GB2051408A - Friction compensating apparatus - Google Patents

Abstract

An electrical circuit provides error signals to compensate for frictional torques in a servomechanism including a servomotor to drive a shaft in a roller or ball bearing in which the frictional torques are related in a linear way to the displacement of the shaft. The circuit includes two integrators connected in the parallel individually integrating with respect to time input signals representative of instantaneous shaft speeds. During a first predetermined displacement the varying output signals of the integrators are combined, subsequently one integrator becoming clamped and providing a constant output signal, until the shaft is displaced by a further second predetermined displacement when the other integrator becomes clamped and provides a constant output signal, the circuit error signals causing the servomotor to drive the shaft so that the different frictional torques in the bearing are overcome precisely.

Description

SPECIFICATION Electrical circuits This invention relates to electrical circuits, and in particular to electrical circuits each to provide, in response to appropriate input signals, error signals to an associated electro-mechanical servomechanism, the servomechanism including a servomotor to drive a movable member mounted either in or on a bearing of a type hereinafter defined, in response to input command signals, the movable member also being driven in response to the error signals from the electrical circuit, so that the different forces are provided to overcome precisely the friction in the bearing.

The provision of input command signals to an electro-mechanical servomechanism is well known, and will not be discussed in this specification .

Further, whilst the present invention relates to the provision of error signals to electromechanical servomechanisms of any suitable kind, including, for example, having a linearly movable member mounted on a bearing comprising recirculating balls, for convenience in this specification and accompanying claims only electro-mechanical servomechanisms each having a shaft mounted in a bearing of a type hereinafter defined will be referred to, although references to such latter forms of servomechanisms will be considered to include references to the other forms of servo mechanism where appropriate.

For an electro-mechanical servomechanism to operate in a precise manner it is essential that the shaft is axially and radially rigidly mounted in the bearing. Thus, the construction of a bearing of the type defined herein is required to be such that the rotatable members of the bearing are compressed at regions adjacent to only portions of their surfaces, and in operation, as the rotatable members rotate, these compressed regions are displaced through each rotatable member. Hence, the rotatable members are required to be of an elastic material.

A shaft mounted in a bearing is shown in the accompanying Fig. 1, the illustrated bearing comprising one example of a bearing of the type defined herein. The illustrated bearing comprises two substantially identical parts, each part having an inner race and an outer race, with steel balls, comprising rotatable members, between the races. Each outer race has a surface co-operating with the balls comprising an axially extending ramp, the ramps of the two substantially identical bearing parts being oppositely directed, the adjacent portions of the ramps comprising the highest portions of the ramps, furthest from the shaft.The axial separation between the inner races can be reduced by clamping means acting upon these inner races, causing the balls to ride up the ramps of the outer races, and to become compressed between the inner and outer races, in the directions of the arrows shown in Fig. 1. Hence, regions of the balls adjacent to only portions of their sur faces contiguous with the races are com pressed. As the balls rotate between the races these compressed regions are displaced through each ball, regions which originally i were not compressed becoming compressed, and vice versa, because of the elastic nature of steel.

However, steel is also viscous in character, as are other elastic materials. Consequently, regions of the rotatable members originally compressed do not become completely re leased when no longer subjected to compres sive forces, because of the viscosity in the rotatable members. This viscous effect is the principle cause of friction within the bearing.

With any bearing of the type defined herein and with the rotatable members compressed, the different frictional forces in the bearing, under different operating conditions, are pre dictable, and can be determined, at least empiricaily.

It is convenient initially to consider when the shaft of Fig. 1 is driven cyclically, with different amplitudes, about a mean position.

Fig. 2 comprises the graph of the rotational displacement H of the shaft from the mean position against the required error torques T, to be provided by the servomotor of the associated servomechanism, to overcome pre | cisely the frictional torques in the bearing, in possible cyclical displacements of the shaft.

Each instantaneous frictional torque is equal and opposite to the required corresponding instantaneous error torque T. Also each in i stantaneous frictional torque is substantially directly proportional to the error in the instan taneous position of the shaft.

Uncorrected cyclical displacements of the shaft between points P, P' on the graph are now now considered, at which points P and P' errors in the positions of the shaft, and the error torques T, have maximum values, but of opposite senses, and reversals of direction of motion of the shaft occur at these points.

When the shaft is driven away from a point P or P', the initial rate of change of error torque T with shaft displacement 8 is large, and approximately constant. Initially, the frictional torques in the bearing are such that the shaft tends tend to be driven by these frictional torques towards the mean position of its cyclical mo tion, because of the viscous effect referred to above. However, after a first predetermined displacement from the point P or P', and whilst the shaft is still moving in the same direction, the rate of change of error torque T with shaft displacement 8 becomes smaller, and is again approximately constant, until the shaft is at the other point P or P' of the graph, at which point a reversal in the direc tion of motion of the shaft occurs.A zero frictional torque position is obtained whilst the rate of change of error torque T with shaft displacement H is at either its larger or smaller value. However, the zero frictional torque position is obtained before the shaft is in its mean position of its cyclical motion. At the zero frictional torque position the error in the instantaneous position of the shaft is zero.

After passing through this zero frictional torque position the frictional torques in the bearing are such that the shaft tends to be driven by these frictional torques towards the zero frictional torque position.

Upon the shaft being driven in the reverse direction in its cyclical motion the same effects as referred to in the preceding paragraph are experienced. However, because the zero frictional torque position again occurs before the mean position of the cyclical motion is obtained the corresponding part of the graph is not coincident with the part of the graph discussed in the preceding paragraph. Thus, the graph has a form similar to that of a socalled hysteresis curve obtained for the magnetisation of ferromagnetic materials.

The illustrated points P, P' are particular points in the cyclical motion characteristic of the shaft, comprising breakaway points. For cyclical displacements of the shaft of amplitudes greater than that represented by the points P, P', for example, displacements to points 0, Q', the error in the position of the shaft, and the required error torque T, remain constant at their maximum values. Parts of the cyclical motion of the shaft beyond the points P and P' are represented partially by the dotted lines of the graph of Fig. 2.

For cyclical displacements of the shaft of amplitudes less than that represented by the breakaway points P, P', for example, between points M, M', the area enclosed by the hysteresis curve comprises the central part of the area enclosed by the hysteresis curve between the breakaway points P, P'. The parts of the curve representing the smaller rate of change of error torque T with shaft displacement 8 are common to the portions of the parts of the curve representing the smaller rate of change of error torque T with shaft displacement 8 between the breakaway points P and P'.However, the first predetermined displacement, between the point M or M', and the point at which the smaller rate of change of error torque T with shaft displacement a occurs, is of the same magnitude as that referred to above in relation to shaft cyclical displacements between the breakaway points P and P'.

The smaller rate of change of error torque T with shaft displacement 8 occurs for a second displacement after the first predetermined displacement. For cyclical motion of the shaft at least between the breakaway points P and P' of the graph, the smaller rate of change of error torque T with shaft displacement a occurs for a second predetermined displacement. For cyclical motion of the shaft between the points M, and M' of the graph, the smaller rate of change of frictional torque with shaft displacement occurs for a second displacement shorter than the second predetermined displacement associated with the breakaway points P and P', but longer than the first predetermined displacement.The first predetermined displacement, and any possible second displacement, each may be small in relation to the amplitude of the cyclical motion of the shaft. The first and second predetermined displacements comprise constants for a particular servomechanism, and may be determined empirically.

For shaft rotational displacements in only one direction, the appropriate part of the hysteresis curve between the points M, M', or P, P', or Q, Q', has to be considered. The extremities of the rotational displacement comprise the positions of reversal of direction of a corresponding cyclical motion of the shaft considered in relation to the graph. For shaft rotational displacements in only the opposite direction, the other part of the hysteresis curve, between the appropriate positions of reversal of direction of the cyclical motion of the shaft, has to be considered.

The error signals from the electrical circuit in accordance with the present invention, and to be provided to the associated servomechanism, are to drive the servomotor of the servomechanism. Hence, the error signals are required to have a parameter, such as voltage, the error signal magnitudes of which parameter are required to represent the different error torques T to be provided to the shaft by the servomotor, and required to overcome precisely the different frictional torques in the bearing in which the shaft is mounted. Further, the senses of the error signals are required to represent the different senses of the frictional torques in the bearing. Thus, the graph of Fig. 2 also comprises the analogue of the shaft displacement 8 against the magnitudes of a parameter of the error signals provided by the circuit.

It is an object of the present invention to provide an electrical circuit to provide, in response to appropriate input signals, error signals to an associated electro-mechanical servomechanism, the servomechanism including a servomotor to drive a shaft mounted in a bearing of a type herein before defined between any required orientations, in response to the error signals from the electrical circuit, so that different frictional torques in the bearing are overcome precisely.

According to the present invention an electrical circuit to provide error signals to an associated electro-mechanical servomechanism, the servomechanism including a servomotor to drive a shaft mounted in a bearing of a type herein defined between any required orientations, includes two integrators connected in parallel with each other, each integrator being arranged to integrate individually a parameter of provided input signals to the circuit with respect to time, the magnitude of the parameter of each such input signal to be representative of a detected instantaneous speed of the shaft, each integrator to be reset at the start of each rotational displacement of the shaft, and each integrator is provided with clamping means, for each rotational displacement of the shaft, the circuit arrangement being such that one integrator provides a constant output signal after the detection of a first predetermined displacement, and the other integrator provides a constant output signal after the detection of a further, second predetermined displacement, each error signal from the circuit being caused by the summation of instantaneous output signals from the two integrators, in response to the receipt of the error signals by the associated servomechanism the corresponding error torques provided by the servomotor precisely to overcome the different frictional torques in the bearing in the shaft rotational displacement, magnitudes of the parameter of the error signals representing the different error torques required to be provided.

The error signals from the circuit in accordance with the present invention are employed in a real time manner to cause the associated servomechanism to overcome precisely the different bearing frictional torques in each shaft rotational displacement.

Each integrator of the circuit may have a conventional construction, including a capacitor connected in parallel with, and a resistor connected in series with, an amplifier.

Usually the clamping means for each provided integrator comprises a pair of oppositely directed Zener diodes connected in parallel with the amplifier of the integrator, the integrator becoming clamped and providing a constant output signal, by becoming saturated in operation.

The first and second predetermined displacements, in each rotational displacement of the shaft, referred to above in relation to an electrical circuit in accordance with the present invention, comprise the first and second predetermined displacements referred to above in relation to the graph of the accompanying Fig. 2.

When the shaft is required to be displaced by more than the first predetermined displacement, the clamping means of said one integrator causes the previously varying output signals of the integrator to become constant.

This constant integrator output signal is combined with the still varying output signals of said other integrator, to cause the corresponding error signals from the circuit, until the shaft has been displaced further by the second predetermined displacement. Then the clamping means of said other integrator causes the output signal of this integrator to become constant, and this constant signal is combined with the constant output signal from said one integrator, to cause the corresponding constant error signal from the circuit. Thus, the characteristic of the rotational displacement of the shaft against the magnitudes of the parameter of the error signals from the circuit, resembles the hysteresis curve of the graph of the accompanying Fig.

2.

If the shaft is required to be displaced by less than the sum of the first and second predetermined displacements, the second integrator does not saturate before it is reset, for rotational displacements of the shaft less than between the breakaway points P and P' of the graph, after the first predetermined displacement there is a second displacement smaller than the second predetermined displacement According to another aspect the present invention comprises the combination of an electrical circuit as referred to above, and an electro-mechanical servomechanism including a shaft mounted in a bearing of a type defined herein, error signals from the circuit being provided to the servo-mechanism.

The present invention will now be described by way of example with reference to the accompanying drawings, in which: Figure 1 is a section of a shaft mounted in a bearing of the type defined herein, the shaft and the bearing comprising part of an electromechanical servomechanism of a conventional construction, and to operate in response to input command signals in a known manner Figure 2 is a graph of when the shaft of Fig. 1 is driven cyclically with different amplitudes, about a mean position, the graph being of the rotational displacement a of the shaft from the mean position against the required error torques T, to be provided by the servomotor (not shown) of the servomechanism, to overcome precisely the frictional torques in the bearing, and Figure 3 is a circuit diagram of one embodiment of an electrical circuit in accordance with the present invention, the circuit to provide error signals to the servomechanism, to cause the servomechanism to drive the shaft, so that the different frictional torques in the bearing are overcome precisely.

As shown in Fig. 1 a shaft 10 is mounted in a bearing 11 of a conventional construction. The shaft is axially and radially rigidly mounted in the illustrated bearing 11, the illustrated bearing comprising one embodiment of a bearing of the type defined hereinbefore in this specification.

The illustrated bearing 11 comprises two substantially identical parts, each part having an inner race 1 2 and an outer race 13, with steel balls 14, comprising rotatable members, between the races. Each outer race has a surface co-operating with the balls comprising an axially extending ramp 15, the ramps 1 5 of the two substantially identical bearing parts being oppositely directed, the adjacent portions of the ramps comprising the highest portions of the ramps, furthest from the shaft.

Each inner race 1 2 has a surface with a channel 1 6 in which the balls rotate about the shaft 10. The axial separation between the inner races 1 2 can be reduced by clamping means, comprising a nut 1 7 co-operating with a screw threaded reduced diameter part 1 8 of the shaft, and acting against an enlarged diameter part 1 9 of the shaft, the clamping means acting upon these inner races causing the balls 14 to ride up the ramps 1 5 of the outer races 13, and to become compressed between the inner and outer races, in the directions of the arrows shown in Fig. 1.

Hence, regions of the balls adjacent to only portions of their surfaces contiguous with the races 12, 13, are compressed. As the bails rotate between the races these compressed regions are displaced through each ball, regions which originally were not compressed, becoming compressed, and vice versa, because of the elastic nature of steel.

The shaft and bearing are part of an electromechanical servomechanism of a conventional construction, the shaft being driven by a servomotor (not shown), in response to input commands. The provision of input commands to an electro-mechanical servomechanism is well known, and will not be discussed in this specification.

Fig. 2 comprises the graph of the rotational displacement H of the shaft against the required error torques T, to be provided by the servomotor of the servomechanism, to overcome precisely the different frictional torques in the bearing in rotational displacements of the shaft. The form of the graph is described above at the beginning of this specification.

The present invention relates to electrical circuits, each such electrical circuit to be connected to a servomechanism of a construction referred to above, and in particular to a servomechanism having a shaft mounted in a bearing of the type defined herein. In response to appropriate input signals, the electrical circuit is to provide terror: signals, to be supplied to the servomechanism so that the different frictional torques in the bearing are overcome precisely. Hence, the error signals are required to have a parameter, such as voltage, the error signal magnitudes of which parameter are required to represent the different error torques T to be provided to the shaft by the servomotor, and required to overcome precisely the different frictional torques in the bearing in which the shaft is mounted.Further, the senses of the error signals are required to represent the different senses of the frictional torques in the bearing. Thus, the graph of Fig. 2 also comprises the analogue of the shaft displacement 8 against the magnitudes of the parameter of the error signals provided by the circuit.

An embodiment of an electrical circuit in accordance with the present invention, is .shown in Fig. 3.

The circuit includes first and second integrators, 21 and 22, each integrator having a conventional construction, the first integrator 21 having an amplifier Al, with a capacitor Cl connected in parallel therewith, and a variable resistor R1 connected in series therewith, and the second integrator 22 having an amplifier A2, with a capacitor C2 connected in parallel therewith, and a variable resistor R2 connected in series therewith. The two integrators 21 and 22 are connected in parallel with each other, and are connected to a common input 'IN' for the circuit. There is a common output 'OUT' for the circuit. The first integrator 21 is connected to the circuit output via a variable resistor R5, a fixed resistor R6, and an amplifier A3 having a resistive feedback path R7.The second integrator 22 is connected to the circuit output via a variable resistor R8, a fixed resistor R9, and the same amplifier A3, having the resistive feedback path R7, as the first integrator Each error signal from the circuit is caused by the summation of instantaneous output signals from the two integrators, each such instantaneous output signal being modified by the appropriate combination of a variable resistor R5 and R8, and a fixed resistor R6 and R9, before being supplied to the amplifier A3.

The integrators 21 and 22 each comprises a li ,miting integrator, including clamping means comprising a pair of oppositely directed Zener diodes Z1 and Z2, and Z3 and Z4, respectively, connected in parallel with the amplifier Al, and A2 of the integrator.

The input signals provided to the circuit are required to be representative of detected instantaneous rotational speds of the shaft.

Such circuit input signals may be supplied directly from a tachogenerator comprising part of the servo-mechanism. Alternatively, these signals may be supplied by a computer in response to signals provided by a synchro resolver.

The integrators 21 and 22 simultaneously integrate these circuit input signals with respect to time.

The characteristic of the shaft displacement 8 against the corresponding magnitudes of a parameter of the output signals of each integrator has an intermediate part comprising a line through the origin, and with a positive slope. Each extremity of the characteristic comprises a straight line representative of a constant magnitude of the parameter, and starting at a predetermined value for the shaft displacement 8, the two extremities being representative of equal magnitudes of the param eter, but of opposite senses, and the starting points of the two extremities are at equal predetermined values for the shaft displacement 0, but of opposite senses. The slope of the intermediate part of the characteristic is inversely proportional to the time constant of the integrator.

For each rotational displacement of the shaft, the circuit arrangement is such that one integrator provides a constant output signal after the detection of a first predetermined displacement, and the other integrator provides a constant output signal after the detection of a further, second predetermined displacement. The first and second predetermined displacements, in each rotational displacement of the shaft in relation to the manner of operation of the circuit, comprise the first and second predetermined displacements referred to above in relation to the graph of Fig. 2.

The integrators are arranged to be reset at the start of each rotational displacement of the shaft.

When the shaft is required to be displaced by more than the first predetermined displacement, the clamping means of the first integrator 21 causes the previously varying output signals of the integrator to become constant.

This constant integrator output signal is combined with the still varying output signals of the second integrator 22, to cause the corresponding error signals from the circuit, until the shaft has been displaced further by the second predetermined displacement. Then the clamping means of the second integrator causes the output signal of this integrator to become constant, and this constant signal is combined with the constant output signal from the first integrator, to cause the corresponding constant error signal from the circuit, until the integrators are reset before the next rotational displacement of the shaft.

Thus, the characteristic of the rotational displacement of the shaft against the magnitudes of the parameter of the error signals from the circuit, resembles the hysteresis curve of the graph of Fig. 2.

If the shaft is required to be displaced by less than the sum of the first and second predetermined displacements, the second integrator does not saturate before it is reset, for rotational displacements of the shaft less than between the breakaway points P and P' of the graph, after the first predetermined displacement there is a second displacement smaller than the second predetermined displacement.

Hence, the error signals from the circuit are employed in a real time manner to cause the associated servomechanism to overcome precisely the different frictional torques in each shaft rotational displacement.

Also as is indicated above in relation to the graph, the first predetermined displacement and any possible second displacement, each may be small in relation to a rotational dis placement of the shaft. Further, the first and second predetermined displacements comprise constants for a particular servo-mechanism, and may be determined, at least empirically, and so that some first and second predeter mined displacements in relation to the operation of the circuit are similarly determined.

Hence, the magnitudes of the circuit ele ments, of the circuit of Fig. 3, are chosen so that the characteristic has the required form.

In particular, the variable resistors can be adjusted, with the circuit in operation, the resistors R1 and R2 being used to determined at what shaft displacements the integrators 21 and 22 saturate, and the resistors R5 and R8 being used to determine the magnitude of the parameter of the output signals from the integrators when saturation of the integrators occurs.

In this manner the rates of change of the output signal magnitudes with the displacement of the shaft during both the first and second predetermined displacements, in each rotational displacement of the shaft, are determined to represent in the required manner the rates of change of the error torque T with the displacement of the shaft during both the first and second predetermined displacements.

The sense of each rotational displacement of the shaft is represented by the sense of the input signals to the circuit, and by the sense of the corresponding error signals from the circuit. In response to the sense of the error signals from the circuit, and supplied to the servomechanism, the servomotor applies torques to the shaft in the appropriate directions to overcome the different frictional torques in the bearing in each rotational displacement of the shaft.

The bearing of the associated electro-mechanical servomechanism may have any convenient construction for a bearing of the type defined herein. The rotatable members of the bearing may not comprise balls.

The associated electro-mechanical servomechanism may be of any suitable kind, including, for example, having a linearly movable member, mounted on a bearing comprising recirculating balls, instead of the movable member comprising a shaft.

Claims (5)

1. An electrical circuit to provide error signals to an associated electro-mechanical servomechanism, the servomechanism including a servomotor to drive a shaft mounted in a bearing of a type herein defined between any required orientations, the circuit including two integrators connected in parallel with each other, each integrator being arranged to integrate individually a parameter of provided input signals to the circuit with respect to time, the magnitude of the parameter of each such input signal to be representative of a detected instantaneous speed of the shaft, each integrator to be reset at the start of each rotational displacement of the shaft, and each integrator is provided with clamping means, for each rotational displacement of the shaft, the circuit arrangement being such that one integrator provides a constant output signal after the detection of a first predetermined displacement, and the other integrator provides a constant output signal after the detection of a further, second predetermined displacement, each error signal from the circuit being caused by the summation of instantaneous output signals from the two integrators, in response to the receipt of the error signals by the associated servomechanism the corresponding error torques provided by the servomotor precisely to overcome the different frictional torques in the bearing in the shaft rotational displacement, magnitudes of the parameter of the error signals representing the different error torques required to be provided.

2. A circuit as claimed in claim 1 in which the clamping means for each provided integrator comprises a pair of oppositely directed Zener diodes connected in parallel with an amplifier of the integrator, the integrator becoming clamped, and providing a constant output signal, by becoming saturated in operation.

3. A conbination of an electrical circuit as claimed in claim 1 or claim 2, and an electromechanical servomechanism including a shaft mounted in a bearing of a type defined herein, error signals from the circuit being provided to the servomechanism.

4. An electrical circuit substantially as described herein with reference to the accompanying drawings.

5. A combination of an electrical circuit, and an electro-mechanical servomechanism including a shaft mounted in a bearing of a type defined herein, error signals from the circuit being provided to the servomechanism, and the combination being substantially as defined herein with reference to the accompanying drawings.