November 1999

Most automobiles use backlighting to illuminate dashboard displays, such as speedometers, gauges, indicators, sound systems, and climate control. This approach provides excellent aesthetics and enhanced readability, which is a positive safety factor. Historically, the design of backlit dashboard displays was largely a trial-and-error process. But the newest generation of illumination design and analysis software reduces design cycle time and cost, and also enables the production of more sophisticated and efficient systems.

Light Pipe Operation and Design

Figure 1. Typical light pipe illumination system.

In a typical automotive display illumination system, the output from a light source is coupled into one or more thick plastic light pipes, which direct the light to various graphics (Figure 1). The light pipes are fabricated from a clear plastic -- typically polycarbonate or acrylic -- by injection molding. An efficient design utilizes total internal reflection (TIR) to confine the light within the light pipe, eliminating the need for any optical coatings. At present, color-filtered incandescent bulbs are most commonly used. Light-emitting diodes (LEDs) are now becoming popular because they offer higher electrical efficiency (with the exception of white LEDs), unfiltered colored light, lower heat generation, and a longer lifetime.

The ability to illuminate several graphics with a single source is the primary advantage of light pipes, since this minimizes the number of light sources required. This is not an insignificant consideration, given that many automobiles now have more than 100 different light sources. Light pipes allow the source output to be steered around various mechanical obstacles within the dashboard. They can also be used to alter the luminous distribution of the source, so as to achieve greater illumination uniformity, or specific light levels on a given graphic.

Unlike imaging optics, which generally have been designed and optimized with sophisticated computer ray-tracing programs for about 30 years, the design of plastic light pipes has traditionally been performed by trial and error. The initial design was generated using basic geometrical concepts (angle of incidence equals angle of reflection) to ensure that incidence angles were high enough to meet the TIR condition, and flat "fold mirror" surfaces were incorporated when needed to change propagation direction. Then an actual plastic prototype part was produced by either injection molding or a combination of machining and hand polishing (polishing is required to produce a surface smooth enough to support TIR that is also free of significant scattering). Finally, this prototype was tested to validate the design and indicate where improvement was needed.

Both injection molding and machining/polishing are time-consuming processes. Furthermore, injection molding tooling can cost $10,000-$40,000. Because of these time and cost constraints, a light pipe would typically undergo only two or three design iterations before going into production. Limited analysis together with a small number of prototype cycles resulted in systems that were not highly optimized.

Using Computing Horsepower

Computer modeling of an illumination system is based on the same basic principles as imaging optical system design. Specifically rays are traced through the system, and either reflect, refract, or both (i.e., a beamsplitting surface) at each material interface. This changes the direction and flux of each ray as it propagates to the next surface. The analysis and optimization of most imaging optical systems is accomplished by tracing several hundred or several thousand rays at a limited number of field positions. In contrast, performing a useful analysis of a relatively simple illumination system, such as the light pipes under discussion, may require the tracing of millions of rays. Thus it is only with the recent advent of powerful desktop computers and software that computer-aided illumination system design has become commonplace.

Figure 2. LightTools system construction screen shot

LightTools^®, by Optical Research Associates, is one illumination system design and analysis software package that takes advantage of inexpensive and fast computing horsepower. Unlike traditional imaging optical design programs, where systems are specified surface by surface, LightTools' CAD-like interface enables the user to build three-dimensional models quickly from a toolbox of common shapes (Figure 2). Boolean operations, including intersection, subtraction, and union, and trimming facilitate the creation of complex geometries. Both optical and nonoptical components, such as mounts and mechanical constructs, can be included and ray-traced in the model. The program also models complex surface and volume-emitting light sources.

Once a light source or input ray bundle is defined, LightTools uses nonsequential Monte Carlo ray tracing to determine the path of the light through the system. The optical effects at every surface a ray encounters are taken into account; thus even the interaction of light with component edges and mechanical structures is analyzed. Polarization, scattering, and surface reflection effects, as well as the performance of thin-film optical coatings, can also be considered. Simulation output consists mostly of plots of the illuminance or intensity distribution at any number of arbitrarily chosen surfaces.

More Sophisticated Design

A coauthor, Dr. John Van Derlofske, has used LightTools for the past five years at DaimlerChrysler and currently at the Lighting Research Center at Rensselaer Polytechnic Institute, in the design of light pipes for dashboard illumination. During that time the software has been found to produce computer models of sufficient accuracy to largely eliminate the expensive and time-consuming step of constructing physical prototypes. Typically, five to 20 iterations of a design are run on the computer before producing a prototype for final verification.

In addition to providing time and cost savings, nonimaging optical design software has also enabled the exploration of more sophisticated design forms for light pipes, particularly that of aspheric surfaces.

To understand the benefit that an aspheric surface can deliver, consider a simple light pipe designed to perform the commonly required task of redirecting an incandescent source through a 90^o angle to provide illumination evenly over an extended area.

Figure 3. Right-angle light pipe analysis shot, showing the flux difference between a flat fold mirror and an aspheric fold mirror.

The most obvious design approach is a light pipe with a 45^o fold surface (Figure 3). This design efficiently performs the basic function of redirecting the light, since most of the source's output meets the condition for TIR. It does not, however, transform the nonuniform input flux distribution into one that is more uniform. Thus the illuminated graphic appears much dimmer at the edges than in the center. Replacement of the flat fold surface with a curved surface results in a component that can both bend the light and alter the input flux distribution to produce greater uniformity.

To create and optimize designs built around aspheric surfaces, the shape of the curve is defined using a standard polynomial expansion. Changes to surface curvature, position, tilt angle, conic constant, and fourth-order deformation coefficient are then input to optimize the design. Light-source modeling proceeds along similar lines. The structure of the source is first built inside the program, and the source model is compared to either measured or manufacturer-specified characteristics, or both.

The results comparing the output of a light pipe using a flat fold surface and an optimized aspheric surface are shown in Figure 3. Two million rays were traced to produce this output, which clearly shows the superior uniformity of the aspheric system. The output also demonstrates that a small amount of overall throughput was sacrificed to achieve this critical goal, because the curved shape of the surface causes TIR to fail for a greater number of rays than the flat surface does.

Previously various surface parameters were modified manually, and then a ray-trace was performed to assess the impact of changes. Automated optimization techniques using LightTools' newly added macro programming, however, are now being developed. The first step in this process is to use the desired final flux distribution to define a performance merit function. Macro commands are then used to alter surface parameters and perform ray-tracing to optimize the design's merit function iteratively.

A trial-and-error approach has long been used to design optics for illumination systems, such as film and video projectors, flat-panel displays, interior vehicle lighting, architectural lighting, segmented mirrors, sign lighting, machine vision systems, and medical imaging systems. Now a new generation of illumination system design and analysis software enables a much more rigorous and analytical approach to be taken.

For more information, please contact the coauthor of this article, John Van Derlofske, Ph.D., the Head of Transportation Lighting at the Lighting Research Center at Rensselaer Polytechnic Institute, 877 25th St., Watervliet, NY 12189; (518) 276-8717. Coauthor David W. Kuntz is a managing partner at TMS, 171 High St., Laguna Beach, CA 92651; (310) 377-5393.

Top