Lighting products used in illumination or luminaires are used to illuminate an environment with electric light sources. Generally, a product has, at a minimum, a fixture envelope, a light source, and an electrical connection to a power source. Examples include downlights, troffers, outdoor area and streetlight luminaires, under-cabinet luminaires, chandeliers, and others. The interest in using white solid-state light sources for illumination applications started in the mid-1990s. Today, the technology, specifically the inorganic light-emitting diode (LED), has matured to a point that these solid-state lighting (SSL) products, both luminaires and integral replacement lamps (i.e., those containing the electronics for the replacement lamp that are not otherwise present in the incumbent luminaire), are able to compete well with some traditional technologies in certain applications.
Even though the replacement lamp is a subcomponent of a luminaire product, in some sense the replacement lamp is also a self-contained product. Therefore, in this chapter we call both the complete luminaire and the integral replacement lamp a product. Ultimately, the lighting product’s (i.e., the luminaire’s or the integral replacement lamp’s) performance in a given application is what matters most to the purchasers and end users of that product. In this chapter we look at each subcomponent of a lamp or luminaire and its performance and then address the luminaire and integral lamp products and issues related to their overall performance. This chapter addresses only products that produce white light, created by either mixing color (red, green, blue, yellow) or downconverting with a phosphor.
Typically, an SSL product consists of several subcomponents, including:
• An LED, an LED array, an integral lamp, or an organic LED (OLED) panel;
• Secondary optics to control the distribution of the light;
• Heat sink, thermal management components, or thermal interface material (TIM); and
• Driver and control devices.
FIGURE 4.1 illustrates these components for a screw-base A-lamp LED replacement.
Some luminaires integrate the light source(s) with the luminaire envelop and other product components, meaning the light source cannot be easily removed and replaced or repaired. Some luminaires with integrated light sources are considered retrofit luminaires and are to be used in whole product replacement. Other luminaires have replaceable lamps (with a screw base or pin base), including A-lamps, linear fluorescent lamps (LFLs), compact fluorescent lamps (CFLs), multifaceted reflector (MR) and parabolic aluminized reflector (PAR) lamps, and others. Figure 4.2 illustrates a luminaire with integrated light sources and a luminaire with a replaceable lamp.
Luminaires with integrated light sources offer a number of advantages compared to luminaires with replaceable lamps. The luminaire designer/manufacturer has more control over the entire product (e.g., electronic components, thermal management design, optics, etc.) and can select the components to optimize performance. In contrast, a developer of replacement lamps must consider all of the possible luminaires within which a product may be installed and design a product to optimize compatibility rather than performance.
After the phase-out of certain types of incandescent lamps between 2012 and 2014, consumer choices for replacing these types of screw-in lamps will include higher efficacy halogen incandescent lamps, CFLs, and LED lamps. Many SSL product manufacturers are producing screw- and pinbased lamp products to replace incandescent, halogen, compact
FIGURE 4.1 An LED equivalent of a screw-base A-lamp showing the component parts. Courtesy of Philips Lighting.
and linear fluorescent, and metal halide lamps. This is an appealing market segment for several reasons. The large number of available sockets appeals to the manufacturing community, and the lower investment required to try these products by directly replacing the older lamp in the existing luminaire appeals to the consumer. Although the LED chips themselves are manufactured by a small number of multinational companies, the assembly of an LED lamp resembles that of any electronic equipment. The investment needed to set up an assembly line is relatively modest, and therefore a very large number of companies can and have entered the industry. The approximately 4 billion medium screw-base sockets in U.S. households represent a very attractive potential market, so the industry development has happened very quickly, in the span of only a few years. At the same time, the industry1 is scrambling to develop meaningful safety and performance standards, and product quality varies over a wide range.
FIGURE 4.3 through Figure 4.5 illustrate examples of LED replacements for incandescent A-lamps, PAR lamps, and linear fluorescent lamps. The lamp on the left of Figure 4.3 uses the remote phosphor concept where the blue LEDs excite the orange phosphor cover (which emits white light), and the lamp on the right uses two phosphor white LEDs2 placed within an envelope that mimics an incandescent A-19 lamp.
FIGURE 4.2 Two types of LED luminaire: (a) with integrated LED light source; (b) with replaceable LED module. Courtesy of Toshiba.
FIGURE 4.3 Sample LED replacement lamps for incandescent A-19 lamps. Courtesy of Philips Lighting.
FIGURE 4.4 Sample LED replacement lamp for incandescent parabolic aluminized reflector lamps. Courtesy of OSRAM SYLVANIA and Paul Kevin Picone/PIC Corp.
FIGURE 4.5 Sample LED replacement lamp for linear T8 fluorescent lamp. Courtesy of Pacific Northwest National Laboratory.
The luminous efficacy of LED replacement products has improved over the past several years and is expected to continue, as illustrated in Figure 3.1 of Chapter 3. Figure 4.6 illustrates examples of performance data for replacement LED-integral lamps (A-lamp, PAR lamp, and linear lamps), reported in 2011. The efficacy values of these replacement lamps are in the range of 40 to 110 lumens per watt (lm/W). Several LED replacement A-19 and PAR lamps are showing very promising results in terms of efficacy.
A few LED replacements for 4-foot linear fluorescent tubes have performance similar to traditional fluorescent lamps, but for many of them the total light output is substantially lower, and the spatial distribution of light is far more concentrated than that of the conventional fluorescent lamps. The narrow spatial distribution and relatively low luminous flux mean that closer spacing of luminaires would be required to achieve the same lighting environment as produced by conventional fluorescent lamps.
There are many challenges for making reliable replacement A-19 lamp replacements. An LED replacement lamp for an incandescent A-lamp requires squeezing all the needed components, LEDs, driver, heat sink, etc., into a light-bulb sized package as shown in Figure 4.3. Heat dissipation is very challenging and could affect the reliability of the LED lamp. At the present time, in early 2012, it is difficult to make long-life, reliable, LED replacements for incandescent A-19 lamps greater than 75 W because of thermal management challenges. There are some high-power PAR replacement lamps that use active cooling in which a fan is employed to move air and remove heat by convection. However, active cooling usually is not desirable in lighting products because of additional failure modes and audible noise issues. Realizing these limitations, an industry group, Zhaga Consortium,3 is developing a standard for a better socket for replacement lamps with better heat dissipation characteristics, among other attributes. Even though the new socket may help lamps and luminaires in the future, it will not help replacement lamps for existing luminaires. Many of the LED A-19 replacement lamps currently in the market (early 2012) cannot be considered as true replacement for the following reasons:
• LED replacement lamps have a larger geometric shape than the incandescent lamp they are meant to replace and may not fit into a luminaire that was designed for incandescent A-19 lamp.
• The spatial beam distribution of the LED replacement lamps is not similar to that of the lamps they are designed to replace. For example, in a common table lamp, LED replacement lamps often will cast light in a more upward direction, leaving the tabletop surface below relatively dark.
• Although a wide variety of LED replacement lamp products are commercially available, their initial purchase price is much higher than that of competing lamp technologies. However, the “Lighting Facts” labels that appear on lamp packages provide consumers an estimate of annual operating costs, which allows rough calculation of payback times.
Retrofit luminaires are SSL products that fit into the spaces occupied by existing luminaires but require complete removal of the existing luminaire for installation. Common types of retrofit luminaires are those for recessed housings, 2Œ × 2Œ or 2Œ × 4Œ recessed troffers, high-bay luminaires, track
3 “Zhaga is a consortium, a cooperation between companies from the international lighting industry. The cooperation is governed by a consortium agreement that defines rules regarding confidentiality, intellectual property, and decision making. Zhaga enables interchangeability of LED light sources made by different manufacturers. This simplifies LED applications for general lighting” (Zhaga Consortium, 2012).
lighting products, pendant lights, and roadway luminaires. Although less constrained by “existing holes in the ceiling,” other LED products that might be categorized as retrofit luminaires include under-cabinet lights, showcase lights, pathway lights, and rope lighting products. These products take on similar forms to existing non-SSL luminaires.
Of this category of products, the most attention has been paid to SSL roadway lighting luminaires. While most of these luminaires produce luminous flux comparable to incumbent technologies, a few are significantly dimmer (National Lighting Product Information Program, 2010; DOE, 2012). Some of the advantages of LEDs, such as long life, high performance in cold environments, and robustness, make SSL very attractive for many roadway applications. As with replacement lamp products, however, the spatial distribution of light is very different from many SSL roadway luminaires than from other types of light sources. This is frequently a disadvantage, because consumers expect a replacement product to behave identically to the preceding technology. Some outdoor luminaires have significant glare, which is not desirable.
One advantage is that the LED luminaire has the opportunity to direct light more toward the task, thus reducing wasted light and helping to control light pollution.
Another application where retrofit luminaires with LED have done well is recessed cans. Websites including those of the ENERGY STAR® program, U.S. Department of Energy’s Lighting Facts, CALiPER, and Gateway programs are places where one can gather information regarding the performance of commercial LED lighting luminaires (Next Generation Luminaires, 2012; DOE, 2012). Figure 4.7 illustrates examples of 2011 performance data for downlight luminaires for commercial lighting applications. As seen, in 2011, the luminaire efficacies of ENERGY STAR®-rated LED downlights are in the range 35 to 85 lm/W. In comparison, CFL and halogen downlights are in the range of 10 to 30 lm/W.
Even though LED luminaires have greater luminous efficacy than traditional light source luminaires, LED luminaires can have greater lamp to lamp color variation, glare, and flicker and cannot be dimmed.
FINDING: While the majority of LED products in the marketplace have better luminous efficacy than traditional lighting technologies, for many of them, other quality factors, such as useful life, color appearance and rendering properties, beam distribution, flicker, and noise, may be inferior to traditional lighting products. Even though the optimistic view is that energy has been saved by using SSL technologies, if other factors such as system life, lamp-to-lamp color variation, glare, flicker, and dimming, do not meet user expectations, they could slow down market adoption of SSL technologies.
FIGURE 4.7 Sample performance in 2011 of commercial LED downlights. SOURCE: See http://www.energystar.gov.
As with most SSL lighting products, retrofit luminaires have higher initial costs than competing technologies. However, they are becoming more widely used in applications where maintenance costs are high.
SSL products in the commercial market employ a variety of LED white light sources, including an array of phosphor-converted LEDs (blue LED chips covered by a coating of phosphor); an array of cool white (i.e., high color temperature) LEDs combined with red LEDs to create a warmer white and feedback control to maintain light output and color; and an LED array with a mixture of multicolored (red, green, blue, etc.) LEDs. These LEDs or LED arrays are mounted on a heat sink to minimize the heat at the LED junction(s) and are powered by an electronic driver that produces power of the form required by the LED. In some cases, secondary optics are used to direct the beam in a specific manner. If the LEDs are packaged as an integral lamp to replace a traditional light source, the lamp envelope (i.e., glass bulb) is designed to mimic the form of the traditional source and includes a specific connector (e.g., an American National Standards Institute (ANSI) standard base).
This section analyzes these subcomponents, their state of the art, and what improvements are needed to produce products with the performance and price necessary for widespread adoption.
LED and LED Array
As described in Chapter 3, white LEDs are commonly made by dispersing phosphor(s) in the encapsulant surrounding the blue (or near-ultraviolet) LED chip. The process of combining phosphors with the LED chip has evolved over the years. Some packages still use the original method of mixing phosphor(s) into an epoxy or silicone medium. Other packages use a layer of phosphor conformally coated on the chip, while newer LED packages and products consist of phosphor layer(s) separated from the LED chip(s), commonly referred to as a remote-phosphor LED or product (Hoelen et al., 2008; Narendran et al., 2005). Remote phosphor-type LEDs minimize heat-induced efficiency loss in phosphors (provided the phosphor conversion efficiency is not very low as well as the absorption of phosphor-converted photons by the blue LED chip). An LED array is created by mounting and interconnecting individual LED devices on a printed circuit board, which is then connected thermally to the heat sink.
A unique feature of OLED lighting is that the device itself can form the installable fixture because of its ability to be fabricated on any particular substrate or shape. Indeed, OLEDs can be fabricated directly on plastic blocks, flexible metal or plastic foils, or glass. In its configuration as an area lighting source, as discussed in Chapter 3, the luminaire itself operates without a significant increase in temperature above the room ambient. That is, in appropriately packaged devices, at a high surface luminance of 3,000 cd/m2, the luminaire temperature rise can be only a few degrees centigrade, creating no local or distributed heat load on the room environment.
In an LED lighting product, secondary optics are needed to tailor the output beam of a lighting product. LED products commonly designed for illumination applications have LEDs arranged in several different ways together with secondary optics. These designs include an LED array placed inside reflector(s) and behind total internal reflection (TIR) lenses. These methods help the collection and distribution of light in a specific manner. Refractive optics, commonly referred to as lenses, reflective optics, or reflectors, are generally designed as non-imaging optics to be used in illumination products for beam shaping. Researchers have designed and used complex optics to achieve difficult beam shapes (Tsais and Hung, 2011).
Typically, no secondary optics are required for OLED panels.
Lens materials are usually made from glass, polymers, epoxies, or silicones. Material selection is very important, especially when designing long-life products. Some optical materials degrade when exposed to radiation (more specifically, short wavelengths like ultraviolet (UV) and “blue” radiation) and heat. This spectrally dependent light output deterioration is one of the main ways that LEDs degrade.
Thermal management is very important to enable reliable, long-life LED products, and the thermal management components in an LED product constitute a large fraction of product cost. A high-temperature LED junction can negatively impact LED life and optical performance, and as discussed in Chapter 3 in the section “An LED Primer,” this places considerable demands on the plastic lens and encapsulant material. At higher p-n junction temperatures, the amount of photons emitted decreases and the spectral power distribution shifts to longer wavelengths. Furthermore, the degradation of the encapsulant and the LED chip, over time, decreases the luminous flux. Electrical energy not converted to light contributes to the heat at the p-n junction. To keep the LED junction temperature low, all heat transfer methods, including conduction, convection, and radiation, must be considered. Heat conducted to the environment from the p-n junction encounters several interfaces and layers. Therefore, to keep the junction temperature low, the thermal resistance of every layer and interface must be very low.
Thermal Management Component and Strategies
An LED chip is typically encapsulated in a transparent material, such as epoxy, polymer, or silicone. These materials have very low thermal conductivities. As a result, the majority of the heat produced at the p-n junction is conducted through the metal substrate below the chip and not through the transparent encapsulant. Usually, a high-power LED is mounted on a metal-core printed circuit board (MCPCB). When creating a product, an LED (or an array of LEDs) mounted on an MCPCB is attached to a metal heat sink using a TIM. Usually these heat sinks have extended surfaces, such as fins, which dissipate the heat to the environment by convection and radiation. Currently, a few manufacturers have started to mount the LED directly onto the heat sink to further reduce the thermal resistance from the junction to the environment and also to reduce the overall cost.
Common thermal interface materials are solder, epoxy, thermal grease, and pressure sensitive adhesive. Parameters that can influence thermal resistance include: surface flatness and quality of each component, the applied mounting pressure, the contact area, and the type of interface material and its thickness. Adding conducting particles and carbon nanotubes (CNTs) to TIM to reduce thermal resistance has been studied (Fabris et al., 2011).
Most manufacturers exploit both conduction and convection methods to reduce LED junction temperature. Usually the heat sinks have a very large metal surface area, and, as a result, the integral lamp or the entire luminaire is much heavier than its traditional counterpart. Figure 4.8 shows typical weights for incandescent, CFL, and LED lamps of different types.
To make the weight of LED products comparable to traditional lamps, lightweight materials, like polymers and composites, with very high thermal conductivity are needed. The thermal conductivity of plastic materials can be increased by using fillers such as ceramics, aluminum, graphite, and so on. Injection-molded polymer parts of high thermal conductivity are an economical approach for cooling
FIGURE 4.8 Weight comparisons among incandescent (INC), compact fluorescent (CFL), and LED lamps for A19, PAR20, PAR30, and PAR38 lamp types. SOURCE: Narendran (2012).
While these passive cooling methods work well for certain types of SSL products, higher power LED lighting products (1,500 lumens and above) pose significant thermal management challenges. Passive heat sinks are not sufficient to keep the LED junction sufficiently cool. Therefore, to achieve desired lumen values in a small form factor (e.g., A-lamp, PAR lamp, MR 16, etc.), active cooling may be required to dissipate the heat. Even though mechanical fans have been used in some high-power LED lighting products (Ecomaa Lighting Inc., undated; Peters, 2012), they are not desirable for many reasons, including short life, acoustic noise, attraction of dust, and increased energy use. Over the past several years, other active cooling techniques have been investigated for managing the heat in high-power electronics, including synthetic jet and piezoelectric fan technologies. Synthetic jet technology uses a moving diaphragm that produces air movement by suction and ejection of air. Rapidly fired pulses of air are directed to where cooling is needed, such as heat sink fins, to improve cooling efficiency. Piezoelectric fans have several advantages, including longer life, lower acoustic noise, and lower power demand (Zhang et al., 2011). These techniques have shown promise and are worthwhile for further development for high-power LED cooling (Acikalin et al., 2007). Even though active cooling may be necessary for some products in some applications, for the majority of the applications, passive cooling is more desirable.
There is a strong interaction among LED device efficacy, the requirement placed on the thermal management system, and the cost of SSL. Increased efficacy reduces the heat generated per lumen, allowing either a shrinking of the necessary heat sink, and thus a reduction in cost and weight, or an increase in lumen output for the same physical luminaire.
FINDING: LED efficacy strongly leverages cost, physical size, and weight of SSL luminaires.
RECOMMENDATION 4-1: The Department of Energy should place a high priority on research directed at increasing the efficacy of LEDs.
Thermal Management for OLEDs
One of the advantages of OLEDs is that the thermal management challenge is less stringent than for LEDs because their heat density is very low because of their large surface area. Indeed, it is found that at a surface luminance of 3,000 candlea per square meter (cd/m2), OLED panels typically operate in room environments cooled only by natural convection, at 5-7°C above room temperature. However, many applications require high-intensity spot sources. In fact, very high luminances (>10,000 cd/m2) have been demonstrated for OLEDs, but, unfortunately, their operational lifetime scales roughly inversely with current (and therefore, brightness). Also, it is known that for every 10°C increase in temperature the OLED lifetime decreases by approximately 30 percent (see Chapter 3). Hence, thermal degradation becomes a limitation at very high brightness. Up until now, this has prevented the application of OLEDs to high intensity or specular lighting applications. Further research, however, directed at developing molecular materials and device architectures that are more robust and, therefore, can more easily withstand these extreme operating conditions can result in a much expanded application domain for OLEDs. Continuous improvements in brightness and lifetime, however, are being made by researchers across the globe. If such high-brightness-spot OLED sources are successfully developed, the other useful features of this lighting technology may eventually dominate the SSL market.
FINDING: OLEDs are typically low-intensity, large-area lighting sources. However, numerous applications require more intense, specular lighting as afforded by LEDs. The lifetime of OLEDs are negatively impacted by high currents used to generate high brightness.
RECOMMENDATION 4-2: The Department of Energy should invest in research that can lead to small area but high-intensity lighting systems with organic light-emitting diode for use in directional illumination applications.
The light output of an LED is proportional to its drive current, which is typically direct current (dc), and this current is supplied at a relatively low voltage. To provide the appropriate dc current and voltage, an electronic circuit known as a driver is inserted between the alternating current (ac) line voltage and the LED. This electronic driver can be incorporated within a lamp product, as for the A-lamp LED replacement, or as a separate device located external to the luminaire.
Integral Drivers in LED Replacement Lamps
The LED replacement lamp has an integral driver, illustrated in Figure 4.1, that enables the lamp to be connected directly to the line voltage socket. The medium screwbase lamp offers very little space for the built-in or integral driver, so thermal challenges for the electrical components can be significant. The drivers utilize electrolytic capacitors for energy storage on the dc side of their ac-to-dc converters, and they are likely going to be the weakest link in these products and limit the product lifetime. The maximum temperature ratings of these capacitors are typically in the 105°C to 125°C range, at which temperature their life ratings are 5,000 to 10,000 hours. Each 10°C reduction in operating temperature increases the capacitor life by roughly a factor of 2, giving the driver designer the challenge to maximize the
For these reasons, the rated wattage of available LED lamps is still fairly small. The luminous efficacy of the LED devices has increased to a point where a lamp with a light output equal to a 60 W incandescent lamp consumes only 10 W. However, because of the thermal challenges, a 100 W equivalent lamp has not yet become available. As a result, screw-in LED lamps that are used as incandescent replacements in existing installations are at present limited to the lumen output of a 60 W incandescent lamp, although a 75 W equivalent has recently been made available.
Commercial-grade LED luminaires are not constrained to use any particular form factors for the components. This is because luminaire replacement in commercial buildings is easier in dropped ceiling-type construction where there is ample room for luminaire housing and components, and replacement is, therefore, more readily performed. The drivers in these luminaires are typically separate from the LED module, and the luminaire may be designed so as not to present a thermal problem for the drivers. There is a trend in the industry to design “universal” drivers that produce either constant voltage or constant current output to the LED with input voltage ranging from 100 Vac to 277 Vac, which covers almost all global requirements. Having said that, today manufacturers of these drivers—quite often the same companies that also produce ballasts for fluorescent lamps—produce a wide array of products with differing specifications, while they are jockeying for market position. Standardization in the industry has started in some areas (see, for example, NEMA (2010a) and emerging standards by the Zhaga Consortium (2012)), especially for the interconnections between different components within the SSL luminaire (e.g., for standardizing electrical, mechanical, and thermal connections of the LED luminaire, including the LED module, the heatsink, the driver, and any lighting controls).
Most currently available drivers for interior lighting applications have relatively low output power, up to around 40 W. Some higher-output drivers, typically rated around 100 W, do exist for outdoor and industrial high-bay applications. It is likely that higher-output drivers will be needed in both interior and exterior applications in the future, when higher-light-output LED modules become available. The construction complexity of these products is similar to electronic ballasts for fluorescent lighting, and their assembly can be performed anywhere in the world. The potential number of different output configurations of LED drivers is much larger than for fluorescent lamps, mainly because fluorescent lamps are quite standardized, while LED designs are not. This may lead to fragmentation in the market in the short term, with manufacturers of the different component parts of an LED luminaire forming loose alliances to make certain that the products work together. It is also reasonable to expect a high level of obsolescence of driver designs, with older designs being replaced by those having different features during the time that the industry remains without standards. Indeed, the early designs from different manufacturers have been quite unique and not compatible with one another, so direct replacement of components within the LED luminaire, and sometimes even the replacement of the entire luminaire, can be challenging. Leading companies have recognized the need to rapidly develop standards, at least for the interconnections between the various components within the LED luminaire. As a result, the Zhaga Consortium was formed to develop these standards.
FINDING: Because of the large number of different ways to construct an LED lamp, industry has recognized the need for some levels of standardization and has organized to develop such standards.
The long expected life of an LED light engine will put pressure on the driver designer to produce designs that have equally long life ratings. Just as with integral drivers in incandescent replacement lamps, the weakest link in a nonintegral LED driver is the electrolytic capacitor that is used for energy storage in the ac-to-dc converter that is part of the driver. Research into other types of energy storage devices, perhaps ceramic capacitors with high capacitance and small size, may become necessary, and funding for it should be considered. Another way to solve this problem may be to create a new building infrastructure, where the ac-to-dc conversion is performed centrally, nearer to the utility entrance to the building. This enables the building to use only a few larger power ac-to-dc converters that may not be as cost constrained as in the case when the conversion is performed in every luminaire. At least one industry group, the EMerge Alliance,4 has been formed to investigate the possibility of a new electrical infrastructure and start the development of standards in this area. This application is currently limited to commercial buildings that use a dropped ceiling consisting of a ceiling grid and ceiling tiles, because it is envisioned that the elements of the ceiling grid are going to become the electrical conductors.
Drivers for OLEDs
The driver industry for OLEDs is at its infancy. It can reasonably be expected that the driver would look similar to, if not be the same as, a driver for LEDs, because the electrical requirement of both light engines are very similar. However, only a few experimental OLED luminaires have been produced so far, and experience driving them is limited. Both
LEDs and OLEDs are current-driven devices and can work using either dc or ac supplies. Given that the line source is an ac voltage supply, the OLED driver must convert voltage to current. Luminance then, is controlled by the current in a nearly linear fashion. Hence, there is a roughly linear dependence of luminance on current. The current-versus-voltage (I-V) characteristic of an OLED follows a power law, I ~ Vm, where the dimensionless ratio m = B/T has values between 3 and 7. Here, B is a constant, and T is the temperature. This is in contrast to that of an LED, where I ~ exp(-AV/T), where A is a constant. Hence, as an LED brightness increases, the voltage required to achieve a given brightness changes. However, given the relatively constant temperature characteristic of OLED operation, the voltage-to-luminance conversion required in the driver is simplified compared to that of an LED, which requires aggressive cooling to remove heat at the highest brightness. However, these products are far from being ready for any kind of standardization.
Nevertheless, the large capacitance of an area device, coupled with a somewhat different current-voltage relationship for OLEDs versus LEDs present as yet largely unexplored challenges in the development of versatile, efficient, and electronically robust control electronics for the former technology. Given the relatively early stage of development of OLED lighting, it is important that issues of lighting control be investigated earlier rather than later to understand what challenges must be met to produce low-cost control systems. In particular, the unusual form factor of OLEDs suggest that there might be opportunities for integrating such electronics in unusual ways with the luminaires that will provide advantages over conventional SSL and other lighting systems.
FINDING: OLEDs are still in their infancy. While the driver electronics may have many similarities to that of LEDs, there are some essential differences in their operating performance because of the large capacitive load presented by OLEDs.
Lighting controls for electric lighting have existed almost as long as incandescent lamps themselves in the form of switches and rheostat dimmers that were used primarily in theatrical applications. The modern lighting control industry had its beginning in the early 1960s, when the first wallbox5 solid-state dimmer for incandescent lamps was commercialized. Since that time, several lighting control devices have been developed by many companies, such as automatic time switches and sensors that are used for detecting the presence of people (“occupancy sensors”) or ambient levels of daylight (“photosensors”). These devices do not connect directly to the luminaire, so the development of SSL technology has practically no effect on the design of the former or compatibility with SSL luminaires.
In the case of incandescent or incandescent halogen lamps, the lighting control device that connects directly to the luminaire is either a switch or a dimmer. Switches are available in two forms, mechanical and electronic. A mechanical switch consists of metal switch leaves that open to form an “air gap” (a physical disconnect) to disrupt electric current to the luminaire and turn the lights off. The actions of opening and closing the switch leaves cause electric arcs to be formed that typically last no longer than a few milliseconds, but over time this arcing causes the switch contacts to erode, thus eventually causing the switch to fail. The introduction of electronic ballasts for fluorescent lighting in the 1980s was followed by some reports of switch failures, which were the result of increased “inrush current” during switch turn-on compared with traditional lighting loads. This led the industry to develop a switch-ballast compatibility standard (NEMA, 2011), which is in use today. SSL drivers that meet that standard are not expected to cause problems with mechanical switches.
Electronic switches also exist in the market. These are used especially in components that are part of a lighting product, in so-called smart switches or smart dimmers. Such devices use a semiconductor switch, typically a device called a TRIAC, which can be turned off without creating an airgap to the load. This technology is useful especially in wallbox switches that may be remotely turned on and off, for example, by using a sensor or handheld device. Typical electrical wiring practices do not bring a neutral wire to the wallbox, and, therefore, the microcontroller that controls the operation of the smart switch or dimmer has to be kept “alive” by allowing a small amount of current to pass through the lighting load when the lamps are in the off state. (Without such current, the microcontroller would shut off, and there would be no way to turn the switch or dimmer on.) For conventional lamps this “keep alive” current does not cause inconvenience to the end user because incandescent or fluorescent lights do not emit any light when the current is that small. However, more care has to be taken in the design of controls for SSL devices. Because of the low wattage of the lamps, even small currents can cause visible light output, often in the form of flickering when the lights are intended to be off. The small current charges the capacitors typical in the design of the drivers for these lamps, and the capacitors in some designs discharge periodically through the lamps, causing the flicker. Industry standards are being developed to address this issue, but there are SSL control products on the market today that cause this problem.
Dimmers for incandescent lighting are available with analog and digital designs. The analog designs are similar to mechanical switches in the sense that they employ a mechanical switch producing an airgap when the dimmer is off. However, digital dimmers are in this sense essentially the
5 The term wallbox refers to a wall-mounted electrical box that houses the wiring connections for electrical devices such as light switches, light dimmers, and receptacle outlets.
Additionally, incandescent dimmers can also be categorized into two main types: “leading edge” (also called forward phase-cut) and “trailing edge” (also called reverse phase-cut). The former uses a TRIAC as the semiconductor switch, and the vast majority of incandescent dimmers in residential buildings are of this type because of the lower cost of the design. The TRIAC is turned on when it receives an electrical pulse at its “gate” (one of the terminals of the device) and stays on until the electric current falls below the TRIAC’s “holding current” very near the end of the half cycle of the ac wave form, which in the United States operates at 60 Hz and ideally (and very closely in practice, too) has the form of a sine-wave. The earlier the TRIAC is turned on in the half cycle, the brighter the lamp operates. The resulting voltage waveform at the lamp is illustrated in Figure 4.9. The design of the dimmer converts the user action—such as moving a slider up and down—to the proper timing of this TRIAC gate pulse. The operation of the TRIAC is illustrated in Figure 4.9. Electronic switches that employ a TRIAC simply turn it on at the beginning of the half cycle to operate the lamp at full on.
The TRIAC works very well with incandescent lighting because the TRIAC’s holding current (below which the TRIAC will not remain on) is much smaller than the current in even the lowest wattage incandescent lamps, such as a 25 W lamp. However, with CFLs, and even more so with SSL devices that require very low power, the current required to operate the lamps may be smaller than the holding current, and this can lead to observations of flickering or other improper operation. In addition, other problems have been observed with keeping the TRIAC reliably in conduction with certain LED lamps. In some cases, the problems manifest themselves when the total load (number of lamps) is actually increased, meaning that the minimum load requirement in an incandescent dimmer is not the only condition that needs to be satisfied. Leading edge dimmers that are used in commercial grade lighting controls sometimes provide a continuous gate signal rather than a pulse, and this action is very successful in keeping the TRIAC on even with smaller loads. Additionally, trailing edge dimmers, which use transistor switches that require a continuous signal to keep them on and are not characterized by a holding current, may also avoid this problem. Trailing edge dimmers were originally designed for low-voltage (incandescent) lighting using an electronic transformer to step the 120 Volts ac (Vac) line voltage to the 12 Vac required by the lamps. These transformers were developed to make them lighter, smaller, and also often less expensive than core and coil wound transformers, and they utilize capacitors on the “front end” for energy storage inside the device. When a capacitor suddenly experiences a high voltage, a large inrush of current occurs, and many such electronic transformers are not compatible with leading-edge dimmer designs. Trailing edge designs reverse the process of switching by turning the transistors on in the beginning of the half cycle and turning them off at some point before the end of the half cycle. Front end capacitors do not cause problems for this mode of operation, so these types of dimmers are more compatible with SSL drivers such as those used in medium screw-base incandescent replacement lamps.
The lighting controls industry is developing new dimmer designs specifically for LED lamps that are used as replacements for incandescent lamps. It is reasonable to expect that the lamps will operate well with the new designs, but the industry has estimated that there are more than 150 million leading-edge dimmers installed in U.S. homes, and it is probably impractical to expect to replace them all as LED lamps become more popular. NEMA (2010b) has developed a new standard, NEMA SSL 6-2011, to address the retrofit issue,
FIGURE 4.9 Waveforms illustrating leading-edge dimming control.
and lamps that are designed to comply with this voluntary standard should operate with the existing dimmers and avoid such problems as flickering. Such controls, however, may not provide the same low-end dimming range that consumers are used to with incandescent lamps. In addition, the U.S. Environmental Protection Agency has asked the industry to develop a testing standard for ENERGY STAR® lamps regarding compatibility with controls.6 Together, these initiatives should mitigate most of the compatibility problems, or at least provide consumers reasonable options. Finally, NEMA and the Zhaga Consortium have started another standard development (NEMA SSL 7-2012) to address the need for future LED lamps and future dimmer designs to provide significantly better dimming performance. These standards will place new requirements on both the lamp and the dimmer design.
In addition to controlling light levels through dimmers as described above, there are other industry standard “protocols” to provide dimming signals to luminaires. These include 0 to 10 volts DC signals where higher voltages correspond to higher light levels, digital standards such as Digital Addressable Lighting Interface (DALI),7 wireless standards such as Zigbee, and various proprietary standards by individual manufacturers. In each of these cases, the luminaire contains a separate fluorescent or high-intensity discharge (HID) ballast or a separate driver for SSL products that has been designed to be compatible with whichever method of signaling is used. This means that the method of signaling the dimming information is decoupled from the actual dimming function performed by the ballast or driver, and there is no need for new compatibility standards.
With the phase-out of incandescent lamps taking place in Europe, there has been a proposal to develop a new standard for communicating dimming information on the power line (the so-called “power line carrier” method) to CFL and LED lamps that would replace incandescent lamps. This is thought to be necessary by some industry representatives because phase-cut dimming is interpreted to be permissible by European standards only when used with incandescent lamps (International Electrotechnical Commission, 2009). The new standard is now under development, and some people in the industry expect it to be published in late 2012 or early 2013. This would also mitigate any concerns about total harmonic distortion (THD) and power factor (PF) that have been expressed by power utilities. For more detail, see the discussion below, “Electric Power Quality.”
FINDING: LED replacements for incandescent lamps may not work with all existing control infrastructure, especially dimmers.
RECOMMENDATION 4-3: Industry should develop standards for LED drivers and future generations of lighting controls that will ensure that all LEDs that are designated “dimmable” work well with all new dimmers in the future. In the meantime, SSL products should indicate on their labels that they may not function correctly with presently installed controls.
LED lamps offer an additional control opportunity, that of controlling the color of the light output. There have been some proprietary products in the market that offer this control function,8 but no industry standards are yet under development. It is not yet clear how much value the market gives to such color control, so the development of these standards may not happen in the immediate future, if at all. The functionality will probably begin with luminaires that have a separate driver using digital communication (such as an enhanced DALI that includes additional control functions or wireless protocol). Whether screw-in incandescent lamp replacements ever develop this functionality remains to be seen.
An important difference between incandescent lamps and both CFLs and LEDs under dimming conditions is that incandescent (and halogen) lamps become warmer, exhibiting a color shift (in terms of color temperature, measured in Kelvin) toward the red, when dimmed. The other light sources do not inherently change their color temperature, which is often viewed as an undesirable feature by residential consumers. Humans are biologically biased to prefer red light in lower ambient conditions because of the same shift in sunlight toward the late evening hours. With proper controls and devices emitting the appropriate colors, LED lamps and luminaires can mimic the color shift performance of incandescent lights when dimmed in this manner. It should be noted that there is at least some patent protection9 for this functionality, possibly leading to limited choices for consumers.
Standards play an important role in the development and deployment of new technologies as discussed in the Chapter 5 section, “Testing and Measurement Standards.” During the past several years, several standards have been created, most notably IES LM-79, “Approved Method: Electrical and Photometric Measurements of Solid-State Lighting Products” (IES, 2008a) and IES LM-80, “Approved Method: Measuring Lumen Maintenance of LED Light Sources” (IES, 2008b), the latter used in conjunction with IESNA TM-21 to extrapolate estimates of lumen maintenance. More
6 Alex Baker, personal communication with Nadarajah Narendran, Committee on Assessment of Solid State Lighting, August 2012.
7 See IEC 62909.
8 See for example controls offered by Philips (Color Kinetics).
9 See for example U.S. Patents 7,038,399; 7,014,336; and 6,636,003.
Difficulties can arise with standards when the test conditions do not match those of the installed application. As an example, downlight luminaires are typically measured according to the IES LM-79 standard that requires the surrounding ambient temperature to be at 25°C. In practice, the ambient temperature generally will be higher when the light source is recessed in a luminaire. In addition, the temperature is dependent on where the luminaire is located and will be higher on upper floors where the fixture is surrounded by insulation material. This potential for higher ambient temperatures was less of an issue in the past when downlights used incandescent and halogen technologies whose performance was less sensitive to changes in temperature. But in the case of LEDs, increases in junction temperature can alter the performance. Past studies have shown that in some cases the light output reduction from the product is significant, more than 30 percent (Narendran et al., 2008), versus the LM-79 data sheet. Practitioners expecting a certain performance may be disappointed if they strictly rely on the product’s LM-79 data. While the use of testing standards has worked for incumbent lighting technologies, LM-79 may not work for SSL products because of the latter’s sensitivity to heat. Test procedures specific to the application environment are an ideal solution but much more costly than a single procedure for all applications. A compromise solution would be for manufacturers to publish data with de-rating factors for use in typical applications.
Another important aspect of standards is their quality; that is, their ability to produce reliable and realistic information about performance. Today manufacturers commonly use the IES LM-80 procedure to test the lumen depreciation of individual LEDs, but then use those data to rate the entire product life. (Labeling programs also use LM-80 test data.) In reality, a product has many more components than just the LED. Electronic drivers with electrolytic capacitors are known to have a short life, especially at high temperatures. Products claiming a life of 25,000 to 50,000 hours may not live up to such claims as a result. LM-80 test results are more appropriate for LED package manufacturers to provide to product manufacturers, not to product end users. Even though white papers have started to point out this issue (Next Generation Lighting Industry Alliance, 2011) and research is under way to develop test procedures to predict whole product life more accurately (Davis, 2012; Lighting Research Center, 2012), early adopters of LED lighting may be disappointed when products do not live up to the claims on their labels, based as they are on LM-80 results. Some SSL product manufacturers have started offering warranties for their products. This too is challenging because the terms and conditions for product replacement or cash reimbursement can be difficult to define and settle.
Other examples of industry challenges with standards include the current color standards (e.g., ANSI C78.377) that were borrowed from the CFL industry. Manufacturers grouping LEDs to single bin for a given correlated color temperature (CCT) product according to American National Standards Institute (ANSI) C78 tolerance area may find the color variation between LEDs very large, to a point that it is not acceptable for general lighting applications. Presently, some manufacturers are using tighter bins to avoid visible color difference between products.
FINDING: Additional standards or revisions to standards are needed to resolve unknowns that will otherwise be left to consumers and other lighting decision-makers to resolve, specifically test procedures and/or de-rating factors that account for higher temperature environments, where performance may vary from LM-79 data, and alternatives to LM-80 that can predict whole product life more accurately. In the case of the latter, research is under way to develop test procedures to predict whole product life more accurately.
RECOMMENDATION 4-4: (a) Manufacturers should publish data for photometric quantities and life per industry standards and de-rating factors for use in typical applications. (b) IESNA should develop a test procedure to predict whole product life more accurately. (c) ANSI should revise the color binning standard to ensure imperceptible color differences between two adjacent light sources.
In the United States, power quality is a subject of voluntary industry standards, except for electromagnetic compatibility of some lighting equipment, which is regulated by the Federal Communications Commission (FCC) at frequencies corresponding to radio and television transmissions.
The Institute of Electrical and Electronics Engineers (IEEE) sets voluntary standards for distortion of the voltage waveform in the utility supply to buildings (IEEE 519) in order to ensure that electrical and electronic equipment in the building has a reasonably clean supply of power (correct frequency, voltage, and lack of distortion). Distortion of the sinusoidal voltage waveform is of most concern and is expressed in terms of a parameter known as “total harmonic distortion” (THD),10 which is typically limited to about 5 percent. On the other hand, industry voluntary standards set limits to the distortion in the current waveform drawn by the equipment connected to the electric supply. The distortion is
10 Total harmonic distortion (THD) of the supply voltage is equal to the square root of the sum of the squares of the amplitudes of the voltage harmonic frequencies above 60 Hz divided by the amplitude of the fundamental 60 Hz voltage. A high THD (>33 percent) causes problems in three-phase power systems, because usually the dominant harmonic current is the third harmonic. The third harmonic currents add in the neutral wire of the electrical system, and in cases of high THD one can have a situation where the current flowing in the neutral wire exceeds the rating of the wire, causing overheating.
expressed as THD limits for the current, calculated in the same way as for the supply voltage; for commercial and industrial lighting equipment, the THD limit is set at approximately 30 percent by ANSI standards. For fluorescent and HID lamp ballasts these limits are defined in ANSI C82.77-2002. The displacement of the current waveform is expressed in terms of PF, which is usually defined as the ratio of the real electric power flowing in the product to the apparent power.11 ANSI standards and other voluntary standards also define PF limits for lighting equipment, typically 0.9 for commercial and industrial equipment. These limits have been in place for several decades and, because of a lack of reported problems, seem to be appropriately set. There are currently no THD or PF standards for SSL products, so it would seem appropriate that similar limits be set for SSL drivers for commercial and industrial applications as for fluorescent ballasts. It should be noted, however, that, for residential lamps with integral ballasts and medium screw bases, ANSI C82.77-2002 specifies very loose standards—PF is required to be greater than 0.5 and THD less than 200 percent. Note, however, that the PF for an incandescent lamp is 1 (i.e., “perfect”) and its THD is 0. Therefore, the impact of the residential standard has been minimal as the penetration of screw-base CFLs is limited. As LED lamps become more ubiquitous as replacements for discontinued incandescent lamps, the effects of the liberal PF and THD limits may not be so benign. The ANSI standard for residential screw-base lamps should match that of commercial and industrial applications, as for fluorescent ballasts.
Modern lighting equipment, such as electronic ballasts for fluorescent lighting or drivers for SSL devices, also generate some electrical energy in the radio frequency bands. These types of equipment are termed unintentional radiators by the FCC, and the FCC sets limits for the conducted (i.e., along the electrical wires) and radiated (i.e., into the air) emissions of such equipment.12 Separate limits are set for residential and non-residential applications, with the residential limits being significantly stricter than non-residential ones, presumably to protect the consumer’s ability to receive AM radio broadcasts in the home.
In the European Union, the Low Voltage Directive (2006/95/EC) of the European Parliament sets limits for PF, THD, and radio frequency emissions for lighting equipment and does so by reference to standards published by the International Electrotechnical Commission (IEC) and the Comité International Spécial des Perturbations Radioélectriques (CISPR; in English, Special International Committee on Radio Interference). All of these limits are mandatory in member countries, and the PF and THD limits tend to be stricter in Europe (THD limits for current are in the range of 30 percent) than they are in the United States and apply to a broader class of lighting products, such as lighting controls. On the other hand, CISPR does not distinguish between residential and non-residential emission limits, and the European requirement falls between the FCC’s residential and non-residential limits. Other countries typically follow either the European model or the U.S. model.
Historically, there has been a tug of war between the electric utilities and the lighting industry about the importance of stricter limits on power quality metrics, in particular on THD and PF, because of concerns about incompatibility between an increasing number of electronic loads in buildings. However, the reported number of incidents claiming poor performance because of power quality problems has remained low, while the number of installed electronic ballasts has increased. Electronic ballasts, introduced to the market in the 1980s, now account for more than 80 percent of sales of all linear fluorescent lamp ballasts in the United States. The current limits, therefore, appear to be appropriate.
FINDING: There are existing standards for THD and PF for electronic ballasts for linear fluorescent lamps, but at present there are no such residential standards for LED drivers that are external to the lamp. Standards for low-wattage, integrally ballasted CFLs with medium screw bases in residential applications allow low PF and high THD.
RECOMMENDATION 4-5: For external solid-state lighting drivers in general, industry should adopt the same total harmonic distortion and power factor standards that are in place for electronic ballasts for linear fluorescent lamps. Industry should revisit the standards for low-wattage medium screw-base lamps to determine their impact on power quality before applying them for light-emitting diode lamps, and these standards should match those for commercial and industrial applications.
A recent limited survey of consumer prices for a variety of lamp types (A19, MR16, PAR20, and PAR38) at a Home Depot store in New Jersey indicates that the initial cost of LED lamps ranges from 3.5 times to 15 times (PAR38) that of halogen lamps (PAR20). However, when the total cost of ownership is calculated using an electricity rate of
11 The power factor (PF) of the equipment is equal to the electric power dissipated in the equipment expressed in watts divided by the product of the amplitude of the supply voltage and the amplitude of the electric current drawn by the equipment expressed in volt-amperes. In the past when most ballasts used in lighting were magnetic coils, the main effect to reduce PF came from the phase angle difference between the supply voltage and the current drawn by the equipment, which is why PF can be thought of as the displacement of the current relative to the voltage. With modern electronic ballasts this effect is smaller, and increasing THD also decreases PF. For example, in the absence of any displacement PF, a THD of about 44 percent corresponds to a PF of 0.9. PF is of particular interest to the electric utilities because they bill their customers based on delivered real power. However, the transmission line capacity is expressed in terms of amperes of current, so a low PF product will limit the utility’s ability to generate revenue.
12 47 CFR Part 15 and 47 CFR Part 18.
|Halogen Incandeso Price||LED||Summary|
|Lamp Type||Power (W)||price (incl. 7 percent sales tax)||Life (h)||Cost to Owna||10.000 hr Costb||Power (w)||price (incl. 7 percent sales tax)||Life (h)||Cost to Owna||10.000 hr Costb||Savings Over 10.000 his||Savings (%)||Initial Cost Ratio|
|T3||150||$4.42||2,000||$37.42||$187.10||not available||not appliesible|
aEnergy rate $0.11/kWh.
bNot including labor to install lamp(s).
$0.11/kWh, LED lamps save between 35 and 58 percent over 10,000 hours of operation, which corresponds to about 10 years in typical residential use. Table 4.1 summarizes the results of this survey. The LEDs that were chosen for this comparison are the closest available in light output to the halogen lamps that they would replace. The “cost to own” is the price of the lamp plus the energy cost over the lifetime of the lamp. This calculation is limited to 10,000 hours, even though most of the life ratings shown on the LED packaging are actually longer than that. The initial price ratio in the final column is just the ratio of the prices of one LED to one halogen lamp (measuring “sticker shock”) even though more than one halogen lamp needs to be purchased to reach 10,000 hours of use. Finally, it is also worth noting that LED alternatives are not available for all lamp types at this time, such as the T3 tubular lamp that is used for example in some bathroom vanity lights and floor lamps. A calculation of lifecycle costs of LEDs and various fluorescent lamps, taking account of discount factors and expected improvements in LED performance, is included in Chapter 6.
Purchasing a product while the technology is still evolving is always challenging, especially when the life of the product is very long. Having said that, people are now accustomed to upgrading computers and cell phones in 2 to 5 years because they see value in the new product’s functions. The same cannot be said for lighting. Until now, people have typically changed a light bulb only when the previous one has failed. Unless the payback period is very short, many would find it difficult to justify investing in LED lighting products as replacements for traditional light bulbs, as promoted by the SSL industry. As a result consumers take a “wait and see” approach, even though the currently available LED products could save them significant amounts of energy.
Nevertheless, SSL offers new methods to light our spaces. SSL technologies can be embedded into many types of architectural elements due to their small size and long life to meet the needs of desired tasks or ambiance for the occupant. Responding to this opportunity, researchers and industry groups have been attracted to the concept of creating mini direct current (dc) grids within buildings for lighting and some appliances (as well as power production from photovoltaic systems) while maintaining an alternating current (ac) power grid to transmit power from the generation site to end-user sites without much loss (EMerge Alliance, 2012; Narendran, 2012; Thomas et al., 2012).
A dc-powered SSL infrastructure that allows for rapid reconfigurations of lighting systems using LED-lighted panels that snap in and out of a modular electrical grid, makes it as easy to redesign lighting as to move furniture, providing value to the end users. Such concepts not only allow for greater energy savings, but also can improve lighting in our built environments.
FINDING: The power requirements and flexible physical configurations of SSL make attractive the concept of a new dc building lighting infrastructure.
RECOMMENDATION 4-6: The SSL industry should collaborate with other industries such as building materials and construction to explore the challenges and potential benefits of developing and adopting standards for a new dc electrical infrastructure.
Acikalin, T., S.V. Garimella, A. Raman, and J. Petrosk. 2007. Characterization and optimization of the thermal performance of miniature piezoelectric fans. International Journal of Heat and Fluid Flow 28(4):806-820.
Davis, J.L. 2012. Building a System Reliability Model for SSL Luminaires. Paper presented at The Seventh Annual DOE Solid-State Lighting Market Introduction Workshop, Pittsburgh, Pa., July 17-19.
DOE (U.S. Department of Energy). 2012. CALiPER Program. Available at http://www1.eere.energy.gov/buildings/ssl/caliper.html. Accessed August 2, 2012.
Ecomaa Lighting, Inc. Undated. New 80W LED PAR Lamp with Fan Inside. Available at http://ecomaa.en.ecplaza.net/ecomaa-new-80w-led-par-258939-1872324.html. Accessed September 5, 2012.
EMerge Alliance. 2012. Our Mission. Available at http://www.emergealliance.org/About/OurMission.aspx.
Fabris, D., M. Rosshirt, C. Cardenas, P. Wilhite, T. Yamada, and C.Y. Yang. 2011. Application of carbon nanotubes to thermal interface materials. Journal of Electronic Packaging 133:020902.
Hoelen, C., H. Borel, J. de Graaf, M. Keuper, M. Lankhorst, C. Mutter, L. Waumans, and R. Wegh. 2008. Remote phosphor LED modules for general illumination: Toward 200 lm/W general lighting LED light sources. Paper presented at Eighth International Conference on Solid State Lighting. Bellingham, Wash.: SPIE.
IES (Illuminating Engineering Society). 2008a. Approved Method: Electrical and Photometric Measurements of Solid-State Lighting Products. New York: Illuminating Engineering Society.
IES. 2008b. Approved Method: Measuring Lumen Maintenance of LED Light Sources. New York: Illuminating Engineering Society.
International Electrotechnical Commission. 2009. Electromagnetic compatibility (EMC), Part 3-2: Limits for harmonic current emissions (equipment input current ≤16 A per phase). Geneva, Switzerland.
Lighting Research Center. 2012. SSL Recent Projects: Performance, Metrics and Test Methods. Available at http://www.lrc.rpi.edu/programs/solidstate/completedProjects.asp#Perform. Accessed August 3, 2012.
Narendran, N. 2012. Updating a Legacy: Trading up Edison’s Light Bulb and Electrical Infrastructure. Paper presented at 13th International Symposium on the Science and Technology of Lighting, Troy, New York. Sheffield, U.K.: Foundation of the Advancement of the Science and Technology of Lighting Sources.
Narendran, N., J.P. Freyssinier, and J. Taylor. 2008. LED luminaire performance: Changing traditions can set the right expectations. LEDs Magazine, June, pp. 33-39.
Narendran, N., Y. Gu, J.P. Freyssinier-Nova, and Y. Zhu. 2005. Extracting phosphor-scattered photons to improve white LED efficiency. Physica Status Solidi A—Applications and Materials Science 202(6):R60-R62.
National Lighting Product Information Program. 2010. Streetlights for collector roads. NLPIP Specifier Reports 13(1). Troy, N.Y.: Rensselaer Polytechnic Institute.
NEMA (National Electrical Manufacturers Association). 2010a. Electronic Drivers for LED Devices, Arrays, or Systems. Rosslyn, Va.: National Electrical Manufacturers Association.
NEMA. 2010b. Solid-State Lighting for Incandescent Replacement-Dimming. Rosslyn, Va.: National Electrical Manufacturers Association.
NEMA. 2011. Performance Testing for Lighting Controls and Switching Devices with Electronic Drivers and Discharge Ballasts. Rosslyn, Va.: National Electrical Manufacturers Association.
Next Generation Lighting Industry Alliance. 2011. LED Luminaire Lifetime: Recommendations for Testing and Reporting Solid-State Lighting Product Quality Initiative. Washington, D.C.: U.S. Department of Energy.
Next Generation Luminaires. 2012. Next Generation Luminaires. Available at http://www.ngldc.org/index.stm. Accessed August 2, 2012.
Peters, L. 2012. Are MR16 LED lamps ready for the 50W-halogen switch? LEDs Magazine. June.
Thomas, B.A., I.L. Azevedo, and G. Morgan. 2012. Edison revisited: Should we use DC circuits for lighting in commercial buildings? Energy Policy 45:399-411.
Tsais, Y.-P., and C. Hung. 2011. Design of optical components for a light emitting diode zoom illumination system. Optical Review 18(2):224-230.
Zhang, K., D.G.W. Xiao, X. Zhang, H. Fan, Z. Gao, and M.M.F. Yuen. 2011. Thermal performance of LED packages for solid state lighting with novel cooling solutions. Pp.1/7-7/7 in 2011 12th International Conference on Thermal, Mechanical and Multi-Physics Simulation and Experiments in Microelectronics and Microsystems (EuroSimE). Piscataway, N.J.: Institute of Electrical and Electronics Engineers.
Zhaga Consortium. 2012. Zhaga: Making LED light sources interchangeable. Available at http://www.zhagastandard.org/. Accessed September 5, 2012.