GB2324168A - Optical deflector and beam splitter - Google Patents

Abstract

A deflector is provided which comprises a array of rotating lenses (30, 56) which pass across an optical beam in order to provide an unidirectional scan. The lenses may be modified in order to control the shape of a focal plane. A beam expander (Figures 31, 32, 33, not shown) providing two arcuate opposed beams may supply light to the deflector, in order to eliminate coma. The scanning system may also include beam splitters (Figures 40, 41, not shown). The scanner is for the recording or writing of an image and use in bar code scanning is mentioned.

Description

BEAM EXPANDER OPTICAL DEFLECTOR. BEAM SPLITTER AND SCANNER The present invention relates to a beam expander, an optical deflector, a beam splitter and to a scanner. Such a deflector is suitable for deflecting a laser beam through an arc, and the scanner incorporates the optical deflector. The scanner may be used to read and/or write an image.

Known deflectors typically comprise mirrors on a galvanometer type mounting. Thus the mirror performs a reciprocating motion and must be accelerated and decelerated during each scan. Furthermore, the mirror must be returned to a start of scan position. This reduces the scan duty cycle typically to less than 50%. Alternatively bi-directional scanning may be performed where a first line of an irnage is scanned from left to right and the subsequent line is scanned from right to left. This cyclical scan sequence is than repeated.

According to a first aspect of the present invention there is provided a deflector comprising at least one refractive optical element arranged, in use, to undergo rotation relative to a central axis such that the or each ofthe refractive optical element experiences relative movement with respect to the path of a light beam.

It is thus possible to provide a light deflector (where the term "light" may encompass optical radiation in the infra-red, optical and ultra-violet regions) which can run at a constant rotational rate. This therefore overcomes the problems associated with speed of operation and speed control associated with galvanometer type deflectors.

Preferably a plurality of optical elements are provided in a rotationally symmetric array.

Preferably the optical elements are lenses. Advantageously the lenses have a plane surface.

In a preferred embodiment ofthe present invention, the lenses are plano-convex. Alternatively the lenses may be circular or cylindrical. Cylindrical lenses have the advantage that the lenses may be placed in abutment thereby offering continuous or near continuous deflection, that is a duty cycle in excess of 95% and preferably approaching 100%.

Advantageously the deflector comprises a plurality of refractive elements arranged to form an array exhibiting rotational symmetry and having an inner surface which is defined by a smoothly varying function. Thus the inner surface, which may, for example, be circular or elliptical, does not change direction abruptly. Other forms of lenses such as plano-concave or meniscus lenses may also be utilised. However, in some embodiments of the invention the inner surface does include abrupt changes of direction and may even be discontinuous.

Advantageously, the lenses may have an elliptical or modified elliptical refracting surface.

The lens may be divided into central and peripheral zones. The central zone being defined, in use, as the portion of the lens within which an inside edge of the ray bundle remains during a scan, the inside edge being the side of the ray bundle nearest the central portion or optical axis of the lens. The outer portion of the lens may then have a modified form, still generally elliptical but modified by an exponential curve, which modified the path varies the deflected angle of the outer edge of the ray, thereby controlling the position in space at which the inside and outside rays converge, and consequently controlling the shape of the focal plane of the deflector.

In a further embodiment, the deflector may only have a single lens. Both the inner and outer surfaces of the lens may be curved in order to allow the lens to be formed over a large arc, and hence interact with the light beam over a large angle of rotation. In an exemplary embodiment of a lens having a high duty cycle, the lens extends over a complete revolution. The inner surface of the lens is substantially circular and the outer surface is an approximation of a cycloid curve. The cycloid may be modified in order to control the shape of a focal plane of the lens. In an alternative variant, the lens may exhibit reflection symmetry along one or more axis, thereby giving rise to bi-directional scanning. Thus the outer surface of the lens may, for example, be elliptical or both surfaces of the lens may be defined by non-concentric circles.

Preferably the light beam is focused by the deflector as it exits therefrom. In the case of small diameter beams, this may reduce divergence over short distances from the lens. If the beam is to be focused to a greater distance from the lens, or the beam is to be expanded, a further lens may be introduced between the laser light source and the deflector in order to extend the focal length. This may result in the focal plane becoming curved, but this may be corrected by the introduction of a further lens, such as an F-theta lens after the deflector, or by designing the deflector lens to be aspheric.

Preferably the deflector can achieve wide angles of deflection, typically being 60 or more.

Advantageously the deflector forms a virtual point of origin of the beam. Advantageously this point is substantially fixed in space for each deflection. The virtual point of origin need not correspond to the origin of the laser light source. A virtual point of origin has the advantage that it offers a way of alleviating the errors in reflection angles that occur when using a rotating mirrored polygon of the prior art.

Advantageously a field lens is provided intermediate the deflector and a focal plane. The field lens may reduce the angle of deviation between the beam as it arrives at the focal plane and the local surface normal of the focal plane. If the field lens extends across the full width of the focal plane, then it can provide correction across an entire scan line, thereby ensuring that the scanning beam is always substantially perpendicular to the focal plane.

Preferably a beam expander is provided having an annular output. It is thus possible to reduce or eliminate coma.

According to a second aspect of the present invention, there is provided a scanning system comprising a beam deflector, a focal surface and a field lens intermediate the beam deflector and the focal surface, the field lens reducing the deviation from the local normal to the focal surface at which the light beam impinges on the surface.

According to a third aspect of the present invention there is provided a beam expander comprising first and second co-operating reflectors, at least one of the reflectors having an annular or an arcuate reflecting surface.

It is thus possible to produce an expanded beam having a central portion omitted therefrom.

The reflectors may be mirrors, or reflecting surfaces. Alternatively, total internal reflection of light within an object may be utilised to form the reflectors.

According to a fourth aspect of the present invention there is provided a beam expander having at least one reflector utilising total internal reflection.

A scanner constituting an embodiment of the invention may be used in conjunction with a intense light source, for example from a laser, which is controlled in a switchable manner to write an image. Such a device can even be used to cut or drill an item, for example to form holes in a printed circuit board. The scanner may itself be mounted for rotation such that, as the scanner rotates, it can write across the entire surface of an object, such as a circuit board held in ajig.

According to a fifth aspect of the present invention, there is provided a method of modifying the surface profile of a focussing element so as to obtain a desired focal surface when the focussing element is moved with respect to a light beam passing there through comprising the steps of:1. defining the width of the beam; 2. analysing the range of relative motion between the light beam and the focussing element over a working range in order to identify a first portion of the lens through which a portion of the light beam always passes, this region forming a first unmodified portion of the lens; 3. defining the shape of the desired focal surface; 4. iteratively for a series of relative positions of the lens with respect to the light beam calculating the paths of two spaced apart rays of the light beam as they pass through the lens, calculating their point of intersection, modifying the profile of a portion of the lens outside the first unmodified portion, recalculating the ray paths and accepting those modifications which reduce the distance between the focal point and the desired focal plane; and 5. repeating step 4 until the actual focal plane approximates the desired focal plane to within an acceptable defined tolerance.

According to a sixth aspect of the present invention, there is provided a beam expander arranged, in use, to provide an exit beam from which a first portion is omitted.

Preferably the first portion is a central portion of the beam. In this way the effects of optical coma in subsequent optical components can be reduced or eliminated.

Preferably the beam expander comprises first and second co-operating optical elements.

Preferably the first and second optical elements are mirrors. It will, however, be realised that refractive optical elements can in principle be used in place of reflective optical elements within the beam expander. This would however give rise to a physically larger design.

Preferably the first mirror comprises one or more of a conical reflector, two conic portions arranged back to back, a plurality of conic portions arranged in an array, or a plurality of mirror elements arranged in an array. A feature of each of these arrangements is that the or each reflector intersects the or part of an incoming optical beam and reflects it towards an associated part of the second reflector.

Preferably the second mirror comprises an annular mirror, an annular portion of a parabolic or similar section of a mirror, a plurality of such annular or parabolic portions, a plurality of part annular or parabolic portions having their virtual optical axis matched, for example coincident with, the optical axis of an associated portion of the first mirror.

The reflecting surfaces of the first and second mirrors may be formed by depositing a reflective layer on a glass or plastics or other suitable surface. Preferably the mirror is formed using total internal reflection within a optical element.

By arranging for the first element to be concentric with the second element and arranged such that the first and second elements can undergo relative axial displacement it is possible to vary the width/diameter of the expanded beam and/or to vary the focussing thereof when the second optical element is designed to have a focussing action. This may be advantageous when it is desired to scan an undulating surface.

The use of profiled ie conical or part conical mirrors enables the energy from the missing first portion of the expanded beam to be directed into the remaining portion or portions of the beam, thereby maintaining optical efficiency.

According to a seventh aspect of the present invention, there is provided a beam splitter arranged to separate two beams which are slowly diverging, characterised in that the beams are reflected from a surface positioned such that a first beam experiences reflection at the surface and such that a second beam does not.

Advantageously the second beam undergoes refraction at the surface.

Thus the properties of total internal reflection can be used to separate the first and second slowly diverging beams. The reflection can occur at any abrupt change of refractive index and consequently may be internal (both beams already in a block of material and one of the beam allowed to exit the block) or external (both beams incident on a block of material and only one beam allowed to enter the block).

In general, optical modulators comprise opto-acoustic crystals which modulate a light beam by Bragg scattering. The modulators have two beams emerging therefrom, typically separated by only 1.5 or so. One of the beams has a high modulation index and is regarded as the modulated beam. The other beam is essentially unmodulated and needs to be separated from the modulated beam. Since total internal reflection is highly responsive to the angle of incidence it can be used to reflect one of the beams but not the other. Beam separators using this technique exhibit extinction ratios in excess of 1000.

A scanner having the ability to write an image onto a photo-sensitive sheet may incorporate one or more of the aspects of the present invention.

The present invention will further be described, by way of example, with reference to the accompanying drawings, in which: Figure 1 schematically illustrates the divergent properties of a plano-concave lens; Figure 2 schematically illustrates how the lens illustrated in Figure 1 can be used to deflect a light beam; Figure 3 schematically illustrates the divergence obtained by a square array of four planoconcave lenses.

Figure 4 schematically illustrates the divergence of a light beam having a finite diameter; Figure 5 schematically illustrates how the divergence varies as a function of the position of the lens; Figure 6 schematically illustrates how a finite beam can become distorted during scanning; Figure 7 schematically illustrates the variation of focal length with respect to deflection angle; Figure 8 schematically illustrates the ray paths of plano-convex lens used as a diverging element; Figure 9 schematically illustrates the focusing action of a plano-convex lens; Figure 10 schematically illustrates how a second plano-convex lens preceding the deflector lens can be utilised to increase the focal length; Figure 11 schematically illustrates how a converging lens preceding a plano-concave deflector lens can be used to correct divergence.

Figure 12 is a plan view of a deflector constituting a first embodiment ofthe present invention; Figure 13 is a plan view of a deflector constituting a second embodiment of the present invention; Figure 14 is a schematic sectional view through the embodiment shown in Figure 12 and illustrating the optics for directing a laser beam towards the lenses; Figure 15 shows the profiles of three elliptical lenses used to demonstrate the effect of the eccentricity of the lens on optical performance; Figures 16a to 16c illustrates the lens profiles and amounts of deflection for 25 on input beam rotation; Figure 17 shows the inside and outside ray paths in greater detail; Figure 18 shows the lens profile in greater detail; Figure 19 shows the addition of a horizontal cylindrical lens for z-plane focussing; Figure 20 shows an alternative lens arrangement for pre-focusing and z-plane focussing; Figure 21 shows a further alternative lens arrangement for Z plane focussing; Figure 22 schematically illustrates the deflector and exit light beam; Figure 23 schematically illustrates a deflector of reduced height; Figure 24 schematically illustrates the optical action of a field lens; Figure 25 schematically illustrates the reduction in vertical mis-alignment as a result of introducing a field lens; Figure 26 illustrates a modified field lens of reduced width; Figures 27a and 27b show the non-linear relationship between angular rotation and projection onto a plane; Figures 28 and 29 illustrate the relationship between beam width and dead zone; Figure 30 illustrates optical coma; Figures 31 a and 31b illustrates a beam expander which omits a central beam portion; Figure 32 illustrates a modified beam expander; Figure 33 shows an improved beam expander.

Figure 34 illustrates the path of a non-collimated beam through the beam expander of Figure 31a; Figure 35 schematically illustrates a modified beam expander including a concave second reflecting surface; Figure 36 schematically illustrates a beam expander having curved mirror surfaces to reverse the intensity profile of laser light sources; Figure 37 schematically illustrates a beam expander utilising total internal reflection within a transparent block of material; Figures 38a to 38d illustrate some of the processing steps in the manufacture of the beam expander illustrated in Figure 37; Figures 39a and b illustrate embodiments of lenses giving high duty cycles.

Figures 40 and 41 illustrate embodiments of beam splitters; and Figures 42a and 42b illustrate the effects of varying the values of C and Sf in the first equation of the appendix.

As shown in Figure 1, a plano-concave lens 2 having its concave side facing towards a light source 4 has a diverging action, as illustrated by the ray paths 6. Such a series of lenses 2 may be mounted in a circular array. Furthermore, if the radius of curvature of the concave face 8 is selected to equal the internal radius of the array, then a beam of laser light incident on the inner surface 8 will always be perpendicular to the inner curve surface 8, and consequently will pass through it and into the solid body of the lens without refraction. Only when the light encounters the radially outward facing surface 10 of the lens does it refract. The angle of refraction varies as a function of the lens position as the lens spins with respect to the light beam 12, as illustrated in Figure 2.

At the first position A, the light exits perpendicular to the planar surface 10 and consequently is not deflected. As the lens sweeps between positions B and C the surface 10 becomes inclined with respect to the light beam 12 and consequently the light exits along the paths 14 and 16 becoming increasingly deflected for positions B and C, respectively.

A plurality of such plano-concave lenses may be joined together to form a square array as shown in Figure 3. The array comprises four plano-concave lenses joined such that the planar surfaces of the lenses 2 define a square surface 18 and the concave surfaces co-operate to define a circular surface 19 (the surface 19 may be cylindrical or define part of a sphere). In the above embodiment a ray represented by chain-dot line 20 does not undergo deviation, whereas a ray rotated through 30 with respect to ray 20, and as indicated by dotted line 22 undergoes a deflection at the outer face of the lens of approximately 18.6 .

It should be noted that the light ray may be rotated using a rotating deflector, such as a rotating mirror, or the beam direction may be held fixed and the lens array rotated. Rotating the ray 30 gives a total deflection of 48 degrees, or so, whereas rotating only the lens array through 30 gives nearly 19 of deflection.

So far, the diameter of the light beam has been assumed to be negligible. However, this is in general not the case. Figure 4 illustrates that for a light beam 24 of finite width the lens 2 produces a divergent beam 26. However, as shown in Figure 5, the amount of divergence varies as a function of the angle of the lens with respect to the incident light beam 24.

A further problem associated with real light beams of finite width is that, when refracting an initially circular collimated beam using an array of plano-concave lenses, the beam shape becomes distorted into an ellipse as the deflection angle increases, as shown in Figure 6.

Figure 6 shows the spot shape as it arrives at an arbitrary scan plane, over which the beam is to be scanned.

A scanner having lens arrays of the type shown in Figure 3 or Figure 6 can be used with acceptable performance if the distance from the deflector to the scan plane or target is short and the beam diameter is not too important, for instance in bar code scanning. In this application the deflector can be kept small and offer large deflection angles or high scan rates.

A prototype having six lenses was run at 20,000 rpm giving 120,000 scans per minute at deflection angles of50".

Most scanning application require the beam to be focused to a specific spot size. For the deflectors shown in Figure 3 and 6, the beam must converge out of the deflector. This means that the rays are not parallel within the deflector material. This causes the focal length to vary as the deflector rotates as schematically illustrated in Figure 7. Thus not only are there problems with achieving a constant spot size by virtue of the beam becoming elliptical, but achieving a good focus has, hitherto, been almost impossible for small spot sizes.

If instead a plano-convex lens 30 is used with the planar side facing towards the light source, as shown in Figure 8, two advantages are gained. Firstly the beam will converge as it leaves the lens 30, and secondly the curvature of the lens in relation to the optical path of the laser can be selected so as to achieve a virtual point of origin 32. Figure 8 is similar to Figure 2 in that it illustrates the lens 30 at three superimposed positions and the optical path for a thin light beam passing there through. Positions 30A, 30B and 30C represent the lens at varying angles with resect to the incoming light ray 12, and ray paths 34A, 34B and 34C illustrate the corresponding exit beams.

In order to achieve the virtual point of origin, the radius of the curved surface of the lens 30 should substantially match the radius about which the lens rotates. Over small angles of the deflection, the virtual origin is substantially stationary. If large deflection angles are used, the point of virtual origin may shift by a larger amount, but typically by less than a few tens of microns.

As shown in Figure 9, a light beam 40 of finite width becomes converged as it passes through the plano-convex lens 30. Because the plano-convex lens focuses the beam to a point quite close the lens, this may lead to an artificially large spot size when the beam eventually arrives at a target. However, if a second plano-convex lens 42 is interposed between the light source and the deflector lens 30, the incident beam will be a diverging cone and this has the effect of lengthening the focal point of the deflector lens as shown in Figure 10.

If a plano-concave lens as illustrated in Figures 1 and 2, is used as the deflector, the natural divergence of such a lens can be countered by placing a converging lens 44 close to the deflector lens in order to correct for the divergence, as shown in Figure 11.

Figure 12 illustrates a first embodiment in which 6 plano-concave lenses 30 are arranged in a circular array. The laser light source may, for convenience, be arranged to be coincident with the axis of rotation ofthe array. As the array rotates, each lens 30 passes in front ofthe laser beam thereby causing the laser to scan across a deflection path in the same direction.

Thus the scan is always in the same sense, for example left to right each time. However, if the sense of rotation is reversed, then the direction ofthe scan will also be reversed. Because each scan line is always in the same direction, this simplifies the scanning operation and allows a continuous feed of the media to be scanned in order to minimise scan time.

The deflector may be used in a digital mode where positional information concerning the lens positions is required. However, it is envisaged that the lens array would normally be run at a constant angular speed. In such an arrangement the only limit to deflection frequencies are the mechanical limits imposed by centrifugal forces. If small lenses are used in conjunction with a small beam diameter, for example 100 microns, the optics can be kept to a modest 20mm diameter. Furthermore, the number of lens segments may be increased thereby increasing the scanning frequency. A prototype having nine lens segments was spun at 10,000 rpm producing a scan frequency of 1.5KHz. DC motors can run at speed much higher than this and hence higher scan rates are achievable.

Cylindrical lenses may be used instead of circular ones thereby enabling the lenses 30 to abut one another, as shown in Figure 13. Thus as the laser completes each scan in one lens segment, the next lens offers virtually instantaneous deflection back to the start point For this reason, there is virtually no null point in the deflection process and the arrangement has an exceptionally high duty cycle. In the prototype, having nine segments the duty cycle measured was approximately 0.97, the loss occurring due to minor bevelling at the edges of the lenses.

A minor limitation of using cylindrical lenses is that the beam convergence will exist in a one optical plane only, namely the plane of the lens shape. There will be no convergence in the plane perpendicular to the scan direction and consequently the laser scan spot is not circular, ie some astigmatism is introduced. This can be used to advantage in some scanning optics.

However if a circular spot is a requirement, then circular lenses are required.

Figure 14 schematically illustrates a carriage for the lenses. The carriage 50 is mounted above a motor 52. An upstanding wall 54 of the carriage serves to support the ring of lenses of which one lens 56 is illustrated for simplicity. A laser light source 60 is mounted above the axis of rotation of the carriage 50 and directs a light beam 62 towards a stationary mirror 64 inclined at an angle of 450 with respect to the beam 62. The mirror rotates the light 62 through 900 thereby ensuring that the beam is perpendicular to the axis of rotation and originates from a position coincident with the axis of rotation.

It is thus possible to provide a scanning system having a large duty cycle and utilising refractive optics.

In embodiments requiring a very small spot diameter, say 10 microns, over a relatively large distance, perhaps 500mm, it is conventional to use a beam expander. However, because of diffraction limits, the light spot in such a system may have to start life as a beam as large as 20mm as it exits a beam expander. The converging nature ofthe lenses may be used to ensure that a beam starting with such a large diameter can converge to a fine focal point. However, it will be appreciated that since the focal point will always be at a more or less constant distance from the lens, the focal plane will be curved.

In a further embodiment, a deflector was produced having a nominal 40mm beam diameter focussing over a distance of approximately 80cm. In order to have the internal beam converge, a convex lens surface was chosen. A concave lens surface would cause the beam to diverge within the lens (tracing the ray paths backwards from the focal spot) resulting in a larger bore and hence large overall diameter of the deflector array.

The external face of the lens could have a cylindrical profile, but calculation and experiments have shown that a higher rate of deflection and a smaller internal beam diameter can be achieved by using an ellipse as the external profile. In fact any curve such as a parabola, a hyperbola or a continuous exponential function can be used, or several curves can be used as long as a smooth blend can be made.

As shown in Figure 15, a series of elliptical lenses were produced. Each lens was required to have an exit beam of 40mm diameter and a divergence angle of 3 , thereby focussing the beam onto a radial focal plane at a distance of 814mm when the lens is in a central position in which one of the principal axes of the ellipse is aligned with the undeflected beam.

The lens was formed by the shaded area 70 shown in Figure 15. Three lens shapes were tested, each had a front surface 50mm from a principal axis of a ellipse, and a rear planar surface 20mm from the principal axis of the ellipse. The radius of the other axis was either 40mm, 50mm (circular) or 60mm. The three different lens profiles are shown in Figures 1 6a, l 6b and 1 6c respectively. Figures 1 6a to 1 6c also show the beam deflection for a 25 rotation of the beam with respect to rotational centre of the deflector lens.

For the lens shown in Figure 16a, the inside ray moves only 10 for the 25 input rotation, yielding an equivalent of 15 of beam deflection. This offers better performance than the lenses of Figures 1 6b or 16c. For example, the lens in Figure 1 6c which deflects nearly 20 for the 25 input rotation, thereby offering only 5' of overall deflection. It should be noted that the convergence of the internal beam is greater in the lens illustrated in Figure 1 6a than in Figures 1 6b or 1 6c, but the lens in Figure 1 6c offers a focal performance slightly closer to the radial focal plane wanted.

High deflection rates for small mechanical rotation angles may sound attractive, however when used with width beams they lead to low duty cycle. Experiments have shown that for wide beams a maximum duty cycle is achieved using the lowest number of facets, ie one, two or three. Thus a single lens could be used on a resonance deflector and using scanning on both the forward and reverse strokes. However, for high resolution high speed scanning, it kept small in overall diameter, and the duty cycle optimised, then a lower deflection rate (ie ratio of deflection angle to mechanical rotation angle) is desirable.

Thus far, it has been shown that the elliptical form of the lens determines the deflection rate by virtue of controlling the deflection of the inside ray path. For clarity, the inside ray path is designated 74 in Figures l 6a to l 6c and is the ray path that, as deflection increases in any given direction from the central axis 76, remains closest to that axis 76. However, it is still necessary to force the direction of the outside or leading edge ray 78 such that it intersects with the inside ray at the required focal length.

As shown in Figure 17, the inside ray 80 always passes through the elliptical profile of the lens 82 irrespective of the angle of rotation. This is infact a design choice. It is therefore necessary to modify the path of the outside ray 84 such that it follows the path of the chain line 86 in order to generate the required focal condition. Calculation has shown that the intersection point can be designed to any shape. Thus the deflector can be designed to function as a flat field lens just as easily as it can be made to generate a radial focal plane, as is the case described herein, or even to have an increasing focal length if it is desired to focus the beam onto a convex surface.

The desired focal plane was formed by extending the ellipse with a series of exponential curves.

The purely elliptical profile of the lens shape is taken only up to the width of the internal beam diameter, which corresponds to the point A shown in expanded detail in Figure 18. (It should be noted that this corresponds to the edge of the beam in the zero deflection state). Taking the intersection of the axes in Figure 15 as an origin, point A in this example has the XY coordinates of 20, -41.3 relative to the origin or, if we redefine the origin of the co-ordinate system to correspond to the most forwardly facing part of the elliptical lens, then the coordinate is transformed to 20, 8.05 and the gradient at this point is 40.4". A modifying exponential curve is matched to point A such that is has the same gradient. The lens or beam is then rotated to an arbitrary angle and a ray traced through it, which will intersect to define a point B which has new XY co-ordinates along the gradient. All three values, of which can be established by the combined exponential equation. The resulting refracting ray will intersect with the inside ray to form a focus at the desired point. The exponential function is expressed in a modified form y = S,[(exp(Cx/S)-l] where y and x are co-ordinates C affects the rate of change of the curve SF is a scale factor from mathematical manipulation of the above, it is possible to find the X position of any given gradient o (gradient being expressed as an angle) Xexp-(SF/C) Ln (tan 8/C) In order to superimpose the exponential curve onto the elliptical lens profile, a further origin needs to be defined.

X shift = X exp - X ellipse Y shift = Y exp - Y ellipse Where the subscript shift denotes the shifted co-ordinate systems, and exp and ellipse denote co-ordinates ofthe origin ofthe exponential function and the ellipse, respectively.

The lens profile in the modified portion, ie towards the edges thereof, is modified by an iterative numerical process. Briefly, for a point in the modified area, a small correction to the lens profile is made, the ray path calculated, the intersection point with the inner ray calculated and then the error between the intersection point and the focal plane calculated, based on whether the correction has increased or decreased the error, a new correction is calculated and the process is repeated, in a step wise manner for various angles of deflection until acceptable performance is obtained. The profile of the lens is modified as a series of adjacent portions, each of which blends with the adjacent portions.

Thus the method can be used to derive a focal map and to pull the outside ray to a new position such that it intercepts the inside ray at the required focal point. It is conceivable that as long at the focal plane is a constant shape, that even a cyclic or undulated focal plane can be mapped and scanned.

A more detailed description of the mathematical analysis is given in Appendix 1 which constitutes part of this specification.

As the deflector rotates, the beam will pass through optical material of varying thickness. At a zero deflection angle, the lens is at its thickest, but as the lens progresses towards the maximum deflection angle, the outer rays pass through increasingly thin sections of lens, whereas the inner lens is still passing through the thicker lens material that lies within the elliptical section. Thus an attempt to focus in the Z plane of the deflector lens will cause the focal spot to become distorted and the degree of distortion will vary according to many functions, such as its focal length, beam diameter, deflection angle and the like.

Depending on the spot size and quality required, it is possible to shape the inside face of the deflector into a cylindrically convex lens, as diagrammatically shown in Figure 19. If sufficiently sophisticated manufacturing methods are available, the radius of curvature of the inside profile 90 can be varied with progressive rotation angles to compensate for the optical distortion and changes in the focal length. Alternatively, there are several other methods available. The first of these is to start focusing the light beam prior to its entering the deflector. This is not a significant problem since the beam requires prefocusing in the XY plane anyway. Two cylindrical lenses can be piggy backed, each with different focal lengths to accommodate the different focal requirements of the XY planes and Z planes, respectively.

Such a piggyback lens arrangement is shown in Figure 20 and comprises a first lens 94 having a vertical cylindrical axis adjacent a second lens 96 having a horizontal optical axis. Where the spot quality requirements are very demanding, the decreasing thickness of the glass lens might lead to slight changes in the focal length If this is the case, the external lens needs to be placed in front of the deflector lens and the light left collimated in the Z plane as it passes through the lens. A lens arrangement which achieves this is shown in Figure 21. The correction lens is shaped like a donut with a flattened centre. The lens 100 only effects focusing the Z plane as it is rotationally symmetric about the deflector 102. This solution is compact, and since it can be placed very close to the deflector lens, can also serve as a protective face should the deflector shatter at very high rotational rates. Secondly, by enclosing the rotating deflector lenses, the noise generated by turbulence (windage) is greatly attenuated.

A more sophisticated solution to Z plane focusing also offers a solution to flat bed scanning and uses the phenomenon of spherical aberration. Although it is possible to design the deflector lens to focus onto a flat focal plane, it is equally possible to compensate for the ellipticity of pixels due to the angle of projection onto a flat plane and the two design requirements might not be achievable simultaneously. Also there is a mechanical rotational speed limit associated with any rotating system. In the embodiment discussed here, the focusing a beam needs to have a 40mm start diameter at the deflector surface. This requirement, by definition, means that the beam diameter in the Z plane is 40mm and therefore the lens hight needs to be at least 40mm, as shown in Figure 22. However, the height of the deflector lens may be reduced as shown in Figure 23, by limiting the beam size in the Z plane and focusing the Z axis using a long cylindrical lens 110 close to the target. Thus in the example shown in Figure 23, the beam leaves the deflector with a diameter of 40mm in the XY plane but a height of only 12rnm in the Z plane. During the beam's path to the focal target, the beam remains collimated in the Z plane but continues to focus in the XY plane.

The lens 110 acts as a flat field lens, and essentially is a thin slice through a plano-convex lens but with a lens diameter equal to the width ofthe scan target, as shown in Figure 24. A further advantage of using such a lens, as shown, is that there exists a focal distance from the planar surface of the lens to the target focal plane whereby the distance is constant and the rays emanating from the flat face are perpendicular to a target. Thus the field lens 110 causes the focal plane to be scanned in a telecentric manner (ie light rays always perpendicular to the focal plane). As a result of this continually perpendicular illumination, the focused beam is void of ellipsity. If the inwardly facing side of the lens is the curved surface then the light which remains columnated up to this point in the Z plane is effectively focused by what it sees as a long cylindrical lens to the target. Alternatively, a radially curved cylindrical lens could be used to achieve a similar focusing effect.

Focusing of the Z plane in this manner offers other optical advantages. For example, any small vertical misalignment of the deflector lenses will generate an inter line error during the scan process. Positioning the focusing lens so close to the target lens means that these errors are significantly reduced, as schematically illustrated in Figure 25. Light rays 120 to 122 represent incoming rays of a bundle formed by a first deflector lens, whereas the offset rays 125 to 127 resulting from misalignment in the deflector are both focused to effectively the same point on the focal plane.

In the above embodiment, a radial planar convex flat field lens was chosen to maintain a perpendicular beam so as to obtain a circular pixel at the focal plane. However, other design parameters such as a restricted scan/deflection angle might mean that a small amount of ellipsity is permissible. In this case, the flat field lens can be moved to still offer a flat focal plane but with a slightly angled beam, as shown in Figure 26. A consequence of this arrangement is that the modified lens 130 does not need to be the full width of the focal plane.

The embodiment described having focal flexibility offers many advantages over prior art scan deflection systems.

A. Optical Efficiency Reflective deflectors such as rotating polygons and reciprocating mirrors mounted on galvanometers require that the focusing be done with the use of additional lenses. Apart from the cost saving of having fewer lenses, the fact that the embodiments of the present invention also serve to focus the beam within the deflector means that light losses associated with absorption through glass and losses associated with degrading reflectivity of mirrored surfaces means that less light power is required, this is an important feature when using infra-red or ultra-violet light.

B. Correction of Elliptical Pixel Proiection A feature of the deflector is that divergence angle at the extremes of deflection is slightly greater than the central beam angle. This increase in divergence angle means that the projected pixel will be slightly smaller in the XY plane. If this deflector is intended to project onto a flat focal field, then there is a natural tendency for the pixel to become elliptical as a result of the projection of the circle onto the flat plane at an angle. By increasing the divergence angle at the periphery of the deflection, the pixel size is reduced in this direction thus allowing a uniformly circular pixel to be maintained if required.

The degree of deflection is variable and other features might predominate but there can always be some degree of correction.

C. Scan Linearitv The mechanical rotation of the deflector causes the deflection of the beam. However, for a planar focal plane, the scan rate is not linear. Figure 27a shows the projection onto a flat focal plane of rays at 50 intervals. Thus for a scanner rotating at constant speed, the scan rate towards the edges of the scan plane increases compared to that in the centre, thereby introducing a distortion.

With a refractive deflector, this can be corrected. Thus the angular deflection rate towards the edges of the scan can be reduced. The deflection rate needs to be reduced by approximately 2% for each successive 5 of rotation in order to maintain a linear projection on the focal plane, as illustrated in Figure 27b. However, scanning systems often utilise electronic means to linearise the scan and consequently linearity is not so important as might be supposed.

D. Dutv Cvcle For a single lens, having a working range of +30 of deflection, the duty cycle is approximately 16.67%. Since 300 out of 360" of mechanical rotation results in a deflection. Use oftwo lenses would increase the duty cycle to 33.3%. It is then tempting to suggest that six lenses will increase the duty cycle to 100%. However, use of six lenses can result in a large and heavy deflector which may limit the speed that the motor can be run at. Furthermore, after the deflection from one lens is finished a certain proportion of the deflector must rotate sufficiently to flilly engage the beam into the next lens. Thus there is effectively a dead zone between adjacent lenses. Thus duty cycle is a compromise between the number of deflections that occur per motor revolution, ie the number of lenses used, and the size of the dead zones, and lastly the portion of each deflection that is used. Thus if the deflection lens is capable of 60 of deflection, but the scanner only requires 20 , then 40 of this deflection is wasted. With a refractive deflector, the optics can be designed such that a deflection angle required is in fact achieved therefore limiting the wasted proportion oftime to solely that resulting from the dead zones. For the refractive deflector, the dead zone is to some extent a function of the beam width and, if large beams are to be used with small overall deflector sizes, then the duty cycle is maximised by using the minimum number of lenses. This also makes it easy to control the vertical alignment of the lenses. For small diameter beams, a cluster of 3 or 4 lenses produce duty cycles in excess of 80%. However, if the beam is large then the duty cycle drops, Figure 28 schematically illustrates a three lens array with a wide exit beam. In this example, the lens can only rotate through 24 before the extreme ray comes to the end of the lens. The deflector must then rotate a further 72" before a beam is fully engaged within a single lens. Thus the 72" of rotation effectively represents the dead zone thereby giving a duty cycle of only 40%. Figure 29 illustrates an equivalent arrangement but with a much smaller beam. This time the lens can rotate to 450 before the extremity of the beam leaves the body of the lens. The deflector then needs only to rotate a further 32 before the next lens is fully optically engaged. Thus 3 x 32 = 96" of rotation represents a dead zone, thereby giving a duty cycle of over 73%.

Figures 39a and 39b schematically illustrate alternative deflector arrays having duty cycles in excess of 90% and, when used with relatively thin beams, approaching 100%. The lens shown in Figure 39a comprises a generally circular inner surface 300 and an outer surface 302 which may be circular or elliptical. The inner surface 300 is offset with respect to the outer surface such that the lens has a thin portion 304 disposed opposite a thick portion 306. Thus the surfaces 300 and 302 are inclined with respect to each other intermediate the positions 304 and 306. Thus the item serves as a deflector. Since the deflector exhibits mirror symmetry, the beam is returned back along the path it travelled. Thus the deflector shown in Figure 39a is suitable for bi-directional scanning. The deflector 320 shown in Figure 39b again comprises a substantially circular inner surface 300. However, the outer surface approximates a cycloid curve. The beginning and end of the curve are bridged by a step portion 322. As before, the inner and outer surfaces remain inclined with respect to each other and consequently again give rise to a varying deflection which changes with angle of rotation. The step feature 322 allows the bearn to undergo rapid flyback. Thus this deflector gives rise to a unidirectional scan.

The exterior surfaces 302 and 320 may be modified as hereinbefore described in order to control the shape of the focal plane.

When designing refractive deflection systems, some imperfections in them systems must be considered. One of these is coma which is an optical phenomenon which arises because a central ray of the beam does not necessarily meet the focus of the inside and outside rays of the beams, as schematically illustrated in Figure 30. Here, central beam 140 intersects the scan plane outside the intersection of the extreme beams 142 and 144. In order to overcome this problem a beam expander may be used that generates an annular beam rather than a solid conical beam may be used. In this way, the central ray no longer exists and therefore coma is eliminated. A suitable annular beam expander is schematically illustrated in Figures 3 la and 3 lb. The beam expander comprises an internal 45" mirrored cone 150 which is essentially disposed within a cylindrical 45" mirror 152. In use, a light beam, say having a diameter of 1 0mm is incident upon the cone 150. The light is reflected to 900 thereby forming a disk of light which impinges upon the cylindrical mirror 152. The thickness of the disk of light is equal to the beam radius. The mirror surface reflects the beam through a further 90" thereby generating an annular beam. The mirrored cone 150 may be moved axially with respect to the cylindrical mirror 152 to vary the diameter of the annular beam. In Figure 31 b, the central cone is displaced downwardly compared to the arrangement shown in Figure 31 a thereby producing a larger expanded beam. However, in one embodiment of a deflector described herein a complete circular beam profile is not required since the beam need only be 40mm x 12mm. In order to achieve this, sections of the outer cylinder 152 may be omitted, as shown in Figure 32. It will be noted that since the internal reflector 150 reflects light equally in all directions, some of the light can be lost into the dead space where portions of the outer reflector 152 are omitted. In order to compensate for this and to utilise the entirety of the laser beam, the profile ofthe reflector 150 can be modified from a cylindrical cone to two back-toback part conical sections as schematically shown in Figure 33. This relies on the fact that if the beam is projected onto the side of the cone rather than onto the central axis the reflected light does not form a disc but in fact will generate an arc. Thus by selecting these part conical selections, all of the laser light can be directed towards the corresponding reflective surfaces of the outer reflector. A collimator may be included between the beam expander and the deflector in order to ensure that the beam is truly parallel and angular when it meets the focusing array.

Figure 34 schematically traces the ray paths through an embodiment of the beam expander where the mirror surfaces 150 and 152 are set at angles of 48" and 50 , respectively, from the horizontal. It can be seen that if the incident beam is not collimated, but in fact converges onto the central cone, the resulting annulus of light is not focussed. A subsequent lens 200 may be used in conjunction with the beam expander to direct the beam back towards the optical axis of the system.

A modified beam expander is shown in Figure 35. This comprises a central conical mirror 150 in conjunction with secondary reflective surface 202 in the form of a concave mirror with a central hole. By moving the position of the central cone 150, a variable divergence angle could be achieved thus offering variable resolution. Furthermore, depending on the shape of the mirror surface a constant focal length can be maintained, or alternatively, the secondary reflector 202 may be designed to produce a beam expander having a controllable focal length.

A combination of moving the central cone and a change in the profile of the mirror surface would enable both variable focus spot size and variable focus length to be achieved. The mirror 202 may, for example, be stepped to offer a switchable range of settings. Alternatively, by projecting several concurrent wavelengths of light, a change in focal length of only a few microns could be achieved. This might then be projected into a phosphor with a high hysteresis trigger deforming a 3-D optical memory. The beam expander may also be used as a focussing element within a pain ablation device.

Figure 36 schematically illustrates another version of a beam expander in which both the central primary reflector 206 and the annular secondary reflector 208 have curved reflecting surfaces. Such an arrangement may be used to reverse the laser light intensity distribution within a laser beam (which normally follows a l/e2 intensity distribution). As shown in Figure 36, the beam expander causes the paths of light from the central and circumferential portions of the beam to cross over, thereby causing the exit beam to have a peak intensity at the outer edge 210 of the annulus. Such a beam expander might also find application in laser drilling and/or cutting of ultra small holes where extra beam power at the edge of the hole may result in a sharper edge. Each of the beam expander designs described hereinbefore may also be implemented by relying on total internal reflection. The highest optical efficiency is achieved when both the mirror surfaces are implemented using total internal reflection. Figure 37 schematically illustrates a beam expander 220 formed from one or more blocks of transparent material, such as glass or plastics, and having inner and outer surfaces 222 and 224 each defined, over wholly or partially, by a conical section. In practice, the internal conical surface 222 may be difficult to machine. The surface 222 could be moulded. However, the constraint of manufacturing a fine point 226 may lead to difficulties in using a moulding process. In any event, the desirability of using an output beam having two opposed arcuate portions can more easily be achieved by forming the beam expander as a split element.

As shown in Figures 38a to 38d, the split cone beam expander is formed by initially by manufacturing the complete disc having internal and external conical surfaces, as shown in Figure 38a. The disc is then cut into two halves and a portion removed from each half, as shown in Figures 38b and 38c. Finally the two halves are bonded together as shown in Figure 38d. This enables the fine conical point on the internal surface to be formed, it ensures that the radii of curvature on the inner and outer surfaces of each half can share a common optical axis, thereby stopping the exit beam from being divergent, and it also enables the entire power of the incident light beam to be directed into two arcuate output beams. An optically flat top surface may be adhered to the uppermost surface of the beam splitter if desired. Similarly, one or more lenses may be placed adjacent or abutting the output side of the beam splitter.

The angles of the internal and external cones may be adjusted relative to one another. For example, the secondary reflecting surface may be inclined more steeply than the first reflecting surface thereby directing a disc of light back towards the optical axis of the beam splitter.

It is thus possible to provide a beam expander having a variable exit diameter and in which the central portion of the output beam is masked, thereby avoiding optical coma. It should be noted that the beam expander may be used in reverse in order to reduce the size of an optical beam.

Typically, scanners having a writing capability modulate a laser beam using an acousto-optic crystal. The crystal is caused to vibrate in response to an electrical signal and the vibrations are used to modulate an output beam by virtue of Bragg scattering. The modulators are well known in the art and need not be described further.

The output of the modulator comprises two beams which are only slowly diverging, ie the angle of divergence between the beams is typically between 1.5 and 1.8 . A first one of the beams is highly modulated, whereas a second one of the beams is not significantly modulated and needs to be removed. Prior art beam splitters have included relatively long optical paths thereby enabling the beams to diverge sufficiently for them to be separated by lenses. The applicant has realised that total internal reflection can be used to separate the beams.

Whether or not a light beam incident on a surface undergoes total internal reflection or not depends on whether the angle of incidence is less than or greater than the critical angle for that surface. The critical angle provides a very sharply defined boundary. Figure 40 schematically illustrates a beam splitter in which first and second beams B1 and B2 from a modulator are directed into a prism 350. The beam B2 is the modulated beam which it is desired to maintain. A surface 352 ofthe prism is angled such that the beam B2 undergoes total internal reflection thereat, whereas the beam B1 does not, and exits from the prism 350. If desired, a second prism 354 may be placed adjacent the prism 350 in order to modify the path of the beam B1 such that it continues in substantially the same direction as it was travelling when it entered the prism. The arrangement shown in Figure 40 utilises internal total internal reflection.

As shown in Figure 41, the interface between two surface 356 and 358 of a prism 355 may be arranged to fall between the two beams B1 and B2. One of the faces may be aluminised if it is desired solely to form a mirror, as shown in the drawing. In any event, the properties of the prism enable the two beams to be separated.

It is thus possible to provide a highly effective beam splitter.

It is also possible to provide a scanning arrangement including a field lens.

The beam expander may also be used in conjunction with light sources, such as high power light emitting diodes or laser diodes in order to enable them to be used, in some applications, in place of a laser itself. The controllable beam width and focussing power of some versions of the beam expander enable it to be used to focus a spot onto an undulating surface, for example the skin of an aircraft, in order to ablate the paint therefrom.

The use of a field lens reduces astigmatism. However other techniques may be used to control astigmatism in place of or in conjunction with the field lens. Thus spot shaping may be performed at the beam expander and/or the deflector lens may be arranged to reduce the spot size at the extremes of the scan.

It is thus possible to provide a refractive deflector array capable of achieving high scan rates, good duty cycle and good linearity.

APPENDIX 1. The general exponential expression y = era needs to be modified to the form:

where the effects of c and Stare shown in Figures 42a and 42b. In Figure 42a, each curve shown has a different valve of C, which alters the rate of curvature. In Figure 42b, the five curves of Figure 39a are plotted with two different valves of Sf. Sf varies the scale of the curve, but does not vary it's overall shape. The curves pass through the origin (0,0).

2. The gradient V of this function is given by:

This expression determines the x position on the exponential curve for which any predefined gradient will be found.

3. For the above exponential curve to be superimposed onto the ellipse, the x and y orig need to be defined.

Xshift = Xexp - Xellipse Yshift = Yexp - Yellipse

Y= - B- B2B2X2 Y =B - B2 B2x2 =ellipse 12cllipre (3) exp f B2 ellipse ellipse ellipse A 2 where A and B are the ellipsity constants.

4. Combined equation

le cXUp+Xlhi -1 - Yshat (4) YeXp = 5fkLf 1J Ysh (4) 5. Intersecting the ray and the exponential ;function; The ray coming through the lens is a simple linear expression Y = ax +b where a is the tangent of the ray angle.

This intersects with the previously determined exponential fUnction, equation 4.

| ss (X+X5iOl t (tan OrayX) + b = St;Lsf ] -lJ-Yshif c (x + Xshift) tan#.x + b + Yshift Sf = > + 1 = e Sf c (x + X shift) b Y shift Sf ax = > + + 1 = e - Sf Sf Sf

c.Xshift let d = Sf cx ax b Yshift Sf Sf = > + + 1 = de Sf Sf This can be simplified since most terms are known b Yshift c if + + 1 = Ex B = Sf Sf Sf and C = ax Sf then Ex = debx - cx.

Claims (34)

1. CLAIMS 1. A deflector, characterised by comprising at least one refractive optical element (18, 30,

   56) arranged, in use, to undergo rotation relative to a central axis such that the or each refractive optical element (18, 30 56) experiences relative movement with respect to the path of the light beam.
2. 2. A deflector as claimed in claim 1, characterised in that a plurality of optical elements are provided in a rotationally symmetric array.
3. 3. A deflector as claimed in claim 1 or 2, characterised in that the refractive optical elements are lenses (18, 30, 56).
4. 4. A deflector as claimed in claim 3, characterised in that the lenses have a plane surface.
5. 5. A deflector as claimed in claim 3 or 4, characterised in that the lenses are plano-convex.
6. 6. A deflector as claimed in any one of the claims 1 to 5, characterised in that the lenses have an elliptical or modified elliptical refracting surface.
7. 7. A deflector as claimed in claim 6, characterised in that lenses have modified elliptical surfaces, the surfaces being divided into a central zone and at least one peripheral zone having a modified form so as to control the shape of a focal plane of the deflector.
8. 8. A deflector as claimed in any one of the preceding claims, characterised in that light is focussed by the deflector as it passes there through.
9. 9. A deflector as claimed in any one of the preceding claims, characterised in that the deflector is, in use, arranged to receive light from a substantially stationary virtual point of origin.
10. 10. A deflector as claimed in claim 1, characterised in that the deflector comprises an circular array of plano-concave lenses with their concave surfaces facing radially inwardly.
11. 11. A deflector as claimed in claim 3 or any claim dependent thereon, characterised in that the lenses are cylindrical lenses.
12. 12. A deflector as claimed in claim 3 or any claim dependent thereon except claim 12, characterised in that the lenses are circular type lenses.
13. 13. A beam expander characterised by first and second co-operating reflectors, at least one of the reflectors having an annular or arcuate reflecting surface.
14. 14. A beam expander as claimed in claim 13, characterised in that the reflectors utilise total internal reflection of light within an object.
15. 15. A beam expander characterised by having at least on reflector utilising total internal reflection.
16. 16. A beam expander arranged, in use, to provide an exit beam from which a first portion is omitted.
17. 17. A beam expander as claimed in claim 16, characterised in that the first portion is a central portion of the exit beam.
18. 18. A beam expander as claimed in claim 17, characterised in that the beam expander comprises first and second co-operating optical elements.
19. 19. A beam expander as claimed in claim 18, characterised in that the first and second optical elements are mirrors.
20. 20. A beam expander as claimed in claim 19, characterised in that the first mirror comprises one or more of a conical reflector, two conic portions arranged back to back, a plurality of conic portions arranged in an array, or a plurality of mirror elements arranged in an array.
21. 21. A beam expander as claimed in claim 19 or 20, characterised in that the second mirror comprises one or more of an annular mirror, an annular portion of a parabolic or similar section of a mirror, a plurality of such annular or parabolic portions, or a plurality of part annular or parabolic portions having their virtual optical axis matched with an optical axis of an associated portion of the first mirror.
22. 22. A beam expander as claimed in claim 19, characterised in that the mirror is formed by total internal reflection within a transparent material.
23. 23. A scanner comprising a deflector as claimed in any one of claims 1 to 12.
24. 24. A scanner as claimed in claim 23, further comprising a beam expander as claimed in any one of claims 13 to 22.
25. 25. A scanner as claimed in claim 23 or 24, further comprising a field lens intermediate the deflector and a focal plane of the scanner.
26. 26. A scanner comprising a beam deflector, a focal surface, and a field lens intermediate the beam deflector and the focal surface, the field lens reducing the deviation from the local normal to the focal surface at which the light beam impinges on the surface.
27. 27. A scanner as claimed in claim 26, characterised in that the focal surface is a focal plane.
28. 28. A scanner as claimed in claim 26, characterised in that the focal surface is a virtual surface.
29. 29. A scanner as claimed in claim 23 further comprising a laser light source.
30. 30. A scanner as claimed in claim 23 further comprising a diode light source.
31. 31. A deflector as claimed in any one of the claims 1 to 12, characterised by further comprising means for Z plane focussing.
32. 32. A method of modifying the surface profile of a focussing element so as to obtain a desired focal surface when the focussing element is moved with respect to a light beam passing there through comprising the steps of: 1. defining the width of the beam; 2. analysing the range of relative motion between the light beam and the focussing element over a working range in order to identify a first portion of the lens through which a portion of the light beam always passes, this region forming a first unmodified portion of the lens; 3. defining the shape of the desired focal surface; 4. iteratively for a series of relative positions of the lens with respect to the light beam calculating the paths oftwo spaced apart rays ofthe light beam as they pass through the lens, calculating their point of intersection, modifying the profile of a portion of the lens outside the first unmodified portion, recalculating the ray paths and accepting those modifications which reduce the distance between the focal point and the desired focal plane; and 5. repeating step 4 until the actual focal plane approximates the desired focal plane to within an acceptable defined tolerance.
33. 33. A beam splitter arranged to separate two beams which are slowly diverging, characterised in that the beams are reflected from a surface positioned such that a first beam experiences reflection at the surface and such that a second beam does not.
34. 34. A deflector as claimed in claim 1, comprising a deflector lens having an outer surface which is substantially circular, elliptical or a cycloid curve with a step therein.