Not long ago, consumer electronics giants were clamoring to get out of the microdisplay market. Sub-par technology, combined with high manufacturing costs and a stiff price tag, all contributed to the demise of the microdisplay.
No longer on the back burner, today's microdisplay market is on fire, and they're back with a vengeance. And at the center of it all is the realm of microdisplays that drive high-definition (HD) rear-projection TVs (RPTVs) with screen diagonal sizes upwards of 50 in. and an equally impressive price tag.
Three technologies contend for supremacy in this space: LCDs, Texas Instruments' digital light processing (DLP), and liquid crystal on silicon (LCoS). A smaller fourth category includes near-to-eye displays for camera viewfinders, headmounted displays, and video games (see "The Other Microdisplay," p. 60).
Designers of HD RPTV home theater systems have a wealth of choices at their disposal. In varying degrees, each method offers vast improvements over previous technology, from higher resolutions to brighter images, lower maintenance requirements, increased viewing angles, and much slimmer sizes and lighter weights.
The different microdisplay definitions may seem confusing. Some experts say that microdisplays have a diagonal of less than 1 in. Many of these devices are direct-view displays. By that definition, some cell-phone displays fit into this category. Yet most experts agree that microdisplays use a projection system that provides high-density content of at least 800 by 1000 lines/in. and a magnification system that enlarges the image from a tiny light source.
Most microdisplays are used in HD RPTVs and other projection products, like home and office projectors. Unlike LCDs, plasma panels, and other front-view displays, though, large rear-projection displays don't cost disproportionally more. That's because LCD and plasma panel manufacturers must construct more pixels on larger sheets of glass to create larger displays, which isn't the case with rear-projection display manufacturers.
Market research firm Quixel Research reports that the U.S. HD RPTV microdisplay market surged from $4.7 billion in 2004 to $5.8 billion in 2005. Many other estimates peg the RPTV microdisplay market as high as $8 billion. Though that figure pales in comparison with the overall HD RPTV market (which includes the venerable CRT), microdisplay sales are definitely on the upswing (Fig. 1).
A microdisplay uptick is remarkable, considering Intel Corp. and Philips both gave up on LCoS a couple of years ago. Thomson (under the RCA brand name) and Toshiba (which used Hitachi LCoS panels) gave up, too. A number of factors contributed to these decisions, but they were due in part to technical and scaling limitations, such as display lifetimes. At the time, it seemed that microdisplays were doomed.
So what changed? First, the technology is better. Use of high-intensity LEDs as light sources has led to higher resolution and brightness levels for HD RPTVs. Optics subsystems improved their performance, and manufacturing glitches were overcome. And prices dropped—considerably. As a result, top-tier Japanese HD RPTV suppliers like Sony and JVC are making a strong push into the LCoS microdisplay market.
Another change is standardization. Microdisplays use an optical subsystem between the light source, or "light engine," and the image seen by the viewer. Vendors other than the display chip manufacturers supply many of these optical systems, though, with few standardized formats available. With demand on the rise for the latter, we're now seeing efforts under way for a more standardized approach.
But perhaps the biggest turning point was the market introduction of 1920- by 1080-pixel progressive resolution (1080p) microdisplays. Also referred to as "true HDTV" and "ultra HD," the 1080p progressive format successors to the 720p displays provide not only the maximum lines of resolution possible, but also the best image quality in the HDTV standard.
The average price for LCoS-based HD RPTVs with screen diagonals of 50 in. or greater has fallen from well over $10,000 to $2000-$3000. The 50-in. Sony KDSR50XBR1 and 60-in. KDSR60XBR1 now cost about $2349 and $2999, respectively. Both are based on Sony's SXRD (Silicon Xtal Reflective Display) LCoS technology. JVC's HD-ILA LCoSbased HD RPTVs dropped in price, too. Its 52- and 61-in. H52G786 and H61Z886 cost $1999 and $2250, respectively.
At this year's Consumer Electronics Show (CES), Akai showed off 46- and 52-in. prototype HD RPTVs driven by LEDs with projected prices of $1799 and $2499, respectively. The prototypes offer 1080p resolution, 500 nits of luminance, a 3000:1 contrast ratio, and a National Television Standards Committee (NTSC) color gamut of 140%. Presently, no other display technology can deliver 1080p resolution in a screen of that size at such low prices.
Akai's new offering highlights one growing trend in light engines for HD RPTV: High-intensity LEDs are being used as light sources instead of the usual ultra-high-performance (UHP) high-intensity lamps. UHP lamps have a limited lifetime of only 5000 to 6000 hours and cost hundreds of dollars to replace, while LED lifetimes span 20,000 to 30,000 hours. LEDs also cost less and offer better color gamuts than UHP systems. They're being used as light sources in all three major microdisplay HD RPTV technologies.
This month, Samsung will release 56-in. RP HDTVs that use very bright LEDs from Luminus Devices Inc. Samsung's Phlat-Light LED chip sets are based on Luminus Devices' proprietary photonic lattice technology. The company says they're the only commercially available solid-state light sources bright enough to illuminate large-screen HD RPTVs.
Samsung has yet to release detailed specifications about the chip set. But it created quite a stir at this year's CES, winning Insight Media's CES 2006 Best Buzz Award in the category of Best New Enabling Technology.
On another front, Novalux Inc. entered into a joint development and licensing agreement with Seiko Epson Corp. to use NECSEL-based (Novalux Extended Surface-Emitting Laser) devices for HD RPTVs. Using its expertise in optical engines, Seiko Epson will integrate Novalux's products into Seiko Epson's 3LCD projectors and TVs.
A closer look at LCD, DLP, and LCoS technologies illuminates their strengths and weaknesses. All three differ in how they control the light path, the number of chips needed, the controlling element's response time, and the size of the pixel attainable.
LCDs use three light sources for red, green, and blue. Dichroic filters pass some wavelengths of the light and split the light from a projection lamp into the three primary colors. The light passes through the LCDs and recombines to form the fullcolor image (Fig. 2).
LCDs are transmissive displays. Some, like those from Sony, are made on a high-temperature polysilicon process. They control the light passing through them by using the liquid crystals' polarization effects. Because they require three devices to form full-color displays, they're often called 3LCD displays.
TI's DLP technology is completely different. It's a reflective technology comprising millions of micromirrors that deflect light into and out of the projection path (Fig. 3). Instead of shining a backlight directly through the LCD layer like LCD microdisplays,-light in a DLP system bounces off the mirror behind the liquid-crystal material.
TI developed an optical architecture using arrays of red, green, and blue LEDs with brightness that nearly equals lampbased systems (Fig. 4). The company uses a unique feedback algorithm to drive its DLP engine. The algorithm allows control of LED color-shift variations that affect the white point to a tolerance beyond what's detectable by the human eye. This problem for driving HD RPTV systems is caused by higher thermal loading of the LED drivers.
The LCoS microdisplay approach falls somewhere between the LCD and DLP approaches. It's a transmissive display that uses liquid crystals instead of individual mirrors. As the liquid crystals open and close, the light is reflected from the mirror below or blocked. This modulates the light and creates the image (Fig. 5). Like LCD microdisplays, LCoS systems use three light sources or chips for red, green, and blue, which are projected simultaneously on the screen. Some LCoS engines come with a single light source, though the three-chip configuration is more common.
LCD-based microdisplays have the slowest response times, particularly for single-chip configurations. Response times are about 10 ms, compared to a few microseconds for DLPs. On the other hand, LCD and LCoS displays can achieve pixel pitch sizes of about 8 m, compared to DLPs' 11 m.
With its larger pitch, the DLP approach struggles to achieve high-definition resolutions. Manufacturing a larger-size die also adds to the cost and lowers manufacturing yields. However, DLP engines have better contrast ratios than LCD-based microdisplays.
MANY LCoS SUPPLIERS
The number of LCoS suppliers of chip sets and systems entering the market continues to balloon. According to its manufacturer, MicroDisplay Corp., the singlepanel FHD system offers more brightness levels, higher contrast ratios, and faster response times than other systems on the market. Used in 52-in. 1080p HD RPTVs, they deliver 400 nits of brightness and a contrast ratio of 700:1.
The eHD70 eXTREME digital1080p microdisplay from eLCOS also can deliver 4000:1 contrast ratios. It's been tested on JDS Uniphase's UltreX projection engine, which uses JDSU's proprietary Birefringent Compensating Element (BCE) for contrast enhancement. Also, JVC showed off its proprietary second-generation HD-ILA family of LCoS TVs last year. The company's Direct-Image Light Amplifier (D-ILA) technology powers the three-chip microdisplays.
A STRONGER OUTLOOK
Microdisplays for HD RPTVs face a much better future today as opposed to a couple of years ago. This is particularly true for LCoS and DLP technologies.
Large companies like LG Electronics plan to introduce 71-in. diagonal LCoS models, with SpatiaLight Inc. supplying LG with a tailored LCoS chip set. Hitachi, which previously dropped out of the LCoS market, waits in the wings with 60- and 70-in. LCoS models. Syntax Corp., a supplier of low-cost direct-view LCD TVs, also has a 50-in. LCoS model on the market, acquired through its controlling interest in Brillian Corp.
LCoS pioneer Brillian has been ramping up production of 720p and 1080p LCoS TVs on its Gen II LCoS process. According to the company, this process significantly decreases manufacturing costs.
In addition, Canon, which has investedin surface-emitting display (SED) technology, indicates it wants to enter the LCoS market for TVs with a system it presently uses in front projectors.
Still, Texas Instruments' DLP technology is a major force in many HD RPTV applications, with a market share of at least 25%, according to market analysts. Constant improvements to DLP technology have brought down the prices for DLP-based TVs. In fact, it's widely used on HD RPTVs made by Samsung.
TI's latest DLP improvement, reflected in its Smooth Picture approach, provides higher resolution and lower cost than the original DLP engines. Its digital micromirrors are oriented in a diamond pattern instead of the traditional rectangular pattern, while an optical actuator is used to shift the light path by a half-pixel sideways on alternate frames. As a result, the DLP engine can reproduce all of the pixels in a high-definition image using only half the number of micromirrors.
TI also enjoys a smoothly functioning supply chain with light engines based on its reference design and a long list of systems integrators. The supply chain has been a major problem for LCoS manufacturers. Typically, they deal with many different IP-based light-engine designs that have little commonality in available parts and manufacturing methodologies.
The fact remains that HD RPTVs are comparatively large and bulky compared with large-screen LC plasmapanel displays and have less perceived image quality. But that's the cost of very large screens, and apparently many home-theater enthusiasts don't mind this tradeoff.