Method and apparatus for machine absolute positioning measurement

Abstract

A method and apparatus for absolute position measurement for numerically controlled machines is provided. The invention produces a unique measurement signal for all locations within the range of a movable machine member using two measuring devices. Each measuring device produces a unique measurement signal within a portion of the range of motion that cyclically repeats over the entire range of motion. The measuring devices are so arranged that the measurement signal of one completes no more than one more cycle over the range of motion than the measurement signal of the other. The difference between the two measurement signals is unique for all locations and is used to produce an absolute position signal.

Description

BACKGROUND

This invention relates generally to position measurement for numerically controlled machines. In particular, this invention relates to the measurement of the absolute position of a movable machine member relative to a fixed reference location.

It is desirable for numerical control of movement of machine members to produce position measurement signals which are unique for all locations within a range of movement. This type of measurement, called herein absolute position measurement, has the advantage of providing precise information as to the location of a movable member within its range irrespective of whether the location has been recorded and maintained during a period when the member is not subject to numerical control, such as during maintenance of the numerical control device.

Absolute position measurement for movable members of numerically controlled machines is limited by the opposing requirements of resolution and axis range. Absolute position measuring systems are known for full machine range using plural measuring devices, each measuring device operating at a different resolution and each serving to provide unique measurement over a portion of the machine axis range. When such systems are implemented using measuring devices incorporating a rotating element, substantial gear reductions are required from the machine member drive to the measuring device used to produce the coarsest measurement resolution. When such systems are implemented using linear measuring devices the number of scales required to produce unique measurement signals over the full range of the machine axis becomes excessive.

A further disadvantage of the known absolute measuring systems, is their susceptibility to errors attributable to the measuring devices themselves or to the mechanical drive mechanisms associated therewith. The errors arising from these sources have an adverse affect on the overall machine positioning accuracy.

It is, therefore, one object of the present invention to provide an absolute position measuring apparatus and method operative over the full range of motion of a movable machine member and requiring no more than two measuring devices, each measuring device producing a unique measurement signal which is repeated over the range of the machine member motion.

It is a further object of the present invention to provide an absolute position measuring apparatus and method using two cyclic measuring devices and which is tolerant of errors of the relative position measurement made by the two measuring devices.

It is a still further object of the present invention to provide an absolute position measuring apparatus and method operative over the full range of motion of a movable machine member using a master measuring device and a vernier measuring device, each measuring device producing a cyclic measurement signal as the machine member moves over its range, the vernier cyclic measurement signal being repeated no more than once more within the range of motion than the master cyclic measurment signal.

It is a still further object of the present invention to provide an absolute position measuring apparatus operative over the full range of motion of a movable machine member having two measuring devices each having a rotatable element and each producing cyclic measurement signals, the rotating elements being driven relative to one another so that one measuring device produces no more than one more cycle of its measurement signal than the other measuring device over the range of axis motion to be measured.

Further objects and advantages of the present invention shall be made apparent in the attached drawings and the associated description thereof.

SUMMARY OF THE INVENTION

In accordance with the aforesaid objects, an absolute position measurement apparatus and method operative over the full range of motion of a movable machine member is provided. A master measuring device and a vernier measuring device each producing a measurement signal proportional to displacement of the machine movable member are provided. The measurement signals are unique within a limited range of machine member displacement and the unique values cyclically repeat over the full range of motion of the machine member. The measurement signal of the vernier device completes no more than one more cycle within the machine member range than the measurement signal of the master device. Absolute position is calculated from the measurement signals by computing the difference therebetween and using the result to compute a value representing the number of cycles of the master measurement signal. The computed value will be unique for any location within the machine range and is not dependent on recording and maintaining a location during an interval when the movable machine member is not subject to numerical control. Relative errors of measurement of machine member displacement less than one master measurement signal cycle between the two measuring devices are eliminated by the procedure used to compute the absolute position from the master and vernier measurement signals.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of an apparatus for measuring absolute position in accordance with the present invention.

FIG. 2 is a flow chart of the method used to determine absolute position from two cyclic measurement signals.

FIG. 3 shows two measuring devices having rotating elements and the associated drive mechanism.

DESCRIPTION OF THE PREFERRED EMBODIMENT

For purposes of illustrating the invention a preferred embodiment as applied to an industrial manipulator manufactured by the assignee of the present invention shall be described. It is to be understood that particular details of implementation of the present invention illustrated by the preferred embodiment are not to be interpreted as limitations of the present invention. Rather, the scope of the present invention is defined by the claims appended hereto and all equivalents thereof.

Referring to the block diagram of FIG. 1, two measuring devices, a master measuring device 12 and a vernier measuring device 14 are shown connected to their respective interface circuits 16 and 18. The measuring devices chosen by applicants for use in the industrial manipulator are of the type known as resolvers having inductively coupled rotor and stator elements. As shown, the moveable elements of the resolvers, that is, the rotors, have an alternating current signal applied to windings thereof from which output signals are induced in windings of the stators. The rotor and stator elements may be provided with windings such that the cyclic amplitude variation of the output attributable to angular displacement of the rotor is unique anywhere within a single rotation of the rotor or may be repeated two or more times for a single rotation of the rotor. Each resolver rotor has applied to it a single alternating current signal and the stator produces as an output two alternating current signals in quadrature. That is, the input signal E sin wt to the rotor is resolved into two components to produce output signals of the stator E sin (wt) sin(θ) and E sin (wt) cos(θ) where:

E=magnitude

wt=instantaneous angle of input signal

θ=angular displacement of rotor

Angular displacement θ is computed from the inverse trigonometric function of the ratio of the output signals, that is θ equals the arc tangent of E sin (wt) sin (θ)/E sin (wt) cos(θ).

It will be noted that an alternative connection of resolvers (not shown) may be substituted wherein input signals are applied to the stator windings and an output signal is produced by the rotor winding. In this configuration the orthogonal stator windings are driven by input signals having a relative phase displacement of 90°, E sin (wt) and E sin (wt+90°). The angular displacement of the rotor produces a corresponding phase shift in the output signal E sin (wt+θ) which is detected by phase discrimination to produce a measurement signal representing a fractional portion θ of a full cycle of angular displacement of the rotor.

It will be noted that output signals of the same form are obtained using inductively coupled linear measuring devices such as the Inductosyn devices available from Farrand Controls, Inc.

The interface circuits 16 and 18 provide the input signals to the resolvers and receive the output signals which are sampled and converted to appropriate levels for the following analogue to digital converters 20 and 22. Each of the analogue to digital converters 20 and 22 produce an output signal representing the instantaneous values of the output signals of the movable element of the measuring device. The outputs of the analogue to digital converters are input to angle computation circuits 30 and 32 for computing the angular displacements from the resolver outputs and for scaling the displacements in accordance with the resolution selected for the measured position. It will be appreciated that the output signals of the angle calculating circuits 30 and 32 represent the angular displacement of the rotors relative to the stators and are unique only for the effective cycle of rotation. That is, if a single revolution of the resolver rotor produces two or more repeated cycles of angular displacement in the output signals, then the output signals of the angle calculating circuits 30 and 32 will produce unique signals only for the fractional part of a resolver rotor rotation associated with a single cycle. As implemented in the preferred embodiment, a single A/D converter and angle computing circuit are used for both resolver outputs, the conversion and calculating functions being time multiplexed.

It will be apparent to those skilled in the art that any device capable of producing unique measurement signals which repeat cyclically over the range of motion of the movable machine member would be suitable substitutes for the combination of the resolvers 12 and 14, the interfaces 16 and 18, the analogue to digital converters 20 and 22, and the angle calculating circuits 30 and 32 hereinabove described. Thus, the measuring devices could be potentiometers, rheostats, variable transformers, or any linear measuring device which produces a unique measurement signal which is repeated cyclically over the range of motion of the machine movable member.

The measuring devices 12 and 14 are so arranged that the vernier measuring device 14 produces up to one more cycle of its measurement signal over the range of motion of the machine movable member than does the master measuring device 12. As the machine movable member advances from its reference position to the opposite extreme of its range, the difference between the measurement signals increases. Thus, the difference between the measurement signals always produces a unique representation of the absolute position with respect to the fixed reference. Since the measurement signals represent no more than one full cycle, the production of a unique representation of the absolute position requires that the difference between the measurement signals not exceed one full cycle over the range of motion. However, while it is apparent that the difference in measured angles is proportional to the number of revolutions imparted to the rotors of the measuring devices, mere subtraction of the master measured angle from the vernier measured angle will not insure that the desired difference will result. Rather, it is necessary to accommodate the possibility of producing a negative difference which would indicate that the vernier measurement is the fractional part of the next cycle of its measurement signal as compared to the cycle of the measurement signal of the master measurement.

Continuing then with reference to FIG. 1, the difference calculation is performed by a differencing circuit 24 which makes the necessary compensation to the result to produce a positive output in all cases. If the subtraction produces a negative difference, a value equal to a full cycle is added to the difference. The difference signal represents a fraction of a full cycle and is directly proportional to the number of cycles of the master measurement signal corresponding to the displacement of the machine movable member to its current position from the fixed reference location.

Since the angular difference is proportional to the number of cycles of the master measurement signal, the absolute position is readily computed knowing the difference in the measurement signals. The absolute position is calculated by the absolute position calculation circuitry 26 of FIG. 1 using the computed number of cycles of the master measurement signal. The absolute position is expressed as a value representing the same resolution as the master measuring device.

While the absolute position value can be computed simply by scaling the number of cycles of the master measurement signal, applicants use a somewhat more complex algorithm in the preferred embodiment to eliminate errors in the measurement signals of the vernier relative to the master measuring devices. The flow chart of FIG. 2 illustrates the procedure used in the preferred embodiment to measure the absolute position. This procedure computes the nearest number of full cycles of the master measurement signal corresponding to the distance from the reference location to the current location to synthesize an absolute position value.

Referring to the flow chart of FIG. 2, the procedure begins at step 40 where the value of the master measurement signal X and vernier measurement signal Y, each representing a fractional part of a full cycle, are read. The measurement signals X and Y are scaled to represent values relative to the desired resolution, for example, a number of thousandths of an inch or thousandths of a degree of rotation. At process step 42 the difference is calculated by subtracting the master measurement value from the vernier measurement value. At decision step 44 the result of process step 42 is tested to determine whether or not a negative value has been produced. If not, execution of the procedure would continue at process step 48. If, however, a negative value is detected then execution of the procedure continues at process step 46 where a value equal to one full cycle of the master measurement signal is added. The result either of process step 46 or the result, when positive, of process step 42 is then used by process step 48 to calculate the integer number of cycles of the measurement signal of the master resolver.

The number of cycles of the master measurement signal may be expressed as the sum of full cycles I and a partial cycle. Only the partial cycle is represented by the master measurement signal X. Assuming that the rotor of the vernier device is geared to the rotor of the master device by a relative ratio of (M+1)/M where M is equal to the number of cycles of the master measurement signal completed within the range of the movable machine member, and assuming each rotor revolution of both devices produces a single cycle of the respective measurement signal, then the cycles of the vernier device equal I+X+I/M+X/M. Again, only a portion of a full cycle is represented by the vernier measurement signal Y. When the measurement signal Y of the vernier device is greater than or equal to the measurement signal X of the master device, the sum of the last three terms X+I/M+X/M is equal to the magnitude of the vernier measurement signal. When the vernier measurement signal Y is less than the master measurement signal X, the sum of the last three terms is 1 greater than the magnitude of the vernier measurement signal Y. Therefore, the integer number of cycles I of the master measurement signal may be computed using the known quantities from:

I=M(DIFF)-X

where DIFF is the difference Y-X adjusted to produce a positive result as previously described.

At process step 48 the difference signal is used to produce an integer cycle signal I representing the integer number of cycles of the master resolver. The difference DIFF is multiplied by the number of cycles of the vernier resolver M over the measurement range, then the fractional portion of a cycle represented by the measured value of the master measurement signal X is subtracted from the product. The resultant difference is divided by the scale factor S corresponding to the number of resolutions in a cycle of the master resolver measurement signal. Any relative error in the measurement signals will cause the result to appear as a reminder R in addition to a whole number. At process step 49 the magnitude of the remainder R is measured. If the remainder R is greater than or equal to one half of a cycle of the master measurement signal then the integer value I is incremented by one at process step 50. If the remainder R is less than one half of a cycle of the master measurement signal, no change is made to the integer value I. At process step 52 the product of the integer value I resulting from process steps 48 through 50 and the scale factor S corresponding to the number of resolutions contained within a single cycle of the master measurement signal is added to the value of the master measurement signal X read at process step 40.

The effect of process steps 48, 50 and 52 is to eliminate relative errors of the values read at process step 40. The round off of step 49 and 50 produces an integer value by effectively adding or subtracting a fractional portion of a cycle of the master measurement signal less than or equal to half of a cycle. The integer value I is corrected if necessary and the partial cycle thus eliminated is replaced by the actual measured value of the master measuring device. Thus, the procedure is effective for differential errors in measurements of the master and vernier devices no greater than the fractional part of a revolution corresponding to one cycle of the measurement signal of the master resolver.

As used in the control for the industrial manipulator, the procedure of the flow chart of FIG. 2 is converted to a program for a microprocessor capable of executing integer arithmetic. In this embodiment, the difference calculation circuitry 24 and the absolute position calculation circuitry 26 of the block diagram of FIG. 1 are implemented using the microprocessor and programs therefore to achieve the procedure of the flow chart of FIG. 2. In addition, temporary storage of values used in the procedure is provided by memory locations within a random access memory.

Referring to FIG. 3, a mechanical arrangement of a master and vernier measuring device together with a drive mechanism is shown. Master resolver 60 and vernier resolver 62 are geared to a common drive shaft through their respective reduction gears 64 and 66. Drive gear 68 may be driven by a shaft directly driven by the motor driving the machine movable member or may be driven by a rack and pinion or other suitable means by motion of the machine movable member. In either case, if master resolver 60 and vernier resolver 62 are identical, a relatively small difference in the gear ratios of the master resolver gear 66 to the drive gear 68 and the vernier resolver gear 64 to the drive gear 68 produces the desired difference in the number of cycles of the vernier measurement signal compared to the master measurement signal over the range of motion of the movable machine member. Specifically, assuming the number of cycles of the master measurement signal completed over the range of motion of the movable machine member is M, then the required relative gear reduction of the vernier resolver rotor to the master resolver rotor is (M+1)/M. As stated previously, this simplified gear train presents a significant advantage over the large gear reductions required in known systems to achieve the desired resolution through one measuring device and cover the entire machine axis range by a second or other measuring devices. It will be readily appreciated from the foregoing description that unique measurement signals result from the differential displacement measured by the master and vernier measuring devices rather than by dedicating a single measuring device to a selected resolution and range.

When applied to the manipulator hereinbefore mentioned, the absolute position measuring apparatus is used to measure positions of axes of rotation of the manipulator. The measuring devices are driven by reduction gearing directly from the motor that produces motion of the movable member. Nevertheless, the same apparatus and methods described above are suitable to application to machines having linear axes of motion either by driving measuring devices having rotating elements such as the resolvers discussed herein or using linear measuring devices having two scales and two read heads.

While the present invention has been illustrated by a preferred embodiment, and while the preferred embodiment has been described in considerable detail, it is not intended to limit the scope of the invention to such detail. On the contrary, it is intended to cover all modifications, alterations and equivalence falling within the spirit and scope of the appended claims.

Claims

1. A method for producing an absolute position signal having a unique value for all positions of a movable machine member within its range of motion using measured values of no more than two measuring devices, the method comprising the steps of

(a) producing a master measurement signal representing the displacement imparted to a movable element of a first measuring device, the master measurement signal cyclically repeating over the range of motion of the machine member;

(b) producing a vernier measurement signal representing the displacement imparted to a movable element of a second measuring device, the vernier measurement signal cyclically repeating over the range of motion of the machine member, the number of cycles of the vernier measurement signal being no greater than one more than the number of cycles of the master measurement signal over a range of motion of the machine member;

(c) producing in response to the master measurement signal and the vernier measurement signal, the difference signal representing the difference in displacement represented by the measurement signals; and

(d) producing in response to the difference signal, an absolute position signal representing the unique value corresponding to the position of the movable machine member.

2. The method of claim 1 wherein the step of producing a difference signal further comprises the steps of

(a) subtracting the master measurement signal from the vernier measurement signal;

(b) testing the result of the subtraction to detect a negative difference; and

(c) in response to detecting a negative result, adding a value to the result equal to one full cycle of the master measurement signal.

3. The method of claim 1 wherein the step of producing the absolute position signal further comprises the steps of

(a) producing an integer cycle signal representing the integer number of cycles of the master measurement signal corresponding to the current position of the movable machine member;

(b) adding the integer cycle signal to the master measurement signal; and

(c) scaling the result of the addition to produce the absolute position signal.

4. Apparatus for measuring absolute position of a movable machine member relative to a fixed reference location, the apparatus comprising:

(a) a master measuring device for producing a master measurement signal representing the displacement imparted to a movable element of the master device by motion of the machine member, the master measurement signal cyclically repeating over the range of motion of the machine member;

(b) a vernier measuring device for producing a vernier measurement signal representing a displacement imparted to a movable element of the vernier device by motion of the machine member, the vernier measurement signal cyclically repeating over the range of motion of the machine member, the number of cycles of the vernier measurement signal being no greater than one more than the number of cycles of the master measurement signal over the range of motion of the machine member;

(c) means responsive to the master measurement signal and the vernier measurement signal for producing a difference signal representing the difference in displacement represented by the measurement signals; and

(d) means responsive to the difference signal for producing an absolute position signal representing a unique value defining the absolute position of the machine member.

5. The apparatus of claim 4 wherein the master measuring device further comprises:

(a) an angular measuring means having a rotatable element; and

(b) driving means connected to the movable machine member for imparting rotation to the rotatable element.

6. The apparatus of claim 5 wherein the vernier measuring device further comprises:

(a) an angular measuring means having a rotatable element; and

(b) driving means connected to the movable machine member for imparting rotation to the rotatable element.

7. The apparatus of claim 4 wherein the means for producing a difference signal further comprises:

(a) means for subtracting the master measurement signal from the vernier measurement signal;

(b) means for detecting a negative difference; and

(c) means for adding the equivalent of one cycle of the master measurement signal to the difference in response to detecting a negative difference.

8. The apparatus of claim 4 further comprising means for scaling the measured displacement of the movable elements of the master measuring device and the vernier measurement device.

9. The apparatus of claim 8 wherein the means for producing the absolute position signal further comprises means for scaling the difference signal.

10. Apparatus for measuring absolute position of a movable machine member relative to a fixed reference location, the apparatus comprising;

(a) a master angular measuring device for producing a master measurement signal proportional to the angular displacement imparted to a rotating element thereof;

(b) a vernier angular measuring device for producing a vernier measurement signal proportional to the angular displacement imparted to a rotating element of the vernier measuring device;

(c) means for driving the rotating elements of the master and vernier measuring devices in accordance with motion of the movable machine member, the driving means effective to produce no more than one more cycle of the vernier measurement signal over the range of motion of the machine member then the number of cycles of master measurement signal;

(d) means responsive to the master measurement signal and the vernier measurement signal for producing a difference signal representing the differential of angular displacement of the rotating elements of the measuring devices; and

(e) means responsive to the difference signal for producing an absolute position signal representing a unique value defining the absolute position of the movable machine member.

11. The apparatus of claim 10 wherein the means for producing an absolute position signal further comprises:

(a) means for producing an integer cycle signal representing the integer number of cycles of the master measurement signal corresponding to the current position; and

(b) means for adding the integer cycle signal and the master measurement signal to produce the absolute position signal.