|
An evening panel discussion, which talked about the "Lessons
Learned in the Commercialization of Microdisplays," was a watershed
of information about what works and what doesn't in the process of bringing
a new display technology into mass production. The panelists were chosen
to reflect a diversity of perspectives, and each has broad experience
and battle scars to show for their efforts. The panel consisted of:
Lewis Collier, CEO, DisplayCheck (www.DisplayCheck.com)
Fred Hammett, President, International Marketing and Distribution (IMAD)
Michael Newell, Director of Product Development, Advanced Digital Optics
(www.advopt.com)
Mike Scarpa, Manufacturing Development Manager, Corning Precision Lens
(www.corning.com)
Larry Hornbeck, TI Fellow, Texas Instruments (www.dlp.com)
Jong Tae Lim, Senior Researcher, Samsung Electronics (www.samsung.co.kr)
The panel members have participated in the commercialization of new
microdisplay technology at the device, engine and final product level,
as well as from an Asian customer/supplier and package and test perspective.
Each offered different viewpoints in their presentations and then engaged
in a lively interaction with audience members, such that all understood
the issues much more clearly. By the end of the evening, some themes
were apparent.
Leadership: The leadership that Texas Instruments has shown
in its commercialization of DLP microdisplay technology was well respected
by the panel and audience members. Earlier in the conference, Larry
Hornbeck, generally acknowledged as the inventor of the DLP technology
back in 1987, gave an overview of the development process that led TI
to commercial shipments of DLP systems in 1996. He noted that in 1992,
TI made the decision to invest the resources to fully commercialize
the DLP technology. While TI does not divulge the cost of its research
and development efforts, industry observers think a $1B is not unreasonable.
"I can guarantee you," said Hornbeck, "that if we were
making that business decision in today's market and financial environments,
we would not commercialize the DLP technology." Whether this decision
would be based on today's economic conditions or the hindsight of the
true cost of commercialization was not clear, however.
TI's strategy included development of DLP devices, supporting ASICs,
engines and reference projection systems. It provided these reference
designs to customers early on, initially supplying the projection engine
and slowly moving toward providing a chip & board set, and now a
component chip set.
Over the last year, it has made a huge effort to bring down the bill
of materials (BOM) cost for a rear-projection TV in order to offer a
competitively priced set. To do this, it has worked with many of the
component suppliers throughout the projection supply chain to set price
targets and help solve other infrastructure issues to reach its objectives.
Simultaneously, it has worked with a number of TV set makers to ensure
that products will reach market with the performance and pricing that
will create sales. This effort has culminated with Samsung's first DLP-based
RPTV products now on the market - with more coming soon from other suppliers.
The leadership of Sony and Epson in the commercialization of high temperature
polysilicon (HTPS) microdisplay technology has been a bit different.
As vertically integrated companies, both were able to develop panels,
engines and final projection products and bring them to market under
their own brand names. They worked with Japan-based suppliers to establish
the component supply and manufacturing infrastructure.
With LCOS, no dominant leader has yet emerged. At last count, there
are perhaps 20 companies in various stages of commercialization of LCOS
panels, yet none tower head and shoulders above the others. All are
investing in LCOS, but none at the multi-hundred-million dollar level
that will probably be needed.
So who could lead and why aren't they? The obvious leader that most
panelists pointed to was JVC. But JVC does not appear to be aiming development
of its commercialization efforts at the mass consumer TV market. Instead,
it is focused on the high end of the market where projection systems
cost more, but fewer are sold. JVC's approach appears aimed at commercializing
LCOS technology in a segment where it does not have to compete toe-to-toe
with DLP and HTPS. Once established, then it can contemplate a move
toward more mainstream applications. While this may work for JVC, it
will not garner large market share for LCOS systems - something that
many feel is necessary to sustain a manufacturing infrastructure.
Hitachi is often mentioned too. It is a component and semiconductor
company that manufactures front projectors and has a strong RPTV presence.
It could become the LCOS leader but, so far, has shown no signs of this.
The same goes with Philips. It has component and semiconductor technology,
builds engines and projectors, and has a full big-screen TV lineup through
its consumer electronics division. It is investing in LCOS, but like
Hitachi, seems to be pacing its investments. Both are big enough to
solve problems and become market leaders, but we don't see high-level
leadership and the dollars to match, from either company, yet.
Three-Five must be considered. While it has the desire to be a leader,
and is certainly leading in the test and measurement arena, most agree
it just does not have the financial resources needed to spend its way
into a market dominant position - ala the TI model.
UMC, which provides backplane foundry services for many and is in partnership
with Aurora to produce LCOS panels, could be a leader too. But its involvement
with the LCOS Taiwan Consortium, which was supposed to spur commercialization,
has not borne fruit.
There are other players in the wings that we can't talk about yet that
have the potential to be dominant LCOS market leaders, but none are
ready to make a move.
Can a final product company champion LCOS and help drive commercialization?
Samsung could, but it appears to be more focused on DLP commercialization
right now. InFocus could, but with a strong line of DLP and HTPS-based
products, it probably doesn't see a compelling reason to take a big
risk on LCOS. Thomson/RCA could, and started down this path, but has
pulled way back now. We suspect it just did not see a path to a low
enough bill of materials with LCOS.
Can the horde of Chinese TV makers become a group of champions for LCOS?
Maybe, but they lack technical horsepower to drive the infrastructure
development at this time, so probably not.
Many observed that pockets of expertise exist within companies. One
has the best panel; another the best optics; a third the best electronics,
TV, marketing or engineering team. But no single company possesses all
of these under one organization. No critical mass exists to drive a
company toward leadership.
So the question remains open - who will lead LCOS into the mass consumer
markets it will need to play in if it is to rise above niche status?
Price & Performance: If LCOS-based projection products are
to compete in the market, they must offer performance matched to price.
Thomson's L50000 LCOS TV at $8K does not offer the level of performance
most would expect in a TV priced at this level. Clearly, high contrast,
good color saturation, minimal artifacts, low black levels and good
uniformity are key elements for quality microdisplay-based TV sets.
Panasonic offers HTPS sets and Samsung offers DLP sets in the $3,500
to $4,500 price range that fit the bill. These are at the high end of
the range of consumer pricing, but are expected to come down quite rapidly
over the next few years. These sets use mature engine architectures
and have an established supply chain. The price-down paths are there.
The prospects for LCOS are less clear. First, while the three-panel
ColorQuad engine design is probably the standard architecture at this
time, it is expensive, so getting the BOM for the engine reduced is
a key need for those in the LCOS camp. There continues to be tremendous
innovation in this area, although some, like Chuck McLaughlin, call
it anarchy. The result could very well be high-performance reduced-cost
engine designs for LCOS systems. In addition, architectures that feature
two- and one-panel designs will help lower BOMs even further, as will
the move to smaller-sized panels. But smaller panels require a trade
off in terms of light throughput. "You can't cheat the etendue
God," quipped IMAD's Fred Hammett. Performance is another issue.
It can be very high, but only if expensive components are used. Market-needed
performance with market-needed BOMs remains a challenge for LCOS.
In real estate, the watchword is location, location, location. But for
microdisplay-based systems, the mantra is price, price, price, says
ADO's Michael Newell. "We must work in an environment where next
year's engine price will be 50% to 70% lower than today's prices. And,
customers want free development, Cadillac performance and Yugo pricing.
That's tough." The other panel members concurred with this.
Believers think the solution is within grasp, but the train is leaving
the station. Unless LCOS is on it, it may not get a chance to compete
well against HPTS and DLP in mainstream projection markets.
In the just-released "Microdisplay Forecast & Company Profile
Report" from Insight Media and McLaughlin Consulting Group, microdisplay-based
technology is forecast to rapidly replace CRTs in rear-projection TV
sets, rising to a 70% penetration level by 2006. But the lions' share
of those sets will be HTPS and DLP based.
Partnerships: Choosing the best suppliers and partners seems
obvious, but can sometimes be tricky. Larger companies may not be willing
to make a big bet on a small component supplier with a critical part
- even if it is outstanding technology. No one likes to sole source
anything, yet TI was able to convince dozens of customers to do just
that.
Partnerships can come unraveled for a number of reasons. If problems
or delays ensue, finger pointing can begin.
Most would agree that the entire LCOS industry has not matured as fast
as hoped. Mass production has been just around the corner for three
years, but has not yet arrived for the projection segment.
Where it has seen success is in viewfinders where Displaytech and Kopin
are clearly in a mass production mode. Kopin forged Taiwanese partnerships
early, but also has critical manufacturing operations in the U.S. Displaytech
made a smart partnership with Miyota, a major supplier of CRT-based
viewfinders, which jumped on an opportunity to have the next-generation
technology in-house.
Many view the RPTV market and the interest of Taiwanese component and
Chinese TV makers as the keys to success for LCOS. But the Chinese TV
makers are hedging their bets by working with HTPS, DLP and many LCOS
suppliers. "These Chinese TV makers don't have the in-house expertise
to fully develop and commercialize these technologies," says IMAD's
Fred Hammett. "If LCOS wants to win here, a total system solution
must be presented to these companies."
There will always be problems and triumphs in the supply chain and customer
relations, but winning combinations have not yet emerged in LCOS projection
products. Vertically integrated companies can have an advantage, but
this is no guarantee for success either.
Manufacturing Infrastructure: The manufacturing infrastructure
to support mass production of the microdisplay and the entire projection
or near-to-eye system has to be in place to realize success. For LCOS,
most of the infrastructure is headed for Taiwan and China, but implementing
this has proved more difficult and time consuming than anticipated.
"As you are building up the infrastructure while in low-volume
production, components cost more and there are limited suppliers,"
notes DisplayCheck's Lewis Collier. "In addition, many have adopted
the fabless manufacturing model of the semiconductor industry, but I
am not sure this will work well for LCOS manufacture. Partnership fabrication
or outsourced design engineering with in-house fabrication seem to be
the prevailing choices."
Standards: With 20 companies developing LCOS solutions, there
are no standards. Each display size and package are different. There
is no standard engine architecture or interfaces between components.
This lack of standardization can be costly since additional development
efforts are needed as plans change.
Consortiums and standards committees can help solve this, but only if
individual company interests are managed to help the overall industry.
As mentioned earlier, the LCOS Strategic Partnership in Taiwan is supposed
to be doing that, but does not appear to be making much progress.
Individual companies can create the industry standard by fielding components
and systems that dominate the market. But that brings us back to the
leadership question for LCOS. With leadership or a product home run
will come standards.
System Level Testing: "You can't test the full microdisplay-based
system soon enough," lamented CPL's Mike Scarpa, referring to its
efforts to build the engine for the RCA LCOS TV. Problems that were
not evident at one level were very evident at other levels. Therefore,
it is critical to test components including lifecycle and environmental
testing early on. "System performance does not equate to the sum
of component performances," observes Scarpa.
ADO's Newell observed that "On-screen image performance is the
result of the optics, mechanics, electronics and screen technologies.
Deficiencies in any of these areas can lead to non-optimal image performance.
What is needed are bright and clear images."
Many LCOS development teams are learning this lesson the hard way. It
is not enough to design a panel, or even an engine reference design
for that matter. A team effort from component and subsystem suppliers
working in tandem with the final product maker/brander is absolutely
essential to commercialize a new microdisplay technology. TI did it
with DLP and Intel did it with microprocessors. It can be done with
LCOS, but it will take commitment and lots of funding. Who will take
the risk?
Managing Change: In the process of commercializing a new microdisplay
technology, and associated products, technology often changes rapidly.
This is certainly the case with all microdisplay technologies today.
This can have a serious impact on the development and commercialization
process.
For example, several years ago, LCOS developers focused on 0.9- and
0.97-inch panel sizes aimed at mainstream presentation markets. But
as these panels were slow to enter the market, mainstream panel sizes
began to shift toward 0.7-inch sizes. Just as the larger LCOS panels
were being perfected and readied for commercialization, they were viewed
as trailing edge and not worthy of a price premium. Therefore, many
early generations of LCOS panels never reached a revenue-producing stage
to provide funds for next-generation development. This meant alternative
sources of funding were required, increasing the risk of development.
The same thing has happened at the engine level. Architectures and components
designed for a particular panel are quickly obsolete, as new technology
is developed to improve performance or lower costs. Or, limitations
in performance may be realized late in the development cycle.
Change is inevitable, so managing it is critical. All of these factors
delay product introductions and can require costly retooling and write-offs
of inventory. As Mike Scarpa advises, "Developers need to monitor
market and technology changes and institute formal change management
early. Pace investment with market acceptance of the imager performance."
Managing Expectations: With change comes new levels of expectations.
The performance bar may keep rising during the commercialization cycle,
often because market conditions demand it. "A couple of years ago,
contrast of 300:1 for LCOS systems looked fine, but it is not today,"
notes ADO's Michael Newell.
The consequence can be delays, redesigns and much higher costs. In addition,
responsibility for component and subsystem quality and performance can
become more complex, and hidden costs may emerge. Communication is key,
but no panacea for success.
So what does all this mean? Probably that the bigger LCOS players will
continue to go it alone, picking up whatever market share they can while
waiting for the industry to mature, or a new dynamic or opportunity
to arise. This industry is changing so quickly, however, that things
could be quite different in 6-12 months. Stay tuned - Chris Chinnock
DisplayCheck, Lewis Collier, 401-392-1023, lcollier@DisplayCheck.com
IMAD Fred Hammett, 415-897-2703, fred_hammett@msn.com
ADO, Michael Newell, 805-497-1771, mnewell@advopt.com
Corning Precision Lens, Mike Scarpa, 513-943-2079, scarpamc@corning.com
Texas Instruments, Larry Hornbeck, 214-567-9844, l-hornbeck@ti.com
Samsung Electronics, Jong Tae Lim, [82] -31-200-8975, jongtae@samsung.co.kr
###
|
