This invention relates to magnetic bearings and is in particular concerned with thrust bearings for use in conjunction with rotating shafts.

It is well-known to support a rotatable shaft by means of a magnetic bearing. It is also known to control the axial displacement of such a shaft by means of a magnetic thrust bearing, as opposed to a conventional thrust race. A magnetic thrust bearing comprises a rotor mounted on the rotable shaft whose axial displacement is to be regulated and at least one fixed stator element mounted to associated hardware. The stator element has associated with it the energising windings whereby the position controlling force is generated, and regulated, typically under the control of electronic sensing/positioning circuitry. It is usual to manufacture the rotor component from solid, iron-based alloys and the same material may be

used for the stator element or elements. This is acceptable where the thrust load is fairly static and the rate of change of flux is not a significant factor.

However, where there are dynamic load changes accompanied by a high rate of change of flux, solid magnetic components are undesirable because of the eddy currents developed. These result in high power loss and very significant heating of the bearing components. Accordingly, a laminated construction is preferred for such applications, but if such a construction is to be effective, simple radially-directed laminations are unsatisfactory because of the reduction in magnetic surface area in the progressively radially outward direction. It will be noted that a typical lamination thickness is on the order of 0.1 to 0.25mm. Wedge shape laminations are not a satisfactory solution, because of this fact. One proposed solution is to increase the diameter of the co-operating components, though this does not really address the further problem which arises when the respective rotor and stator laminations do not actually align with one another at their outer periphery. Where there is no alignment, there will be a high reluctance; where there is alignment, there will be a low reluctance. The overall effect is one of rapid changes leading to higher winding losses and to the generation of high frequency transients in the windings.

It is an object of this invention to minimise these problems.

According to this invention, laminated stator and rotor components for a magnetic thrust bearing each comprise a plurality of generally radially directed laminations whose roots lie along the axis of rotation of a shaft located by the bearing and whose distal portions are progressively twisted relative to said axis such that the width of any lamination at any given radial distance from the root corresponds to at least the circumferential separation between adjacent laminations at that distance.

By progressively twisting individual laminations about their point of attachment to a central hub or other support, the effective radial gap between them is minimised, without increasing their thickness. In other words, the effective area of metal as seen in an axial direction is maximised, in contrast to that achieved by radially aligned flat laminations. Preferably the laminations of the stator element are twisted in the opposite sense to those of the rotor, to enhance their mutual alignment in use.

In order that the invention be better understood, an embodiment of it will now be described with reference to the accompanying drawings, in which:-

Figures 1, 2 and 3 illustrate prior art constructions, and

Figure 4 illustrates the present invention.

Figure 1 is a side view of a solid metal thrust collar 1. In use, it is mounted on a central shaft 2, and disposed between a pair of corresponding shaped stator elements provided with energising windings (not shown or further discussed here). The stator elements would in this case also be of solid metal, and provided with annular recesses for the windings. Such an arrangement might be used for light duty applications, particularly where stable and relatively modest axial thrust forces are generated.

Figure 2 shows a simple laminated thrust collar 2 corresponding to Figure 1, but having a plurality of radial laminations 3 supported by a central hub 4. It will be noted that the radially outermost tips of these laminations are separated by generally triangular gaps 5, which would in practice be filled with a magnetically inert material such as an epoxy or phenolic resin composition.

Figure 3 illustrates how the collar of Figure 2 would need to be increased in diameter in order to achieve the same effective magnetic area as that of the solid collar of

Figure 1. As only the diameter is changed, the same reference numerals as Figure 2 have been used.

Figure 4 is in several parts. Figure 4A is a cross-sectional side view taken along the axis of rotation. Figures 4B-4D inclusive are sectional views of the same part of Figure 4A, but as seen at different radial displacements from the axis of rotation.

Thus in Figure 4B, the lamination 10 is seen in plan view on line AA of Figure 4A. At this point, the width of the lamination (in a circumferential direction) is t, and in a conventional lamination its effective magnetic area would be t multiplied by (Re - Rs) the radial depth from the hub 12 to the outer rim. Figure 4C shows the same lamination, but this time in section on line BB. In this case, because of the progressive twist from the disposition shown in Figure 4B, the width t' is now t/sin x, where x is the angle relative to the face of the collar. It will be seen that t' is t. RB/Rs, the effective area being increased over that of Figure B by the factor RB/Rs.

At the radially outermost point, as seen in Figure 4D along line CC, the twist angle is y and the value of t' is now t.Rc/Rs.

RC is greater than either RA or RB, so the effective area is greater and accordingly, the overall performance achieved by progressive twisting is much closer to that of a solid collar such as that of Figure 1 than is that of the simple flat radial assembly of Figure 2.

It will be appreciated that the arrangement of Figure 3 could be used with wedge shaped (in side view) laminations since although the root width is fixed by the hub dimensions, twisting reduces the axial thickness of the assembly. Accordingly, wedge shaped laminations may be used to counteract this. Figure 4 does in fact assume that this particular constructions is used.