GB2314420A - Method of measuring position - Google Patents

Description

2314420 METHOD OF MEASURING POSITION The present invention relates to a

method of measuring position, especially in a length-measuring system.

For various industrial fields of use, for example the measurement of pneumatic/hydraulic pistons, the adjustment of machine tools or generally for an exact positioning of a part, contactless, contamination-insensitive and highly accurate position-measuring devices are needed. Such position-measuring devices comprise, for example, optical or inductive length-measuring systems. Optical systems are usual] y sensitive to contamination and need a medium, which is transparent to light, between the measurement part and the measuring system. Inductive systems usually have only a limited accuracy. Evaluating methods are therefore proposed for positionmeasuring devices to increase accuracy. Such a method operates on, for example, the Vernier principle. In the case of the Vernier principle, an additional line division is used at a length-measuring or angle- measuring instrument and serves for reading off fractions or intermediate values at a uniformly divided main scale instead of estimating them. The Vernier is in that case arranged to be displaceable parallel to the main scale and has a division which is divided more closely or more widely by a fraction than the main scale. If such scales are scanned at a different periodic division, the absolute position of the scanning sensor can be concluded from the phase difference. It is a prerequisite that the two sensors are disposed one beside the other over a respective scale. Although the, absolute position can be concluded from the phase differeifte---with such a position-measuring device, measurement uncertainties with the scanning sensors can result in appreciable errors. Accordingly, modified evaluating methods have been developed, for example by which only the period of the scales is ascertained from the phase difference, whilst the exact position is ascertained by the direct sensor values.

However, measurement errors can arise at the period limits in such methods. It is therefore proposed, in the publication "Magnetoresistiver Sensor mit grossen Einbautoleranzen zur inkrementalen und absoluten Langenmessung" by Dr. Fritz Dettmann and Uwe Loreit of the Institut fOr Mikrostrukturtechnologie und Optoelektronik in WetzIar, to develop the known measurement methods so that the travel or position to be measured is achieved with two periodic scales as in the case of the Vernier principle, wherein one scale has a certain number of periods present over its total length and the other scale has an additional period over the same length.

The scales thus have a linear travel dependence, which is measured by two sensors. For the travel or position measurement, either the scales or the sensors can be moved.

If the phase difference is used only for deciding which period of the scales is, being-. sensed and if one of the sensor signals is then used to ascertain the exact position within the period, the error in the case of evaluation within the periods becomes very small, but large errors, which arise due to an association error, can result at the period limits just as before.

According to the present invention there is provided a method of position measurement in the case of a device with a carrier which has 3 two periodic scales each of the length L and with regions which are of like kind and each form a period, wherein the lengths of the regions are so chosen that m regions are present on one scale and m + 1 regions are present on the other scale, with two sensors which are each associated with a respective scale and supply output signals, which are each displaced in phase relative to the other and evaluated in an evaluating device, wherein either the carrier or the sensors are movable and the sensors define the position to be ascertained, characterised in that the position to be determined of the sensors is ascertained with the aid of an algorithm which at the same time takes into consideration the relationship between the two output signals, which are normalised to their respective maximum, of the two sensors as well as the different period length of the two scales.

A method exemplifying the invention may have the advantage that the measurement errors of the individual sensors are greatly reduced.

It is particularly advantageous that an exact measurement can be performed at the period limits. These advantages are achieved by use of an evaluating algorithm in which it is taken into consideration that a certain relationship exists between the two sensor signals.

This relationship must necessarily exist, because the output signals of the sensors cannot assume any desired values.

An example of the method of the invention will now be more particularly described with reference to the accompanying drawings, in which:

Fig. 1 is a schematic perspective view of measurement scales and sensors for use in carrying out a method exemplifying the inventon; Fig. 2 is a diagram showing the sensor signals and a phase difference therebetween; Fig. 3 is a diagram showing an error course for an evaluation according to the Vernier principle; Fig. 4 is a diagram showing an error course for an evaluation by a modified Vernier principle according to the disclosure of the afore-mentioned publication; and

Fig. 5 is a diagram showing an error course in the case of an evaluation in a method exemplifying the invention.

Referring now to the drawings, there is shown in Fig. 1 a device for ascertaining the position, by which a method exemplifying the invention is performable; the basic device illustrated in Fig. 1 is known from Dettmann and Loreit, op. cit. The device comprises a is carrier T, on which two scales M1 and M2 are applied. These scales can be components of the carrier, such as regions of different magnetisation or differently directed magnetisation or optical elements. The individual scale elements, which each form a respective period, are denoted by E1 in one track and by E2 in the other track.

The scales M1 and M2 are scanned by two sensors SE1 and SE2, respectively, which in the case of magnetic elements are constructed as, for example, Hall sensors or inductive or magnetoresistive pick ups. In the case of optically marked scales, appropriate optical sensors are used, for example light-sensitive diodes, which then co operate with a light-emitting element.

The structuring of the two scales M1 and M2 is such that one scale has m elements El, which are distributed over the measurement length L and form m periods, whilst the second scale has m+1 elements E2, which are distributed over the same length L and form m+1 periods. The scales have a linear travel dependence which is measured with the aid of the sensors SE1 and SE2. For this travel measurement, either the scales or the sensors can be moved, for which purpose either both the scales or both the sensors move in the same manner.

The measurement signals U1 and U2 generated by the sensors SE1 and SE2 are entered in Fig. 2 in dependence on the position of the sensors. In addition, the phase difference PD between the sensor signals U1 and U2, which rises linearly with the position p of the sensors, is entered in Fig. 2.

Since the phase difference PD rises linearly with the position is or travel, the position or travel can be ascertained from the phase difference. This determination corresponds with the Vernier principle. The measurement error F1, which results due to noise, ref erred to the total length L is in the same order of magnitude as the noise of the individual sensor. The error course on application of the Vernier principle is indicated in Fig. 3. It fluctuates in a region of about + 4%.

If thereagainst, evaluation is carried out as proposed in the prior art, thus the phase difference between the two sensor signals is used only for deciding in which period of the scales the sensors are situated and then one of the sensor signals is used in order to ascertain the exact position within this period, the error can be reduced. Measurements have shown that the error course is as indicated in Fig. 4. The error F2 is then very small within the periods, but large errors of up to10%, which arise due to an association error, result at the period boundaries PG, thus in the transition regions between the individual elements E1 or E2 of the scales.

if, the evaluation is carried out by a method exemplifying the invention and described more closely in the following, then as shown in Fig. 5 the error F3 can be significantly reduced at the period boundaries. The error course in the case of this measurement method is apparent from Fig. 5. In that case, the sensor output signals are processed with the aid of an algorithm. The algorithm is based on the recognition that a relationship exists between the values of the two sensors. This relationship is necessarily present because the sensor signals cannot assume any desired values.

If the signal, which is normalised to the maximum of the sensor signals, of the sensor 1 is denoted by S1 and the signal, which is normalised to the maximum, of the sensor 2 is denoted by S2, the result Z of the equation S2 m - S1 (m+l) = Z must be a whole number. Due to measurement errors and inaccuracies, however, this is not the case in practice and no whole number will be obtained from the equation. For that reason, the next whole number k is chosen, for which the following equation applies:

k = [S2 m - S1 (m+l)].

7 The position M to be measured, thus the position, at which the sensors SE1 and SE2 are situated, can be computed with the number k thus ascertained,, in particular by the following equation:

M = L/2 m (m+l) [S1 + (m+l) + S2 m + k (2 m+l)].

If the solution of this equation results in M being negative, the actual position XW to be measured is obtained by adding the overall length of the scales L to the neagtive result of XM, thus XM' = M + L.

If an error computation is performed for the thus ascertained position M or W, the error course illustrated in Fig. 5 is obtained, in which the percentage error F3 of the measurement value is indicated. As is evident from Fig. 5, the error F3 always lies within a range of + 0.2%.

In addition to this very small measurement error, the method has the advantage that a sensor failure can be recognised, since the number k does not change steadily, but can only carry out steps of t m and + (m+l), insofar as the two sensors operate within the frame of a fixed accuracy limit. If one of the sensors fails, k changes by a value of 1 1. It can thus be recognised by checking the change in the magnitude k whether the sensors are functioning correctly or whether one of the sensors is defective.

The described method can be carried out in evaluating equipment A to which the output signals of the sensors SE1 and SE2 are fed.

This evaluating equipment A comprises, for example, a microprocessor, in which the required computations and comparisons take place.

9

Claims (6)

1 A method of measuring position with use of a carrier bearing two periodic scales of the same length and each composed of a plurality of like regions forming periods, wherein the lengths of the regions are such that one of the scales has one more region than the other scale, and two sensors each associated with a respective one of the scales, the method comprising the steps of displacing the carrier and the sensors relative to one another to cause each sensor to scan the respective scale and issue an output signal displaced in phase relative to the signal of the sensor, and evaluating the output signals with evaluating equipment to determine the positions of the sensors relative to the carrier by means of an algorithm taking account of the different period lengths of the scales and the relationship between the output signals of the two sensors, each signal being normalised to its respective maximum.

2. A method as. claimed in claim 2, wherein the positions (XM) are determined by checking whether the result (Z) obtained by the equation Z = S2 m - S1 (m+l) is a whole number and if this is not the case obtaining XM by the equation XM = L/2 m (m+l) [S1 (m+l) + S2 m + k (2 m+l)], - wherein k is the whole number closest to S, L is the length of each scale, m is the number of regions in one scale, m +1 is the number of regions in the other scale, S1 is signal of the sensor associated with the scale having m regions, and S2 the signal of the sensor associated with the scale having m + 1 regions.

3. A method as claimed in claim 2, wherein the step of evaluating comprises checking wherein M is positive or negative and if it is negative a positive position XW is determined by the equation XW = XM + L.

4. A method as claimed in claim 2 or claim 3, comprising the steps of carrying out plausibility checks in respect of successively obtained values of k and recognising a fault in a sensor when the checked values of k do not change in a predetermined manner.

5. A method as claimed in claim 4, wherein the fault is recognised when k changes by plus or minus 1.

6. A method as claimed in claim 1 and substantially as hereinbefore described with reference to Figs. 1 and 2 of the accompanying drawings.