GB2065399A - Determining a dimensional parameter of a workpiece - Google Patents

Abstract

Transducers in contact with a workpiece produce electrical analogues A to E of the positions of end points of a dimension to be determined. The analogues are combined in a summing or difference circuit 5 or 6 to produce a derived analogue representative of a dimensional parameter pertaining to the dimension. An AD converter 8 converts the derived analogue into a corresponding digital signal, thus obviating the need for conversion of the individual analogues A to E. <IMAGE>

Description

SPECIFICATION Determining a dimensional parameter of a workpiece This invention relates to determining a dimensional parameter of a workpiece. By "dimensional parameter" is meant herein a parameter, apertaining to a dimension of a workpiece, which will often represent a deviation in a gauged value of that dimension from a nominal value but which could be the value of the dimension itself, i.e. the algebraic sum of the nominal value and the deviation existing, or could represent a mean position on that dimension or a deviation from nominal in such a mean position.

Following completion of manufacture of workpieces, it is common practice to check various geometric features for compliance with specified tolerance requirements. Particularly when gauging complex mechanical workpieces, it is often necessary to measure a large number of separate parameters, for example several diameters, lengths, widths, tapers, squareness of surfaces etc. and the like.

A technique, known as "multipoint gauging" has been devised for this purpose. The technique requires the study of the geometry of the workpiece to discern which points on the surface of the workpiece need to be examined to enable the required parameters to be computed. It will be appreciated that the measurement at any one point may be used in the compution of one or even several different parameters, and that any one parameter may require the examination of several points.

The measurement at each measuring point on the workpiece is made using an electrical analogue transducer contacting the workpiece at the point in question. Usually, there will be a separate transducer at each such point but in accordance with a modified technique the workpiece could undergo displacement, relative to a transducer, so as to bring one or more further measuring points to the transducer for measurement purposes. Thus, for example, a shaft accurately mounted for rotation about its axis could be rotated to check the roundness of the peripheral surface by taking readings with a single transducer from different pairs of diametrically positioned measuring points. In all cases, the output data from the transducer(s) is processed to determine the required parameters.Such processing involves subtracting or adding the transducer.output data associated with the various measuring pointS. For example, in the usual case where two transducers are placed in contact with the workpiece at opposite ends of a dimension to be determined so that they measure in opposite senses, dimensions or deviations from nominal value are generally determined as the algebraic sum of the respective transducer output data associated with the end points of the dimensions concerned. On the other hand deviations in mean positions or the mean positions themselves can involve determining the algebraic difference of such output data.

In the present state of the art, the several sets of transducer output data are presented, by means of suitable circuitry, as electrical analogues of the deviations of the various measuring points from nominal positionS. The various electrical analogues are individually and sequentially connected by electrical switches to known means for converting electrical analogue signals into a digital form in which they are then presented to the digital computer. The sequential operation of the said electrical switches is normally controlled by the digital computer.The computer then performs the required adding or subtracting operations and any further processing which may be required (for example multipiication of deviation in one nominal length by another nominal length to give an area deviation, or control sorting of the workpieces or quality control of the workpiece production line.) However, such techniques can be disadvantageous as the following example will illustrate. It will be appreciated that in some cases the tolerance applied to a given dimension may be small, for example + 0,01 Omm, while the position of the feature of the workpiece containing this parameter may change from workpiece to workpiece by as much as + 1,5mum. In normal practice, the gauging means should have an accuracy better than one-tenth of the tolerance of the parameter being measured.In the above example the accuracy should be better than j 0,001 mm.

This demands that the discrimination of the system must be better than 0,001 mm. The parameters measured are typically those of length and each will typically have a separate transducer at each end of the length to be measured. The discrimination required in the determination of the position of each end of the length must be better than 0.0005mm. It will be appreciated that the analogue to digital converting means for converting the said electrical analog signal into a digital form must be capable of accepting an electrical change equivalent to 3mm and have a discrimination equivalent to 0,0005mm. This can be expressed as a discriminaton of 1 bit in 6000 bits. Known converting (e.g. a thirteen bit analogue-to-digital converter) means which are capable of such fine discrimination are available but are expensive and slow in operation.

Furthermore, two analogue to digital conversions must be made to enable the dimensional parameter to be gauged.

According to the invention from one aspect there is provided a method of determining a dimensional parameter of a workpiece, wherein the end positions of a dimension under investigation of the workpiece are gauged to produce electrical analogues representative of said end positions, these analogues are combined to produce a derived analogue representative of a dimensional parameter pertaining to said dimension, and the derived analogue is converted into a corresponding digital signal.

According to the invention from another aspect there is provided apparatus for use in determining a dimension of a workpiece, comprising first means, including at least one transducer for positioning in contact with the workpiece, for enabling electrical analogues to be produced representative of the positions of the end points of said dimension, second means arranged to combine these analogues to produce a derived analogue representative of a dimensional parameter pertaining to said dimension, and an analogue-to-digital converter arranged to convert the derived analogue into a corresponding digital signal.

With such a method or apparatus for determining a dimensional parameter of a workpiece, the desired level of discrimination can be achieved, but requiring only a cheaply available analogue-to-digital converter. Essentially, this is because each individual derived analogue signal represents the dimensional parameter being gauged, which in general has a tolerance ascribed to it, and in practice it is usually necessary to consider deviations from the nominal length which only equal three or four times the ascribed tolerance. By way of illustration even a cheaply available 8 bit analogue-to-digital converter is capable of discriminating a deviation of say four times the tolerated deviation from nominal into 256 bits, and each of the bits then represents one sixty fourth part of the tolerance permitted.

It will be appreciated that not only is even a cheap 8 bit analogue-to-digital converter faster in action than say, a 1 3 bit analogue-to-digital converter required to give the necessary discrimination in the known technique referred to abc ze, but that with the invention only one such conversion is required, while by known means two such analogue-to-digital conversions are made to derive the length.

The digital signal is generally suppiied as an input signal to a digital computer programmed to process the signal in some desired manner. Usually, the first means will comprise several transducers and switching means operable, for example by control signals from the computer, to connect the several transducer outputs sequentially and in pairs to the second means, which can take the form of separate adding and subtracting circuits and a switch for selectively connecting the outputs of said circuits to the analogue-todigital converter. The computer will be programmed to perform the required computations from the several input digital signals and to display suitable output data (e.g. number of dimensions out of tolerance) and/or effect quality control of the production line or injection of unsuitable workpieces.

To ensure accuracy in the analogue-to-digital conversion particularly at the zero point, the converter is preferably of a kind arranged to produce a single digit to represent the sign of the input analogue and further digits to represent the modulus of the input analogue.

In a further development the electrical analogues from the transducers and/or the derived analogues are modified, prior to the analogue-to-digital conversion in accordance with a correction factor, for example to correct for errors due to thermal expansion and the like.

Features relating to the geometry of a workpiece can, in general, be determined by gauging of one length or a plurality of lengths.

These are generally linear lengths, but small angles can be checked by measuring a small length of arc (assumed to be linear) subtending the angle concerned. Further, any linear dimension is defined by the distance between its two ends, and not by the position of one end. Some geometric features can be measured by consideration of a single length, for example parameters such as thickness, diameter, length. Other geometric features demand consideration of two lengths, by example areas, taper, conicity. Squareness can be considered as such an example in that it can be found by consideration and measurement (in the manner described above) of the small angle by which one feature deviates from its nominal angular position, and by the small angle by which a second feature, which should be square to the first feature, deviates from its nominal angular position.Some geometric features may involve the measurement of a large number of lengths, for example the average diameter of a cylinder or the maximum thickness of a plate.

In all cases, as indicated above, the various computations such as multiplication of one length by another are carried out in the digital computer.

In the above description it has been assumed that the dimensional parameter being determined is one of size of a feature, where linear length is concerned. However, any method or apparatus in accordance with the invention is equally applicable where the position of a centreline, say of other feature is concerned. For example, if the width of a particular feature exhibits widely varying values, while the centreline of the width has a small tolerance, the derived analogue may be formed (usually by algebraic subtraction) from the analogues of the deviation from the nominal position of each end of the length so as to represent the position of the centrepoint. In this case it is this centre point displacement derived analogue which is presented to the analogue-to-digital converter.The discrimination of the analogue-to-digital converter is then applied to the tolerances of the centreline displacement and not to the larger variations analogues of the thickness deviation from nominal.

Of course, if both the width and the centreline disposition have to be investigated the formation of derived analogue will entail the summation of, and the differences between, the two transducer analogues defining the positions of the two ends of the length representing the width of the feature concernd.

For a better understanding of the invention and to show how the same may be carried into effect, reference will now be made by way of example to the accompanying drawings in which: Figure 1 is a circuit diagram of one embodiment of the invention; Figure 2 is a block diagram of a preferred form of analogue-to-digital converter used in the circuit, and Figure 3 is a diagram of a preferred circuit arrangement for part of the analogue-to-digital converter shown in Fig. 2.

Referring to Fig. 1 analogue signals A-E are individually produced by respective transducers (not shown) in contact with a workpiece at various measuring points or positions on the surface of the workpiece. Each signal (such as A) can be switched, by an appropriate analogue switching device (such as 1 A or 2A), to one or other of two electrically conducting bus lines 3, 4. A summing circuit 5 and a difference circuit 6 are each connected to receive the respective signals (such as A and B) on the two bus lines 3, 4 and, depending upon the state of an electronic switching device 7, the appropriate derived analogue signal (sum or difference) is supplied to a analogue-to-digital converter 8 via a switched gain amplifier 9 in which the derived analogue signal is amplified by a selected one of a plurality of different amplification factors (e.g. 2, 4, 8 and 16).The overall operation is controlled by a digital computer 1 0. Thus the computer controls a circuit 11 for operation the switching devices two at a time so as to select sequentially pairs of signals from the transducers for processing, a circuit 1 2 controlling the state of the switching device 7, a cicuit 1 3 determining the amplification factor of the amplifier 9, and a circuit 14 which presents the digital output of the converter 8 to the computer 10 for the required computation of workpiece geometry and for any further processing or action which may be required.

Where measurement is made of the deviation from nominal value of a dimension, the nominal value of the dimension is the data of measurement and represents a zero value analogue. Since the zero value analogue is the datum it should be the most accurately known value. With conventional 8 bit bipolar analogue-to-digital converter the passage of the analogue value through zero is indicated by the most significant bit (i.e. 10000000) and the accuracy is only that given for the overall circuit accuracy. This is typically + 2 bit.

In one preferred embodiment of the invention, a magnitude detecting circuit is used at the input to an 8 bit analogue-to-digital converter. This circuit detects the difference between the derived analogue and the reference zero analogue, and produces an analogue which always has the same polarity. At the same time it produces an output as a logic level to indicate to which side of zero the original analogue lay. This logic level is passed to the computer as an additional bit to indicated the sign of the derived analogue. It will be appreciated that when used with an 8 bit analogue to digital converter, the computer is now presented with a 9 bit code representing sign and magnitude, and as a result provides discrimination of 51 2 steps.It will be noted that the code presented to the computer is not that normally presented by an analogue-to-digital converter and the computing program may need some modification.

However, the preferred circuit and code have two advantages which become apparent if it is assumed that each bit from the analogue-todigital converter represents by example 1 micron. To move from a minus reading of 1 bit to a plus reading of 1 bit from the system would require a change in the derived analogue of the equivalent of 2 microns which is in agreement with the mechanical engineering tolerance of plus or minus one micron. However, this is not true for all codes used in conventional 8 bit bipolar analogue-to-digital converters. Further, as the derived analogue passes through a zero value, there is a complete change in the code at zero (i.e. in the particular bit indicating the sign of the input analogue signal). This allows preuse setting of the derived analogue zero, or nominal value.

A suitable form of such converter is shown in Fig. 2. Here, the converter comprises, in series, a magnitude circuit 1 5 and a converting circuit 1 6 both connected to a ground plate 1 7 at 0 volts potential, thereby setting the zero point at 0 volts potential. The circuit 1 5 receives the amplified sum or difference analogue signal (An) and accordingly supplies to the computer a signal digit output signal X to represent the sign of An and to the circuit 1 6 an analogue signal having a value equal to the modulus of An, this being converted in circuit 1 6 into a corresponding output digital signal Y which is an 8-bit signal in the illustrated example.

Fig. 3 shows a preferred form for the magnitude circuit 1 5. The input analogue signal An is fed to a voltage follower 20. This provides at point 21 the same signal An but ensures that current can be drawn from point 21 without affecting the magnitude of the signal at 21. Point 21 is connected to an input 22 of an operational amplifier 23 whose other input is connected back to the point 21.

Point 21 is also connected to an input of a further operational amplifier 24 whose other input is connected to the ground plate 1 7 and whose output is connected to the input 22 via diode 25 and also to a first input of another operational amplifier 26 whose other input is connected to the second input of amplifier 24 at zero potential. When the signal at 21 is positive, diode 25 is back-biased and no current flows into input 22. Amplifier 23 then merely transfers + An to its output. At the same time, the amplifier 26 supplies a positive signal X (i.e. "1"). When the signal on 21 becomes negative, the diode 25 conducts until the output voltage of the diode, which is fed back to the first input of amplifier 24, becomes equal to zero to match the voltage on its other input.Operational amplifier 23 then functions as an inverting amplifier so that again + An appears at its output. Accordingly amplifier 23 supplies An whatever the sign of the signal at point 21. On the other hand the amplifier 26 supplies 0 volts (i.e.

logical "0") as signal X.

An example will now be given to demonstrate how the desired level of discrimination can be achieved with the circuit arrangement according to Fig. 1 using an analogue-todigital converter as shown in Fig. 2. Consider the gauging requirement given above in connection with the known gauging technique.

The parameter of length to be measured is required to have a tolerance of + 0.101 mm.

In order that the maximum likely out-of-tolerance can be read, the analogue-to-digital converter would need to be able to accept a larger analogue equivalent than + 0.010 mm, a suitable range being + 0.040 mm say. For the reasons which have been given above, the preferred form of analogue to digital converter is one in which the analogue of the modulus of the value is extracted together with a digit representing the sign of the analogue value. A cheaply available known 8-bit analogue-to-digital converter (i.e.

having 28 = 256 bit positions) would be capable of discrimination to a value of 0.04 . 256 = 0.000156 mm. This represents less than 1% of the tolerance of + 0.010 mm and more than meets the gauging requirement. Of course, in general suitable amplification is given to the derived analogue so that a deviation from nominal of 0.040mm will produce the highest possible bit number which the analogue-to-digital converter can produce.

The factor chosen for amplification of the electrical analogue prior to conversion to digital form is determined by consideration of the engineering tolerance of the parameter being measured. As already indicated, it would normally be such that if the parameter was out of tolerance by an amount equal to three or four times (say) the tolerance, the electrical analogue would reach the limit at which the analogue-to-digital converter could operate. Alternatively, however, it could be chosen such that the change in the parameter necessary to cause a change in the least significant bit from the analogue-to-digital converter represents approximately 5% of the tolerance applied to the parameter.It will be appreciated that the digital computer can correct the digital expression of the deviation from the nominal value of the parameter to allow for any factor applied by amplification of the electrical analogue.

It will be appreciated that the analogue and digital signals handled are representative of deviations in gauged length from nominal values rather than the actual dimensions themselves. This is because generally transducers operating with the level of accuracy required for gauging workpieces exhibit linear operation only over a small range. Thus, if actual lengths need to be computed, the computer will combine stored values of length with the input digital signals from the analogue-to-digital converter. However, it may in some cases be feasible for the transducer output signals to be representative of actual lengths, in which case the computer would not have values of length stored therein.

Moreover, it may be appropriate in particular instances for the transducer analogue signals to be modified in accordance with appropriate cosine factors so that deviations along lines different from the measuring lines of action of the transducers can be gauged. Whilst the use of a separate transducer at each measuring point is to be preferred,the gauging of the geometry of the workpiece can be effected by a combinaton of the use of one or more transducers to take some of the measurements and effecting relative movement as between the workpiece and the transducer(s) to complete the taking of measurements.

It has been assumed hereinabove that all the transducers used, whose outputs are processed to investigate the geometry of the workpiece are engaged on the workpiece either directly or via suitable levers or other mechanical means which are to be considered as being part of the transducers. However, in embodiments of the invention further transducers may be used whose outputs represent measurements made of such distortions or malpositioning of the complete gauge construction so as to correct the computations made to allow for said distortions and malpositions. In a development one such transducer measuring temperature is accurately set in a suitable mount to give a zero analogue at, for example, 20"Centigrade and the analogue signal obtained from it is used to correct all other transducer analogues to allow for transducer errors resulting from a changing temperature and for temperature changes.

Claims (14)

1. A method of determining a dimensional parameter of a workpiece, wherein the end positions of a dimension under investigation of the workpiece are guaged to produce electrical analogues representative of said end positions, these analogues are combined to produce a derived analogue representative of a dimensional parameter pertaining to said dimension, and the derived analogue is converted into a corresponding digital signal.

2. A method according to claim 1, including the further step of processing said digital signal in a digital computer.

3. A method according to claim 1 or 2, including the step of selecting said electrical analogues as a pair from a plurality of electrical analogues from respective transducers.

4. A method according to claims 2 and 3, wherein said computer computes dimensional data from a plurality of said digital signals.

5. A method according to any preceding claim, wherein the or each said digital signal comprises one sign bit and a plurality of magnitude bits.

6. A method according to any preceding claim, including the step of modifying said electrical analogues in accordance with an error correction factor.

7. A method according to claim 6, wherein said factor is representative of temperature change.

8. Apparatus for use in determining a dimension of a workpiece, comprising first means, including at least one transducer for positioning in contact with the workpiece, for enabling electrical analogues to be produced representative of the positions of the end points of said dimension, second means arranged to combine these analogues to produce a derived analogue representative of a dimensional parameter pertaining to said dimension, and an analogue-to-digital converter arranged to convert the derived analogue into a corresponding digital signal.

9. Apparatus according to claim 8, further comprising a digital computer arranged to process said digital signal.

1 0. Apparatus according to claim 8 or 9, wherein said first means comprises a plurality of said transducers, and switching means operable to connect the transducer outputs sequentially to said second means.

11. Apparatus according to claim 8, 9 or 10 wherein said second means comprises an adding circuit and a subtracting circuit and a switch for selectively connecting the outputs of said circuits to said converter.

1 2. Apparatus according to claim 9, or to claim 10 or 11 as appendant thereto, wherein said computer is arranged to compute dimensional data from a plurality of said digital signals.

1 3. Apparatus according to any one of claims 8 to 12, wherein said digital signal comprises one sign bit and a plurality of magnitude bits.

14. Apparatus according to any one of claims 8 to 13, further comprising means for modifying said electrical analogues in accordance with an error correction factor.

1 5. Apparatus according to claim 14, wherein said factor is representative of temperature change.

1 6. A method of determining a dimensional parameter of a workpiece, the method being substantially as hereinbefore described with reference to Fig. 1, optionally as modified by Fig. 2 or Figs. 2 and 3, of the accompanying drawings.

1 7. Apparatus for use in determining a dimensional parameter of a workpiece, the apparatus being substantially as hereinbefore described with reference to Fig. 1, optionally as modified by Fig. 2 or Figs. 2 and 3, of the accompanying drawings.