WO2008129290A1

WO2008129290A1 - Control circuits for an sma actuator

Info

Publication number: WO2008129290A1
Application number: PCT/GB2008/001400
Authority: WO
Inventors: Richard Topliss; Anthony Hooley; Irving Alexander Bienek
Original assignee: Cambridge Mechatronics Limited
Priority date: 2007-04-23
Filing date: 2008-04-22
Publication date: 2008-10-30

Abstract

Control circuits control a shape memory alloy actuation apparatus comprising an SMA actuator. In one control circuit, a bridge arrangement of resistors comprises a first resistor in series with the SMA actuator and, in parallel thereto, a second and third resistor in series with each other. A current source is controlled on the basis of the output of a differential amplifier having differential inputs connected respectively to the node between the SMA actuator and the first resistor and to the node between the second and third resistors. In another control circuit, a digital controller supplies a digital offset signal to a DAC. The controller varies the power of the drive current in response to the output of a comparison circuit whose first input is supplied with a voltage providing a measure of the resistance of the SMA actuator and whose second input is supplied with the output from the DAC.

Description

Control Circuits for an SMA Actuator

The present invention relates to the control of a shape memory alloy (SMA) actuation apparatus using SMA material as an actuator to drive movement of a movable element.

The present invention has particular application to actuation of a relatively small movable elements, for example a camera lens element, particularly a camera lens element of the type used in a miniature camera which may be employed in a portable electronic device such as a mobile telephone or a mobile digital data processing and/or transmitting device.

In recent years, with the explosive spread of portable information terminals sometimes known as PDAs (portable digital assistants) and portable telephones, an increasing number of devices incorporate a compact digital camera apparatus employing an image sensor. When such a digital camera apparatus is miniaturized using an image sensor with a relatively small image- sensing area, its optical system, including one or more lenses, also needs to be miniaturized accordingly.

To achieve focusing or zooming, an actuation arrangement of some type must be included in the confined volume of such a miniature camera to drive movement of the camera lens element along the optical axis. As the camera lens element is small, the actuation arrangement must be capable of providing precise actuation over a correspondingly small range of movement. At the same time it is desired that the actuator arrangement is itself compact given the desire for miniaturization of the camera apparatus as a whole. In practical terms, these points limit the types of actuation arrangement which can be applied.

Similar considerations apply to actuation arrangements for a wide range of other small objects.

Whilst most of the existing cameras rely on variations of the well-known electric-coil motor, a number of other actuation arrangements have been proposed as small drive units for the lens system. Such other actuation arrangements may include transducers based on piezoelectric, electrostrictive or magnetostrictive material, commonly referred to as electro- active devices and one example is an actuator comprising a curved structure of helically coiled piezoelectric bender tape as disclosed in WO-01/47041 which may be used as an actuator for a camera lens as described in WO-02/103451. Another type of actuation arrangement which has been proposed uses SMA material as an actuator. The SMA actuator is arranged on heating to drive movement of the camera lens element. Actuation may be achieved by control of the temperature of the SMA actuator over an active temperature range in which the SMA actuator changes between martensite and austenite phases in which the stress and strain of the SMA actuator changes. At low temperatures the SMA actuator is in the martensite phase, whereas at high temperatures the SMA actuator transforms into the austenite phase which induces a deformation causing the SMA actuator to contract. The temperature of the SMA actuator may be changed by selectively passing a current through the SMA actuator to heat it causing the phase change. The phase change occurs over a range of temperature due to the statistical spread of transition temperature in the SMA crystal structure. The SMA actuator is arranged so that the contraction drives movement of the movable element. The use of SMA material as an actuator for a small element such as the camera lens element of a miniature camera provides the advantages of being intrinsically linear, providing a high power per unit mass, being a low cost commodity item and being a relatively small component.

The present invention is concerned with the control circuit for such an actuator, and in particular with balancing the need to provide for accurate control of position with the need to minimize the cost of the circuit. The present invention is concerned with control circuits which use the resistance of the SMA actuator as a measure of the position. Such use of resistance has a considerable advantage of being accurate and being straightforward and compact to implement, simply by providing additional electronic components supplementing the elements needed to provide the drive current which heats the SMA actuator. In. contrast, direct measurement of the position of the movable element requires a position sensor which is bulky in the context of a miniature device. Also, measurement of the temperature of the SMA actuator is difficult to implement with sufficient accuracy.

When the electrical drive signal to an SMA actuator is based on a measure of the electrical resistance of the SMA element, one approach is to use a precision closed-loop control circuit. This is typically quite complex and involves (1) measuring a measure of the resistance of the SMA actuator, for example the voltage and/or current across the SMA actuator, (2) comparing the measure of resistance to a target value (the closed-loop control input "demand" value), and (3) on the basis of that comparison, modifying the drive signal to the SMA actuator so as to reduce the difference between the measured and demanded resistance.

According to a first aspect of the present invention, there is provided a control circuit for a shape memory alloy actuation apparatus comprising an SMA actuator arranged on heating to drive movement of a movable element, the control circuit comprising: a current source operable to pass current through the SMA actuator to heat the SMA actuator; a bridge arrangement of resistors comprising a first resistor in series with the SMA actuator and a second and third resistor in series with each other and together in parallel with the first resistor and the SMA actuator together; a differential amplifier having differential inputs connected respectively to the node between the SMA actuator and the first resistor and to the node between the second and third resistors, the current source being controlled on the basis of the output of the differential amplifier.

Such a control circuit may be much simplified as compared to a closed-loop control circuit, thereby providing the advantage of cheapness and ease of implementation. In particular this is achieved because the "input demand" to the closed loop control circuit is in the form of another physical electrical resistance, i.e. a physical resistor component. This allows the comparison to be performed by a differential amplifier which controls the current source, for example being directly connected to the current source.

The second aspect of the present invention is concerned with control circuits employing a digital controller to control a current source to vary the power of the drive current supplied thereby through the SMA actuator. In this case, a measure of the resistance of the SMA actuator is fed back to the controller, for use as a measure of position. The digital controller varies the power of the drive current based thereon.

The measure of resistance depends on the nature of the control but is typically a voltage at a point in the control circuit. For example in the case of constant-current drive, the measure of resistance may be the voltage across the SMA actuator, or in the case of constant- voltage drive, the measure of resistance may be a voltage across a resistor in series with the SMA actuator. Similarly, a measure of resistance may be derived from the voltage across other components in the control circuit.

The signal supplied to the digital controller needs to be converted into a digital signal, for example using an ADC. One issue arising is the dynamic range and resolution of the ADC which affects the sensitivity of control. If the voltage providing a measure of resistance has a high fixed component, the dynamic range of the ADC can be wasted leading to the effective resolution being reduced. Sensitivity can be increased by increasing the resolution of the ADC but at the expense of increased complication and cost for the ADC. To reduce this problem, the present inventors have considered using a bridge arrangement consisting of a balancing arm containing a pair of resistors in parallel with an arm containing the SMA actuator and a ballast resistor. In this case, the difference in voltages between the arm containing the SMA actuator and the balancing arm can be derived using a differential amplifier. By selecting the values of the resistors appropriately, the fixed component of the voltage can effectively be removed by the bridge arrangement, which in turn allows the dynamic range of the ADC to be better utilised and the resolution effectively increased. However, such a bridge arrangement has a number of practical problems. Firstly the component count is increased. Secondly the power consumption is increased because of the current flow through the balancing arm. Thirdly, the values of the resistors are in practice difficult to select due to variation in the SMA resistance due to manufacturing tolerances in the mechanical, material, termination and physical properties and also due to short and long term aging. Such selection can be achieved using trirnmable resistors, but that is more expensive and increases the complexity of manufacture.

According to a second aspect of the present invention, there is provided a control circuit for an SMA actuator capable on heating of driving movement of a movable element, the control circuit comprising: a current source arranged to supply a drive current of variable power through the SMA actuator; a digital controller arranged to control the current source to vary the power of the drive current supplied thereby; a DAC, the digital controller supplying a digital offset signal to the DAC and the DAC being operative to convert the digital offset signal into an analog offset voltage signal; and a comparison circuit having two inputs and an output, the first input being supplied with a voltage providing a measure of the resistance of the SMA actuator, the second input being supplied with the digital offset voltage from the DAC, the comparison circuit being arranged to compare the voltages at the two inputs and to produce an output signal at the output representing a comparison between the voltages at the two inputs, the controller being supplied with the output signal from the output of the comparison circuit and being arranged to vary the power of the drive current supplied by the current source in response thereto. In this case, the digital controller supplies a digital offset signal to a DAC converts this into an analog offset voltage signal. A comparison circuit compares the analog voltage signal with a voltage providing a measure of the resistance of the SMA actuator, the controller varying the power of the drive current in response to the output of the comparison circuit. In this case the analog offset voltage can be used to remove the fixed component of the voltage providing a measure of resistance. This in turn allows the dynamic range of the ADC to be better utilised and the resolution effectively increased in a similar manner to a bridge arrangement. However, as the digital offset signal is derived by the controller, it is straightforward for the controller to adapt the offset and the problems discussed above associated with a bridge arrangement of physical resistors are avoided. Ih one type of embodiment, the comparison circuit is arranged to produce a digital output signal at the output representing the difference between the voltages at the two inputs, for example comprising a differential amplifier and an ADC connected to the output of the operational amplifier.

In another type of circuit, the comparison circuit is a comparator arranged to produce a binary output signal at the output representing which of the voltages at the two inputs is higher. In this case, the controller is supplied with the output signal from the output of the comparator and is arranged to increase or decrease the power of the drive current supplied by the current source in dependence on which of the voltages at the two inputs of the comparator is higher. This type of circuit has the advantage of avoiding the need for an ADC converter.

To allow better understanding, an embodiment of the present invention will now be described by way of non-limitative example with reference to the accompanying drawings, in which:

Fig. 1 is a schematic cross-sectional view of a camera incorporating an SMA actuation arrangement;

Fig. 2 is a detailed perspective view of the camera; Fig. 3 is an exploded perspective view of part of the camera;

Fig. 4 is a perspective view of an SMA actuator of the camera;

Fig. 5 is a detailed cross-sectional view of the camera;

Fig. 6 is a schematic diagram of the overall control arrangement of the camera;

Fig. 7 is a diagram of the control circuit; Fig. 8 is a graph of the resistance-length property of SMA during contraction;

Fig. 9 is a perspective view of a modified form of mounting member of the SMA actuator;

Fig. 10 is a diagram of a modified form of the control circuit;

Figs. 11 to 28 are diagrams of possible circuit implementations for the control circuit; Fig. 29 is a flow chart of control algorithm implemented in the control circuit;

Fig. 30 is a graph of the resistance of the SMA over time with the control algorithm of Fig. 29.

There will first be described the structure of a camera 1 incorporating an SMA actuation apparatus. The camera 1 is to be incorporated in a portable electronic device such as a mobile telephone, media player or portable digital assistant.

The camera 1 is shown schematically in Fig. 1. The camera 1 comprises a support structure 2 which has a base portion 3 on which there is mounted an image sensor 4 which may be CCD (charge-coupled device) or a CMOS (complimentary metal-oxide-semiconductor) device. The support structure 2 further comprises an annular wall 5 protruding from the front side of the base 3 on which the image sensor 4 is mounted. The support structure 2 may be made of plastic.

The camera 1 further comprises a lens element 6 which holds a lens system 7 consisting of one or more lenses 8. By way of example, the lens system 7 is shown in Fig. 1 as consisting of two lenses 8 but in general there may be a single lens 8 or plural lenses 8 as needed to provide the desired balance of optical performance and low cost. The camera 1 is a miniature camera with the lenses 8 of the lens system 7 typically having a diameter of at most 10mm.

The lens element 6 is arranged with the optical axis O of the lens system 7 perpendicular to the image sensor 4. In this manner, the lens system 7 focuses light onto the image sensor 4. The lens element 6 is suspended on the support structure 2 by a suspension system 9 consisting of two suspension elements 10 connected between the annular wall 5 of the support structure 2 and the lens element 6. The suspension system 9 guides movement of the lens element 6 along the optical axis O. Such movement of the lens element 6 changes the focus of the image formed on the image sensor 4. The detailed construction of the camera 1 will now be described with reference to Fig.

2 which is a detailed perspective view omitting the base 3 of the support structure 2.

The lens element 6 has a two-part construction comprising a lens carrier 20 and a lens holder 21 mounted inside the lens carrier 20 on an internal screw thread 22 formed inside the lens carrier 20. Typically the lens holder 21 has a diameter of 6.5mm. Fixed to the lower rim of the lens carrier 20 is a metal ring 14 described further below. The lens carrier 20 is connected to the suspension system 9 to suspend the lens element 6. The lens holder 21 mounts the one or more lenses 8 of the lens system 7. Both the lens carrier 20 and the lens holder 21 may be made from moulded plastic .

The suspension system 9 for the lens element 6 will now be described in detail. The suspension system 9 comprises two suspension elements 10 each formed from a respective single sheet of material such as steel or beryllium copper cut into shape. One possibility is hard rolled grade 302 austenetic steel which has the advantage of providing a high yield stress. The suspension elements 10 are mounted at opposite ends of the carrier 20. Whilst only one of the suspension elements 10 is clearly visible in Fig. 2, both suspension elements 10 have an identical construction, as follows.

Each suspension element 10 comprises an inner ring 11 connected to the lens carrier 20. In particular, the inner ring 11 is connected to a respective end surface of the lens carrier 20 so that it extends around the outer circumference of the lens holder 21.

Each suspension element 10 further comprises an outer ring 12 connected to the support structure 2. In particular, the outer ring 12 extends around and is connected to the end surface of the annular wall 5 of the support structure 2.

Lastly, each suspension element 10 comprises four flexures 13 which each extend between the inner ring 11 and the outer ring 12. Thus the flexures 13 are coupled at opposite ends to the lens element 6 and the support structure 2. As viewed along the optical axis O, the flexures 13 are inclined relative to the direction radial of the optical axis O. Thus the flexures 13 extend around the optical axis. The flexures 13 are disposed around the lens carrier 20 at different radial positions with rotational symmetry around the optical axis O. Furthermore, the flexures 13 have a thickness along the optical axis O (that is the thickness of the sheet of material from which the suspension element 10 is made) which is smaller than their width in a direction perpendicular to the optical axis O. The suspension system 9 is designed with an appropriate number of flexures 13 of appropriate width, thickness and length to provide the desired degree of stiffness along the optical axis O and perpendicular thereto. The flexures 13 typically have a thickness in the range from 25μm to lOOμm. The number of flexures 13 may be changed by varying the number of flexures 13 within a suspension element 10 and/or by providing additional suspension elements 10. The flexures 13 are also curved along their length as viewed along the optical axis O with three regions of alternating curvature. By introducing such curvature to the flexures 13, a degree of strain relief is added to the structure. The tendency of the flexures 13 to plastically deform is reduced and instead the flexures 13 have a tendency to bend elastically. By introducing the outer regions having opposite curvature to the central region, the force imbalance is reduced and the stress developed at the joints with the inner ring 11 and outer ring 12 are reduced. Thus the flexures 13 become more compliant in the planar direction without experiencing material failure. This is achieved without an unacceptable compromise to the radial and axial stiffnesses. This allows the suspension system 9 to accommodate the displacement of the lens element 6 radially of the optical axis O caused by mechanical impacts without causing permanent damage to the flexures 13. To limit the displacement in this direction, the camera 1 is provided with a small clearance, for example of the order of 50μm or less, between the lens element 6 and the wall 5 of the support structure 2 so that the wall 5 of the support structure 2 acts as a stop to limit the maximum displacement.

To maximize this effect the three regions of the flexures 13 preferably have unequal lengths and curvature, in particular with the central region having a greater length and a lesser curvature than the outer regions. Advantageously, the central region has a length which is at least twice the length of the outer regions, for example with the ratio of the lengths of the three regions 1:2.5:1. Advantageously, the central region has a curvature which is at most half the curvature of the outer regions, for example with the ratio of length to curvature of each region being substantially the same so that the angles subtended by each region are substantially the same.

Optionally each flexure 13 could be modified to consist of a group of parallel flexures to allow the suspension system 9 to be made more compliant radially of the optical axis by reducing the width of each parallel flexure. The practical limitation to this technique is the minimum width to which the parallel flexures may be manufactured.

The two suspension elements 10 suspend the lens element 6 on the support structure 2 by means of the flexures 13 being coupled between the lens element 6 and the support structure

2. Due to their configuration, the flexures 13 accommodate movement of the lens element 6 along the optical axis O by flexing or bending. When the lens element 6 moves along the optical axis O, the inner rings 11 move along the optical axis O relative to the outer rings 12 with consequent bending of the flexures 13.

As the flexures 13 have a thickness parallel to the optical axis O which is smaller than their width, the flexures 13 are more compliant to bending in their thickness direction than to bending in their width direction. Accordingly, the flexures 13 provide the suspension system 9 with a lower degree of stiffness against movement of the lens element 6 relative to the support structure 2 along the optical axis O, than against movement of the lens element 6 relative to the support structure 2 perpendicular to the optical axis O.

Furthermore, the two suspension elements 10 are spaced apart along the optical axis O and thus the resistence to movement of the lens element 6 perpendicular to the optical axis O also provides resistence to tilting of the lens element 6. Such resistence to off-axis movement and tilting of the lens element 6 is desirable because such off-axis movement and tilting can degrade the optical performance of the lens system 7 in focussing an image on the image sensor 4.

The support structure 2, lens carrier 20 (including the metal ring 14), the suspension elements 10 and two stiffener elements 15 are manufactured as a subassembly as will now be described with reference to Fig. 3. These components are arranged in a stack as shown in Fig.

3. Location pins 16 formed on the support structure 2 and the lens carrier 20 locate in apertures 17 formed in the suspension elements 10. While the complete stack is compressed in a jig, adhesive is dispensed onto the ends of each of the location pins 16, both on the top and bottom of the stack. The preferred adhesive is a cyanoacrylate that is also UV curable. By capillary action the adhesive soaks around the location pins 16, and bonds the different layers to the support structure 2 and the lens carrier 20. Once the adhesive has cured, the subassembly can be removed from the jig. As an alternative to adhesive, it is possible to form the joints be heat staking the location pins 16 to form a plastic head that retains the parts mechanically.

Each stiffener 15 comprises two rings 18 which respectively conform to, and stiffen, the inner ring 11 and the outer ring 12 of a suspension element. The two rings 18 are joined together by sprues 19 which are removed only after the subassembly has been assembled. The use of the sprues 19 helps assembly in terms of jigging the rings 18 of the stiffeners 15, and reduces the component count, and hence part cost. Once the sprues 19 are removed, the lens carrier 20 can be moved upwardly relative to the support structure 2 by an external load. In addition, the camera 1 comprises an SMA actuator 30 which is illustrated in isolation in Fig. 4. The SMA actuator 30 comprises a piece of SMA wire 31 mechanically and electrically connected at each end to a respective mounting member 32, each formed as an elongate piece of metal for example brass. In particular the mounting members 32 are each crimped over the piece of SMA wire 31. To ensure proper electrical connection, during manufacture of the SMA actuator 30 the oxide coating which forms naturally over the SMA wire 31 is removed before crimping.

During manufacture, the SMA actuator 30 is made as a subassembly separately from the remainder of the camera 1. In particular, the SMA actuator 30 is manufactured by holding the mounting members 32 in place, applying the piece of SMA wire 31 taut over the mounting members 32 and then crimping the mounting members 32 over the piece of SMA wire 31. The SMA actuator 30 is then assembled into the camera 1 in the arrangement as follows.

The two mounting members 32 are each mounted onto the outside of the annular wall 5 of the support structure 2 and are fixed in place so that to connect the piece of SMA wire 31 to the support structure 2. As shown in Fig. 2, the mounting members 32 are mounted in recesses 40 provided in the annular wall 5, for example by adhesive, swaging of the wall 5 or some other means.

Furthermore, the piece of SMA wire 31 is hooked over a retaining element 41 which is an integral part of the metal ring 14 fixed to the lens element 6 and protrudes outwardly of the lens element 6. The surface of the retaining element 41 in contact with the piece of SMA wire 31 may be curved to reduce the maximum curvature of the SMA wire. Ia the camera 1, the mounting members 32 are positioned on diametrically opposite points around the optical axis O. The retaining element 41 is positioned mid-way between the two mounting members 32 around the optical axis O. As viewed along the optical axis, the lengths 42 of SMA wire 31 extend at 90° to each other along sides of the camera 1. After the assembly and in equilibrium, the piece of SMA wire 31 can be held in place with a small amount of adhesive, to ensure retention on the retaining elements 41 during operation or drop testing. This may be done after cycling of the SMA wire to help eliminate assembly tolerances.

The retaining element 41 is arranged at a position along the optical axis O which is closer to the image sensor 4 than the portion of the mounting members 32 to which the piece of SMA wire 31 is crimped. As a result, the two lengths 42 of SMA wire 31 formed by half of the piece of SMA wire 31 on either side of the retaining element 41 are held at an acute angle to the optical axis O. Slippage over the retaining element 41 during assembly assists in achieving an equal lengths and tensions for the two lengths 42 of SMA wire 31.

The lengths 42 of SMA wire 31 are held in tension in the camera 1 so that they apply a tensional force having a component along the optical axis O, in particular in a direction biassing the lens element 6 away from the image sensor 4. Thus in the absence of heating of the lengths 42 of SMA wire 31, the lens element 6 is in its closest position to the image sensor 4 within its range of movement. The camera 1 is designed so that this position corresponds to far- field or hyperfoeal focus, which is the most common setting for the camera 1, particularly if an auto-focus function is provided.

In addition, each individual length 42 of SMA wire 31 applies a tensional force having a component perpendicular to the optical axis O. Some of the components of these forces are balanced by the symmetrical arrangement of the two lengths 42 of wire but there remains a net component of force radially of the optical axis O at the retaining element 41, this tending to tilt the lens element 6. However, the tilt is resisted by the suspension system 9 to be sufficiently small to be adequate for many lenses and image sensors. The operation of the camera 1 to drive movement of the lens element 6 along the optical axis O relative to the support structure 2 will now be described.

SMA material has the property that on heating it undergoes a solid-state phase change which causes the SMA material to contract. At low temperatures the SMA material enters the Martensite phase. At high temperatures the SMA enters the Austenite phase which induces a deformation causing the SMA material to contract. The phase change occurs over a range of temperature due to the statistical spread of transition temperature in the SMA crystal structure. Thus heating of the lengths 42 of SMA wire 31 causes them to decrease in length.

In the camera 1 , the lengths 42 of SMA wire 31 are arranged in tension providing a net tensional force along the optical axis O between the lens element 6 and the support structure 2 in a direction moving the lens element 6 away from the image sensor 4. This force acts against the biassing force provided by the suspension system 9 in the opposite direction along the optical axis O. The flexures 13 flex depending on the tensional force from the lengths 42 of SMA wire 31. The flexures 13 as manufactured are straight as viewed radially of the optical axis O. On flexing the flexures 13 remain generally straight although some slight curvature may be generated. Thus the flexing of the flexures 13 provides biassing of the camera lens element 6 in an opposite direction from the tensional force applied by the lengths 42 of SMA wire 31. In other words the suspension system 9 provides the function of acting as a passive biassing arrangement for the lengths 42 of SMA wire 31 as well as the function of suspending and guiding movement of the camera lens element 6. In the unhealed state of the SMA wire 31 in which it is not contracted, the SMA wire

31 is in tension, thereby displacing the lens element 6 away from its rest position in the absence of the SMA wire 31. In this state, the lens element 6 is in its closest position to the image sensor 4 within its range of movement. The camera 1 is designed so that this position corresponds to far-field or hyperfoeal focus, which is the most common setting for the camera 1 , particularly if an auto-focus function is provided.

On heating of the lengths 42 of SMA wire 31 so that the stress therein increases, the lengths 42 of SMA wire 31 contract moving the lens element 6 away from the image sensor 4. The lens element 6 moves over a range of movement as the temperature of the SMA wire 31 increases over the range of temperature in which the transition of the material of the SMA wire from the Martensite phase to the Austenite Conversely, on cooling of the lengths 42 of SMA wire 31 so that the stress therein decreases, the biassing provided by the flexures 13 causes the lengths 42 of SMA wire 31 to expand moving the lens element 6 towards the image sensor 4.

To maximise the movement of the lens element 6 relative to the support structure 2 along the optical axis O, the total stiffness of the flexures 13 of the suspension system 9 is preferably in the range from (a) the total stiffness of the lengths 42 of SMA wire 31 experienced in the austenite phase of the SMA material to (b) the total stiffness of the lengths 42 of SMA wire 31 experienced in the martensite phase of the SMA material, more preferably the geometric mean of values (a) and (b).

It is desired that the total stiffness against movement of the lens element 6 relative to the support structure 2 along the optical axis O, provided by the sum of the stiffnesses of the flexures 13 and the lengths 42 of SMA wire 31, is sufficiently great to minimize the movement of the lens element 6 relative to the support structure 2 under gravity when the camera 1 changes between orientations. For typical lens systems the movement is desirably limited to at most 50μm which for a typical miniature camera this means that the overall stiffness should be at least 1 OON/m, preferably at least 120N/m.

The flexures 13 are designed with an appropriate width to provide the desired stiffness against movement of the lens element 6 relative to the support structure 2 in directions perpendicular to the optical axis O, based on the extent to which the lens element 7 can accommodate off-axis motion and tilting. The stiffness of the lengths 42 of SMA wire 31 is also taken into account but usually provides a smaller contribution.

Another design consideration is to ensure that the maximum stress experienced by the flexures 13 and the lengths 42 of SMA wire 31 do not over-stress the respective materials. The degree of displacement of the lens element 6 relative to the support structure 2 along the optical axis O is dependent on the stress developed within the lengths 42 of SMA wire 31 and also on the acute angle of the lengths 42 of SMA wire 31 with respect to the optical axis O. The strain which may be developed in an SMA wire is limited by the physical phenomenon of the phase change. Due to the acute angles of the lengths 42 of SMA wire 31 with respect to the optical axis O, the lengths 42 of the SMA wire change in orientation when they change in length. This effectively gears the movement so that the degree of displacement of the lens element 6 along the optical axis O is higher than the change in length of the lengths 42 of SMA resolved along the optical axis O. In general the acute angle may take any value but is approximately 70° in the example of Fig. 2.

The position of the lens element 6 relative to the support structure 2 along the optical axis O may be controlled by control of the temperature of the lengths 42 of SMA wire 31. In operation, heating of the lengths 42 of SMA wire 31 is provided by passing a current therethrough which provides resistive heating. Cooling is provided by ceasing the current and allowing the lengths 42 of SMA wire 31 to cool by conduction to their surroundings. The current is controlled by a control circuit 50 which is described further below.

The SMA wire 31 may be made of any suitable SMA material, for example Nitinol or another Titanium-alloy SMA material. Advantageously, the material composition and pre- treatment of the piece of SMA wire 31 is chosen so that the phase change occurs over a range of temperature which is (a) above the expected ambient temperature during normal operation, typically above 70⁰C and (b) as wide as possible to maximise the degree of positional control.

High speed actuation of the lens element 6 is desired in many applications, for example if an auto-focus function is provided. The speed of response of the actuation is limited by the cooling of the lengths 42 of SMA wire 31. The cooling may be speeded up by reducing the thickness of the lengths 42 of SMA wire 31. For the size of cameras and wires under consideration, the cooling time changes approximately linearly with wire diameter. For this reason, the thickness of the lengths 42 of SMA wire 31 is desirably at most 35μm to provide a response which is acceptable for an auto-focus application of the camera 1. Fig. 5 shows in detail the camera 1 except omitting the lens holder 21 for clarity. The additional components of the camera 1 beyond those shown in Fig. 2 will now be described.

The camera 1 has a screening can 44 clipped and bonded over the wall 5 of the support structure 2. The wall 5 is also bonded to the base 3 of the support structure 2. In the direction along the optical axis O, there are clearances between the lens element 6 and the screening can 44 and between the lens element 6 and the base 3 which allow sufficient movement of the lens element 6 along the optical axis O to provide for focussing of the image on the image sensor 4 whilst preventing a degree of movement which would damage the suspension system 9 or the lengths 42 of SMA wire 31. Thus the screening can 44 and the base 3 effectively form endstops for the movement of the lens element 6 along the optical axis. In fact the base 3 has a more complicated construction than is shown schematically in

Fig. 1. In particular the base 3 has a central aperture 45 behind which the image sensor 4 is mounted. For mounting of the image sensor 4, the base 3 has a ledge 45 formed to the rear of the aperture 45 and outside the area of the aperture 45. On the ledge 46 is mounted an image circuit board 47 on which the image sensor 4 is formed facing and aligned with the aperture 45 to receive light along the optical axis O. Optionally the aperture 45 may have an infra-red filter fitted thereacross to improve the image quality, but also as a seal to prevent dust landing on the image sensor 4.

The base 3 further includes a protruding wall 48 disposed outside the ledge 46 and protruding rearwardly. A drive circuit board 49 is mounted on the protruding wall 48 and the drive circuit 50 is formed on that drive circuit board. As an alternative, it is possible to use an image circuit board 47 which is double-sided, with the drive circuit 50 mounted on its underside. Another alternative is to integrate the control circuit 50 into the same chip as the image sensor 4. Alternatively the same processing function could be carried out by another processor in the electronic device outside the camera 1, but already present for other purposes. A camera of identical construction to the camera 1 is described in WO-2007/113478 (co-owned International Patent Application No. PCT/GB07/001050) containing additional disclosure about the construction and manufacture which may be applied to the camera 1. Accordingly, WO-2007/113478 is incorporated herein by reference.

The nature of the control circuit 50 and the control effected thereby will now be described. A schematic view of the overall control arrangement is shown in Fig. 6. The control circuit 50 is connected to the piece of SMA wire 31 and applies a current thereto to control the temperature of the piece of SMA wire 31 which moves the lens element 6 and changes the focus of the image formed on the image sensor 4. The output of the image sensor 4 is supplied to the control circuit 50 to be processed for determination of a measure of the quality of focus. The control circuit 50 is shown in Fig. 7. The control circuit 50 is connected to each of the mounting members 32 which provide electrical connection to the piece of SMA wire 31 by means of the crimping thereof. The electrical connections 55 between the control circuit 50 and the mounting members 32 are formed by a conductive adhesive (e.g. silver-filled epoxy). It is undesirable to solder the control circuit 50 to the SMA actuator 30 because of potential damage caused by heating during the soldering process, or because of flux emissions caused by the soldering process.

The control circuit 50 supplies a current between the two mounting members 32. The control circuit 50 controls the degree of heating of the piece of SMA wire 31 by varying the power of the current flowing therethrough. The control circuit 50 varies the power of the current in response to the resistance of the piece of SMA wire 31 which is used as a measure of the position of the lens element 6. Other measures of position such as the temperature measured by a temperature sensor or a direct measure of the position of the lens element 6 output by a position sensor could be used, but a resistance sensor is advantageous because it does not increase the package size of the camera due to being implemented merely by additional components in the control circuit 50.

The physical phenomena behind the use of resistance are as follows. On heating of the SMA, there is an increase of resistivity with temperature as for most materials. This occurs inside and outside the range of temperature over which the phase-change occurs (the phase transition range) and hence over which the SMA contracts. However inside the phase transition range two further effects occur. Firstly, the Austenite phase has a higher resistivity than the Martensite phase which tends to increase resistance with temperature. However, an opposing effect is that the change of geometry, involving a reduced length and increased cross-sectional area, tends to reduce resistance with temperature. This opposing effect is significantly greater than the other effects. Thus, during heating from low temperature, when the phase transition range is reached and the SMA starts to contract, after an initial rise of resistance the geometrical effect rapidly dominates with the result that during the major part of the contraction the resistance of the SMA actuator decreases. This occurs until the phase change has occurred in nearly all of the SMA so that the degree of contraction falls allowing the resistance to rise.

Thus, SMA has a property that resistance varies with length during heating and contraction along a curve of the form shown in Fig. 8 which is a graph of resistance of the SMA against length of the SMA, corresponding to the position x of the lens element 6, the length increasing as the SMA contracts corresponding to increasing temperature. Thus across the phase transition range, the lens element 6 moves across a positional range Ax; due to the contraction of the SMA. The resistance rises across a small initial part of the positional range Δx to a local maximum 60 having a resistance value Rmax. The resistance falls across the major part of the positional range Δx to a local minimum 61 having a resistance value Rmin, whereafter the resistance rises across a small final part of the positional range Δx.

Due to this property of SMA material, the control circuit 50 implements control based on the measured resistance as follows. From an unheated state, the control circuit 50 heats the piece of SMA wire 31 until the local maximum resistance value is detected. This is used as an indication that contraction has started to occur. In fact a small amount of contraction has already occurred. However the local resistance maximum 60 can be easily detected, whereas the start of the positional range Δx cannot. Accordingly, the local resistance maximum 60 is used and this is so close to the start of the positional range Δx that the loss of movement is not significant.

Thereafter the control circuit 50 heats the piece of SMA wire 31 using the measured resistance as a measure of position. The local minimum resistance 61 is used to indicate the end of the positional range Δx. In fact, a small amount of contraction is still available. However the local minimum resistance 61 can be easily detected, whereas the end of the positional range Δx cannot. Accordingly, the local minimum resistance 61 is used. This is so close to the end of the of the positional range Δx that the loss of movement is not significant. Furthermore use of the positional range Ax above the local minimum resistance 61 can reduce the lifetime of the piece of SMA wire 31 as described further below.

The control circuit 50 may use pulse-width modulation (PWM). In this case, the control circuit 50 applies a pulse-width modulated current pulses (which may be of constant current or constant voltage) and varies the duty cycle in order to vary the power of the current applied and hence the heating. Use of PWM provides the advantage that the amount of power supplied may be accurately controlled with a fine resolution. This method provides a high signal-to-noise ratio, even at low drive power. The PWM may be implemented using known PWM techniques. Typically, the control circuit 50 will continually supply a pulse of current, for example with a duty cycle varying in the range from 5% to 95%. When the duty cycle is at a low value within this range, the average power in the piece of SMA wire 31 is low and so the wire cools even though some current is being supplied. Conversely, when the duty cycle is at a high value in the range, the piece of SMA wire 31 heats. The resistance is measured during the current pulse, for example after a short, predetermined delay from the start of the pulse. During heating of the piece of SMA wire 31 from a cool state below the phase transition range, the resistance varies with position in the manner shown in Fig. 8 in a manner which is consistent from sample to sample and in successive heating cycles. However, during cooling the resistance changes along a curve of similar form but the variation of resistance is less repeatable from sample to sample and there is variable hysteresis as compared to the heating. This does not prevent the use of resistance as a measure of position during cooling altogether, but does reduce the accuracy of the control. This problem is avoided by the control circuit 50 following a predetermined and repeated motion in which positional control is only effected during heating of the sample as described below.

The control circuit 50 includes the following components. The control circuit 50 includes a drive circuit 53 which is connected to supply current to the piece of SMA wire 31. The drive circuit 53 may be a constant-voltage current source or a constant-current current source. For example, in the latter case the constant current might be of the order of 120mA.

The control circuit 50 further includes a detection circuit 54 arranged to detect a measure of the resistance of the SMA actuator 30. Various measures of resistance are possible. In the case that the drive circuit 53 is a constant-current current source, the detection circuit 54 may be a voltage detection circuit operable to detect the voltage across the SMA actuator 30 which is a measure of the resistance of the piece of SMA wire 31. In the case that the drive circuit 53 is a constant-voltage current source, the detection circuit 54 may be a current detection circuit operable to detect the current through the SMA actuator 30 which is a measure of the resistance of the piece of SMA wire 31. For a higher degree of accuracy the detection circuit 54 may comprise a voltage detection circuit and a current detection circuit operable to detect the voltage and current across the SMA actuator and to derive a measure of resistance as the ratio thereof.

The controller 52 is a digital device, for example implemented by a suitable microprocessor, and controls the drive circuit 53 to supply the drive current. The controller 52 receives the measure of resistance measured by the detection circuit 54 and varies the power of the drive current in response thereto.

Typically, the control circuit 50 varies the power of the current using the measured resistance of the SMA actuator as a feedback signal to drive the measured resistance to a target value. In the case of PWM control, the duty cycle of the PWM current is varied. The controller 52 may implement a number of control algorithms to vary the duty cycle. One possibility is proportional control in which the duty cycle is varied by an amount proportional to the difference between the detected resistance and the target resistance. As the piece of SMA wire 31 heats across the active temperature region, the decrease in resistance is sensed and used in a feedback control technique. The stability of the feedback control is maintained by the inherent proportional-integral action of the piece of SMA wire 31 itself during heating. The overall feedback response is dominated by the response of the whole of the heating of the piece of SMA wire 31. Such a proportional control feedback loop provides for accurate control of position. The piece of SMA wire 31 may have some non-linearities in its response. Such non- linearities may be limited by incorporating precompensation in the control circuit 50. One option is for the precompensation to consist of a gain or offset modifier on the output signal supplied to the drive circuit 53, for example based on the demand and the history of the demand signal. This is most beneficial if insufficient feedback is present to control the piece of SMA wire 31.

A difficulty is created by the resistance of the electrical connections 55, particularly as they are formed by conductive adhesive which has variable and quite large electrical resistance compared to solder, as well as having a significant temperature coefficient. The detection circuit 54 actually measures the total resistance of the SMA actuator 30 and the electrical connections 55. Thus the variable and temperature dependent resistance of the electrical connections 55 causes significant problems of precision when attempting to provide accurate positional control.

This problem is overcome by modifying each mounting member 32 as shown in Fig. 9 and by modifying the control circuit as shown in Fig. 10, as will now be described. Each mounting member 32 is provided with two separate terminals 33 arranged adjacent one another and protruding from the remainder of the mounting member 32 with a gap therebetween. Separate electrical connections 56 and 51 axe, made to each terminal, formed as before by conductive adhesive. The first electrical connection 56 of each mounting member 32 is connected to the drive circuit 53. The second electrical connection 56 of each mounting member 32 is connected to the detection circuit 54. Furthermore the detection circuit 54 is a voltage detection circuit operable to detect the voltage across the SMA actuator 30. This is used to provide a measure of the resistance of the SMA actuator 30. Preferably the drive circuit 53 is a constant-current current source so that the voltage across the SMA actuator 30 is a direct measure of resistance although alternatively a further detection circuit could be arranged to detect the current through the SMA actuator 30. As the detection circuit 54 is a voltage detection circuit it draws much less current than the current supplied by the drive circuit 53, typically by a significant extent. For example, the input resistance of the detection circuit 54 is typically at least ten times the resistance of the SMA actuator 30 at 25 °C. This means that the voltage dropped across the electrical connections 57 due to their resistance is lower than the voltage dropped across the electrical connections 55. Typically the resistance of the electrical connections 57 has a negligible effect. Thus the detection circuit 54 detects the voltage across the SMA actuator 50 precisely and independently of the resistance of the electrical connections 57. In this way, variation in the resistance can be excluded from the resistance measurement and thus precision positional control of the SMA actuator 30 achieved. The implementation of the control circuit 50 will now be discussed with reference to

Figs. 11 to 24 which show various implementations for the control circuit 50.

A simple way to drive (thermally activate) the SMA wire 31 shown in Fig. 11 is for the drive circuit 53 to comprise a voltage supply 100 providing a drive voltage Vs and a switch 101 via which the voltage supply 100 is connected to the SMA wire 31. When the switch 101 is open there is no electro-thermal heating and the SMA wire 31 is easily extended by external forces (if in its full or partial Austenite phase or a compact Martensite phase) or is already relatively extended, and displays low stiffness (e.g. low Young's Modulus). Conversely, when the switch 101 is closed, the voltage supply 100 drives a drive current Is through the switch 100 and the SMA wire 31 which heats the SMA wire 31. Provided the magnitude of the drive voltage Vs of the voltage supply 100 is sufficiently high, this will cause sufficient heating to raise the temperature of the SMA wire 31 sufficiently to contract the SMA wire 31.

The heating and contraction of the SMA wire 31 in turn causes some change in the resistance Rsma of the SMA wire 31. Once into the active region, the usual negative temperature coefficient of the SMA wire 31 causes the drive current Is to increase, causing further heating and further lowering of resistance. This positive thermal feedback makes precise control of the drive current Is difficult with such a simple drive circuit 53. Note .that the true control variable is the power Ps delivered into the SMA wire 31, where Ps = Vs²/Rsma which is clearly a function of Rsma, which in turn is a function of temperature. However, in practice, the circuit will roughly stabilise and the SMA wire 31 will continue to provide a short stiff mechanical form until it is either pulled very hard by external forces, cooled externally, or the power is shut off again by opening the switch 101.

Fig. 12 shows an alternative simple control circuit 50 in which the drive circuit 50 comprises a current source 102 providing a drive current Is connected directly to the SMA wire 31, and a switch Sl provided to shunt the current away from the SMA wire 31 when closed, hi practice, with a non-ideal current source 102, the switch 103 could usefully be placed in series with the SMA wire 31 , and in this case heating of the SMA wire 31 occurs when the switch 103 is closed rather than open.

With this arrangement, with the switch 103 closed, the SMA wire 31 does not electro-heat. Upon opening the switch 103, the drive current Is flows through the SMA wire 31 and heats it much as described for Fig. 11, the major difference being that the magnitude of the drive current Is is now determined by the current source 102, and not by the resistance Rsma of the SMA wire 31, so there is no positive feedback. In fact the power delivered falls as the SMA wire 31 heats up through its active region so there is negative feedback. Thus precise current control is much simpler. The power Ps delivered to the SMA wire 31 is Ps = Is².Rsma which for a constant Is clearly reduces with reducing resistance Rsma (i.e. with increasing temperature in the active region).

Fig. 13 shows a variant of the control circuit 50 of Fig. 11 wherein the voltage source 100 provides a variable drive voltage Vs, thus allowing some control over the power delivered to the SMA wire 31. hi practice this arrangement is useful only for setting a maximum current and almost useless for precise positioning of the SMA wire 31, largely because of the positive feedback inherent in the system.

Fig. 14 shows a variant of the control circuit 50 of Fig. 12 wherein the current source 102 provides a variable drive current Is, thus allowing some control over the power delivered to the SMA wire 31. In practice this arrangement is useful for setting a maximum current and somewhat useful for positioning of the SMA wire 31 (because of the elimination of positive feedback) but the inherent hysteresis in the SMA material makes its use poor for precise positioning, as the actual contraction of the SMA wire 31 extension achieved for a given drive current Is is dependent on the mechanical load, the ambient temperature, hysteresis, and also the particular dimensions and composition of the SMA wire 31 itself.

The implementations of the control circuit 50 shown in Figs. 13 and 14 can be simply converted into digitally controllable versions by making the switches 101 and 103 digitally controllable (e.g. by use of an FET or BJT, or even a relay or other type of electro-mechanical switch) and driving the switches 101 and 103 with a pulse-width-modulated (PWM) switching waveform. In this way the average value of the PWM (determined by its mark-space ratio or MSR), a value between 0 and 1, effectively multiplies the value of Vs or Is, so far as the heating efficacy is concerned, and some proportional control of the SMA wire 31 is possible with a fixed voltage or current source and a digital PWM drive.

Fig. 15 shows an implementation of the control circuit 50 providing simple continuous (non-pulsed, analogue) precision drive. The drive circuit 53 comprises a PFET 104 as a supply connected in series with the SMA wire 31. In particular, the SMA wire 31 is connected to a power supply line Vss at one end, and to ground via a ballast resistor 105 and the PFET 104. The drive circuit 53 further comprises an operational amplifier 106 connected to control the PFET 104 in amplified voltage-follower mode (i.e. negative feedback around the PFET 104). This guarantees that the source voltage V2 of the PFET 104 is clamped to the control voltage Vo supplied by the controller 52 after digital-to-analogue conversion by a DAC 107.

The detection circuit 54 comprises the ballast resistor 105 and a capacitor 108 connected between (1) the junction of the SMA wire 31 and the ballast resistor 105 and (2) an ADC 109 at the input of the controller 52. The capacitor 108 inputs a voltage Vi into the controller 52 which is just the AC component of the voltage Vl at the junction of the SMA wire 31 and the ballast resistor 105.

The controller 52 outputs an output voltage Vo to the operational amplifier 106 of the drive circuit 53 having a DC component Vdc and an AC component Vac. The mean drive voltage to the SMA wire 31 and the ballast resistor 31 is just the mean component of the output voltage Vo which is Vdc. The drive current Is through the SMA wire 31 thus has a mean value Im where

Im = (Vss-Vdc)/(Rsma+R2) The AC component lac of the drive current Is is given by lac = Vac/(Rsma+R2)

Thus the AC component of the voltage Vl which is passed as Vi to the controller 52 via the capacitor 108 is

Vi = lac . Rsma = Vac.Rsma/(Rsma+R2) As the controller 52 knows the value Vac (because it is generated therein), and because R2 is a known fixed value, the controller 52 can thus compute the resistance Rsma of the SMA wire 31 as follows:

Rsma = Vi. R2/(Vac - Vi)

In practice the amplitude of Vdc is made to be a small fraction of Vss (e.g. < 1/10 and preferably <l/100), thus allowing continuous resistance monitoring of the SMA wire 31. In this way, smooth and almost noiseless precision control of the SMA wire 31 is possible with very little switching or other AC noise imposed on the voltage supply Vss. Because this is a class-A amplifier type, its power efficiency is relatively low.

Fig. 16 shows an implementation of the control circuit 50 wherein the drive circuit 53 comprises a switch 110 implemented for example by a FET and controlled by a PWM control signal from the controller 52. The switch 110 ia connected in series with the SMA wire 31 and a ballast resistor 111 between a voltage supply (magnitude Vs) and ground.

The detection circuit 54 comprises the ballast resistor and an ADC 112 the input of which is connected to the junction between the ballast resistor and the SMA wire 31 so that it receives the drive voltage Vs across the SMA wire 31. The digital output of the ADC 112 is supplied to the controller 52.

When the switch 112 is closed, the drive current Is through the ADC 112 is predominantly determined by the drive voltage Vs, the resistance Rsma of the SMA wire 31 and the resistance Rbal of the ballast resistor 111 in accordance with the equation

Is = Vs/(Rbal+Rsma) The voltage Vsma is

Vsma == Is .Rsma

The ADC 112 is either controlled by the controller 52 to sample its input at a time when the PWM control signal closes the switch 110, or samples continuously and is only read by the controller 52 when the switch 110 is closed. The ADC 112 then produces an output code C which is a ratio between Vsma and Vs, typically for an n-bit ADC as per

C = int(2n.Vsma/Vs) for 0 <= code <= 2n -1 As

Vsma/Vs= Rsma/(Rbal+Rsma) the controller 52 may compute the resistance Rsma in internal code units Csma, given the ADC output code C and a known fixed value for Rbal as per

Csma =C.Cbal/(C-l)

The controller implements a simple control algorithm to regulate the duty cycle of the PWM control signal to the switch 110 so as to produce the desired value of the resistance Rsma. Thus precision control of the SMA wire 31 is achieved. The resistance Rbal is chosen to satisfy one or more of several conditions. One possible condition is to limit the absolute maximum current Irnax through the SMA wire 31 as per

Rbal = Vs/Imax -Rsma_min where Rsma_min is the smallest value of the resistance Rsma of the SMA wire 31 reached in operation over the operating temperature range. Another possible condition is to make the power lost in the ballast resistor 112 small compared to the power delivered to the SMA wire 31 so

Rbal « Rsma_av where Rsma_av is usefully some average value of the resitance Rsma of the SMA wire 31 over the operating temperature range. Other similar configurations wherein the switch 110 is connected between ground and the series-connected resistor pair, and/or the relative order of series connection of the ballast resistor and SMA wire 31 are equally possible.

Fig. 17 shows an implementation of the control circuit 50 which is a development of the implementation of Fig. 16. The implementation of Fig. 16 has a large, essentially fixed DC component of voltage at the input to the ADC 112. Thus much of the dynamic range of the ADC 112 is effectively wasted just sensing that offset, and the small changes in the input voltage to the ADC 112 caused by the changes in resistance Rsma with temperature form only a small portion of its dynamic range. This limits resolution.

In contrast in the implementation of Fig. 17, the detection circuit 54 is modified as follows. A balancing arm containing a pair of resistors 113 is connected in parallel with an arm containing the SMA wire 3 land the ballast resistor 111. To allow the balancing arm to largely cancel the fixed offset component of voltage on the resistance Rsma, the detection circuit 54 further comprises a differential amplifier formed by an operational amplifier 114 with system gain determined by a feedback resistor 115, with the inputs of the operational amplifier 114 connected respectively to the junction between the SMA wire 31 and the balancing resistor 111 and to the junction between pair of resistors 113 in the balancing arm. In this way, the differential amplifier both differentially senses and amplifies the changing component voltage at a terminal of the SMA wire 31.

The use of the balancing arm and differential amplifier makes much better use of the dynamic range of the ADC 112 and allows either higher resolution resistance measurement and/or a simpler ADC 112 (with a smaller number of bits). Thus there is an improvement in the cost or performance or both.

The operation of this circuit is otherwise much as described for Fig. 16. The resistance values of the resistors 113 in the balancing arm are chosen such that the voltage at their common connection point, arising from the ratio of the resistance values, is approximately equal to the average voltage across the SMA wire 31, and such that the resistors draw as little supply current as possible, consistent with that current being sufficiently high relative to the input current of the operation amplifier 114, preferably at least 1 to 10 times higher.

Fig. 18 shows an implementation of the control circuit 50 which is effectively a constant-current analogue of the implementation of Fig. 16.

Here the drive circuit 53 comprises a constant-current current source 116 to which the controller 52 supplies a PWM control signal. The detection circuit 54 comprises an ADC 117 the input of which is connected across the SMA wire 31, eliminating the need for a ballast resistor as the current is now controlled directly. The ADC 117 samples the SMA wire 31 voltage and because the SMA wire 31 current Is is known, being fixed by the current source 116, the voltage sensed is directly proportional to the resistance Rsma of the SMA wire 31. This makes the internal computations much simpler, without any need for multiplies or divides, and therefore potentially cheaper on simple processors or controllers.

Fig. 19 shows an implementation of the control circuit 50 which is a development of the implementation of Fig. 18 in the same manner that the implementation of Fig. 17 is a development of the implementation of Fig. 16. Thus the implementation of Fig. 19 is modified as compared to Fig. 18 by the detection circuit 54 being the same as the detection circuit of Fig. 17, except that it is supplied with the voltage across the SMA wire 31. The use of the balancing arm in the detection circuit 54 provides essentially the same advantages as the implementation of Fig. 17 as described above. A potential weakness of the control circuits 50 of Figs. 17 and 19 is that precise offsetting is difficult typically requiring either (a) select-on-test or trimmable offset resistors, or (b) very tight control of the resistance of the SMA wire 31 (and associated ballast resistor 111 where fitted). In practice the SMA wire 31 resistance is difficult to control precisely because of mechanical, material and termination issues, including long- and short-term ageing of the SMA wire 31.

Fig. 20 shows an implementation of the control circuit 50 which is a modification of the implementation of Fig. 19 in which this potential weakness is overcome. Ih particular, an offset voltage which cancels most of the static voltage across the SMA wire 31 is not derived from a pair of resistor 113 in the balancing arm. Instead, the detection circuit 54 comprises a DAC 117 whose input is supplied with a digital offset signal from the controller 52. The DAC 117 converts the digital offset signal into an analog offset voltage signal which is supplied to the input of the operational amplifier 114 of the differential amplifier. Thus the offset voltage is selected by the controller 52 rather than being derived from the balancing arm, thereby overcoming the potential weakness discussed above. The controller 52 may determine the digital offset signal at run-time, e.g. every time the control circuit 50 is turned-on or reset, by running a short offset-determination algorithm, and may almost completely offset depending on the resolution of the DAC 117. This in turn allows a higher loop gain for the differential amplifier and therefore greater sensitivity to the changing resistance of the SMA wire 31 with temperature, resulting in higher resolution positioning accuracy, and/or a reduction in the necessary bit-width of the ADC 112. Overall, higher performance and/or lower cost are achieved. In this case, the current source 116 is set (i.e. its fixed current set) by a single fixed resistor. Whilst the current source 116 may also be programmable digitally by the digital controller 112 as an alternative, this possibility is prone to glitch and noise issues, as any sort of digital system fault could allow a dangerously high current value to be set that might damage the SMA wire 31 or the surrounding system. A current-set resistor is not prone to these problems. AU other embodiments incorporating a current-source may also optionally be current-set in the manner described here.

A useful optimisation of this control circuit 50, especially where it is manufactured in an Application Specific Integrated Circuit (ASIC), is to use the same voltage reference device to supply both the ADC 112 and the DAC 117. This not only saves parts/chip-space, but can also be used to cancel out at least one source of error.

Fig. 21 shows an implementation of the control circuit 50 which is a modification of the implementation of Fig. 20 to eliminate the need for the ADC 112 to sense and thus indirectly measure the resistance of the SMA wire 31. In particular, in the detection circuit 54, the differential amplifier and ADC 112 are replaced by a comparator 118 which compares the voltage across the SMA wire 31 (strictly speaking the voltage across the SMA wire 31 voltage plus the small voltage across the switch 110 which may be made small or negligible, or at least relatively constant or that corresponding to a constant switch on-resistance) with the analog offset voltage signal from the DAC 117. The comparator 118 outputs a binary output signal representing which of the two voltages is higher. The binary output signal is supplied to the controller 52 which increases or decrease the power of the drive current in dependence thereon.

In operation, the controller 52 drives the switch 110 with a PWM control signal of some duty cycle Ml, and sends a demand code-word to the DAC 117 corresponding to some desired length of the SMA wire 31 and therefore indirectly to some resistance of the SMA wire 31. If the output of the comparator 118 indicates that the voltage across the SMA wire 31 voltage is lower than the analog offset voltage signal, then the controller 52 increases the duty cycle to Ml+delta(Ml). Conversely if the output of the comparator 118 indicates that the voltage across the SMA wire 31 voltage is lower than the analog offset voltage signal, then the controller 52 reduces the duty cycle. In this way a negative feedback loop is created which rapidly causes the duty cycle of the PWM control signal to approach a value at which the SMA wire 31 voltage is very close (within one or at most two LSB of the DAC 117) to the analog offset voltage signal output from the DAC 117. Thereafter any changes to the analog offset voltage signal caused by the controller 52 sending new demand codes to it, causes the SMA wire 31 voltage to track the analog offset voltage signal. Thus the resistance and length of the SMA wire track that demand code. A suitable tracking algorithm within the digital-controller can be arranged to optimise loop control speed, control precision, and overshoot/undershoot. Fig. 22 shows an implementation of the control circuit 50 which is a development of the implementation of Fig. 17 in the same manner that the implementation of Fig. 21 is a development of the implementation of Fig. 19. Thus the implementation of Fig. 22 is modified as compared to Fig. 17 hy the detection circuit 54 heing the same as the detection circuit of Fig. 21, except that it is supplied with the voltage at the junction between the SMA wire 31 and the ballast resistor 111. The operation of the control circuit 50 is equivalent to that of Fig. 21. As compared to Fig. 21, the control circuit 50 of Fig 22 is slightly less complex but gives less good control of the SMA wire 31 drive current.

Fig. 23 shows an implementation of the control circuit 50 which employs continuous drive, i.e. essentially class-A analogue drive, not PWM drive.

The drive circuit 53 comprises a constant-current or constant-voltage current source 119 constituted by an amplifier, controlled by an analog control signal derived by a control DAC 120 converting a digital control signal output by the controller 52. The drive current from the current source 119 is supplied to the SMA wire 31 through a ballast resistor 121. The detection circuit 54 is identical to that of Fig. 20. m operation a drive current of known voltage or current (controlled by the control DAC 120) is driven into the SMA wire 31 and an offset-setting algorithm is first run to set the analog offset voltage signal and thus to optimise the available gain of the current source 119 and ADC resolution. Thereafter and until a further offset-determining cycle is needed, the setting of the analog offset voltage signal is kept fixed. The ADC 112 of the detection circuit 54 supplies to the controller 52 a code proportional to the resistance of the SMA wire 31, which in turn is varied by the controller 52 through writing appropriate codes into the control DAC 120 controlling the current source 119. The presence of the ballast resistor 121, in the case where a voltage amplifier is used, allows calculation of the SMA wire 31 current needed for SMA wire 31 resistance calculation.

Fig. 24 shows an implementation of the control circuit 50 which is a modification of the implementation of Fig. 23 in which the detection circuit 54 is the same as that of Figs. 21 and 22. The control and operation is equivalent similar to that of Figs. 21 and 22 with the difference of using continuous drive not PWM control. Some further detailed circuit implementations for the control circuit 50 are shown in

Figs. 25 to 28.

The circuit implementation shown in Fig. 25 is cheap but has limited performance. In particular, the drive circuit 53 is a constant-current current source implemented using a simple arrangement of bipolar transistors 120. The detector circuit 54 is a voltage detection circuit and is formed as a simple bridge arrangement of a pair of diodes 121 and a resistor 122.

The circuit implementation shown in Fig. 26 is more accurate but is more expensive. In particular, the drive circuit 53 is a constant-current current source implemented by a MOSFET transistor 123 controlled by an operational amplifier 124. The detector circuit 54 is a voltage detection circuit and is implemented by a bridge arrangement of two resistors 125, the output of which is amplified by an operational amplifier 126. The operational amplifier 126 allows the AfD converter of the controller 52 to make use of its full dynamic range.

The third circuit implementation shown in Fig. 27 uses a drive circuit 53 which is a constant-voltage current source implemented by a field-effect transistor. The voltage detector circuit 54 is a Wheatstone bridge arrangement formed by three resistors 55 which form a bridge circuit with the piece of SMA wire 31 and an operational amplifier 56 which measures the voltage differential between the two legs of the bridge circuit as a measure of the resistance of the piece of SMA wire 31. This circuit implementation has the advantage of a relatively low component count and also has good power supply rejection immunity. A power down mode is available by simply switching off the field-effect transistor forming the drive circuit 53.

The fourth circuit implementation shown in Fig. 28 uses the resistance of the SMA actuator 30 as a measure of position but is also simple to and configured as follows.

To avoid the complexities of resistance measurement, instead an auto-balanced Wheatstone bridge circuit is constructed around the SMA actuator 30 using three resistors Rl to R3. The first resistor Rl is in series with the SMA actuator 30 between the voltage line Vcc and earth, to form a first arm of the Wheatstone bridge. A second arm of the arm of the Wheatstone bridge is formed by the second and third resistors R2 and R3, which together are in parallel with the first arm. The resistance values of the resistors Rl to R3 is chosen such that when the bridge is balanced, i.e. when R1/R2 = R3/Rsma, then the SMA actuator 30 has a length equal to the desired actuation length.

A current source Ql supplies current through the two arms of the Wheatstone bridge. The current source Ql is a MOSFET device operating in a linear mode, providing a very low cost circuit. The current source Ql could have other forms, for example being a bipolar transistor or a more complicated arrangement of transistors.

The nodes between the SMA actuator 30 and the first resistor Rl and between the ^■ second and third resistors R2 and R3 are connected to the differential inputs of a differential amplifier ARl having a high gain. The differential amplifier ARl could be an operational amplifier, or a simple two-device differential amplifier (e.g. two bipolars or two FETs), or even simpler but with reduced accuracy a single transistor (preferably bipolar). The output of the differential amplifier ARl is used to control the current source Ql, with polarity of gain chosen to provide negative loop feedback. As power is turned on, current will flow through the SMA actuator 30 and current source Ql, causing heating of the SMA actuator 30. Once the SMA actuator 30 is heated to its active region (i.e. where its temperature coefficient of resistance is substantially negative because of SMA element contraction with temperature rise), then its resistance will fall, and the output of the differential amplifier ARl will change value in such a manner as to reduce the drive to current source Ql and thus to reduce the heating (the temperature coefficients of the three resistors Rl to R3 are assumed small, or substantially zero, as for any good quality resistor). At the point where the Wheatstone bridge is very nearly balanced, the differential input to the differential amplifier ARl becomes near to zero and its output becomes small enough to just maintain the required supply current needed to maintain balance.

Thus, the control circuit 50 drives the SMA actuator 30 to the desired resistance relatively independently of ambient temperature or mechanical load on the SMA actuator 30. In this way resistance control is achieved by resistance feedback, that being effectively the quantity (strictly resistance ratio) that is fed back to the error terminal of the closed loop control circuit. This results in a very robust, stable, reliable and low cost control circuit 50. The output of the differential amplifier ARl is connected to the gate of the current source Ql through a damping circuit formed by a first capacitor Cl and a fourth resistor R4.

Typically the resistance value of the first resistor Rl will be chosen to limit the maximum controlled current allowed through the SMA actuator 31. Then given the desired controlled resistance Rsc of the SMA actuator 30, then the resistance values of the second and third resistors R2 and R3 are chosen such that Rl/Rsc = R2/R3. As the current through the second and third resistors R2 and R3 is used only to bias one of the inputs of the differential amplifier ARl, then the total resistance value of the second and third resistors R2 and R3 can be maintained as high as possible subject to the ratio R2//R3 (i.e. R2 in parallel with R3) being significantly lower than the input impedance of the differential amplifier ARl to avoid accuracy problems from amplifier loading. Such a choice avoids the wasting of unnecessary power in R2 and R3.

The other components of the control circuit 50 are as follows. Offset voltages are provided to the differential amplifier ARl to allow the finding of the maximum resistance Rmax of the SMA actuator 30 as follows. The controller 52 includes a PWM circuit 59 which provides a PWM signal PWM Input The signal PWM Input is supplied through a filter circuit consisting of a second capacitor C2 and fifth and sixth resistors R5 and R6, to the node in the second arm of the Wheatstone bridge between the second and third resistors R2 and R3. The fifth and sixth resistors R5 and R6 of the filter circuit are set such that the minimum duty cycle of the signal PWM Input corresponds to a desired resistance of 21Ω (minimum possible resistance of the SMA actuator in this particular actuator design), and the maximum duty cycle of the signal PWM Input corresponds to a desired resistance of 29Ω. The control circuit 50 further includes a linear region detection circuit comprising a second MOSFET Q2 and a seventh resistor R7 in series, the output of the differential amplifier ARl being connected to the gate of the second MOSFET Q2. The output signal Linear is supplied to the controller 52 and goes low when the current source Ql is active.

The control circuit 50 further includes an overdrive detection circuit comprising a third MOSFET Q3 and a eighth resistor R8in series, the node between the current source Ql and the SMA actuator 30 being connected to the gate of the third MOSFET Q3. The output signal Overdrive is supplied to the controller 52 and goes low when the SMA actuator 30 has been heated beyond its minimum resistance Rmin, whereafter little or no further actuation is possible and beyond which damage to the SMA actuator 30 can sometimes occur. The operation of the control circuit 50 is implemented by the controller 52 as follows.

The following calibration cycle is applied. The signal PWM Input is first set high, which guarantees that the SMA actuator 30 is inactive. The duty ratio of the signal PWM Input is then decreased, 8bits or so at a time, until the output of the differential amplifier ARl increases, as detected by a low on the signal Linear. The current value of the duty ratio of the signal PWM Input is saved, as P WMstart. Then the duty ratio of the signal PWM Input is increased, bit by bit, until the signal Linear switches high. This value of duty ratio of the signal PWM Input is saved, as PWMrmax.

After this calibration cycle, the duty ratio of the signal PWM Input is set to the value of P WMstart. After 100ms to allow the SMA actuator 30 to stabilize, the duty ratio of the signal PWM Input is changed to the value PWMrmax + Pdelta, where Pdelta is pre-programmed to a macro position.

If the signal Overdrive line pulls low continuously, this is an error condition, associated with control failure. This circuitry may be omitted if there is sufficient margin between normal operation and overdrive operation. The focal position of the lens element 6 changes as follows.

The focussed object distance for a given displacement (from the infinity position) is given by the equation:

EFL * ( 1 + EFL / Displacement ) where EFL is the effective focal length of the lens element 6 and Displacement is the position of the lens element 6. Typical values for EFL are: 3.85e-3 for a Largan lens used in a current 8.5mm camera 1; 3.55e-3 for an equivalent Sekonix lens; and 2.54e-3 for a Genius lens matched to a 1/6" sensor.

Considering performance when the second drive signal has a constant current, at room temperature, the end stop which mechanically sets the far focus position is positioned to +/- 15μm tolerance. Several schemes are then envisaged, with focal length tolerances calculated for each scheme. With no temperature compensation, and no settle time compensation, for operation between 0⁰C and 50⁰C, using a constant current source of 4OmA, the position settles to: lOμm to 60μm at 0⁰C

160μm to 230μm at 25⁰C. 330μmto 370μm at 50°C

Thus at 25 °C, the camera focal distance is 6.4cm to 10.6cm. Over the range 0⁰C to 50⁰C, the lens element 6 is poorly controlled, and moves from lOμm to 370μm, overall focal distance varying from infinity to 4cm.

With ambient temperature compensation as described above (power of second drive signal reducing with ambient temperature), and no settle-time compensation, the focal position spread of focal distance would be 6.2cm to 33cm, over 0⁰C to 50⁰C. The camera takes 2s to stabilise to 30μm accuracy.

Using a simple constant-current drive waveform results in a long time-constant, as the actuator reaches its final operating temperature exponentially over several seconds in practice. It is possible to speed this up dramatically by using a more sophisticated open-loop digital control of the constant current source. In this way, the actuator achieves steady state position within 100ms, to a tolerance of +/-3μm. With such settle-time compensation and ambient temperature compensation as described above (power of second drive signal reducing with ambient temperature), the focal position spread of focal distance would be 6.2cm to 33cm, over O⁰C to 50⁰C, within 100ms.

The constant-current drive technique can be given improved accuracy by doing one calibration-on-test operation. Essentially this involves (during the module test-cycle) controlling the lens element 6 so as to tightly focus on a target held at the desired macro-focus position (e.g. 10cm), and measuring the required drive current at ambient temperature. This value is then stored somewhere in the processor control system and used to modify the default pre-set current value, with or without temperature compensation. In this way ambient temperature camera-to-camera variability is almost completely eliminated. The expected performance is for the actuator to achieve steady state position within 100ms, to a tolerance of +/-30μm with a focal position spread of focal distance at 25°C of from 8.6cm to 10.3cm, and over O⁰C to 50°C of from 6.2cm to 13.3cm.

Use of the drive circuit of Fig. 28 employing resistance feedback is anticipated to give good performance from 0⁰C through 25°C to 50⁰C, with positional stability of 180μm +/- 40μm possible over this temperature range. It would not operate correctly significantly below 0°C. The expected variation of focal length is thus 7.1cm to 1 lcm over 0°C to 50°C.

The control circuit 50 may implement an autofocus algorithm. In this case, the control may be based on a measure of the focus of the image, for example a modulation transfer function or a spatial frequency response, derived by the controller 52 from the image signal from the image sensor 4. A wide range of suitable measures are known and any such measure may be applied. In this case, there is a limitation that the derivation of the measure of focus is slow. To combat this, during a scan across many focus positions, at the desired focus position determined from the measure of focus, the control circuit 50 may determine the resistance value. Then at the end of the scan the lens element 6 is driven back to the same position of the basis of that resistance value rather than the focus measure. In this case, an image signal from the image sensor 4 is used to derive the primary feedback parameter, any drifting in the absolute values of the measure of the position as a secondary parameter over repeated cycles and age is irrelevant, as there is no perceivable change over the course of a single auto-focus cycle. In a given camera 1, the resistance might vary from 10Ω at high temperature to 12Ω at low temperature, and then over the course of several 100k cycles, this may change to 15Ω at high temperature and 20Ω at low temperature. However, for any given cycle, best focus will correspond to a specific resistance to a sufficient degree of accuracy. Thus it is only necessary to return to this specific resistance, irrespective of its absolute value.

An example of a control algorithm which may be performed by the controller 52 is shown in Fig. 29 and will now be described. By way of illustration, reference is also made to Fig. 30 which illustrates an example of the change in resistance of the lens element 6 with time. This control algorithm uses the local maximum resistance 60 as a reference.

Ih step Sl, power is supplied to the camera 1 and the control circuit 50. The unheated piece of SMA wire 31 is in the Martensite phase. In step S2, a command to capture an image is awaited. After receiving the command, there is performed a focus detection operation 01, followed by a focussing operation 02. The focus detection operation Ol comprises the following steps. In step S3, the control circuit 50 heats the piece of SMA wire 31 from its unheated state. Initially the heating is achieved by the control circuit 50 supplying PWM current with the maximum duty cycle. Thus the resistance rises as shown by the curve 71. As the heating continues, a local maximum resistance 72 is encountered, corresponding to the local resistance maximum 60 in Fig. 8.

During the heating, the controller 52 monitors the voltage across the piece of SMA wire 31 detected by the detector circuit 54 as a measure of the resistance of the piece of SMA wire 31 to detect the local maximum resistance 72,

On detection of the local maximum resistance 72, in step S4 there are derived a series of target values 73, 74. Firstly an upper target value 73 is derived from the resistance value of the detected local maximum resistance 72. The upper target value 73 may be the resistance value of the local maximum resistance 72 detected in step S3, but is more preferably that resistance value less a predetermined decrement where the greater slope of the curve shown in Fig. 8 allows for more accurate positional control. Then a predetermined number of further target values 74 at predetermined decrements below the upper target resistance value are derived. In Fig. 30, a limited number of further target values 74 are shown for ease of illustration, but in general there may be any number of target values 73, 74. Indeed a particular advantage of the camera 1 is that it is possible to achieve accurate positional control to a large number of positions.

The target values 73, 74 may be disposed linearly across the range, but there could alternatively be an unequal spread for example concentrated in a particular part of the range.

Ih steps S5 to S7, there is performed scanning across the series of target values 73, 74. This is achieved by setting in step S5 successive ones of the series of target values to the feedback control loop so that the piece of SMA wire 31 is heated to that one of the target values. Thus in Fig. 30, the resistance is driven to successive plateaus 75 each at the level of one of the target values 73, 74. As the target values 73, 74 successively decrease, the temperature of the piece of SMA wire 31 is raised monotonically.

In addition, during the seeking of the further target values 74, step S5 may employ a safety routine as described below.

After the measured resistance has reached a given target value 73, 74 in step S5, then in step S6 an image is captured by the image sensor and a measure of the quality of the focus of the image signal output by the image sensor 4 is derived and stored in the memory of the controller 52. In step S7, it is determined whether there are any remaining target values 74 in the series. If so, and provided the local resistance minimum has not been detected, the method returns to step S5 so that the process is repeated for each of the target values 73, 74. In this way, as the piece of SMA wire 31 is heated during the scanning, the quality of focus of the image signal is monitored. ????

The safety routine which forms part of step S5 will now be considered. The series of target values 73, 74 are intended all to be above the predicted resistance value of the local resistance minimum 61, based on the expected properties of the piece of SMA wire 31. However, there is a risk that a target value is below the actual resistance value of the local resistance minimum 61, for example due to manufacturing tolerances in the components of the camera 1 or physical changes in the SMA wire over its lifetime. If this does occur, then there is a possibility that as a result of the feedback loop, the controller 52 could cause the SMA to continue to be heated in an attempt to seek a target value 74 that is unattainable. This could damage the piece of SMA wire 31. Thus a safety routine is performed as part of step S5 in which the measured resistance is monitored to detect the local resistance minimum 76, corresponding to the local resistance minimum 61 in Fig. 8. If this is detected, then the controller 52 immediately reduces the power supplied to the piece of SMA wire 31. Then a new target value 78 is set a predetermined increment above the resistance value of the detected local • resistance minimum 74.

Fig. 30 illustrates an example in which the final further target value 74 is below the actual resistance value of the local resistance minimum 76. In this case, the heating in step S5 to seek the final further target value 74 causes the local minimum resistance 76 to be reached. This is detected by the safety routine and the power is reduced which causes the resistance to fall back through another local minimum resistance 77. Thereafter a new target value 78 is set a predetermined increment above the detected local minimum resistance 76. The new target value 78 is set to the feedback control loop so that the piece of SMA wire 31 is heated to drive its resistance to a plateau 79 at the level of the new target value 78.

Furthermore, if the local resistance minimum 76 is detected, thereafter the remaining target values in the series are not used. Similarly, the resistance value of the detected local resistance minimum 76 is stored and thereafter in step S4, when a series of target values 73, 74 are derived any target values 73, 74 below the stored resistance value are rejected from the series.

In step S8, the stored measures of focus quality are used to derive a focus value of the control signal at which the focus quality is at an acceptable level. Most simply this is done by selecting one of the plurality of test values having the best measure of focus quality. As an alternative, it is possible to predict the value of the resistance which would provide the best focus from the test values using a curve-fitting technique. Thus the focus value need not be one of the test values. The curve fit may be a simple mathematic equation such as an Mth order polynomial where M>1 or instead could be chosen as a best-fit to a curve taken from a library of curves premeasured from representative scenes. The focus value is stored in the memory of the controller 52 for subsequent use. Fig. 30 illustrates an example for the stored focus value 80.

Instead of determining the focus value after steps S5 to S7, it could alternatively be determined on-the-fly during steps S5 to S7.

The focussing operation O2 comprises the following steps. In step S9, flyback occurs. Ih particular, the control circuit 50 allows the piece of SMA wire 31 to cool back into the Martensite phase. This may be achieved by applying a PWM current with a minimum duty cycle, although it could alternatively be achieved by applying no current at all. The transformation into the Martensite phase indicating the end of the flyback phase can be detected by the controller 52 monitoring the voltage measured by the detector circuit 54. Alternatively, the flyback phase can simply be maintained for a pre-determined time selected to be sufficiently long to allow the piece of SMA wire 31 to cool under any expected operating conditions. The flyback is shown by the curve 81 in Fig. 30.

Next, in step SlO the control circuit 50 heats the piece of SMA wire 31 to return it to the position corresponding to the focus value determined and stored in step S8. This is achieved by the control circuit 52 applying the feedback control technique with the stored focus value 80 being used as a target value so that the measured voltage across the piece of SMA wire 31 used as the feedback signal is driven to that stored focus value 80. The temperature rise is again monotonic, as in the focus detection operation 01. Thus in Fig. 30 during the heating the resistance changes as shown by the curve 82 and then is driven to a plateau 83 at the level of the stored focus value 80. As discussed above, as a result of the flyback technique achieved by the inclusion of step S9, the focus value 80 is approached on the heating cycle and thus the problem of hysteresis in the piece of SMA wire 31 is overcome. Accordingly the lens element 6 is known to be at the position corresponding to the stored focus value 80.

As the image is now properly focussed, in step Sl 1 an image is captured by the image sensor 4. The captured image is stored in a memory.

Whilst the embodiments described above relate to a camera incorporating an SMA actuation arrangement which drives movement of a camera lens element, the SMA actuation arrangements described can equally be adapted to drive movement of an object other than a camera lens element.

Claims

1. A control system for a shape memory alloy actuation apparatus comprising an SMA actuator arranged on heating to drive movement of a movable element, the control system comprising: a current source operable to pass current through the SMA actuator to heat the SMA actuator; a bridge arrangement of resistors comprising a first resistor in series with the SMA actuator and a second and third resistor in series with each other and together in parallel with the first resistor and the SMA actuator together; a differential amplifier having differential inputs connected respectively to the node between the SMA actuator and the first resistor and to the node between the second and third resistors, the current source being controlled on the basis of the output of the differential amplifier.

2. A control system according to claim 1, wherein the current source is a linear device.

3. A control system according to claim 2, wherein the current source is a MOSFET arranged to operate in its linear region.

4. A control system according to any one of claims 1 to 3, wherein the differential amplifier is an operational amplifier.

5. A control system according to any one of claims 1 to 4, wherein the differential amplifier has a high gain.

6. A control system according to any one of claims 1 to 5, wherein the output of the differential amplifier is connected to the input of tfie current source.

7. A control system according to claim 6, wherein the output of the differential amplifier is connected to the input of the current source through a damping circuit for damping the output of the differential amplifier.

8. A control system according to any one of claims 1 to 7, further comprising an adjustment circuit arranged to apply an adjustment voltage to the node between the second and third resistors.

9. A control system according to claim 8, wherein the adjustment circuit comprises a filter circuit arranged to filter a pulse-width modulated signal applied thereto.

10. A control system according to claim 9, wherein the adjustment circuit further comprises a pulse-width modulation circuit arranged to derive a pulse-width modulated signal and apply it to the filter circuit.

11. A control system according to any one of claims 1 to 10, wherein the movable element is a camera lens element, and the shape memory alloy actuation apparatus is arranged on heating to drive movement of a the camera lens element along the optical axis.

12. A control system according to claim 11, wherein the camera lens element includes one or more lenses having a diameter of at most 10mm.

13. A control system for an SMA actuator capable on heating of driving movement of a movable element, the control system comprising: a current source arranged to supply a drive current of variable power through the SMA actuator; a digital controller arranged to control the current source to vary the power of the drive current supplied thereby; a DAC, the digital controller supplying a digital offset signal to the DAC and the DAC being operative to convert the digital offset signal into an analog offset voltage signal; and a comparison circuit having two inputs and an output, the first input being supplied with a voltage providing a measure of the resistance of the SMA actuator, the second input being supplied with the analog offset voltage from the DAC, the comparator being arranged to compare the voltages at the two inputs and to produce an output signal at the output representing a comparison between the voltages at the two inputs, the controller being supplied with the output signal from the output of the comparison circuit and being arranged to vary the power of the drive current supplied by the current source in response thereto.

14. A control circuit according to claim 13, wherein the comparison circuit is arranged to produce a digital output signal at the output representing the difference between the voltages at the two inputs.

15. A control circuit according to claim 14, wherein the comparison circuit comprises a differential amplifier and an ADC connected to the output of the operational amplifier.

5

16. A control circuit according to claim 13, wherein the comparison circuit is a comparator arranged to produce a binary output signal at the output representing which of the voltages at - the two inputs is higher, the controller being supplied with the output signal from the output of the comparator 1.0 and being arranged to increase or decrease the power of the drive current supplied by the current source in dependence on which of the voltages at the two inputs of the comparator is higher.

17. A control circuit according to claim 16, wherein the comparator is arranged to produce 15 the output signal as a digital signal.

18. A control circuit according to claims 13 to 17, wherein the current source is a PWM current source arranged to supply a pulse-width modulated drive current having a variable duty cycle to vary the power of the drive current.

20

19. A control circuit according to claim 18, wherein the controller is arranged to vary the power of the drive current supplied by the current source in response to the output of the comparison circuit whilst the PWM current source is supplying a pulse of current.

25 20. A control circuit according to claims 13 to 17, wherein the current source is a constant- current current source arranged to supply a drive current having a variable voltage to vary the power of the drive current.

21. A control circuit according to claim 20, wherein the first input of the comparison 30 circuit is supplied with the voltage across the SMA actuator.

22. A control circuit according to claims 13 to 17, wherein the current source is a constant- voltage current source arranged to supply a drive current having a variable current to vary the power of the drive current.

35

23. A control circuit according to claims 22, wherein the control circuit further comprises a sensing resistor in series with the SMA actuator, and the first input of the comparator is supplied with the voltage appearing between the sensing resistor and the SMA actuator.

24. A control circuit according to claims 13 to 23, wherein the controller supplies a variable a digital offset signal.

25. A control circuit according to claims 13 to 24, wherein the digital controller and the DAC are integrated into a common integrated circuit chip.

26. A control circuit according to claims 13 to 25, wherein the movable element is a camera lens element, and the shape memory alloy actuation apparatus is arranged on heating to drive movement of a the camera lens element along the optical axis.

27. A control circuit according to claims 13 to 25, wherein the camera lens element includes one or more lenses having a diameter of at most 10mm.