A Non-Contact Sensor

Technical Field

The present invention relates to a non-contact sensor, more particularly a non-contact sensor based on magneto-resistive resistors arranged into a sensing element and a manufacturing method thereof.

Introduction

Original Equipment Manufacturers (OEMs) are increasingly using sensors for sensing angular or linear motion, for example, manufacturers in the automotive industry are driven by the need to accurately control key vehicle parameters to achieve emission standards and optimise performance and safety. Sensors have to operate in harsh environments within the engine and drivetrains and are required to be robust, highly accurate and low-cost. Magneto resistors (MR) used in MEM (micro-electro mechanical) sensors are widely used as non- contact sensors for high-precision contactless angular and linear measurements in chassis applications e.g. ABS, power steering and acceleration; as well as being useful in many applications outside the automotive industry. Magnetoresistance is the property of a material to change the value of its electrical resistance when an external magnetic field is applied to it. There is a variety of effects that can be called magnetoresistance, some of them occurring in bulk non-magnetic metals and semiconductors, e.g. geometrical magnetoresistance, others in magnetic metals, e.g. negative magnetoresistance in ferromagnets or AMR (anisotropic magnetoresistance). Current MR Sensors are based on AMR Barber poles on aluminium strips with output signal being determined by the angle between the AMR and the magnetic field. Recently giant magneto-resistive (GMR) sensors have emerged as powerful tools for ultrasensitive, multiplexed, real-time electrical read-out due to the large change in resistance compared to AMR when devices comprising GMRs are subjected to a magnetic field. The main application of GMR is magnetic field sensors, which are used to read data in hard disk drives, biosensors, microelectromechanical systems (MEMS) and other devices. The essential feature of a conventional GMR element comprises at least two ferromagnetic metal layers separated by a non-ferromagnetic metal layer. The GMR effect has been found in a variety of systems, such as Fe/Cr, Co/Cu, or Co/Ru multilayers exhibiting strong non- ferromagnetic coupling of the ferromagnetic layers. This GMR effect has also been observed for these types of multilayer structures, but wherein the ferromagnetic layers have a single crystalline structure and thus exhibit uniaxial magnetic anisotropy. The physical origin of the GMR effect is that the application of an external magnetic field causes a reorientation of all of the magnetic moments of the ferromagnetic layers. This in turn causes a change in the spin-dependent scattering of conduction electrons and thus a change in the electrical resistance of the multilayered structure. The resistance of the structure thus changes as the relative alignment of the magnetic moments of the ferromagnetic layers changes. In the absence of applied magnetic field the magnetic moments from each ferromagnetic layer are antiparallel to their next nearest layer. As the magnetic field is applied the magnetic moments in each of the neighbouring ferromagnetic layers align so that all of the magnetic moments become parallel and the magnetisation of the multilayer is saturated which leads to an increase in the electrical conductivity of the multilayer - this is known as "interlay er exchange coupling". Another important aspect of GMR is "spin- dependent scattering"; the number of available quantum states that the electron can scatter into depends on the relative direction of the electron's spin and the magnetic field within the ferromagnet. The higher the number of available states, the higher the probability of scattering, and the higher the electrical resistance. Having the spin antiparallel to the magnetic field creates more available states for electron scattering, so the electrical resistance is increased if the spin is parallel to the magnetic field.

It is ideal that the output from an MR sensor varies linearly with the change in direction of the magnetic field, e.g. angular movement. Without this linear relationship between the output of the MR sensor and the rotation of the magnetic field, correlating the output of the MR sensor to rotation can be difficult without having the necessary processing power. For example, in the case of an automotive controller such as throttle control, ideally the output from the GMR sensor varies linearly with the angular movement of the throttle pedal such that the processor can determine how much throttle is being applied. Known MR type sensors comprise four MR resistors arranged as a Wheatstone bridge. Whereas AMR resistors exhibit a change of resistance of less than 3%, GMR resistors are preferable since they achieve a change in resistance of the order of 10% to 20%. A Wheatstone bridge configuration provides an easy to use voltage output that is proportional to the magnetic field applied but insensitive to any variations in the absolute resistance of the MR device. Typically magneto-resistive metals or resistors making up the Wheatstone bridge are laid on a substrate e.g. sputtered. A typical configuration of a Wheatstone bridge GMR sensor known in the art comprises two Wheatstone bridge circuits that are laterally offset from each other. The signals from the Wheatstone bridge circuits are manipulated to produce a substantially linear relation between the output of the MR sensor and the change in orientation in the magnetic field. The outputs from each of the Wheatstone bridge circuits vary with the angular movement of a magnet spaced apart from the Wheatstone bridge circuits. The rotational magnetic field interacts with the GMRs to cause a change in their respective resistances which in turn causes the outputs, v ₁ and v2, from each Wheatstone bridge circuit to follow a sinusoidal wave.

In order to achieve linear relationship with a high degree of precision it is important that the magneto-resistive resistors in each Wheatstone bridge circuit experience the same magnetic flux in a uniform magnetic field. Traditionally the Wheatstone bridge circuits are manufactured on a micro-scale. The change in resistance of each magneto-resistive resistor is small and because at such small scales, the flux of magnetic field crossing each magneto- resistive resistor is small, the precision in sensing angular displacement is compromised. Moreover, no two Wheatstone bridges can be made identical due to limitations in manufacturing, particularly on a micro-scale. As a result, changes in resistance in one Wheatstone bridge can be different to the changes in another Wheatstone bridge for the same change in magnetic field strength and/or direction. Consequently this leads to a variation of the signal from each of the Wheatstone bridge circuits which requires compensation in order to derive a linear relationship between angular displacement and sensor output, leading to further processing needs.

A mathematical algorithm is used convert the two sinusoidal wave outputs into a linear output that is proportional to the angular rotation of the magnetic field. One of the essential requirements to achieve a linear output from the MR sensor is that the output from one Wheatstone bridge circuit should ideally be orthogonal to the other; i.e. there is a phase change of 90° between the outputs, v ₁ and v ₂ . When the outputs, v ₁ and v ₂ , are orthogonal and everything else being perfect, the output from the MR sensor which is given by the functional term arctan (v ₂ /v ₁ ) varies linearly with the angle of rotation of the magnetic field. To achieve this orthogonality between the outputs v ₁ and v ₂ , the Wheatstone bridge circuits are orientated such that the resistance elements of the two circuits are such that one circuit is at an angle of 45° with respect to the other about the axis of rotation of an external magnetic field, which axis is normal to the circuit plane. Rotation of the magnetic field would therefore lead to a change in resistance of the GMRs differently in each Wheatstone bridge circuit leading to the output from one Wheatstone bridge circuit being out of phase with the output from the other Wheatstone bridge circuit. When the phase difference is 90° the output signals are said to be orthogonal. In reality, v ₁ and v ₂ are not perfectly orthogonal and, therefore, the arctan of the ratio between v ₂ and v ₁ results in a non-linear output. Therefore, determination of the angular rotation sensed by the sensor (e.g. degree of pedal depression) is more challenging.

Ideally, in the absence of any external magnetic field, the resistances of each of the MR resistors in their respective Wheatstone bridge circuit are such that each Wheatstone bridge circuit is in a perfectly balanced state, i.e. the voltages v ₁ and v ₂ across the bridge circuits are zero and each of the MR resistors in each of the Wheatstone bridge circuits are exposed to the essentially uniform background magnetic field. For a typical MEMs device, the AMR or GMR resistors are laid at a micro-scale onto a chip. However, the dimensional and material chemistry tolerances and the resultant electrical property tolerances achievable at the micro-scale are limited. Moreover, the MR resistors in each Wheatstone bridge circuit do not experience the same level of change in magnetic field when the magnet forming part of the sensor is rotated. Thus, when MR sensors are used in MEMs devices such sensors give low signal to noise ratios and a non-linear response, requiring corrective processing by electronic circuitry, typically using an application-specific integrated circuit(ASIC).

As a result complicated mathematical algorithms as discussed below are needed to provide a linear response by compensating for inherent discrepancies in the Wheatstone bridge sensors. In addition to correcting for the non-orthogonality of the output signal from each Wheatstone bridge circuit, the output signals are corrected for offset and amplitude. The output voltage, v ₁ and v ₂ from each of the Wheatstone bridges is given by the equations: where:

The offset signal for v ₁ and v ₂ is given by:

Temperature corrected offset is represented by:

KT are parameters that are established by experiment in a production

The output voltage, v ₁ and v ₂ is corrected for the temperature corrected offset by subtracting the calculated temperature corrected offset values determined by equations E and F above from the output values, v ₁ and v ₂ . The offset corrected output signals, v ₁₁ and v ₁₂ , are then corrected for amplitude and phase by using discrete Fourier transform as shown below:-

where

e non-ort ogona s gna s correcte us ng a matrix formulation as shown below:

Through this processing, the signals, v ₁ and v ₂ , are made orthogonal. The arctan of the ratio of results in a linear output function as demonstrated in the equation below:-

The cost and power demand of the signal processing electronics to correct the output signals from the non-contact sensor are higher than those for the sensor element. Moreover, the requirement to process complex mathematical algorithms to provide a linear response removes the possibility of using cheaper, less sophisticated processors. Even if more sophisticated processors such as ASICs are used to derive a linear response as per the above requirements, any errors introduced during the laying down of the resistors during production may propagate through such complex algorithms and mathematical equations which can cascade into significant uncertainty in the derived linear relationship. There has been a need for a non-contact sensor that does not require extensive processing power, yet which is low cost to manufacture and provides a sufficiently accurate and linear response with a sufficiently high signal to noise ratio.

Harada et al (US 2011/0246133 Al) and Hinz (US 2010/085038 Al) teaches a non-contact rotation sensor comprising sets of rotationally arranged magneto-resistive elements and Hall elements laid over a given surface area of a circuit board. Each magneto-resistance element provides a magneto-resistance effect with respect to the magnetic field. Each Hall element provides a Hall effect with respect to the magnetic field. The plurality of magneto-resistance elements are arranged in the magneto-resistance element region so as to cause a phase difference between output signals of the magneto-resistance elements. The plurality of Hall elements are arranged in the Hall element region so as to cause a phase difference between output signals of the Hall elements. The magneto-resistance element region and the Hall element region at least partially overlap with each other. The detection circuit includes a comparison section, an angle computing section, and an output section. The comparison section compares an output level from each Hall element with a predetermined threshold value level, and provides a comparison result for each Hall element. The angle computing section calculates a calculation angle corresponding to the relative rotation angle according to an output signal from each magneto-resistance element. The output section compares the calculation angle with a predetermined threshold value, and provides a comparison result for each magneto-resistance element. The output section outputs a signal corresponding to the relative rotation angle based on the comparison result of the output section and the comparison result of the comparison section. Harada et al attempts to miniaturise the non- contact sensor over a given surface area of the circuit board. The number of sensing elements based on magneto-resistive elements is limited to two as Harada et al aims to improve the rotation angle accuracy by comparing signals from the sensing element based on the magneto-resistive elements with signals generated by Hall elements that are partially overlapping the magneto resistive elements. Hinz also attempts to reduce the dimensions of the mechanical components and the number and size of the sensory and electronic components. Hinz (US 2010/085038 Al) teaches a non-contact angle sensor comprising eight approximately sector-shaped magneto-resistive elements of which four are connected in the form of a bridge circuit. The magneto-resistive elements form a first and second bridge circuit. The elements of the first bridge circuit alternate with the second bridge circuit. As with Harada et al, the non-contact sensor in Hinz is taught to comprise only two sensing elements generating only two signals which are combined to generate one linear output.

In an alternative arrangement, Uenoyama et al (US 6366079 Bl) and Sony Corp (GB 2034053 A) also relate to a non-contact sensor that includes sensing elements consisting of magnetoresi stive material. Uenoyama et al specifically relates to a pair of magnetoresistive resistors forming a first sensing element in a bridge circuit, and another pair of magnetoresistive resistors form a second sensing element in the same bridge circuit. Sony Corp more specifically relates to apparatus for sensing relative displacement of a magnetic grating using sensing elements consisting of anisotropic magnetoresistive stripes. In Sony Corp the first and second sensing elements are arranged over a given area of the circuit board such that the tracks from the sensing elements are interlaced. Likewise, the tracks forming the sensing elements in Uenoyama at al are arranged in a comb-shape. Laying the tracks in an interlaced fashion or comb-shape taught in Sony Corp (GB 2034053 A) and Uenoyama et al (US 6366079 Bl), respectively, limits the number of sensing elements that can be laid over a given surface area of the circuit board. In both cases, the number of sensing elements that can be laid over a given area of the circuit board is two. Any of the embodiments described in Sony Corp can be applied to a rotation sensor. A magnetic grating is affixed to the rotating disk (onto which the sensor elements are placed). The divisions of the magnetic grating are disposed along the radii of the disk and, unlike the magnetoresistive linear displacement sensors described in Sony Corp; the rotation sensor requires that bias magnetic field must be oriented at an angle relative to the radii of the disk. In an attempt to increase the number of signals generated from a given non-contact sensor, Thomas et al (US 2009/0174395 Al and US 2011/0248708 Al) and Yabusaki et al (US 2010/0156397 Al) take a different approach to Harada et al, Hinz, Uenoyama and Sony Corp whereby they generate/derive a second signal from a first signal wherein the second signal has a quadrature relationship with the first signal. A rotational sensor requires the generation of a pair of sine/cosine signals in order to determine the signal, and this is conventionally done by placing two Hall elements with a 90° offset, however any misalignment of the elements results in a phase error between the signals which leads to an angular error. Thomas et al (2009 and 2011) and Yabusaki et al relates to a rotational sensor which minimises the effects of sensing element misalignment with respect to a magnet. Thomas et al (2009 and 2011) and Yabusaki et al disclose a rotational sensor which generates a linear output from phase shifted waveforms generated by magnetic sensors (sensing elements made up of linear Hall effect devices and/or magneto-resistance resistors can be used) using signal processing. A signal inversion module can be coupled to a signal generation module for inverting the first waveform to provide a first inverted waveform and for inverting the second waveform to provide a second inverted waveform. A signal processing module coupled to the signal inversion module provides an algebraic manipulation of the subset of the first and second waveforms and inverted waveforms to generate a linear output voltage. Thomas et al (2009 and 2011) and Yabusaki et al teach that a single waveform, which can be sinusoidal, is used to generate a corresponding cosine signal. Trigonometric identities can be applied to the sinusoidal input to construct its corresponding cosine signal.

It is apparent from the above documents that the maximum number of sensing elements in a given area of the circuit board is two which results in two measured signals. At least two signals are necessary to produce a meaningful result from the sensing element, such as to measure the angular rotation of the device. The generation of a second "virtual" signal from a first "real" signal of a sensing element requires additional processing power to derive the virtual signal from the measured signal. Therefore, there is a drive towards a non-contact sensor which can accommodate multiple sensing elements to generate multiple signals in a given area of the circuit board without the need for additional processing power. Summary of the Invention

The present invention mitigates the above problems by zoning the sensor elements on the sensing board. Zoning refers to the sensor elements being arranged in inner- and outer-zones on circuit board whereby inner- and outer- refer to the inside and outside over a given surface area of the circuit board, respectively, i.e. relative positions of the sensing elements with respect to the other sensing element and relative to a sensitive axis of the sensor. Zoning provides the present invention with the flexibility to arrange a sensor consisting of multiple sensing elements on a circuit board of a given surface area.

The present invention provides a sensor assembly comprising at least one sensor and a magnet, said at least one sensor comprises a first sensing element which in itself comprises a plurality of rotationally symmetrically arranged sets of resistors lying within an outer zone of the sensor, at least one resistor of the sets of resistors being a magneto-resistive resistor. The at least one sensor also comprises a second sensing element which in itself comprises a plurality of rotationally symmetrically arranged sets of resistors, a portion of the second sensing element lying in a zone radially inward of the outer zone, at least one resistor of the sets of resistors being a magneto-resistive resistor. The sets of resistors of the first and the second sensing elements are rotationally symmetrical about a sensitive axis. The magnet overlaps the first and second sensing elements and has a rotational axis that is co-axial with the sensitive axis, wherein the rotational axis of the magnet is perpendicular to the magnetic axis of the magnet. The magnetic axis is the extension of a line between the north and south poles of a magnet about which there is rotational symmetry of the magnetic field in absence of an external field, i.e. a straight line joining the two poles of a magnet. Preferably, the magnet is spaced apart from the sensor with its magnetic axis substantially perpendicular to the sensitive axis (within the confines of manufacturing tolerance) so as to optimise the interaction of the magnetic lines of flux of the magnet with the magneto-resistive resistors of the sensing elements. Proper positioning of the magnet with respect to the sensor so that the magnetic lines of flux are substantially symmetrically arranged around the resistors of the sensor ensures a more balanced sensor. Thus, the output signal generated from the sensor is more attributable to the rotation of the magnet about the sensitive axis whilst minimising errors as a result of any imbalance between the position of the magnet and the resistors of the sensing element. A rotation of the magnet about the sensitive axis will have an influence on the sensor. Optionally, the present invention also provides a sensor comprising:

a plurality of sets of resistors, each set arranged to form a respective sensing element and comprising at least one magneto-resistive resistor; wherein at least one of the sensing elements is spatially arranged to overlap or coincide with the area occupied by another of the sensing elements.

By having at least partial spatial coincidence or overlap between sensing elements, the magneto-resistive resistors in each of the sensing elements are exposed to more closely the same magnetic field and therefore experience more closely the same changes in magnetic field; which in turn reduces the need to correct for discrepancies arising where each of the sensing elements are exposed to different areas of the magnetic field or different lines of magnetic flux. Preferably, the magneto-resistive resistors of said one sensing element alternate with the magneto-resistive resistors of the other sensing element. Having the magneto-resistive elements in one sensing element alternate with the magneto-resistive resistors of the other sensing element provides further improved similarity of changes of magnetic field experienced by the magneto-resistive resistors in each of the sensing element. The alternating magneto-resistive resistors of the two sensing elements may be concentrically arranged. Preferably, the magneto-resistive resistors in each of the plurality of the sensing elements are rotationally symmetrically arranged. Due to the symmetrical arrangement of the magneto-resistive resistors about a common axis making up a sensing element, the present invention permits multiple sensing elements to be laid down more compactly on a circuit board. This is made possible since at least one of the magneto- resistive elements in one of the multiple sensing elements lies between the other of the magneto-resistive resistors in another of the sensing elements. In prior art sensors, multiple sensing elements each formed by a Wheatstone bridge are laterally offset from each other, thereby occupying more surface area which in a MEMS type sensors can present challenges in terms of space. The pattern of magneto-resistive resistors making up a sensing element in the present invention permits multiple sensing elements to be formed coincidently or overlapping in the same space, without the need to space them out one beside another. Preferably, each magneto-resistive resistor is formed in a distinct sector and arranged in an alternating manner with similar resistors of another sensing element, in a concentric circular fashion. Preferably, each sector of the sensor comprises one of the plurality of resistors of the first sensing element nested within one of the plurality of resistors of the second sensing element. This arrangement of a plurality of sensing elements permits at least one concentric sensing element to be angularly offset relative to the other by 45° about their mutual axis. Each of the sensing elements in the sensor thereby tends to be exposed to the correct changes in magnetic field, as a sensor magnet is rotated. This enables the output signals from each of the sensing elements to be more accurately controlled than in prior art set-ups, more particularly it allows the output signals from each sensing element to be controlled so that they are more closely orthogonal. Preferably, the sensor magnet is rotatable about the mutual axis to generate a substantially sinusoidal output from each of the sensing elements.

Preferably, the magneto-resistive resistors comprise tracks laid down on a substrate for example, printed circuit board. This can therefore be manufactured at the macro scale so that dimensional tolerances can be accurately controlled. The macro scale allows the sensing elements to maximise the magnetic flux incident over each sensing element whilst ensuring that the flux over each element remains substantially correct. To keep costs down, preferably the substrate comprises an FR-4 board commonly used in electronic circuit boards. Since individual layers of the materials used to form the tracks of the magneto- resistive resistors are in the order of only 0.5nm to 10nm thick and typically the surface roughness of an FR-4 type circuit board is in the order of 5μm from peak to peak, the tracks could be disrupted by the peaks and troughs in the FR-4 board surface. To mitigate this effect, preferably a smooth coating layer is added to the FR-4 board prior to forming the tracks of the magneto-resistive resistors on the circuit board. Preferably, the smooth coating has a surface roughness of 5nm or less; preferably lnm or less; more preferably 0.5nm or less. Preferably, the smooth coating comprises a dielectric material; preferably a solder resist; more preferably poly (methyl methacrylate) (PMMA). The responsiveness of the sensor and the ability to detect a measureable value is dependent upon the magnitude of the change in the resistance of the magneto-resistive elements when exposed to a changing magnetic field. This in turn is dependent upon the surface area of the magneto-resistive elements covered by the magnetic field. Thus, the greater the surface area of the magneto-resistive elements that is exposed to a given magnetic field, the greater the magnitude in the resistance change when subjected to a changing magnetic field. A macro- scale sensor is therefore advantageous for this reason also. Preferably, the tracks of the magneto-resistive resistors comprise meanders; more preferably, the meanders comprise parallel portions, for high compactness. Forming a pattern of meanders substantially increases the length of the magneto-resistive resistors and hence increases their overall resistance. The higher the overall resistance, the larger the change in resistance for a given change in the magnetic field to which the magneto-resistive resistors are exposed. Preferably, the tracks double back on themselves. This enables the pattern of connections between individual resistors, and between the resistors and voltage input and signal output leads, to be arranged around the periphery of the sensing element; again facilitating a compact and relatively easy to manufacture sensor design. More preferably, the tracks comprise contact pads arranged peripherally of the sensing elements. It is important that the physical characteristics of the magneto-resistive resistors in absence of a magnetic field in each of the sensing elements result in a balanced state, i.e. zero output voltage. If resistive properties of one sensing element behave differently to those of another sensing element with changing magnetic field, then this is reflected in the output voltages from each of the sensing element leading to discrepancies in the output signal. These discrepancies would need to be compensated, resulting in the requirement for further signal compensating processing and ultimately more processing power. To mitigate this effect, any of the sensing element geometries and individual magneto-resistive resistor layouts and geometries described above may be adopted, individually or in any suitable combination. Preferably, the magneto-resistive resistors comprise giant magneto-resistive (GMR) resistors that are able to provide a greater change in resistance with changing magnetic field in comparison to anisotropic magneto-resistive resistors (AMRs).

While exemplary embodiments refer to the use of GMRs, other magneto-resistive resistors may also be used in place of or in combination with MRs, AMRs, and/or GMRs. For example, tunnelling magneto-resistive resistors (TMR), colossal magneto-resistive resistors (CMR), and extraordinary magneto-resistive resistors (EMR), could also be employed. As conventional MR sensors are manufactured mounted on top of a processor chip (e.g. in the form of a MEMS device), the cost of an MR sensor system can only be reduced if the cost of auxiliary electronic components, e.g. ASICs, can be reduced. The present applicant has mitigated this problem, by preferably mounting the sensor outside of the chip. This permits the use of high precision manufacturing processes at the macro level for the production of non-contact position sensors for applications requiring position sensing for example, automotive controls. The magneto-resistive sensor can be manufactured using e.g. photo-resistive, plasma thin-film deposition to achieve high precision and therefore to achieve high signal response quality. This is possible because the high precision manufacturing technology will allow for tight control of the resistivity, geometry, positioning and orientation of the sensing elements and their constituent magneto-resistive resistors, as well as the components of such resistors, e.g. individual material (e.g. metal) layers. Through the use of the high precision manufacturing processes made possible by mounting the sensor outside of the chip, high signal to noise (S/N) ratio and highly linear response through relatively simplified signal conditioning can be achieved using minimal processing. This removes the necessity for an ASIC, enabling this to be replaced with less sophisticated and cheaper surface mount components that are readily available; leading to significant cost reductions. Regulations stipulate that any system relating to the control of a vehicle must contain a redundant sensor system in case of component failure. Preferably, the sensor provides a redundant output signal that is substantially similar to the output signal of the sensor. A redundancy circuit, acting as a fail-safe, is of the utmost importance in the application of accelerator or brake pedals in motor vehicles. If there is a discrepancy (above an acceptable threshold) between the output signal and a redundant signal, an alarm can be raised to alert users of the potentially dangerous fault. Preferably, the sensor provides duplicate output signals so as to provide this redundancy. More preferably, the redundant output signal is a duplicate of the output signal, however, due to the minor differences in the locations of the sensing elements the redundant output signal behaves as a check signal which can vary from the output signal by up to the threshold value as is compliant with regulations. To accommodate a redundant sensor system that is similarly responsive to the changing orientation of the magnetic field, preferably a redundant sensor system is included within the sensing system. More preferably, one of the magneto-resistive resistors in one of said plurality of sensing elements is nested within another of said magneto-resistive resistors in another of said sensing elements. More preferably, the resistors of each of the plurality of the sensing elements are co-planar or arranged in different horizontal planes i.e. arranged in a stack.

The resistors of the two sensing elements are both arranged such that they are rotationally symmetric and concentric. The rotational symmetry allows more Wheatstone bridges per unit area leading to increased sensing density; an important feature for applications wherein a rotational angle sensing circuit is required in limited space. The increased sensing density, therefore, means that the present invention does not need a flux concentrator to direct the magnetic flux of the magnet. An advantage of having multiple sensing elements on the sensing board (circuit board) within a given sensing area is that it enables the possibility of having a redundancy sensor system without requiring a second sensing board. A concentric arrangement also has the effect of requiring only a single magnet which is positioned co- axially to the sensor plane. The concentricity of the arrangement allows each resistor of a single sensing element to experience substantially equal magnetic flux as each of the other resistors of the same sensing element which insures that the resistance changes for each resistor in a single sensing element for a given relative rotation of the magnet. Although the resistors of the first sensing element are equidistant to the sensitive axis and the resistors of the second sensing element are equidistant to the sensitive axis, the distance between the resistors of the first sensing element and the sensitive axis is not necessarily equal to the distance between the resistors of the second sensing element and the sensitive axis. This may imply that the first and second sensing elements are in an outer- and inner-zone, respectively. Optionally, the pattern of resistors in the sensor is arranged such that the first sensing element is outwardly demarcated by the second sensing element such that the outer zone comprises the first sensing element and the second sensing element forms an inner zone.

The magnetic field provides substantial coverage such that each of the plurality of resistors of each sensing element experiences a magnetic field H causing a change in resistance for a given relative rotation of the magnet with respect to the sensor, such that a relative rotation of the magnet with respect to the sensor is detected as an angle change. The magnet is positioned such that the rotational axis of the magnet is axially co-aligned with the sensitive axis, and at least partially overlaps the planar surface of the sensor. In order for the output signal from one sensing element to be substantially orthogonal to the signal of the other sensing element, the sensing elements can be offset, with respect to each other, by an angle of 45° about the rotational axis (sensitive axis). Preferably, the magnetic lines of flux generated by the magnet overlaps the first and second sensing element and having a rotational axis that is co-axial with the sensitive axis, wherein the rotational axis of the magnet is perpendicular to the magnetic axis of the magnet.

An effective sensing area of a sensing element is defined as the portion(s) of the sensing element in which the resistor tracks contain "substantial meandering". Optionally, one of the plurality of resistors of the first sensing element is nested on three sides by either one or two of the plurality of resistors of the second sensing element. The first sensing element (in the outer zone) is demarcated on three of its four sides (its two radial sides and its innermost side) by the second sensing element. More particularly, the pattern of the resistor of the second sensing element is such as to define an area enclosed by three sides which is occupied by a resistor of the first sensing element. This pattern repeats around the sensitive axis of the sensor so that the resistors of the first sensing element area nested within the second sensing element. Extending radially from the meandering portions that define the effective sensing area are contacts that permit electrical contact with the resistors. For the purpose of the present invention, the term "substantially meander" is construed to comprise a pattern of repeated order of tracks that doubles back on itself.

By having effective sensing elements in inner zones and outer zones allows for the present invention to maximise the area of the sensing board attributable to the effective sensing area of each resistor of a sensing element on a given sensing board, wherein the area of the sensing board is shared between at least two sensing elements. Therefore high quality signals may be produced, including that of the aforementioned redundant sensor system (a duplicate or check signal) in case of failure - a necessary feature if used as a brake or throttle pedal in a motor vehicle in order to comply with regulations.

Although laying down the tracks at a macro level permits the use of high precision production methods, thereby removing some of the imprecisions inherent in micro-scale sensors, no two sensing elements will be entirely identical; thus leading to a phase change between the outputs from two separate sensing elements being different to π/2 required to produce a linear output. The above problem is mitigated by providing a sensor according to a further aspect of the present invention comprising:

a. set of resistors including at least one magneto-resistive resistor and arranged to form a sensing element, wherein the sensing element produces a first alternating signal;

b. a signal processor for generating a second alternating signal;

characterised in that;

the processor derives the second alternating signal from the first alternating signal such that the second alternating signal is orthogonal to the first alternating signal.

In order to provide a linear output, it is important that the signals used to derive the linear output are alternating; preferably in the present invention these alternating signals are largely sinusoidal in form. However, when taking the voltage across pairs of resistors in any of the sensing elements described above (Wheatstone bridge), the achievable resolution of the output sinusoidal signal is low. To improve upon this resolution, outputs are taken from each resistor pair of a given sensing element and amplified such that the variations in the output signals are more pronounced. By combining these amplified signals, e.g. by subtraction, similar noise in each signal is at least partially cancelled. This fulfils a long-felt need as it allows digital signals to be used in the field of angle sensing circuits which was previously unadvisable due to the high levels of noise when compared to analogue signals (analogue signals typically provide a higher resolution but they require a more costly and slower signal processing method compared to signal processing with a PIC microcontroller). The resulting difference of the amplified signals thus yields a sinusoidal signal of higher resolution than if the voltage output was taken across the pairs of resistors. This sinusoidal signal can therefore be used as a higher resolution version of vl or v2 described earlier.

A PIC chip is capable of generating a second (virtual) alternating signal (although preferably sinusoidal in form) from a first alternating signal. The present invention provides, a method of generating a linear signal from an output signal and a physical displacement using a sensor, comprising the steps of:

a. generating a first signal from at least one sensing element; b. deriving a second signal from the first signal; c. performing a trigonometric operation on the ratio between the first and second signal so as to generate a signal that is directly proportional to the angular displacement of the rotating object.

Preferably, the trigonometric operation can be an arctan function. Preferably, the second signal is a translation of the first signal such that the second signal is substantially orthogonal to the first signal. As the second (virtual) alternating signal is derived from the first alternating signal, a phase change of π/2 can easily be imparted to the second (virtual) alternating signal, thus removing errors incurred through variations in production of the two separate sensing elements required in the prior art whereby outputs from two separate sensing elements are combined to generate a single desired linear output. By virtue of the alternating nature of the two signals, the arctan operation on the ratio of the two signals results in a linear output suitable for determining the displacement of a pedal. The first signal used to generate the second (virtual) alternating signal can be determined or derived from either taking a potential difference across both pairs of resistors in a given sensing element based on the use of sensing elements in the prior art, Wheatstone bridge or by taking the potential at the junction of the two resistors in each pair of resistors in a given sensing element according to the present invention. Deriving a second "virtual" alternating signal from a first alternating signal removes the need to rotationally offset one sensing element at an angle of 45° with respect to another sensing element, as discussed above, in order to establish substantially orthogonal output signals. Optionally, where the first and second sensing elements of the sensor assembly are not rotationally offset (i.e. offset angle substantially 0°), a first real signal produced by the first sensing element can be used to generate a first virtual signal which is substantially orthogonal to the first real signal, and a second real signal produced by the second sensing element can be used to generate a second virtual signal which is substantially orthogonal to the second real signal. Deriving a virtual signal from a physically measured signal simplifies the deposition of the resistive tracks on the circuit board and, thereby, simplifies the manufacturing process, in particular to providing a redundant signal discussed above. Since the response time is an imperative feature to various embodiments of this invention (such as the angular sensor in an accelerator or brake pedal for a vehicle) from a safety point of view, hence the present invention permits the use of a PIC microcontroller which can easily handle the simple algorithm required to generate a translated (by π/2) "virtual" signal from a "real" measured signal, and then combine these two orthogonal waveforms to generate a linear voltage output curve by employing an inverse trigonometric function (preferably arctan). The present invention, by using a PIC microcontroller and amplifying the original waveforms, can generate a linear voltage output more quickly than a typical analogue signal processor and with greater accuracy than a typical digital signal processor.

Brief Description of the Drawings

Further preferred features and aspects of the present invention will be apparent from the claims and the following illustrative description made with reference to the accompanying drawings in which: Figure 1 is a schematic view of a pedal assembly according to an embodiment of the present invention, with part of the non-contact sensor shown separated.

Figure 2a is a schematic view of a pivot pin for a pedal assembly according to another embodiment of the present invention.

Figure 2b is a top/bottom view of a magnet with respect to the rotational axis of the sensor assembly.

Figure 2c is a side view of the magnet according to an embodiment of the present invention.

Figure 3 is a top and bottom view of a non-contact sensor motherboard according to another embodiment of the present invention.

Figure 4 is an expanded view of a sensing board which may be used in the embodiment of Figure 3.

Figure 5 is a schematic exploded view of layers of magneto-resistive metals forming the magneto-resistive tracks according to another embodiment of the present invention. Figures 6a-d are schematic illustrations of various different forms and patterns of GMR resistors laid down onto a sensing board according to different embodiments of the present invention.

Figure 7a is a circuit diagram comprising two sensing elements according to another embodiment of the present invention.

Figure 7b is a circuit diagram depicting a GMR sensor according to another embodiment of the present invention.

Figure 8a is a plot of the output signals from two sensing elements according to some embodiments of the present invention.

Figure 8b is a plot of the offset-compensated signals obtained from the output signals shown in Figure 8a.

Figure 8c is a plot of the offset-compensated signals shown in Figure 8b after amplification.

Figure 8d is a plot of the signals normalised for input to a PIC chip or similar standard processor, and the processed output of an arctan of their ratio to produce a linear function.

Figure 8e is a plot of the linear function shown Figure 8d that has been vertically translated for output from the PIC chip.

Figure 9 is a flow chart of a method according to another embodiment of the present invention for the removal of an offset potential using a statistical mapping method.

Figures 10a is a plot of the output signals from a sensing element of the sensor according to another embodiment of the present invention.

Figure 10b is a plot of the offset-compensated signals obtained from the output signals shown in Figure 10a.

Figure 10c is a plot of the amplified offset-compensated signals of Figure 10b.

Figure 10d are plots of V _real , difference between the amplified signals centred to about the x- axis; V _virtual derived from V _real ; and the output of the arctan of their ratio to produce a triangular function. Figure 10e are plots of the triangular function shown in Figure 10d after bounding and a vertical translation of the bounded triangular function.

Figure 11 is a circuit diagram showing circuits for compensating the outputs from each sensing element according to an embodiment of the present invention. Figure 12 is a flow chart showing the method for the generation of a virtual signal from a real signal used to generate a linear output of Figure 10f according to an embodiment of the present invention.

Detailed Description

The present invention relates to a non-contact sensor apparatus, more particularly a non- contact sensor based on magneto-resistive (MR) resistors or elements whereby when one component of the system is moved or displaced with respect to a magnet 18 a change in resistance in the MR resistors occurs, and the signals derived from the sensor can then be used to determine the magnitude of the displacement to a high level of precision. More specifically, the magneto-resistive elements are arranged to form sets of resistor pairs whereby their combination forms at least one magneto-resistive sensing element.

The sets of resistor pairs may be arranged such that the sensing element forms a Wheatstone bridge circuit.

A Wheatstone bridge circuit configuration provides an easy to use voltage output that is proportional to the magnetic field applied but insensitive to any variations in the absolute resistance of the magneto-resistive resistors. Each resistor of the resistor pairs may comprise a magneto-resistive resistor. However, the present invention is not limited to each resistor of each resistor pair being based on a magneto-resistive resistor; at least one of the resistors of the resistor pair can be a non-magneto-resistive resistor.

Non-contact sensors of the present disclosure can be used to sense angular displacement in any application that requires sensing of displacement. For example in a pedal assembly that can be mounted to a vehicle control platform so as to determine the extent to which the pedal has been depressed. For example, the pedal can be used for brake or throttle control.

Figure 1 shows a commercial embodiment of a pedal assembly 01 mountable to a vehicle. The pedal 02 is pivotally mounted to a base 03 such that when pressed by the vehicle operator, the pedal 02 rotates about a pedal rotation axis X-X. The base 03 comprises securing points 13 to secure or fix the base 03 to the vehicle body. The pedal 02 is rotatably mounted to the base 03 by means of a pivot pin 15 (see also Figure 2a) journaled in a pivot aperture (not shown) in the base 03. A splined surface 09 of the pivot pin is received in a complementary splined mounting aperture in an actuating arm 11 of the pedal 02, to ensure that the pedal 02 and pivot pin 15 rotate together about the axis X-X. The pivot pin 15 comprises a housing 17 for accommodating a permanent magnet 18. The magnet 18 is shaped so as to have rotational symmetry about the rotation axis X-X. To provide rotational symmetry, the magnet 18 is substantially or generally circular or disc in shape as shown in Figures 1 and 2a-c. However, the magnet 18 can have other shapes (for example quadrilaterals, hexagons, etc.) that permit the rotational symmetry about the hinge pin axis X-X which is equivalent to the sensitive axis 28. The north and south poles of the magnet 18 preferably lie, centred symmetrically, on opposite sides of the rotation axis X-X and are shown diametrically opposed and shaded differently from one another in Figure 2a. Therefore the magnetic axis 27 is substantially perpendicular to the axis of rotation of the magnet. The degree by which the magnetic axis 27 is substantially perpendicular to the axis of rotation of the magnet will depend on the manufacturing tolerances of the sensor, in particular the proper alignment of the magnet 18 with respect to the planar surface of the circuit board comprising the magneto-resistive sensing elements. In a particular example of the present invention, as shown in Figure 2c, a face of the magnet 18 having the largest cross-sectional area lies in a plane substantially perpendicular to the sensitive axis 28, i.e. the magnetic axis is substantially parallel to the planar surface of the sensor board.

Rotation of the pedal 02 about the axis X-X causes the magnet 18 to rotate about the sensitive axis 28 and thereby causes a change in magnetic field orientation relative to sensing elements on a sensing board 21 mounted to a motherboard 22 in a sensor housing 20. In the finished assembly 01, the sensing elements are also centred around or about the rotation axis X-X. Together with the magnet 18 they form the main components of the non- contact sensor. The magnet 18 rotates in direct proportion to the rotation of the pedal 02. Alternatively, the magnet 18 could be coupled in a manner such that rotation of the pedal 02 causes a rotation of the magnet 18 with a different known relationship: for example, by means of meshed gears, a belt and pulleys, a crank and connecting rod, a Geneva mechanism, a flexible drive shaft, or any other suitable mechanical transmission which imparts rotational movement of the magnet 18 when the pedal 02 is depressed. Figure 3 shows the sensing board 21 mounted and secured to the non-contact sensor motherboard 22. The motherboard 22 is mounted in the housing 20 using a snap-in mount (not shown). It should be understood that any suitable mount or housing may be used and as such are within the scope of the present invention. Electrical connections 25 between the motherboard 22 and the housing 20, as in Figures 1 and 3, permit outputs from the sensor 20 to be relayed to the vehicle control platform. The mount is positioned so that the sensing plane (i.e. major surface) of the sensing board 21 faces the exposed poles of the magnet 18 and is orientated such that the sensing plane is substantially perpendicular to the rotation axis X-X. The sensing board 21 is mounted spaced apart from the magnet 18 but disposed sufficiently close to it so that on rotation of the magnet 18, the change in magnetic field orientation is detectable by the sensing board 21, resulting in a measureable change in resistance. Preferably, the spacing is between 2 and 8 mm, but may also lie outside this range. A spacing effective to provide a consistent output signal is dependent on the magnetic field strength of the magnet 18.

The motherboard 22 comprises a plurality of electrical components that can be used in conjunction with the sensing board 21 in the determination of the magnitude of pedal displacement. The motherboard 22 may comprise diodes 31, operational amplifiers 32, capacitors 33, printed carbon resistors 34 and processor chips (PICs) 35 as shown in Figure 3. A printed carbon resistor 34 comprises thick- film carbon printed onto the motherboard 22. One purpose of the carbon resistors 34 is to generate an artificial signal that combines with the signal from the sensing board 21 so as to reduce the output from the sensing board 21 for input into a PIC 35, as described in more detail later. As each of the resistors forming sensing elements on the sensing board 21 behaves differently when subjected to a magnetic field, the output from the sensing board 21 can vary from one board to another. The length of one of the printed carbon resistors 34 can be tailored to the specific sensing board 21 properties produced e.g. during a particular production run, or for each board individually if necessary.

Referring to Figure 4, magneto-resistive resistor tracks 41 are laid down on the sensing board 21 to form a Giant Magneto-Resistive (GMR) sensor 45 comprising sensing elements. More generally, each sensing element may comprise an electrical circuit including at least one MR resistor. The arrangement shown comprises at least one such MR resistor on the sensing board 21, preferably a plurality of MR resistors preferably all on the sensing board 21, or more preferably a plurality of GMR resistors preferably all on the sensing board 21.

The sensing element circuit may comprise, for example, a Wheatstone bridge with a fixed reference voltage applied across the input terminals and an output voltage signal taken from across the output terminals and used to derive a displacement signal which varies substantially linearly with angular displacement of the magnet 18. Where four MR or GMR resistors of substantially equal geometry and equal "zero magnetic field" resistance are used, arranged substantially rotationally symmetrically about the magnet rotation axis X-X, the output voltage signal from the Wheatstone bridge will vary substantially sinusoidally as the magnet 18 rotates. The sensing board 21 is a relatively large scale fabrication, which enables the dimensions, positioning and material composition of its constituent resistors to be accurately controlled to provide the required precision in resistance values and resistor geometry; so that a usefully sinusoidal output voltage signal can be obtained.

Alternatively the sensing element may comprise a pair of resistors in series, with an input reference voltage applied across their outer end terminals and an output voltage signal taken from their junction. The output voltage may be determined with reference to any suitable system potential, for example (but not necessarily) the potential at one or other of the input terminals. In other words, the sensing elements may be formed as "half Wheatstone bridges", functioning as a potential divider, hereafter referred to as "resistor pair" sensing elements. It has been found that if the output signals from a pair of such resistor pair sensing elements are individually compensated for voltage offset, individually normalised for peak- to-peak voltage and then combined (preferably by subtraction), a particularly "clean" sinusoidal output signal results. Preferably the output signal from each resistor pair is amplified, conveniently as part of the peak-to-peak voltage normalisation step. It is thought that combining signals from a pair of resistor pairs in this way provides noise cancellation. Resistor geometrical and compositional accuracy therefore becomes less critical, so that fabrication of the sensing element on a smaller scale (e.g. even on the micro scale, e.g. as part of a MEMS device) becomes a possibility; with a usefully sinusoidal output signal still obtainable. On a smaller scale, eccentricity of individual resistors and sensing elements becomes less critical in avoiding undesirable distortion of the substantially sinusoidal signal, so it is even possible for the sensing elements to no longer spatially correspond or overlap. For example, side-by-side, non-spatially corresponding and non-congruent sensing element arrangements may be possible at smaller scales including at the micro-scale, capable of producing acceptably sinusoidal output signals. However a smaller scale generally means a smaller sensing element output signal and a lower signal to noise ratio.

An output signal which varies sufficiently linearly with rotational displacement of the magnet 18 may be obtained in any suitable way from a sufficiently accurately sinusoidal sensor signal, e.g. from an accurately sinusoidal signal obtained from a Wheatstone bridge sensing element or obtained from a set of two resistor pair sensing elements as described above. For example, the sinusoidal sensor signal may be transformed using a suitable trigonometric function, such as an arcsine or arccosine transformation. However, better linearity may result if a pair of such sinusoidal voltage signals are obtained, which are both of substantially equal amplitude and have matching voltage offsets, but which are substantially orthogonal. The ratio of these signals is then subjected to an arctangent transformation to obtain the output signal substantially linearly proportional to the angular displacement of the magnet 18. The substantially orthogonal sinusoidal signals may be obtained from two Wheatstone bridge sensing elements as described above; or from two sets of pairs of resistor pair sensing elements; or from one Wheatstone bridge sensing element and one pair of resistor pair sensing elements. In each case, one Wheatstone bridge sensing element or pair of resistor pair sensing elements is offset from the other by an angle of 45 degrees about the rotational axis X-X, so that its output signal is orthogonal to that of the other.

The positioning of the various sensor elements spatially coincident or overlapping with each other ensures that they all experience closely similar variations in the magnetic field; albeit with a different phase angle, depending upon their relative angular position about the rotational axis X-X. Preferably the resistors in the sensing element or sensing elements giving rise to a particular substantially sinusoidal signal are equi-angularly spaced about the rotational axis X-X; preferably all these resistors having the same geometry and all at the same radial distance from the axis X-X. Where two substantially sinusoidal, substantially orthogonal signals are used as described above, preferably all of the resistors in all of the sensing elements giving rise to these signals are preferably of the same geometry and are preferably equi-angularly spaced about and at equal radial distances from the axis X-X. As an alternative to using a sensing element/set of sensing elements at 45 degrees to another to obtain substantially orthogonal, substantially sinusoidal signals as described above, substantially orthogonal, substantially sinusoidal signals may instead be derived from a single Wheatstone bridge sensing element or from a pair of resistor pair sensing elements by processing the substantially sinusoidal signal thus obtained in order to produce the orthogonal signal (a "ghost" or "virtual" signal). For example, the substantially sinusoidal signal obtained from the sensing element or elements may be transformed by an arcsine function [or arccosine function]; π/2 is then added to or subtracted from the result if working in radians (± 90, if working in degrees); then this result is transformed by a sine [or cosine] function to obtain the orthogonal signal. The two orthogonal signals may then be treated as described above by taking the arctangent of their ratio so as to obtain the substantially linear angular displacement output signal.

As a safety feature, sets of sensing elements may be ganged or grouped together on the same sensor board 21, to provide dual outputs capable of being independently processed by separate electronics/processors, e.g. on the motherboard 22. Dual independent linear outputs both of which correspond to angular displacement of the magnet 18 can therefore be obtained. This provides redundancy, and error and component failure checking capabilities. The second set of signal processing electronics on the motherboard is identified by reference 82 in Figure 3. In any of the cases described above where there is a need to offset-compensate a substantially sinusoidal voltage signal, any suitable circuit or signal processor may be used to do this. Conveniently, a simple analogue circuit such as a voltage dividing resistor network may be used. For example individual sensing boards 21 or equivalent sensor element assemblies, or individual production batches of sensor boards/sensor element assemblies, may be calibrated with regard to voltage offset compensation by selecting appropriately rated trim resistors for use in the voltage divider; or by adjusting carbon film resistors 34 or trim potentiometers in the voltage divider. Similarly, in any of the cases described above where there is a need to normalise amplitudes of substantially sinusoidal voltage signals, any suitable circuit or signal processor may be used to do this. Conveniently a simple analogue circuit such as an operational amplifier 32 may be used, having an appropriately selected gain control resistor or appropriately adjusted carbon film resistor 34 or trim potentiometer used as its gain control resistor. Thus individual sensing boards 21, or equivalent sensing element assemblies, or individual production batches of sensing boards/sensing element assemblies, may be calibrated for signal amplitude normalisation. In any of the cases described above where there is a need to apply trigonometric transformations to a sinusoidal or angular signal, any suitable circuit or signal processor may be used. Conveniently, an appropriately programmed, inexpensive, general purpose digital signal processor, such as a PIC 35, may be used.

The change in resistance of the GMR sensor 45' s GMR resistors 42 is dependent on the physical characteristics of the tracks 41, e.g. cross-sectional and surface area of the tracks. Moreover, other magneto-resistive elements can be used in the present invention such as anisotropic magneto-resistive elements. However, GMR resistors 42 are preferable because of their ability to achieve a greater variation in resistance for a given change in the magnetic field.

The tracks 41 are laid down on the sensing board 21 preferably consisting of any substrate for example FR-4 board with a smooth coating 46 (see Figure 5), or a ceramic with or without such a coating, according to need. Typical circuit boards such as untreated FR-4 are too rough so that any tracks 41 (which can comprise individually deposited layers as little as a few nanometres thick) built up directly on the FR-4 board, would tend to be disrupted and broken up by the surface roughness of the FR-4. The coating 46 provides a smoother substrate for laying down the tracks 41. The sensing board 21 is fabricated as a flip chip which interconnects to the motherboard 22 via electrical contact pads 48. In the particular example shown in Figure 4, the electrical connection between each end of a track 41 and a corresponding contact pad 48 is made via a solder blob or join 47 located substantially around the perimeter of the GMR sensor 45. The flip-chip sensing board 21 has a width in the range of 0.5mm to 25mm; preferably 8mm to 15mm, more preferably 11mm. The smooth coating 46 preferably comprises a dielectric material; preferably a solder resist and more preferably PMMA. The coating 46 may have a surface roughness of 5μm or less; preferably Ι μm or less and more preferably 0.5μm or less.

In comparison to depositing the tracks 41 at a micro-scale, the tracks 41 on the sensing board 21 are deposited on a macro scale, enabling greater precision manufacturing processes to be used for tight control of the geometry and material composition of the tracks 41, and hence their resistance, position and orientation. Each of the tracks 41 may have a width 56 in the range of substantially 25μιη to 500μιη and preferably 25μιη to 250μιη. In the particular example of Figure 4, the width 56 of each track 41 is substantially equal to the minimum gap 57 between adjacent track portions.

A track 41 preferably comprises a pattern of meanders which comprise parallel portions that may be substantially arcuate, centred on the rotation axis X. Indeed, the tracks 41 may double back on themselves. The tracks 41 meander to occupy an area that is generally 'V shaped and mirror symmetrical. Each resistor 42 of a given sensing element is formed by the tracks 41 and occupies a distinct sector 51 of the sensing board 21 (see the exploded portion of Figure 4). As also shown in the exploded portion of Figure 4, a resistor 42e with contact pads 48el and 48e2 is nested within a resistor 42a with contact pads 48al and 48a2. Together the resistors 42a, 42e and the insulating gaps 57 between adjacent track portions occupy the entire sector 51. As the track width 56 is substantially equal to the insulation gap width 57, the resistors 42a, 42e together occupy substantially half of the area of the sector 51. As the track width 56 is relatively narrow, the resistors 42a, 42e can be made relatively long and still fit compactly within the sector 51.

To maximise the resistance of each resistor track 41, the pattern of the track 41 is designed such that the resistors 42 have near to the longest length possible so that each track 41 occupies the greatest possible surface area of the sensing board 21. By forming the tracks 41 in each sector 51 into a pattern of meanders, as discussed above, the length and thus the resistance of the resistor tracks 41 is increased. In the particular example shown in Figure 4, the tracks 41 are laid down in a tree-like pattern. A larger surface area of the tracks 41 increases the sensitivity to the change in magnetic field since more lines of magnetic flux are incident over the tracks 41 and individual resistors 42.

As shown in Figure 5, the GMR resistors 42 comprise at least two ferromagnetic metal layers 58 (five shown, but there may be as many as 20 or more) separated by non- ferromagnetic metal layers 59. The ferromagnetic conductive layers 58 may comprise, for example, a metal such as cobalt or iron and the non-ferromagnetic conductive layers 59 may comprise, for example, a metal such as copper or chromium. The thicknesses of the ferromagnetic conductive layers 58 and the non-ferromagnetic conductive layers 59 are in the range of O. lnm to 10nm; preferably 0.5nm to 1.5nm and more preferably 0.7nm to 1.2nm to ensure the presence of interlayer exchange coupling, e.g. copper, which is usually an excellent conductor, has a very high resistance due to electron scattering when it is only several atoms thick, and the resistance changes depending on the relative orientation of the electron spins surrounding the non-ferromagnetic conducting layer.

The resistors 42 may be formed using a number of suitable techniques including but not exclusively limited to photo-resistive, plasma or thin- film deposition technologies. A coating of magneto-resistive material, or successive alternating ferromagnetic/non- ferromagnetic layers, may be deposited on the smooth coating of the sensing board 21, or on one another, or vapour deposition or sputtering. The tree-like patterns of the resistors 42 tessellating to form the GMR sensor 45 may be created by removing selected portions of the coating/layers. Such selective removal may use laser cutting, water jet cutting, etching away parts between a printed protective mask, or the like. Alternatively, printing technologies can be used to directly apply a magneto-resistive coating to form the pattern and thus remove the necessity for further cutting methods. For example, the MR resistors can be fabricated and printed using the technique as taught in US9000764 (KARNAUSHE KO, DANIIL ET AL). Yet alternatively a printed "non-stick" mask may be used to prevent deposition of the magneto-resistive/ferromagnetic/non-ferromagnetic layers in the areas of the pattern which are to form the insulating gaps. The use of laser cutting allows selective portions of the coating to be removed on a small scale at a high precision and accuracy than other known removal methods, e.g. water jet cutting. A batch of sensors may be formed on a parent board using any of these techniques, with the parent board then cut up to form individual sensing boards 21.

These techniques allow the manufacturing process to achieve high levels of precision. For example laser cutting permits the minimum gap 57 between resistor tracks 41and the tracks themselves to be fabricated to a width of 25μm. This precision in fabrication provides a number of cascading advantages including: the minimisation of track width minimises cross-sectional area as well as increasing the length of the resistor tracks 41 that can be laid within a sector 51 thus maximising the absolute resistance of each sector 51 in the sensing board 21. The absolute resistance of a resistor is defined as the resistance of the resistor 42 in the absence of a magnetic field. Higher resistor resistances result in lower currents being present in the sensing elements which consequently causes component heating to be reduced, therefore increasing the life and improving consistency and reliability of performance of the sensor 45 and other neighbouring auxiliary components, e.g. amplifiers 32, capacitors 33, and processors 35.

Resistance heating is further mitigated by the scale at which the GMR sensor 45 is designed. At the macro scale of sensor 45 compared to the micro-scale used in prior art MEMS devices, heat dissipation is faster. The scale of the GMR sensor 45 has the added advantage that no flux concentrator is required to concentrate the magnetic flux from the magnet 18 into the resistors 42 as is required by conventional AMR or Hall Probe sensors to achieve similarly precise measurements. This results in lower component complexity and cost.

In the GMR sensor 45 shown in Figure 6a, eight resistors 142a-h are arranged to form two Wheatstone bridge type sensing elements 119a and 119b, which are identical to each other, except that sensing element 119a is rotated 45 degrees about the axis X (or sensitive axis) with respect to sensing element 119b. The resistor tracks of sensing element 119a are shown in black, whereas the resistor tracks of sensing element 119b are shown crosshatched in grey. The resistors of the two sensing elements, 119a and 119b, are all concentrically arranged about the axis X(sensitive axis) in eight adjoining sectors. The resistors 142a, 142b, 142d, 142c of sensing element 119a alternate with the resistors 142e, 142g, 142h, 142f of sensing element 119b. A circuit diagram of the resistors arranged to form the two sensing elements 119a and 119b is shown in Figure 7a. The sensing element 119a comprises four resistors 142a-d connected to form the first Wheatstone bridge. The sensing element 119b comprises four resistors 142e, 142f, 142g, 142h connected to form the second Wheatstone bridge. The 45 degree rotational offset between the two sensing elements 119a and 119b ensures that the output signals from each of the sensing elements 119a, 119b are substantially orthogonal (i.e. the two signals are π/2 out of phase). Output signal potential difference V ₁ is taken between (i) the junction of resistors 142a, 142b and (ii) the junction of resistors 142c, 142d using potential differentiator 61a. Output signal potential difference V ₂ is taken between (i) the junction of resistors 142e, 142f and (ii) the junction of resistors 142g, 142h using potential differentiator 61b. The ratio V ₁ / V ₂ is used to determine the angular displacement of the magnet 18 which in turn provides displacement of the pedal 02. The input reference voltage 80 is applied between (i) the junction between resistors 142a, 142c and the junction between resistors 142f, 142g and (ii) the junction between resistors 142b, 142d and the junction between resistors 142e, 142h. Any number of resistors may be alternatingly arranged to form different sensing elements whilst retaining sensing element concentricity and rotational symmetry, thus maintaining the system function of producing a potential difference that reflects a changing magnetic field.

For example, the plurality of resistors of each of a first and a second sensing element are arranged in sectors; optionally each sector comprising one of the plurality of resistors of the first sensing element is nested within one of the plurality of the resistors of the second sensing element. In an alternative example of the present invention, at least one said set of resistors in the first sensing element is rotationally offset with respect to another said set of resistors in the second sensing element. Further alternatively, one of the plurality of resistors of the first sensing element is nested on three sides by two of the plurality of resistors of the second sensing element, i.e. one of the plurality of resistors of the first sensing element is nested in the area surrounding three sides by two of the plurality of resistors of the second sensing element. Further alternatively, the first sensing element is outwardly demarcated by the second sensing element such that the outer zone comprises the first sensing element and the second sensing element forms an inner zone.

In the arrangement shown in Figure 6b, eight rotationally symmetrically arranged sets of resistors forming a first pair of Wheatstone bridge sensing elements are arranged in an inner zone 250a, whose resistor tracks are shown with a solid fill. The inner zone 250a comprises an effective sensor area 250c and a path 250d leading to a contact for electrical connection to an electrical source. The path (connector) 250d is defined as the inner zone 250a subtracted by the effective sensor area of the inner zone 250c (250d = 250a - 250c). The path 250d may be made of the same material as the effective sensing area 250c but due to insubstantial meandering, and therefore a relatively small surface area, it makes a negligible contribution to the change in resistance of a resistor. A further eight rotationally symmetrically arranged sets of resistors whose tracks are shown with a cross-hatched fill are arranged in an outer zone 250b, i.e. the sensing elements and their constituent resistors are arranged in concentric rings. Thus, the patterns of the individual sensing elements are arranged such that one of the sensing elements 250b is outwardly demarcated by an inwardly located sensing element 250a, i.e. the outer zone comprises a sensing element and the inner zone comprises another sensing element. The resistors are arranged in eight identical sectors 251, each containing a resistor 242a of the inner zone 250a (comprising 250c and 250d) and a resistor 242b of the outer zone 250b. The electrical connections between the resistors 242a, reference voltage and potential differentiators of the inner zone 250a may be the same as is shown in Figure 7a; referring to Figure 6a for the individual positions of the resistors in the inner zone. Likewise, the electrical connections between the resistors 242b, reference voltage and potential differentiators of the outer zone 250b may be the same as is shown in Figure 7a; referring to Figure 6a for the individual positions of the resistors in the outer zone. Regulations stipulate that any safety critical system must contain a redundancy system in case of component failure. This manifests in the provision of a second, independent set of signal processing components 82 mounted on the motherboard 22 and corresponding, independent further sensor elements on the sensing board 21. Preferably, the redundant output signal is substantially equivalent to the original signal, i.e. the redundant signal replicates or substantially duplicates the original signal. The inner and outer zones 250a and 250b shown in Figure 6b may be used to provide sensing elements for a pair of such independent sensing systems and yet still share the same magnetic field of the magnet 18 in a compact manner.

The example shown in Figure 6b shows four Wheatstone bridge circuits of four resistors each of which can produce a total of four signals. The signals can be provided as input signals to the virtual signal generator. As a result, the four measured signals can effectively generated a total of eight signals from a single sensing board as currently shown in Figure 6b which can be processed to yield four linear outputs. Equally, the examples shown in Figures 6c and 6d show two Wheatstone bridge circuits of four resistors each whereby the two Wheatstone bridge circuits can produce a total of 2 signals. As with the example of Figure 6b, the signals produced by the Wheatstone bridge sensor elements of 6c and 6d may be provided as input signals to produce orthogonal virtual signals. Alternatively, the patterns of resistors can be fabricated such that the resistors of the inner 250a and outer 250b zones are not placed within the same sectors 251 of the sensor 45. The first and second zones 250a and 250b are shown as co-planar, but sensors in parallel planes or other offset arrangements are also possible. Again, here a double zoned system is considered but by extension, any number of zones of resistor tracks 41 could be used. Figure 6c shows an arrangement of resistor tracks similar to those of Figure 4. In Figure 6c, one of the plurality of resistors (e.g. 42h) of a sensing element is nested on three sides by one of the plurality of resistors of another sensing element (e.g. 42d); more specifically the area enclosed by one of the plurality of resistors of a sensing element is occupied by one of the plurality of resistors of another sensing element. Equally, as shown in Figure 6d, one of the plurality of resistors (e.g. 44e) of a sensing element is nested on three sides by two of the plurality of resistors (e.g. 44a and 44d) respectively of another sensing element; more specifically the area enclosed by two resistors (e.g. 44a and 44d) of a sensing element is occupied by one of the plurality of resistors (e.g. 44e) of another sensing element. For example in a clockwise direction (see Fig. 6d), the resistor 44e in the outer zone is nested between the resistors 44d and 44a; the resistor 44f in the outer zone is nested between the resistors 44a and 44b; the resistor 44g in the outer zone is nested between the resistors 44b and 44c and the resistor 44h in the outer zone is nested between the resistors 44c and 44d. In both cases, a portion of one sensing element lies in a zone radially inward of the sensing element forming the outer zone.

By nesting on three sides, increases the resistor surface area on the sensing board which allows for the use of a more compact sensing board, and therefore a smaller and less costly magnet as the radius of the concentrically arranged sensing elements can be reduced implying that a smaller magnetic field is required to have the same effect in the resistance than if a larger radius concentric arrangement was employed.

The corresponding circuit diagram for resistors 42a-42d in Figure 6c and 6d is shown in Figure 7b. Resistors 42a and 42b together form a first resistor pair sensing element 343a. Similarly, resistors 42c and 42d together form a second resistor pair sensing element 343b. An input reference voltage 80 is applied between the end terminals of the resistor pairs 343a, 343b which are connected in parallel. A first output voltage signal Vn is taken between the junction in resistor pair 343a (i.e. the junction between resistors 42a and 42b) and an arbitrary system potential (in this case the reference potential at the junction between resistors 42b and 42d is shown). Likewise, a second output voltage signal V ₁₂ is taken between the junction in resistor pair 343b (i.e. the junction between resistors 42c and 42d) and an arbitrary system potential (in this case again the reference potential at the junction between resistors 42b and 42d). The signals V ₁₁ and V ₁₂ may each be compensated for offset voltage and normalised for amplitude as described above. The signals thus treated may then be combined (subtracted) to obtain a substantially sinusoidal output signal. This signal may then be transformed by a trigonometric function, with or without first generating a "ghost" signal, all as previously described; to obtain a linear output corresponding to the angular displacement of the magnet 18.

The nested resistors 42e-42h may be interconnected similarly to the resistors 42a-42d, as indicated in Figure 7b. Thus these nested resistors may be used to provide a second, independent output channel for a second, independent linear output corresponding to the angular displacement of the magnet 18. The resulting GMR sensor therefore has redundancy allowing output error and component failure checking.

In Figures 6a-6d the outer edge of the sensing board 21 and the contact pads 48 are omitted for simplicity. However the tracks of the individual resistors which are clustered together to form the GMR sensor 45 again preferably terminate at the periphery of the cluster, so that the connector pads can be formed around the periphery of the sensor board, for compactness and ease of connection. Figures 6a-d show some of the many possible variations of tessellating patterns possible to arrange multiple Wheatstone bridge sensing elements onto a sensing board, however the present invention is not limited by the examples shown in Figures 6a-d. For example, whilst the pattern of resistors in Figs 6a-d are arranged to provide up to four sensing elements, a similar 'nested' arrangement of resistors can be adopted to provide any number of sensor elements.

The rotation of the magnetic field is dictated by the rotation of the magnet 18 about the pivot pin axis X-X. This rotation changes the orientation of magnet's field passing through each magneto-resistive resistor 42a-h. As the orientation of the magnet's magnetic field changes, the magnetic moments of the ferromagnetic layers 58 also re-orientate. In the case of a GMR resistor, this in turn causes a change in the spin-dependent scattering of conduction electrons and thus a change in electrical resistance of the zoned resistor structure shown schematically in Figure 5. As the resistance of the resistors changes, in the case of the two Wheatstone bridge sensing elements shown in e.g. Figure 7a, the potential differences V ₁ and V ₂ measured at the bridge outputs, exhibit a substantially sinusoidal behaviour as shown in Figure 8a. As can be seen in this Figure, the signals are offset from the system ground potential 78 (referred to as 0 volts). This offset is a result of the manufacturing process used to create the resistor tracks 41. Although high precision techniques are used in the manufacturing process there still remains (but to a lesser extent) some uncertainty in track dimensions that is discernible during track resistance measurement. These discrepancies in the resistance between tracks 41 in the absence of a magnetic field, produces the offset in potential difference as shown in Figure 8a.

In order to determine the offset potential, the sensing board 21 is subjected to conditions experienced during operation. The magnet 18 used in the pedal assembly 01 is rotated above the sensing board 21 as it would be during operation and the potential differences Vi and V ₂ are measured versus angular displacement of the magnet. The midline value 72i and 72 ₂ , of each signal V ₁ and V ₂ respectively, is then calculated using equation (1):

Subtracting the midline potential difference 72i, 72 ₂ from the signals V ₁ and V ₂ respectively centre each signal at the ground potential, thereby compensating the signals for their different fixed voltage offset components. However this would also result in the minima 731, 73 ₂ of the signals being less than the ground potential 78. This poses a potential problem to processing the signal as only positive potentials can be passed into the PIC 35. For example, a typical PIC may be rated 0 - 5 volts. To compensate the offset signal, an offset adjustment potential is assigned a value such that the minimum potential of the offset- compensated signal will always remain greater than the system ground potential 78. Hence this adjustment potential must be less than either of the signal minima 731, 73 ₂ ; but preferably only slightly less, in order to optimise the resolution and accuracy of the PIC processing calculations. A number of ways to achieve positivity in the potential exist, either through modelling the variation in resistance statistically or determining it experimentally. One empirical approach is to determine the signal minima 731, 73 ₂ and subtract these values from each recording of voltage V ₁ and V ₂ . A safety factor (e.g. 0.2V) can then be added to this result so as to ensure that the resultant Vr and V never become negative, this process is given explicitly in the equations below:

and Where: = offset-compensated value of V ₁ voltage signal

= offset-compensated value of V ₂ voltage signal

minimum value of V ₁ voltage signal

minimum value of V ₂ voltage signal

δ = a safety factor

Conversely, this approach could be performed by subtracting each recorded V ₁ and V ₂ from the respective signal maxima. As discussed above, a safety factor (e.g. 0.2V) can then be added to this result so as to ensure that the resultant Vr and V ₂ ' never become negative.

Another empirical approach that can be utilised in order to ensure that the potential being input into the PIC chip 35 is positive is to experimentally determine the amplitude of the sinusoidal behaviour of each output. The output from each sensing element is subsequently added to half of this amplitude, a safety factor is also added as below:

Where: offset-compensated value of V ₁ voltage signal

offset-compensated value of V ₂ voltage signal

Amp. ₁ = amplitude of the sinusoidal V ₁ voltage signal

Amp. ₂ = amplitude of the sinusoidal V ₂ voltage signal

δ = a safety factor

Figure 9 shows one method to remove the offset potential from the each of the sensing board 21 's output signals using a statistical mapping method. In step 90, the absolute resistance of each individual track is determined by, for example, measurement using a multi-meter in the absence of an influencing magnetic field. Measurements of the magneto- resistive resistor's resistance are then repeated but at different orientations of the magnetic field so as to determine the relationship between the resistance of a magneto-resistive resistor and differing orientations of magnetic field. For example, repeat measurements are taken at orientations whereby the reading, which gives an indication of resistance e.g. current or voltage, from the multi-meter is either a maximum or minimum. By knowing the relationship between the resistance of each GMR resistor 42a-h for known orientations of the magnetic field, the maximum magnitude of variation in the GMR resistor' resistance can be determined through calculation in step 91.

In step 92, to provide for the inherent uncertainty in the variation of the signal due to the physical characteristics of the tracks, an error function, e.g. a Gaussian distribution, is fitted to the repeated measurements of the resistance variation. This error function fitting minimises the effect that limited repeat measurements in production testing have on the subsequently calculated offset adjustment potential. This variation in resistance is proportional to the change in potential difference relative to the zero-field value, taken from a given sensing element (Wheatstone bridge or resistor pair sensing element) for a given orientation of magnetic field. That is, the variation in resistance mimics the amplitude of the sinusoidal behaviour of the output signal from a given sensing element. This variation can be envisaged to have an upperbound output signal and a lowerbound output signal such that for a given orientation of magnetic field, the upperbound signal lies at the midpoint resistance for an arbitrary angular displacement plus a standard deviation dependent on the error function fitted. Conversely the lowerbound signal lies at the midpoint resistance for an arbitrary angular displacement minus a standard deviation dependent on the same fitted error function.

In step 93, the total maximum resistance of the track can therefore be determined by summing the absolute resistance and the maximum variance (represented by the upperbound of the resistance measurement when the magnetic moments of a resistor are all aligned with the direction of current flow) in the resistance. Thus by calculation using Ohm's law, the maximum amplitude and offset of the potential differences V ₁ and V ₂ can be known with an assigned uncertainty.

To ensure that the signal input to the PIC 35 is positive, a statistical outer bound on the offset adjustment potential (O.A.P) 74 is thus calculated (Step 94) to be: where the offset p.d. is equivalent to the midline p.d. 72 as determined in equation (1) above.

Experimentally, by recording the voltage signal from each sensing element and determining the minimum output signal from each of the sensing elements, the offset adjustment potentials 74i and 74 ₂ required to remove the offset potentials from the signals V ₁ and V ₂ are determined. Conversely, if the maximum output signal from each of the sensing elements was measured, then the offset p.d. in equation (6) would be added to the half maximum amplitude to generate the outer bound O.A.P. However, for the purpose of explanation, the outer bound O.A.P is based on the subtraction of equation 6. During operation, the interaction of the magneto-resistors' magnetic moments with the external magnetic field can give extreme responses that may not present themselves during production testing. Thus, determining the offset adjustment potentials by calculation is preferable so as to ensure a positive input signal to the PIC 35.

To apply the offset adjustment potential 74, a potential divider is used. The potential divider comprises carbon resistors 34 printed on the non-contact sensor motherboard 22. To obtain the positive signal for input into the PIC 35, the outer bound O.A.Ps are subtracted from the original input signals i.e. raw signals V ₁ and V ₂ (Step 95).

During operation, the interaction of the magneto-resistors' magnetic moments with the external magnetic field can give extreme responses which may lie outside of the more statistically probable responses. Therefore the maximum amplitude used in step 93 is assigned a value so as to encompass an acceptably low probability of occurrence, e.g. max. amplitude + 1 standard deviation defines an upper bound which encompasses 68% of the possible values for the maximum signal amplitude as defined by the Gaussian function. The inversion of the signal as a result of calculating the offset-compensated signal by equation (7) above does not require a rectification but could be rectified if needed. The result is shown in Figure 8b. The offset potential can be removed from the signals V ₁ and V ₂ through the use of the printed carbon resistors 34, the offset-compensated signals W and V ₂ ' are passed into the PIC 35. Alternatively a method exists whereby no signal processing is required to ensure that the potentials are always positive. As discussed above, if the all of the resistors in a sensing element 19 have the same resistance, the output signal from a sensing element arranged according to sensing element 119a as shown in Figure 7a will be substantially zero. To ensure that the output signal when the GMR sensor 45 is exposed to an external magnetic field is always positive, the absolute resistance value of one resistor in the sensing element can be adjusted. During production testing, a laser cutter is used to tailor, e.g. trim the length of one or more resistor tracks 41 in a sensing element thus changing its absolute resistance. The resistance value(s) is adjusted such that (using resistors within sensing element 119a according to Figure 7a as an example):

The advantage of performing the resistor track trim with a laser cutter is that the absolute resistance of the resistor track being trimmed can be measured in-situ whilst the trimming occurs, whereas using a water-jet cutter would require measurement of the resistance separately from the cutting process.

In order to increase the resolution of signal that is manipulated by the limited bit-rate PIC 35, the signals can be amplified such that the signal variation ranges

substantially between the positive values that a PIC 35 can manipulate, this results in the capability of using a lower bit-rate PIC 35 without loss of precision. In the present example of the invention this is between 0 - 5 volts as shown in Figure 8c. Once the signals W and V ₂ ', whether they are amplified or not, are input into the PIC 35, they undergo a translation such that the midpoint of the potentials' variation lies substantially about 0 volts.

Within the PIC 35, a trigonometric operation is performed between the two inputted signals so that they yield a linear output. To perform this operation, it is necessary that the amplitude 76 of the input signals are substantially identical. To achieve this, the signals are treated individually and normalised by dividing the offset-compensated signal 75 corresponding to each of the first or second sensing element by a fixed multiple of the maximum 77 value of the offset-compensated signal from each respective input signal, Vr and V ₂ '. For example, a normalisation may be performed such that each normalised signal varies between -1 and 1 volt as depicted in Figure 8d. After normalisation has occurred, the PIC is configured to perform a trigonometric operation on the ratio of the two signals. For example, an arctan trigonometric operation may be performed as defined below:

Since the phase relationship between the output signals from each of the sensing elements remains unaltered throughout the compensating operation, arctan of the ratio of the offset- compensated signals yields a linear output function of the form shown in Figure 8d. This output function must be translated so that all values of the signal are positive as the PIC 35 cannot output negative values. As described in Figure 12, step 99 translation is achieved by vertically shifting the signal by a suitable value such that the signal remains always positive. This value is a predetermined value that is pre-programmed into the PIC chip, further detail of this value is discussed below. The translated linear output voltage signal is shown in Figure 8e, in this case the gradient of the saw-tooth function is positive, if instead the relationship was used, the saw-tooth function would have a negative

gradient instead. This translated output 81 can be transformed such that the output can be mapped to an end-user preferred transfer function and thus relayed to the control platform as to provide a measure of the initial displacement. Since the only processing undertaken within the PIC 35 is the normalisation, trigonometric operation and translation, the processing power required is therefore substantially reduced in comparison to existing methods described in the introduction. Moreover, such a reduction in the operational steps to yield the linear output 81 improves the signal/noise ratio. It should be understood that the normalisation could occur over any potentials that the PIC chip 35 can manipulate. This is commonly between -5 and 5 volts. However, the reduction in the operational steps to yield a linear output according to the present invention allows a less sophisticated processor to be used than would otherwise be necessary if substantially more processing is required.

In practice, the physical angular displacement of the pedal 02 is usually less than 45°, more commonly 25°, typically 13° or less. The saw tooth function as shown in Figure 8d provides linearity over a range well beyond the angular range required for a typical pedal. By truncating the peaks and troughs of the saw tooth function to within an acceptable range for use with a pedal, precision can be improved. To perform this truncation, the saw tooth function may be bounded between an upper bound and lower bound angle determined by the range that the pedal can be displaced. This can be performed before or after the vertical translation of the saw tooth function, i.e. before a translated output 81 is made or after the translated output 81 is generated as shown in Figure 8d. Instead of taking the potential difference across sets of resistors in Wheatstone bridge circuits, potential differences can be taken at the junction between the resistor pairs 343a, 343b and a local system ground 378 in sensing element 319 as depicted in Figure 7b. Taking the potential differences, across Wheatstone bridge sensing elements as shown in

Figure 7a can introduce a loss in precision in the sinusoidal form output of each sensing element 119a and 119b.

In order to address and reduce this imprecision, the circuitry shown in Figure 7b samples the potential difference, between the junctions in each of the resistor pairs 343a

(left) and 343b (right) in comparison to the local system ground 378. Because the output is taken from each set of resistors, 343a and 343b, separately, the mean potential of each of the signals will approximately be a half of the input voltage source 80 when taking the resistance of the resistors in each pair 343a and 343b to be substantially the same. For example, a 5 volt voltage source 80 will result in a 2.5 volts mean output from each resistor pair. To maximise the resolution of the input signal to the PIC, it is preferable to amplify the output signals from each resistor pair 343a and 343b. The change in magnetic field causes the output signal from each resistor pair 343a and 343b to oscillate at an offset as depicted in Figure 10a. During amplification of the output signals, it is preferable to

remove this offset which has the effect of inverting and dropping the signals substantially close to zero as depicted in Figure 10b such that subsequent amplification of the output signals generates signals substantially covering the range of voltages that can

be inputted into the PIC 35 as depicted in Figure 10c.

In order to determine the precise offset that must be removed from the output signals, during manufacture the resistances of the sensing elements 343a and 343b are measured and the possible ranges for the potentials, when exposed to different orientations of the

magnet's magnetic field, are determined. A voltage divider circuit 384 comprised of trimmed carbon resistors 334 is then used to create a constant arbitrary voltage at the upper bound of the signal range. In comparison to the constant arbitrary voltage

being generated using a voltage divider circuit 384; alternatively, the constant arbitrary voltage can be generated by an auxiliary voltage source. The signal taken into a differential amplifier 332 with respect to the constant arbitrary voltage This has the effect of inverting the signal and offsetting them to a lower voltage before being amplified to give offset-compensated signal Vir, for input into the PIC 35. This process is performed for each output from sensing element individually using custom voltage divider circuits 334 with custom values of the custom arbitrary voltages 84. In the case of the example shown in Figure 10b, the output signal from each resistor pair 343a and 343b is amplified to produce an input signal to the PIC with a range substantially covering 0 to 5 volts. To amplify the output signals for input into the PIC, these signals are amplified separately by at least one amplification circuit 331 as illustrated in Figure 11. In the given example, the potentials Vir and Vi2' do not have a sinusoidal profile; however, a subtraction of one signal, Vir, from the other signal, V ₁₂ ', would generate a single signal of substantially sinusoidal form for input into the PIC chip 35 having a form similar to the sinusoid taken across the resistor pairs as discussed for use with a Wheatstone bridge circuit configuration of the sensing element 119. This mimics the potential difference between the resistor pairs as measured in the set-ups of the prior art and that attained when using sensing elements arranged in the example of Wheatstone bridge circuits according to the prior examples of the present invention discussed above but with greater precision or reduced noise on the subtraction. The result of removing this offset and amplification is that higher resolutions of signals are input into the PIC 35 when compared to signals derived using systems based upon earlier examples of the present invention that initially takes a potential difference between pairs of resistors. A subsequent advantage of this higher resolution is that even while using a lower bit-rate PIC 35 the linear output of the PIC 35 will be more precise. A second independent aspect of the present invention removes the requirement for there to be separate sensing elements for generating two output signals orthogonal to one another. Instead, two signals may be generated from one sensing element 119a, thereby removing the need for a second sensing element 119b to generate the second orthogonal signal. For example, a system based on a sensing element producing two signals as discussed above are subtracted within the PIC 35 and then centred around 0 Volts by a translation to produce a sinusoidal input signal 487, imilar to that attained in the

potential differentiator of the earlier examples. The PIC 35 is used to create a virtual (or "phantom") signal 486, The virtual signal 486 comprises a signal generated by the PIC 35 such that when performing a trigonometric operation of the ratio between the real signal 487 and the virtual signal 486 a linear output can be produced. As discussed above, preferably the trigonometric operation is an arctan function. This new approach allows the reduction of the GMR resistors from 16 to 8 resistors for a twin zoned sensing board 221 and from 8 to 4 resistors for a single zoned sensor 121 producing considerable cost savings through the reduction of the number of passive components required whilst enabling the effective sensing area of each sensing element to be increased thus increasing the accuracy of performance of each sensing element.

The process to produce the linear output is illustrated in the flow diagram of Figure 12. Step 96a determines the arguments (i.e. the angle) of the sinusoidal input signal with the use of an arcsine function. Step 96b adds π/2 to the result of step 96a, arcsin so as to

perform a horizontal translation. Step 96c performs a sine function on the result of Step 96b to generate the virtual signal 486 which is orthogonal to the input function by nature of the addition of π/2 in Step 96b as shown in Figure 10d. When considered together the steps 96a-c perform the operation as given in the equation below (equation 10). Equally in practice, cosine and arccosine functions could be used and would generate a virtual signal 486 as when using sine and arcsine functions; similarly a subtraction instead of addition of π/2 may also be implemented.

Step 97 performs an arctan operation on the ratio between 486 as below, generating a periodic triangular function 488, as shown in Figure 10d:

There exists higher uncertainty (i.e. noise) in the periodic triangular function at the maxima and minima of the function due to the shape of Since pedals typically operate

over a limited angular range, this permits the sensor to operate over a narrow range. Thus, the higher uncertainty associated with inflection areas of the signal, can be

avoided and stop it propagating into the desired linear output by truncating the periodic triangular function 488 at the maxima and minima, as provided for in step 98. For example, the periodic triangular function 88 may be truncated such that the maxima and minima are bounded between the limits given by:

A negative voltage cannot be output from a PIC chip 35 and thus in step 99, the bounded, periodic triangular function V _ou t (i) is translated to ensure that the function V _ou t (i), is positive. For example, this may be accomplished by vertically translating the periodic triangular function, V _ou t'(i) by π/2 as shown in Figure 10e.

The magnet 18 is situated within the pivot pin housing 17 such that when the pedal 02 is in the idle position, the poles of the magnetic field are aligned with the points of inflection 489 of the input signal V _rea i 487. i.e. such that in the idle position, the magneto-resistive resistors' resistance is equal to the absolute resistance of the track. A result of this is that the displacement of the magnet 18 is assumed to represent either the positive or the negative gradients of the bounded, translated periodic triangular function V _out (i)- In step 100, the translated output V _ou t' can be transformed such that the output of the sensor is mapped to an end-user preferred transfer function and thus relayed to the control platform so as to provide a measure of the initial displacement.

As such, the linear relationship between angular displacement and sensor output can be known with a high precision over a substantial angular displacement as shown by the saw- tooth function 490 or the truncated triangular function 490a. A signal representing displacement of the pedal 02 can then be returned to the vehicle control platform. The techniques of steps 98, 99 and 100 can be implemented to any of the examples described in this disclosure in order to improve the accuracy and precision of the linear relationship between the output signal from the sensor and the displacement of the pedal. For example, the linear relationship may be generated to a high precision over an angular displacement of approximately 60°. In practice, the physical angular displacement of the pedal 02 is usually limited to 45°, more usually 25°, often 13° maximum.

As discussed with the earlier examples, to comply with safety codes, a second sensing element must always be included to provide redundancy; however, for systems based on the second aspect of the present invention there is no need for an angular displacement between the two physical sensing elements. The requirement is removed by generating a virtual signal 486 derived from a real signal 487 associated with the first sensing element 119a. As a result of this, a greater number of possible resistor track arrangements become available. The sensing elementl l9a may comprise different combinations of patterns e.g. as described within the earlier examples such as double or greater zoning where two or more sensing element form concentric rings around the rotation axis X. The sensing board 21 resistor tracks 41 could utilise any combination of the earlier examples.

In Figures 6a, 6b , 6c and 6d, the sensing elements 119a and 119bare shown to be co-planar, i.e. lie within the same plane normal to the rotation axis X-X. However, it should be appreciated that the circuits may lie in different horizontal planes such that on rotation of the magnet 18, the change in magnetic field orientation is detectable by the sensing elements 119a and 119b resulting in a measureable change in resistance. It is also permissible within the scope of the present invention, to have any number of sensing elements arranged in a fashion according to any of the above examples so as to produce a series of linear outputs, each linear output having a precision over a given angular displacement, the accumulation of which produces a precision over a greater angular displacement. For example, a system may comprise a greater number of sensing elements each angularly offset from the other whereby precision measurement is achieved over 360° rotation of the magnet 18. Such measurements over a 360° rotation can be used in precisely determining the angular displacement of a steering wheel, e.g. in self-parking systems; or other angle sensing systems, rotation counters and even rotation speed measurement devices.

It should be understood that the scope of the present invention is not limited to the measurement of the displacement of a pedal system 01. The sensor can be used to determine or detect the angular or linear displacement of a wide range of systems where displacement measurement is required, e.g. the magnitude of trigger depression for a drill. Linear displacement of, for example, a drill trigger is transformed to an angular movement of a magnet 18 that can be rotatably housed within the sensor above the sensing element; sensing the angular displacement of the magnet 18 then provides a measure of the original linear displacement. It should be understood that the magnet 18 is not limited to being housed within a pivot pin instead it is rotatably mounted to any housing that allows rotation and thusly could be coupled to the initial linear displacement in a manner such that the initial linear displacement causes a rotation of the magnet 18 with a relationship that is directly proportional or indirectly proportional: for example, by means of meshed gears, a belt and pulleys, a crank and connecting rod, a Geneva mechanism, a flexible drive shaft, one or more cams, swashplates or any other suitable mechanical transmission which imparts rotational movement to the magnet 18.

Features of the present invention

1. A sensor comprising:

a plurality of sets of resistors, each set arranged to form a respective sensing element and comprising at least one magneto-resistive resistor;

wherein at least one of the sensing elements is spatially arranged to overlap or coincide with the area occupied by another of the sensing elements. 2. A sensor assembly comprising

a. a set of resistors including at least one magneto-resistive resistor and arranged to form a sensing element, wherein the sensing element produces a first alternating signal; b. a signal processor for generating a second alternating signal;

characterised in that;

the processor derives the second alternating signal from the first alternating signal such that the second alternating signal is orthogonal to the first alternating signal.

3. A sensor comprising

sets of resistors comprising a first pair of resistors in series and a second pair of resistors in series to form a sensing element; wherein the sensing element produces;

a) a first signal corresponding to the potential difference between the junction of the two resistors in the first pair and a first arbitrary system potential and

b) a second signal corresponding to the potential difference between the junction of the two resistors in the second pair and a second arbitrary system potential,

characterised in that:

the first and second signals is combined so as to produce an alternating signal. 4. A sensor of feature 3 wherein the first signal is combined with the second signal through a subtraction between the two signals.

5. The sensor of any of the features 2 to 4, comprising a plurality of said sensing elements.

6. The sensor as defined in feature 1 or 5, wherein at least one of the resistors in one of said plurality of sensing elements lies between the resistors in another of said sensing element.

7. The sensor as defined in feature 6, wherein the resistors of said one sensing element alternate with the resistors of the another sensing element.

8. The sensor as defined in any of the preceding features, wherein the sensor provides duplicate output signals so as to provide redundancy.

9. The sensor as defined in feature 6 or 8, wherein one of the resistors in one of said plurality of sensing elements is nested within another of said resistors in another of said sensing element.

10. The sensor as defined in any of the features 5 to 9, wherein the resistors in each of the plurality of the sensing elements are rotationally symmetrically arranged.

11. The sensor as defined in feature 10, wherein the plurality of the sensing elements are arranged substantially concentrically.

12. The sensor as defined in feature 11, wherein the resistors of each of the plurality of the sensing elements are co-planar or arranged in different horizontal planes.

13. The sensor as defined in feature 12, wherein at least one said set of resistors in one sensing element lies at an angle of substantially 45° with respect to another said set of resistors in another sensing element.

14. The sensor as defined in feature 12 or 13, wherein the said plurality of sensing elements is substantially disposed in a rotatable magnetic field. 15. The sensor as defined in any of the features 10 to 14, wherein the magnetic field is rotatable about the axis of symmetry. 16. The sensor as defined in any of the preceding features, wherein at least one of the resistors in at least one of the sets of resistors comprises tracks on a substrate.

17. The sensor as defined in feature 16, wherein the substrate has a width in the range of 0.5mm to 25mm; preferably 8mm to 15mm, more preferably 10mm to 12mm, most preferably about 11mm.

18. The sensor as defined in feature 16 or 17, wherein the substrate comprises an FR-4 board. 19. The sensor as defined in any of features 16 to 18, wherein the substrate comprises a smooth coating forming an underlay for the tracks.

20. The sensor as defined in features 19, wherein the coating has a surface roughness of 5μm or less; preferably Ι μm or less; more preferably 0.5μm or less.

21. The sensor as defined in feature 19 or 20, wherein the coating comprises a dielectric material; preferably a solder resist; more preferably PMMA.

22. The sensor as defined in any of the features 16 to 21, wherein the width of each of the tracks is in the range of 25μm to 500μm; preferably 25μm to 250μm.

23. The sensor as defined in feature 22, wherein the width of each of the tracks is substantially equal to the minimum gap between adjacent track portions. 24. The sensor as defined in features 16 to 23, wherein said tracks of said at least one resistor in at least one set of resistors comprises meanders.

25. The sensor as defined in feature 24, wherein the meanders comprises parallel portions. 26. The sensor as defined in feature 25, wherein the parallel portions are substantially arcuate. 27. The sensor as defined in any of the features 16 to 26, wherein the track forming at least one of the resistors doubles back on itself.

28. The sensor as defined in any of the features 16 to 27, wherein the track forming at least one of the resistors is mirror symmetrical.

29. The sensor as defined in any of the features 16 to 28, wherein the resistors are generally ' V shaped in outline.

30. The sensor as defined in any of the preceding features, comprising contact pads arranged peripherally of the sensing element.

31. The sensor as defined in any of the preceding features, wherein the at least one magneto- resistive resistor comprises a giant magneto-resistive resistor (GMR). 32. The sensor as defined in feature 31, in which the GMR comprises ferromagnetic and non- ferromagnetic layers having a thickness in the range of 0.5nm to 10nm; preferably 0.5nm to 1.5nm; more preferably 0.7nm - 1.2nm.

33. The sensor as defined in any of the preceding features, further comprising a potential differentiator; said potential differentiator is arranged to combine two signals from the same sensing element.

34. The sensor as defined in any of the preceding features, comprising a processor chip that is arranged to combine two signals from the same or different sensing elements.

35. The sensor as defined in feature 34 whereby the processor chip is configured to perform a trigonometrical operation upon the ratio between two alternating signals. 36. The sensor as defined in feature 35, wherein the processor is configured to calculate the arctan of the ratio of the two alternating signals.

37. The sensor as defined in any of features 34 to 36, wherein the at least one of the sensing elements is external to the processor chip.

38. The sensor as defined in any of the preceding features further comprising;

a. a differential amplifier having a first input signal corresponding to the first alternating signal from at least one sensing element and a second input signal corresponding to a constant arbitrary voltage signal; wherein the at least one differential amplifier is arranged to:

i) offset the first signal with respect to constant arbitrary voltage;

ii) amplify the offset signal to give an offset-compensated signal.

39. The sensor as defined in feature 38 further comprising;

a. another differential amplifier having a first input signal corresponding with another alternating signal from the same or another sensing element and a second input signal corresponding to the same or another constant arbitrary voltage signal; wherein the another differential amplifier is arranged to:

i) offset the first signal with respect to constant arbitrary voltage;

ii) amplify the offset signal to give a another offset-compensated signal.

40. The sensor as featured in feature 38 or 39, wherein the offset-compensated signal corresponding to each sensing element is normalized by dividing the offset-compensated signal value by a fixed multiple of the maximum value of the offset-compensated signal.

41. The sensor as defined in any of the features 38 to 40, further comprising one or more voltage dividers arranged to produce the constant arbitrary voltage(s) from a voltage source.

42. The sensor as defined in any of the preceding features, wherein the sensor is an angle sensor or a linear sensor. 43. The sensor as defined in feature 42, wherein the sensor comprises a cam mechanism for converting linear motion into a rotary motion.

44. The sensor of feature 43, wherein the cam mechanism comprises a cam surface and a cam follower engaging with the cam surface so as to convert linear motion of the cam follower into rotary or angular motion of the magnet.

45. A method of producing a sensor as defined in any of features 1 to 44, comprising the step of depositing a coating of magneto-resistive material on a substrate by printing, by vapour deposition or by sputtering.

46. The method of feature 45, comprising forming a pattern of magneto-resistive material by masking selective portions of the magneto-resistive material coating.

47. The method of feature 46, comprising forming a pattern of magneto-resistive material by removing selective portions of the magneto-resistive material coating.

48. The method of feature 47, wherein removing selective portions of the magneto-resistive material coating comprises laser cutting or water jet cutting or etching.

49. A system controller comprising a sensor as defined in any of the features 1 to 48.

50. An automotive controller, comprising a system controller as defined in feature 49.

51. The automotive controller as defined in feature 50, wherein the automotive controller is a brake or throttle control.

52. A method of sensing the degree of angular displacement of a rotating object using a sensor as defined in any of the features 1 to 51, comprising the steps of,

a. generating a first signal from at least one sensing element;

b. deriving a second signal from the first signal; c. performing a trigonometric operation on the ratio between the first and second signal so as to generate a signal that is directly proportional to the angular displacement of the rotating object. 53. The method of feature 52, wherein the trigonometric operation is an arctan function.

54. The method of feature 52 or 53, whereby the second signal is a translation of the first signal such that the second signal is substantially orthogonal to the first signal.