Cree’s CEO Chuck Swoboda delivered a keynote address that chronicled the company’s focus on LEDs for general lighting, provided an update on the latest LED technology, and predicted a quicker uptake of solid-state lighting (SSL) than many analysts expect. Swoboda noted that LED epitaxy technology is nearing peak efficiency but suggested gains will still come in light extraction, phosphors, and packaging.
Of the major LED chip makers, Cree alone is almost exclusively focused on general illumination applications for LEDs. The company does supply LEDs into other applications such as mobile phones, but R&D is specifically targeted at general lighting.
Home Depot sells the Cree CR6 downlight for $50
Cree CR6 downlight
Swoboda discussed the decision the company made about five years ago to focus on general lighting. He said very succinctly, “It seemed like the biggest market and the hardest to do.” And there is no tempering of his expectations and goals. He predicted that Cree can obsolete every type of light source over time with LED technology.
The CEO is invariably frank about the difficulty of proliferating SSL technology in lighting as opposed to other areas. He noted that when you supply an LED to a company developing a mobile phone, the application is well defined and the engineers know precisely how to design in the LEDs. However, said Swoboda, “Lighting is a generic term that has hundreds of different unique problems that are often solved differently.” The luminaire and fixture makers face that difficulty in transitioning to LED light sources and Cree understands that it must help them to succeed. Swoboda said, “Every application is different and we have to spend more money to make it easier for our customers to use our product.”
The company is pursuing several paths to help the customer, for example building out a global network of customer application centers. And Cree is developing a much broader product line than you will find from most LED makers, said Swoboda, with the intent of matching the components and packaging to unique application requirements.
The system side of Cree
Swoboda also pointed to the company’s system business as being very important in driving success both in terms of component development and charting a course that the component customers can follow. He explained that Cree bought the LED systems business that is now Cree LED Lighting to help move the market as a demonstration vehicle.
The company is certainly selling luminaires and striving to make the systems business a profitable one according to Swoboda, but the larger focus is on accelerating the LED component sales that make up two thirds of Cree’s revenue. He said that the company’s lighting business is being used to establish a roadmap that other luminaire makers can follow.
Swoboda also discussed the value that the systems business has brought to the LED design and manufacturing process. “You would be surprised what the benefit is when your R&D guy that has to make an LED component is forced to sit across the table almost every day and argue with his sister division about why his LED doesn’t solve a lighting problem,” he said. “What you quickly find out is the semiconductor guy quickly learns that lighting is not like every other market.” He believes the interaction has led to the component engineers thinking differently about the LED, and in turn to components that solve general-lighting-centric problems.
The goal of course is quicker adoption of LEDs into lighting. According to Swoboda, LEDs will comprise a meager 4% of the general lighting market in 2010. During his Q&A session a member of the audience asked Swoboda to comment on the premise that LEDs would pervade 80% of the lighting market by 2020. Swoboda quickly said, “I hope it won’t take that long. I want to know how we can do it in 2015.”
Driving LED adoption
Swoboda discussed what he sees as the key to driving LED adoption -- looking at the technology and system design as intertwined elements. He suggested that many see LED system design, LED components and modules, and LED chips as three separate pieces rather than considering the implications that any one of the pieces has on the other two.
The SSL industry also needs to recognize where the opportunities lie according to Swoboda. He noted that fixtures and luminaires that target industrial and commercial installation currently afford the greatest market potential saying, “This is not about the light bulb.” He was quick to note that Cree is developing technology that enables retrofit light bulbs, but that the application focus would give preference to the largest applications.
The path to success sounds straightforward. Swoboda said, “Our goal for LED lighting is this: To help people design products that are better in every single respect to [legacy technology].”
Swoboda acknowledged that costs are a problem saying, “Today we ask our customers to buy a product and live with a ten year payback.” Were a vendor to approach Cree about buying into technology with a ten year payback for use in its factory, Swoboda said he would tell that vendor to “come back when it’s two.” While acknowledging the cost issue, Swoboda also stressed that SSL success has to come from value, saying “It’s not meant to be cheaper than the light bulb it’s meant to be better.”
LED component advancements
The road to superior lighting products invariably runs through advances in the LED component technology – and that means brighter, lower-cost LEDs. Swoboda recalled that, “When we got up to about 100 lm/W, many people said there was no reason to make brighter LEDs.” The point is that the efficiency was passing that of legacy sources, so some people questioned the need for even brighter LEDs.
The Cree XM-L LED delivers 160 lm/W
Cree XM-L LED
Swoboda reminded the audience, however, that delivering good LED lighting is very much a system-level problem. He said, “Brighter LEDs are actually the way to make the most cost-effective lighting product” because they provide flexibility in the design. For instance, a design can reduce drive current -- simplifying the thermal design and extending the luminaire lifetime. But Swoboda also noted that “In some cases, if you don’t have color control perfect, you don’t have a product.”
Swoboda repeatedly emphasized the myriad lighting applications, the need for a broad product line, and the need for low component costs, but these factors seem to be at odds with one another. Historically low cost equates with high-volume manufacturing of a single product yet Cree has a broad product portfolio. Swoboda said, that Cree can manufacture a die in high volume to help drive costs down while using different package designs to yield components for different applications.
Swoboda did note that a move to larger wafers would soon start to drive cost down since a 6-inch wafer provides double the number of die compared with a 4-inch wafer. He pledged that Cree would be first in volume production with a 6-inch line for nitride (InGaN-based) LEDs.
There is a broad industry push to 6-in wafers underway. Rubicon Technology announced back in August that it was supplying 6-inch sapphire substrates for volume production of LEDs under a one year contract. It’s widely believed that the unnamed customer is an Asian manufacturer. Still the manufacturing message from Cree is clear that it has a roadmap toward lower component costs that’s rooted in semiconductor technology and that’s a proven way to drive costs down.
Indeed Cree is confident that LEDs are no longer an obstacle to cost-effective SSL. The industry does need to continue to drive the cost out of the system design. But Swoboda suggested that branding and marketing are as significant roadblocks to success as is the technology.
Cree CEO Chuck Swoboda
Cree CEO Chuck Swoboda
Swoboda pointed out a common mistake people in the SSL segment are making. He said “They either want it to be really expensive as a super-premium product or they want to make it as cheap as what it’s replacing. It’s neither. It’s a value sell.”
To make his point, Swoboda discussed the widely-publicized partnership between Cree and Home Deport in selling the CR6 down light. Cree approached the move as an awareness campaign and Swoboda said, “Our only objective was to get them to put the demo in the store.”
Cree hoped to test what would move the SSL market. And while he didn’t share details on product sales, Swoboda asserted that the product has done very well at a price ($50) that was broadly considered too high. He suggested that once people can see and touch SSL technology, they understand the value.
“I think there is much more demand for LED lighting in the near term than anyone has thought about,” Swoboda said. “If the customer is comparing [LED lighting] to CFL it’s an even easier sell.” Earlier he had noted that the CFL had taught consumers that energy-efficient lighting was equivalent to a compromise in quality but that LEDs will change that perception.
Efficiency and applications
The Q&A session following the presentation was dominated by a discussion of LED efficiency, applications where LEDs are winning, and the challenges for luminaire makers.
Swoboda affirmed that Cree is shipping the XLamp XM-L family of LEDs that he believes offers an industry-best 160 lm/W. He added that the company is doing better in the lab with prototypes operating in excess of 200 lm/W. Asked when Cree would ship such products, Swoboda would not commit but noted that the company has a track record of delivering production components 18 months to three years after lab demonstrations.
The inevitable next questions centered on what would deliver the gains in brightness. One questioner suggested that bigger die would be the answer. While acknowledging that Cree has some LEDs based on larger die including current versions of the XM-L, Swoboda suggested that “die size is a short term solution.” In a separate presentation, Cree applications engineer Yuming Chen detailed the fact that larger die size creates the need for much larger secondary optics to prevent the light losses associated with the etendue effect.
Instead, Swoboda believes that there will be advancements in chip design, phosphor technology and packing that deliver better efficiency. He stressed that by chip design he does not mean the actual epitaxy process. He believes that the industry is nearing peak efficiency in semiconductor devices. But he believes that there are other elements of chip design that can improve light extraction. Swoboda predicted that Cree will deliver LEDs with efficiency in the 220 to 230 lm/W range while noting that the company will be nearing the theoretical limit of converting energy into light at that point.
The stress on efficiency is as much about thermal issues as it is brightness per se. As discussed in our prior article on the keynote, brighter LEDs simplify system design. Swoboda said, “When that thermal load comes down, everything gets easier.”
Swoboda was also asked about a number of applications and when a “tipping point” might be reached. He believes that street lights tipped when LEDs hit 100 lm/W. The same level has allowed “a segment” of the retrofit bulb market to tip. But he noted that “converting a category of lighting is a really long process.”
Swoboda was quick to highlight a potential market for SSL that has not been viewed favorably of late. He boldly predicted that LEDs would succeed sooner than expected in linear-fluorescent retrofit applications.
Fast design cycles
An audience member, clearly representing a luminaire maker, expressed frustration at the fact that the LED market is moving so quickly with new products and packages coming to market constantly. The questioner noted that before the luminaire maker can get a product to market and into volume manufacturing that the LED maker invariably had a new brighter product – perhaps in a new package – that was replacing the old one.
Swoboda sympathized with the plight of the luminaire maker while opining that companies should go ahead and complete each design rather than stopping and changing light sources in the middle of a project. He used an LED street light manufacturer as an example of why. He noted that the company is on its fifth generation of its street light design, and that much more than the LED components were changed in each generation. He said, “They had to change [LED] platforms but they changed everything else each time too. Their product today is three to four times brighter, twice as efficient, and one fifth the cost as in gen one.”
Swoboda made a bold statement that applies both to luminaire makers and to the LED component makers. To succeed he said, “You’ve got to want to obsolete your own products.”
Swoboda said, “I don’t think there’s any doubt that the world is going to LED-based lighting technology.” And he succinctly summed up the reasoning behind Cree’s general-lighting focus saying, “If you get there early it works out a lot better than if you get there late.”