GB2183843A - Load sensing pad - Google Patents

Abstract

A load sensing pad for use in position sensing apparatus comprises a rigid plate (1) mounted at each corner via U-shaped beams (2) to a base (6). The beams (2) are resilient so that the plate (1) is biased towards the rest position. The application of a force on the plate, e.g. by a writing instrument, causes the plate to move against the bias. Four sensing devices, each comprising a pair of cooperating first (13) and second (14) members connected independently to the plate (1) and the base (6) respectively are provided. One member (14) is connectable in an electrical circuit and is responsive to a magnetic field generated by the other member (13). Relative movement between the members (13, 14) caused by movement of the plate (1) causes a change in the electrical condition of the electrical circuit. Each first member (14) may comprise a Hall-effect or coil magnetic field sensor while each second member (13) comprises a permanent magnet or electro-magnet. Circuits for calculating the position of application of the force from the outputs of members (14) are described. The apparatus can be used for graphic inputs to a computer or signature verification. <IMAGE>

Description

SPECIFICATION Load sensing pad The invention relates to load sensing pads, for example for use in position sensing apparatus, of the kind comprising a substantially rigid load member mounted to a support member and biased towards a rest position, whereby the application of a load on the load member causes the member to move against the bias; and sensing means for sensing the applied load. Such load sensing pads are hereinafter referred to as of the kind described.

Load sensing pads of the kind described find application in a wide variety of fields including for example input devices for computers where the input is to be in the form of writing, charts, graphs, or drawings and also signature verification.

An example of a load sensing pad of the kind described is described inEP-B-0025282. In this load sensing pad a faceplate made of glass is mounted on three fulcra and the forces applied at the fulcra are converted into electric signals by employing strain gauges, piezoelectric elements or the like as force detectors. It can be shown mathematically that if the applied force is monitored at three different points then the exact position at which force is applied to the load member can be determined.

One of the problems with known load sensing pads of the kind described is that the magnitudes of the output signals are small and typically it is necessary to use high gain amplifiers to obtain electrical outputs which can be utilised for subsequent processing. In addition, the load sensing pads rely on the direct measurement of the forces applied at the fulcra.

In accordance with the present invention, a load sensing pad of the kind described is characterised in that the sensing means comprises at least three sensing devices, each device comprising a pair of cooperating first and second members connected independently to the load member and the support member respectively, one of the first and second members being connectable in an electrical circuit and responsive to an electro-magnetic or magnetic field generated by the other member, whereby in use relative movement between the first and second members caused by movement of the load member causes a change in the electrical condition of the electrical circuit.

In this invention, direct measurement of strain or forces is not required. Instead, the force at each position is monitored by sensing deflection of the load member at that position.

Preferably, the second member is connected in the electrical circuit. This means that no electrical connection is required with the first member if this comprises, for example, a permanent magnet thus enabling the load member to be mounted completely independently of the sensing devices.

In one example, the member connected in the electrical circuit comprises a Hall effect transducer while the other member comprises a magnetic field generator such as a permanent magnet whereby relative movement causes a change in the current flowing in the circuit. In other examples, the member connected in the electrical circuit may comprise a coil while the other member comprises a magnetic field generator whereby relative movement between the members causes a current to be induced in the coil.

Although the first and second members could be mounted so that the first member moves along an arc intersecting the second member it is preferable if the arc along which the first member moves extends alongside the second member. This latter arrangement ensures that a maximum change in flux coupling the first and second member occurs.

The load member may be mounted to the support member in a number of different ways.

Preferably, at least one resilient mounting member is connected to the load member and secured to the support member. For example, a pair of substantially U-shaped arms could be used, the load member being mounted between the legs of the arms or alternatively a pair of leaf springs could be used, the arms or leaf springs being mounted at their centres to the support member.

In order that the invention may be better understood, some examples of position sensing apparatus incorporating load sensing pads in accordance with the present invention will now be described with reference to the accompanying drawings, in which: Figures 1 and 2 are a side elevation and plan respectively of one example of a load sensing pad; Figure 3 illustrates a portion of a modified load sensing pad; Figure 4 is a circuit diagram of an amplifier for amplifying the output from a Hall effect transducer; Figure 5 is a circuit diagram illustrating initial processing of the output signals from four Hall effect transducers; Figures 6-8 illustrate three different arrangements for processing the signals generated by the circuit shown in Fig. 5; and Fig. 9 illustrates a microprocessor based system.

The load sensing pad shown in Figs. 1 and 2 comprises a rectangular steel plate 1 constitut ing a load member which is mounted on a pair of U-shaed steel beams 2, 3. Central portions of the beams 2, 3 are mounted in slots 4 of an integral upstanding boss 5 of a base 6. Each beam 2, 3 has an integral leg portion 7-10. Each leg portion 7-10 supports a ball-ended machine screw 11 which is received in a cooperating hemispherical suspension socket 12 of the plate 1.

Each leg portion 7-10 has, at its lower end, a permanent magnet 13 bonded to it. Adjacent each magnet 13 is mounted a Hall effect transducer 14 on a support 15. It will be seen in Fig.

1 that the permanent magnets 13 are are spaced from the transducers 14 so that the plate 1 is mounted completely independently of the transducers 14 and electrical circuitry (not shown) to which the transducers are connected.

When the load sensing pad is to be used in signature verification apparatus, a strip of paper or other writing material will be positioned above the plate 1 and a user will sign his name on the paper. During the act of signing, the force of the pen on the paper will be sensed by the plate 1 which will pivot accordingly about the boss 5. This pivotal movement will be accompanied by movement of the leg portions 7-10 of the beams 2, 3 by different amounts depending on where force is applied. It can be shown in fact that the amount of deflection is related to the size of the applied force.

Each Hall effect transducer 14 is positioned in a circuit (to be described below) and generates an output signal (current) relating to the strength of the magnetic field incident on the transducer. Thus, since deflection of a leg portion 7-10 will cause the appropriate magnet 13 to move along an arc beside the transducer 14, a higher magnetic flux will impinge on the transducer 14 causing a larger signal.

Fig. 3 illustrates-an alternative sensing system. Part of the arm 2 is illustrated in Fig. 3 and in this case the permanent magnet 13 is mounted on an undersurface of the beam 2. The magnet 13 is in alignment with an electrical coil 16 connected in an electrical circuit. In this example, movement of the beam 2 will cause a current to be induced in the electrical circuit due to relative movement between the magnet 13 and coil 16. In another modification (not shown) the coil 16 could be replaced by a Hall effect transducer.

The area of the plate 1 is defined in terms of a coordinate system having an origin at the bottom left hand corner of the plate (as seen in Fig. 2). The width of the plate (Y) and the length of the plate (X) are indicated in Fig. 2.

The location of any point P on the plate 1 can be defined in terms of the reaction at each suspension point labelled A-D in Fig. 2. Thus, the coordinates of the point P (XP, YP) are defined as: XP= (RB+RC) X YP= (RA+RB) Y RA+RB+RC+RD RA+RB+RC+RD where RA-RD=reactions at suspension points A-D respectively.

Although the equations above are based on monitoring reactions at four suspension points, it is also possible to determine the coordinates of P by monitoring deflections at just three suspension points and modified mathematical expressions similar to those above can be determined.

Each Hall effect transducer is capable of producing output signals in the region of several hundreds of millivolts. However, the displacement and hence the output is limited to a lower level to preserve a better degree of linearity. In view of this, the transducer output signal is amplified to produce a level acceptable to subsequent A/D converters. A typical circuit for this is shown in Fig. 4 for a single ended transducer. A nominal gain of 50 is suitable, but other gains are feasible.

This should be contrasted with conventional strain guage arrangements. Typical strain guages produce signals of a few millivolts, because the strain is low. The signals thus require amplification in the region of 1000 times to produce a suitable input to an anlogue to digital converter.

The signal from the transducer is fed to a capacitor 17 and then via a resistor 18 to the inverting input of an amplifier 19. The non-inverting input is grounded. The output of the amplifier 19 is coupled into a feedback circuit comprising a pair of series connected resistors 20, 21 in parallel with a capacitor 22.

The lower frequency response is set to 0.001 Hz so that neglible phase shift occurs at the frequency of interest e.g 1 Hz to 15 Hz. The filter effectively removes both the dc offset of the transducer and variations with temperature. The long setting time normally associated with filters of this frequency could be a problem when the circuit is first switched on. This is overcome by an automatic reset which enables the capacitor 17 to reach its final dc charge state more quickly. The reset could be accomplished using transistors or a switch operated by a computer.

A differential output Hall effect transducer may also be used with the advantage of a higher differential output and better characteristics. In this case, a circuit similar to Fig. 4 is used but having differential filtered input.

The amplifier circuits could be constructed from discrete components or on a thick fiim hybrid package mounted directly under the plate 1 on the base 6.

An alternative processing system based on the use of a microprocessor is illustrated in Fig. 9.

This system extends the frequency response to dc and introduces zero phase shift. Fig. 9 illustrates the processing of signals from one Hall effect transducer 14 and similar arrangements are associated with the other Hall effect transducers although a common microprocessor will be used. The output from the Hall effect transducer 14 is fed to the non-inverting input of an operational amplifier 50. The output of this amplifier 50 is fed to an analogue to digital converter 51 whose digital output is fed to a microcomputer 52. The microcomputer 52 supplies a digital input signal to a digital to analogue converter 53 which applies a corresponding analogue voltage to the inverting input of the amplifier 50.

In practice, the D/A converter 53 is continuously- updated when the load sensing pad is not in use. The transducer offset will then be exactly cancelled by the signal from the converter 53, within one bit resolution of the converter. When the load sensing pad is used, a signal will be detected. This could be from the insertion of a magnetic card, the pressing of a keyboard pad or the detection of analogue pressure on the load sensing pad. On receipt of this signal, the data entered into the D/A converter 53 will be frozen. Inputs to the load sensing pad will then pass into the computer 52 without cancellation. In effect, the dc offset of the signal would be removed without adverse effects of phase shift, settling time or voltage decay.

As illustrated in Fig. 9, the D/A converter could contain either an internal or external storage register and could be fed from the computer 52 or directly from the A/D converter 51 (shown by a dashed line).

The computer 52 carries out calculations on the input digital data to determine the coordinates XP, YP, as previously described.

Fig. 5 illustrates the connection between each Hall effect transducer 14 (labelled A-D) and respective amplifier circuits 23-26 each similar to that shown in Fig. 4.

The output signal from the amplifier 23 is fed to conventional adder circuits 27-28. The output signal from the amplifier 24 is fed to the circuits 27, 28 and and adder circuit 29. The output signal from the amplifier 25 is fed to the adder circuits 28, 29; and the output signal from the amplifier 26 is fed to the adder circuit 28.

The output signals from the adders 27-29 are related to the mathematical functions set out in Fig. 5.

These output signals can be processed in a number of different ways.

One simple processing method is illustrated in Fig. 6 where the signals from the adders 27, 28 are fed to respective analogue divider circuits 30, 31 where each is divided by the output signal from the adder circuit 28. The output signals from the divider circuits 30, 31 are directly related to the coordinates YP, XP of the point of applied force by virtue of the equations set out above.

Fig. 7 illustrates a digital divider circuit. In this case, the signals from the adder circuits 27-29 are fed to respective voltage to frequency converters 32-34 the output signals from which have frequencies directly related to the input signals. The output signal from the converter 32 is fed to an incrementing input of a counter 365 while the signal from the converter 33 is fed to an incrementing input of a counter 35 while the signal from the conveter 33 is fed to an incrementing input of a counter 36. The output signal from the converter 34 is fed to two binary rate multipliers 37, 38 whose outputs are connected to decrementing inputs of respective counters 35, 36. The count n (varying typically from 0 to 255) from each counter 35, 36 is applied to respective multipliers 3q, 38 after normalisation.The counts (n) from the counters 35, 36 are digital representations of the coordinates of the point P.

Each multiplier 37, 38 accepts a fixed or variable frequency (fi,) and multiples it by a fixed or variable parallel and digital code (n after normalisation by division by 2N where N is the bit length of the counters). The weight of the code can be any length but is typically eight bits. For instance, if a binary number of 1100,0000 was applied this would represent a decimal value of 0.75 (ie n/28). Thus, if the input frequency was 100 kHz then the output frequency would be 75 kHz. It will be seen that each multiplier thus acts as a frequency multiplier (with the multiplicand between 0 and 1).

In this case, when a pen is placed on the pad, the frequency representing (RA+RB) will cause the counter 35 to count up. After each increment, the current count is applied to the multiplier 37 so that the input frequency to the multiplifier 37 (RA+RB+RC+RD) Hz is multiplied by the value n. The value of the frequency fount from the multiplier 37 increases and is fed to the decrementing input of the counter 35. It therefore follows that the rate of counting up decreases since the net input is (fup-fdown) Hz. The counter will cease continuous counting when the two frequencies are equal.

Now fdown =(in counter)xfjn, neglecting constants.

=n x (RA+RB+RC+RD) Hz and fUp=(RA+RB) Hz.

Thus (R+R)=n (RA+RB+Rc+RD)

The value of YP is calculated and converted to a parallel code.

A similar analysis applies to the counter 36 and multiplier 38.

Fig. 8 illustrates a further system for processing the output signals from the circuit shown in Fig. 5. In this case, the output signals are fed to a four channel A/D converter 39 whose serial output is fed to a microcomputer 40. In addition, an output signal from a temperature transducer 41 monitoring the ambient temperature is fed to the converter 39. The microcomputer 40 thus receives in series digital representations of the signals from adder circuits 37-39 and the temperature from the transducer 41. The temperature transducer 41 is used to allow compensation for any deviation in characteristics which might affect the overall accuracy of the system.

The computer 40 then calculates from the digital input values, using the equations previously set out, the coordinates of the point P.

It will be appreciated that the equations defining the coordinates of the point P are essentially independent of the applied force. However, if the applied force is too low then the valid data may have poor resolution because the full scale range of the converter is set for the maximum forces encountered. In this application the range can be varied by using different reference voltages applied to the converter 39. For example, a two bit code would allow the selection of four different analogue voltages giving ranges for very low, low, medium, and high forces. This code can be derived from the computer 40 which could monitor an initial pressure and, if necessary, request the user to repeat the movement while changing the reference voltage via an analogue switch 42.

When the apparatus is to be used for signature verification, the two bit code could be derived from a previous history of the signer. For example, the code could be stored on a magnetic or intelligent card and read by a suitable card reader connected to the computer 40 before the signer signs his name.

The computer 40 can also provide an output relating to the total signing force.

The coordinates determined by any of the methods described above could be used to control a display device which simply repeats the pattern being drawn on the plate 1. Alternatively, the sets of coordinates can be periodically sampled and stored to allow the pattern to be subsequently processed or analysed.

Claims (9)

1. A load sensing pad of the kind described characterised in that the sensing means comprises at least three sensing devices, each device comprising a pair of cooperating first and second members connected independently to the load member and the support member respectively, one of the first and second members being connectable in an electrical circuit and responsive to an electro-magnetic or magnetic field generated by the other member, whereby in use relative movement between the first and second members caused by movement of the load member causes a change in the electrical condition of the electrical circuit.

2. A pad according to claim 1, wherein the second member is connbctable in the electrical circuit.

3. A pad according to claim 1 or claim 2, wherein the member connectable in the electrical circuit comprises a Hall effect transducer and the other member comprises a magnetic field generator.

4. A pad according to claim 1 or claim 2, wherein the member connected in the electrical circuit comprises a. coil and the other member comprises a magnetic field generator.

5. A pad according to any of the preceding claims, wherein the first member is mounted for movement along an arc which extends alongside the second member.

6. A pad according to any of the preceding claims, wherein at least one resilient mounting member is connected to the load member and secured to the support member.

7. A pad according to claim 6, wherein the load member is mounted between the legs of generally U-shaped arms mounted at their centres to the support member.

8. A load sensing pad substantially as hereinbefore described with reference to any of the examples shown in the accompanying drawings.

9. Signature verification apparatus incorporating a load sensing pad according to any of the preceding claims; and processing means responsive to the electrical condition of the electrical circuits to monitor force applied to the load member.