a displacement sensor 11 in a shape of a circular rod (rod-like shape) constituting a sensor for measuring a load. The displacement sensor 11 is of a noncontact type and a detecting face provided at a front end face (lower end face) thereof is made to be opposed proximately to an outer peripheral face of a sensor ring 12 outwardly fitted to be fixed by a middle portion in an axial direction of the hub 2. The displacement sensor 11 outputs a signal when a distance between the detecting face and the outer peripheral face of the sensor ring 12 is changed in correspondence with an amount of the change.
In the case of the rolling bearing unit having the load measuring apparatus of the background art constituted as described above, based on the detecting signal of the displacement sensor 11, the load applied to the rolling bearing unit can be calculated. That is, whereas the outer ring 1 supported by the suspension of the vehicle is pressed to a lower side by a weight of the vehicle, the hub 2 fixedly supporting the wheel is going to stay at a position as it is. Therefore, the more the weight is increased, the more the deviation between the center of the outer ring 1 and the center of the hub 2 is increased based on elastic deformations of the outer ring 1 and the hub 2 as well as the rolling members 9a, 9b. Further, the more the weight is increased the shorter the distance between the detecting face of the displacement sensor 11 and the outer peripheral face of the sensor ring 12 provided at the upper end portion of the outer ring 1. Hence, when the detecting signal of the displacement sensor 11 is transmitted to a controller, the radial load applied to the rolling bearing unit integrated with the displacement sensor 11 can be calculated from a relationship or a map previously calculated by an experiment or the like. Based on the load applied on the respective rolling bearing units calculated in this way, ABS is properly controller, further, a failure in a loading state is informed to a driver.
Further, according to the background art structure shown in
In using the rolling bearing unit integrated with the above-described rotation speed detecting apparatus, when the sensor rotor 13 is rotated along with the hub 2 fixed with the wheel and the detected portion of the sensor rotor 13 is made to run at a vicinity of the detecting portion of the rotation speed detecting sensor 15, an output of the rotation speed detecting sensor 15 is changed. A frequency of changing the output of the rotation speed detecting sensor 15 in this way is proportional to a rotational number of the wheel. Therefore, when an output signal of the rotation speed detecting sensor 15 is transmitted to a controller, not illustrated, ABS and TCS can pertinently be controlled.
Although the rolling bearing unit having the load measuring apparatus of the first example of the above-described background art structure is for measuring the radial load applied to the rolling bearing unit, also a structure of measuring an axial load applied to the rolling bearing unit is described in JP-A-3-209016 (hereinafter, described as “Patent Reference 2”) or the like and is known in the background art.
Further, respective load sensors 20 are attached to a plurality of portions of an inner side face of the fixed side flange 17 and portions thereof surrounding screw holes 19 for screwing bolts 18 for coupling the fixed side flange 17 of the knuckle 16. In a state of fixedly supporting the outer ring 1a by the knuckle 16, the respective load sensors 20 are supported between an outer side of the knuckle 16 and an inner side face of the fixed side flange 14.
In the case of the load measuring apparatus of the rolling bearing unit of the second example of the background art structure, when an axial load is applied between the wheel, not illustrated, and the knuckle 16, the outer side face of the knuckle 16 and the inner side face of the fixed side flange 17 strongly press the respective load sensors 20 from both faces thereof in an axial direction. Therefore, by totalizing measured values of the respective load sensors 20, the axial load applied between the wheel and the knuckle 16 can be calculated. Further, although not illustrated, JP-B-62-3365 (hereinafter, described as “Patent Reference 3”) describes a method of calculating a revolution speed of a rolling member and measuring an axial load applied to a rolling bearing from a vibration frequency of an outer ring corresponding member a rigidity of which is partially reduced.
In the above-described case of the first example of the background art structure shown in
Further, in the above-described case of the second embodiment of the background art structure shown in
Further, according to the rotation speed detecting apparatus for detecting rotation speeds of various rotating members used in the load measuring apparatus of the rolling bearing unit, when a rotational center of the member the rotation speed of which is to be detected and a geometrical center of an encoder do not coincide with each other, an accuracy of detecting the rotation speed is deteriorated. In order to prevent the accuracy of detecting the rotation speed from being deteriorated by such a cause, it is conceivable to eliminate an influence by a deviation of the two centers by summing together detecting signals of a pair of the rotation detecting sensors arranged at positions of two portions of the encoder on sides opposed to each other in the diameter direction. However, in that case, two pieces of the rotation detecting sensors are needed, by that amount, the cost and an installing space are increased thereby and therefore, there is also conceivable a case in which the apparatus is difficult to be adopted.
As a technology for removing a noise component having a comparatively low frequency, there is known an adaptive filter operated by an LMS algorism described in Haruo Hamada, “Fundamental of an adaptive filter (part 2)” proceeding of the Japan Acoustic Society, vol. 45, No. 9, (Corp.) Japan Acoustic Society, 1989, p. 731-738 (hereinafter, described as “Nonpatent Reference 1”). Further, with regard to an outline of an adaptive filter, there have been known: Chuo university, electric, electronic, information, communication engineering department, Chao research lab., “What is adaptive filter”, [online], [searched on August 29, Heisei 15], internet <URL: http://www.elect.chuo-u.ac.jp/chao/forB3/dsp/volterra/filter.htm 1> (hereinafter, described as “Nonpatent Reference 2”), The MathWorks, Inc., “Outline and application of adaptive filter”, [online], [searched on August 29, Heisei 15], internet <URL:http://www.mathworks.ch/access/helpdesk/jhelp/toolbox/filte rdesign/adaptiv2.shtml> (hereinafter, described as “Nonpatent Reference 3”), The Mathworks, Inc., “Example of adaptive filter using LMS algorism”, [online], [searched on August 29, Heisei 15], internet, <URL: http://www.mathworks.ch/access/helpdesk/jhelp/toolbox/filterdesi gn/adaptiv9.shtml> (hereinafter, described as “Nonpatent Reference 4”) and the like. Further, also with regard to a synchronizing type filter which is a kind of an adaptive filter, there has been known, for example, Haruo Hamada, other 3 “Synchronizing type adaptive filter and active application to active noise/vibration control”, proceeding of the Japan Acoustic Society, 3-5-13, (Corp.) Japan Acoustic Society, March, Heisei 4, p. 515-516 (hereinafter, described as “Nonpatent Reference 5”). Further, with regard to a technology of suppressing vibration of an engine by a synchronizing type LMS algorism, it has been known by being described in Shigeki, Sato, other 4 “Development of active mount”, automobile technology, (Corp.) The Society of Automobile Engineers of Japan, Vol. 53, No. 2, February 1999, p. 62-66 (hereinafter described as “Nonpatent Reference 6”. However, in the background art, the above-described adaptive filter is used centering on so-to-speak active noise control for reducing low frequency noise by emitting a sound wave having a phase inverse to that of low frequency noise. That is, in the background art, the adaptive filter is used only for reducing low frequency noise such that low frequency noise emitted from a duct of an air conditioner to a room is reduced, or exhaust sound or running sound at low frequency entering inside of a compartment of a passenger vehicle is reduced, further, external noise at a low frequency entering from an outside of a headphone is reduced or the like. The technology described in Nonpatent Reference 6 is aimed at restraint of vibration of an engine. In other words, with regard to the technology of the adaptive filter which has been known in the background art by being described in Nonpatent Reference 1 or the like, it has not been conceived at all to promote an accuracy of detecting a rotation speed by utilizing the encoder regardless of a whirling movement of the encoder. Further, it has not been particularly taken into consideration to promote the accuracy of detecting the rotation speed by other type of filter.
In view of the above-described situation, it is an object of the invention to provide a rotation speed detecting apparatus and a load measuring apparatus of a rolling bearing unit capable of being constituted at low cost, without posing a problem in durability or an installing space, and capable of measuring a rotation speed of a rotating member while ensuring an accuracy needed for a control.
Preferably, it is an object thereof to provide a rotation speed detecting apparatus capable of measuring a rotation speed of the rotating member without producing a delay over time by being applied to a field of detecting a rotation speed which is quite different from an acoustic field or the like applied in the background art.
The object of the invention is achieved by the following constitution.
A rotation speed detecting apparatus of the invention comprises an encoder fixedly supported by a rotating member, rotated along with the rotating member and changing a characteristic thereof alternately in a circumferential direction, a rotation detecting sensor provided with a detecting portion thereof in a state of being opposed to a detected face of the encoder, and a calculating unit for calculating a rotation speed of the rotating member based on a detecting signal transmitted from the rotation detecting sensor and changed periodically.
Particularly, according to the rotation speed detecting apparatus of the invention, the calculating unit includes a filter circuit for removing an influence of a variation of the detecting signal of the rotation detecting sensor constituting an error in calculating the rotation speed of the rotating member {for example, owing to an incoincidence between a rotational center and a geometrical center of the encoder}.
As the filter circuit, preferably, an adaptive filter is used.
According to the rotation speed detecting apparatus of the invention constituted as described above, for example, even when the rotational center of the rotating member and the geometrical center of the encoder do not coincide with each other, the rotation speed of the rotating member can accurately be calculated. That is, even when the centers do not coincide with each other and the variation based on the incoincidence is produced in the detecting signal of the rotation detecting sensor, the variation can be canceled. Therefore, by accurately grasping various states based on the rotation speed of the rotating member, a swift and proper measure can be carried out.
Particularly, when the adaptive filter is used as the filter circuit, a delay of a signal processing in accordance with canceling the variation can be nullified and various controls utilizing the rotation speed can swiftly be carried out.
When the invention is embodied, preferably, an error component in the detecting signal constituting an object to be removed as the influence of the variation by the filter circuit is constituted by a rotation primary component of the encoder.
A variation width of the rotation primary component of the encoder is liable to be increased in comparison with other component and therefore, by canceling the variation of the component by the filter circuit, the accuracy of detecting the rotation speed of the rotating member can effectively be increased.
Further, as the filter circuit, other than the adaptive filter, one or a plurality of kinds of digital filters or analog filters, or a low pass filter, or a notch filter can respectively be used.
In a case of using the filter circuits, when a filter of a degree number fixed type for changing a cut off frequency in accordance with a rotation speed of the rotating member is used, the detecting signal can effectively be processed by the filter circuit even in a use in which the rotation speed of the rotating member is changed.
Further, when the invention using the adaptive filter and the filter circuit is embodied, preferably, a tap number of the adaptive filter is made to be equal to a number of pulses per one rotation of the encoder.
Further, as the adaptive filter, it is preferable to use the adaptive filter operated by a synchronizing type LMS algorism.
When constituted in this way, a number of times of calculation processings necessary for the detecting signal of the rotation detecting sensor at each pulse of the encoder is considerably reduced to be sufficiently able to process by a calculating unit (CPU) of low cost in which a calculation speed thereof is not particularly fast.
Further, preferably, an average value of a filter coefficient of the adaptive filter is calculated and a DC level of the detecting signal of the rotation detecting sensor is corrected based on the average value.
In this case, as the average value of the filter coefficients, it is preferable to use an average value of filter coefficients sampled at arbitrary two points present at equal intervals in the direction of rotating the encoder (positions on sides opposed each other by 180 degrees), or an average value of 4 points or more of filter coefficients constituting a plurality of combined data constituted by a combination of pairs of filter coefficients sampled at arbitrary two points present at equal intervals in the direction of rotating the encoder.
When constituted in this way, even in a case in which the adaptive filter operated by a synchronizing type LMS algorism is used, the adaptive filter prevents the DC level of the detecting signal of the rotation detecting sensor from being canceled and various states based on the rotation speed of the rotating member can accurately be grasped.
Further, preferably, the adaptive filter is arranged in parallel with a main signal path (main route) for transmitting the detecting signal of the rotation detecting sensor. Further, along therewith, an error component constituting an amount of the variation of the rotation detecting sensor calculated by the adaptive filter is subtracted at a downstream portion of the main signal path. Further, by such a constitution, an influence of the variation of the detecting signal of the rotation detecting sensor is removed.
When the adaptive filter is arranged in parallel with the main signal path in this way, by a constitution different from a constitution in which a filter is arranged (inserted) in series with the main signal path and a characteristic of the filter is made to be variable by some method which is generally used in a background art, the influence of the variation of the detecting signal of the rotation detecting sensor can easily and sufficiently be removed. Further, although in a case of a filter of a notch filter or the like inserted in series therewith, there is a possibility of producing a delay over time in a main signal, by arranging the filter in parallel therewith, there is not a concern of producing a delay over time in the main signal.
Further, preferably, as the adaptive filter, a digital filter or an analog filter operated by a steepest descent method is used. Further, further preferably, as the adaptive filter, a digital filter or an analog filter operated by an LMS (least squares mean) algorism (an operation rule minimizing a squares mean error based on the steepest descent method).
When the adaptive filter operated by the steepest descent method (further preferably, LMS algorism) is used, the adaptive filter can be finished in a state of minimizing the variation based on the incoincidence between the rotational center of the rotating member and the geometrical center of the encoder. Therefore, the error based on the variation can easily and sufficiently be reduced.
Further, preferably, the reference signal constituting the input of the adaptive filter (signal related to the variation of the output signal of the rotation detecting sensor based on whirling) is generated by itself by a processing circuit of the detecting signal of the rotation detecting sensor opposed to the encoder in a which a number of times of changes in the characteristic in one rotation has been known, or a processing circuit for calculating the rotation speed of the rotating member based on the detecting signal.
When constituted in this way, the reference signal can be generated at low cost and by saving a space. That is, in a case of an active noise control which is generally known as a use of an adaptive filter in a background art, a frequency and a waveform of external noise to be reduced are not necessarily known. Therefore, it is necessary to generate a reference signal for producing sound for canceling the external noise (sound having a magnitude the same as that of the external noise and a phase of a waveform deviated therefrom by 180 degrees) based on external noise collected by a microphone provided separately (producing the reference signal by a signal inputted from outside). In contrast thereto, in the case of the invention of using the adaptive filter as the filter circuit, by the adaptive filter, the variation signal of the detecting signal of the rotation detecting sensor based on whirling of the encoder is reduced. Further, the number of times of changes in the characteristic in one rotation of the encoder is previously known and therefore, by observing the number of pulses of one rotation of the encoder, without particularly providing a sensor for measuring whirling, the reference signal related to the variation can be generated. The reference signal can be constituted by any waveform of a sine wave, a triangular waveform, a sawtooth wave, a rectangular wave, a pulse wave constituting one period by one rotation of the encoder.
Further, preferably, with regard to the variation of the detecting signal of the rotation detecting sensor, a low pass filter is provided frontward or rearward of the adaptive filter for averaging a second variation based on a cause different from that of the variation based on whirling of the encoder (so-to-speak accumulated pitch error) constituting the variation the influence of which is removed by the adaptive filter and having a period shorter than that of the variation (first variation) based on the whirling.
In the variation of the detecting signal of the rotation detecting sensor in accordance with rotation of the encoder, other than-the variation (first variation) (of a low frequency wave) having a comparatively long period based on the whirling, there is a variation (second variation) (of a high frequency wave) having a comparatively short period by a pitch error of a change in the characteristic in the circumferential direction. It is difficult to reduce the variation of the high frequency wave by the adaptive filter. However, the variation of the high frequency wave can be corrected by a low pass filter of an averaging filter or the like for executing an averaging processing of a moving average or the like. Therefore, when the low pass filter of the average filter or the like is provided frontward or rearward of the adaptive filter as described above, not only the variation (first variation) of the detecting signal of the rotation detecting sensor based on whirling of the encoder referred to as so-to-speak accumulated pitch error but also the variation (second variation) of the detecting signal of the rotation detecting sensor based on the pitch error of the change in the characteristic of the encoder can be reduced.
Further, when the invention using the adaptive filter as the filter circuit is embodied, preferably, the rotating member fixedly supporting the encoder is constituted by a retainer provided between a pair of bearing rings constituting a rolling bearing unit and rotated in accordance with revolution of rolling members retained in a plurality of pockets.
The variation of the detecting signal of the rotation speed detecting sensor based on whirling the encoder is produced owing to the incoincidence between the rotational center and the geometrical center of the encoder. Further, the incoincidence is produced also by an integrating error or the like. However, the incoincidence between the two centers based on the integrating error can be restrained to a practically unproblematic degree by increasing the integrating accuracy.
However, in a case in which the encoder is supported by the retainer, even when geometrical centers of the retainer and the encoder are made to coincide with each other completely, the incoincidence between the rotational center and the geometrical center of the encoder is produced. The reason is that as described above, a clearance is present between a rolling face of each rolling member and an inner face of the pocket of the retainer.
Therefore, when the rotation speed of the retainer is measured by utilizing the encoder fixedly supported by the retainer, it is important to deal with the variation of the detecting signal of the rotation detecting sensor based on the incoincidence between the rotational center and the geometrical center of the encoder.
Further, particularly, when the detected face is one side face in an axial direction of the encoder, it is important to embody the invention of using the adaptive filter as the filter circuit.
In a case in which the encoder is fixedly supported by a portion of the retainer and in a case in which the geometrical center and the rotational center of the encoder do not coincide with each other, even when the detected face of the encoder is constituted by any face (regardless of a peripheral face, the one side face in the axial direction), the detecting signal of the rotation detecting sensor is varied based on the incoincidence. However, when the encoder and a detecting portion of the rotation speed detecting apparatus are arranged at inside of a limited space in the rolling bearing, a degree of freedom of design is increased by constituting the detected face by the one side face in the axial direction of the encoder.
Further, the embodiment of the invention using the adaptive filter as the filter circuit, preferably, a load measuring apparatus of a rolling bearing unit is conceivable.
The load measuring apparatus of the rolling bearing unit is provided with a stationary ring, a rotating ring, a plurality of rolling members, a pair of rotation speed detecting apparatus and a calculating unit.
The stationary ring among is not rotated even in being used.
Further, the rotating ring is arranged concentrically with the stationary ring and is rotated in being used.
Further, the respective rolling members are rollably provided between stationary side tracks and rotating side tracks formed by respective two rows at portions of the stationary ring and the rotating ring opposed to each other by respective pluralities of pieces thereof and making directions of contact angles inverse to each other between the two rows.
Further, the rotation speed detecting apparatus is for detecting the rotation speed of the pair of retainers retaining the two rows of the rolling members.
Further, the calculating unit calculates the load between the stationary ring and the rotating ring based on the rotation speeds of the pair of retainers detected by the respective rotation speed detecting apparatus.
When the invention of using the adaptive filter as the filter circuit is applied to the load measuring apparatus of the rolling bearing unit, the above-described respective rotation speed detecting apparatus are constituted by the above-described structure.
Further preferably, the rotating ring is constituted by a hub rotated along with a wheel in a state of being fixed to the wheel of an automobile.
Further, in notations in the drawings, notations 1, 1a designate outer rings, notations 2, 2a designate hubs, notations 3, 3a designate rotating side flanges, numeral 4 designates a hub main body, numeral 5 designates a nut, numeral 6 designates an inner ring, numeral 7 designates an outer ring track, numeral 8 designates an inner ring track, notations 9a, 9b designate rolling members, notations 10, 10a designate attaching holes, numeral 11 designates a displacement sensor, numeral 12 designates a sensor ring, numeral 13 designates a sensor rotor, numeral 14 designates a cover, notations 15, 15a designate rotation speed detecting sensors, numeral 16 designates a knuckle, numeral 17 designates a fixed side flange, numeral 18 designates a bolt, numeral 19 designates a screw hole, numeral 20 designates a load sensor, notations 21a, 21b designate retainers, numeral 22 designates a sensor unit, numeral 23 designates a front end portion, notations 24a, 24b designate revolution speed detecting sensors, numeral 25 designates rim portion, notations 26a, 26b designate revolution speed detecting encoders, numeral 27 designates a rotation speed detecting encoder, numeral 28 designates an adaptive filter.
A detailed explanation will be given of a rotation speed detecting apparatus and a load measuring apparatus of a rolling bearing unit according to respective embodiments of the invention in reference to the drawings as follows.
The outer peripheral face of the hub 2 constituting the rotating ring is formed with the inner ring tracks 8, 8 of the double row angular type respectively constituting rotating side tracks. Further, the inner peripheral face of the outer ring 1 constituting the stationary ring is formed with the outer ring tracks 7, 7 of the double row angular type respectively constituting the stationary side tracks. Further, the hub 2 is rotatably supported by the inner diameter side of the outer ring 1 by rollably supporting the rolling members (balls) 9a, 9b between the inner tracks 8, 8 of the double row angular type of the hub 2 and the outer ring tracks 7, 7 of the double angular type of the outer ring 1 in a state of dividing the respective rolling members (balls) 9a, 9b in double rows (2 rows) and maintaining the rolling members by the retainers 21a, 21b by respective pluralities of pieces thereof at respective rows. Under the state, the rolling members 9a, 9b of the respective rows are provided with contact angles αa, βb (
The attaching hole 10a is formed in the state of penetrating the outer ring 1 at the middle portion in the axial direction of the outer ring 1 and at the portion between the double rows of outer ring tracks 7, 7 constituting the above-described rolling bearing unit. Further, the sensor unit 22 is inserted to the attaching hole 10a from an outer side to an inner side in the diameter direction of the outer ring 1 and the front end portion 23 of the sensor unit 22 is projected from an inner peripheral face of the outer ring 1. The front end portion 23 is provided with a pair of the revolution speed detecting sensors 24a, 24b respectively constituting the rotation detecting sensors and a single piece of the rotation speed detecting sensor 15a.
The respective rotating speed detecting sensors 24a, 24b thereamong are for measuring revolution speeds of the rolling members 9a, 9b arranged in the double rows. The respective revolution speed detecting sensors 24a, 24b arrange respective detecting faces thereof at two side face in the axial direction of the hub 2 (left and right direction of
For that purpose, the detecting faces of the respective revolution speed detecting sensors 24a, 24b are made to be proximately opposed to faces opposed to each other constituting detected faces of the respective revolution speed detecting encoders 26a, 26b. Further, it is preferable to constitute distances between the detected faces of the respective revolution speed detecting encoders 26a, 26b and the detecting faces of the revolution speed detecting sensors 24a, 24b (detection clearance) to be larger a pocket clearance constituting a clearance between inner faces of pockets of the respective retainers 21a, 21b and rolling faces of the respective rolling members 9a, 9b and equal to or smaller than 2 mm. When the detection clearance becomes equal to or smaller than the pocket clearance, in a case in which the respective retainers 21a, 21b are displaced by an amount of the pocket clearance, there is brought about a possibility of rubbing the detected face and the detecting face and therefore, the case is not preferable. On the contrary, when the detection clearance exceeds 2 mm, it is difficult to accurately measure rotation of the respective revolting speed detecting encoders 26a, 26b by the respective revolution speed detecting sensors 24a, 24b.
On the other hand, the rotation speed detecting sensor 15a is for measuring the rotation speed of the hub 2 constituting the rotating ring and a detecting face thereof is arranged on a front end face of the front end portion 23, that is, an inner end face thereof in the diameter direction of the outer ring 1. Further, the rotation speed encoder 27 in a cylindrical shape is outwardly fitted to be fixed at a middle portion of the hub 2 and between the inner tracks 8, 8 of the double rows. A detecting face of the rotation speed detecting sensor 15a is made to be opposed to an outer peripheral face of the rotation speed detecting encoder 27 constituting a detected face. A characteristic of the detected face of the rotation speed detecting encoder 27 is changed alternately and at equal intervals in a circumferential direction to make the rotation speed of the hub 2 detectable by the rotation speed detecting sensor 15a. Also a measurement clearance between an outer peripheral face of the rotation speed detecting encoder 27 and the detecting face of the rotation speed detecting sensor 15a is restrained to be equal to or smaller than 2 mm.
Further, as the respective encoders 26a, 26b, 27 having various structures utilized for detecting a rotation speed of a wheel can be used in order to provide a signal for controlling ABS or TCS in the background art. For example, the respective encoders 26a, 26b, 27 made of multiple magnets arranging N poles and S poles alternately and at equal intervals at the detected face (side face or outer peripheral face) can preferably be used. Incidentally, an encoder simply made of a magnetic material or an encoder in which an optical characteristic is changed alternately and at equal intervals over a circumferential direction (by combining with a rotation speed detecting sensor of an optical type) can be used.
In the case of the embodiment, as the respective revolution speed detecting encoders 26a, 26b, there is used a permanent magnet in a shape of a circular ring in which side faces thereof constituting a detected face is arranged with S poles and N poles alternately and at equal intervals. The respective revolution speed detecting encoders 25a, 26b are coupled to fix to side faces of the rim portions 25, 25 of the respective retainers 21a, 21b produced separately by adhering or molded by insert molding by setting the revolution speed detecting encoders 26a, 26b at inside of a cavity when the respective retainers 21a, 21b are molded by injection molding. It is selected which of the methods is adopted in accordance with cost a required coupling strength or the like.
Further, rotation detecting sensors for a magnetic type can preferably be used as the respective revolution speed detecting sensors 24a, 24b and the rotation speed detecting sensor 14a all of which are sensors for detecting the rotation speed. Further, as the rotation detecting sensor of the magnetic type, an active type integrated with a magnetic detecting element of a hole element, a hole IC, a magentoresistive element (MR element, GMR element), MI element or the like can preferably be used. In order to constitute the rotation detecting sensor of the active type integrated with the magnetic detecting element, for example, one side face of the magnetic detecting element is butted to one end face in a magnetizing direction of a permanent magnet directly or by way of a stator made of a magnetic material (when an encoder made of a magnetic material is used), and other side face of the magnetic detecting element is made to be opposed to the detected faces of the respective encoders 26a, 26b, 27 directly or by way of a stator made of a magnetic material. Further, in the case of the embodiment, the encoder made of the permanent magnet on the side of the sensor is not needed.
In the case of the load measuring apparatus of the rolling bearing unit according to the invention, detecting signals of the respective sensors 24a, 24b, 15a are inputted to a calculating unit, not illustrated. Further, the calculating unit calculates one or both the radial load and the axial load applied between the outer ring 1 and the hub 2 based on the detecting signals transmitted from the respective sensors 24a, 24b, 15a. For example, when the radial load is calculated, the calculating unit calculates a sum of revolution speeds of the rolling members 9a, 9b of the respective loads detected by the revolution speed detecting sensors 24a, 24b. Further, the calculating unit calculates the radial load based on a ratio of the sum to the rotation speed of the hub 2 detected by the rotation speed detecting sensor 15a. Further, when the axial load is calculated, the calculating unit calculates a difference between the revolution speeds of the rolling members 9a, 9b of the respective rows detected by the revolution speed detecting sensors 24a, 24b. Further, the calculating unit calculates the axial load based on a ratio of the difference to the rotation speed of the hub 2 detected by the rotation speed detecting sensor 15a. An explanation will be given in this respect in reference to
nc={1−(d·cos α/D)·(ni/2)}+{1+(d·cos α/D)·(no/2)} (1)
As is apparent from Equation (1), the revolution speed nc of the respective rolling members 9a, 9b is changed in accordance with a change in the contact angle α (αa, αb) of the respective rolling members 9a, 9b, and as described above, the contact angles αa, αb are changed in accordance with the radial load Fr and the axial load Fa. Therefore, the revolution speed nc is changed in accordance with the radial load Fr and the axial load Fa. Specifically, with regard to the radial load Fr, the larger the radial load Fr, the slower the revolution speed nc since the hub 2 is rotated and the outer ring 1 is not rotated. Further, with regard to the axial load, the revolution speed of the row for supporting the axial load becomes fast and the revolution speed of the row which does not support the axial load becomes slow. Therefore, the radial load Fr and the axial load Fa are calculated based on the revolution speed nc.
However, the contact angle α related to a change of the revolution speed nc is changed not only by the radial load Fr and the axial load Fa while being related to each other but also changed by the preloads F0, F0. Further, the revolution speed nc is changed in proportion to the rotation speed ni of the hub 2. Therefore, the revolution speed nc cannot accurately be calculated unless all of the axial load Fa, the preloads F0, F0, the rotation speed ni are relatedly taken into consideration. The preloads F0, F0 thereamong are not changed in accordance with an operating state and therefore, it is easy to exclude the influence by initial setting or the like. In contrast thereto, the radial load Fr, the axial load Fa, the rotation speed ni of the hub 2 are always changed in accordance with the operating state and therefore, the influence cannot be excluded by the initial setting or the like.
In view of the above-descried situation, in the case of the embodiment, as described above, when the radial load is calculated, by calculating the sum of the revolution speeds of the rolling members 9a, 9b of the respective rows detected by the respective revolution speed detecting sensors 24a, 24b, the influence of the axial load Fa is reduced. Further, when the axial load is calculated, by calculating the difference between the revolution speeds of the rolling members 9a, 9b of the respective rows, the influence of the radial load Fr is reduced. Further, in any of the cases, by calculating the radial load Fr or the axial load Fa based on the ratio of the sum or the difference to the rotation speed ni of the hub 2 detected by the rotation speed detecting sensor 15a, the influence of the rotation speed ni of the hub 2 is excluded. However, when the axial load Fa is calculated based on the ratio of the rotation speeds of the rolling members 9a, 9b of the respective loads, the rotation speed of the hub 2 is not necessarily needed.
Further, although there are variously present methods for calculating one or both of load(s) of the radial load and the axial load based on signals of the respective revolution speed detecting sensors 24a, 24b, the methods are not related to the gist of the invention and therefore, a detailed explanation thereof will be omitted.
However, in calculating any load by any method, it is important to accurately calculate the revolution speeds of the rolling members 9a, 9b of the respective rows based on the detecting signals of the respective revolution speed detecting sensors 24a, 24b for increasing the accuracy of measuring the load.
In contrast thereto, the detecting signals of the respective revolution speed detecting sensors 24a, 24b (signals indicating the revolution speeds based thereon) include a variation having a comparatively high frequency as described above based on an error of a magnetizing pitch (pitch between an S pole and an N pole contiguous to each other in the circumferential direction) of the detected face and a variation having a comparatively low frequency as described above in accordance with whirling movements of the retainers 21a, 21b. When the variations are not processed (reduced), the revolution speeds of the rolling members 9a, 9b of the respective rows cannot accurately be calculated and therefore, accuracies of measuring the radial load or the axial load are deteriorated. Hence, in the case of the embodiment, by the adaptive filter as shown by
First, an explanation will be given of reason of producing two kinds of the variations in reference to
When the detecting face of the revolution speed detecting sensor 24a (24b) is made to be opposed to the side face of the revolution speed detecting encoder 26a (26b), the detecting signal (signal indicating the revolution speed based thereon) of the revolution speed detecting sensor 24a (24b) is sinusoidally changed as shown by a chain line α of
Further, although the pitch between the S pole and the N pole aligned at a side face of the revolution speed detecting encoder 26a (26b) is to stay to be the same inherently, owing to a magnetizing error or the like brought about in fabrication, there is case in which the pitches differ from each other although a difference therebetween is small. Further, based on the error, the detecting signal of the revolution speed detecting sensor 24a (24b) is varied. A period of the variation based on the error of the magnetizing pitch is far shorter than a period of the variation based on the whirling movement. For example, when a characteristic (a repetition of S pole and N pole) of the side face (detected face) of the revolution speed detecting encoder 26a (26b) is changed by 60 times over the total periphery of the detected face, the period of the variation based on the error of the magnetizing pitch becomes about 1/60 of the period of variation based on the whirling movement.
The detecting signal (signal indicating the revolution speed based thereon) outputted from the revolution speed detecting encoder 26a (26b) is as shown by a bold line β in
First, an explanation will be given of a reduction in the variation having the low frequency by the adaptive filter shown in
On the other hand, in order to operate the adaptive filter 28, a reference signal x having a correlation with the variation amount dn on whirling becomes necessary. When the reference signal x can be obtained, the adaptive filter 28 forms an FIR (finite impulse response) filter (filter having finite impulse response time=filter nullifying impulse response in finite time) having a characteristic the same as a transfer characteristic of an actual signal flow “dn→d” by self learning. Further, when a cancel signal y{=y(k), mentioned later} provided as a result of calculation by the adaptive filter 28 is subtracted from the output signal d of the revolution speed detecting sensor 24a (24b), the subtraction becomes equivalent to that the variation amount dn by the whirling is removed from the output signal d of the revolution speed detecting sensor 24a (24b) (d−dn). When the variation amount dn is removed in this way, the adaptive filter 28 does not filter the output signal d transmitted on a main route of signal (upper half portion of
In the case of the embodiment, the reference signal x is generated by itself by a circuit of processing the output signal of the revolution speed detecting sensor 24a (24b) opposed to the revolution speed detecting encoder 26a (26b) based on a number of times of a change in the characteristic during one rotation of the revolution speed detecting encoder 26a (26b), or a processing circuit for calculating the revolution speed of the respective rolling members 9a (9b) based on the detecting signal. Therefore, cost required for generating the reference signal x can be reduced. That is, when a structure of an active noise control which has been known as a use of the adaptive filter in the background art is applied to a structure for accurately calculating the revolution speed of the respective rolling member 9a (9b) as it is, whirling of the revolution speed detecting encoder 26a (26b) is detected by a separately provided sensor of a displacement sensor, a rotation speed sensor or the like, and a detecting signal of the sensor is used as the reference signal x of the adaptive filter 28. Naturally, the invention can also be embodied by such a structure, cost and an installing space are needed by an amount of providing a sensor separately.
In contrast thereto, in the case of the embodiment, the reference signal x is obtained without using the detecting signal of the sensor provided separately in this way, by the adaptive filter 28, the variation amount dn of the output signal d of the revolution speed detecting sensor 24a (24b) based on whirling of the revolution speed detecting encoder 26a (26b) is reduced. That is, the number of times (number of S poles and N poles) of changing the characteristic during one rotation of the revolution speed detecting encoder 26a (26b) is previously known. Therefore, by observing a number of pulses by one rotation of the revolution speed detecting encoder 26a (26b), without particularly providing separately a sensor of a displacement sensor, a rotation speed sensor or the like, the reference signal x having the correlation with the variation dn can be generated. Specifically, the influence of whirling of the revolution speed detecting encoder 26a (26b) is constituted by a waveform having a main component by a primary component of rotation thereof, for example, when the revolution speed detecting encoder 26a (26b) is for 60 pulses per rotation, the waveform can be generated by itself as a sine wave, a triangular wave, a sawtooth wave, a rectangular wave, a pulse wave or the like constituting one period by 60 data.
The waveform of the reference signal x can also be generated by a processing circuit (CPU) for calculating the revolution speed of the respective rolling members 9a (9b) and can also be generated by an electronic circuit portion (IC) attached to the revolution speed detecting sensor 24a (24b). At any rate, the cancel signal y calculated based on the obtained reference signal x is subtracted from the output signal d of the revolution speed detecting sensor 24a (24b) to calculate a modified signal e{=e (k), mentioned later} indicating the actual rotation speed dd. The modified signal e calculated in this way is transmitted to a processing circuit for calculating the revolution speed of the respective rolling members 9a (9b) to be utilized for calculating the revolution speed, further, utilized also as information learnt by the adaptive filter 28 by itself.
Further, a processing for providing the modified signal e by calculating the cancel signal y at the portion of the adaptive filter 28 and subtracting the cancel signal y from the output signal d of the revolution speed detecting sensor 24a (24b) is executed based on Equations (2) through (4) shown below.
In Equations (2) (3) (4), notation k designates a data number in time sequence, notation N designates a tap number of an FIR filter as the adaptive filter 28. Further, notation w designates a filter coefficient of the FIR filter, notation wk designates a filter coefficient used when k-th data is processed, notation wk+1 designates a filter used when a successive data sequence ((k+1) -th) is processed, respectively. That is, in the case of the embodiment, the FIR filter becomes an adaptive filter successively properly updating the filter coefficient by Equation (4). Null may substitute for the filter coefficient Wk used first in starting the operation since the filter coefficient is adapted by itself when the operation is started, however, a desired filter characteristic may previously be calculated and a value thereof may substitute therefor. Further, a filter coefficient finally used in a preceding processing may be stored to storing means of EEPROM or the like and may be used when restarting.
Further, notation μ in Equation (4) is a value for determining an updating amount when the filter coefficient is made to be proper by itself referred to as step parameter, and normally becomes a value of about 0.01 through 0.001, however, actually, the value can be set by investigating acceptability of adaptive operation beforehand, or may successively be updated by using Equation (5) shown below.
Further, also α in Equation (5) becomes a parameter for determining an updating amount for making the filter coefficient proper by itself, α may fall in a range of 0<α<1 and α is set more easily than μ, mentioned above. Further, in the case of the embodiment, the reference signal x is generated by itself and therefore, a value of a denominator in Equation (5) is known and an optimum value of μ can also be calculated beforehand. From a view point of reducing a calculation amount, it is preferable that μ is previously calculated by Equation (5) and the filter coefficient is made to be proper by itself by Equation (4) by constituting a constant by μ.
As described above, a modified signal e indicating the actual rotation speed dd is calculated by subtracting the cancel signal y calculated by the adaptive filter 28 from the output signal d of the revolution speed detecting sensor 24a (24b). Further, based on the modified signal e calculated in this way, the revolution speed of the respective rolling members 9a (9b) can accurately be calculated. Further, in an actual case, in the output signal d of the revolution speed detecting sensor 24a (24b), there is present a second variation based on the pitch error having a period shorter than that of the variation based on whirling of the revolution speed detecting sensor 24a (24b). Hence, by providing a low pass filter of an averaging filter or the like for averaging the second variation before or after the adaptive filter 28, despite the second variation, the revolution speed of the respective rolling members 9a (9b) is made to be calculated accurately. Structure and operation of a low pass filter of an average filter or the like for restraining the variation having the high frequency is well known in the background art and therefore, a detailed explanation thereof will be omitted.
In contrast thereto, a broken line b shows a result of correcting data after moving average indicated by the bold line a by using the adaptive filter (in correspondence with the modified signal e). As is apparent from the broken line b, although the calculated value is varied immediately after starting the adaptive filter, the filter coefficient is adapted by itself after an elapse of a short period of time and a calculation result is converged into a constant value of substantially 100 min−1. Therefrom, it is known that by using both of the average filter and the adaptive filter, even when the encoder having the pitch error and a large deviation between the rotational center and the geometrical center (whirled) is used, the rotation speed of the rotating member can accurately be calculated.
Further, in calculating two pieces of the lines a, b shown in
Although all of the above-described respective equations (2) (3) (4) are simple equations utilized for making the adaptive filter proper in the above-described first embodiment, in the actual application, there is conceivable a case in which a calculation amount becomes problematic. For example, when the tap number of the adaptive filter is set as N=60, a total of 241 times of operation of 60 times of multiplication in Equation (2), one time of subtraction in Equation (3), 180 times of 120 times of multiplication and 60 times of addition in Equation (4) need to be executed at each pulse of the encoder. Therefore, a calculation amount necessary for calculating a revolution speed of double rows of rolling members provided at a single piece of rolling bearing unit becomes 482 times/1 pulse. Although the calculation amount (number of times of operations) is not unable to be processed physically, it is necessary to use comparatively expensive CPU having a fast processing speed. For example, when rotation speeds of wheels for an automobile (4 pieces of wheels) are detected for controlling a vehicular running stabilizing apparatus of ABS, TCS, VSC or the like, it is necessary to use 4 pieces of expensive CPU (or, high speed CPU capable of executing four operations of 241 times×2×4=1928 times) to cause to increase cost of the vehicular running stabilizing apparatus and therefore, the constitution is not preferable.
In view of such a situation, in the case of the embodiment, it is intended to enable to use low cost CPU by considerably reducing a calculation amount by using the synchronizing LMS algorism. However, when the adaptive filter is operated by the synchronizing type LMS algorism, the adaptive filter cancels not only a whirling component of the encoder but also a DC component indicating the rotation speed when constituted in this way. A phenomenon of canceling the DC component in this way is significant when the synchronizing LMS algorism is used. Hence, in the case of the embodiment, the DC level indicating the rotation speed is made to be able to be detected accurately by providing a function of nullifying the output value of the adaptive filter.
First, a principle of operating the synchronizing type LMS algorism will be explained. In the block diagram shown in
In Equation (6), a position j at which the reference signal x becomes the impulse of a value of 1 is shifted one by one to the right side in accordance with progress of the time sequence k and when the position j is shifted to “N−1”-th on the rightmost side, in a next time sequence, a new impulse value is expressed at 0-th on the leftmost side. That is, the reference signal x becomes a data sequence for only circulating the position of the impulse having the value 1 from 0-th to (N−1)-th. When Equation (6) is applied to Equations (2) (4), Equations (7) (8) are provided as follows.
Whereas when the adaptive filter 28 is operated by the normal LMS algorism which is not a synchronizing type, as described above, it is necessary to repeatedly execute calculations shown in respective Equations (2) (3) (4), when the adaptive filter is operated by the synchronizing type LMS algorism, only calculations shown in Equations (7) (8) and Equation (3) may be executed. For example, in a case in which the tap number N of the adaptive filter 28 is set to 60, when the adaptive filter 28 is operated by the normal LMS algorism, a total of a number of times of operation for each pitch of the encoder becomes 214 times as described above. In contrast thereto, when the adaptive filter 28 is operated by the synchronizing type LMS algorism, the operation is executed only substituting data in Equation (7), and four operations of a total of 3 times of one time of subtraction in Equation (3), 2 times of 1 time of multiplication and 1 time of addition in Equation (8) maybe executed for each pulse of the encoder. That is, by adopting the synchronizing type LMS algorism, in comparison with a case in which the synchronizing type LMS algorism is not adopted, a number of times of operation can be reduced by about 1/80.
However, when the synchronizing type LMS algorism is adopted for operating the adaptive filter 28, in order to prevent also the DC component constituting the signal expressing the rotation speed from being canceled, it is necessary to correct the null point of the adaptive filter 28. The null point correction will be explained as follows.
When measured date with regard to the rotation speed as shown by the bold line a in
When the synchronizing type LMS algorism for operating the adaptive filter is adopted and a measure is not carried out particularly, as shown by the chain line c, not only the variation component based on whirling of the encoder but also the DC component indicating the rotation speed are canceled to nullify the output value. This is a phenomenon in which a filter coefficient W of the adaptive filter 28 carries away a DC level by the adaptive operation, as a result, the output signal y of the adaptive filter 28 carries away the DC level. In order to resolve the problem, in the case of the embodiment, as shown by
Next, an explanation will be given of a method of calculating the DC level from the average value of the filter coefficient W. A filter coefficient of the adaptive filter 28 is varied as shown by
Meanwhile, the object to be canceled as the error, that is, a waviness based on whirling the encoder is constituted by a rotation n degree component constituted mainly by a rotation primary component. Further, in the case of the embodiment, the tap number N of the adaptive filter is made to be equal to a number of pulses per one rotation of the encoder and therefore, the filter coefficient W becomes a periodic function having a period of L (=60). An average value of two arbitrary points set with an interval of N/2 (=30) therebetween becomes equivalent to an average value of total points of N (=60). Hence, when the average of the two points is calculated to constitute the DC level indicating the rotation speed, also the number of times of operations can considerably be reduced, which is advantageous in view of low cost formation of CPU. When a concern remains in reliability by the average of only the two points, other than the two points, two arbitrary points set with an interval of N/2 (=30) therebetween are selected and an average value of a total of 4 points is calculated. Further, although not illustrated, even when the filter coefficient W is a periodic function of rotation n degree, the above-described average value can similar be calculated by pertinently increasing a number of points for calculating an average point and pertinently setting intervals thereof.
For example, when a digital low pass filter is used, a sampling frequency of filter calculation is set not to a fixed frequency but to a sampling frequency in accordance with the rotation speed of the encoder. When the sampling frequency is set in this way, the cut off frequency can be changed in accordance with (in proportion to) the rotation speed of the encoder. Specifically, data for the filter calculation may be sampled at each time of outputting a pulse signal from a sensor opposed to the encoder.
Y′(k)=a0·X(k)+a1·Y′(k−1)+a2·Y′(k−2) (9)
Y(k)=b0·Y′(k)+b1·Y′(k−1)+b2·Y′(k−2) (10)
In
When the output signal of the sensor changed in accordance with rotation of the encoder is processed by a low pass filter functioning as in the flowchart shown in
Y′(k)=X(k)−α·Y′(k−N/A) (11)
Y(k)={(1+α)/2}·{Y′(k)+Y′(k−N/A)} (12)
In
Further, term Y′ (k−N/A) signifies Y′ calculated by N/A pieces past of current time (processing number k). In order to calculate Y′ (k) at current time point in Equation (11), a value of Y′ (k−N/A) multiplied by α is subtracted from the input X (k). Past Y′ (k−N/A) is stored to a memory or the like integrated in a processing circuit constituting the notch filter. Although in an initial state of calculation, past Y′ (k−N/A) is not present, the calculation may be started by substituting null therefor, or a pertinent value may be stored to the memory previously as an initial value.
The output Y of the notch filter is calculated by using newest Y′ (k) and past Y′ (k−N/A) as shown by Equation (12). In this case, by pertinently specifying the constant A specifying a notch frequency by combining with the pulse N per one rotation of the encoder, there is constituted a notch filter of a so-to-speak frequency following type (degree number fixed type) following a frequency changed by an increase or a reduction in the rotation speed. For example, when A=2, there is constituted a notch filter for removing a rotation primary error component. Further, when the notation primary error component is restrained by the notch filter in this way, a response delay can more be reduced than in a case of using the low pass filter to be able to carry out a control for ensuring a running stability of the vehicle by detecting the slip force produced at the contact portion of the wheel and the road face in a state of abruptly changing a lane in running at high speed.
However, even in the case of the notch filter, although the response delay is smaller than that of the low pass filter, the response delay is invariably present and there is a possibility of posing a problem by the response delay. For example, a case of detecting a road face grip force at an instance of avoiding a hazard jumped out abruptly by fast steering. In order to be able to deal with even a case in which the response delay is hardly (not at all) permitted in this way, a method of correcting an error by using the adaptive filter is effective as in the first embodiment and the second embodiment. It is determined which filter is used in accordance with a case of requesting the fastest response. Also a structure of using both of a filter having a fast response and a filter having a slow response can be adopted depending on cases including the case of using both of the adaptive filter and the low pass filter as described above.
Although an explanation has been given of the invention in details and in reference to the particular embodiments, it is apparent for the skilled person that the invention can variously be changed or modified without deviating the sprit and the range of the invention.
The application is based on Japanese Patent Application (Japanese Patent Application No. 2003-320058) filed on Sept. 11, 2003, Japanese Patent Application (Japanese Patent Application No. 2003-379536) filed on Nov. 10, 2003, Japanese Patent Application (Japanese Patent Application No. 2004-126311) filed on Apr. 22, 2004, and the content is incorporated herein by the reference.
The rotation speed detecting apparatus of the invention is not limited to the load measuring apparatus of the rolling bearing unit for measuring the load applied to the rolling bearing unit for supporting the wheel of the automobile as shown in the embodiments but can be utilized for detecting rotation speeds of rotating members of various rotating machine apparatus. In this case, a member for fixedly supporting the encoder is not limited to a member having a possibility of deviating a rotational center and a geometrical center as in a retainer but may be a rotating member in which the rotational center and the geometrical center are not deviated from each other as in a rotating shaft or the like. In this case, it is not necessary to particularly promote an accuracy of attaching an encoder to the rotating member to thereby reduce cost required in integration. Further, an encoder which can be used when the invention is embodied is not limited to a so-to-speak multipole magnet encoder in which S poles and N poles are alternately arranged in a rotational direction but includes encoders having various structures for providing information of rotation speeds such as a tone wheel, a gear, a slit disk and the like. Further, also a rotation detecting sensor is not limited to that of a magnetization detecting type but those of various structures of an optical type, an eddy current type and the like can be used.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2003-320058 | Sep 2003 | JP | national |
| 2003-379536 | Nov 2003 | JP | national |
| 2004-126311 | Apr 2004 | JP | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/JP04/12700 | 8/26/2004 | WO | 3/13/2006 |
