WO2016139458A1

WO2016139458A1 - Calibration of dimensional measuring apparatus

Info

Publication number: WO2016139458A1
Application number: PCT/GB2016/050528
Authority: WO
Inventors: Kevyn Barry Jonas
Original assignee: Renishaw Plc
Priority date: 2015-03-02
Filing date: 2016-03-01
Publication date: 2016-09-09
Also published as: EP3265749A1; CN107532893A; JP2018511046A; GB201503490D0

Abstract

A dimensional measurement apparatus is further calibrated in order to improve an initial error map or error function of the apparatus. A calibrated production workpiece (14) is measured on the apparatus (10), and the measurements are compared (44,90) with calibrated values to produce error values. One or more updated error maps or functions are produced (46,92) from combinations of some or all of these error values with all or part of the initial error map or function. Then it is determined (48,94) whether an updated map or function produces better correction, and if so it is selected for future measurements. By selecting an error map or function which is based on a combination of new and existing error values, and which has been determined to give better correction, the improved map or function is learned intelligently during the course of normal production measurements.

Description

CALIBRATION OF DIMENSIONAL MEASURING APPARATUS

Field of the Invention This invention relates to the calibration of dimensional measuring apparatus. It may be used for calibrating coordinate measuring apparatus such as coordinate measuring machines (CMM), comparative gauging machines, machine tools, manual coordinate measuring arms and inspection robots. Description of Prior Art

After workpieces have been produced, it is known to inspect them on a coordinate measuring apparatus (such as a CMM or a comparative gauging machine) having a movable member supporting a probe, which can be driven within a three- dimensional working volume of the machine.

The CMM (or other coordinate measuring apparatus) may be a so-called Cartesian machine, in which the movable member supporting the probe is mounted via three serially-connected carriages which are respectively movable in three orthogonal directions Χ,Υ,Ζ. This is an example of a "serial kinematic" motion system. Alternatively, the measuring apparatus may be a non-Cartesian machine, for example having a "parallel kinematic" motion system comprising three or six extensible struts which are each connected in parallel between the movable member and a relatively fixed base member or frame. The movement of the movable member (and thus the probe) in the Χ,Υ,Ζ working volume is then controlled by coordinating the respective extensions of the three or six struts. An example of a non-Cartesian machine is shown in International Patent Applications WO 03/006837 and WO 2004/063579. It is known to calibrate such coordinate measuring apparatus by producing an error map or error function relating to the measurement errors experienced at different places within its Χ,Υ,Ζ working volume. This error map or error function is then used to correct measurements made on workpieces.

Our International Patent Application WO 2013/1401 18 describes a method in which such an error map or error function is built up gradually, as series of nominally identical workpieces are measured by the measuring apparatus during normal production use.

In WO 2013/140118, a first series of nominally identical workpieces includes one workpiece which is calibrated. Error values are determined by comparing the measurement values of that workpiece with its corresponding calibration values. These error values are not merely used to correct the measurements of further workpieces in the same series. They are used to populate the error map or error function of the measuring apparatus, which can then be used when other, different workpieces are measured.

Furthermore, subsequent series of nominally identical workpieces, which are different from the first series, also include workpieces which have been calibrated. Error values from measurements of these calibrated workpieces of subsequent series are also used to further populate the error map or error function of the measuring apparatus.

One advantage of the method described in WO 2013/140118 is that it is not necessary to carry out a full calibration of the apparatus to produce an error map over its entire working volume, which is normally a time-consuming operation, perhaps taking several days. Instead, the apparatus "learns" its error map over time, during its normal day-to-day use for measuring workpieces. Whenever a further calibrated workpiece is measured, the error values derived from it are used to further populate the error map or error function. This improves the calibration of the measuring apparatus. Nevertheless, a problem exists to improve the calibration even further. Summary of the Invention

The present invention provides a method of further calibrating a dimensional measuring apparatus which is calibrated by an initial error map or error function, the method comprising:

measuring a production workpiece on the measuring apparatus;

comparing the measurements of the production workpiece with calibration values for the production workpiece, obtained from a source external to said measuring apparatus, to produce one or more error values;

determining one or more updated error maps or error functions which combine some or all of the error values with all or part of the initial error map or function;

characterised by:

determining whether one of the updated error maps or error functions gives better correction of measurement errors than the initial error map or error function; and

if an error map or error function is determined to give better correction, then selecting that error map or error function for use in correcting the

measurements of one or more further workpieces.

At least in preferred embodiments, the method of the invention thus selects an error map or error function which is based on a combination of error values, rather than blindly incorporating all error values into the error map or error function. The combination of error values has been determined to give better correction of errors than would otherwise be the case.

The production workpiece may be one of a first series of nominally identical workpieces produced by a production process. The initial error map or error function may have been performed in a conventional manner, or it may have been produced by comparing measurements of a calibrated workpiece with

corresponding calibration values. Or it may have been produced by an earlier iteration of a method according to the present invention. Thus, the apparatus may "learn" its error map over time, during its normal day-to-day use for measuring workpieces.

One or more further workpieces may be measured, and the measurements thereof corrected using the selected error map or error function. The one or more further workpieces may include production workpieces from the first series of nominally identical workpieces. And/or the one or more further workpieces may include production workpieces from a second series of nominally identical workpieces produced by a production process, which are different from the workpieces of the first series.

The present invention also provides software programs for performing the above methods, a controller for a dimensional measurement apparatus incorporating such a software program, and dimensional measuring apparatus programmed with such a software program. Software programs may be recorded on machine readable media such as discs or memory devices, or stored on a remote server for downloading.

An "error map" as discussed in this specification may include, for example, a look-up table of values for the correction of subsequent measurements.

Brief Description of the Drawings

Preferred embodiments of the invention will now be described by way of example, with reference to the accompanying drawings, wherein:

Fig 1 shows operative parts of a comparative gauging machine with a motion system using parallel kinematics;

Figs 2 and 3 are flow charts of two preferred methods of calibration of a machine of Fig 1; and

Fig 4 is a flow chart giving more detail of part of the methods of Figs 2 and 3. Description of Preferred Embodiments

Fig 1 is an illustration of parts of a coordinate measuring apparatus. The apparatus is a comparative gauging machine 10 as sold by the present applicants Renishaw pic under the trademark EQUATOR. It comprises a fixed platform 30 connected to a movable platform 32 by a parallel kinematic motion system. In the present example, the parallel kinematic motion system comprises three struts 34 which act in parallel between the fixed and movable platforms. The three struts 34 pass through three respective actuators 36, by which they can be extended and retracted. One end of each strut 34 is mounted by a universally pivotable joint to the movable platform 32, and the actuators 36 are likewise universally pivotably mounted to the fixed platform 30.

The actuators 36 each comprise a motor for extending and retracting the strut, and a transducer which measures the extension of the respective strut 34. In each actuator 36, the transducer may be an encoder comprising a scale and readhead, with a counter for the output of the readhead. Each motor and transducer forms part of a respective servo loop controlled by a controller or computer 8. The parallel kinematic motion system also comprises three passive anti-rotation devices 38, 39 which also act in parallel between the fixed and movable platforms. Each anti-rotation device comprises a rigid plate 39 hinged to the fixed platform 30 and a parallel, spaced pair of rods 38 which are universally pivotably connected between the rigid plate 39 and the movable platform 32. The anti- rotation devices cooperate to constrain the movable platform 32 against movement in all three rotational degrees of freedom. Therefore, the movable platform 32 is constrained to move only with three translational degrees of freedom X, Y, Z. By demanding appropriate extensions of the struts 34, the controller/computer 8 can produce any desired X, Y, Z displacement or X, Y, Z positioning of the movable platform.

The principle of operation of such a parallel kinematic motion system is described in our US Patent No. 5,813,287 (McMurtry et al). It is an example of a tripod mechanism (having the three extending struts 34). Other motion systems e.g. with tripod or hexapod parallel kinematic mechanisms can be used. Taken together, the transducers of the three actuators form a position measuring system. This determines the X, Y, Z position of the movable platform 32 relative to the fixed platform 30, by appropriate calculations in the controller or computer 8. These calculations are known to the skilled person. Like all measuring apparatus, however, the position thus determined by the position measuring system is subject to errors. Methods are discussed below for calibrating the position measuring system for these errors.

Typically an analogue probe 16 having a deflectable stylus 20 with a workpiece contacting tip 22 is mounted on the movable platform 32 of the machine, although other types of probes (including touch trigger probes) may be used. The machine moves the probe 16 relative to a workpiece 14 on a table 12 in order to carry out measurements of features of the workpiece. The X, Y, Z position of a point on the workpiece surface is derived by calculation from the transducers in the servo system, in conjunction with the outputs of the analogue probe 16. This is all controlled by the controller/computer 8. Alternatively, with a touch trigger probe, a signal indicating that the probe has contacted the surface of the workpiece freezes the X, Y,Z position value calculated from the output from the transducers and the computer takes a reading of the coordinates of the workpiece surface. If desired, for gauging operations during normal production use, automatic means such as a robot (not shown) may place each of a succession of substantially identical workpieces from a production run in at least nominally the same position and orientation on the table.

The parallel kinematic measuring apparatus of Fig 1 is only one example of a type of measuring machine which can be used in the present invention. Other examples include measuring apparatus with serial kinematic motion systems, such as a conventional Cartesian CMM with three serially-connected carriages which are movable orthogonally in XYZ directions. This could be computer controlled or manually operated. Another possible serial kinematic machine is an inspection robot or a manual articulating arm, with multiple articulating arm members connected serially by multiple rotary joints. Whichever type of machine is used, typically it is placed in a workshop environment in order to inspect production workpieces from an automated manufacturing process.

In use, the controller or computer 8 in Fig 1 contains a program which causes the probe 16 to scan the surface of the workpiece 14. Or for a touch trigger probe it causes it to contact the surface of the workpiece at a plurality of different points, sufficient to take all the required dimensions and form of the workpiece for the inspection operation required. This controller/computer may also be used to run programs which control the calibration methods which will be described below. The calibration methods will be described with reference to the comparative gauging machine 10 of Fig 1, but the same methods can be performed on other measuring apparatus such as the serial kinematic machines mentioned above.

Fig 2 illustrates a first example of such a calibration method. The machine 10 has an initial error map or error function, derived by an initial calibration which is performed in a conventional manner in step 40. This may be a preliminary step performed by the manufacturer of the machine, before or during its installation at the user's premises. Because it may not be part of the method performed by the user, step 40 is shown in broken lines. However, it is also possible for this initial calibration to be performed by the user after installation of the machine.

For the conventional initial calibration in step 40, typically the machine is used to make numerous measurements of dimensionally calibrated reference standards, at numerous locations in the working volume of the machine. The reference standards are preferably calibrated in a manner which is traceable to appropriate national or ISO standards. They may for example be ring gauges, reference spheres, gauge blocks such as length bars or step gauges, straight edges, etc. Or another calibration artefact may be used, such as a "forest of balls", comprising a number of spheres mounted to a base plate fixture on stems or stalks. These spheres are accurately spherical, have accurately known dimensions, and they are mounted so as to be spaced in three dimensions with accurately known

relationships to each other. The fixture is placed in the working volume of the coordinate measuring apparatus and the spheres are measured using the apparatus to move the probe. By comparison with the known dimensions and spacings of the spheres, this produces a coarse map of the measurement errors experienced at a grid of points spread over part or all of the Χ,Υ,Ζ working volume of the machine. It is also possible to make measurements using a telescoping ball bar or a laser interferometer as a reference standard, as is conventional.

The initial error map in step 40 comprises first error values derived by comparing such measurements to the corresponding known calibrated values of the reference standards, at various locations within the machine' s working volume.

Alternatively an initial error function may be derived from such error values. The initial error map (and the other error maps discussed in this specification) can be created as a lookup table which indicates errors in the X, Y and/or Z directions for a given X, Y, Z coordinate position in the working volume of the machine. An error function may for example be a polynomial function which enables the calculation of errors in the X, Y and/or Z directions for a given X, Y, Z coordinate position.

The initial calibration need not be to a high accuracy, and it may not cover all locations within the working volume of the machine. The purpose of the following steps is to further calibrate the machine, improving the error map or error function.

In step 42, a calibrated workpiece is placed on the table 12 of the machine 10, as shown at 14 in Fig 1. The calibrated workpiece is one of a first series of nominally identical workpieces received from a production process, which are to be measured on the machine as part of an inspection process. By way of example, the workpieces in the first series might be con rods (connecting rods) for an automotive internal combustion engine.

Suitably the calibration of the calibrated workpiece of the first series (e.g. a con rod) may have been performed by measuring all its desired dimensions which are to be inspected, for example on a separate, more accurate coordinate measuring machine (CMM). This produces a set of calibrated values for the workpiece. The more accurate CMM may be located in a laboratory environment, whereas the machine 10 of Fig 1 could be located on the production floor, close to the machine tools or other production machines which manufacture the workpieces.

During the measurement in step 42, all of the dimensions to be inspected of the calibrated workpiece (e.g. con rod) are measured again on the machine 10, in the conventional manner by moving the probe 16 around the workpiece. This produces a set of raw measurement values, corresponding to the calibrated values. In step 44, the raw measurement values are compared to the corresponding calibrated values, producing a second set of error values. Both the raw

measurement values (from step 42) and the second error values (step 44) are stored by the computer or controller 8.

It will be appreciated that the calibration of the workpiece may take place after it has been measured on the machine 10 in step 42, rather than before. This still produces calibrated values which are compared to raw measurement values in step 44, to produce the second set of error values.

In step 46, a second error map or error function is created from a combination of some or all of the first and second error values, stored in steps 40 and 44.

Alternatively, if the initial calibration produced an error function and no initial error map is available, then error values may be synthesised from the error function and combined with some or all of the second error values. As is well understood by a skilled person, algorithms may be applied to remove outliers in the error values, or to average or weight some of the values. In practice, it may be desirable to produce not just a single instance of such a second error map or error function, but multiple further error maps or error functions. These are produced in step 46 from multiple different combinations of some or all of the available error values.

The second error map or error function may in practice give better or worse results than the initial error map or function of step 40. That is, when measurements are corrected using the second error map or error function, the results may be more or less accurate than when corrected using the first error map or error function.

Likewise, if there are multiple further error maps or error functions, one may give better results than another.

In step 48, therefore, it is determined which of the error maps or error functions (which combinations of error values) gives the best results. This is described in more detail below, with reference to Fig 4.

The error map or error function thus determined is selected for subsequent use in measuring production workpieces (step 50). For example, further workpieces from the first series of nominally identical production workpieces (e.g. con rods) are placed on the table 12 of the machine 10 (Fig 1). These workpieces are not calibrated, but their dimensions to be inspected are merely measured using the probe 16, giving corresponding raw measurement values. The raw measurement values are then corrected by applying the selected error map or error function. It is also possible to use the selected error map or error function to correct measurements of different workpieces, such as a piston for an automotive internal combustion engine.

As indicated at step 52, when it is desired to manufacture and inspect some different series of nominally identical workpieces (e.g. pistons or crankshafts for an automotive internal combustion engine), then steps 42-50 are repeated. One workpiece of the new series is calibrated and measured, as in step 42, and the raw measurement values are stored in the computer 8. By comparing these raw measurements with the calibrated values (step 44), further error values are created. A further error map or error function is created (step 46) by combining some or all of these error values with error values from any of the previous error maps or functions. In step 48 a choice is made as to which error map or error function should be used for future inspection of production workpieces, as described below with reference to Fig 4. This choice can select from any of the available error maps or functions, including the initial map or function from step 40, and those produced in step 46 using combinations of error values from various workpieces.

Note that the further error map or error function will preferably combine some or all of the error values from each location or orientation, in order to maximise the coverage of the working volume of the machine. This further error map or error function is then tested in step 48 to see whether it gives better results and should be selected for future use.

The above method of Fig 2 starts from a conventional initial calibration of the machine (step 40). Referring to Fig 3, a method will now be described which does not require a conventional initial calibration. This method may also be used in combination with the Fig 2 method, for subsequent improvement of the machine's error map or error function.

Steps 80 and 82 of Fig 3 are similar to steps 42 and 44 of Fig 2. In step 80, a calibrated workpiece (such as a con rod) is placed on the table 12 of the machine 10. The workpiece (e.g. con rod) has been calibrated as described above in relation to Fig 2, and it is now measured on the machine 10 giving raw

measurement values. These are compared to the corresponding calibrated values in step 82, producing a first set of error values. Both the raw measurement values (step 80) and the first error values (step 82) are stored by the computer or controller 8.

In step 84, a first error map or error function of the machine 10 is created from a combination of the first set of error values. This may then form an initial error map or error function, which will be used in a manner comparable to the error map or function of step 40 of Fig 2. If this is the first calibration of the machine, then all the error values may be used. If there is already a previous conventional initial calibration, then the first error map or function might be formed from a combination using only some of the error values, as in step 46 of Fig 2. As previously, algorithms may be applied to remove outliers in the error values, or to average or weight some of the values. Next, in step 86, the method continues with normal production measurements of the remainder of the first series of nominally identical workpieces (e.g. con rods), as they are manufactured. These workpieces are not calibrated, but their dimensions to be inspected are merely measured on the machine 10 of Fig 1, giving corresponding raw measurement values. These raw measurement values are then corrected by applying the error map or error function created in step 84.

At some future time, it is desired to use the machine 10 to measure a different, second series of nominally identical production workpieces. By way of example, the workpieces of the second series might be pistons for an automotive internal combustion engine. A calibrated workpiece (e.g. a piston) from the second series is placed on the table 12 of the machine 10. It is calibrated in the same way as above, by measuring all the desired dimensions to be inspected, e.g. on a separate, more accurate CMM, producing a set of calibrated values. In step 88, all the dimensions to be inspected of the calibrated workpiece (e.g. piston) of the second series are measured again on the machine 10, producing a set of raw measurement values corresponding to the calibrated values. In step 90, the raw measurement values are compared to the corresponding calibrated values, to produce a second set of error values. As previously, both the raw measurement values (step 88) and the error values (step 90) are stored by the computer or controller 8. Again as previously, the calibration of the workpiece (e.g. piston) may take place after the measurements on the machine 10, rather than before. In step 92, a second error map or error function is created from a combination of some or all of the error values stored in steps 82 and 90. As previously, error values may be synthesised from an error function if necessary, e.g. if they were not stored in step 82. Again, algorithms may be applied to remove outliers in the error values, or to average or weight some of the values. As in step 46 of Fig 2, it may be desirable to produce multiple further error maps or error functions, from multiple different combinations of some or all of the available error values. As in Fig 2, these second or further error maps or error functions may in practice give better or worse results than the first error map produced in step 84. That is, the results may be more or less accurate than when corrected using the first error map or error function. In step 94, therefore, it is determined which of the error maps or error functions gives the better results. As for the corresponding step 48 in Fig 2, this is described in more detail below, with reference to Fig 4. The error map or error function thus determined is selected for subsequent use in measuring production workpieces.

Next, in step 96, the method continues with normal production measurements of the remainder of the second series of nominally identical workpieces (e.g.

pistons), as they are manufactured. As above, these workpieces are not calibrated, but their dimensions to be inspected are merely measured on the machine 10 of Fig 1, giving corresponding raw measurement values. These raw measurement values are then corrected by applying the error map or error function selected in step 94.

As indicated at step 98, when it is desired to manufacture and inspect some different, third or subsequent series of nominally identical workpieces (e.g.

crankshafts for an automotive internal combustion engine), then steps 88-96 are repeated. This creates further error maps or error functions. In step 94 a choice is made as to which should be used for future inspection of production workpieces, as described below with reference to Fig 4.

At step 52 in Fig 2 and at step 98 in Fig 3, it is suggested to repeat the procedure with a different calibrated workpiece. However, rather than measuring a different calibrated workpiece from a new series of nominally identical workpieces, it is possible to repeat the measurements of some or all of the dimensions to be inspected of a previous calibrated workpiece, but located in a different position and/or orientation on the machine 10. For example, the calibrated con rod previously used in step 42 (Fig 2) or step 80 (Fig 3) could be measured again in a different position or orientation. This produces further error values which are stored in step 44 or 90, and which may then be used to create a further error map or error function (step 46 or 92). Note that the further error map or error function will preferably combine some or all of the error values from each location or orientation, in order to maximise the coverage of the working volume of the machine. This further error map or error function is then tested in step 48 or 94 to see whether it gives better results and should be selected for future use.

In step 46 (Fig 2) and steps 84 and 92 (Fig 3), error maps or error functions are created from combinations of some or all of the error values stored in steps 40 and 44 or 82 and 90 (possibly including error values synthesised from an error function). It would be possible to create an error map or error function which merely combined all of the available error values. However, the purpose of the determination at step 48 or 94 is to find a combination of the error values which produces good results (more accurate correction of the raw measurement values), possibly also removing outliers in the sets of error values. For this, it is desirable to produce multiple error maps or error functions, from numerous different combinations of the available error values. For each error map or function, a combination is made from a different sub-set comprising only some of the available error values. The error values of the initial or first error map (or synthesised from the initial or first error function) may be combined with only some of the second error values produced in step 44 or step 90. Or error values from only a part of the initial/first error map may be combined with some or all of the second error values.

Thus, the determination which takes place in step 48 or 94 can select from numerous such error maps or error functions, created from numerous different combinations of the error values. If sufficient computing power and time is available, it would be possible to create and use error maps or error functions from all possible combinations of the error values. Alternatively, to save computing resources, combinations may be chosen selectively, for example favouring combinations which have a denser spread of error values (and/or lower error values) in a central zone of the machine's working volume, where most measurements take place.

Fig 4 illustrates a method which can be used at step 48 of Fig 2 or in step 94 of Fig 3, in order to determine which of two or more error maps or error functions should be selected for future production measurements.

In step 60, the method takes raw measurement values of the calibrated workpieces as stored in step 42 (Fig 2) or in steps 80 and 88 (Fig 3). It also takes the first error map or error function, i.e. the initial error map or error function (Fig 2) or the error map or error function which has been created in step 84 (Fig 3). It uses this error map or error function to correct the raw measurement values. Where possible, it is preferable to operate on raw measurement values from more than one of the calibrated workpieces. Or, if the raw measurement values come from one particular calibrated workpiece, they may be corrected using an error map or error function which derives wholly or in part from a different calibrated workpiece.

In step 62, the accuracy of the correction performed in step 60 is assessed. This may be done by calculating a set of residuals between the corrected results and the corresponding calibration values. In steps 64 and 66, the steps 60 and 62 are repeated, using a second, different one of the error maps or error functions created in steps 46 and 92. This gives a set of residuals which assess the accuracy of the second error map or error function. As indicated at step 68, steps 64 and 66 may be repeated for the other error maps or functions created in steps 46 and 92, giving respective further sets of residuals.

Then, in step 70 a decision is made as to which of all the tested error maps or error functions gives the best results. This may be an automatic decision by the computer or controller 8, based upon which error map or error function gives the lowest residuals in steps 62, 66. For example, the sets of residuals for each error map or function may be compared by a least squares calculation, i.e. determining which set of residuals has the lowest sum of its squares. If desired, a weighted least squares method may be used, for example giving greater weight to residuals in a central zone of the working volume of the machine where most measurements take place.

Alternatively, step 70 may present the residuals calculated in steps 62, 66 to a skilled operator, e.g. as a display on a computer screen, and invite him/her to select a preferred error map or error function from those tested. This enables the operator to take into account other factors when selecting an error map or a function. For example, one of the error maps or error functions may give slightly poorer residuals over the entire working volume of the machine, but could be selected because it has better residuals in a central zone where most measurements take place. It is possible to store multiple error maps or functions, and

subsequently to select an appropriate one of them depending on the measurement requirements of a particular workpiece or series of workpieces to be measured.

If the residuals are to be presented to an operator, they may be processed into a suitable form to assist his or her selection. For example, they may be presented as a "heat map" (a 2D or 3D graphical representation in which the values of individual residuals are represented as colours, e.g. red for large residuals, yellow/orange for medium residuals, green for small residuals).

Error maps or error functions may have been derived from measurements of specific workpieces in specific locations in the machine's working volume (e.g. a con rod in one location, a piston in a second location, and a valve housing in a third location). In this case their heat maps may appear as coloured graphical representations of the workpieces concerned in their respective locations. If the operator knows that the machine will be used to measure both pistons and valve housings in the near future, he/she may decide to select an error map or error function which offers an acceptable compromise for both, rather than the best error map/function for pistons or the best for valve housings.

Finally, in step 72, the error map or error function that is determined is selected for use in future production measurements which take place in step 50 (Fig 2) or steps 86, 96 (Fig 3).

Thus, in the preferred methods described above, the apparatus "learns" its underlying error map or error function over time, during its normal day-to-day use for measuring workpieces. The error map or error function is based on combinations of error values which have been determined to give better correction of errors than would otherwise be the case. During use as a comparative gauging machine, the comparison of a specific workpiece against a corresponding calibrated workpiece takes place on top of this underlying error map/function. Eventually, the operator may have sufficient confidence in the accuracy of the underlying error map/function that he/she decides to use the machine to measure absolute coordinates and dimensions, in the traditional manner of a coordinate measuring machine, rather than just for comparative gauging measurements.

The preferred methods described above may be combined with the techniques described in our International Patent Applications Nos. WO 2013/021157,

WO 2013/140118 or WO 2014/181134. Those techniques produce temperature- dependent error maps or error functions. In the same way, the error maps or functions produced in Figs 2-4 above may be dependent on temperature. For example, in steps 42, 80 and 88, the temperature of the calibrated workpiece may be measured when the calibrated workpiece is measured on the apparatus of Fig 1. This temperature value is stored with the corresponding error values in steps 44, 82 and 90. Then, in steps 46, 84 and 92, combinations of error values are chosen which relate to the same or a similar temperature (to within a pre-determined temperature tolerance). This produces a set of error maps or functions which relate to respective temperatures. When production workpieces are measured, their temperature is monitored, and the appropriate error map or function is used to correct the measurements.

Claims

1. A method of further calibrating a dimensional measuring apparatus which is calibrated by an initial error map or error function,

the method comprising:

measuring a production workpiece on the measuring apparatus, the production workpiece being one of a first series of nominally identical workpieces produced by a production process;

determining one or more updated error maps or error functions which combine some or all of the error values with all or part of the initial error map or the initial error function;

characterised by:

determining whether one or more of the updated error maps or error functions gives better correction of measurement errors than the initial error map or error function; and

measurements of one or more further workpieces.

2. A method according to claim 1, wherein the one or more updated error maps or error functions are determined by combining only some of the error values with all or part of the initial error map or with the initial error function; or by combining the error values with only part of the initial error map.

3. A method according to claim 1 or claim 2, including measuring one or more further workpieces, and correcting the measurements thereof using the selected error map or error function.

4. A method according to any one of the preceding claims, wherein the one or more further workpieces include production workpieces from the first series of nominally identical workpieces.

5. A method according to any one of the preceding claims, wherein the one or more further workpieces include production workpieces from a second series of nominally identical workpieces produced by a production process, which are different from the workpieces of the first series.

6. A method according to any one of the preceding claims, further including: measuring a production workpiece from a second series of nominally identical workpieces on the measuring apparatus;

comparing the measurements of the production workpiece from the second series with calibration values therefor, obtained from a source external to said measuring apparatus, to produce one or more further error values;

determining one or more further updated error maps or error functions which combine some or all of the further error values with all or part of a previously determined error map or function;

determining whether one or more of the further updated error maps or error functions gives better correction of measurement errors than a previously determined error map or error function; and

measurements of one or more further workpieces.

7. A method according to claim 6, wherein the workpieces of the second series are different from the workpieces of the first series.

8. A method according to claim 6, wherein the workpieces of the second series are differently located on the apparatus from the workpieces of the first series.

9. A method according to any one of the preceding claims, wherein the initial error map or error function is produced by measuring a production workpiece on the measuring apparatus; and comparing the measurements of the production workpiece with calibration values for that production workpiece, obtained from a source external to said measuring apparatus.

10. A method according to claim 9, wherein the production workpiece used to produce the initial error map or error function is one of said first series of nominally identical workpieces.

11. A method according to claim 9, wherein the production workpiece used to produce the initial error map or error function is one of a series of nominally identical workpieces which are different from the workpieces of the first series.

12. A software program for a dimensional measuring apparatus, the software program being configured to cause the measuring apparatus to perform a method according to any one of the preceding claims.

13. A software program for a dimensional measuring apparatus according to claim 12, configured to cause the measuring apparatus to perform the steps of determining and selecting one of the updated error maps or error functions by presenting information about the error maps or error functions to an operator, and receiving a selection of an error map or error function from the operator.

14. A controller for a dimensional measuring apparatus, incorporating a software program according to claim 12 or claim 13.

15. A dimensional measuring apparatus, programmed with a software program according to claim 12 or claim 13.