Brushless electric motors are well-known and are used in a wide variety of applications, providing the advantages that there are no brushes to wear, and there is no commutator arcing.

A typical brushless motor comprises a stator assembly having a toothed iron core and multiple phase windings around the teeth, and a rotor assembly including an even number of permanent magnets.

Commutation circuitry controls the energisation of the different phase windings, generating magnetic fields which interact with those of the rotor magnets to produce rotation. AC and DC brushless motors are known. In DC brushless motors the commutator circuitry switches the currents in each phase between discrete steps, i. e. between +1,0 and-I where the magnitude of the current I is determined according to a desired speed or power output of the motor.

In AC brushless motors, the current supplied to each phase is sinusoidal. AC commutation circuitry is typically more sophisticated than its DC counterpart, and requires a high resolution feedback of the rotor position with respect to the stator.

Many different brushless motor configurations are known, with internal or external rotors and with different combinations of numbers of rotor magnets (poles) and winding phases. A common configuration is three-phase, four-pole and an example of such a motor is shown schematically in Fig. 1. The stator assembly

1 and rotor assembly 2 are generally cylindrical, but the motor is depicted linearly for simplicity.

The stator assembly 2 comprises a stator core of ferrous material having a plurality of teeth 11 radially extending from a generally annular portion 13. The generally annular portion 13 forms the stator core"back iron". The teeth 11 extend either radially inwards or outwards depending, of course, on whether the rotor is internal or external respectively. The stator assembly also comprises three-phase windings in the slots 12 defined by the teeth 11, the three-phases being labelled A B and C.

The rotor assembly 2 comprises a generally annular portion including a steel ring (cylinder) 21 to which four permanent magnets 22 are attached. The magnets are arranged such that their polarities alternate. The steel ring 21 forms the"back iron"of the rotor assembly, providing a low reluctance path for magnetic flux between adjacent magnets.

There are three times as many teeth as magnets, and the"width"of three adjacent teeth corresponds approximately to the"width"of one rotor magnet pole.

The windings of each phase form a string of coils, each coil being wound around (encompassing) three adjacent teeth. One of the coils corresponding to Phase B is shown schematically in Fig. l (b), encompassing nominal teeth 5,6 and 7. The coils of the different phases overlap, such that any one tooth is encompassed by coils of all three phases. This winding configuration is known as a"lapped winding".

Typically, the coil strings of the three phases are connected in"star"configuration, as shown in Fig. 2, although other configurations e. g. delta, are also known.

Sensing apparatus 6 generates a signal indicative

of the position of the rotor with respect to the stator, and this signal is used by commutation circuitry to control excitation of the windings.

In order to produce rotation, one known commutation method involves six steps, with two of the phase coils being energised at a time (i. e. in each step). This so-called six-step commutation method is shown schematically in Fig. 3. The degrees referred to in the figure are electrical degrees. In a four-pole motor, there are two full electrical cycles for one full rotation of the rotor with respect to the stator, and so 60 electrical degrees correspond to an angular rotation of 30°.

In the six-step method, from a nominal starting point with the rotor positioned at zero electrical degrees with respect to the stator, the commutation circuitry first energises phase coils A and B. Then, when the rotor has rotated 60 electrical degrees the commutation circuitry switches, now energising phases A and C. The method continues similarly, switching for every 60 electrical degrees. The current in each phase is thus switched from +I to 0 to-I to 0 etc, and hence a motor controlled in this way is a DC brushless motor.

Clearly, in order to perform this method, the position signal received by the commutation circuitry must be indicative of the rotor position to within 60 electrical degrees. In other words, as the rotor rotates through 180 actual degrees, the position signal must have six distinct states.

One of the ways in which rotor position has been sensed in the past has been to couple the rotor to drive the shaft of an encoder, or to mount encoding apparatus on the rotor itself. The encoder apparatus generates pulses as the rotor rotates and includes a

pulse counter to provide an indication of the rotor position. The encoder apparatus can be chosen to give the desired resolution, and high resolution encoders are usually employed with AC brushless motors.

However, the addition of encoder apparatus increases the complexity, cost and weight of the combined motor and commutation circuitry apparatus.

Another position sensing technique has employed Hall effect sensors, and in particular so-called digital Hall effect devices."Digital"Hall effect devices are well-known and incorporate a Hall effect sensing element and a latch (e. g. Schmidt trigger) circuit. The output from the device has two states, and is switched between these states when the component of magnetic flux density sensed by the sensing element crosses respective thresholds.

An example of a known digital Hall device, used in the prior art and suitable for use in embodiments of the present invention, is shown in Fig. 4. This device is available from Allegro Microsystems Inc, Massachusetts USA.

The active area, i. e. the Hall effect sensing element, is sensitive to magnetic flux B in the direction shown in Fig. 4 (A). In operation, the output is off until the strength of the magnetic field perpendicular to the surface of the chip exceeds the threshold or operate point (BoP). When the field strength exceeds Bop the output switches on. The output switches off when magnetic field reversal results in a magnetic flux density below the off threshold (Brp). This is illustrated in the transfer characteristics graph shown as Fig. 4 (c). The device latches, i. e, a south pole of sufficient strength will turn the device on. Removal of the south pole will leave the device on. The presence of a north pole of

sufficient strength is required to turn the device off.

It would be desirable to use digital Hall devices as described above to detect rotor position by sensing the magnetic fields of the rotor magnets directly.

However, there are numerous problems. The first is how to sense the magnetic field of the rotor magnets without being affecting by the changing fields generated by the three-phase stator windings.

In the example of Fig. 1, even if space permitted the location of Hall devices in the region between the rotor assembly and the stator assembly, the Hall effect device would be influenced by both the stator and rotor magnetic fields. Secondly, if the digital Hall device were positioned on the opposite side of the ring 21 from the magnets 22, it would be "screened"from the field of the magnets 22, i. e. substantially all magnetic flux would be confined to the back iron 21.

One attempt to solve this problem has been to cut slots in the ring 21 to expose digital Hall effect devices mounted on the opposite side to the fields of the rotor magnets. Disadvantages of this approach, however, are that the ring is weakened, and the reluctance of the total magnetic circuit is increased thereby decreasing the magnetic field in the air gaps between the stator teeth and the rotor magnets and so reducing torque for a given winding excitation.

Another attempted solution has been to arrange digital Hall devices outside the stator assembly to detect fringe fields from the rotor magnets at the axial ends of the rotor assembly. These fringe fields are, however, of reduced magnitude and are easily distorted. This can lead to irregular shifts in commutation points and a corresponding decrease in the

motor torque.

A further attempted solution has been to couple the rotation of the rotor assembly to another array of magnets (e. g. a NSNS ring magnet), outside the stator assembly, and arrange digital Hall devices to detect the fields of, and switch in response to the rotation of, this array. A disadvantage of this approach however, is that the weight, complexity and cost of the entire motor and commutation apparatus is increased.

A known problem with DC brushless motors is torque ripple, i. e. variations in the torque produced, effectively resulting from the differences between the drive voltage and the back EMF in each phase winding.

A method for reducing torque ripple in a DC brushless motor is described in US 4,758,768. The improved method comprises a total of twelve commutation steps, the six steps shown in Fig. 3 plus the intermediate steps shown in Fig. 5. The phase coils are energised according to the sequence Step 1, Step lA, Step 2, Step 2A etc, the switching occurring every 30 electrical degrees.

Therefore, in order to implement this method, rotor position sensing to within 30 electrical degrees is required. US 4,758,768 discloses apparatus for providing the necessary rotor position feedback to the commutation circuitry. This apparatus is shown in Figs 6 and 7. The motor includes a housing 90 out of which the rotor shaft 27 protrudes. A four pole ring magnet 65 is mounted on the shaft 27 outside the stator assembly, and an array of six digital Hall devices 61 are mounted on an end plate 19, fixed with respect to the stator, to detect the fields of (and switch in response to rotation of) the ring magnet 65.

A disadvantage of this arrangement is that by

adding the ring magnets, the weight, complexity and cost of the motor is increased.

A known problem with brushless motors such as the example shown in Fig. 1 is detent torque. This is caused by the tendency of the rotor magnet edges to align with the stator tooth edges.

An attempted solution to this problem in conventional lapped winding motors has been to skew the teeth with respect to the rotor magnets. This does, however, increase the complexity of the stator core and the stator windings.

Electrically powered vehicles are well known in which electric motors (including both internal and external rotor brushless motors) are arranged to provide drive (i. e. propulsion and/or steering).

One such example is shown schematically in Fig 8.

In this example the electric vehicle is an electric wheelchair controlled by a user by means of a joystick. Left and right wheels 5 are driven by electric motors 10, with steering being achieved by deferential drive of the motors. A controller controls the supply of power from a battery to the motors 10 according to signals received from the joystick.

Electric wheelchairs are known in which the motor 10 is arranged inside the wheel 5.

Japanese patent application number 63115210 discloses an electric wheelchair comprising an outer (ie. external) rotor brushless dc motor arranged in a wheel. The disclosed wheel and motor is shown schematically in Fig. 9. A wheel 5 supports a tyre 51, and the motor includes a rotor assembly having a rare earth magnet 22 and a back surface yoke 21 made of pure iron on the outer periphery supported by a hub 52. The rotor assembly rotates with respect to the stator assembly on a bearing 19. A stator 1 is a

driving coil which is wound onto a stator yoke 13 and the stator winding is arranged to enable a three phase all-wave drive. The application describes how since the outer rotor brushless dc motor is arranged in the wheel, a lightweight motor-driven wheelchair having high performance and reliability can be realised.

Noise can also be reduced.

Electric motor driven vehicles are known that incorporate a safety brake which is spring applied and released by an electro-magnet. An example of a known wheel and safety brake arrangement is shown schematically in Fig. 10.

The safety brake is typically only applied when the vehicle has come to a halt. Dynamic (i. e. regenerative) braking is employed to slow the vehicle down. The safety brake prevents undesired rolling of the vehicle, and it is usually an important design feature that in the event of electrical power supply failure or interruption the safety brake is applied.

Thus, the safety brake is usually biased to apply a braking force to the wheel 5, usually by means of a ferrous member 81 urged into contact with a surface of the wheel 5 by a spring 82.

The brake assembly includes an electro-magnet having an iron yoke 84 and windings 83. When the windings of the electro-magnet are excited the electro-magnet attracts the ferrous member 81, providing an attractive force Fa opposing the force Fb applied to the member by the spring 82. Thus, by attracting the ferrous member against the spring and away from the wheel 5, substantially free rotation of the wheel is enabled. To maintain free rotation, the current I must continue to flow in the electro-magnet windings 83. In the event of a fault interrupting the current supply the electro-magnet releases the ferrous

member 81 and the spring again brings it into contact with the wheel, providing a frictional force which inhibits rotation.

According to a first aspect of the present invention there is provided an electric motor comprising a stator assembly and a rotor assembly, the stator assembly including a stator core having X teeth, and multiple phase windings comprising X coils, each coil being wound around a single different respective one of said teeth, where X = 2mP, where m is an integer and P is the number of phases and the rotor assembly having Y magnets for interacting with magnetic fields generated by the stator assembly to provide rotation, where Y = X + 2n and is positive, where n is an integer and both X/Y and Y/X are not integers. Thus, a motor embodying this first aspect differs from a conventional motor in which the number of stator core teeth is P times the number of rotor magnets (where P is the number of phase windings i. e. phases). In the prior art designs such as that shown in Fig. 1, the edges of all rotor magnets could align with edges of the stator teeth simultaneously, resulting in a large detent torque. In a motor embodying this first aspect, however, there is a mismatch between the number of stator teeth and the number of rotor magnets, and the number of teeth is not an integer multiple of the number of magnets (and vice versa). This produces a"Vernier effect"i. e. a beating pattern between the teeth and rotor magnets which are spaced around the rotational axis R at different angular intervals, and ensures that not all of the magnets can align with the teeth simultaneously, so reducing detent torque (cogging).

Preferably n = 1 so that the numbers of teeth and magnets differ by just two. In this arrangement only

two rotor magnet leading edges may align with teeth edges at one time. This provides the advantage that detent torque is greatly reduced.

Furthermore, unlike the case with conventional lapped windings, in this first aspect of the present invention each tooth is wound (i. e. encompassed) by only one coil, that coil corresponding to one of the multiple phases. The coils do not overlap. For a given number of rotor poles, the stator core of the present aspect has fewer teeth than conventional arrangements, and those teeth are correspondingly wider, as are the slots in between.

These features mean that winding of the stator windings is facilitated, increased packing factors can be achieved, and more efficient use can be made of winding material as less material is used in the end portions of the coils axially outside the slots.

Advantageously, the reduction in detent torque resulting from the stator tooth and rotor magnet configuration enables the inventive motor to operate smoothly even at low speeds.

The rotor assembly may be substantially external to the stator assembly, and in this form the inventive motor is particularly suited for application as a brushless in-wheel motor for an electrically powered wheelchair, forming a direct replacement for conventional brushed gear motors. The inventive motor may be arranged to drive the wheel or tyre of the wheelchair directly, providing the advantage that the elimination of a gearbox and brushes gives quieter and more efficient performance, excellent reliability and simpler construction. The low speed performance of a gear motor can be matched as a result of the"Vernier effect"reducing detent torque, and the number of teeth and poles can be chosen to reduce torque ripple.

Detent torque is caused by the tendency of rotor magnet edges to align with the stator tooth edges. In embodiments of this first aspect of the present invention, the"Vernier effect"almost eliminates detent torque by reducing the number of aligned edges at any angle to a minimum (2).

The employment of the"Vernier effect"has other important implications for the inventive motor. It is now possible to produce a motor with a high magnet pole count suitable for low speed operation having a number of stator teeth only two greater than the number of rotor magnets. In contrast, conventional lapped winding design requires the number of stator teeth to be three times the number of rotor magnets (for 3-phase stator windings).

Advantageously the motor may be three-phase (i. e.

P = 3) and the stator core may have 36 teeth and the rotor may have 34 poles. The larger the number of teeth and poles, the higher the electrical frequency for a given rotational speed, and hence the smoother the low speed operation.

The stator teeth of the inventive motor require no skew to reduce detent torque as with conventional lapped winding design and so the core construction is simplified.

In the inventive motor, the windings span one tooth each, reducing the end winding length and overhang to a minimum. This gives reduced winding resistance and a more compact and easier to produce design.

Advantageously, the coils wound around any two diametrically opposed stator core teeth may correspond to the same respective phase. In the case of three phase windings, a respective one-third of the total number of coils may correspond to each phase, and the

coils of each phase may be split into two diametrically opposed sets.

Splitting the phase coils diametrically provides the advantage that the forces on the rotor assembly caused by interaction with fields generated by the stator are symmetric, and hence so are the loading on the bearings on which the rotor assembly rotates.

The motor may further comprise a Hall effect device arranged in one of the slots defined by the stator core teeth to detect the magnetic fields of the rotor assembly magnets and provide a signal indicative of the rotational position of the rotor assembly with respect to the stator assembly.

The rotor assembly may be external to the stator assembly, with the stator core teeth extending radially outwards. This coupled with the fact that for a given number of rotor poles the inventive motor has fewer stator teeth and hence larger slots, facilitates the location of the Hall effect device to detect, and be switched by the fields of, the rotor magnets. Thus, position detection can be achieved without the need to adapt the rotor or stator or to provide an encoder or an additional magnet array coupled to the rotor.

According to a second aspect of the invention there is provided an electric motor comprising: a stator assembly including multiple phase windings and a stator core having a plurality of radially extending teeth defining a plurality of slots; a rotor assembly comprising a plurality of magnets for interacting with magnetic fields generated by the stator assembly to provide rotation; and a Hall effect device arranged in one of said slots to detect the magnetic field of said rotor assembly magnets and provide a signal indicative of

the rotational position of the rotor assembly with respect to the stator assembly.

The stator core teeth may have flared end portions proximate the rotor magnets, and the Hall effect device may be positioned inside the slot, constrained by the reduced slot entrance width and abutting surfaces of the faired portions of the adjacent teeth.

The teeth may be adapted to constrain and/or locate the Hall effect device closer to the rotor magnets.

The Hall effect device may be a digital Hall effect device and/or may comprise a latch circuit.

The electric motor may be a three phase motor, and may include six Hall effect device arranged in respective slots to provide an indication of the rotor assembly position relative to the stator assembly to within thirty electrical degrees.

Switching signals for commutation circuitry controlling excitation of the three phase windings may be produced by the array of six Hall effect devices for every 30 electrical degrees of rotation.

The electric motor may have thirty-six teeth and thirty-six slots, the six Hall effect devices being arranged in nominal slots 1,4,7,10,13 and 34.

The electric motor may have an external rotor assembly and the stator core may be hollow.

Advantageously, the hollow may be used to house control electronics and/or at least part of a brake assembly.

According to a third aspect of the present invention there is provided drive apparatus for an electrically powered vehicle, the drive apparatus including: a motor comprising a stator assembly and a rotor

assembly, the stator assembly including a ferrous stator core; and a brake assembly biased to inhibit rotation of the rotor assembly, and including brake release means operable to permit substantially free rotation of the rotor assembly, the brake release means comprising a ferrous member and an electromagnet operable to exert an attractive force on the ferrous member, the electromagnet comprising windings and a ferrous yoke, the ferrous yoke including a portion of the stator core.

The drive apparatus may comprise a motor in accordance with other aspects of the present invention.

The portion of the stator core thus decreases the reluctance of the brake electromagnet magnetic circuit and yields an increased attractive force for a given brake windings excitation (i. e. number of ampere turns).

Advantageously, the portion of the stator core may be a portion which is unsaturated by magnetic flux generated by the stator windings when the stator windings are carrying nominal maximum current.

Advantageously, the portion may be stator core iron that is magnetically redundant, in as much as it does not contribute significantly to the stator magnetic circuits generating flux to cause the rotor to rotate. In the case of external rotor motors, the portion may be a portion at the centre of the stator core, e. g. an annular portion at an inner surface bounding a hollow.

A surface of the stator core may be substantially cylindrical, and the portion of the stator core may be a substantially annular portion bounded by the

cylindrical surface.

The cylindrical surface may be the surface of a cavity in the stator core, the surface being spaced a constant distance radially from the nominal axis of rotation of the rotor assembly.

Advantageously, the electro-magnet windings may comprise a solenoid, and the solenoid may be arranged at least partially inside the cavity with its longitudinal axis being substantially co-linear with the axis of rotation.

Utilising magnetically"spare"stator core iron and locating the brake assembly at least partially inside the cavity in the core gives improved brake attractive force for a given excitation, and reduces the overall weight and size (footprint) of the combined motor/brake apparatus.

Preferably, at least part of the brake assembly may be arranged inside a cavity in the stator core.

At least a portion of the electro-magnet windings may be arranged inside the cavity.

Advantageously the brake assembly may comprise a disc or plate coupled to rotate with the rotor assembly, and a member urged into contact with the disc or plate by at least two springs to provide a frictional force to inhibit rotation of the disc or plate.

If one of the springs fails, some braking, albeit reduced, is still applied to the rotor.

The urged member may comprise the ferrous member, and the electro-magnet may be arranged to attract the urged member against the springs.

The disc or plate may have a plurality of splines extending radially outwards for engaging with corresponding slots in the rotor assembly.

By coupling the disc or plate to the rotor with

outwardly extending splines rather than with splines on an internal shaft, the braking torque applied to the rotor assembly can be increased.

According to a fourth aspect of the present invention there is provided drive apparatus for an electrically powered vehicle, the drive apparatus including an electric motor having a stator assembly and a rotor assembly, the rotor assembly including a generally annular portion (member) arranged to encircle at least a portion of the stator assembly, the generally annular portion comprising a plurality of magnets arranged to interact (for interacting) with magnetic fields generated by the stator assembly to provide rotation of the rotor assembly, and the drive apparatus further including a tyre mounted on (supported by) the generally annular portion.

Alternatively, the generally annular portion may be adapted to support a tyre, or may have an outer surface providing a surface on which the vehicle rolls.

The drive apparatus may comprise an electric motor or drive apparatus in accordance with other aspects of the present invention.

The generally annular portion may have a substantially annular groove formed in an outer surface for locating or facilitating mounting of a tyre.

The generally annular portion may comprise a ring of ferrous material and the magnets may be attached to an inner surface of the ring facing the axis of rotation of the rotor assembly. The magnet may be substantially flat rectangular sheets of magnetised material and/or may comprise NeBFe.

Alternatively, the magnets may be magnetised regions formed in a substantially cylindrical member.

According to a fifth aspect of the present invention there is provided a wheel for an electrically powered vehicle, the wheel including a generally annular portion for supporting a tyre, the generally annular portion comprising a plurality of magnets arranged to interact (for interacting) with magnetic fields generated by a stator assembly on the vehicle to rotate the wheel.

The generally annular portion may be adapted to support and/or facilitate mounting of the tyre, or alternatively may have an outer surface providing a surface on which the vehicle rolls.

Electric motors in accordance with the various aspects of the present invention may include rotor assemblies which are arranged to rotate on two bearings, the bearings sealing a volume between the rotor assembly and stator assembly from the motor's environment.

Embodiments of the present invention will now be described with reference to the accompanying drawings in which: Fig. 1 is a schematic diagram of a known stator and rotor assembly configuration, suitable for use in embodiments of the present invention; Fig. 2 is a schematic diagram of a known arrangement of phase coils and commutation circuitry suitable for use in embodiments of the present invention ; Fig. 3 is a schematic representation of a known commutation method suitable for use in embodiments of the present invention; Fig. 4 shows a diagram of a known digital Hall effect sensor and its switching characteristics, the

device being suitable for incorporation in embodiments of the present invention; Fig. 5 shows the additional commutation steps for a known twelve step commutation method suitable for use in embodiments of the present invention; Figs. 6 and 7 show a known arrangement for detecting rotor position with respect to the stator; Fig 8. is a schematic diagram of an electric vehicle in accordance with the prior art and embodiments of the present invention; Fig. 9 is a schematic diagram of a know electric motor and wheel for an electric wheelchair, the combination being suitable for incorporation in embodiments of the present invention; Fig. 10 shows schematically a known brake arrangement suitable for incorporation in embodiments of the present invention; Fig. 11 shows schematically the arrangement of the rotor assembly, and stator assembly including stator windings in a motor in accordance with an embodiment of the present invention; Fig. 12 shows a schematic cross section of a stator assembly in accordance with an embodiment; Fig. 13 shows a schematic cross section of the rotor assembly and stator assembly of the embodiment of Fig. 12; Fig. 14 is a schematic diagram of part of the cross section of an embodiment of the present invention; Fig. 15 is a schematic perspective view of part of the embodiment of Figs. 14 (b) and (c); Fig. 16 is a schematic cross section of part of a further embodiment; Fig. 17 is a schematic representation of an embodiment of the present invention;

Fig. 18 is a perspective view of a wheel in accordance with an embodiment of the present invention; Fig. 19 is a schematic diagram of a rotor assembly suitable for use in embodiments of the present invention; Fig. 20 is a schematic diagram of part of a rotor assembly suitable for use in embodiments; Fig. 21 is a schematic diagram of part of the rotor assembly of an embodiment of the present invention; Fig. 22 is a perspective view of part of a wheel or rotor assembly in accordance with an embodiments of the present invention; Fig. 23 shows cross sections of various drive apparatus in accordance with embodiments of the present invention; Fig. 24 is a schematic diagram of an embodiment; Fig. 25 is a schematic diagram of drive apparatus in accordance with an embodiment of the present invention.

Fig. 26 is a schematic diagram of the cross section and perspective view of drive apparatus in accordance with a further embodiment; Figs. 27 and 28 are schematic cross sections of the stator cores of embodiments of the present invention; Fig. 29 is a schematic cross section of drive apparatus in accordance with an embodiment; Fig. 30 is a schematic cross section of a further embodiment; Fig. 31 is a schematic diagram of the inventive brake apparatus suitable for use with the embodiment shown in Fig. 30; and Fig. 32 is a diagram of the front face of the

front plate of the drive apparatus shown in Fig. 30.

Referring now to Fig. 11, in this embodiment of the present invention, the stator core 1 has 12 teeth 11 and the rotor assembly 2 has 10 magnets (i. e. poles), both the teeth and poles being distributed uniformly around the rotational axis, albeit at different angular frequencies. As with the motor shown in Fig. 1, although the rotor assembly and stator assembly are generally cylindrical, Fig. 9 shows the motor linearly for convenience.

The fact that the number of rotor magnets is two less than the number of stator teeth results in a "Vernier effect"or beating, such that the leading edges of the rotor magnets align with edges of the stator teeth at a maximum of two positions P. This reduces detent torque significantly.

The stator windings are three phase, and each stator tooth 11 is wound with a coil from just one of the phases. The coils do not overlap. Diametrically opposed teeth (e. g. tooth 1 and tooth 7), are wound with coils corresponding to the same phase to even the magnetic pull on the rotor and so even the load on the rotor bearings.

The rotor assembly 2 of the inventive motor may be arranged either internally or external to the stator assembly 1.

The windings corresponding to phase C are shown in Fig. 11 (b). These windings essentially comprise a string of four coils surrounding teeth 5,6,11 and 12 respectively. As can be seen from the figure, when current flows in phase C windings it circles adjacent teeth (e. g. teeth 5 and 6) in opposite directions, providing the necessary field distribution for interaction with the rotor magnets 22 whose polarities also alternate around the rotor.

Moving on to Fig. 12, in this embodiment the stator core comprises 36 teeth extending radially outward from a central iron hub 13. The centre of the stator core is hollow, and the stator teeth 11 flare at their ends. The stator windings are three-phase, connected in star configuration. The star point is labelled on the figure, as are the start points of each phase winding. The stator core teeth are individually wound with a single respective phase and there are 12 teeth wound with each phase (i. e. one- third of the total number). These 12 teeth are divided into two diametrically opposed sets of six to ensure that the load on the rotor bearings is balanced.

The slots 12 between the teeth 11 are numbered arbitarilly from 1 to 36.

The flaring of the teeth is a known desirable feature, and helps to ensure that the tips of the teeth do not saturate. The flared teeth help to decreased the overall reluctance of the magnetic circuit comprising adjacent teeth, the stator hub 13, the rotor magnets and the rotor back iron, and so increase the efficiency of the motor.

In Fig. 13, the stator core is shown without the stator windings, but with the rotor assembly in place.

Although the stator assembly has 36 teeth, the maximum number of the 34 rotor magnets that can align exactly with the stator teeth at any one time is just 2.

Detent torque is therefore dramatically reduced.

The motor exhibits extremely smooth low speed operation and is particularly suited to application as an in-wheel motor for an electrically powered vehicle such as an electric wheelchair.

The external rotor configuration enables the torque producing radius (i. e. the radius of the air gap between the rotor magnets and stator teeth) to be

made as large as possible to provide maximum torque for a given stator winding excitation.

The stator core is hollow to reduce the overall weight of the motor, and this hollow space can advantageously be used to house associated circuitry.

In further embodiments, this space can be used to house, in part, a brake assembly.

Fig. 14 is a schematic diagram of part of the cross section of the embodiment shown in Figs. 12 and 13. The figure shows two adjacent teeth 11 having flared end portions 111 acting to reduce the width of the entrance to the slot 12. Windings corresponding to phase A are shown substantially filling the slot 12. The windings, although comprising insulated conductors, are further insulated from the stator core by lining material W. The lining material is wrapped over the top of the windings and serves to bias a digital Hall device 61 against surfaces of the flared portions 111. The width of the Hall device is greater than the width of the slot entrance. The active sensing element 611 of the Hall effect device 61 is sensitive to magnetic fields in the direction labelled n in the figure and so is insensitive to magnetic flux generated by the currents in the phase A windings. The magnetic flux generated by the phase windings in the centre of the slot has no radial component. Thus, the Hall device is able to sense or switch in response to the magnetic fields of the rotor magnets and so provide an indication of the rotor position with respect to the stator.

The flux distribution shown in Fig. 14 is highly schematic. The back iron 21 of the rotor assembly is the form of a steel hoop and the rotor assembly magnets 22 are flat sheets of magnetic material bonded to the hoop.

In further embodiments, the stator teeth may be adapted to enable location of the Hall effect sensor closer to the rotor magnets. This adaption may take the form of modifications to the flared portions of the teeth.

Conveniently the stator core may be formed from a stack of laminations to reduce eddy current losses.

In such an arrangement, a number of the laminations may be adapted to enable closer location of the Hall effect sensors. In one such arrangement, the tips of the flared teeth are removed using a punch of suitable diameter D as shown in Fig. 12 (b). The diameter D is chosen so that the Hall effect sensor sits as close as possible to the rotor magnets, but is still constrained by the teeth tips (as shown in Fig.

14 (c)).

The embodiment shown in Figs. 12,13 and 14 (a) does in fact include 6 Hall effect devices, each located as shown in fig. 14 (a), and positioned in nominal slots 1,4,7,10,13 and 34. This arrangement enables a position signal to be generated by the Hall effect device array which gives the position of the rotor with respect to the stator to within 30 electrical degrees. This signal can be used by commutation circuitry and advantageously the commutation may involve 12 steps.

Thus, with the combination of the reduced detent torque, the large number of poles, and 12 step commutation, this embodiment can provide especially smooth rotation even at low speeds.

One electric motor embodying the present invention has the configuration shown in Fig. 11, with a conventional internal rotor 2 and shaft output. The motor has a 12 slot stator and 10 pole rotor. Using

commercial available design software a balance was achieved between the torque producing radius (rotor OD (outer diameter)), the magnets thickness and the winding area (a function of slot depth and tooth width).

In this example, the stator (12 slot) has a three-phase winding star connected wound as three coil strings, each string wound directly into the slots and configured as two plus two i. e. two coils diametrically opposite two coils of the same phase to produce a magnetic balance for each phase, equalising magnetic pull, minimising noise and bearing load. The second and third phases repeat phase one but displaced such that phase 1 uses slots 1,2 and 7,8. Phase two uses slots 3,4 and 9,10. Phase 3 uses slots 5,6 and 11,12. When one phase or two phases are energised the 10 pole rotor magnets try to align. Because they are not the same pitch as the stator, only the two diametrically opposite poles truly align with the relevant stator teeth. Towards the end of the active arc (i. e. the extent of the energised coils) the magnets become more out of line. Where they are completely out of line the stator teeth are not excited (the windings around these teeth correspond to the unused phase). In practice, this arrangement gives close to the conventional"aligned magnet"performance (i. e. the performance achieved with conventional motors in which the number of stator teeth = 3 x the number of rotor poles) but with very little cogging.

The back EMF is near sinusoidal and with conventional six step commutation the torque ripple is expected to be near 13%.

Thus, in the example shown in Fig. 14 (b) the nominal width of the entrance to the slot 12 defined

by the teeth 11 (and in particular the flared end portions 111) is G. In this example G = 3mm. The stator core 1 is constructed from a number of laminations, i. e. it is formed from a stack of substantially similar laminations whose teeth have the form shown in Fig. 14 (b)). In order to enable the Hall to be positioned closer to the rotor magnets, a number of the laminations are adapted. This adaption takes the form of removing one millimetre from each tooth tip using a punch of diameter D, where D = 5mm in this example, to yield a modified slot entrance width of between 3 and 5mm.

Fig. 14 (c) shows the Hall effect sensor 61 seated in the modified slot entrance, still constrained by the end portions of the teeth. The sensor may be held in this position partly by the resilience of the windings and/or the folded insulation material W beneath the sensor in the slot 12.

Figure 15 shows a schematic perspective view of a stator core 1 formed from a stack of laminations. A portion P of the laminations have been adapted as shown in figs. 14 (b) and (c), and this portion P is situated at one end of the stack. This arrangement enables the Hall effect device 61 to be inserted into the modified slot after the windings have been completed, from the end of the stack.

In another embodiment of the present invention, as shown in Fig. 16, the rotor assembly is internal to the stator assembly and the stator teeth 11 project radially inwards. In this example, the stator teeth 11 do not flair, but are adapted to locate and hold the Hall effect device 61 at the"top"of the slot 12 above the windings. The adaption takes the form of slots or grooves 121 machined near the tips of the

teeth 11.

Fig. 17 shows another embodiment of the present invention. The figure shows drive apparatus for an electrically powered vehicle, the drive apparatus having an internal stator assembly 1 and a so-called external rotor, which rotates with respect to the stator assembly 1 about an axis R. The rotor includes a generally annular portion comprising a ring 21 to which a plurality of permanent magnets 22 are attached. The permanent magnets 22 are uniformly spaced around the inner surface of the ring 21 to interact with magnetic fields produced by the stator assembly 1 to produce rotation.

The air gap g between the stator assembly 1 and the rotor magnet 22 is made as small as possible to maximise torque.

The ring 21 is arranged to have a minimum thickness, determined by strength and rigidity criteria, and a tyre 51 is mounted directly on the ring 21 of the generally annular portion. In effect, the tyre is mounted directly onto the motor rotor.

This provides numerous advantages. For example, it enables the torque producing radius r (the inner radius of the rotor magnet ring) to be made as large as possible, enabling maximum torque to be produced for a given magnetic flux density in the air gap g.

Also, by using the rotor assembly as the wheel for supporting the tyre 51 the number of components in the drive apparatus can be reduced, and its weight reduced also.

Fig. 18 shows a perspective view of the wheel and tyre of the example shown in Fig. 17. There are in fact 12 permanent magnets 22 attached to the inner surface of the ring 21, with alternating polarity.

This wheel is designed for operation with a stator

assembly having a 36 tooth core and conventional three phase lapped windings. In other examples, the number of rotor magnets may be chosen for interaction with one of the inventive stator configurations, such that the number of poles and stator teeth differ by 2 to give reduced magnetic cogging.

Fig. 17 does not show all of the components of the rotor assembly. Clearly, some support means and bearings are required to constrain the annular member including the rotor magnets to rotate about the axis R. Some of the rotor assembly may protrude into the stator assembly. However, this does not affect the fact that this type of configuration is generally referred to as an external rotor arrangement, as the ring of rotor magnets encircles part of the stator assembly.

In the example shown in Figs. 17 and 18, the rotor magnets 22 are in the form of discs of magnetic material. The ring 21 is formed of steel, and provides both support for the tyre 51 and"back iron"for the rotor. In other embodiments, the ring 21 may serve a purely mechanical function, and may be formed from non-ferrous material. The material may be chosen to reduce the weight of the vehicle incorporating the drive apparatus, and in particular may enable significant reductions in wheel mass and moment of inertia.

Fig. 19 shows schematic perspective and cross sectional views of the annular portion of a rotor assembly suitable for use in the embodiments of the present invention. The annular portion includes a steel cylinder 21 having an inner surface (i. e. facing the nominal axis of rotation) to which a plurality of permanent magnets 22 are attached. The poles of the permanent magnets are the broad faces of the strips,

and the magnets are arranged with alternating polarity around the cylinder 21. A cross section of part of the annular member of Fig. 19 is shown in Fig. 20 (a). In this figure, just two of the strip magnets 22 are shown, and their shape is shown in schematic perspective view in Fig. 20 (b). The cross sections of the magnets 22 are curved with the radius of curvature matching that of the cylinder 21. Thus, in this example the rotor assembly comprises 2 different sets of permanent magnets, both having the same geometry but one set having its north pole on the outer broad face and the other set having the north pole on the inner broad face. These types of magnets are known as"tile"magnets.

Fig. 21 shows part of the cross section of a rotor assembly annular portion suitable for use in embodiments of the present invention. In this example, rather than using curved tile magnets, the rotor magnets 22 are flat rectangular sheets. In the past, it was thought that the use of such magnets would distort the flux pattern and lead to back EMF wave forms which resulted in unacceptable torque ripple.

However, the present inventors have determined that flat magnets may be used with more than acceptable results. The use of flat magnets provides the advantage that only one type of magnet is required, as shown in Fig. 21 (b). This reduces costs and simplifies manufacture. Rather than needing two polarities of magnets, the flat magnets can simply be reversed as appropriate to give the desired alternating field pattern around the ring. In this example the magnets 22 are bonded to the inner surface of the cylinder 21.

In small diameter motors, the air gap g may have to be increased to accommodate the flat rotor magnets.

However as the motor diameter increases the flat

magnets become better approximations to the tile magnets, and no adaption of the stator teeth may be necessary. The flat rotor magnets are particularly suitable for motors in accordance with embodiments of the present invention in which the rotor is external, and directly supports a tyre. In these examples, the diameter of the rotor magnet ring is as large as possible.

In the examples shown in Figs. 17-21, the rotor magnets were attached to a surface of the cylinder or ring 21. Fig. 22 shows an alternative arrangement, in which the ring or cylinder 21 is formed from magnetizable material, and the"magnets"are magnetised regions formed in the body of the ring 21.

The magnetised regions are shown shaded in the figure, and as before their polarities alternate around the ring 21. This arrangement can provide extreme mechanical reliability, as the magnets are formed in the ring itself and are not separate components which must be attached.

Fig. 23 shows four other embodiments of the present invention. In the example of Fig. 23 (a), a wheel rotates around a stator assembly 1 about a rotational axis R. The wheel includes a steel ring 21 and a plurality of magnets 22 attached to the inner surface of the ring. On an outer surface of the ring 21 a tyre 51 is mounted.

Fig. 23 (b) shows the cross section of part of a rotor assembly surrounding a stator assembly 1. The rotor assembly 2 includes a generally annular portion including a steel ring 21 and a plurality of permanent magnets 22. The magnets 22 are let into recesses in the steel ring 21 so that they are flush with its inner surface. The steel ring 21 is adapted to seat and support a tyre. The tyre rests in contact with an

outer surface 213 of the steel ring 21, the edges of the ring forming rims. Letting the magnets into the ring enables the torque producing radius to be further increased, and the air gap between the magnets and the stator assembly is made as small as possible. The stator assembly core is hollow, defining a cylindrical void 101 centred on the rotational axis R.

Fig. 23 (c) shows an alternative embodiment in which the rotor assembly annular portion is formed of magnetizable material, and the rotor magnets 22 are magnetised regions in the annular portion. The annular portion includes formations 211 at its axial ends which provide rims for mounting a tyre on the annular portion.

Fig. 23 (d) shows another embodiment with an external rotor assembly 2 and an internal stator assembly 1. Permanent magnets 22 in the form of flexible strips of magnetic material are attached to the inner surface of a generally annular member 21.

The magnets 22 and member 21 form the general annular portion of the rotor assembly 2. The inner surface of the member 21 is substantially spherical and an outer surface 212 provides a surface on which the vehicle rolls (i. e. it is typically in contact with the ground). The outer surface 212 is adapted to improve traction. It may have a tread formed on it, or be adapted in other suitable ways.

Fig. 24 is a highly schematic diagram of part of drive apparatus embodying one aspect of the present invention. The drive apparatus includes a brake assembly which is biased to inhibit vehicle motion, and the brake is electrically released. Brake release is achieved by means of applying an attractive force to a ferrous plate 81 using an electromagnet. The electromagnet includes windings 83 and an iron yoke.

The yoke comprises an iron member 841 and a portion 131 of the stator core 13,11.

When the windings 83 are excited, magnetic flux is generated around the magnetic circuit (mc) which includes the portion 131 the iron member 841, at least part of the ferrous plate 81 and the two air gaps, and an attractive force FA is developed between the ferrous plate 81 and yoke.

The attractive force on the plate 81 is determined by the number of ampere turns provided by the excited windings and the reluctance of the magnetic circuit.

The excited windings 83 and iron member 841 in isolation (i. e. positioned away from the stator core) would of course still be able to apply an attracted force to the plate 81. However, by using the portion 131 of the stator core to increase the amount of ferrous material in the yoke and decrease the reluctance of the magnetic circuit, the flux density in the air gaps, and hence the attractive force on the plate, for a given excitation is increased.

The brake is biased to slow the vehicle by means of a spring, which provides a braking force F3. The brake apparatus is arranged so that the attractive force FA opposes the braking force FB. In order to release the brake, it is necessary to provide sufficient ampere turns to the magnetic circuit so that an attractive force equal to the braking force is applied to the plate. In practice, the excitation is arranged to produce an attractive force greater than the braking force, and plate 81 moves slightly towards the yoke. Thus, the air gaps are reduced, and the reluctance of the magnetic circuit is decreased. Thus, the number of ampere turns required to maintain the attractive force at a level equal to or greater than

the braking force is reduced. In general, the excitation required to release the brake initially is greater than the excitation required to keep the brake "off".

As mentioned above, utilising a portion 131 of the stator core as yoke material decreases the reluctance of the brake magnetic circuit and so reduces the number of ampere turns required both to release and hold the brake in the off state.

This aspect of the present invention therefore provides the advantage that fewer turn windings can be used to provide the same attractive force as previous arrangements, reducing cost and weight.

Alternatively, for the same number of turns, the winding current required to release and hold the brake can be reduced, thereby reducing power consumption.

This is particularly advantageous for battery powered vehicles.

Also, by utilising core iron in the brake magnetic circuit a stronger attractive force can be achieved for a given number of ampere turns. Thus, stronger springs can be used to bias the brake to hold the vehicle, thus increasing safety, especially on steep slopes.

Also, if a certain amount of iron is required in the yoke of the brake electromagnet to produce a particular attractive force with a given excitation, then by utilising a portion 131 of the core iron the total amount, and hence weight, of iron in the combined motor and brake assembly is reduced.

Advantageously, the portion 131 of core iron used in the brake yoke may be iron that is"under-used"in the stator magnetic-field-generating magnetic circuit, i. e. iron that is far from being saturated when the stator windings are carrying nominal maximum current.

The portion 131 may in fact be"spare iron"that makes a negligible contribution to the generation of stator fields to produce rotation, and which is present in the core for other design reasons (e. g. to facilitate attachment to other components or to the vehicle itself).

By using"spare"or under-used stator core iron in the brake magnetic circuit, the amount of additional iron 841 required to enable desired brake disengagement characteristics to be achieved is reduced, and hence so is the overall weight of the vehicle.

In the case of external rotor motors with hollow cores, the stator iron at the centre, i. e. forming an annular portion adjacent to the inner cylindrical surface of the hollow, is generally magnetically under-used (understressed), and has the lowest flux density of all parts of the stator core for a given stator winding excitation.

In such motors, it is desirable to use this inner ring of core material as material shared with the brake yoke. An example of such an arrangement is shown in Fig. 25.

Fig. 25 (a) shows a schematic cross section of the rotor assembly 2, stator assembly 1 and brake assembly. The brake electro-magnet includes windings 83 in the form of a solenoid, wound on a"top hat" shape iron member 841. The iron member 841 provides the inner and back iron of the brake electromagnet yoke. The"outer"yoke iron is provided by a portion 131 of the stator core. This portion is generally annular and adjacent the inner surface of the cylindrical hollow 101 in the core. The shared portion 131 is an innermost portion of the core back iron 13.

When the solenoid windings 83 are excited,

magnetic flux F is generated in the magnetic circuit including part of a ferrous plate in the form of a disc. An attractive force FA is therefore developed between the disc 81 and brake yoke.

Fig. 25 (b) shows a schematic front view of the stator assembly and brake electro-magnet. In this figure flux F generated by the stator windings is shown highly schematically around magnetic circuits including the core teeth 11 and part of the core back iron 13.

The rotation-producing flux pattern is such that even when the stator windings are carrying maximum current, the flux density at the"hub"of the stator core is low. Thus, the hub material 131 can be utilised in the brake yoke without impairing motor performance.

Fig. 27 is a schematic cross sectional view of the stator core and brake electromagnet of drive apparatus in accordance with another embodiment. In this example, the thickness t of the core back iron 13 is more than sufficient to ensure that the back iron does not saturate when the stator windings are carrying full current. The back iron 13 is thicker than magnetically necessary so that holes H can be provided to enable the stator assembly to be bolted to the vehicle. In effect, a portion 131 of the core iron is magnetically spare and this spare portion forms part of the yoke of the electro magnet, which further comprises a solenoid 83 and ferrous member 841.

Fig. 26 shows schematic cross section and perspective views of another embodiment of the present invention. In this drive apparatus, the motor has an internal rotor assembly 2 inside an external stator 1.

The stator assembly includes a stator core having

inwardly projecting teeth 11 from an outer ring of back iron 13. An outermost portion 131 of this back iron 13 forms part of the yoke for a brake electromagnet, which further comprises a solenoid 83 and ferrous member 841. The brake electro magnet is arranged to attract a ferrous ring 81 to disengage a brake on the wheel.

Fig. 28 shows a schematic cross section of the stator core of a similar embodiment. The stator core is formed from a stack of laminations. The thickness t of the core back iron 13 is thicker than magnetically necessary to enable holes H to be provided for tie bars which hold the stack of laminations together. An outer portion 131 of the back iron 13 is, therefore, magnetically under-stressed and advantageously is used as yoke material for an electro magnet arranged to disengage the brake.

Fig. 29 shows drive apparatus in accordance with an embodiment of the present invention. The drive apparatus is, in fact, a combined electric motor, wheel and brake for an electrically powered vehicle.

The motor has an internal stator assembly 1 and an external rotor assembly 2. The stator assembly 1 includes a stator core which comprises a plurality of teeth 11 extending radially outwards from the stator core back iron 13 towards the rotor assembly permanent magnets 22. The motor has a 34 pole rotor and a 36 pole stator such that only two of the edges of the rotor magnets and the stator slots align at any one time, thereby minimising cogging without skewing the stator teeth. The windings are arranged into 12 teeth per phase (the windings are three phase) split into 6 teeth either side of the stator to keep the radial forces symmetrical. The three phase windings comprise three strings of 12 coils. Each coil spans 1 tooth.

With two phases on (line to line excitation when connected in star configuration) the mismatch of the rotor and stator teeth, because of the dissimilar number of teeth on each, lie in the unexcited stator area.

The stator and rotor assemblies are in fact the same as those in the embodiments shown in Figs. 12 and 13.

The stator core is hollow and part of the brake assembly protrudes into this space. The brake assembly includes brake energising coils 83 inside an iron member 841. Holes H are formed in the back iron 13 of the stator core, and via these holes the stator assembly is bolted to the brake assembly iron member 841. A portion 131 of the stator core iron adjacent the inner surface of the hollow forms part of the iron yoke of the brake electro magnet. Thus, the portion 131 contributes to reducing the reluctance of the brake apparatus magnetic circuit.

The brake energising coils 83 protrude into the hollow at the centre of the stator assembly.

The brake assembly further includes a brake plate 811 which is fixed with respect to the iron member 841 and hence the stator assembly. A ferrous brake actuator plate 81 is urged by 6 actuator springs 82 to apply a braking force to a brake disc 812 via friction plates 89. Thus, the brake disc 812 is trapped beneath the fixed brake plate 811 and the actuator plate 81 and ordinarily, in the absence of current in the brake coils, the springs 82 bias the brake assembly to provide a frictional force which inhibits rotation of the brake disc 812 with respect to the stator assembly. Using a number of actuator springs provides the advantage that in the event of failure of one spring a braking force, albeit reduced, is still

applied to the brake disc. The brake disc 812 is coupled to rotate with the rotor assembly by means of a splined stubshaft connected to the rotor assembly end plate 29.

With no power supply to the brake coils, rotation of the rotor assembly is inhibited. When the brake coils 83 are energised, however, an attractive force is generated between the actuator plate 81 and the brake electro magnets which act to oppose the springs 82. Thus, by exciting the brake coils, the frictional force between the friction plates and brake disc 812 can be reduced to zero (or substantially zero) and hence free rotation of the rotor assembly is permitted.

In this example, both the portion 131 of the stator core and the steel bolts used to attach the brake iron member 841 to the stator assembly contribute to the brake electro magnet yoke material and so decrease the reluctance of the magnetic circuit and increase the attractive force applied to the actuator plate 81 for a given excitation.

In one embodiment, the brake assembly is arranged such that without the contribution to the brake yoke from the stator core, the maximum force applicable to the actuator plate 81 is insufficient to overcome the force provided by the plurality of springs. Thus, for the brake assembly to function, it must be attached to the stator assembly.

In the embodiment of Fig. 29, the brake assembly and stator assembly are attached to a stator hub 18 which in turn is adapted for attachment to the electric vehicle. An inner bearing 19i is seated on a surface of the stator hub 18 and an outer bearing 19o is seated on a surface of the brake iron member 841.

The rotor assembly 2 rotates on these two bearings

only, and the bearings seal a volume V between the stator assembly and rotor assembly from the outside environment.

The rotor assembly 2 includes a generally annular member which comprises a steel ring 21 to which a plurality of permanent rotor magnets 22 are attached.

The magnets 22 are let into the inner surface of the steel ring 21. An outer surface of the steel ring includes a groove G which in this example is v-shaped.

A tyre 51 is mounted on the steel ring 21 and sits (i. e. is located by) on the groove.

An inner end plate 291 is attached to one axial end of the steel ring and provides an inner tyre rim 211i. The inner end plate 291 includes a seating for the inner bearing 19i.

An outer end plate 29 is attached to the opposite axial end of the steel ring 21 and provides an outer rim 211o. The outer end plate is attached to the splined subshaft and so is coupled to the brake disc 812. The splined subshaft is seated on the outer bearing 19o.

The drive apparatus further includes two brake release bolts which can be used to release the brake manually, i. e. in the absence of power supply to the brake coils 83. This enables the vehicle to be manually manoeuvred without power supply.

Fig. 30 shows another example of drive apparatus embodying the present invention. The drive apparatus includes a brushless in-wheel motor. The motor is a brushless synchronous AC motor that has been configured to give high torque at relatively low speeds and with low torque ripple. This is achieved in part by having a multipole design to keep the electrical frequency relatively high at low speeds for averaging of any torque ripple effect.

The original application of the brushless in- wheel motor system was as a direct replacement for brushed gear motors in powered wheelchairs.

Significant advantages were seen in using a direct drive brushless system in that the elimination of gearbox and brushes would give quieter more efficient performance, extreme reliability and simpler construction, if the low speed performance of the gear motor could be matched. The main concern in this respect was the torque ripple and detent torque.

Detent torque is caused by the tendency of rotor magnet edges to align with stator tooth edges. The motor design of the embodiment in Fig. 30 was optimised very successfully to almost eliminate detent torque using the Vernier effect to reduce the number of aligned edges at any angle to a minimum (2). The motor in Fig. 28 includes a 36 tooth stator and a 34 pole rotor.

The employment of the Vernier effect has other important implications for the motor. It was now possible to produce a motor with high magnet pole count (34) suitable for low speed operation having a number of stator tooth only two greater than the number of rotor magnets. (Conventional lapped winding design would require the number of stator teeth to be three times the number of magnets). Also, the stator teeth require no skew to reduce detent torque as with conventional lapped winding design.

Also, the windings of the motor shown in Fig. 30 span one tooth each, reducing end winding length and overhang to a minimum. This gives reduced winding resistance and a more compact and easier to produce design.

The inventive 36/34 motor has a sinusoidal back EMF which can be shown to precipitate a 13% ripple

with conventional brushless dc drive systems (e. g. six step).

A brushless ac drive system with the inventive motor can give the desired performance, virtually eliminating torque ripple by matching the motors back EMF characteristic. Usually this would require complicated, expensive and space consuming position sensors, which would compromise the motor design. This would require analogue Hall devices (analogue Hall devices require either a gap in the stator to expose a pair of magnets, or an auxiliary commutation magnet), or an encoder. Furthermore, a conventional brushless ac system would require considerable microprocessor power.

In order to reduce torque ripple to an acceptable level however, it is possible to use the twelve step commutation method described above. In the embodiment of Fig. 30 this is achieved, and the smooth operation of a brushless ac system is approached, using digital Hall devices usually associated with a brushless dc system, avoiding compromise of the motor design and without the use of conventional high resolution position sensors. Six digital Hall effect sensors are positioned between the stator teeth such that they produce a Grey code giving position to within 30 electrical degrees. The Hall effect devices are located in slots as shown in Fig. 12, Fig. 13, Fig.

14 (c) and Fig. 15.

The stator core of the embodiment of Figure 30 is hollow and an important advantage of this motor is that drive electronics can be integrated into the void inside the stator. The drive apparatus can therefore be manufactured in the form of a compact unit and a plurality of units could be linked in modular form.

The prime function of the drive apparatus shown

in Fig. 30 is to propel the electric vehicle, which in this case is an electric wheelchair, forwards and reverse, and by differential driving of two such pieces of apparatus, to provide steering. The drive apparatus also supplies the main, and normally the sole dynamic braking for the wheelchair.

The motor case (i. e. the rotor assembly in this external rotor motor) forms the wheel and one tyre rim (the inner rim 211i). The outer rim 211o is provided by a decorative outer front plate 292. The outer face of this decorative plate 292 is shown in Fig. 30.

The rotor assembly 2 (i. e. the motor case) comprises a generally annular portion, including permanent magnets 22 and a steel ring 21, an inner plate (main casting) 291 and a front plate 29. The inner plate 291 and front plate 29 include seatings for inner and outer bearings 19i, 19o respectively and the rotor assembly rotates around the stator assembly on these two bearings. The inner and outer bearings seal a volume V between the stator assembly and rotor assembly from the outside environment.

The motor bolts directly to the wheelchair chassis by means of threaded holes in an extension boss 18.

There is an electrically and manually released spring applied, fail safe brake integrated within the drive apparatus that is normally applied only when the chair is at rest or the power is turned off.

On dynamic braking, the motor regenerates power back to the batteries of the vehicle.

With regard to safety aspects, the motor has long thermal time constants that help to even temperature changes minimising temperature rise during short term high loads, i. e. hill climbing. Combined with a high efficiency surface temperatures are kept low. Motor

faults such as open/shorted windings or commutation errors are detected by the controller, and in the event of a fault the chair is brought to a stop. In addition, the design of the motor prohibits run away in the event of a controller fault.

Should a battery lead or brake become open circuit, the brakes are applied and the drive removed.

A smooth case design without spokes adds to the safety.

The brake is applied with several springs 82, ensuring some, albeit reduced, braking in the event of a spring failure.

The motors are designed to propel the chair smoothly and quietly, without a gearbox, and its associated noise and leakage problems, to a maximum of 7 mph. Braking is primarily by controlled regeneration by the motor until the chair is stationary when the fail safe brake is de-energised and applies. In this way the chair is brought to a controlled stop without skidding on a good dry surface, and mechanically brakes when stationary. The brake can be hand released for manoeuvring the chair without power.

The design of the tyre rim is such that a standard 3.00 x 8 tyre of 14 inch diameter or a low profile tyre of 12 inch diameter can be fitted, and can be either pneumatic or solid (foam filled). Tyres can be changed without removing the wheel.

With the tyre deflated the decorative front motor plate 292 can be removed for brake adjustment and cleaning. There is no need to remove the wheel from the chair.

The friction plate 89 can be replaced on removal of the inner front plate 29 which also acts as a carrier for the outer bearing.

The whole wheel is designed to IP 54 and has

smooth surfaces for easy cleaning.

The rotor assembly 2 comprises a mild steel tube 21 carrying NeBFe magnets on its inside, and is secured into an aluminium/magnesium"main casting"291 which forms the inner tyre rim and inner bearing outer housing. The front plate 29, also cast aluminium/magnesium, attaches to the outer axial end of the steel tube 21, forming bearing housing for the outer bearing and enclosing the motor and brake assembly.

This front plate 29 also provides engagement for the brake friction disc 812 (or discs), by means of recesses or slots 298 arranged to receive splines projecting outwardly from the brake disc 812. The outer removable decorative front plate 292 provides the outer rim 211o and cosmetic features. The bearings 19 provide free running of the wheel around the motor stator and brake, and the inner bearing and outer cosmetic plate complete the environmental seal.

The brake is fail-safe and is bolted to the motor stator and via an extension boss 18 through the inner bearings, the boss providing means to bolt the drive apparatus to the chair frame. The brake also supports the front (outer) bearing. The brake consists of a friction disc 812 or discs trapped between a spring loaded plate 81 and a second fixed plate 811, both held from rotation relative to the motor stator and therefore the chair.

Application of an electric current to the brake coil 83 attract the spring loaded plate 81 releasing the friction disc 812. The friction disc 812 is free to move axially over a limited range but outwardly splined to the rotor, thus enabling the wheel to rotate freely.

There is provision for a manual release operating

on the sprung loaded plate 81 accessed through holes H1 in the mounting boss 18 and brake electro magnet iron member 841.

The brake electro magnet windings 83 protrude into the cylindrical space inside the stator assembly and the yoke of the brake electro magnet comprises the ferrous member 841 and an inner portion 131 of the stator core. The bolts BT securing the stator assembly to the brake iron member 841 also contribute to the iron in the brake electro magnet magnetic circuit and so decrease its reluctance.

Thus, the portion 131 of the stator core provides additional iron for the part of the brake electro magnet yoke lying radially outside the coil 83 and increases the attractive force on the actuator plate 81 for a given coil excitation.

The torque producing radius of the motor is made as large as possible by mounting the rotor magnets 22 directly on the member 21 carrying the tyre and the stator assembly is made hollow to reduce weight.

There is sufficient radial depth of core back iron 13 to ensure that the iron in the magnetic circuits around the teeth does not saturate when the windings are carrying maximum current.

By integrating the brake assembly at least partially into the stator assembly void, the drive apparatus is made more compact.

The stator assembly includes a laminated steel core and has a three phase winding such that proper excitation of the phases causes rotation of the rotor.

There are 12 stator teeth per phase and three phases.

The windings are close wound to the stator core for good thermal conductivity and low resistance and varnish impregnated for further temperature uniformity and mechanical security.

Each phase is split into two, the first six coils diametrically opposite to the second six. In this way. equal magnetic pull is achieved to avoid bearing side load and mechanical noise. This arrangement also allows the rotor to have two magnetic poles less that the stator, creating a Vernier effect for low magnetic cogging.

The motor commutation is achieved with Hall effect switches. These are situated between selected stator teeth such that the rotor magnets cause them to switch. Suitable arrangements are shown in Figs.

12,13,14 and 15. The rotor position is thereby signalled to the controller, and the supply can be sequenced to the windings. The Hall sensors are semiconductor switches operated by magnetic flux. With the sensors positioned as shown in Figs, 12 and 13, the twelve step commutation method can be applied to give a particularly smooth and quiet operation of the motors.

A mixed pin"D"connector is used for the motor, situated at the centre of the mounting boss 18 to be at the motor axis and high enough to avoid submergence in up to 150mm deep puddles. Three motor leads use 40 amp pins, the Hall sensors use 8 low current pins and the brake 2 pins. The connector is fixed to the rotor rear boss and sealed in with a silicone sealant.

Advantages provided by the drive apparatus shown in Fig. 30 include: 1. The apparatus is a direct drive system. There is no gear box or associated problems.

2. There is space for the safety brake and/or control electronics within the wheel, and hence the size of the apparatus can be reduced.

3. The design provides smooth and continuous rotation at speeds down to 1 RPM.

4. The drive apparatus exhibits very low torque ripple.

5. The apparatus provides quiet operation for indoor use.

6. The design gives maximum motor power density and flexibility. It uses digital Hall effect devices as commutation position sensors inserted into the existing stator tooth gaps. This avoids the need to accommodate analogue Hall devices, which require a separate magnet ring or a gap in the stator, or an expensive absolute encoder, which would occupy space at the centre of the stator ring reserved for the safety brake.

7. The apparatus provides reduced cost.

Processor loading is reduced when compared with systems using other feedback devices (e. g. an encoder or analogue Hall devices). This allows a cheaper processor to be used.

The front plate 29 includes an angled recess 2921 for the tyre valve, so that the valve does not protrude from the decorative front plate.

Fig. 31 (a) shows a cross sectional view of brake apparatus suitable for use with drive apparatus similar to the embodiment shown in Fig. 30. The brake coils 83 are wound in an annular recess in a ferrous member 841. The brake apparatus is designed for partial insertion into the cylindrical hole in the centre of the stator assembly of an external rotor motor. The space S is intended to be filled by iron from the stator assembly core. The brake apparatus includes a member 811 which is fixed relative to the iron member 841. A plurality of springs 82 urge an actuator plate 81 to trap two brake discs 812 against the fixed member 811. A front view of the apparatus is shown in Fig. 31 (b). The fixed plate 811 can be seen,

as can the outwardly protruding splines 8121 from the brake discs 812. These splines are designed to locate inside corresponding slots or recesses in the rotor assembly of the motor.

Use of two brake discs improves the braking performance.

The brake is thus spring applied.

The nominal pull in voltage is 24 volts and the minimum pull in voltage is 17 volts. The pull in current at 24 volts is 1.1 amps, and the pull in current at 17 volts is 0.78 amps. The hold in current at 12 volts is 0.55 amps (serious). The minimum holding voltage is 5 volts, and the mass of the brake apparatus is 2.1 kg. The diameters shown in the figure are in millimetres.

With regard to the embodiment of Fig 30, an external rotor design has been chosen specifically to get the force producing magnetic circuit to the largest diameter possible for highest torque production, and to use the rotor magnetic return path as part of the tyre rim. The motors is external rotor, 3 phase, 34 pole rotor and 36 toothed stator with Hall sensor commutation. (Brushless dc trapezoidal).

The windings are 3 phase star connected with options for delta connection. Star connected gives low speed high toque, and delta gives around double speed at half torque.

Each of the three phases consists of 12 coils, each coil is wound round 1 stator tooth and reversed on alternate teeth. This produces alternate north and south poles on consecutive teeth. However the 12 coils are split and wound on 6 teeth on one side of the stator and the remaining 6 wound on 6 teeth spaced 180° away. This configuration is to equalise the

magnetic pull between the rotor and stator avoiding excessive bearing loads and mechanically generated noise.

Normally coil strings are wound on adjacent bobbins with small loops of wire between each to facilitate insertion between the teeth of the stator laminations. Also, conventionally, where few large diameter wires are specified for each coil, they are made up of several smaller diameter wires. This to enable the wire to be inserted between the flared stator teeth and to reduce eddy currents in the wire resulting in losses and heat at high speed.

An alternative flareless tooth design does not require separate bobbins for winding-one coil string can easily be split into 6 and placed over the stator tooth. For in wheel applications, the flareless design allows insertion of large diameter wire and multi-parallel paths are unnecessary.

With regard to lamination design, in combination with the 34 rotor poles, the stator has 36 teeth.

This gives a vernier'action between the rotor and stator so that only 2 magnets and stator teeth edges align. Zero excitation detent (cogging) is therefore very low. The choice of magnet and tooth widths for these motors helps to reduce cogging. (Motor current is required to overcome detent before rotation occurs). High cogging causes poor low speed performance. the alternative tooth tip design for ease of winding has no tooth tip flares. A resultant increasing in cogging can be reduced by skewing the teeth of the stator. A decrease in back emf due to the reduction in tooth tip width and consequence increase in magnetic reluctance can be compensated for by a modest increase in the number of turns per coil.

Hall effect sensors are used to cause the amplifier to switch the windings in the appropriate sequence. These are positioned in the motor and are switched by the rotor magnets. 3 sensors may be used for conventional trapezoidal operation.

The external rotor design allows an internal bore in the stator. This is convenient for an integral brake. The motor allows for the addition of a brake substantially within the wheel width. The motor includes a brake that shares (magnetically) the motor state back iron achieving the required performance with an iron circuit that would not normally meet the specification outside the motor.

Features of the motor/brake combination include: a) The motor provides an 8"rim for a 3.00x8 pneumatic tyre allowing the option for a low profile 8"tyre to be fitted. b) The front of the motor is recessed to allow the tube valve to be below flush with the tyre outer face c) The front cosmetic plate is removable for brake servicing and for tyre replacement d) The brake is serviceable by means of friction plate replacement (adjustable airgap possibility to maintain pull off response), once the motor front plate has been removed e) There is no need to remove the wheel from the chair to change tyres or service the brake.

Each feature disclosed in this specification (which term includes the claims) and/or shown in the drawings may be incorporated in the invention independently of other disclosed and/or illustrated features.