Energy efficiency in solid-state lighting apps

Energy efficiency in solid-state lighting apps

Many factors influence the overall efficiency of today’s solid-state lighting systemsBY DAVID NEAL
Avnet LightSpeed
Peabody, MA
www.em.avnet.com/lightspeed

Solid-state lighting (SSL) technology has great promise for significant gains in energy efficiency and huge overall energy savings. According to a 2005 study titled Energy Efficient Lighting Technology Options, this country uses 765 terawatt-hours of electricity annually for lighting. By increasing the efficiency of our lighting systems, we can have significant impact on our nation’s energy usage.

At the recent 2007 Department of Energy (DOE) Voices For SSL Efficiency Workshop in Boston, attendees learned about the programs that the DOE has in place to support the introduction of SSL fixtures. Two key ongoing DOE activities are the development of Energy Star performance requirements for SSL fixtures and the development of testing programs to verify fixture performance. These requirements are still in draft form, a copy of which can be obtained at www.em.avnet.com/lightspeed. The reason the DOE is investing so heavily in SSL technology is that while other light sources have reached their limits of efficiency, technology based on LEDs is expected to double or triple efficiency over the next 10 years.

Efficiency numbers

Since the introduction of the standard incandescent bulb, we have rated our light sources based on the power they consume (wattage), but not on the light output of the source. This is obviously a bad idea, and hopefully incandescent lamp packaging labeling will change soon, if incandescents stay around much longer. In illumination applications, the amount of light radiating from a point source in all directions is often measured in lumens.

Light source performance, or efficacy, is the ratio of the power consumed and the light output from the source and is measured in lumens/watt. A typical frosted 60-W incandescent light bulb from Philips has a light output of 890 lumens, and therefore has a source efficacy of 15.8 lm/W.

The actual performance of a fixture can be significantly lowered due to the type of enclosure that the bulb is in. For example, placing a standard incandescent bulb in a recessed-down light, with a reflector, will result in 50% fixture efficiency due to reflector losses. The overall fixture efficacy would then be 7.5 lm/W.

These calculations can be done for the other types of light sources as well. However, when it comes to fluorescent, compact fluorescent (CCFL), and LED based light fixtures, one must also factor in ballast losses and off-line ac to dc constant current control losses. Table 1 shows a comparison of the fixture efficacy of various light sources for a down light application. Next, we’ll look at the same calculation for an LED-based fixture.

able 1. Efficacies of various non-LED light sources for down-light applications

LED lamp efficiency calculations

Nearly every week we see announcements from LED manufacturers highlighting improvements in the lm/watt performance. The DOE has research and development programs that are targeting to increase the efficiency of SSL luminaires from the current 30 lm/W to over 150 lm/W for commercial lighting applications.

Using the current draft of Energy Star requirements for a SSL recessed down light fixture as the starting point, we can calculate the overall efficacy of an SLL fixture using components available today. The specification has a minimum light output of 500 lm for a fixture with a greater than four inch aperture, 3000K correlated color temperature (CCT), and a minimum luminaire efficacy of 35 lm/W.

Let’s use a typical, warm white LED as the basis of a down light application. This LED is specified at 3000K CCT, 350 mA, and a maximum forward voltage drop of 3.8 v. LEDs are “binned” or sorted into intensity bins, and we select a part that yields a minimum of 60 lm @ 350 mA. This is under test conditions, which means the LED junction temperature is 25C and the current is usually pulsed for 25 ms. In real life operation, the junction temperature will rise to a steady state value based on the thermal design of the fixture. With any LED, as the temperature rises, the light output decreases. Operating the LEDs at 60C results in a 10% decrease in light output when compared to the tested values. This means that the actual light output will be 54 lm from each LED in steady state operation.

To achieve the required light output for the fixture, we would need at least 10 LEDS of this value, and we choose to put two strings of five LEDs in parallel. A LED-18W DC 700 ma constant current driver from Light Tech will be able to drive the two strings at 350 mA. It has an efficiency of 80%, and can provide up to 24 volts on the output.

Calculating the power required for 10 LEDs operating at 350 mA and 3.8 V yields 13.3 W. The source efficacy would then be 54 lm *10 LEDs, or 540 lm divided by the power required, or 41.5 lm/W.

If you take into account the driver efficiency, you can calculate the worst case fixture efficacy of 33.2 lm/W. This is just under the current Energy Star requirements, and near the performance of a reflective CCFL fixture, but more than four times the performance of an incandescent fixture.

Variations on the LED theme

This example uses a warm color temperature LED that has 60 lm output at 350 mA with a CCT of 3000 K. Several LEDs manufacturers have released announcements highlighting devices that exceed 100 lm at 350 mA. These are cool white LEDs, with a CCT of 6,500 K, that use a different mixture of phosphors to generate the white light.

The performance of cool white LEDs has doubled over the past two years. If there is a similar rate of performance increase in the efficiency of the warm white LEDs, one can expect that within the next year an 80 lm/W warm white LED will be available.

The selection of the off line driver in the above example also plays an important part in the overall fixture efficacy. Today this is price versus performance selection. Linear drivers are the lowest cost, but have efficiencies that are in the 50% to 60% efficiency range.

Switch-mode drivers are the most efficient, typically in the 80% to 85% efficiency range. In order to achieve the higher the efficiency, more complex circuitry is required and therefore, the more the driver will cost.

Fig. 1. The AN2212 from Seoul Semiconductor yields 120 lm at 4 W at 3000K color temperature.

Another LED option is the Acriche EcoLight from Seoul Semiconductor (www.zled.com). This LED requires no off-line power conversion. Shown in Fig. 1 is the model AN2212, which at 3,000K delivers 120 lm at 4 W, or 30 lm/W. The single emitter AN2214 is a 2-W device that delivers 65 lm at 3,000K CCT. This device has a source efficacy of 32.5 lm/W before any thermal losses. The technology was developed by integrating dozens of light-emitting cells onto one die and features a 30,000-hour lifetime. ■

for more on lighting, visit http://electronicproducts-com-develop.go-vip.net/opto.asp