AU2005287781A1

AU2005287781A1 - Method and arrangement for compensating errors in displacement or angle sensors

Info

Publication number: AU2005287781A1
Application number: AU2005287781A
Authority: AU
Inventors: Oliver Glitza; Cord Hoersten; Matthias Moerbe; Ralf Noltemeyer; Klaus Sekot; Axel Wenzler
Original assignee: Bishop Innovation Pty Ltd; Robert Bosch GmbH
Current assignee: Bishop Innovation Pty Ltd; Robert Bosch GmbH
Priority date: 2004-09-22
Filing date: 2005-09-15
Publication date: 2006-03-30
Anticipated expiration: 2025-09-15
Also published as: WO2006032241A1; AU2005287781B2; EP1794548A1; DE102004045811A1

Description

-'U Lsox 1 j 335 (Mall) Aldridge & Co Ltd 14 Fairbum Grove (Courier) Johnsonville PATENT, LEGAL, & TECHNICAL TRANSLATIONS Wellington, NEW ZEALAND From:- Telephone: (64 4) 478-2955 Danish, Dutch, Esperanto, Flemish, French, German, Facsimile: (64 4) 478-2955 Italian, Norwegian, Portuguese, Spanish, Swedish... E-mail: [email protected] William R. Aldridge W A. ATCL DM Tcr.. DW FWZFA KAAT Consulting Lnguist & Translator Gillian M. Aldridge-Heine Admiistrator Thursday, 1 March 2007 My ref: CallawrieCM/Tr1691 I, WILLIAM RUPERT ALDRIDGE, MA Hons, ATCL, Dip. Tchg., FNZEA, DBEA, NAATI III, Consulting Linguist & Translator of Wellington, New Zealand, HEREBY CERTIFY that I am acquainted with the German and English languages, and am a competent translator from German to English, and I FURTHER CERTIFY that, to the best of my knowledge, ability, and belief, the attached translation, made by me, is a true and correct translation of PCT/DE2005/001616 a W02006032241 As WITNESS MY HAND AND SEAL Bridge & CO. -1MAR 207 } WIngtn, NZ- Translation from German WO 2006/032241 l'CI/I2005/001616 Method and Arrangement for Compensating Errors in Displacement or Angle Sensors Prior Art The invention relates to a method or arrangement for the 5 compensation of errors in displacement- or angle-sensors, particularly when performing angle-of-rotation detection on axles or shafts, in accordance with the generic part of the main claim. For example, in order to detect the torque acting on a motor 10 vehicle's steering shaft while the steering wheel is being turned, it is necessary to measure very small changes of angle, in both directions of turning. This can be done using incremental angle sensors that analyze an angular position on the basis of impulses which are produced optically, 25 magnetically, or otherwise, by the turning of the steering wheel, and which are then detected by suitable means. To increase measuring-accuracy and the unambiguous angle measurement range, it is normal to analyze a plurality of such incremental - and generally periodic - measurement 20 values; and this results in a number of phase measurement values, from which the parameter being measured (e.g. angle of rotation, an angle-difference, or the distance to a target), will be determined. For example, WO 00/28285 describes an optical sensor for determining the position of moving surfaces; this sensor detects an optical pattern consisting of sections of high and low reflectivity. WO 00/36377 describes a signal processing method for this purpose that is based on 2 WO 2006/032241 PCT/DI):2005/001616 assigning certain linear detector-arrays to the optically detectable code or pattern in such a way that specifically one linear detector-array is provided for each such code track consisting of sections of high and low reflectivity. 5 A method described in DE 101 42 449 Al is proposed for analyzing such phase or angle measurements, for example, in the case of more than two phase signals. With this method, using one or more linear detector-arrays, signals for determining displacements can be obtained with a 10 displacement sensor, or signals for determining angles of rotation can be obtained with an angle-sensor. This prior art method produces a single-valued phase or angle measurement, for determining the position of the code bearing object, from N multi-valued phase signals. These 15 measured phase values are transformed computationally by a linear transformation method, and processed with a given weighting. Thus, in this prior art, a method is described that produces a highly accurate, robust, single-valued phase measurement 20 from N multi-valued, and possibly disturbed, phase measurements. That method is applicable to e.g. an optical angle-sensor in which N parallel tracks are applied to a cylinder. On each of the N tracks (i=l... N) there are n, periods of an item of phase information, which is represented, e.g. in the optical case, by n, periods of light-dark transitions. Other sensor principles, e.g. magnetic or capacitive, are also possible in this regard. Also, instead of being on a cylinder, the tracks of the sensor can be applied to a plane, e.g. in the case of a w0 displacement sensor. The prior art devices or methods rely on the code being able to be detected and analyzed by the sensor at any time, from the information-bearing sections of the surface of the 3 WO 2006/03224 I PCT/D:i2005/00 1616 object whose position (displacement or angle) is to be detected. This will no longer necessarily be so if relative de-positioning of the sensor and the code occurs, e.g. due to mechanical tolerances in the underlying system. For 5 instance, with an opto-electric angle- or torque-sensor, that would be the case if the cylindrical code-disks fastened to a rotary shaft were shifted axially so far that they could no longer be mapped, by means of the lens of the opto-electric angle- or torque-sensor, onto the respective 10 linear detector arrays of the detector. Advantages of the Invention With the above-mentioned generic method for detecting the movement or angle of rotation of moving mechanical parts, it is possible to analyze code signals produced by scanning a 15 number of code-tracks on the moving part - arranged adjacent to one another and at right angles to the direction of movement - by means of respective linear detector arrays of a stationary sensor that are assigned to said code-tracks. In the present invention, the known method is advantageously 20 further developed by arranging additional linear detector arrays in such a way that, at the maximum tolerance allowed to the moving part perpendicular to the direction of movement of each code-track, at least one linear detector array of a detector will be able to be employed. 2! If the code consists e.g. of a number of tracks - e.g. four tracks - on the circumference of a code-disk of an angle sensor, then normally one linear detector-array on the stationary detector will be assigned uniquely to each track. This allows for shifting of the code mapped onto the w0 detector by L/2 in the axial direction, where L is the width of a code-track in the image region, if L is large relative to the size of the linear detector-array in the axial 4 WO 2006/032241 PCT/DE2005/001616 direction. In the prior art, a shift of further than this will mean that the sensor principle will no longer work; and one-to-one assignment of the code-track to a linear detector-array means that each code-track can only be 5 properly read, i.e. analyzed, by one specific linear detector-array. With the present invention, however, by adding further linear detector-arrays, together with suitable circuitry for them, or by adapting the signal analysis of the linear )O detector-arrays, it now becomes possible for a code-track to also be analyzed by one or more of the adjacent linear detector-arrays. Thus, with suitable circuitry for the linear detector-arrays, it is possible, advantageously, to achieve one-to-plural assignment, and to allow shifting of )5 the code, and hence the object to be measured, by more than L/2. To perform the inventive method it must be ensured, as described above, that the code-tracks are at all times readable by linear detector-arrays without requiring any a expensive mechanical arrangements, e.g. limitation of axial shifting of a rotating object to be measured, by means of expensive mechanical bearings. Another expensive solution increasing the code-track width and the resultant pitch of the linear detector-arrays - can also be avoided with the present invention. Thus, thanks to the invention, neither the code region on the object to be measured nor the detector area need be increased - both of which solutions would be undesirable and both of which would increase the cost of the sensor. 30 The inventive solution is essentially based on making the detector more tolerant to shifts in the axial direction, which is achieved by having a suitable arrangement of linear detector-arrays, and an electronic-processing method adapted 5 WO 2006/032241 P(T/D1)2005/001616 thereto, thus making an electronic solution possible - which is generally more economic than a mechanical solution or than increasing the size of the code-track and/or detector. In an advantageous form of embodiment, the pitch of the 5 linear detector-arrays is less than the width of the code tracks; and the inner linear detector-arrays, as a function of position within the tolerance allowed, can each detect one of two adjacent code-tracks; and at each extreme position of the tolerance there is an outer linear detector m array, next to a code-track. In a first variant of the proposed embodiment, the pitch of the linear detector-arrays is smaller, by a given factor, than the code-track width L, said factor being given by the number of code-tracks divided by the number of linear detector-arrays. In a second form of embodiment, the pitch of the linear detector-arrays is less than the code-track width L, by the width of the linear detector-arrays perpendicular to the direction of movement - i.e. in the axial direction, in the case of an angle sensor. 20 In another form of embodiment, it is also possible to add an auxiliary code-track at one or both of the outer code tracks; in which case, each auxiliary code-track will also be detectable by the detector's outer linear detector-array at the extreme tolerance positions of the object being measured. In a sensor arrangement according to the invention, for performing the method described above, it is advantageous to arrange the code-tracks on code-disks on a rotary shaft, in order to be able to detect the rotating shaft's angle of :w rotation and/or its torque. It is preferable that such code tracks be formed by light-and-dark markings on the code- 6 WO 2006/032241 PCT/DI:2005/00 1616 disks, and that the linear detector-arrays consist of opto electric sensor elements. Drawing Examples of embodiments of a sensor arrangement for performing the inventive process are explained with reference to the drawings, in which: Fig. 1 shows the basic arrangement of four code-tracks, and four linear detector-arrays assigned respectively thereto, in a prior-art angle sensor; 10 Fig. 2 shows the geometrical relationships at extreme positions in the arrangement shown in Fig. 1, in the case of shifts of the code-tracks and the linear detector-arrays relative to each other within the tolerance allowed; 15 Fig. 3 shows an arrangement according to the invention, with five linear detector-arrays for four code tracks; Fig. 4 shows the geometrical relationships of the respective linear detector-arrays and code-tracks in Fig. 3; Fig. 5 is another embodiment-example, with five linear detector-arrays; Fig. 6 is an embodiment-example with additional auxiliary code-tracks; and Fig. 7 is an embodiment-example showing dual linear detector-arrays in the central code-track region.

7 WO 2006/032241 PCT/D1'2005/001616 Description of the Examples Fig. 1 is a diagrammatic view of an unrolled code-disk with four code-tracks on it. The width of each code-track 2, 3, 4, 5 is L. These code-tracks 2 to 5 each have a linear s detector-array 6, 7, 8, 9 assigned to them in the usual manner. The pitch of the linear detector-arrays 6, 7, 8, 9 is P. With this prior-art arrangement, it is possible, for example, to determine the angle of rotation from the movement of code tracks 2-5 along the linear detector-arrays 1o 6-9. If, due to the mechanical axial tolerance of this arrangement, i.e. shifting in the y-direction, perpendicular to the direction in which the code-disk 1 moves, the code tracks 2-5 and the linear detector-arrays 6-9 are both 25 shifted relative to one another, possibly even in opposite directions, then it must be ensured that the code can still be read at all times, i.e. that the code tracks 2-5 can be scanned e.g. optically by the corresponding linear detector array 6-9. 2o The parameters used here - P for the pitch between the linear detector-arrays, L for the code width, and PD for the width of the linear detector-array, all in direction y - can be seen in Fig. 2, which shows the geometrical relationships of the arrangement in Fig. 1 at two extreme positions: i.e. the right-hand position shows shifting, as between code tracks 2-5 and linear detector-arrays 6-9, equal to an axial tolerance of AT. Thus, Fig. 2 shows a conventional detector layout e.g. for four code-tracks, with code-track images, shifted in the y-direction, in the "lower" (left-hand) and "upper" (right-hand) extreme positions. The numbers on the right-hand side of the linear detector-arrays 6-10 indicate which code tracks 2-5 can be analyzed by the linear detector-array concerned.

8 WO 2006/032241 PCT/D:2005/001616 L, P, and AT are related to one another as follows: P = f(L) and AT = g(P), where f and g symbolize functions, and L and P each represent the area required for the code region on the object to be measured (e.g. a rotating shaft); and S = 5 h(P) is the area required for the detector, with h being a function. According to the invention, L and S in particular can now be minimized for a given value of AT. For the arrangement shown in Fig. 2, the following equations apply: 10 P = L, AT = P - PD, and S = 3P + PD. From Figs. 3 and 4, it can be seen that, by adding another linear detector-array 10 to the four code-tracks, the axial 15 tolerance AT can be increased, by modifying arrays 6-10 of the detector so that they can now read and analyze more than 1 code-track - e.g. two code-tracks. Linear detector-array 7 in Fig. 4 can analyze code-tracks 2 and 3, linear detector array 8 can analyze code-tracks 3 and 4, and linear 20 detector-array 9 can analyze code-tracks 4 and 5. A vernier arrangement as shown in Fig. 3, with e.g. five equally spaced linear detector-arrays 6-10 results in a layout as shown in Fig. 4, for scanning four code tracks 2-5. With this arrangement, shifting of the code-tracks results in equally-spaced change-overs, which is thoroughly desirable, because it means that the effect of changing-over on the resulting measurement-values will be kept small. In this regard, the following equations apply: P = 4/5 L, w AT = 3/2 P -PD S = 4P + PD.

9 WO 2006/032241 ICT/DI:2005/001616 In an arrangement with seven linear detector-arrays (not shown), the following equations would apply: P = 4/7 L, AT = 5/2 P -PD, 5 S = 6P + PD. Instead of the side condition P = 4/5 L as in Fig. 4, it can also be required that P + PD = L, resulting in a maximum distance between the linear detector-arrays, and the maximum possible space gain, as shown in Fig. 5. With this 20 arrangement, however, the change-over of the linear detector-arrays 6-10 no longer occurs equi-spatially, but isochronically. Here the following equations apply: P = L -PD, 15 AT = 2P -3PD, S = 4P + PD. From Fig. 4, it can be gathered that the maximum shift of the code-tracks is always when either code-track 2 or code track 5 drifts out of the detector field and thus can no 20 longer be analyzed. This leads to the example in Fig. 6, in which the code is extended by appending an auxiliary code track 2' (left-hand diagram) to code-track 2 and an auxiliary code-track 5' to code-track 5 (right-hand diagram). In this case, the maximum shift is when code-track 2 2' becomes contiguous with linear detector-array 6 or code track 10 becomes contiguous with linear detector-array 5', if e.g. it is required that so-called "dual linear arrays" be used, i.e. with one linear detector-array being able to process a maximum of two adjacent code-tracks. The appended w code segments - auxiliary code-tracks 2' and 5' - are therefore shorter than the main code tracks 2-5. Another 10 WO 2006/032241 PCT/)F2005/00 116 feature in this regard is that now linear detector-arrays 6 and 10 as will have to be designed as "dual linear arrays". The following equations then apply: P = 4/5 L, 5 AT = 2P - PD, S = 4P + PD. Fig. 7 shows another example, in which the three inner linear detector-arrays 7-9 of Fig. 4 or Fig. 6 are in the form of dual linear arrays, i.e. these arrays are each aO divided into two separate arrays that are directly adjacent to each other. The following equations then apply: P = L - PD, AT = 2 PD, 1! S = 4P + 4PD.

Claims

1. A method for detecting the movement or angle of rotation of moving mechanical components, in which - code signals are analyzed that are produced by scanning a number of code-tracks, disposed adjacent to one another and perpendicular to the direction of movement, on the moving component, by means of linear detector-arrays (6, 7, 8, 9, 10) of a stationary sensor that are assigned respectively to said code-tracks (2, iO 3, 4, 5; 2', 5'), characterized in that - additional linear detector-arrays (10) are arranged in such a way that, at the maximum tolerance (AT) allowed to the moving component, perpendicular to the 11 direction of movement of each code-track (2, 3, 4, 5; 2', 5'), at least one of the linear detector-arrays (6, 7, 8, 9, 10) can be assigned thereto.

2. A method as claimed in claim 1, characterized in that: - the linear detector-arrays (6, 7, 8, 9, 10) are 20 adjacent to one another, with a pitch (P) smaller than the code-track width (L), and - the inner linear detector-arrays (6, 7, 8, 9, 10), as a function of position within the tolerance (AT), can each detect one of two adjacent code-tracks (2, 3, 4, 5) ; and at each extreme tolerance position, there is an outer detector-array (6, 10) adjacent to a code-track (2, 5).

3. A method as claimed in claim 2, characterized in that: 12 WO 2006/03224) I>CI/l2005/00 1616 - the pitch (P) of the linear detector-arrays (6, 7, 8, 9, 10) is smaller, by a given factor, than the code track width (L), said factor being the number of code tracks (2, 3, 4, 5) divided by the number of linear 5 detector-arrays (2, 3, 4, 5).

4. A method as claimed in claim 2, characterized in that - the pitch (P) of the linear detector-arrays (6, 7, 8, 9, 10) is smaller than the code-track width (L) by an amount equal to the width (PD) of the linear detector w0 arrays (6, 7, 8, 9, 10) perpendicular to the direction of movement.

5. A method as claimed in any of the above claims, characterized in that - an auxiliary code-track (2', 5') is appended to at , least one of the outer code-tracks (2, 5), said auxiliary code-tracks being also detectable, at the extreme tolerance positions (AT), by the outer detector arrays (6, 10).

6. A sensor arrangement for performing a method as claimed .mo in any of the above claims, characterized in that - the code-tracks (2, 3, 4, 5; 2', 5') are arranged on code-disks on a rotating shaft, and serve for detecting the rotating shaft's angle of rotation and/or its torque.

7. A sensor arrangement as claimed in claim 6, characterized in that 13 WO 2006/032241 PCI/I.)2005/001616 - the code-tracks (2, 3, 4, 5; 2', 5') are formed by light and dark markings on the code-disks, and the linear detector-arrays consist of opto-electronic sensor elements.