Iterative steps ensure successful LED lighting
White LED technology is evolving rapidly. Pat Goodman and Christos Sarakinos discuss the iterative process that all developers should follow when designing a lighting system.
Light-emitting diodes (LEDs) have been commercially available since the 1960s and have, over time, become commonplace. Indeed, one would be hard pressed to think of a modern electronic device that does not use an LED. Mobile telephones, televisions, computers and even coffee makers all use LEDs for tasks such as "on/off" indication or backlighting of keypads. But while these LEDs are useful, they are of limited value to the general lighting market due to their low current and low-power handling capabilities.
Unlike indicator LEDs, power LEDs are designed to operate at higher currents and in turn produce far more light. Since their introduction by Philips Lumileds in 1999, power LEDs have become potential replacements for most conventional illumination light sources. Because the technology is so different to the familiar filament bulb, many wonder how LEDs create white light, how the technology is evolving and what is required to design a solid-state lighting system. This article addresses these questions.
The colour of white light
The colour of white emitted by a white light source depends on the relative strength of each constituent wavelength, known as the source's spectral composition. Colour temperature (CT), a simplified way of classifying the colour of white light, relates a white light's tint when compared with the colour of light emitted by a blackbody radiator as it is being heated. For this reason, CT is measured in degrees Kelvin.
A lower CT means that the white light is warmer and contains more red, while a higher CT represents light that is cooler and bluer in appearance. In comparison, the CT at sunrise and sunset is around 2200?K, while that of the Sun at midday is around 5500?K.
Very few light sources match a blackbody radiator exactly. To accommodate for deviations from this ideal source, a factor known as the correlated colour temperature (CCT) is commonly used to describe white light. As the name suggests, CCT reflects the nearest colour temperature. It is worth noting that a CCT is calculated based on its chromaticity coordinates when compared with the ideal blackbody radiator (also known as the Planckian locus).
When designing a lighting system, one of the first elements that you should consider is the CCT. Studies indicate that people feel more comfortable when the ambient lighting is similar to the Sun's CT at that time of the day. For that reason, cool CCT is favoured in work environments during the day, while warm CCT, the colour of mornings and evenings, is favoured at home, in restaurants and similar environments.
Figure 1A simple method to describe white light was established in 1931 by the International Commission on Illumination (CIE). The resulting CIE 1931 chromaticity diagram illustrates the way that the human eye experiences light using an x-y coordinate system. In the graph, which is shown in figure 1, the Planckian locus (the black line in the middle) is visible as are the relative positions of all visible colours and a 2D description of each colour point (x, y).
Figure 2Power LED manufacturers expand the area around the Planckian blackbody locus and define x-y boundaries representing "bins" of CCT, as shown in figure 2. These are typically called binning charts and describe the CCT and tint of the white LEDs that each manufacturer produces.
Binning and labelling is the process of separating LEDs based on their various characteristics and assigning them to a specific bin with a unique code. Bin selection can be an important consideration when designing a lighting system.
Creating white light
There are several ways of making white light with power LEDs. One method, which most televisions, computer monitors and other large-area displays employ, is to combine red, green and blue light to make white. Another approach is to combine blue or ultraviolet light with a phosphor (or several phosphors) to make white. The latter, when using a single phosphor, is known as the binary complimentary technique.
There are several ways of combining the phosphor and the LED. One method involves dispensing a drop of phosphor onto the light-emitting semiconductor. As the drop is thicker in the middle of the semiconductor chip, less of the blue light is emitted and the CCT is warmer in appearance. As we move along the top surface of the semiconductor, more blue light escapes near the edges and there is a shift in CCT to a cooler range. This approach is frequently called a "glob" top and can be recognized by the telltale variance in CCT over the viewing angle of the part.
Figure 3A second method conformally coats the phosphor onto the light-emitting semiconductor, which provides a much more uniform CCT over the viewing angle. The two approaches are shown in figure 3.
A recent breakthrough, Lumiramic phosphor technology (Lumiramic), provides even better CCT control than conformal coating. Lumiramic technology combines a ceramic phosphor plate with a thin-film flip-chip. This process is said to minimize the tint variation and produce white LEDs with a specific CCT more consistently. This reduction in variation enables high-volume usage of power LEDs in luminaires.
Capability and technology
LED technology is advancing rapidly and identifying state-of-the-art LEDs requires constant attention to the many releases and claims from scientists and manufacturers. It can be challenging to see through all of the marketing hype and appreciate what characteristics define a good power LED.
Perhaps the first characteristic to consider is the light output or flux. Over the past two years, light output from white power LEDs has more than doubled, but just how much light is emitted from a white power LED can be difficult to understand.
Datasheet numbers are based on testing at a specific current, temperature and time period, and these are often very different to the conditions found in the end application. It is critical that all flux claims and data-sheet numbers are considered and evaluated for what they are – a starting point. By applying readily available tools and calculations, it is possible to determine how much light will be emitted in an actual application before it is even built.
The efficiency of a power LED is an important measure as today even more attention is paid to the energy-saving potential of solid-state lighting. Efficiency, expressed as lumens/watt is a calculable result.
Because power LEDs vary significantly between manufacturers, it is again critical to calculate the performance in the intended application. It is not uncommon to discover that a comparison of datasheet values results in entirely different conclusions when comparing LEDs in the target application. Crucially, the efficiency of an LED can be dramatically different to the efficiency of a complete solution employing that LED. Drive electronics, thermal engineering, optical design and operating environment all have an impact on a system's overall efficiency.
The final point to consider, and possibly the most important, is the reliability of the power LED. While the myth of 100,000?h usable lifetime is still repeated, some power LED manufacturers are providing information that details lumen maintenance of 70% at 50,000?h. One company, Philips Lumileds, has gone a step further and provides complete reliability data with a 90% confidence basis. This means that developers can design arrays of devices or systems with the same understanding that they have for today's conventional light sources.
From idea to application
The question asked most frequently is: how do I design my system? Strictly speaking, there is a preferred way to achieve a robust design. It is an iterative process as the design is tweaked and has proven successful with many solid-state lighting companies.
• Define the light output requirement
For a general illumination system, the first step is to determine the size of the area that needs to be illuminated, the distance between the source and the surface and the desired light density on the surface. From here, the light output requirements of the source can be calculated easily.
• Calculate your target efficiency
This step is critical because the result will dictate the drive current. Remember, all LEDs become less efficient as they are driven harder and get hotter.
• Select a power LED
This selection must be based on both the light output requirements and the drive current. Great care must also be taken to select an LED with the appropriate reliability. For instance, a disposable product might not require the same lumen maintenance or reliability as a medical application. The most common misconception is that all power LEDs are created equal. The simple fact is that they are not. Lifetime claims need to be substantiated by data.
• Simulate the system
The fourth step is to simulate the system, taking into account the various relationships between current, voltage, light, heat and lifetime. Future Lighting Solutions provides online tools that calculate the usable light of a system and allow comparison of LEDs and systems under different conditions (see http://www.futureelectronics.com/promos ... ablelight/
). This simulation will help to determine the number of power LEDs required, the optimal drive current, thermal requirements, and much more.
The final step is to develop and test a prototype. Ideally, this should be done using off-the-shelf and readily available components. First, the prototype should be tested to ensure that the light output and efficiency requirements have been met. It should then be tested to ensure that the thermal design is suitable and will enable the desired reliability.
Power LEDs are continuing to displace other light sources in the general illumination market. The comparison, measurement and implementation in applications are all significantly different to the well-understood process for conventional light sources. Following a few simple steps and utilizing tools specifically designed to help evaluate power LEDs will allow for simpler, faster and ultimately more successful development of new and never before possible solid-state lighting applications.