by Matthew Brennesholtz

A couple of months ago, I was asked to work on an LED report and predict 5 years in advance how LEDs will be used in projection systems. Although I have spent much of my professional life designing projectors, I was surprised at what I learned.

Before I began the study, I thought of LEDs primarily as indicator lights. I was aware that there were many LEDs in my house, especially in my computer and its peripherals. There are probably hundreds of LEDs in an ordinary American home. Try turning out the lights and going into your kitchen, or near your computer or home-theater system. In my house, the only piece of electronics without LEDs is my 1972 stereo amplifier. (It is completely dark since the incandescent light burned out years ago.)

As I started to look into the LED market, I found that these devices were currently used for applications where I had not realized the light came from LEDs. The backlight of your cellular-telephone keyboard is probably an LED. All new traffic lights are LED arrays, as are the stop lights on high-end auto-mobiles. LEDs are moving into area lighting, in competition with incandescent and fluorescent lamps.

For many of these applications, the long life of LEDs is the most important property. For example, a traffic light made from LEDs is far more expensive than the equivalent incandescent lamp, but it does not need to be replaced very often, if at all. Given the extremely high cost of replacing a lamp in a traffic light – when the cost of a lift truck, flagmen, and perhaps a policeman for additional traffic control is considered, not to mention the disruption of traffic – the cost of a 100,000-hour LED traffic-light array begins to look very reasonable. And, as a further benefit, if an LED in an array goes bad, the rest of the array continues to operate normally.

My main interest in LEDs, however, was for projection displays. Several years ago, a colleague of mine at Philips Research looked into the possibility of using LEDs to replace the standard UHP lamp, and his conclusion was that LEDs were at least a factor of 10 away from being usable in projection systems. So I dismissed the idea of LEDs as a projection light source and did not think of it again until I was asked to work on this report.

Technology marches on, however. In the semiconductor industry, they have Moore's law; in the LED industry they have Haitz's law. Both "laws" are phenomenological predictions of future performance based on past experience. Moore's law predicts that transistor density in ICs will double every 18 months while Haitz's law predicts that the luminance of an LED package will double every 18–24 months. Moore's law has perhaps another two decades to go before it runs into limits set by physics. LED experts have warned me that LED output is much closer to the theoretical limits. Even so, when comparing the predictions of Haitz's law to the theory of LEDs, I came to believe it should be possible to match the prediction of Haitz's law through at least 2009, and perhaps for a couple of years after that.

LED projection is possible today for certain specialized markets. For example, Mitsubishi has been planning to release a "Pocket Projector" late in 2005. Mitsubishi has been cagey about its exact light output, but it will probably be in the 30–40-lm range.

When will LEDs enter the projection mainstream? The lowest lumen requirement for a current mainstream projector is in consumer rear-projection television (RPTV), which requires at least 350 lm to be interesting to manufacturers or consumers. Current LED demonstration projectors – built with selected LEDs and hand-tweaked in the lab – can produce 200 lm or more. But by 2008, it should be possible to manufacture a 350-lm DLP RPTV system in quantity with LED illumination. Because of their polarization requirements, LCD and LCoS projectors with LED light sources will not be possible for a year or two after that.

When LED projection arrives, it will bring the consumer three main benefits. First, the colorimetry, particularly the red, will be improved. Second, the LEDs will last the life of the set and not need replacing. This is a major issue with consumers – although many of them do not realize it yet – because lamps in RPTV sets need replacing every 2–3 years, at a cost of $300–450 each. Finally, the manufacturer's bill-of-materials (BOM) cost for the LEDs will be lower than for the UHP lamp, and much lower than the cost of a xenon lamp with colorimetry equivalent to that of the LEDs. There is a bright future ahead for LEDs in the projection business.