Light sources represent the input quantity for all design processes of projection and illumination systems and often there are features of the light source that are unsatisfying in the optical engineer's view. However, just as the design of an optical system may be understood and improved with the help of modeling tools, the development of a light source may also greatly benefit from physical models of the discharge. This paper tries to provide a basic understanding on how HID lamps work and how they can be modeled. This covers the major processes within the plasma discharge as well as the heat balance of the discharge within the burner and in particular the influence of the burner shape on the discharge behavior. Questions concerning the arc shape or the presence of a liquid melt at certain areas of the burner wall which are of vital interest to the optical designer will be addressed together with an assessment of the possibilities to influence these phenomena. Finally, a discussion on how and to which degree of reliability light technical properties can be obtained from a lamp model is presented.

In our contribution we demonstrate a wide variety of ray tracing software applications for the design of VIP short-arc discharge video projection lamps. On the basis of simulations we derive design rules for the lamp itself and for its optical environment. Light Tools software acts as a means to understand the collection efficiency of a VIP lamp with an elliptical reflector and as an instrument to prove the conclusions.

Short arc lamps are a key component for projection systems to achieve highest efficiency for small display sizes. Consequently UHP-lamps are now standard for high efficient projection systems. UHP lamps combine a very high luminance with a good spectrum and lifetime. The optical performance of UHP lamps can be further improved: A dichroic coating on one half of the UHP burner is applied to focus all light into one hemisphere. This allows for extremely compact reflector systems and an improvement by 20-30% in light collection. The design of projection optical systems requires reliable optical models of the light source. Two different approaches for the model description of lamps will be presented. 1) A physical model based on the geometrical description of all surfaces of the lamp. Volume emitters are used to describe the light generation. 2) A model based on high resolution goniophotometric measurements of the light source. Simulations using both types of models are verified by measurements and the benefits and shortcomings of both approaches are discussed. Special emphasis will be placed on the role of the light bulb as the first optical element in the system. The relatively thick quartz bulb of these high pressure lamps acts as a lens and shapes the beam of light.

Techniques to improve source modeling are presented: filament flux weighing, depositions on the arc envelope interior, and electrode degradation. Filament sources provide more light from the center in comparison to the ends. Additionally, the helix interior is hotter due to increased absorption, and thus the flux emission is greatest here. These effects for linear filaments are modeled in software with the ancillary use of camera images of lit appearance. The result is that the source luminance is more accurately modeled. This technique, called flux weighting, is described and software examples using reflectors are presented and compared to those that do not use flux weighting. Software models of arc sources that employ camera images of the arc provide accurate representations of the source radiance. However, these models do not include arc source aging. Aging effects include degradation of the electrodes and the depositions on the interior of the envelope. These phenomena lead to a decrease typically in the luminance from the source. Camera images of the lit and unlit appearance of arc sources are presented and their effect on the arc output is discussed. Additionally, software examples using reflectors are presented and compared to those that do not use these techniques.

Automotive lighting devices generally have to meet high standards. For example to avoid discomfort glare for the oncoming traffic, luminous intensities of a low beam headlight must decrease by more than one order of magnitude within a fraction of a degree along the horizontal cutoff-line. At the same time, a comfortable homogeneous illumination of the road requires slowly varying luminous intensities below the cutoff line. All this has to be realized taking into account both, the legal requirements and the customer's stylistic specifications. In order to be able to simulate and optimize devices with a good optical performance different light source models are required. In the early stage of e.g. reflector development simple unstructured models allow a very fast development of the reflectors shape. On the other hand the final simulation of a complex headlamp or signal light requires a sophisticated model of the spectral luminance. In addition to theoretical models based on the light source's geometry, measured luminance data can also be used in the simulation and optimization process.

This paper discusses the reverse engineering of filament-based light sources for computerized optical analysis purposes, especially problems raised by source tolerances. The H12 automotive headlamp bulb is used as a case study. Notes on reverse engineering sources consist of: a statement of the challenges involved, our source-modeling methodology, and useful values and procedures pertaining to simulated sources and optical ray-tracing software. A multi-model approach is outlined and consists of: gathering tolerance information from specification sheets, modeling for source tolerances with eight key models, and output comparisons between nominal and toleranced versions of the H12 source. Recommendations for including source tolerances in non-imaging illumination designs conclude.

Source modeling is one of the key elements that enables software simulations to predict the output of an illumination system. A successful method to model a source is to use measured spatial luminance distributions over a range of view angles to synthesize rays that are then used in a Monte Carlo illumination simulation. The measurements are typically performed using a camera at numerous discrete view angles. Important issues arise when synthesizing rays at angles between the measured angles, especially when the source being measured incorporates optical elements such as a reflector.

Accurate optical modeling of illumination systems requires, among other things, an accurate characterization of the light sources used in the system. The problem is that there is no single acceptable definition of 'accurate' that applies to all circumstances. What is determined to be acceptably accurate must also be tempered with analytical efficiency, meaning that the ideal light source model must contain just enough accuracy to produce acceptable results in the quickest manner possible. Finally, the modeling technique must also be practical to implement given that the technique needs to be applied to the large number of light sources in use today. There are many light source modeling techniques already used in practice and it is worth considering the appropriateness of the various techniques for the applications to which they are being applied. This paper starts with a description of the different aspects of light source models and their modeling techniques. The importance of the different aspects is then summarized for a range of applications. The goal of the paper is to present a set of guidelines that can be referenced by designers interested in selecting the most appropriate type of light source model for their range of applications.

In lamp development and quality control we are often faced with small differences between lamps in a batch or changes in lamp characteristics during life that we want to understand. The paper describes a practical and powerful tool to guide the interpretation, viz. a photometry map in which the luminous flux of a lamp is plotted versus the electrical power input in a relative way. Due to the physical properties of tungsten, there are only a limited number of directions in such a plot. At constant voltage, two principal directions are shifts related to variations in filament dimensions that go in the direction of +2.5% increase in luminous flux for +1% power increase, and shifts related to variations in convective losses that go in the direction of 8% drop in flux for 1% increase in power. Other relevant directions are related to external resistive circuit elements and blackening. A number of cases in which these plots have been applied are described. They show that the maps give clues to complex lamp phenomena. However, the agreement between the experimentally observed shifts and some predicted values is unsatisfactory.

Integrating spheres are often used as calibration sources providing uniform radiance within a solid angle and/or uniform irradiance at a distance. The best performance in such a system can be achieved if one is able to evaluate as well as predict the important characteristics of the sphere system's output, such as the spatial and angular distributions of radiance over the exit port, or the distribution of irradiance at the external plane of calibration. We have developed the algorithms and specialized software based on Monte Carlo techniques to solve the problem of radiation transfer inside an integrating sphere containing several point sources and conical annular baffle. The new algorithm employs backward ray tracing coupled with the shadow rays technique. As a timesaving procedure, the axial symmetry of the sphere and the superposition principle are used to substitute the sum of single source radiation fields rotated through a specific angle, for the radiation field of the complete multiple source sphere. The random (due to the stochastic character of the Monte Carlo method) component of uncertainty for the radiance or irradiance results is less than 0.1%. The results of numerical experiments are used to establish the performance variation as a function of the reflectance and specularity of the sphere wall, the number of radiation sources, the type of baffle used, and the angular distribution of their radiant intensity.

The complexity and specialization of optical applications is calling for tailor-made solutions as can be seen in the multitude of new specialized light sources present on the market. The OSRAM LINEX™ barrier discharge system is an example, how application related requirements are already addressed on the light source level. After introduction of the technology the paper explains the lamp's optical features by means of numerical simulation which are in agreement with experimental data. Several examples like LCD backlighting or automotive signal applications demonstrate how overall system performance can be enhanced through design changes already on the light source level.

With the increased interests in mass production of LCOS and DLP™ crodisplay-based projectors and televisions, the image panel sizes become smaller and the pixel counts become higher. While the image panels are being miniaturized, the screen sizes required by the users are increasing. The two factors together require an increasing amount of light passing through the small image panels, thus increase the brightness requirement of the light sources. Most lamp developers approach the issue by reducing the arc length of the lamp, thus increasing the brightness of the arc. But, traditional reflectors used to collimate or focus light from the arc degrade the brightness of the arc. In this paper, a novel approach is used to increase the brightness of an illumination system by preserving the brightness of the arc using a dual paraboloid reflector system, thus relaxes the requirement of have a shorter arc gap lamp. To illuminate these smaller image panels, a patented dual paraboloid reflector system has been developed to collect and focus light from an arc lamp onto the image panel without loss of brightness. This optical platform provides the control and etendue efficiency that has been missing in standard illumination systems. The dual paraboloid reflector 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 rectangular tapered light pipe (TLP) resulting in a unity magnification, i.e. 1:1 imaging, with conserved brightness. 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. 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. ASAP models of the system and experimental results will be presented. The advantages of this system when applied to polarization recovery systems, polarization recycling systems, and color recycling systems will be discussed.

HiPerVision is a new automotive signaling range of lamps (clear and colored) developed by Philips. These lamps offer car life service, reduced size and - consequently - new design opportunities. HiPerVision aims at progressively replacing P21W lamps and at being an economic alternative to LEDs. All lamp dimensions are significantly smaller than P21W's. The HiPerVision 16W lamp produces less heat than the P21W lamp (9 W less dissipation at 13.5 V), what enables a reduced reflector size, the use of low cost plastic or a combination of both. With a luminous flux of 300 lm (instead of 460 lm for P21W), the legal requirements can be easily fulfilled because of the smaller dimensions and tolerances. In order to illustrate the lamps benefits, a complete automotive taillight with 4 functions was designed and made. This paper describes the reflector design process for that taillight with HiPerVision. According to a current styling trend, the reflectors are based on Pillow Shaped Facets and on a clear front lens with no optical structure. With this design method, the whole reflector area is filled with sparkling light. The basic shape of the reflector was used to optimize heat management. By changing the shape and/or number of the pillows the desired light distribution was made. The HiPerVision lamp was measured with a Luminance Goniometer. The measurements were converted to ASAP ray sets as input for accurate simulations. The legal requirements were easily met which was confirmed by actual measurements. The total depth of the complete designed taillight was 53 mm, which is small compared to existing P21W based designs. If the lamp is placed transversal the requirements are still met and the depth of the complete taillight could be reduced to 33 mm, which is comparable with a taillight based on LEDs.

We suggest an optical system for beam homogenization and speckle reduction of spatially highly coherent Laser beams. The new method is applicable to laser beams with moderate temporal coherence. Based on the finite temporal coherence the spatial coherence is reduced prior to the homogenizer component. The new design was experimentally tested for ArF - Excimer laser at 193nm. For this Deep UV application we used silicon micromirrors in combination with fused silica microlenses with a pitch of 150 μm . In contrast to former fly's eye homogenizers for laser beams the new method employs a large number of sub-apertures even for lasers with high spatial coherence. This results in a speckle free and uniform intensity distribution in the target plane. Furthermore the pupil filling can be increased drastically. Experimental results for a DUV microscopy application are presented and discussed.

Many illumination tasks require large area light sources, e.g. LCD backlighting or general lighting. Depending on the application these sources have to fulfill criteria concerning uniformity, angular distribution and color of light output, brightness, efficiency, flatness and cost. After discussing established solutions with regard to these requirements, light guides which incorporate light sources like light emitting diodes or thin fluorescent lamps in cavity like recesses are introduced. The advantages of this scheme are flatness, scalability in area and reliance on established sources. This will be demonstrated for large area light sources utilizing high power LEDs or thin fluorescent lamps.

The divergence of the laser beam radiated from a light source affects the resolving power of the optical pickup head. The high power semiconductor laser does not have Gaussian light intensity distribution, which can be easily proved by scanning the photo-detector in two directions. The optical resolving performance is directly measured by using the so-called Target Chip, which consists of gratings with different spatial frequencies ranging from 555 up to 1667 lines per millimeter on the synthesized quartz substrate, sputtered with aluminum forming a 70% reflective layer. The line depths of the gratings are constant at 0.1 micrometers and the line widths are 0.3 micrometers. The pickup head to be tested is situated on a table and light is focused on the grating through the substrate. The motion of the chip generates the corresponding output signal from the pickup head. This demonstrates the relationship between the divergence from the light source and the resolving power.

Several models were presented in literature to explain the photoluminescence of porous silicon, suggesting carriers originate in Si nano-structure and recombine on the surface. None of the models have a clear justification on such carriers generation and carriers dynamics with respect to the mass conservation law, since the carriers are known to originate in one place and recombine in other. In this investigation a qualitative model, namely, the Virtual Band Model (VBM) is proposed to fill in the gabs which are raised in literature.