In support of the U.S. Department of Energy SuperTruck program—an initiative that aims to increase the fuel efficiency of Class 8 trucks by 50%—a lighting research team developed a group of low-energy, high-output light emitting diode (LED) lamps for forward lighting that could be retrofitted into the current incandescent headlamp housing and lens.
The use of LEDs, which are a 2π source, requires a different strategy in the optics from those used with a 4π source, like a traditional incandescent, halogen, or high-intensity-discharge (HID) light source. A thermal management system also is necessary to facilitate the long life of the LED source, according to researchers from Grote Industries LLC.
Since the SuperTruck program monitors the electrical load of the lighting, a representation of power usage of a production LED headlamp system was the most critical parameter. The design team used the opportunity to test new methods of photometric concepts and thermal concepts.
A reflector design approach was used to increase the coupling efficiency. The LED was mounted in front of the reflector, with the LED pointed back into the reflector.
The source chosen for this truck application are white LEDs for their efficacy, life, color temperature, and robustness. There were many different styles of high-power white LEDs reviewed. For the low beam, an M x N (i.e., 1 x 4) array of die was chosen. Because multiple dies are needed to achieve the light output, an array of dies are already aligned, removing the need to line up images from multiple sources to form the cutoff required for a Visional Optically Aimed (VOA) lamp.
LED thermal management
The energy input into the LED lighting system is about one-third that of the incandescent bulb system. LEDs are generally about 20% efficient in converting electrical energy into light. Since this efficacy is much higher than the 3 to 4% of the incandescent bulb, the heat produced by the LED system is about one-fourth of the heat produced by the incandescent bulb system.
However, LED system longevity is only achieved if the diode junction stays below a temperature of 400 to 425 K. Even at an ambient of room temperature, the energy input required by the LED system can quickly raise the junction temperature to three or four times the safe junction temperature and lead to premature LED failure. Heat management in the LED system quickly switches from the need to have high-temperature material around the incandescent bulb to the need for heat-conducting material at the base of the LED package to transfer the heat away from the LED to the outside of the lamp.
The existing headlamp system is sealed in a large plastic housing. Such a size is ideal for the heat issues of an incandescent system, but not for that of an LED system. The large volume of this lamp gives the option of circulating the heat within the enclosure or transferring it to the outside.
The task of transferring the heat to the outside of the lamp is complicated by the need to move the reflector/LED to aim the beam pattern properly, while still providing a thermal path to the outside of the lamp. Consequently, the team investigated and tested multiple methods of transferring the heat.
First, ground straps of flexible braided copper wire of gauge 0 were used as heat pipes. These straps were soldered to a copper bar that was attached to the LED mount on one end and the heat fin on another. The fin could be mounted inside the housing or outside as a replacement cap to the access doors. There were two problems with this method. The heat transferred to the ground strap was less than the same gauge of solid bar, and the flexibility of the straps was diminished when the strap length was reduced to cut thermal resistance. This flexibility was reduced below the usable level of the lamp.
Another method was a ball and socket joint. The geometry of the lamp made this method unworkable. However, since the lamp only moved up and down, a method of two parallel plates sliding across one another was possible. In this method, the parallel plates are mounted perpendicular to the axis of rotation and sliding along each other. One plate is fixed to the reflector and the other to the heat fins on the housing with enough surface area to conduct the heat energy away from the LEDs. Though this method worked, there was concern that sufficient contact between the surfaces would not keep its integrity over the life of the lamp.
The last method tried was two parallel disks with wave springs that transferred the heat from an inside disk to an outside disk. The inside disk was connected to the LED heat pipe and could move as the lamp is aimed. The outside disk replaced the plastic access caps that were used to replace the incandescent bulbs.
The electronics uses a switching power supply with a thermal pull-back feedback system. The LED was reflowed onto a metal core board on the mount in front of the reflector. A negative temperature coefficient (NTC) resistor was mounted on the same metal core circuit board with the LED array. The NTC resistor provides a temperature feedback to the switcher. The switching power supply monitored the NTC resistance and adjusted the current to the LED to assure that in hot ambient conditions, the LED does not produce enough heat to exceed the junction temperature rating. Full testing of the thermal characteristics was not completed at the time of publication.
To expedite the design process, two switching power supply systems were modified for this application. The boards were potted into aluminum housing and mounted within the lamp housing.
The optical design is a complex reflector for both the high beam and the low beam. The original low beam was a 60-mm (2.4-in) projector. An attempt was made to make a small complex reflector. Since the source is a 2π emitter, one of the best collimators would be a parabolic mirror with the source directed into the mirror. As long as the reflector extends to the latus rectum and the source axis is aimed at the vertex of the parabola, all of the light emitted will be controlled by the reflector.
When using the intersection of the latus rectum and the parabola as a limit of the reflector width, there is a fixed relationship between the focal length and the distance to the intersection of the latus rectum with the reflector. For a given opening, this fixes the greatest distance from the source to a point on the reflector to half of the opening. Given a source of fixed dimension, the smallest possible image size projected into space is determined by the geometric projection.
This image size was adequate but the simulation showed a larger image. The simulated hot spot was too large for the low beam to make a legal lamp in part because a sufficiently sharp cutoff was not possible. Other predictions more closely matched measured data. The high beam reflector width was 120 mm (4.7 in) along with a less restrictive specification to allow this type of design (LED pointed toward reflector) to be used on the high beam. For a source to be placed at the focal point of a parabola means that it must be supported in space along with giving a path for power to come into the source and heat to be removed. This support was referred to as a “spider.” This spider was reduced to two large vertical legs to minimize the interference with the light pattern and increased in cross-sectional area to improve the heat conduction away from the source.
To make the low beam in the small package size, the image of the source projected into space must be made smaller. Two methods to achieve this are by either masking the source with some type of shield or by increasing the distance from the source to the reflector. The mask would block some of the light reducing the efficacy. The source could be rotated axially and use a short focal length parabola as is used with filament lamps. This would leave some of the light going out uncontrolled and not contributing to the light pattern. A rotation of the source less than 90° could be used so that the 90° direction of light intersected at or near the edge of the parabola.
When using this design, the source is shifted off the axis by the width of the source and a mirror source is added for the other side of the parabola. The two sources form a “V” shape and the mount that supports them also has the “V” shape.
Measured light output
The first low beam yielded 387 lm from a 508-lm source, yielding an optical efficacy of 76%. This meets the predicted output of 75% and an overall lamp efficacy of 3.4%. The predicted output was 550 lm, which yielded an overall efficacy of 5.3%. The source output was lower than what was used in predictions.
The high beam yielded 583 lm out of the available 885 lm from the LED source. The optics has an efficacy of 66% based on light output. The output gives a 6.5% source efficacy and the lamp efficacy of 4.3%.
There are several methods that can be used to calculate fuel savings based on reduced power consumption. All use various assumptions regarding operating time, alternator efficiency, engine efficiency, and other parameters. Depending on the method, fuel savings based on the LED headlamp can vary from 1 gal (3.8 L) to over 20 gal (76 L) per year.
For the low beam, the LED efficacy showed an improvement of 1.0 to 3.4%, while for the high beam, the LED efficacy showed an improvement of 2.0 to 4.2%. The fuel savings alone would not compensate for the higher cost of this LED lamp. However, the longer life of the LED lamp can lower the system cost over the life of the vehicle. In the overall analysis, the LED version results in a brighter color light, a lower system cost, and increased fuel savings.
This article is based on SAE International technical paper 2013-01-0753 written by Albert Bolander, Timothy Brooks, and Peter Thompson of Grote Industries LLC, as part of the Automotive Lighting Technology session taking place at the SAE 2013 World Congress in Detroit. The research was made possible through the cooperation and support of the Volvo Group.