Open Access
Add to BibSonomyAbstract
By having a single optical element combine the light of several high luminance LEDs, a high luminance light source can be formed, with shape and emission characteristics adaptable to nearly all illumination problems. The illuminance distribution of this virtual source facilitates the generation of the desired intensity pattern via its imaging-stye projection into the far field. This projection is achieved by one refractive and one reflective freeform surface, both calculated simultaneously by the 3D SMS method, which is herein demonstrated for an LED automotive headlamp.
©2006 Optical Society of America
1. Introduction
The dramatic increase in the luminosity and efficacy of high luminance LEDs makes them the future source of choice in lighting, projection, and numerous other applications. But many times a single LED chip can't provide sufficient flux, requiring use of several LED chips. One of the most challenging applications for high-brightness LEDs is vehicular headlamps. Until recently LEDs have been applied successfully in all exterior automotive lightning devices except full headlamps. The long life time of LEDs, their compactness, instant switch-on, lower power consumption, high color temperature, and unique styling potential for a high-tech automotive look that also distinguishes from competitors, are major drivers for the development of LED front lighting. Because of the high price of the LEDs the most efficient optical system is desirable. Many LEDs also display visible color variations caused by deviations in phosphor layer thickness on the blue chip. Some optical designs transport these color variations into the headlight pattern and thus form unwanted bluish or yellowish zones. Optical designs for high luminance LEDs must take into account mechanical issues, particularly the placing tolerances of LEDs and the color and illuminance variations across their luminous surface.
Conventional incandescent optical designs often use a single free-form or faceted mirror to form a headlight beam pattern [1]. The design is usually adjusted considering the light reflected at different points of the reflector. The reflected light projects an image of the source to the far-field (pin-hole image). The lower the source luminance, the higher must be the size of those projected images (for fixed reflector aperture and total flux).[2, 3] Design approaches developed for incandescent lighting will fail for LEDs because of their comparatively low luminance (a typical automotive tungsten halogen filament displays approximately 48 cd/mm2, as the filament is surrounded by air, compared to today’s Luxeon LED’s 8.5 cd/mm2 with an index of refraction of the chip encapsulant of n=1.5) and their high cost per lumen. For large LED chips, the source images can be so large that the control of rotation and size of their pin-hole images offers great advantages when forming a full headlamp beam pattern. The SMS [4, 5, 6, 7, 8, 9] (Simultaneous Multiple Surface) design method provides such control. The SMS method is the most advanced design tool in non imaging optics providing devices that perform very close to the theoretical limits: high collimation and/or prescribed intensity patterns and very small aspect ratios (depth to diameter) are achievable. Next we present an optical design based on the SMS 3D method.
2. Boundary condition for LED headlights
Today’s HID (high intensity discharge) headlights deliver 900lm onto the road. Taking into account that the available white LEDs reach up to 100 lm per package more than a dozen LED per headlamp are required for illuminance levels similar to HID. The luminance of LEDs is much lower than that of halogen filaments or HID arcs. To produce high flux values, large LED area is needed (typically each chip is 1 to 1.6 mm2). The resulting etendue [11] of these light sources makes large lit aperture necessary to meet the flux and intensity goals of a headlamp. Car producers, however, need extremely compact designs.
Fig. 1. Munsell representation of typical (hypothetical) low beam pattern as seen shone against a wall at infinite distance; axes in degrees (right hand drive ECE low beam)
The regulatory requirements for a high-beam lamp consist basically of a small very high intensity hotspot with an elliptical pattern surrounding it. If sufficient aperture area is available and the optical system efficiency is high, forming the pattern is relatively straight forward as opposed to the low beam functions. The low beam light pattern emitted by a headlight does not only have to comply with the typically 20 legal test points at different angles (with maximum or minimum or both intensity/illuminance values), but also with gradient minima, as well as manufacturer’s additional requirements. Furthermore a beam pattern must have a homogenous appearance (no visible “lines” or “holes” in the pattern), as well as no visible color variations. The most important aspects of the right-hand drive ECE [11] low-beam pattern are shown in Fig. 1.
3. Design concept
In this novel LED headlight concept, the light of several flat-exit-surface, high-luminance LEDs is mixed first to create a secondary source with tailored dimensions and angular spread. This secondary source is transformed by coordinated pair of surfaces, one refractive (R) and one reflective (X), into the headlight far field pattern (in further notation the XR stands for two optical surfaces that produce one reflection and one refraction of the input rays). The elements are:
- Several LED light sources, in this case 3, placed on a common PCB and heat sink
- A tailored light guide, our novel LED combiner, that is in optical contact with the LED’s emissive surfaces. This guide gathers the light from three LEDs and creates a single prescribed illuminance distribution at its exit aperture. This guide has also the function of mixing the light from the 3 LEDs alleviating the effect of LED binning
- A refractive free-form (without any kind of symmetry) lens that is in optical contact with the LED combiner
- A reflective free-form back surface
The illuminance distribution of the LED combiner at its exit aperture 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 free-form surface, as calculated by the 3D SMS method [4, 5, 6]. The entire design process had the objective of producing the right hand drive ECE pattern requirements [11].
The LED assumed for this architecture has a flat exit aperture of about 1.2×1.2 mm2. It consists of an InGaN blue chip with a conformal phosphor coating (top emitting thin film type LEDs, such as the Osram OSTAR). When the LED’s rated flux is measured in air, the LED combiner will increase this nominal output by typically 30–50%, because the optical contact to the LED reduces the usual light losses of flat top LEDs, due to total internal reflection (TIR) light-confinement and eventual absorption in the LED package. The coupling of the LED to the guide can be achieved by an optical gel, an index-matching fluid, or a UV-curable optical adhesive.
Fig. 2. Schematic view of a Combiner RX module (left) and the full LED head lamp concept (right). The mirrors can be placed in many different configurations. The combiners can be covered to be invisible for the end user.
The same concept is used for the low and high beam function, although each element is tailored to each function. In the low-beam case, a step-like feature in the light guide exit aperture is to be projected into the far field. Small images are directed to the hotspot and the larger ones will create the pattern spread. A certain number of identical modules, consisting of 3 LEDs each and their combined optics, are placed as designers desire to form a full low or high beam (Fig. 2).
3.1 The LED combiner
The LED combiner (Fig. 2 and Fig. 3) is a special case of an optical manifold [12, 12], one that provides etendue-limited combination of a plurality of phosphor-coated white LEDs. It is attached directly to the LED surface, thereby reducing Fresnel losses there, thus increasing LED efficiency. In the embodiment shown, the combiner gathers the light from three LEDs via TIR, and creates a simply connected virtual source that spatially combines the light from the LEDs, thereby reducing the color and flux variations between the different LEDs, as well as the illuminance and color variations across the emitting face of each LED.
In this way the exit aperture of the LED combiner acts optically as a large virtual LED (without the thermal problems associated with such large LED). The contour of the LED combiner exit aperture defines sharp illuminance edges, which are used by the secondary optical elements to create the vertical intensity gradient of the headlamp intensity pattern. A detailed explanation of the design procedure of the LED combiner can be found in [12].
Fig. 3. Schematic view of the LED combiner (left and center, dimensions are in mm)) and the source (1 and 2)/ combiner (3)/ free-form lens (4) (right);
3.2 The XR SMS Design
In the SMS design method, the optical prescription is stated in the form of the desired incoming and outgoing wavefronts. The simplest SMS design couples two pairs of wavefronts, and thereby creates two free form surfaces (which can be reflectives or refractives). The effect of extra optical surfaces (prescribed prior to the design), that can be placed before, after, or in between the SMS surfaces, is taken into account by simply propagating the source and target wavefronts through those surfaces. The resulting wavefronts, now refracted or reflected at the non SMS surfaces, are then used in the SMS calculation. Besides the obvious physical limitations (i.e. etendue conservation [11]), the only condition for a successful SMS design process is the absence of any caustics of the wavefronts that are to be coupled, particularly in the vicinity of the SMS surfaces themselves.
In the XR design the two optical surfaces to be calculated by the SMS method are a refractive “dome” (which together with the LED combiner forms a single dielectric solid) and a reflectively coated mirror surface. Besides the two pairs of wavefronts (WFi1/WFo1 and WFi2/WFo2 (Fig. 4, center)) needed as input parameters for the SMS design procedure, the two optical path lengths between the two corresponding pairs of wavefronts and a “seed” curve on one of the surface sto design (X or R) have to be chosen. This curve can also be calculated with the SMS procedure using for example WFi1/WFo1 and a third pair of wavefronts WF4i3/WFo3 (Fig. 4, left). One of the SMS surfaces will start to “grow” from this curve. The vertical seed curve, derived from the corresponding outgoing wavefronts, will define the vertical extend of the projected source images.
The input wavefronts WFi1 and WFi2 are spherical surfaces emitted from the corners of the rectangular exit aperture of the LED combiner light guide. The two outgoing wavefronts project one source edge exactly to the horizon (WFo1) and the other source edge to a fixed angle below the horizon. The other two exit wavefronts (WFo1 and WFo2) are chosen in accordance with source luminance to prescribe the light emission. The SMS calculation now couples WFi1 with WFo1 and WFi2 with WFo2, or, in other words, ensures that all rays emitted from the edges of the light guide will leave the optical device exactly as dictated by the two outgoing wavefronts. The schematic design, as depicted in Fig. 4, would create a well defined rectangular image of fixed angular extent. In a more general low beam design, the outgoing wavefronts are free from surfaces (not shown) that contain detailed information on how the light will be emitted.
Strictly speaking, the SMS design method only controls the light emitted by two, or partially a third point of the source. In most practical cases, however, all other rays emitted by the source from non-design points behave “well” in the sense of not showing excessive pin hole image distortion. The calculations are carried out with our proprietary software. SMS points generated using the 4 design wavefronts from a point on the seed curve are called chains. The seed curve can be sampled at as many points as desired to create many chains. The full design is eventually defined by all the SMS points, which are finally interpolated by two surfaces; one of them containing the seed curve.
Fig. 4. Schematic design procedure for the XR SMS design. Left: design to create a seed curve. Center: design that creates SMS “chains” of points in space. Right: Schematic view of the far field creation with a free-form combiner/ XR. Some pinhole images of different points on the reflector are shown
To create a smooth pattern with varying intensities, small source images are directed to the hotspot and the larger ones to create the pattern spread (Fig. 4, right). In the case of the low beam, a sharp edge and a step-like feature in the light guide exit aperture (not shown) are to be projected into the far field, where they create an elbow and a vertical gradient. Simple calculations show that the required sharpness of those edges in a mass produced plastic injected LED combiner/lens is well within standard edge radii [7]
4. Results
Both, the ECE low and high beam designs are based on 15 LEDs, 75 lm each, 1.2×1.2 mm2 emitting surface, no cover glass and 97% mirror reflectivity (multi-layer optical film reflectors [14, 15]).
The bulk material used for the LED combiner/ SMS lens is polycarbonate (PC). Detailed ray-tracing simulations predict excellent total optical efficiency (76%/77.2% respectively for low and high beam) and intensities (52.9 lux for low beam and 115 lux for high beam measured at 25 m distance) that promise exceptional driver viewing (Fig. 5. and Fig. 6.). The design meets all legal ECE R112 (for high beam) and ECE R98 (for low beam) test points. The low beam design has a wide beam spread to mimic HID head lamps. Also very low glare values, perfect cut-off, good shoulder/step, a hotspot close to horizon (only 0.9 deg down) as well as the low foreground light are provided. A detailed analysis of tolerances shows that an LED to light guide placing error up to +/-0.2mm has virtually no effect on the light pattern. The figures of manufactured optics described in this paper can be found in [16].
Table 1. Results of ray-tracing simulations for low and high beam
5. Summary
High and low beam headlamp optical designs based on LEDs have been completed. Each of these designs contains an LED combiner that mixes the fluxes coming from several LEDs to form a secondary source with defined boundaries. Simulated and measured results confirm that the high efficiency and etendue conserving properties of this concept make it applicable to many high luminance LED design concepts where good uniformity and high intensities are paramount. The presented XR SMS design uses such a LED combiner to create an LED headlamp with a high quality light distribution and a very sharp cut-off that is tolerant to LED optical misalignments and illuminance variations across the LED surfaces. Legal ECE low and high beam designs with >75% total optical efficiency respectively (without cover lens) and patterns similar to HID headlights have been achieved. Very high hotspots, low glare values, a sharp cut-off and wide beam spread make this headlight an extremely safe and comfortable lighting device. The presented design uses a loss free concept to create the gradient and cut-off, by imaging features of the optics itself to form the low beam patterns. The design is tolerant to typical LED placement errors and LED illuminance characteristics.
Acknowledgments
This study was supported by EU project TST3-CT-2003-506316: “Integrated communicating solid-stage light Engine for use in Automotive Forward lighting and information exchange between vehicles and infrastructure”. The authors thank William A. Parkyn, for his help in editing the manuscript
References and links
1. Jung-Hyang Park and Jong-Youb Sah, “Design of Reflector Optics with Smooth Surface for Automotive Lamps,” in SAE 2001 World Congress: Lighting TechnologySAE Technical Papers, 2001
2. V.I. Oliker and O. von Tempski, “On the design of reflectors with prespecified distribution of virtual sources and intensities,” Inverse problems 14, 661–678 (1998). [CrossRef]
3. H. Ries and J.A. Muschaweck, “Tailoring freeform lenses for illuminations,” in Novel Optical Systems Design and Optimization IVR. Winston , Editor, Proc. SPIE4442, 43–50, (2001). [CrossRef]
4. P. Benítez, R. Mohedano, and J.C. Miñano, “Design in 3D geometry with the Simultaneous Multiple Surface design method of Nonimaging Optics,” in Nonimaging Optics: Maximum Efficiency Light Transfer VR. WinstonJ. Koshel , eds., Proc. SPIE, 3781, 12–19 (1999). [CrossRef]
5. P. Benítez and J.C. Miñano, et al, “Simultaneous multiple surface optical design method in three dimensions”, Opt. Eng. 43, 1489–1502, (2004) [CrossRef]
6. O. Dross, P. Benitez, and J. C. Miñano et al, “Review of SMS Design Methods and Real World Applications,” in Nonimaging Optics and Efficient Illumination SystemsR. Winston and J. Koshel, eds., Proc. SPIE5529, (2004) [CrossRef]
7. O. Dross, A. Cvetkovic, J. Chaves, P. Benitez, and J.C. Miñano, “LED Headlight Architecture that creates a High Quality Beam Pattern independent of LED Shortcomings,” in Nonimaging Optics and Efficient Illumination Systems II, R. WinstonJ. Koshel , eds., Proc. SPIE5942, 126–135 (2005).
8. US Patent 6,639,733: “High Efficiency Non-Imaging Optics” Patent #, Issued Oct. 28, 2003. Inventors: Juan C. Minano, Pablo Benitez, Juan C. Gonzalez, Waqidi Falicoff and H. John Caulfield.
9. US Patent: “Three-Dimensional Simultaneous Multiple-Surface Method and Free-Form Illumination-Optics Designed Therefrom” (pending). Inventors: Pablo Benitez and Juan C. Minano.
10. R. Winston, J.C. Miñano, and P. Benítez. Nonimaging Optics, Elsevier Academic Press, San Diego, CA, 2005
11. For ECE specifications (i.e. Reg 112) see http://www.unece.org/trans/main/wp29/wp29regs.html
12. J. Chaves, W. Falicoff, O. Dross, J.C. Miñano, P. Benítez, and W. A. Parkyn. “Combination of light sources and light distribution using manifold optics,” in Nonimaging Optics and Efficient Illumination Systems IIIR. Winston and P. Benítez, eds., Proc. SPIE6338-21. 63380M-1–63380M-10 (2006)
13. Patent: “Optical manifold for light-emitting diodes” (international patents pending). Inventors: Julio Cesar Chavez, Waqidi Falicoff, Juan C. Minano, Pablo Benítez, Oliver Dross, William A. Parkyn, JR, Roberto Álvarez
14. M. Weber, C. Stover, L. Gilbert, T. Nevitt, and A. Ouderkirk “Giant Birefringent Optics in Multilayer Polymer Mirrors” Science, 287, 2451–2455 (2000) [CrossRef] [PubMed]
15. For reflector characteristics see www.3m.com
16. Oliver Dross, J.C. Miñano, Pablo Benitez, A. Cvetkovic, and Julio Chaves, “Non-imaging optics combine LEDs into one bright source,” in SPIE Newsroom, http://newsroom.spie.org/x3596.xml (2006).
