My column "Microdisplay Industry Close To Igniting" (April 3, p. 62) explained how liquid-crystal-on-silicon (LCOS) microdisplays are finally reaching the point of initial mass production. This technology will open up new product categories in both front- and rear-projection systems. Many believe LCOS will ultimately become the low-cost solution in a lot of these systems. But to make this a reality, new optical-engine architectures must be developed and optimized for these display technologies.
Impressed, I just returned from the Society for Information Display (SID) conference, which featured a large number of new optical architectures and optical devices enabling the growth of LCOS-based projection engines. These optical engines, which include one or three microdisplays, electronics, lamps, optics, and projection lens, form the basis for many projection products.
For instance, OCLI, Santa Rosa, Calif., presented a three-panel optical architecture geared toward LCOS microdisplays. Called the OCLI Prism, it's a modified Philips prism configuration promising improvements in performance and manufacturing costs.
The Philips prism configuration was designed many years ago to support three-channel camera systems. It consists of three uniquely shaped prisms with multilayer coatings used to separate color. For example, white light enters the prism where the first stack reflects red, but passes blue and green light. The multilayer stack at the next surface reflects blue, but passes green. Two of the color channels also use total internal reflection to direct the light to the end of the prism where an image capture device, like a charge-coupled device, is mounted.
For a projection engine, though, the Philips prism design is used for both color separation and recombination. While the basic architecture is similar to a camera system, a polarizing beam splitter (PBS) is placed in front of the prism assembly. The PBS sends polarized light into the prisms where the multilayer coatings again separate it into red, green, and blue channels. But for a projection engine, a reflective LCOS light-valve is placed at the end of the prism, instead of a CCD. Also, light, whose polarization has been changed by the LCOS display, must travel back through the prism system and out the PBS to the projection lens. This places a constraint on the coatings design and can result in only fair system contrast.
In the new modified Philips design, the third prism (P3) was redesigned in favor of an element that's the same shape as the second prism (P2), resulting in lower system mass and weight, and reduced manufacturing costs. With only two uniquely shaped prisms, factory tooling is reduced by 1/3 and greater economies of scale are possible for the P2/P3 prism. Plus, the design eliminates the quirky compensator foil used with the P3 prism in the conventional design, and allows the use of a multilayer coating for phase compensation, like the other surfaces of the assembly.
Nikon Corp., Tokyo, Japan, showed its version of this new prism design, called the Cactus 700 Light Engine, and claims to have come up with the design independently (but perhaps that's a bit of a stretch). The company developed this engine architecture to support 0.9-in. LCOS panels from JVC, Kanagawa, Japan. Why the engine was never commercialized wasn't disclosed.
Now, Nikon is redesigning the optical components for the 0.78-in. SXGA panels from Three-Five Systems, Tempe, Ariz. The company expects engine developer's kits to be ready by the forth quarter of this year, with mass production in the second quarter of next year. Pricing isn't yet set, but Nikon expects a 60-in. system using the engine to have a retail cost of $4000 to $5000, with the engine at 1/4 to 1/3 of that.
Meanwhile, Advanced Optical Engineering, Westlake Village, Calif., demonstrated another engine design that will be manufactured by Samsung Electro-Mechanics Company Ltd. (SEMCO), Suwon City, Korea. It uses SXGA panels from Three-Five too. This architecture produces a rather unusually shaped engine with color separation and recombination accomplished in a proprietary three-element system. It should be ready for production by the end of this year.
In still another design for a three-panel projection engine, color separation is accomplished by four PBS cubes and a polarization retarder stack that replaces the multilayer coating stack. These elements, made by ColorLink, Boulder, Colo., rotate the polarization of light in one color band while leaving the polarization of the other colors unchanged. The PBS can then pass or reflect the various color bands. The result is the separation in the red, green, and blue channels so that LCOS light valves can modulate the light.
The main advantage of this approach is a low-cost passive retarder stack and the wide acceptance angles it can support. Optical systems with f/2 designs are possible with this approach, which means that faster optical systems can allow more light to be coupled into the system, thus producing brighter, more efficient products. In addition, ColorLink claims that less-expensive glass can be used in the PBS elements, thereby lowering system cost.
Finally, new architectures are in development for systems where red, green, and blue light is sequentially presented to a single LCOS display panel. Displaytech showed a prototype at the SID show. Such designs represent next-generation optical engines and will undoubtedly be the subject of a future column.