The arc in an optical system is usually the brightest point in the system. For many applications including projection displays, the brightness of the arc plays an important role in determining the brightness of the projected image at the screen and the size of the imaging panel that can be used. Brightness cannot be increase outside the arc, but had been increased in many applications by spherical retro-reflectors, in which the light collected by the retro-reflector is focused back into the arc itself, thus increasing the brightness of the arc. Using the unique property, unit magnification, of the Dual Paraboloid Reflector (DPR) system, the arc can be manipulated in many different ways and the arc size is preserved after multiple reflections using DPR. The more times the image of the arc passes through the arc itself, the brighter the arc is. This increase in brightness allows a longer arc, lower brightness lamp to be used in higher brightness applications. In this paper, several configurations will be described where the brightness of the arc can be increased. Ray tracing results will also be presented.

A novel backlight concept suitable for LED's has been designed using the flow-line design method, which allows controlling both the illumination uniformity and light extraction without scattering the light. This contrasts with conventional LED backlight optical designs, which are based on the use of a light guide with Lambertian scattering features that break the guidance and extract the light. Since most of Lambertian scattered light is re-guided, the average ray path in conventional backlights is long and multiple bounces are needed, which may lead to low efficiency. On the other hand, the new design presented here is not only efficient but also provide a relatively high collimation of the output beam (an output beam within a 10 degrees half-angle cone, with total theoretical efficiency over 80% including Fresnel and absorption losses). Wider beams can be controlled by design or obtained by adding a holographic diffuser at the exit. The new design offers other very interesting practical features: it can be very thin, can be made transparent (which widens its applications, including front lighting), can mix the colors from several LED's or recover reflected polarization for LCD illumination.

An afocal system keeps parallel any two parallel rays emitted by a pair of point sources. If they have the same intensity at identical angles, this design will ensure that the far-field intensity will be higher than that of the bare sources as well as giving same the intensity to both, so that color balance is preserved. If these sources are of different color, it is possible to mix them in the far field while increasing the intensity in a prescribed way. If a third LED is placed at the midpoint between the other two, its intensity pattern will still be close to the one created for the other two souces. This enables the pixel to color-mix red, green and blue LEDs while increasing their apparent intensity and preserving the color-balance of the video signal across the far field. This enables large active-video screens to redirect and intensify their light towards the audience, instead of just spreading it out uselessly.

In this paper, we investigate how to improve the uniformity of the spatial distribution of the illuminance at the output plane for angle-to-area-converting, light-piping systems through the introduction of cyclical surface features. A superposition approach is used for studying uniformity. Improvements in uniformity for square-to-circle and rectangle-to-circle lightpipe configurations are demonstrated for a short package length.

Lightpipes are used to transfer light from the source to a desired target. The lightpipe shape typically conforms to the necessary path, thus bending of the lightpipe is required. A number of different methods of bending the lightpipe have been developed, from linear, discontinuous bends to smooth, common circular bends to bends that expand or contract the cross-sectional size of the lightpipe over the path. In this paper we develop a set of parameters to describe the overall shape of an in-plane lightpipe section. These parameters include the thickness, radius of bend, index of refraction, and ratios of sets of these parameters. The transfer efficiency from the source to target is used to quantify the utility of parameterized lightpipes. Etendue is used to highlight the results. More complex lightpipes, such as those with several bends along their path, can be developed from the combination of parameterized sections. Finally, this parameterization can be used to automate the development of lightpipe geometry within optical analysis and design software.

The edge ray theorem has become a tenet for illumination design. For idealized configurations this theorem has proved extremely powerful and has led to new methods of design. However, many real life sources have no clean delimitation and thus the boundary of the phase space is not easy to specify. Optical efficiency expressed by the fraction of the light captured by the optical system and transferred to the target on one hand and average radiance at the target on the other hand become two conflicting objectives in illumination design. Optimal design generally requires a single scalar merit function. Two different objectives require a metric which specifies the relative valuation before optimality may even be defined. In practice the relative valuation depends on the application and seldom may be specified in advance. It is, however, possible to refer to a slightly weaker criterion, called pareto-optimality which allows to postpone the decision on the relative valuation. In this contribution we show how to derive the upper limits of efficiency and average radiance for all paretooptimal illumination designs using a given source. If an optical system observes the edge ray principle such that the boundary of the source follows a iso-radiance hyper-surface in phase space, then this system is pareto-optimal. To calculate the characteristic curve one needs to sort phase space according to radiance. We show an example of an illumination lens which was inspired by phase space sorting. We have found this tool extremely useful because it allows to outline the range of options before actually embarking on the optical design.

Numerous optical and electromagnetic applications require numerical design of reflecting surfaces in 3D with capabilities to redirect the input energy flow and reshape the energy radiation intensity of a source into a prescribed output irradiance distribution over a specified target surface. In the geometrical optics approximation, a systematic application of the ray tracing equations and energy conservation law reduces the problem, in many cases, to finding numerical solutions to nonlinear, second order partial differential equations. If the severe limitation of rotational symmetry is not assumed then the resulting equations are very far from being standard and require significant efforts for their theoretical investigation and reliable numerical solution. In recent years a quite general approach combining geometric techniques with methods from calculus of variations has been developed and applied to a rigorous and unified investigation of several classes of such equations. Moreover, this approach allows implementations in provably convergent numerical algorithms. In this paper I outline this approach in the problem of designing a reflecting surface capable of redirecting the energy flow from a point source so that the reflected rays have directions specified in advance as a subset on the far-sphere and the output irradiance density is also pre-specified in advance as a function of the reflected direction. A numerical example illustrating the solution is also presented.

The design of illumination lenses is far easier under the regime of the small-source approximation, whereby central rays are taken as representative of the entire source. This implies that the lens is much larger than the source's active emitter, and its entire interior surface is nowhere close to the source. Also, a given source luminance requires a minimum lens area to achieve the candlepower necessary for target illumination. We introduce two-surface aspheric lenses for specific illuminations tasks involving ceiling-mounted downlights, lenses that achieve uniform illuminance at the output aperture as well as at the target. This means that squared-off lenses will produce square spots. In particular, a semicircular lens and a vertical mirror will produce a semicircular spot suitable for gambling tables.

Successful design of LED light sources requires application of principles of non-imaging optics and their integration into optical design software. We propose some design tools for CPC-like concentrators, which are flexible to cover both different LED types and diverse applications of LED for illumination. The methods are based on the optimization using standard ray-tracing either according to the edge-ray principle (with special ray aiming and integration over the projected images) or a Monte-Carlo procedure (with number of rays, sufficient for resolving concentrator features). The algorithms take skew rays into consideration, what results in 5...15% higher efficiency compared to concentrators designed according to the theoretical CPC equation. The specifics of optimization algorithms (compared with conventional imaging optical design algorithms) are discussed. We also propose different optimization criteria, specific for diverse light source configurations. Examples of solid and hollow concentrators, optimized for efficiency or light intensity were manufactured by single point diamond turning. Measured characteristics of devices are in good agreement with the design goals.

New concerns regarding the unwanted effects of lighting our nighttime environment (i.e., light pollution), and new efforts to design energy efficient roadway lighting installations, are changing the requirements of roadway lighting. These factors necessitate the need for different luminous intensity distributions with greater optical control than have traditionally been produced by roadway lighting fixtures. New requirements affecting the optical design of roadway lighting are presented. An example of a nonimaging optical design solution that addresses these requirements is given. Conclusions are then made about the challenges that lie ahead in the optical design of roadway lighting fixtures.

The lighting systems of a car provide a variety of challenges from the point of view of illumination science and technology. Engineering work in this field has to deal both with reflector and lens design as well as with opto-mechanical design and sensor technology. It has direct implications on traffic safety and the efficiency in which energy is used. Therefore, these systems are continuously improved and optimized. In this context, adaptive systems that we investigate for automotive applications gain increasing importance. The properties of the light distribution in the vicinity of the cut-off line are of key importance for the safe and efficient operation of automotive headlamps. An alternative approach is proposed to refine the description of these properties in an attempt to make it more quantitative. This description is intended to facilitate intercomparison between different systems and/or to study environmental influences on the cut-off line of a system under investigation. Designing projection systems it is necessary to take a delicate trade-off between efficiency, light-distribution characteristics, mechanical boundary conditions, and legal requirements into account. Considerations and results on optical properties of three-axial reflectors in dependence of layout parameters will be given. They can serve as a guideline for the optical workshop and for free-form optimization.

A new design method of free-form Kohler integrator array optics is presented. Only few solutions to the integrator design problem are known, which apply for specific and simple source and targets (for instance, flat integrator lenslet arrays when the source and target are squares located at infinity). The method presented here find more general solutions and the resulting optics is formed by two arrays of free-form optical surfaces (which can be either reflective of refractive). The contour curves of the array units are also obtained from the design. Two types of Kholer integrators will be defined, depending if they integrate only along one direction across the source (one-directional integrators) or in the two directions (two-directional integrators). This design method has been applied for an ultra-compact high efficiency LED low beam head lamp producing a legal pattern independently of the chip luminance variation and permitting LED position tolerances of ±200 microns. The ray tracing proves that the high gradient (0.32) and its vertical position in the pattern remain invariable when chip is moved.

One of the most challenging applications for high brightness LEDs is in automotive headlights. Optical designs for a low or high beam headlights are plagued by the low flux and luminance of LEDs compared to HID or incandescent sources, by mechanical chip placement tolerances and by color and flux variations between different LEDs. Furthermore the creation of a sharp cutoff is very difficult without baffles or other lossy devices. We present a novel LED headlight design that addresses all of the above problems by mixing the light of several LEDs first in a tailored light guide called LED combiner, thereby reducing color and flux variations between different LEDs and illuminance and color variations across the LED surfaces. The LED combiner forms a virtual source tailored to the application. The illuminance distribution of this virtual source facilitates the generation of the desired intensity pattern by projecting it into the far field. The projection is accomplished by one refractive and one reflective freeform surface calculated by the 3D SMS method. A high quality intensity pattern shape and a very sharp cutoff are created tolerant to LED to optics misalignment and illuminance variations across the LED surface. A low and high beam design with more than 75% total optical efficiency (without cover lens) and performance as latest HID headlights have been achieved. Furthermore it is shown that the architecture has similar tolerance requirements as conventional mass produced headlights.

This paper documents the development of the Multi-Mode External Lighting System for Aircraft (M²ESA), a solid-state near-IR and visible light emitting diode-based programmable system designed to replace existing incandescent navigation lights on the exterior of military aircraft, and tailored for use with night vision goggles. Integrated systems of optics, electronics and mechanical structures were designed that were compatible with legacy aircraft systems, and which thus conformed to rigid configuration requirements and severe volume constraints. The genesis of the concept, evolution and general architecture of the system, top-level performance and environmental requirements, integration on the designated aircraft platform (the F-15), and general results of flight demonstration assessments are described.

When developing an LED illumination system, the designer is often restricted to a constrained working volume. This can lead to efficiency loss, thermal issues, and performance restriction. It becomes important to understand the etendue of the source and optics. Also, the optics should be designed so as to maximize the efficiency of the system. Along with discussion of these issues, a case example will be presented where incandescent position lights on the F-15 fighter are replaced with LED systems that have both visible and near infra-red functions.

We consider the problem of designing an ultra-high-numerical-aperture rotationally symmetric monochromatic imaging concentrator that could be used for optical detection and tracking. This problem combines requirements for both high-efficiency flux transfer and low focal-plane aberrations. A solution has been previously demonstrated using the 2D SMS technique, resulting in the RX concentrator design. We review the RX concentrator and analyze its performance. We then describe a numerical test of the RX concentrator for optimality, performed using a global optimization procedure. We present the results of this test, which produced a counter-example possessing significantly improved imaging and nonimaging performance. The solution space was then extended to a globally optimized cemented-doublet comprising three aspheric surfaces, of which two are refractive and the third reflective. Compared with the original RX concentrator, the optimized doublet provides a threefold reduction in the mean aberration level within the field of view, combined with a significant improvement in flux-transfer efficiency, without any increase in the axial lens thickness.

Recently proposed aplanatic imaging designs are integrally combined with nonimaging flux boosters to produce an ultra-compact planar dielectric-filled concentrator that performs near the étendue limit. Such optical devices are attractive for high-efficiency multi-junction photovoltaics at high flux, with realistic power generation of 1 W from a 1 mm² cell.

We explore compact, imaging (aplanatic) designs capable of reconstituting the power density of ultra-bright lamps at a remote target, with specific application to optical fiber light transport. The solutions presented here, for concentrating incoherent light from an extended near-field source, closely approach the thermodynamic limit for optical performance at high collection efficiency, as confirmed with raytrace simulations. Our investigations are motivated by the prospect of arc-discharge lamps as effective alternatives to lasers for many surgical procedures where high intensity heating rather than monochromaticity is needed. Additional applications include LED-fiber and fiber-fiber coupling, as well as projection systems. Consideration is restricted to pure reflective (mirrored) systems since refractive elements can introduce inadmissibly large chromatic aberrations. The tailored aplanatic mirror contours constitute monotonic functions that can be solved analytically, which facilitates rapid surveying of a wide range of design options. Several near-field designs suitable for coupling high numerical aperture light sources into optical fibers are presented.

A reflector design method that follows the standard reflector design principle but uses equations to match incident angles and reflected angles is presented. Using this method, errors introduced by traditional graphic approach are minimized if not eliminated in many cases. The resolution of the reflector profile is no longer limited. A calculation method that is capable of generating reflector profiles in (x,y) form is also presented. The methods and all results are presented for symmetric reflector configurations.

Efficient homogeneous illumination of rectangular or circular areas with LEDs is a promising application for doublesided microlens arrays. Such illumination schemes employ a primary optics - which can be realized with a concentrator or a collimation lens - and a secondary optics with one or more double-sided microlens arrays and a collection optics for superposing the light from the individual array channels. The main advantage of this design is the achievable short system length compared to integrating lightpipe designs with subsequent relay optics. We describe design rules for the secondary optics derived from simple ABCD-matrix formalism. Based on these rules, sequential raytracing is used for the actual optics system design. Double-sided arrays are manufactured by polymer-on-glass replication of reflow lenses. With cylindrical lens arrays we assembled high-brightness RGB-illumination systems for rectangular areas. Hexagonal packed double-sided arrays of spherical lenslets were applied for a miniaturized circular spotlight. Black matrix polymer apertures attached to the lens array helped to avoid unwanted straylight.

Conventional linear light sources irradiate cylindrically, into 360° in a circle orthogonal to the lamp axis. This makes it difficult to optically couple lamp-output to a much narrower range of output directions. Light-emitting diodes, however, emit into a hemisphere or less of solid angle, and thus are much more amenable to proper lensing. Recent LED tape systems have ±60° emission angles, due to recessing the emitting chips in reflector cups. This turns out to be highly beneficial to linear lenses, enabling nearly 100% collection efficiency. Here we present a new class of nonimaging linear lenses with two aspheric profiles, designed according to the amounts of source-light received along longitudinal strips on the interior surface of the linear lens. In the small-source approximation, the light falling on each strip is imaged relatively intact in the far field, although the source itself is not. This enables the design of extruded linear lenses achieving uniform illumination of nearby planar targets, as in shelf and cove lighting.

Phosphor-conversion (PC) LEDs are the leading type of white solid-state lighting (SSL), due to the high efficacy of the yellow wavelengths of blue-stimulated photoluminescence. Conventional phosphor-conversion LEDs have the photoluminescent phosphor in immediate contact with the blue LED. Major types are the thin conformal phosphor and the thick or in-cup phosphor. The first trades away efficiency for increased luminance, while the latter gains efficiency at reduced luminance. In both cases the phosphor suffers from the elevated temperature of the blue chip, particularly the thermal quenching that reduces phosphor quantum efficiency. Also, the inevitable 15% Stokes heat of the phosphor conversion of blue light to longer-wave yellow light adds to the chip's heat load, as does much of a conformal phosphor's back-emission into the chip. It would be preferable to relocate the phosphor away from the chip illuminating it. Although remote phosphors have recently been showcased, their phosphor is much larger than the chip, greatly reducing luminance. A new design is presented of a Dual-Optic-based remote phosphor configuration with minimal increase in phosphor etendue over that of the source, as well as greatly improved spatial uniformity. Moreover, the yellow phosphor back-emission is recycled with a blue-pass mirror that re-illuminates the phosphor to increase its luminance. The result is a new white-light source with superior luminance, efficacy, and uniformity.

Nonimaging optics needs to address the interesting effects upon white-LED luminance of scattering within a photoluminescent phosphor, and how strong scattering leads to luminance recycling of TIR-trapped phosphor-emission. This paper analyzes LED optical systems that extract light by multiple internal reflections and varying degrees of bulk scattering. The luminance values of such devices can greatly exceed those predicted by the luminance-conservation law of etendue, formulated for non-scattering, non-recycling optical architectures. To illustrate this, the results of extensive modeling of LED architectures via a commercial raytracing package are described and analyzed. The analysis includes the effects of bulk scattering within the phosphor, and reveals the crucial role of diffuse reflectance, within the LED itself below its emitting layer. The study shows how it is possible to achieve an increase luminance in an LED via use of flat-windows over an LED as opposed to the traditional approach of dome-covers, albeit with some loss of overall luminosity extraction. The paper includes a discussion of luminance-luminosity tradeoffs and a summary of analytical and numerical methods for modeling optical systems involving bulk scattering.

Amonix has become the first company to begin production of high concentration silicon solar cells where volumes are over 10 MW/year. Higher volumes are available due to the method of manufacture; Amonix solely uses semiconductor foundries for solar cell production. In the previous years of system and cell field testing, this method of manufacturing enabled Amonix to maintain a very low overhead while incurring a high cost for the solar cell. However, recent simplifications to the solar cell processing sequence resulted in cost reduction and increased yield. This new process has been tested by producing small qualities in very short time periods, enabling a simulation of high volume production. Results have included over 90% wafer yield, up to 100% die yield and world record performance (η =27.3%). This reduction in silicon solar cell cost has increased the required efficiency for multi-junction concentrator solar cells to be competitive / advantageous. Concentrator systems are emerging as a low-cost, high volume option for solar-generated electricity due to the very high utilization of the solar cell, leading to a much lower $/Watt cost of a photovoltaic system. Parallel to this is the onset of alternative solar cell technologies, such as the very high efficiency multi-junction solar cells developed at NREL over the last two decades. The relatively high cost of these type of solar cells has relegated their use to non-terrestrial applications. However, recent advancements in both multi-junction concentrator cell efficiency and their stability under high flux densities has made their large-scale terrestrial deployment significantly more viable. This paper presents Amonix's experience and testing results of both high-efficiency silicon rear-junction solar cells and multi-junction solar cells made for concentrated light operation.

We report results of ultra-high-flux experiments on tandem and triple-junction solar cells, with a real-sun probe predicated on mini-dish fiber-optic concentrators. We focus on the sensitivity of cell efficiency to a wide range of flux levels and distributions. Our experiments also revealed pronounced reversible photovoltaic hysteresis at high concentration, and provide a non-destructive method for assessing tunnel diode characteristics.

The materialization of a recent conceptual advance in high-flux photovoltaic concentrators into first-generation prototypes is reported. Our design strategy includes a tailored imaging dual-mirror (aplanatic) system, with a tapered glass rod that enhances concentration and accommodates larger optical errors. Designs were severely constrained by the need for ultra-compact (minimal aspect ratio) modules, simple passive heat rejection, liberal optical tolerances, incorporating off-the-shelf commercial solar cells, and pragmatic considerations of affordable fabrication technologies. Each unit has a geometric concentration of 625 and irradiates a single square 100 mm² triple-junction high-efficiency solar cell at a net flux concentration of 500.

A novel photovoltaic concentrator has been developed in the framework of the European project "High efficiency silicon solar cells concentrator". In this project, front-contacted silicon solar cell have also been designed and manufactured by the project leader (the French LETI). This silicon cell concept is potentially capable to perform well (24% efficiency has been predicted) for much higher concentration levels than the back-contacted cells (and, of course, than the two-side contacted cells). The concentrator is formed by one lens of squared contour with flat entry surface and large-facet Fresnel exit surface, and a secondary that encapsulates the solar cell. On the contrary to the conventional Fresnel lens plus nonimaging secondary concentrators, the primary and secondary are designed simultaneously, leading to better concentration-acceptance angle product without compromise with the compactness. The grid lines in the front-contacted cells are aluminium prisms (which contact the p+ and n+ emitters, alternatively), acting as a linear cone concentrator that concentrates Cg =1.52× in the cross sectional dimension of the prisms. The secondary concentrator has a refractive rotational symmetric top surface that is crossed with two linear flow-line TIR mirror. Then, in the cross section normal to the aluminium prisms, the secondary provides a 2D concentration of Cg =12×, while in the cross section parallel to the prisms it provides a 2D concentration of Cg =24.16× as the grid lines in this dimension. Therefore, the cell is rectangular (1:2.08 aspect ratio), being the grid lines parallel to the shorter rectangle side. The total 3D geometrical concentration is 24.16×(12×1.52) = 455× for the square aperture and rectangular cell, and gets a design acceptance angle α=±1.8 degrees. Injection moulded prototypes are have been manufactured and measured, proving an optical efficiency of 79%. Computer modelling of the concentrator performance will also be presented.

Concentrating photovoltaics have the greatest performance for high solar efficiency, but current designs feature large high-profile tracking platforms that are expensive and laborious to install and maintain, and are vulnerable to even moderate winds. Moreover, their air cooling can only handle relatively low concentrations, and then only at elevated temperatures that can only shorten operating life. A previously described new approach has been implemented in a prototype that has been operating since last winter. This is an array of 450X concentrator utilizing 15mm-square water-cooled triple-junction cells, in lensed troughs that utilize flotation both for structural support and azimuth tracking. The troughs tilt as low as 24° solar altitude, for elevation tracking. The troughs float in only 14" of water, in circular ponds that can be close packed to cover 70% of the land with lenses, one-two orders of magnitude more effective land coverage than conventional concentrating systems, needing only two square miles per peak GigaWatt. The sealed troughs are mass-produced off-site and quickly installed by a small crew without any cranes or heavy machinery. Current costs are already under $2/Watt, and should drop in half for large-scale city-powering applications. The lens and auxiliary optics are discussed, and the water cell-cooling system is thermally analyzed.

Spectroscopic measurements have been undertaken for a range of different quantum dot (QD) types and transparent host materials for use in a novel solar energy-concentrating device, a Quantum Dot Solar Concentrator¹ (QDSC). A QDSC comprises QDs seeded in materials such as plastics and glasses that are suitable for incorporation into buildings where photovoltaic cells attached to the edges convert direct and diffuse solar energy into electricity for use in the building. High transparency in the matrix material and QDs with a large Stokes shift are essential for an efficient QDSC. An optimum matrix material for a QDSC has been determined based on absorption characteristics and an optimum commercially available QD type has been chosen using steady-state absorption, photoluminescence and photoluminescence excitation spectroscopy of QDs in solution and solid matrices.

The design purpose of general lighting apparatus is how to induce the illuminating direction of light source efficiently and provide people comfortable vision. The lighting device has a reflective area. However, the reflective light was often caused a glare and a light pollution. It causes eyes discomfort and eyestrain. This problem has been overcome by the novel device in this study. This paper proposes a novel LED light apparatus with innovative reflector type, designed to reduce the glare for light comfort. The apparatus is consist of a reflective wall with many circumferential micro-structured and a circumferential light source as plurality of LED. In addition, a Fresnel type reflector can replace the traditional convex reflector. Multiple LED light sources will be guided to the reflective wall and Fresnel type reflector with in device, so ray of light will be well spreaded and sufficiently directed to or focused on the desired area. The micro-structured reflection type will enhances the light illumination efficiency by applying appropriate design of micro-structured reflecting layer.

Design a Fresnel lens for a concentrator to collect more sunlight onto the solar cell due to the efficiency and cost. Since 1970, the non-imaging concentrator was used for solar energy; most of them were reflecting mirrors. The non-imaging optical system provides large aperture and forgiving imaging requirements. The Fresnel lens used in non-imaging optical system was usually called non-imaging Fresnel lens. In this research, the Fresnel lenses were refracting optical elements but diffracting ones. According to the method of Ralf Leutz and Akio Suzuki _[2], using minimum deviation and minimum dispersion to design a non-imaging Fresnel lens, which obeys the edge ray principle. Use optical software TracePro to simulate the non-imaging Fresnel lens, and each pitch size was 0.3mm and 200mm focus distant. Discusses the losses of non-imaging Fresnel lens and find out the relation of efficiency and F-Number. The optical concentration ratio could reach 15X (2-D) and 230X (3-D).

Illumination system is an important component of projection display system, and decides the brightness and uniformity of the whole system. This paper will present a description of a simplification design and discuss the performance of LED illuminators with respect to light output, uniformity, and color performance.

This paper represents comparison analysis of schemes for illumination channel in a light microscope - Kohler illumination, projection and critical illuminations. It is proved that these schemes are particular cases of the proposed common scheme for microscope illumination. It is calculated uniformity of irradiance distribution on sample surface produced by different particular cases of this common scheme with a halogen lamp or a uniform extended light source. It is showed that the high uniformity can be reached not only in Kohler illumination.

Canesta, Inc. has developed a CMOS based technique that yields 3D information using only one 2D array sensor and is commercializing this 3D camera. The camera is composed of a light source (illumination system) and a CMOS based 2D array camera. The illumination system can generate 4 types of illumination fields of view with IEC eye classification class 1 or class 1M by simply switching diffusers. In this paper, the design of the illumination system is described and the performance results are presented.