News and New Products
Voices: Avnet LightSpeed’s Cary Eskow: illuminating energy-efficient design
By Suzanne Deffree, Managing Editor, News -- EDN, 11/27/2008
Cary Eskow is director of LightSpeed, the solid-state-lighting- and LED (light-emitting-diode)-business unit of Avnet Electronics Marketing. A 28-year Avnet veteran, he leads Avnet’s national team of illumination-focused engineers experienced in thermal, drive-stage, and optics design and has worked closely with LED manufacturers, advanced analog IC, and secondary optics vendors since receiving his first patent using LEDs two decades ago. He has published more than 30 articles and columns on HB-LED (high-brightness LED)-system design and is a member of the Underwriters Laboratory-8750 LED Standards Technical Panel. Eskow recently spoke to EDN about designing for the HB-LED market. Excerpts of that interview follow.
Why is the HB-LED market in the spotlight right now?
Initially, people are looking toward HB LEDs because of energy savings. The single greatest usage of energy in any building at around 25% tends to be lighting. Because of the physics involved and the way incandescents work, typically, an HB LED has three or four more times light output for any given amount of power input.
In California, Title 24 mandates that all structures have to have a certain amount of energy-efficient lighting. [You can achieve] energy efficiency by using compact fluorescents or other types of devices like that, but they all have mercury. There are about 13 states that have legislation pending or in force that restricts the use of or mandates special handling of mercury-containing lamps. So, at some point, you have a convergence of legislation that mandates energy-efficient lighting and other legislation that increases awareness of hazardous materials [such as] mercury. The only real lighting source that satisfies both is HB LEDs.
What’s the design challenge?
The challenge with LEDs is, with the exception of infrared LEDs, they get hot. Based on certain things that happen with the semiconductor, that heat cannot escape in the form of infrared energy. The most significant design challenge—the one that people typically don’t anticipate when they begin an HB-LED design—is pulling the heat out of the device.
HB LEDs are more energy-efficient than incandescent and fluorescent lights, but, at the system level, does that advantage fade because of power loss?
On the system level, it is it still far more advantageous to use an LED over an incandescent. An incandescent doesn’t have any converter, but maybe 5 to 7% of the energy that goes into a bulb comes out as white light. In an LED-based system, yes, when you convert from the ac power to the constant-current dc typical low voltage that an LED requires, there may be 10 or 15% of loss in the circuitry. However, you gain that [loss] back because so much of the input energy going into an LED is realized and radiated as useful light, not as infrared or ultraviolet or other things that you don’t want.
If you compare compact fluorescents or regular fluorescents, they are pretty efficient and are very close to what an LED can provide as far as lumens, which is the way you measure light per watt. However, there are some big negatives with fluorescent tubes and compact fluorescents. One of them is lifetime. They don’t last a long time. The second thing is that they are filled with mercury gas—not a lot, but it’s in there, and that’s why they are considered hazardous wastes.
Can you outline the Energy Star regulations and how they apply to solid-state lighting?
There were many lessons we learned from the failure of compact fluorescents to [achieve] wide acceptance. The first is quality. So [the Department of Energy] came up with a set of specifications that include quality, exactly how much light must be put out in a given area in a given situation, the efficiency, even the manufacturer’s warranty. If all these requirements are met, that product can be Energy Star-certified. … That’s not the energy efficiency of the LED. It’s the amount of energy going in from, say, the wall outlet to the amount of light on the specific area you are interested in.
It’s really a quality-control system, not just a matter of energy efficiency. And it’s difficult to achieve, by the way. So, when a solid-state-light product is Energy Star-certified, it’s been well-designed. The result is a product that will last, that will build a consumer confidence, and [that will] actually have the energy savings that were promised but perhaps never really realized with other technologies, such as compact fluorescents.
Why do you think it is important for engineers to work with distributors such as Avnet on specialized system-level design assistance when it comes to HB LEDs?
Today, what is lacking isn’t necessarily the components themselves. A number of companies have rapidly advanced the art of producing efficient HB LEDs. Furthermore, the ICs and other components required for a constant-current-HB-LED supply are certainly not difficult to use. What is missing, however, is knowledge of how to design an illumination system. You can liken this to FPGAs [field-programmable gate arrays] back in 1985—when their adoption was paced by the knowledge of how to apply them. That’s where we are now with solid-state lighting.
The most valuable commodity related to HB LEDs isn’t sourced by any supplier; it’s the experience and system-level knowledge of how to best leverage HB LEDs in your specific application. This [knowledge] usually entails a proper understanding of secondary optics, comprehensive modeling of the application’s thermal requirements, careful HB-LED-device selection, and experience in derating HB-LED-performance data.
Of course, data-sheet interpretation, modeling, and careful design are not new disciplines for engineers, but doing this [work] with light and LEDs can be a major task. Avnet’s approach is to offer customers a dedicated team of optical, thermal, and electrical engineers who work with these issues exclusively, day in and day out. In six or seven years, HB-LED products will be common commodities, and system-level optical/thermal knowledge will be pervasive. It just isn’t yet.















