Sensor device and method for determining an absolute position

Abstract

A sensor device (10) for determining an absolute position of a first object relative to a second object, comprising a measuring embodiment (14) connected to the first object, with an absolute coding (16) that encodes respective absolute positions as a code word from a first plurality of code elements of a bit length, a scanning device (18) connected to the second object, with a second plurality of scanning elements (20) for generating a respective scanning signal by scanning the measuring embodiment (14), wherein at least two scanning elements (20) redundantly scan a code element with a mutual offset of a fraction of the bit length, a binarization unit (24) for generating a binary zero or one from the scanning signals, and a control and evaluation unit (26) designed to reconstruct the code word from the binary zeros and ones and thus determine the absolute position, characterized in thatthat the absolute coding (16), the scanning device (18) and / or the binarization unit (24) is designed such that an effective binarization threshold for the assignment of a scanning signal to a binary zero or one is shifted in favor of either the binary zero, so that a binary zero is deliberately generated rather than a binary one, or the binary one, so that a binary one is deliberately generated rather than a binary zero.,

Description

[0001] The invention relates to a sensor device and a method for determining an absolute position according to the preamble of claims 1 and 12, respectively.

[0002] A distinction is made between linear and rotary encoder systems. A linear encoder determines a displacement along an axis. A rotary encoder, angle sensor, or encoder, on the other hand, is used to detect a rotation angle or angular position, for example, of a drive element's shaft. An important application area is motor feedback systems, where the rotary encoder in a servo motor reports the actual speed back to the control system.

[0003] Various technologies are used for the underlying sensor principles. Optically transparent and non-transparent areas are provided for transmitted light in a transmission arrangement of a light transmitter and a scanning sensor, while reflective and absorbing areas are provided for incident light in a reflection arrangement of the light transmitter and scanning sensor. These areas, or code elements, can be openings, reflective and diffractive structures, prisms, structures varying in transmissivity, reflectivity, shade of gray, or color, and other optically distinguishable elements. Together, the code elements form a code track or measuring track. Magnetic encoder devices, for example, with a Hall sensor, detect corresponding magnetic structures or coding permanent magnets. Other physical principles are suitable for position measurements, including capacitive and inductive encoder devices.

[0004] There are a number of suitable coding methods for determining absolute positions, such as Gray codes, code tracks offset according to the vernier principle, or pseudorandom codes. Pseudorandom codes, in particular, make it possible to generate an absolute position from a single measuring track. The following example uses optical scanning, without thereby excluding the alternative technologies already mentioned. To capture the light and shadow pattern generated by the code, photodiodes are arranged in a row in the measuring direction so that they each capture at least one bit of the code. The spacing thus corresponds to one bit or code element. The binarization threshold, with which the scanning signal from each photodiode is converted into one of the two binary states zero or one, is in the middle or at half the maximum level in order to distinguish between the light and dark areas as robustly as possible.

[0005] In principle, a number of photodiodes is required that corresponds to the number of bits in the codeword to be read. However, this leads to problems because the photodiodes are not aligned with a code element during operation. Depending on the relative position between the code and the photodiodes, partial illumination (e.g., half illumination) occurs, resulting in undefined detected bit states. Contrary to the requirement of absolute position determination, which is to be able to unambiguously determine the codeword at any time, this leads to ambiguous decoding.

[0006] A conventional solution is to use twice the number of photodiodes in two rows offset from each other in the measurement direction by, for example, half a bit length, as channels A and B. In each relative position, there is then a channel that is not or hardly affected by the unfavorable partial illumination and thus captures unique bits of the code. Switching back and forth between the two channels A and B occurs at the rate of a relative movement of half a bit length, with the switching times obtained from an additional incremental track.

[0007] Aside from the additional effort required for an incremental track and its scanning, this approach relies on sufficiently precise switching times between the channels. If the switching point is inaccurate, the wrong channel may be activated, resulting in the absolute code not always being clear and, consequently, sometimes being read incorrectly. With a random code, this results in undefined jumps to a code word that, in the worst case, is far away, and thus a potentially very large absolute position error.

[0008] In accordance with the above half-bit cycle, switching must under no circumstances be delayed beyond this amount due to tolerances. The bit length and the length of an increment on the incremental track should ideally match. This results in a so-called phase margin of ± 90° or ± π / 2, based on the period of an increment. The phase measurement on the incremental track, and thus the specification of the switching times, must be at least this precise, taking into account all mechanical tolerances and other influencing factors. Furthermore, in real-world applications, the phase margin cannot be fully utilized because fluctuations due to changes in the system's ambient and operating conditions must be taken into account. A conventional switching system is therefore limited in its design, size, and resolution, or complex and costly adjustment processes are required to circumvent these limitations.However, dynamic effects during operation are not prevented by adjustment.

[0009] DE 199 39 643 A1 discloses a device and method for determining the position between two relatively movable parts. Their code division is divided into 250 sections, each with an 8-bit code. A specific reference mark with the example value 254 is interspersed between each two 8-bit codes for synchronization. Each code element is oversampled with such a large number of pixels that the problem of partial illumination hardly arises. This requires a very high level of effort for the pixel row, its reading, and its evaluation.

[0010] DE 20 2019 101 115 U1 presents a device for determining position, length, or angle that provides more than two readout areas or photodiodes for each code element. In contrast to DE 199 39 643 A1, oversampling remains minimal. Nevertheless, more than two photodiodes are required per code element, increasing the complexity of the device.

[0011] US 2003 / 0187609 A1 discloses a position transducer with an absolute track and an incremental track. The incremental track is used to determine the location of the sampling detectors of the absolute tracks relative to a respective code element. The signal from the sampling detectors is binarized using two different thresholds. If agreement is reached, the binarization result is retained. Otherwise, a binary value is assigned based on the incremental position and the neighboring code elements.

[0012] EP 0 085 951 A2 presents a method and device for detecting measured quantities, in which a mark is projected onto a surface with a large number of photodetectors. The intensity distribution of the mark is determined sequentially using fixed threshold values, and then a center of gravity of the distribution is determined, the position of which is a measure of the quantity to be measured.

[0013] US 2014 / 0166865 A1 describes an absolute encoder with a binarization threshold. During calibration, the light source intensity is set so that the binarization threshold actually lies between the maxima and minima of the evaluated signal.

[0014] It is therefore an object of the invention to improve the determination of an absolute position.

[0015] This object is achieved by a sensor device and a method for determining an absolute position of a first object relative to a second object according to claim 1 and 12, respectively. The two objects are movable relative to one another in a longitudinal or rotational direction, and the corresponding position or angular position is to be measured. In the case of a linear movement, the sensor device is also referred to as a linear sensor; in the case of a rotational movement, it is referred to as a rotary sensor, rotary angle sensor, or encoder; and when used in a servo motor, it is particularly referred to as a motor feedback system. The first object is then preferably a shaft, and the second object is a stationary part or housing of the rotary sensor. Derived variables, such as a speed or angular velocity, can be generated from the position or angular position.

[0016] A measuring embodiment with an elongated or circular measuring track or code track with an absolute coding is connected to the first object. The absolute coding is, for example, a pseudo-random code in which a code word from a first plurality of code elements or bits of a bit length, denoted by m, encodes a respective absolute position. Each of the m code elements represents a binary state of zero or one, such as light and dark in an optical encoder device. A scanning device with a second plurality of scanning elements is connected to the second object and is thus movable relative to the measuring embodiment. The terms first object and second object are interchangeable. Each code element has n≥2 scanning elements, so that each code element is recorded with n-fold redundancy, but with an offset of a fraction of the bit length.Preferably, n=2 applies, whereby one scanning element detects the front part of a code element and one scanning element detects the rear part. The alternative n>2 is also conceivable, particularly in combination with the procedure according to DE 20 2019 101 115 U1 mentioned in the introduction, although this requires additional effort and the invention already manages with n=2.

[0017] A binarization unit generates a binary zero or one from the respective scanning signals of the scanning elements. A control and evaluation unit reconstructs the code word from the sequence of these binary zeros and ones and thus determines the desired absolute position. The control and evaluation unit comprises, for example, one or more circuits or one or more digital computing components, such as at least one microprocessor, FPGA (Field Programmable Gate Array), or ASIC (Application-Specific Integrated Circuit).

[0018] The invention is based on the fundamental idea of unbalanced binarization of the sampled signals into a respective binary zero or one. "Unbalanced" means that there is a deliberate deviation from the binarization threshold, which is conventionally set at 50% of the maximum level. Preferably, a specially adapted decoding rule, which will be explained later, is used to then reconstruct the codeword. This deliberately generates a binary zero rather than a binary one, or in an alternative inverse embodiment, a binary one rather than a binary zero, instead of balancing these as is conventional.

[0019] Several terms require some attention here. First, the preference for a binary zero or one refers to the individual code element. It is not about a codeword that contains more zeros than ones, or vice versa; the invention makes no assumptions in this regard. Furthermore, an effective binarization threshold is shifted. This can be an actual binarization threshold of the binarization unit, but this is only a preferred embodiment. Alternatively or additionally, sampling signals can already be generated by the code elements or by the sampling device that would be more likely to produce binary ones or binary zeros even with a 50% binarization threshold of the binarization unit. This will be explained in more detail below. Finally, a kind of change in direction with regard to the shifting of the effective binarization threshold must be noted.If the effective binarization threshold is shifted toward 100%=1, this favors binary zeros, since only the now relatively rarer suprathreshold samples still yield a binary one. Conversely, an effective binarization threshold shifted toward 0 favors the binary one. The independent claims formulate this from the perspective of the favored states, not the effective binarization threshold itself, which would be equally possible.

[0020] The invention has the advantage that, in contrast to the prior art described in the introduction, switching between two channels is no longer necessary. The absolute position is read directly and unambiguously from the absolutely coded track. This eliminates the need for an incremental track, although an additional incremental system can of course still be provided, for example, to increase resolution. However, the absolute position determination according to the invention is self-sufficient and does not depend on the incremental system. Errors due to unfavorable switching times are no longer possible. The invention is robust against undefined or ambiguous states in transition positions, i.e., when a scanning element detects portions of two adjacent code elements. With n=2, the same number of scanning elements as in the switching method is sufficient.

[0021] Preferably, a threshold of the binarization unit is shifted from a central position between the minimum level of the sampling signal and the maximum level of the sampling signal. The minimum level is typically zero, so that the central position is assumed to be 50% for simplicity. In this embodiment, the effective binarization threshold is shifted by the binarization unit performing unbalanced binarization, i.e., making its distinction between binary one and zero in an unbalanced or off-center manner with a threshold not equal to 50%. The circuit of the binarization unit is constructed or parameterized with the unbalanced threshold. This preferably includes the case in which the binarization threshold is set to 50% of an original sampling signal, but the sampling signal is then further amplified or attenuated before binarization. The threshold of the binarization unit preferably corresponds to the effective binarization threshold.However, it is also conceivable that several effects are cumulated, which will be explained immediately, in which case the threshold of the binarization unit may differ from the effective binarization threshold.

[0022] The effective binarization threshold is preferably shifted by a level of the scanning signals, wherein the level is set by the configuration of at least some code elements, at least some scanning elements, and / or a signal generator of the scanning device. This is an alternative or supplement to shifting a threshold of the binarization unit. To manipulate the scanning signal and change its level, one or more of the following measures are conceivable: changing the width of the code elements transversely to the measuring direction, changing the scanning elements, for example, their area or geometry, or changing a signal generator of the scanning device. This is, for example, a light source that, in interaction with the code track, generates the light-dark pattern to be detected and whose signal strength or illumination characteristics can be changed.

[0023] The control and evaluation unit is preferably designed to form groups of zeros of consecutive binary zeros and groups of ones of consecutive binary ones and to reconstruct the code word from the groups as follows: - Determine a bit number corresponding to the group by dividing the length of the group by the number of scanning elements scanning a code element together, - If the effective binarization threshold is shifted in favor of binary zero, round the number of bits up for groups of ones and down for groups of zeros, and conversely, if the effective binarization threshold is shifted in favor of binary one, round the number of bits up for groups of zeros and down for groups of ones, and - in the code word, while maintaining the original order of the groups, set a number of code elements corresponding to the rounded number of bits for each group of zeros to zero and a number of code elements corresponding to the rounded number of bits for each group of ones to one.

[0024] This is an example of a decoding rule adapted to the unbalanced effective binarization threshold according to the invention. First, the binarized sampling signals generated from the codeword detection are divided into blocks of consecutive binary zeros or ones, referred to as zero groups and ones groups. For each group, the number of zeros or ones it contains is counted, and this is referred to as the bit count. This bit count is further corrected to correctly account for code elements in the transition between two groups, where a sampling element samples two different code elements simultaneously in numerous relative positions of the two objects. First, an embodiment is considered in which the effective binarization threshold favors the binary zero. Then, the ones groups are rounded up, bit count = ceil(bit count), and the zero groups are rounded down, bit count = floor(bit count).It should be noted that the rounding is preferably not the usual rounding as stated, but rather is incremented once with ceil and decremented once with floor, regardless of the value of the first decimal place, and then all decimal places are omitted. In an inverse embodiment, in which the effective binarization threshold favors binary one, the rounding is reversed. Thus, the ones are rounded down, bit number:= floor(bit number), and the zeros are rounded up, bit number:=ceil(bit number). In both embodiments, the reconstructed codeword is created, regardless of the preference for binary zeros or ones, by combining binary zeros for each zero group and binary zeros for each one group, according to the original arrangement of the groups, to form the reconstructed codeword.

[0025] Preferably, all scanning elements scan the measuring scale together, and / or the control and evaluation unit reconstructs a code word from the binary zeros and ones of the scanning signals of all scanning elements. Consequently, there are no different channels as in the conventional solution described in the introduction, and consequently no switching between them. All scanning elements are active, and the reconstruction refers to all scanning signals. In particular, according to the advantageous decoding rule presented in the previous paragraph, the number of consecutive binary zeros or binary ones is counted across all scanning elements or their scanning signals. It remains conceivable to provide additional scanning elements that are not involved in the reconstruction, for example, as reference elements.

[0026] The effective binarization threshold preferably lies in the interval [51%, 99%], [55%, 95%], [60%, 90%], [1%, 49%], [5%, 45%], or [10%, 40%], relative to a maximum level of the scanning signals. An effective binarization threshold above 50% favors the binary zero, while an effective binarization threshold below 50% of an inverse embodiment favors the binary one. Furthermore, the reference point is a maximum level or a central effective binarization threshold at 50% from which the shift is made. With an effective binarization threshold of 50%, even with the slightest and, in practice, completely unavoidable noise effect, a scanning element would randomly deliver a binary zero or one if it detects two different code elements in a transition position. The greater the distance to 50%, the greater the robustness.An effective binarization threshold of 0% or 100%, on the other hand, would ignore the measurement, so a certain margin must be maintained for this as well. The effective binarization threshold can, in principle, be different for the individual sampling signals or sampling elements, but is preferably the same for all, as otherwise, a different robustness would result for each code segment.

[0027] The effective binarization threshold preferably lies in the interval [70%, 80%] or the interval [20%, 30%] or is essentially 75% or 25%. This maximizes the distance just explained to a central value of 50% and to the extreme values of 0% and 100%. An optimum of a maximum noise or tolerance margin is 75%, or in an alternative inverse embodiment, 25%. This optimum is particularly preferably essentially realized, which means that this is the desired and set value, but tolerances of a few percentage points, one percentage point, or less than one percentage point are still conceivable. With this understanding of tolerances, the formulation can essentially be omitted.

[0028] The scanning device preferably has a second plurality of scanning elements, which is at least twice the first plurality. With m code elements, there are therefore at least m*n scanning elements, more preferably exactly m*n code elements, whereby, as already mentioned above, n=2 preferably applies, with just twice as many scanning elements as code elements, or 2m scanning elements. Thus, a complete code word consisting of m code elements can be read with n scans per code element.

[0029] The scanning elements preferably form a row arrangement, in particular with equal spacing between them. The row arrangement is aligned in the measuring direction, i.e., along the measuring track of the absolute coding. It is preferably a single row. Multiple rows offset transversely to the measuring direction are alternatively conceivable, provided that all scanning elements can still detect the code word. The scanning elements are preferably evenly spaced from one another within the row or within the rows. In this case, there are no differences in the detection of different code elements.

[0030] The scanning elements preferably each detect a portion of a code element of equal length. They are preferably of equal length in the measuring direction, i.e., in the longitudinal or circumferential direction, depending on the type of encoder device. Even more preferably, n scanning elements each divide a code word equally among themselves, so that the detected portion corresponds to the quotient of the bit length and the number n of scanning elements that scan a code element with a mutual offset. Alternatively, longer portions of a code element can also be detected, for example, using a zipper arrangement of several offset rows next to one another.

[0031] The regular arrangements explained are preferred, but the invention is not initially limited to them. Thus, scanning elements can capture portions of code sections of different lengths and / or have different mutual spacing. This results in different effective binarization thresholds for each scanning signal, which was already mentioned above as possible, but not necessarily advantageous.

[0032] The scanning device preferably detects the same single code track made up of code elements of the absolute coding. This is to emphasize again that the invention operates autonomously with absolute coding, and that this absolute coding can in turn be designed, for example, as pseudo-random coding with just a single track. There are preferably no further tracks on the measuring embodiment, in particular no incremental track for defining switchover points between two channels of scanning elements, because the invention does not require such channels. If further code tracks are provided on the measuring embodiment for other purposes, for example an incremental system for increasing resolution, the scanning device is not interested in them; it does not generate any scanning signals from them and does not require the information from any further code tracks because it determines the absolute position autonomously from the code track with code elements of the absolute coding.

[0033] The method according to the invention can be further developed in a similar manner and thereby exhibits similar advantages. Such advantageous features are described by way of example, but not exhaustively, in the subclaims following the independent claims.

[0034] The invention will be explained in more detail below with regard to further features and advantages, using exemplary embodiments and with reference to the accompanying drawings. The figures of the drawing show: Fig. 1 a schematic representation of a rotary encoder; Fig. 2 an explanatory diagram of the scanning of an absolute code in an ideal relative position between scanning elements and code elements; Fig. 3 a representation similar Fig. 2, in which the scanning elements are shifted to the code elements by half their width; and Fig. 4 a representation similar Fig.2, in which the scanning elements are shifted to the code elements by three-quarters of their width.

[0035] Fig. 1 shows a schematic representation of a sensor device 10 in one embodiment as an encoder or rotary encoder. The sensor device 10 has a code disk or measuring scale 14 rotating with a shaft 12, and on the measuring scale 14 is a code track or measuring track 16 with an absolute coding. In the absolute coding, code elements (not shown) with two different states alternate with one another. One example is a pseudorandom code, but the invention is not limited to this; any binary codes are possible. M code elements detected together result in a code word from which a unique absolute position can be deduced. The measuring scale 14 is preferably circular and has only the one measuring track 16.

[0036] On the circumference of the measuring scale 14 or the measuring track 16, a scanning device 18 is arranged with a plurality of scanning elements 20, which in the optical embodiment of a sensor device 10 shown as an example are designed, for example, as photodiodes and are preferably arranged in a row. The scanning elements 20 are preferably integrated on an opto-ASIC (Application-Specific Integrated Circuit) and receive the light from a light transmitter 22 assigned to the scanning device 18, which light penetrates the measuring track 16. The scanning elements 20 each generate a scanning signal whose amplitude or level depends on the state of the respective detected code element.

[0037] A binarization unit 24 converts the sampled signals into binary ones and zeros by evaluating them with a binarization threshold. The binarization unit 24 preferably comprises analog circuit elements, such as a threshold comparator. Alternatively, the sampled signals can be digitized with higher resolution and then computationally evaluated with the binarization threshold.

[0038] A control and evaluation unit 26 first generates the code word currently detected by the scanning device 18 from the binary zeros and ones, and then uses this to determine the absolute position of the measuring scale 14. This absolute position or variables determined therefrom, such as speed information after differentiation, can be provided at an output. The clear separation of the scanning device, binarization unit 24, and control and evaluation unit 26 into Fig.1 is for illustrative purposes only; the functionality can be distributed across common or additional modules.

[0039] The representation of the encoder device 10 in Fig. 1 is very schematic. Therefore, the illustrated design of the scanning device 18 with its scanning elements 20 and the associated light transmitter 22 is kept very simple. Particularly with regard to the dimensions and specific positions of the components, the sensor device 10 can differ greatly in practice from Fig.1. As an alternative to a transmissive design, a reflective design is also conceivable, in which the scanning device 18 and light transmitter 22 are located on the same side of the measuring scale 14 and the measuring track 16 has reflective properties. The design as an optoelectronic sensor device 10 is also exemplary; alternatively, a magnetic, inductive, or capacitive sensor device 10 can be provided according to the invention, or another suitable physical operating principle can be used to detect scanning signals. In a magnetic system, for example, instead of transparent and non-transparent areas of the measuring track 16, north poles and south poles alternate, which are detected by Hall sensors or the like as scanning elements 20 of the scanning device 18.

[0040] Furthermore, the encoder device 10 can be designed as a linear encoder or a longitudinal measuring system instead of a rotary encoder. The measuring scale 14 is then elongated rather than circular, the measuring track is arranged on a straight line instead of a circle, and the movement occurs in a longitudinal direction instead of a rotation.

[0041] Fig. Figure 2 shows an explanatory diagram of the scanning of a code word of the absolute coding of the measuring track 16, wherein the absolute coding is exemplified as a binary pseudorandom code. The code elements of the code word are shown in black or white, based on an optical detection principle, and thus also represent two binary states of an absolute coding for any sensor principle. As shown above the code word with curly brackets, individual or multiple light and dark code elements or bits follow one another.

[0042] To scan the code word, n=2 scanning elements 20 of the scanning device 18 are provided for each bit or code element encoded therein. In principle, more scanning elements 20 per code element are conceivable, i.e., n>2, particularly in combination with the procedure according to DE 20 2019 101 115 U1 cited above, but according to the invention, two scanning elements 20 per code element are sufficient. If the code word comprises a total of m bits or code elements, n*m scanning elements 20 are required.

[0043] Each scanning element 20 generates a scanning signal which is converted in the binarization unit 24 into a binary zero or binary one, which is Fig.2 is written into the respective scanning element. The adjacent binary zeros and binary ones are then grouped into segments and counted, as indicated by curly brackets below the scanning elements 20. There is no division into two channels A and B as in the prior art described above, nor any switching between them; the scanning signals and the binary states of all scanning elements 20 determined from them are taken into account.

[0044] In Fig.2, the codeword of the measuring track 16 and the scanning elements 20 are in an ideal relative position. The light-dark transitions of the codeword lie exactly at the transition between two scanning elements 20. Each scanning element 20 thus completely covers either a light or a dark code element, and the binary states are unambiguous. With this 100% coverage of the code elements, decoding is very simple. The length of the segments or contiguous sequences of binary zeros and binary ones is divided by the number n of those scanning elements 20 that together cover a code element, here n=2. This results in the decoded code 28 shown below. For the first segment of the codeword with two dark code elements, for example, four consecutive binary zeros were detected. This results in four divided by n=2 and thus two binary zeros at the beginning of the decoded code 28. This decoding continues analogously for the remaining segments.

[0045] Fig. 3 shows a similar representation as Fig.2, except that the scanning elements 20 are now shifted relative to the code elements of the measuring track 16 by half the width of the scanning elements 20. This results in a 50% coverage at the light-dark transitions of the code word. With the conventional central binarization threshold, the binary state of a scanning element 20a at a light-dark transition would be undefined. The level of the scanning signal would be just at the threshold, and the binary state would randomly flip in one direction or the other due to minimal tolerances. Thus, the number of contiguous binary zeros and ones would be variable. For example, one, two, or three contiguous binary zeros or ones could be detected for one code element, and three, four, or five contiguous binary zeros or ones could be detected for two code elements.This would make the decoding ambiguous, and depending on the coding, for example in a pseudorandom code, the difference in one bit can cause arbitrarily large errors in the determined absolute position.

[0046] Therefore, according to the invention, the conventional binarization threshold is shifted from its central position at 50%. An optimum is achieved with a binarization threshold of 75%, as this allows the greatest symmetrical tolerance for n=2 scanning elements per codeword. With this binarization threshold of 75%, the binary state one is only generated when the level of the scanning signal of the scanning elements reaches 3 / 4 of the maximum level. In an optical embodiment, therefore, a scanning element must be illuminated by 3 / 4. This is the case with the 50% coverage according to Fig.3 is not the case. Thus, a scanning element 20a clearly generates a binary zero at a light-dark transition of the codeword. Tolerances are neglected in this consideration, but would be accommodated by the 75% threshold.

[0047] The binarization threshold at 75% favors the binary zero, since a binary one is only generated when the sampled signal reaches the high portion of three-quarters of the maximum level. In an inverse embodiment, the optimum for the binarization threshold is 25%, which then favors the binary one, but otherwise functions entirely analogously. It is also not absolutely necessary to set the binarization threshold to the optimum. Rather, it is sufficient to choose a binarization threshold >50% or <50%. Outside the optimum, the tolerance margin is then smaller than 50% or the extreme values 0% and 100%.

[0048] The binarization threshold is most clearly understood as the actual binarization threshold of the binarization unit 24. However, it is sufficient to shift an effective binarization threshold from the conventional central position. This can also be achieved using numerous other parameters, which are presented at the end of the description.

[0049] Again Fig.As can be seen from Figure 3, in the case of a relative position of the scanning elements 20 to the code elements with 50% coverage and a binarization threshold of 75%, no ambiguous states arise. In a dark segment, the number of binary zeros is indeed greater than the number of binary ones in a light segment of the same length. This is a direct consequence of the fact that the binarization threshold shifted to 75% favors the binary zero. However, one contiguous binary one unambiguously represents a one in the decoded code 28, three contiguous binary zeros represent one zero, three contiguous binary ones represent two ones, five contiguous binary zeros represent two zeros, and so on.

[0050] Fig. 4 shows again a similar representation as Fig.2, whereby the scanning elements 20 are now shifted by three-quarters of their width relative to the code elements of the measuring track 16. This results in an overlap of 75% and 25% respectively. Even with the binarization threshold still shifted to 75% as an example, ambiguous states now arise. During a transition from dark to light, an associated scanning element 20a remains in a binary state of zero, because an even smaller portion of the light code element is detected. During a transition from light to dark, however, an associated scanning element 20b detects just three-quarters of the light code element. The level of the scanning signal is thus exactly at the binarization threshold, which here is still set to the optimum of 75% as an example, and the binary state will tip unpredictably in one direction or the other due to noise, which is marked by a question mark.

[0051] As a result, the length of the segments is no longer unique. Unlike the conventional approach with a 50% threshold, however, this can be compensated for by a suitable decoding rule, meaning that the uniqueness can be restored. This can be Fig. 4 based on the two possible lengths of the segments entered. In each segment, there is only a single question mark for the respective scanning element 20b at the light-dark transition, thus avoiding the overlap that would cause ambiguity in the decoding.

[0052] For a segment of binary zeros, it is certain that there are at least as many binary zeros as in the case of a 50% threshold. The corresponding scanning element 20a at the dark-to-light transition will always generate a binary zero due to the binarization threshold of 75%. The number of binary zeros can only increase by one due to the question mark of the corresponding scanning element 20b at the light-to-dark transition, but can never increase by two or become less than in the case of the 50% threshold. Consequently, there are two or three binary zeros for a single zero in the codeword, four or five binary zeros for two consecutive zeros, and so on, so that each number of binary zeros in a segment can be uniquely assigned to a number of zeros in the decoded code 28.

[0053] Analogously, there are at most as many binary ones as in the case of a 50% threshold. A binary zero is still always generated for the corresponding scanning element 20a at the dark-to-light transition. The number of binary ones can only decrease by one due to the question mark of the scanning element 20b at the light-to-dark transition, but can never decrease by two or increase by more than in the case of the 50% threshold. Consequently, there are one or two binary ones for a single one in the codeword, two or three binary ones for two consecutive ones, and so on, so that each number of binary ones in a segment can also be uniquely assigned to a number of ones in the decoded code 28.

[0054] A special case arises when the first or the last of the scanning elements 20 forms a segment with a single binary zero. If this concerns the first of the scanning elements 20, this 1-segment is exceptionally translated into a single zero at the beginning of the decoded code 28, and everything else is shifted by one scanning element 20 or half a code element. If it concerns the last of the scanning elements 20, this can exceptionally be translated into a single zero at the end of the decoded code 28, provided that the decoded code 28 has not yet reached the specified bit length. The latter case is the case in the example of the Fig. 4 occurred: The last segment can contain one or two binary zeros depending on the state of the question mark, and even with only one binary zero, if the question mark assumes the binary state one, a zero is appended in the decoded code 28.

[0055] To restore uniqueness, the decoding rule for segments with binary zeros differs from the decoding rule for binary ones. The number of consecutive binary zeros or ones in the segment is still counted together, and this number of bits is divided by the number n = 2 of scanning elements per code element, Bit number = [Length of segment] / n. However, a distinction is then made. For a segment with binary zeros, the number of bits is corrected by rounding off any decimal places, Bit number = floor(Bit number). For a segment with binary ones, however, the number is rounded up, and this happens at any value of the first decimal place, not only at a value greater than or equal to five: Bit number = ceil(Bit number). Finally, the special case of a single binary zero at the beginning or end must be considered. For the example n = 2 and a binarization threshold of 75%, the following decoding rule results: Length of segment with consecutive binary zeros Length of segment with consecutive binary ones Decoded in number of zeros / ones 2 or 3 (special case 1 at the beginning / at the end) 1 or 2 1 4 or 5 3 or 4 2 6 or 7 5 or 6 3 8 or 9 7 or 8 4 ... ... ...

[0056] The binarization threshold of 75% ensures optimal symmetric tolerance distance to 50% and the boundary value of 100%. However, with the same decoding rule, the weaker condition of >50% and <100% is sufficient, although a certain distance should preferably be maintained to avoid becoming too sensitive to noise, such as a binarization threshold in the interval [55%, 95%], [60%, 80%], or [70%, 80%].

[0057] In other embodiments, the situation is inverted, and a binarization threshold of <50% is selected. This then leads to a preference for binary ones, and thus adjustments must be made for the analogous inverted case: The above rounding rules and thus the values in the first two columns of the table are to be swapped, and the edge cases at the beginning and end of the codeword now concern light segments with a binary one instead of dark segments with a binary zero. The optimum in the inverted case is 25%; generally, the condition <50% and >0% is sufficient, with other exemplary suitable binarization thresholds in the interval [5%, 45%], [10%, 40%], or [20%, 30%]. In principle, 20 different binarization thresholds can be selected for each sampling element, which then creates an individual tolerance margin and may require the decoding rule to be adapted locally, depending on the sampling element 20.

[0058] It has already been pointed out several times that the binarization threshold of the binarization unit 24 is not the only way to shift the assignment of scanning signals to binary states and to favor binary zeros or binary ones. This is referred to as the effective binarization threshold, which can be influenced by a non-centered binarization threshold of the binarization unit 24, but does not have to be. There are a number of other ways to shift the effective binarization threshold, which can be combined with one another and / or with a shift of the binarization threshold of the binarization unit 24. These include the design of the scanning elements 20, in particular their geometry, for example, using triangular photodiodes. Furthermore, the geometry, in particular the width, transmissivity, or reflectivity of the code elements of the measuring track 16, can be manipulated.Depending on the sensor principle, there is also a signal generator, in the case of optical detection the light source 22. Its signal strength or specifically the lighting characteristics can also be changed.

Claims

[1] Transducer device (10) for determining an absolute position of a first object relative to a second object, a measuring embodiment (14) connected to the first object with an absolute coding (16) which encodes respective absolute positions as a code word from a first plurality of code elements of a bit length, a scanning device (18) connected to the second object, having a second plurality of scanning elements (20) for generating a respective scanning signal by scanning the measuring standard (14), wherein at least two scanning elements (20) each redundantly scan a code element with a mutual offset of a fraction of the bit length, a binarization unit (24) for generating a binary zero or one from the scanning signals, and a control and evaluation unit (26) which is designed to reconstruct the code word from the binary zeros and ones and thus to determine the absolute position, characterized by , that the absolute coding (16), the scanning device (18) and / or the binarization unit (24) is designed such that an effective binarization threshold for the assignment of a scanning signal to a binary zero or one is shifted in favor of either the binary zero, so that a binary zero is deliberately generated rather than a binary one, or the binary one, so that a binary one is deliberately generated rather than a binary zero. [2] Transducer device (10) according to claim 1, wherein a threshold of the binarization unit (24) is shifted from a central position between the minimum level of the sampling signal and the maximum level of the sampling signal. [3] Transmitter device (10) according to claim 1 or 2, wherein the effective binarization threshold is shifted by a level of the scanning signals, wherein the level is set by designing at least some code elements, at least some scanning elements (20) and / or a signal transmitter (22) of the scanning device (18). [4] Encoder device (10) according to one of the preceding claims, wherein the control and evaluation unit (26) is designed to form groups of zeros of consecutive binary zeros and groups of ones of consecutive binary ones and to reconstruct the code word from the groups as follows: Determine a bit number corresponding to the group by dividing the length of the group by the number of scanning elements (20) scanning a code element together, if the effective binarization threshold is shifted in favor of binary zero, round the number of bits up for groups of ones and down for groups of zeros, and Conversely, if the effective binarization threshold is shifted in favor of binary one, round the bit number up for zero groups and down for ones and In the code word, while maintaining the original order of the groups, set a number of code elements corresponding to the rounded number of bits for each group of zeros to zero and a number of code elements corresponding to the rounded number of bits for each group of ones to one. [5] Encoder device (10) according to one of the preceding claims, wherein all scanning elements (20) scan the measuring embodiment (14) together and / or the control and evaluation unit (26) reconstructs a code word from the binary zeros and ones of the scanning signals of all scanning elements (20). [6] Transducer device (10) according to one of the preceding claims, wherein the effective binarization threshold lies in the interval [51%, 99%], [55%, 95%], [60%, 90%], [1%, 49%], [5%, 45%] or [10%, 40%], based on a center position at 50%. [7] Transducer device (10) according to one of the preceding claims, wherein the effective binarization threshold lies in the interval [70%, 80%] or the interval [20%, 30%] or is substantially 75% or 25%, relative to a center position at 50%. [8] Transducer device (10) according to one of the preceding claims, wherein the second plurality of scanning elements (20) of the scanning device (18) is at least twice the first plurality. [9] Sensor device (10) according to one of the preceding claims, wherein the scanning elements (20) form a series arrangement, in particular with a uniform mutual spacing. [10] Encoder device (10) according to one of the preceding claims, wherein the scanning elements (20) each detect a portion of a code element of equal length, in particular a portion which corresponds to the quotient of bit length and number of at least two scanning elements (20) which scan a code element with mutual offset. [11] Encoder device (10) according to one of the preceding claims, wherein the scanning device (18) detects the same single code track of code elements of the absolute coding (16). [12] Method for determining an absolute position of a first object relative to a second object, in which a measuring embodiment (14) connected to the first object with an absolute coding (16) which encodes respective absolute positions as a code word from a first plurality of code elements of a bit length is scanned to generate a respective scanning signal by a scanning device (18) connected to the second object with a second plurality of scanning elements (20), of which at least two scanning elements (20) each redundantly scan a code element with a mutual offset of a fraction of the bit length, wherein a binary zero or one is generated from each of the scanning signals and the code word is reconstructed from the binary zeros and ones and the absolute position is thus determined, characterized bythat an effective binarization threshold for the assignment of a sample signal to a binary zero or one is shifted in favor of either the binary zero, so that a binary zero is deliberately generated rather than a binary one, or the binary one, so that a binary one is deliberately generated rather than a binary zero.

Images