New Projection Technology Around The Corner
From Silicon Light Machines
by Michael Karagosian
© 2000 MKPE Consulting All rights reserved worldwide
Published in the November 2000 issue of System Contractor News
If you haven't heard about Silicon Light Machines (SLM), and you're interested in high
quality projection, then you'll want to read on. The story of SLM and their recent
acquisition by Cypress Semiconductor is as interesting as their technology.
Founded in 1994 by David Bloom, a former professor at Stanford University, the original focus of
Silicon Light Machines was the development of the Grating Light Valve(tm) (GLV(tm))
technology for high quality display applications. GLV technology is based on a
"grate" of ribbon-like mirrors. Imagine a comb, with each tooth of the comb a
flexible mirror, and the comb itself resting upon a mirror. Now imagine a beam of
single-wavelength light shining on the comb and the exposed strips of underlying mirror.
Let's further imagine that everything lines up so that without flexing the teeth of the
comb, the reflected beam is uniform without interference. By flexing the comb elements
downwards by a quarter of a wavelength, we can effectively cancel the beam through wave
interference. This is the essence of GLV(tm) technology.
GLV(tm) technology significantly departs from the moving-mirror or shuttered mirror
methods of light control. In moving mirror technology, shades of gray for each pixel is
created by flipping a mirror thousands of times a second. In shuttered mirror technology,
shades of gray are created by gating an LCD shutter in front of the mirror. GLV
technology, though, is more analog in nature. Flexing the mirrors anywhere between the
rest position and 1/4 wavelength provides the means to create an infinite range of grays.
Further, the response of a GLV array is close to the logarithmic response of the human
eye, making this an ideal technology for displays.
One of the problems of moving mirror and shuttered mirror technology is the difficulty
in scaling them. For instance, doubling the image resolution in current DLP moving mirror
technology requires creating a chip with double the mirrors, negatively affecting both
yield and cost. GLV technology gets around the scaling problem by imaging only a vertical
column of pixels at a time. As progressive vertical columns are generated, they are
scanned across the screen by a much slower moving mirror. With the ribbon mirrors capable
of very high speed operation, the vertical scan rate can be quite fast. GLV HDTV displays
have been demonstrated with 100Hz refresh rates.
Flexible aspect ratios are a side benefit of this technology. If a wider image is
needed, the vertical scan width need only be extended. No additional mirrors are needed.
If more resolution is required, the vertcal array of ribbons is extended. While this also
requires a doubling of mirrors, the numbers tell the story. For GLV, doubling the
resolution of a 1920 x 1080 HDTV display requires an increase in the linear GLV array to
2160 pixels from 1080, an increase in pixel count for the array of 1080 pixels, while
increasing the total pixel count of the display to 4M pixels from 2M pixels. For DLP
Cinema, doubling resolution of the 1280 x 1024 display requires a new array of 2560 x 2048
pixels, an increase in pixel count for the array of 1.3 million. Get the picture? Doubling
resolution with GLV can take place at literally a tiny fraction of the cost for a similar
doubling for moving mirror or shuttered mirror technology.
There was a small but non-fatal flaw to SLM's plans to develop a digital cinema
projector. Since analog gray scales can only be created with a single wavelength of light,
lasers are required as the light source. Obviously, for RGB color, lasers of three
wavelengths are needed. SLM has depended on outside sources for the lasers, which
unfortunately stood in the path of high brightness projectors. Obtaining lasers of the
correct wavelength and intensity to support digital cinema has been a problem, as the
market size is too small to excite a laser manufacturer into developing a special product
for SLM. Developing smaller displays appear to be well within reach of available laser
On the other side of the coin, using lasers makes GLV projection technology very
efficient. Since this is a reflective technology and there are no stray light frequencies
to heat up the GLV device, very high light intensities can be modulated, making this an
ideal choice of technology for the big screen. Images produced with GLV technology are
said to have high contrast, wide dynamic range, and deep, accurate colors, making it all
the more attractive for high quality digital projection.
Given this knowledge, Sony attempted to purchase Silicon Light Machines earlier this
year. For whatever reason, their offer was not accepted. However, in July, Sony entered an
exclusive licensing agreement with Silicon Light Machines to use GLV technology for
displays. Interestingly, Sony has announced that they will first develop front projectors
based on GLV for the industrial market, and later thin rear projection devices for the
consumer market. Since front projectors require more powerful lasers, Sony may have some
laser technology of their own up their sleeve. If this technology proves successful, one
can easily imagine it as a replacement for the aging Trinitron CRT, making a major impact
on the consumer display market.
The original Grating Light Valves were manufactured by Cypress Semiconductor.
Interestingly, this was a backdoor project within Cypress, taken on secretly to fill some
unused capacity of one of their fab lines. A former process engineer of Cypress was hired
by Silicon Light Machines, and a combination of his skill and relationship within Cypress
allowed Silicon Light Machines to create GLV devices using a standard CMOS process.
Fabricating these chips using a standard CMOS process is a big advantage when it comes to
adding additional circuitry to the GLV chip. Upon learning of this backdoor project, T. J
Rodgers, president of Cypress, was ready to axe it, but first was talked into visiting
Silicon Light Machines. Being a home theatre enthusiast, this was a done deal. Upon
viewing a demonstration of the technology, Rodgers was converted. He became a member of
the board and Cypress Semiconductor became an investor in SLM.
Following the Sony licensing deal, Silicon Light Machines changed their focus from
applying GLV technology for display devices to applying it to optical networking products
and printers. One can imagine that modulating light for high-speed fiber optic
transmission to be a natural application for GLV technology. Seeing the fit in their own
product line, and pleased to see the new revenues that the Sony deal produced, Cypress
immediately offered to buy SLM. Their offer was accepted in late July. As Rodgers himself
has pointed out, this is not one of those "grand strategies" that one plots and
sits in wait for the right moment to execute. The marriage of Cypress and SLM is one of
those deals that just sort of stumbled along, kept alive by a guy who loves quality home