Creating a lighting luminaire for parking garages
Summary of the project
The University of Colorado (CU) offers a course in optical design for its illumination curriculum as part of their Architectural Engineering program. Students of this program generally work as either lighting designers or product development engineers at luminaire manufacturers.
The final project for this course has the students experience the full design process, including completing a competitive analysis, developing beam requirements, designing an optic, building and testing it. The hardest part is always building an accurate prototype. Students were allowed to choose between building a reflector or 3D printing a lens optic. For this project, the design was for a parking garage luminaire and 4 students chose to design a reflector while 6 students chose to design lens optics.The light source used for the project was a Cree XQ-E LED, which is a 1mm^2 chip on a 1.6 x 1.6mm base.
Unique project features
Students experienced designing for additive manufacturing
3D printed optics were still much more accurate than the reflector prototypes
The parking garage luminaire required a square distribution with a peak beam angle around 65°. The optical design was done in Photopia and the lens prototypes were manufactured by Luxexcel.
The actual prototype manufactured by Luxexcel
The printed lens is shown below. The 2nd image shows the stretched image of the LED chip (the yellow region) from a view angle into the lens at about the angle where the peak intensity is produced. The luminous intensity at each angle in the beam is directly proportional to the amount of luminous area the lens projects, so the LED image is either compressed or stretched depending on whether you are trying to increase or decrease the LED's default Lambertian distribution. One of the difficulties of this project was that very slight deviations in the lens curvature can affect the luminous area projected, which is why these types of lenses are so sensitive to the exact physical geometry obtained.
The next 2 images show the actual illuminance pattern obtained. The general trend of a higher concentration of light at the beam fringe was seen in all 6 lenses that were printed. In this case, the extra punch in the beam center was also a bit exaggerated from what was expected.
So the main difference between the simulated Photopia beam and the physical beam is the illuminance gradient between the beam center and the peak beam angle. The results in Photopia show a smooth transition from the beam center outward, while the physical lens shows a peak in the center with a slight drop in illuminance moving outward until the point where the illuminance spikes. The 2 plots below show the simulated and measured ("Tested") intensity distributions for both the 0 and 45° planes in the distributions. The 0° plane is a vertical slice of the beam you see in the images above and the 45° plane is along the diagonal of the square pattern.
The cause of the difference in the measured intensity distribution is from a slight difference in the printed 3D lens shape compared to the design, either on the inside surface, outside surface or both. Small geometric deviations will cause significant distortions in a beam of this type. In this case, the beam was generally smooth, but just more focused to the widest angle.
In the data there is a minor change in the surface normals of the lens shape and therefore the angles of the refracted light. A dip in the lens reduces the light in the central part of the beam since it refracts the light to the sides. So the flatter that section of lens becomes, the more light that will end up in the beam center. That's exactly what is seen in the picture of the beam pattern where there is a punch in the center. Also, the small hemispherical cutout into which the LED is inserted has some differences that slightly change the beam entering the lens.
The printed lenses proved to be very helpful to the students as they not only saw the physical confirmation of their design performance, but also learned just how sensitive this type of design is and the challenges involved in producing accurate prototypes. The goal was for them to experience the full design cycle including developing the beam requirements, designing the optic, building and testing it. Four students chose to build reflectors and had mixed success since they are difficult to fabricate accurately given the facilities available. Although the printed lenses showed a higher concentration at the beam fringe, they were still much more accurate than the reflector prototypes. This is why CU wants to continue to better understand the limits of what can be printed. This double sided lens style with optically active surfaces on both sides was interesting since it opened up a new lens type for printing.
These small, wide beam optics are some of the most sensitive types of optics used in lighting in regards to their exact surface geometry, so small deviations can cause significant distortions in the beam. Still, 3D printed optics do offer significant benefits in manufacturing speed, flexibility and affordability. So knowing the limits of 3D printed optics will help us design things differently next time while at the same time the accuracy of the 3D printing process continues to improve at a rapid pace. CU greatly appreciates the opportunity to cooperate with Luxexcel on this and future projects.