Illumination design has historically been more efficient to perform using prototypes than by using computer design tools. That situation has changed with better understanding of the limitations of illumination systems and dramatic improvements in computing power and simulation accuracy. In particular, software base optimization of illumination systems is now possible. Design trends and how they relate to making an illumination designer's job fun will be discussed.

This paper discusses the impact of light coherence on the defocusing properties of a novel retinal projection display. The display is based on a liquid crystal display (LCD) illuminated by partially coherent light from a LED which projects the LCD-image directly onto the eye's retina. It is shown that the increase of the coherence level of the illumination light enhances on the one hand the contrast of a defocused image. On the other hand, however, the perceived image quality is affected by the occurrence of intermediate defocusing peaks as well as by coherence effects, such as edge-ringing. Measurements of both the modulation transfer function (MTF) and of the text readability when defocusing are presented using a bench model of the display system. The experimental results reveal that for text readability the best DOF can be achieved not for fully coherent illumination, but for partially coherent light of a coherence level σ ≈ 0.35-0.5. Further, it is shown that the optimum σ-level depends slightly on the targeted text size.

This paper begins by reviewing the current state of development of LEDs, their existing markets as well as their potential for energy conservation and their potential for gaining market share in the general illumination market. It discusses LED metrics such as chip size, lumens per watt, thermal resistance, and the recommended maximum current rating. The paper then goes on to consider the importance of non-imaging optics for both optically efficient and extremely compact LED lighting systems. Finally, microstructures useful for controlling the fields-of-view of LED lighting systems are considered and described in some detail. An extremely efficient and cost effective microstructure, called kinoform diffusers, is shown to have very unique properties that make this technology almost ideal for shaping the output beams of LED lighting systems. It concludes by illustrating some general illumination LED lighting systems

Given a set of available LEDs or other light sources with known spectrum, total luminous flux (lumen) and efficacy (lumen/Watt), price etc. we show how to select that combination which yields light of desired (photometric) color and, in addition, maximizes efficacy, luminous flux , color rendering index or other objectives.

This study investigates illuminators composed of light emitting diode (LED) array sources and side-emitting light guides to provide efficient general illumination. Specifically, new geometries are explored to increase the efficiency of current systems while maintaining desired light distribution. LED technology is already successfully applied in many illumination applications, such as traffic signals and liquid crystal display (LCD) backlighting. It provides energy-efficient, small-package, long-life, and color-adjustable illumination. However, the use of LEDs in general illumination is still in its early stages. Current side-emitting systems typically use a light guide with light sources at one end, an end-cap surface at the other end, and light releasing sidewalls. This geometry introduces efficiency loss that can be as high as 40%. The illuminators analyzed in this study use LED array sources along the longitude of a light guide to increase the system efficiency. These new geometries also provide the freedom of elongating the system without sacrificing system efficiency. In addition, alternative geometries can be used to create white light with monochromatic LED sources. As concluded by this study, the side-emitting illuminators using LED sources gives the possibility of an efficient, distribution-controllable linear lighting system.

An analytical model of light propagation in rectangular light pipes is presented. Light pipe illumination systems are an efficient means of collecting, transporting, and distributing light. One area where light pipe illumination systems are successfully employed is in transportation display lighting, such as instrument panel illumination. In these applications the transportation industry takes advantage of injection molding to manufacture light pipe systems at relatively low costs. One historical drawback to using light pipe illumination systems is the design effort associated with iterative prototyping cycles and evaluation. The model presented here is a graphical method of describing ray propagation in light pipes. The model describes ray direction vector space in spherical coordinates. With this model, the angular "mode" of a ray can be defined at any plane in the system. The angular mode propagation space describes input coupling, flux transport, and output coupling in light pipe illumination systems. The model of flux propagation described here is therefore a tool to aid in "first order" design and layout of light pipe illumination systems.

For reasons both fluid-dynamic and stylistic, volumetric constraints on vehicular luminaires grow more exacting. For full design-freedom of luminaire placement and shape, new designs are needed that have shallow depth and are capable of emitting a beam that makes a net angle with the local surface normal. Automotive headlamps, fog-lamps, and daylight-running lamps may need to project their illumination patterns onto the road from a position on sloped front surfaces. A conventional paraboloid, however, must be recessed behind a sloped window, thus using up space inside the vehicle-skin. A conventional TIR lens, with its output beam centered on its axis of circular symmetry, will also have to intrude into the vehicle interior, and shine through a sloped window. Instead, the luminaire should be thin enough to mount on a vehicle’s skin without needing a hole to be cut into it, a luminaire also capable of emitting its beam substantially off the local normal. To this end, two new TIR lenses are introduced here that generate off-normal beams. In one, a circular TIR lens takes on an internal tilt of its symmetry axis to produce a collimated output beam with high tilt, nearly 45° from the surface normal of the lens exterior. In the other, an off-axis linear TIR lens can be made with an internal tilt to the reflected rays. When used with LEDs, this new linear lens can be combined with exterior transverse lenslets, tailored to meet an intensity prescription.

Injection-molded optical components are used often for commercial illumination systems. This paper discusses methods of how to model the tolerance aspects of such components. Tolerance aspects include surface roughness, source-to-optic position and rotation errors, and surface slope errors. It is noted that all of these tolerance investigations cannot correctly account for errors in the injection-mold process. A method to model deformations induced in the injection-mold process is proposed. The method is based on the laser scan of an injection-molded part, which allows the rebuilding of the surface from the point cloud. This method, while quite accurate, is time consuming, so a second algorithm based upon approximation with a Harvey scatter model is developed that takes over an order of magnitude less in time. It is shown that the approximate model provides results within a few percent if comparisons are done in the far field. Near-field results require the rebuild method that uses the measured point cloud. Additionally, illumination systems comprising multiple interactions with the component surface (e.g., lightpipes) can use the approximate Harvey model.

Techniques for illumination of Digital Micro-mirror Devices (DMD's) are reviewed and evolution of the illumination design discussed by example over three generations of system. The use of a TIR prism for illumination and imaging beam combining is explained and equations for definition of the prism block provided.

The color characterization of a DLP (digital light processing) PJ (projection) TV is studied by changing the picture modes. After measuring the chromaticity coordinates, luminance, and spectral radiance of red-green-blue and white as a function of the digital-analog convert value, the contrast ratio, white luminance, chromaticity constancy, channel independency are analyzed. And then colorimetric models for producing desired colors on DLP PJ TV are studied. The performances of GOG (gain-offset-gamma) model and 1-D LUT (look up table) model are evaluated by comparing Δ Eab between measured and predicted values for 56 target colors.

We propose and demonstrate a single DLP projection system with high illumination efficiency by the moving color stripe method. White light from the lamp is split and focused as color images by the color filter and lens cells of spiral lens wheel (SLW). Fly eye lens and relay lens superpose color bars on the light valve, and then 3 color strips are scrolled linearly by rotating SLW. As a result, the system output is evaluated as 1.7 times compared to a typical single panel DLP system.

Examples of realized sophisticated lighting products (daylighting devices and luminaires) will be presented, where complex systems are used. These systems are built up on reflective and refractive border surfaces (reflectors, lenses etc.). The surfaces are preferably composed on symmetrical and analytical geometries, whose optical behaviour is well known. In the case of complex surfaces they are computed point by point by special calculation methodes to fulfill special photometric requirements, which are not solvable with conventional design methodes.

Our paper treats the conversion of light beams with circular cross-section into light beams of square cross-section as well as the conversion of light beams with rectangular cross-sections of different aspect ratios. We calculate the possible concentration ratio,and introduce symmetry-breaking microstructures in order to mix high with low skewness without affecting the axial component of the k -vector. A typical example for its application is as secondary concentrator-homogenizer in 3D-photovoltaic (PV)concentrator systems,for square PV-cells. For the performance of modern multi-junction cells it is crucial to homogenize the incoming radiation in the secondary, both in location and color. We compare the performance of homogenizers with and without microstructures and show the advantage of adding this feature. Several designs are modeled and the performance is compared by Monte-Carlo ray-tracing. With wall microstructures,light uniformity and concentration is significantly better.

A novel luminaire utilizes repeated Fresnel reflections by angled surfaces to transform a small collimated input beam into a controlled output pattern with a high degree of polarization, either linear or radial. Applications to backlighting, front-lighting, optical communications and automotive lighting will be discussed.

3-D tailoring is a constructive method for the design of free-form optical elements for illumination. The light of a point source is redirected in a controlled manner to cast a prescribed irradiation pattern on a target surface. Free parameters can be used to control the shape of the surface resulting from the tailoring process. Every change in the parameters may lead to an entirely different design. Hence the choice of parameters is crucial for the technical feasibility and the visual appearance of the luminaire. Examples of free parameters are the chosen caustics, trimming of the surface, the choice between mirror and lens optics, and the mutual orientation of source and optical elements.

Due to antiquated technologies (calculation methods, regulations, lighting and luminaire concepts, production techniques) current outdoor lighting causes a lot of problems like light pollution, glare, energy waste etc. New types of luminaires, and in consequence new outdoor lighting concepts, can be created by combining advanced calculation methods for optical surfaces with recent production technologies and novel light sources such as short arc metal halide lamps. Light emitted from this small Etendue light sources can precisely be redirected by 3D-curved surfaces manufactured with injection molding, milling and aluminium metallization. The required optical design may use techniques like complex surface calculations and 3D-Tailoring. An innovative concept based on the latest findings in visual perception research is to focus the light of such short arc light sources onto a facetted secondary mirror which provides the desired illuminance distribution on a facade or a public place. These systems are designed to fulfill lighting requirements as well as providing visual comfort. Thus lamps with improved color rendering, luminous efficacy and increased lifetime are used and glare is minimized by splitting the reflector into many facets (light spot decomposition). A few examples of realized projects will be presented where such complex facetted surfaces are used to reach a special quality of light. Using novel techniques like 3D-Tailoring, each facet can be designed to individually create the desired (e.g. uniform) illuminance distribution on the target surface - in this case, a large facade. For this particular application, we chose to impose a square boundary for each facet, in order to tile the rectangular aperture of the secondary mirror without compromising efficiency.

With the increased interests in mass production of LCD, LCOS and DLP microdisplay-based projectors and televisions, the image panel sizes become smaller and requires more efficient coupling of light from the source to the image panel. At the same time, the demand for high-energy efficient general illumination system also requires efficient coupling of light from a light source into fiber optics. To illuminate these smaller image panels and fibers efficiently, a patented dual paraboloid reflector system has been developed to collect and focus light from an arc lamp onto the targeted application without loss of brightness. Arc lamps with longer arc lengths can be used, which are usually easier to make and have longer life. The dual paraboloid reflector1,2 system consists of two parabolic reflectors placed symmetrically facing each other. The first parabolic reflector collects and collimates light into a parallel beam. The second parabolic reflector intercepts the parallel beam and focuses the light into a lensed rectangular tapered light pipe (TLP) resulting in a unity magnification, i.e. 1:1 imaging, with conserved brightness. Due to the unique nature of 1:1 imaging of the system, together with the retro-reflector, the folding of the arc to increase brightness will also be described. The TLP transforms the focused light into an output with the needed area, shape, and numerical aperture. It also acts as a homogenizer so that the intensity profile at the output surface is uniform and eventually provides a uniform intensity profile at the screen or at the input face of the fibers. The reflection of light twice in the dual paraboloid reflector system provides high IR and UV rejection ratios, resulting in less degradation of the optical components and fibers. ASAP models of the system and experimental results will be presented. The shape of the etendue curve also provides higher efficiency in using polarization recovery system. Several patent-pending light-pipe-based polarization recovery and recycling systems, will be discussed. Calculation and experimental results will also be presented.

The key idea of Fresnel optics is to decouple the global slope from the local slope by breaking up the optical surface into small facets. The size of the facets is irrelevant as long as they are larger than the wavelength of light, so that the system behaves according to geometrical optics, and at the same time small compared the overall size of the optical surface. From the point of view of phase-space conservation, Fresnel optics suffer from a basic shortcoming. The phase-spaces of incoming and outgoing radiation beams need not automatically be equal. This results in either a dilution of radiation or losses or both. On the other hand, decoupling local from global slope allows to tailor the overall shape of the Fresnel lens independently from designing the individual facets. We show that it is possible to closely match incoming and outgoing radiation beams with a particular choice of the global shape of the Fresnel surface. This shape imultaneously minimizes dilution and blocking.

Snell's law allows to find the slope of an optical surface needed to redirect a given incoming ray into a given outgoing ray. Since a prism comprises two surfaces the problem of redirecting one ray with a prism is underdetermined. In a range of situations it is possible to determine a prism such as to simultaneously match two given incoming rays into two given output rays. This allows to tailor 2D Fresnel optics for finite sources and targets. If source and target subtend equal angles as seen from the Fresnel lens, then the facets are symmetric resembling the minimum deviation configuration, which also minimizes chromatic aberration based an the dispersion in the material of the lens.

The methods of optimization for illumination optical systems with application of Total Internal Reflection are considered. The problem of optimization includes: a) the optimal values of the angles for reflecting triangles in order to have maximum efficiency; b) the calculations of the angles for the case of both primary light (from the bulb) and secondary light (from the reflector). Developed methods can be applicable for automotive lamps, aviation lighting and any other illumination systems.

The light guide plate in LCD backlight has a large number of surfaces that make patterns. They make simulation time long in the non-sequential Monte Carlo ray tracing simulation. We present an efficient way to handle the printed pattern in light guide plate; we used the area ratio function of pattern as a probability function of meeting at the pattern surface. This reduced the calculation loops of simulation,and made the great improvement of simulation speed.We developed a simulation tool to apply our new algorithm,and compared the simulation result of our approximate ray tracing method with that of real tracing,and also compared the simulation speed of ours with that of commercial one.