WO2006064442A1

WO2006064442A1 - Method and apparatus for measurement of straightness of a straightedge using a multi-probe sensor

Info

Publication number: WO2006064442A1
Application number: PCT/IB2005/054172
Authority: WO
Inventors: Antonius M. Rijken; Martinus A. M. Cuppen; Erik J. M. Janssen; Theo A. M. Ruijl
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2004-12-16
Filing date: 2005-12-12
Publication date: 2006-06-22

Abstract

In an apparatus for measuring position errors in a machine having a movable element (2) and a straightedge (3) and a system for measurement of the straightness of the straightedge, said measurement system comprises a multi-probe (4a, 4b, 4c) device (4) for sequentially measuring along the straightedge (3) using a carriage (4) moving along a guide way (G (x) ) . The carriage (4) is moved along one surface (S (x) ) of the straightedge (3) to take a set of measurement points, whereafter a second set of measurement points, with an offset 0, is taken, and the two sets of measurement points are mapped together.

Description

METHOD AND APPARATUS FOR MEASUREMENT OF STRAIGHTNESS OF A STRAIGHTEDGE USING A MULTI-PROBE SENSOR

The invention relates to a sequential multi-probe method for measurement of straightness of a straightedge using a multi-probe device for sequential measurements along the straightedge using a carriage moving along a guide way.

The invention also relates to an apparatus for measuring position errors in a machine having a movable element, a straightedge and a measurement system for measurement of the straightness of the straightedge, said measurement system comprising a multi-probe device for sequentially measuring along the straightedge using a carriage moving along a guide way and a calculator for calculating the straightness of the straightedge.

The invention also relates to a measurement system for measurement of the straightness of the straightedge, said system comprising a multi-probe device for sequentially measuring along the straightedge using a carriage moving along a guide way and a calculator for calculating the straightness of the straightedge.

Machine tools and multi-axis machinery require a high standard precision in line with the development of high precision engineering. High precision in manufacturing can only be accomplished if it is possible to measure and calculate accurately the errors of machine components.

Coordinate measurement machines are used for 1- 2 or 3 dimensional inspection of work pieces such as machine parts. A work piece is typically secured to a fixed table, and a measuring probe is used which is movable in one, two or three dimensions. To measure the position of a point on the work piece, the probe in brought into contact with the point or in other ways for instance capacitively a measurement is made, and measuring scales or other sensors on the machine are read. Probes may be of any type, the probes may be contact probes which make contact with the straightedge, or they may be optical probes, or they may be contactless probes based on Eddy currents or capacitance. The position of the point is typically expressed as X, Y and/or Z coordinates within a working volume of the machine. To measure a distance between two points, the points are measured successively, the coordinates of the points are read, and the distance between the points is calculated from the coordinates. State of the art coordinate measuring machines typically have features such as high resolution measuring systems, electrical contact probes, motor drives, computer controlled drives and computer acquisition and processing of data.

One type of coordinate measuring machine is known as a moving bridge machine. A bridge moves in the Y direction along guide ways on a table. A carriage moves in the X direction along guide ways on the bridge. Scales associated with each of the movable elements indicate the positions of the movable elements in three axial directions.

The accuracy of a coordinate measuring machine is limited by inaccuracies in the scales or other measuring devices, and by faults in the guide ways or other elements which define machine motions. One approach to increasing accuracy is simply to improve the construction techniques and to reduce tolerances of the system so that errors are reduced. However, the reduction of errors becomes progressively more expensive as required accuracies increase and as the size of the work pieces increase.

Typically the 'individual machine fault' of an individual machine is measured with (often) laser measurement tools and stored. Then all kind of efforts are made to keep conditions the same, so that the 'individual machine fault' stays the same. This requires a very good control over conditions such as temperature and humidity, the use of often (very) expensive materials such as Zerodur and invar to reduce as much as possible any deviation of the established 'individual machine fault'. Even then, the measurement procedure has to be regularly repeated after any service activity that could have affected the configuration.

Various techniques to measure and calculate such errors are known. In such techniques often a precision straightedge is used as a gauge to measure the error of machine components or machine movements. In order for the measurement to be accurate the straightness of the precision straightedge itself has to be precisely and accurately measured, i.e. the straightedge surface error has to be accurately measured.

Li et al describe in SPIE vol. 2101 Measurement technology and intelligent Instruments (1993), page 483 describe a sequential-three-points method for measurement of the straightness of precision straightedges. "Multi-probe' measurement means, within the concept of the present invention, that at least three probes are used. In such measurement a carriage in which three probes are provided is moved along a guide way and sequentially sets of points (at least three) are measured on the straightedge. Using the sets of measured points, Li et al state that it is possible to calculate the straightness of the straightedge, independent of the errors in guide way or yaw error of the carriage. The known method is restricted due to the fact that the probes are separated at a distance L from each other, the so-called 'probe interval'. In principle the sampling interval (the distance between measuring points) and the distance between the probes has to be equal. This restricts the spatial resolution of the measurements. To overcome this restriction Li et al has proposed a method in Measurement Technology and Intelligent Instruments, SPIE VoI 2101, pag 483-487, 'on-line measurement of the straightness of precision lathes'.

In this method different probe intervals are used. Using different probe intervals and mapping profiles together a higher spatial resolution is obtainable.

However, as will be explained below, a systematic error in the method introduces an error in mapping the profiles. This systematic error leads to a systematic error in the calibration.

In fact, all known prior art systems for calibrating coordinate measuring machines have been relatively complex and expensive. In addition, calibration procedures are lengthy, complex, expensive and subject to error.

It is an object of the invention to provide a method, apparatus and system as described in the 'field of the invention' paragraph with an increased accuracy and/or which is relatively simple and/or is better suited for on-line measurements. To this end the method, apparatus and system are characterised in that the carriage is moved along a surface of the straightedge to take a set of measurement points, and is subsequently moved again along the same surface to measure a set of measurements points in between the previously measured points, whereafter the sets of measurement points are mapped together by assuming that an average line through said sets of measurement points is the same.

The invention is based on the insights that:

1. Ideally, the sequential multi-probe measurement should be free of systematically errors in guide way or carriage. However, the inventors have realized that still a systematic error persists. This systematic error changes as probe intervals change, so that mapping profiles together taken with different probe intervals, as proposed by Li et al, is prone to error, because of the unknown and possible different systematic errors.

2. By moving the carriage more than once along a surface of the straightedge, without changing the probes mutual position, and using an offset between the respective sets of measurements points, profiles are taken that do not comprise the error identified in point 1. A priori the two sets of measurements are independent measurements, without common points and thus it is difficult to combine the two (or more) sets of measurements points. However, it has been found that the assumption that the average line through the profiles is the same allows the profiles to be combined. The average line through the profiles can be obtained by for instance a least square method. Below, the word "profiles' will also be used to denote the respective sets of measurements points. An average line through a profile is found be plotting a profile (a set of measurements points) as a function of distance travelled along the straightedge, and fitting a straight line through the measurements points using e.g. a least square error fit. Two profiles will each give a straight line. The two profiles are mapped together by shifting up or down and/or rotating one of the profiles until the two straight lines through the profiles match. When more than two profiles are taken, e.g. four, the four profiles are matched, and the same manner of matching is used.

Preferably p sets of measurement points are taken and the distance between neighbouring points at different sets of measurement points is such that the distance between points taken at different sets is O=L/p where L is the distance between neighbouring probes in the carriage. In such embodiments the measurement points are equally distributed over the straightedge which increases the accuracy.

In preferred embodiments measurement points are taken at opposite surfaces of the straight edge. This allows eliminating a further systematic error, as will be explained below.

The apparatus and measurement system are characterized in that they are arranged to move the carriage along a surface of the straightedge to take a set of measurement points, and to subsequently move the carriage again along the same surface to measure a set of measurements points in between the previously measured points, and to map the sets of measurement points together by assuming that an average line through said sets is the same.

In preferred embodiments the apparatus and measurement system are arranged to take p sets of measurement points wherein the distance between points taken at different sets is O=L/p where L is the distance between neighbouring probes in the carriage. In more preferred embodiments the apparatus and system comprises means to take measurements at opposite surfaces of the straightedge.

In practice this means that the apparatus and system comprise means to take two (or p) sets of measurement points with a controlled offset between the sets of points means to calculate the average lines, and to map the sets of measurement points together.

These and further aspects of the invention will be explained in greater detail by way of example and with reference to the accompanying drawings, in which Fig. 1 is a schematic drawing of a x-y moving machine. Fig. 2 shows an example of a precision gauge straightedge on a machine as shown in figure 1. Figs. 3 and 4 illustrate the three probe sequential method.

Fig. 5 illustrates a systematic error Fig. 6 illustrates a measurement. Figs. 7 A and 7B illustrate the method. Figs. 8 to 12 illustrate embodiments of the invention.

The present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which preferred embodiments of the present invention are shown. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiment set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Like numbers refer to like elements throughout.

Fig. 1 is a schematic drawing of a x-y moving machine 1. The machine has a moving part 2 for moving a part A over a table B in two perpendicular directions x and y. Such machine may be any kind of precision machinery. Machine tools and multi-axis machinery require a high standard precision in line with the development of high precision engineering. High precision in manufacturing can only be accomplished if it is possible to measure and calculate accurately the errors of machine components. The current solution to straightness calibration is to measure the 'individual machine fault' of the individual machine with (often) laser measurement tools and store it. Then all kind of efforts are made to keep conditions the same, so that the 'individual machine fault' stays the same. This requires a very good control over conditions such as temperature and humidity, the use of often (very) expensive materials such as Zerodur and invar to reduce as much as possible any deviation of the established 'individual machine fault'. Even then, the measurement procedure has to be repeated at regular intervals, e.g. bi-yearly, and after any service activity that could have affected the configuration.

In order to be able to measure accurately a gauge for calibration must be available. However, this gauge has to be measured also. Ultimately, therefore, the accuracy of the gauge (calibration) determines the accuracy of manufacturing.

Often a straightedge is used for straightness calibration. Fig. 2 shows schematically a position of such a straightedge 3.

To measure the straightness of the straightedge various methods are known, one of which is the so-called three probe method, illustrated in fig. 3. A measuring device 4 is used, in which three probes 4a, 4b and 4c are provided. The measuring device is moved along a guide way G(x) and at intervals the position S(x) is sequentially measured. The measured values are sent to and registered by a calculator C. Thus, sequentially measurements using three probes are taken, which is the reason that this method is called the sequential three probe method. A relatively large number of unknown parameters play a part. First of all the to be measured suria.ce S(x) is a priori unknown. Secondly the guide way G(x) is unknown, thirdly, the yaw angle γ of device 4 is unknown. The calculator calculates S(x) in dependence on the measured values. The "calculator" may, within the concept of the invention be any piece of hardware or software with which the necessary calculations may be done. At each point n, n+1, n+2, n+m etc. probe 4a measures a distance a(n) between the guide way G(x) and the surface S(x) at position n.

a(n)=G(n)-S(n).

Likewise probe 4b measures a distance b(n) between the guide way and the surface at position n+1 and probe 4c measures a distance c(n)

b(n)= G(n)-S(n+1)+Lγ(n), where γ(n) is the yaw error at position n c(n)=G(n)-S(n+2)+2Lγ(n).

Adding and subtracting gives:

a(n)-2b(n)+c(n)=-S(n)+2S(n+l)-S(n+2). a(n+l )-2b(n+l )+c(n+l )=-S(n+l )+2S(n+2)-S(n+3). a(n+2)-2b(n+2)+c(n+2)=-S(n+2)+2S(n+3)-S(n+4). Etc.

Through two points (for instance the starting points (S(I) and S(2)) a line can be drawn, and thus these points S(I) and S(2) can be set at zero. This will allow to establish all points S(n) using the above equations. It is remarked that the above equations are independent of the guide way G(x) and the yaw error γ(x).

Thus, by measuring with three probes at n points it is possible to determine the values of S(x) independent of the guide way G(x) and the yaw error γ(x). This determination is done in the calculator.

The sequential three probe measurement method is based on such calculations.

The known method is restricted due to the fact that the probes are separated at a distance L from each other. In principle the sampling interval (the distance between measuring points) and the distance between the probes has to be equal, which follows from the above equations. This restricts the spatial resolution of the measurements. To overcome this restriction Li et al has proposed a method in Measurement Technology and Intelligent Instruments, SPIE VoI 2101, pag 483-487, 'on-line measurement of the straightness of precision lathes'.

In this method different probe intervals L are used. Using different probe intervals L and mapping profiles together a higher spatial resolution is, according to Li et, al obtainable.

The known method, however, has a number of shortcomings. It requires either the use of several carriages with probes, each carriage having a different distance between the probes, or it requires, if one carriage is used to change regularly and accurately the distance between the probes. In both cases the method is cumbersome and time-consuming.

Both possible embodiments of the method proposed by Li et al have also a fundamental shortcoming, apart from the above cited practical problem.

If G(x) and γ(x) would be only unknown quantities, the sequential three (or more) probe method would indeed allow to eliminate the unknown quantities G(x) and γ to establish S(x) and thereby measure the straightness of the straightedge, which can then be used as a calibration for other straightness measurements for instance of a work piece.

This could then be done at various separations L which would enable to increase the spatial resolution by mapping the profiles together. However, the inventors have realized that a seemingly small systematic error occurs, which is illustrated in figure 5.

The centre probe 4b may be offset by an amount δ in respect of a straight line through the outer probes 4a and 4c. This offset is a fixed offset, independent of the value n. The above mentioned equations now become:

a(n)=G(n)-S(n). b(n)=G(n)-S(n+l) - δ +Lγ(n), c(n)=G(n)-S(n+2)+2Lγ(n).

Adding and subtracting gives:

a(n)-2b(n)+c(n)=-S(n)+2S(n+l)+δ-S(n+2). a(n+l )-2b(n+l )+c(n+l )=-S(n+l )+2S(n+2)+δ-S(n+3). a(n+2)-2b(n+2)+c(n+2)=-S(n+2)+2S(n+3)+δ-S(n+4).

Etc.

The unknown parameter δ influences the outcome of the equations. In fact an error in measurement occurs which is cumulative, i.e. whereas at the first point the error is small (δ) at the n⁴¹¹ point the error is approximately n(n-l)δ. Even though the systematic error itself δ may be small, the error in measurement may be large, due to the cumulative nature in which δ influences the outcome of the equations. The size of the device has a tendency to increase and the number of measurement points also, so that this systematic error Δ becomes appreciable. Figure 6 illustrates measured values (Δ+S(x)) as determined by the common sequential three probe method. The measured values actually comprise two components, namely the true values for S(x), i.e. the bends in the supposedly straight straightedge, and the systematic error Δ, which grows more or less in a quadratic form as a function of the measurement points n. One could presume that, knowing that the systematic error grows in a particular manner as a function of distance travelled along the straightedge one could isolate this systematic error. However, bends usually also follow a more or less quadratic curve, so that the two contributions cannot be separated, at least not easily. When the distance L is changed, between measurements either different carriages have to be used, or the same carriage has to be used, but the distance L between the probes has to be changed. Both possibilities are cumbersome and expensive, and in both cases the different profiles will be measured with different systematic errors δ, and thus the contribution Δ changes. In fact, if the distance is changed, the systematic error δ will change every time this is done in an a priori unknown manner, so that even a very accurate determination of δ for a particular arrangement would not help. Because the profiles for different probe separations L differ with an unknown quantity Δ for each of the profiles, mapping the profiles together is prone to error.

Figures 7A and 7B illustrate a method in accordance with the invention.

In figure 7 A one set of measurement points is taken, in figure 7B a second set of measurement point is taken wherein the starting point of the measurement is offset a distance O form the starting point of the first measurement. The outcome of the two sets of measurement points are two different curves Sl(x) and S2(x). In contrast to the method as proposed by Li et al the systematic error δ (in the probe) and Δ (in the curves) are the same, and thus the problem of a change in systematic error does not occur. However, since the two measurements do not have points in common, the two profiles Sl(x) and S2(x) cannot a priori be linked to each other. The concept of the invention is that, even though this is true in principle, it is possible to link the two profiles by assuming that, on average, the two profiles Sl(x) and S2(x) follow the same line, i.e. if a line ai +a₂x is drawn through the profiles, these two lines are the same. Thus the two lines are made to coincide. Doing so, it is possible to combine the two profiles and increase the spatial accuracy. Since the two profiles have, in contrast to the method proposed by Li et al one and the same systematic error Δ, a source of inaccuracy is eliminated.

In figure 7 A the offset O is 1/2L. It is preferred that, if a full measurement comprises p sets of measurements points, wherein each set of measurement points is offset with a distance O from a neighbouring measurement, then it holds (within practical accuracy) O=L/p. The measurement points of the different profiles are then as good as possible distributed over the straightedge, which increases the spatial accuracy.

Figure 8 and 9 illustrate the method and device in accordance with the a preferred embodiment of the invention. The device 4 with the probes 4a, 4b and 4c is moved along one surface S (x) of the straightedge 3, and measurements are taken. The device is then brought to the opposite suria.ce of the straightedge 3, to measure the opposite surface S'(x) of the straightedge 3.

The device 4 may, at opposite surfaces of the straightedge be oriented, such that probe 4a faces probe 4a or faces probe 4c, i.e. the same probes face each other, or different probes face each other. In figure 8 the same probes face each other, in figure 8 different probes face each other, i.e. the sequence of probes is reversed, "facing each other' means that when the centre probe is positioned at the same coordinate along the straightedge, the coordinates of the 4a probes are substantially the same (figure 8) or the coordinates of the 4a and 4c probes are substantially the same (figure 9).

Figure 10 illustrates the same measurements but now taken at the opposite surface of the straightedge 3. The measured values (Δ+S'(x)) actually comprise two components, namely the true values for S'(x), i.e. the bends at the opposite surfaces, and the systematic error Δ. The systematic error is due to the error δ which has not changed. The values for S'(x) are of opposite sign for those of S(x), since what was convex at one surface of the straightedge, is concave and vice versa at the other side of the straightedge. Again, using only the measurements of either side of the straightedge the two contributions cannot be separated. However, using both measurements, it is possible to separate the two contributions, i.e. the systematic error and the true curvature of the straightedge, to the measurements, as is illustrated in figures 11 and 12. Adding and subtraction the measurements gives, on the one hand the value for the systematic error Δ and on the other hand the value for the curvature S(x). This scheme works, provided that the thickness th (figure 8, 9) of the straightedge can be better controlled than the curvature (bends) of the straightedge. The latter is, however, almost always the case. The thickness of the straightedge can be controlled to a very high accuracy, and is also hardly dependent on other parameters such as temperature and humidity, that may have an appreciable influence on the straightness of the straightedge. It is to be noted, that one could think that the scheme could only work if the two guide ways (G(x)) and the yaw error (γ(x)) are the same for the opposite surfaces of the straightedge. This would, if true, pose a serious limitation on the method, since it is hardly likely that this would be the case. However, this is not the case, G(x) and γ(x) drop out of the equations at both sides of the straightedge 3, and independent of each other. It is essential, though, that the same probe is used at the opposite surfaces of the straightedge, and the position of the probes in respect to each other remains the same, so that the error δ remains the same. Using two different probes simultaneously at opposite surfaces does not fall under the scope of the invention, nor does using the same probe two at the same surface, but in different orientations.

In short the invention can be described as follows: In an apparatus for measuring position errors in a machine having a movable element (2) and a straightedge (3) and a system for measurement of the straightness of the straightedge, said measurement system comprises a multi-probe (4a, 4b, 4c) device (4) for sequentially measuring along the straightedge (3) using a carriage (4) moving along a guide way (G(x)). The carriage (4) is moved along one surface (S(x)) of the straightedge (3) to take a set of measurement points, whereafter a second set of measurement points, with an offset O, is taken, and the two sets of measurement points are mapped together.

It will be clear that within the framework of the invention many variations are possible. It will be appreciated by persons skilled in the art that the present invention is not limited by what has been particularly shown and described hereinabove. The invention resides in each and every novel characteristic feature and each and every combination of characteristic features. Reference numerals in the claims do not limit their protective scope. Use of the verb "to comprise" and its conjugations does not exclude the presence of elements other than those stated in the claims. Use of the article "a" or "an" preceding an element does not exclude the presence of a plurality of such elements.

For instance, in preferred embodiment a three-probe method is used, but more than three probes could be used. On the one hand an additional error would be introduced, however, also additional information would be available. In preferred embodiments a three probe method is used. A straightedge may be provided in one direction or two or more directions.

The method of the invention works, since the deviations in thickness are much better controllable than the deviations in straightness. The dimensions of the straightedge are for instance typically 5 mm (thickness) by 2-3 meters (length). The thickness of the straightedge can during manufacturing be controlled to within micrometers. One type of systematic error would be a systematic change in thickness of the straightedge along its length. Such a systematic, known change in thickness along the length of the straightedge can be accounted for in the measurements by accurately measuring the variation of thickness of the straightedge along the straightedge before putting it in the apparatus or while in the apparatus, and accounting for such variation when comparing the measurements at opposite surfaces of the straightedge. A simple look-up table comprising such systematic deviation as a function of position along the straightedge would suffice.

Deviations due to temporal influences on the thickness of the straightedge, such as temperature variations, would then not be accounted for, but such errors are orders of magnitude less than the deviations of straightness. The straightedge may have the form of a lath or a plate, or any suitable shape or form.

The probes do not need to be all on one side of the straightedge.

For instance, a number n (n≥l) of probes could be probing a first surface of the straightedge, while simultaneously m (m≥l), wherein n+m> 3, probes probe the opposite surface of the straightedge. The method would then comprise taking measurements along one direction, whereafter, the carriage is rotated so that the m probes probe the first surface, and the n probes probe the opposite surface, and the measurements are repeated.

In principle, when both sides of the straightedge are measured, either the carriage may be transferred to an opposite side of the straightedge leaving the straight edge in position, or the carriage is left in position and the straight edge is turned around so that the opposite surface faces the carriage.

The straight edge is convex or concave (i.e. having a form departing from a true straight edge) due to - an intrinsic curvature in the straightedge a curvatrure in the straightedge due to the mounting of the curvature.

Turning the straight edge around does not change the intrinsic curvature of the straightedge (and thus what was convex in respect of the carriage becomes concave and vice versa), however, the mounting of the straight edge may be changed due to the running around of the straightedge which may introduce a change in curvature and such change may lead to an error. Thus it is preferred that the straightedge is left unchanged and the carriage is transferred to the opposite surface of the straightedge.

The present method, apparatus and system is suitable for on-line measurement. An apparatus as described regularly travels along the straightedge, during such travels back and forth, the carriage, when it has reached a final position, is changed in position so that the probes which before such change were probing one surface, are thereafter probing the opposite surface. Provided that during such movement back and forth along the straightedge, the temporal influence, such as temperature and humidity do not change to much, the method allows on-line accurate calibration of straightness.

Claims

CLAIMS:

1. A sequential multi-probe method for measurement of straightness of a straightedge (3) using a multi-probe (4a, 4b, 4c) device (4) for sequential measurements along the straightedge (3) using a carriage (4) moving along a guide way (G(x)), wherein the carriage (4) is moved along a surface (S(x)) of the straightedge (3) to take a set of measurement points, and is subsequently moved again along the same surface to measure a set of measurements points in between the previously measured points, whereafter the sets of measurement points are mapped together by assuming that the average line through said sets of measurement points is the same.

2. A sequential multi-probe method as claimed in claim 1, wherein p sets of measurement points are taken and the distance between neighbouring points at different sets of measurement points is such that the distance between points taken at different sets is O=L/p where L is the distance between neighbouring probes in the carriage. In such embodiments the measurement points are equally distributed over the straightedge.

3. A sequential multi-probe method as claimed in claim 1, wherein measurement points are taken at opposite surfaces (S(x), S'(x)) of the straight edge (3).

4. A sequential multi-probe method as claimed in claim 3, wherein at opposite surfaces (S(x), S'(x)) of the straightedge (3) the carriage (4) is so oriented that the same probes (4a, 4b, 4c) face each other.

5. A sequential multi-probe method as claimed in claim 3, wherein at opposite surfaces (S(x), S'(x)) of the straightedge (3) the carriage (4) is so oriented that the sequence of probes (4a, 4b, 4c) is reversed.

6. A sequential multi-probe method as claimed in claim 1, wherein three probes (4a, 4b, 4c) are used.

7. An apparatus for measuring position errors in a machine having a movable element (2), a straightedge (3) and a measurement system for measurement of the straightness of the straightedge (3), said measurement system comprising a multi-probe (4a, 4b, 4c) device (4) for sequentially measuring along the straightedge (3) using a carriage (4) moving along a guide way (G(x)) and a calculator (C) for calculating the straightness of the straightedge (3), wherein the apparatus is arranged to move the carriage along a surface of the straightedge to take a set of measurement points, and to subsequently move the carriage again along the same surface to measure a set of measurements points in between the previously measured points, and to map the sets of measurement points together by assuming that an average line through said sets is the same.

8. An apparatus as claimed in claim 7, wherein the apparatus is arranged to take p sets of measurement points and wherein the distance between neighbouring points at different sets of measurement points is such that the distance between points taken at different sets is O=L/p where L is the distance between neighbouring probes in the carriage. In such embodiments the measurement points are equally distributed over the straightedge.

9. An apparatus as claimed in claim 7, wherein the apparatus has means to transfer the carriage (4) to an opposite surface (S'(x)) of the straightedge (3) and to move the carriage along the opposite surface of the straightedge (3) to take measurements.

10. A measurement system for measurement of the straightness of the straightedge (3), said measurement system comprising a multi-probe (4a, 4b, 4c) device (4) for sequentially measuring along the straightedge (3) using a carriage (4) moving along a guide way (G(x)) and a calculator (C) for calculating the straightness of the straightedge (3), wherein the apparatus is arranged to move the carriage along a surface of the straightedge to take a set of measurement points, and to subsequently move the carriage again along the same surface to measure a set of measurements points in between the previously measured points, and to map the sets of measurement points together by assuming that an average line through said sets is the same.

11. A measurement system as claimed in claim 10, wherein the measurement system is arranged to take p sets of measurement points and wherein the distance between neighbouring points at different sets of measurement points is such that the distance between points taken at different sets is O=L/p where L is the distance between neighbouring probes in the carriage. In such embodiments the measurement points are equally distributed over the straightedge.

12. A measurement system as claimed in claim 10, wherein the measurement system has means to transfer the carriage (4) to an opposite surface (S'(x)) of the straightedge (3) and to move the carriage along the opposite surface of the straightedge (3) to take measurements.