PUBLICATIONS : PAPER

Comparison of Microdisplay Based rear Projection Televisions
by Charles McLaughlin | chuck@mcgweb.com | Telephone: 650 323 7155

1. BACKGROUND

During the past decade the performance and price of projection systems based on microdisplays have demonstrated dramatic improvements. Ultraportable front projection systems can deliver nearly 1000 lumens of light at XGA definition and street prices of such systems are rapidly falling to $2,000.

The market leading low priced presentation projectors use two alternative microdisplay technologies: transmissive liquid crystal with a high temperature poly-silicon on quartz backplane (HT p-Si) or the reflective digital micromirror device (DMD)™ , made with CMOS. These incumbent microdisplay technologies are now being challenged by several new approaches, including reflective liquid crystal on a silicon backplane (LCOS) and transmissive liquid crystal with a low temperature poly silicon on glass backplane (LT p-Si).

In addition to new microdisplay technologies, developers have been demonstrating new single imager projector architectures that have the promise of leading to even lower systems costs. While the color field sequential DMD system and the larger 3 to 6 inch spatial color transmissive liquid crystal with an amorphous silicon on glass backplane (a-Si) are the only commercial single imager designs, single microdisplay projectors have been demonstrated using color field sequential LCOS1, spatial holographic microlens LCOS2, and a spatial system than employs color scrolling and a LCOS imager3.

While the presentations market continues to offer a major growth opportunity for the projector developers, the potential use of microdisplay technology in the television market offers, at minimum, an incremental market opportunity, and at maximum, a huge potential additional market. But for the home market to amount to much, the microdisplay based televisions will have to offer much better value to consumers than the current rear projection CRT TVs. Otherwise, rear projection televisions will remain a North American niche market with demand of 1 million units per year and modest growth. Unless the new microdisplay televisions offer better value, their makers will be fighting for a share of a modest market against deeply entrenched competitors

What does better value mean? Published market research as well at several studies completed by McLaughlin Consulting Group4 5 (MCG) indicate a hierarchy of consumer preferences for NTSC televisions. Not surprisingly, for the American market two characteristics lead all consumer preference lists: price and size. As shown in Table 1, secondary preferences include brightness, contrast, image quality, tuning options, cabinet depth, weight, and sound system. Well down the list are features such as power and safety.

Table 1. North American Consumer Preferences for Television Features

Feature Category

Historical Factors
Future Factor

First Order

Price
Screen Size
16:9 video capability

Second Order

Brightness
Image Quality
Tuning Features
Internet interactivity
High definition

Third Order

Weight
Form Factor
Power
Environmental
Safety
Flat Panel

While it is easy to argue that all of this will change in the new world of digital video, there is little consumer data available to support any differences in preferences. Our survey of market insiders indicate that one new parameter may go to the top of the list: wide screen. Here the North American market will be following the trend already apparent in Europe and Japan: wide screen televisions, coupled with digital video sources, offer value. However, it is important to note that consumers tend to compare picture height as opposed to screen diagonal. For example a widescreen television with a 59 inch diagonal has the same picture height as a 48 inch NTSC 4:3 screen and will be perceived my most consumers as being of equivalent size.

A more tenuous opinion of market insiders is that high definition and internet compatibility will also be perceived as adding value, making a progressive scan SDTV or HDTV preferred to an NTSC model. How much consumers will be willing to pay for these features is an open question, but as shown in Table 2, experts believe that consumers can justify price premiums of 30-50% for widescreen HDTV performance.

Table 1. North American Consumer Preferences for Television Features

Nominal Screen Size
Typical NTSC Price
Competitive DTV Price
Premium
NTSC
DTV
 
 
 
 
27"
33"

$230

$290

$60
(26%)
 
 

$490 *

$650

$160
(33%)
32"
39"

$700

$930

$230
(33%)
 
 

$1,200 *

$1,690

$490
(41%)
48"
59"

$2,000

$2,800

$800
(40%)
 
 

$2,550 *

$3,800

$1,250
(49%)
61"
75"

$3,800 *

$6,100

$2,300
(61%)

* includes enhancements such as digital comb filter, picture in picture, surround sound, velocity scan modulation and s-video

Suffice to say for the purposes of this analysis, that unless developers offer microdisplay based televisions with improved value, overall market growth for projection television will be modest and competition will be brutal. To stimulate market growth, the new generation of projection televisions should target some or all the following value points:

  • Value priced sets for under $2,000; high performance sets for < $3,000.
  • Picture quality equivalent to direct view CRT televisions: high brightness, wide viewing angles, and excellent color and video.
  • Wide screen and digital video compatibility

2. PROJECTION TELEVISION TARGETS

A weakness of current CRT rear projection systems is their limited viewing angle and brightness that results from the low light output of the CRT projection engine (150 to 300 lumens). Current receivers must incorporate a high gain rear projection screen in order to achieve even marginally acceptable brightness levels (>300 nits) at normal viewing angles. The high gain screen results in a very limited half brightness angle of less than 20° compared to a direct view CRT of 60°. Ideally, the microdisplay based projection television will achieve both a higher normal brightness (>500 nits) and a broader average half brightness angle of 30 to 40° by employing a screen with a gain of less than 2. The screen lumens required for a range of 16:9 widescreen diagonals is shown in Figure 1, assuming 500 nit normal brightness and 85% screen transmittance.

Figure 1. Screen Lumens for Projection Televisions

Boosting the projection engine output to 500 &endash; 600 lumens allows the use of a gain 4 screen to achieve 500 nit brightness. Such a TV would show a higher brightness level than current CRT RPTV, but would still have a limited viewing cone. To show both a brightness and viewing angle improvement requires a projection engine capable of 700-1,200 lumens. Simply incorporating the current low cost ultraportable projectors (500-1,000 lumens) in a new family of widescreen HDTVs will offer increased brightness over CRT RPTV models, but viewing angles will still be limited, especially when compared to the direct view CRT benchmark. To make rear projection technology more competitive, projection engine output should be pushed to more than 1,000 lumens

The bottom line is that microdisplay based rear projection TVs need to incorporate projection engines that have more lumen output than current ultraportable projectors. Further, prices must be targeted at under $3,000 and ultimately less than $2,000 if rear projection sets are achieve a more significant share of the television market.

3. MODELING RPTV PERFORMANCE AND PRICE

A number of researchers have published papers describing the methodology for modeling projector light throughput6 7 and documenting the performance of current components. Over the past several years, MCG has published a series of studies on microdisplay, lamp, and projector throughputs and costs8 9 10. Cost forecasts are based both on surveys of market participants, as well as bills of material analysis for key components.

In order to focus on the impact of microdisplay technology and projector architecture developments on rear projection TV performance, the throughput and costs for a variety of prospective designs were compared using the same lamp and screen.

All comparisons are based on the 120W UHP high pressure mercury (HPM) lamp, the highest power lamp currently in production with a half life in the range of 10,000 hours. Further, our modeling of lamp costs predicts that the combination of the lamp and power supply could be available at high volumes at a price of approximately $100 in the near term. While extension of the UHP technology to higher operating voltages and output is a high priority for Philips and other developers, to date the 150 W HPM lamps have not demonstrated long lifetime and few are confident that they will in the near term. Likewise, few think that a long life high power, metal halide lamp will become available. As a result, about the only likely lamp for a production TV is a HPM, and about the highest output that can be assumed is 8,000 lumens.

Referring back to Figure 1, 500 screen lumens are required for a gain 2 40-50 inch widescreen set and 1,000 lumens for a 60-70 inch receiver with good performance and > 2,000 lumens for a striking large screen. Given the 8,000 lumen output limitation of a long life HPM lamp, the overall projection system must operate at an efficiency of > 10% to be acceptable, and > 20% to be exceptional.

4. RPTVS USING INCUMBENT TECHNOLOGIES

Current three imager 0.9 inch p-Si based ultraportables powered by 120 W HPM lamps are specified to deliver 500-700 lumens, with the current best of class projector, the In Focus Litepro 755 and others, claiming 1,000 lumens, an efficiency level of about 13%. Output of single chip DMD projectors that use the 120W HPM lamp, are considerably lower with the best of class promising 800 lumens (InFocus LitePro 750), about a 10% overall efficiency. List pricing for such projectors is more than 2 times the levels targeted for televisions.

Using the MCG throughput and cost model, the performance and the 2002 high volume (>5,000 per month) bill of material (BOM) costs for rear projection, 1280x720 definition, televisions are shown in Table 3. The model assumes that the microdisplay component factory price is 2 times cost. It is forecasted that CRT technology will offer the lowest cost solution with the field sequential DMD offering an acceptable level of advantage, and the higher cost 3 channel p-Si designs a better level of performance with costs about 20% greater than the DMD. Street prices for all of the technologies could be under $2,000 at high volume.

Table 3. Comparison of Leading Technologies

Technology

3 x 7 " CRT
1 x 1.0" DMD
3 x 09." p-Si

2002 BOM Cost

$792.48
$963.02
$1,148.30

Screen Lumens

200
479
895

Brightness (Gain 2 60")

109
261
488

Comments for each of the approaches are as follows:

CRT. Despite the pricing of current HDTV receivers (Å$5,000), our survey of big screen manufacturers shows widespread optimism that future high volume costs for HDTV models will be only moderately higher than current NTSC RPTVs. While makers are bullish on costs, they were unwilling to predict key performance features such as true definition, brightness, or lifetime.

DMD. RPTVs are projected to have considerably lower light output then current ultraportable projectors because stiffer requirements for color saturation will dictate the use of a color wheel without a clear (white) segment. Pricing for the DMD imaging module is forecasted to be less than half of current levels principally due to more aggressive and lower margin pricing policies.

p-SI. Continuous improvement of the p-Si imaging technology and further cost reductions result in the highest performance system. Costs, however are also significantly higher than those for the DMD.

5. LCOS PERFORMANCE AND COSTS

Reflective LCOS HDTV imagers are being developed by a range of companies and front projection units with 3 imagers, targeted at the presentation market, are commercially available. In addition several single imager architectures have been announced and prototyped. Table 4 presents a comparison of the LCOS designs.

Table 4. Comparison of Alternative LCOS Architectures

Technology

3 x 1.0"
1 x 0.9"
Field Sequential
1 x 0.9"
Scroll
1 x 0.9"
Hologram

2002 BOM Cost

$1,079.13
$840.41
$982.14
$944.04

Screen Lumens

888
324
586
586

Brightness (Gain 2 60")

484
176
319
319

Comments on the alternative CMOS approaches follow:

3 Imager Systems. The 3 chip reflective architecture is predicted to offer similar performance to p-Si designs at a lower cost due solely to the projected lower cost of the CMOS backplane technology.

Single Imager Systems. Cost for a single chip color field sequential LCOS system approach the lowest cost CRT designs and output more than 1.5 times as many lumens. While costs for the scrolling and hologram architectures are predicted to be about the same as those of the DMD and higher than the field sequential design, output is nearly 600 lumens, significantly higher than the DMD.

LCOS designs promise to be very competitive with the incumbent technologies. 3 Imager solutions should match the performance of 3 imager p-Si technology and cost somewhat less. Single imager architectures offer the lowest cost microdisplay solution with the scrolling and hologram design approaches projected to be very competitive with DMD designs in both performance and price.

 6. OTHER IMAGING TECHNOLOGIES AND ARCHITECTURES

Developers are also pursuing large area light valves (>1.5 inch diagonals) using both electron beam addressed reflective light modulators and transmissive low temperature poly silicon. Large modulators offer only a modest increase in light throughput when used with a single long life 120W HPM lamp. When used with higher voltage MH lamps (>250W) light throughput increases dramatically to more than 2,000 lumens. The Sharp prototype RPTV with three 2.5 inch CGS™ low temperature p-Si imagers demonstrates just how good a high brightness, wide viewing cone projection display can look. Unfortunately, manufacturers of higher voltage (>200W) small arc (<2 mm) MH lamps are not optimistic that lamp lifetime can be extended to > 10,000 hours. The use of multiple lamp architectures has been demonstrated as a potential path to higher throughput systems.

7. CONCLUSIONS

The screen lumens and bill of materials costs for a range of projectors are arrayed in Figure 2. The single imager architectures offer the best compromise of throughput and cost, but non of these architectures will deliver a dramatic improvement in RPTV image quality. Only the more expensive three imager systems approach 1,000 lumens in throughput.

Figure 2. Comparison of Low Cost Projectors

Direct extensions of DMD and HT p-Si microdisplay technology are likely to result in wide screen high definition rear projection televisions that deliver 2 to 4 times as much light to the screen and therefore can be both brighter and offer larger viewing cones than CRT based rear projection sets. Bill of material costs for such systems will be higher than those of CRT receivers, but will be under $1,000. The performance and price of the microdisplay based TVs will close the gap between the image quality of rear projection and direct view displays, but viewing cones will still be noticeably smaller.

The emerging LCOS microdisplays will offer stiff competition for the market leading technologies. Single imager LCOS architectures will offer the lowest cost designs and performance can equal that of single imager DMD system. The availability of these LCOS devices will drive DMD prices to much lower levels. Single imager LCOS displays in small screen sizes (Å 40 inches) could compete with the largest direct view CRT televisions and could dramatically increase the size of the RPTV market if priced at levels < $1,500.

For microdisplay rear projection systems to compete directly with the image quality of direct view CRT, the sets must offer both high brightness and large viewing cones with 1,000 to 1,500 screen lumens required. A breakthrough in small arc lamp technology is required to achieve this result at both low cost and long life time. Extensions of current HPM technology will result in systems with 1,000 lumens, but the leap to >1,500 lumens requires a significant step forward.

 8. REFERENCES

1. M.D Wand, W.N. Thurmes, M.R. Meadows, and R.T. Vohra, "Chronocolor FLC devices for high-resolution projection displays", SPIE Projection Displays IV, vol. 3296, pp. 13, 1998.
2. JVC Press Release, 1999.
3. J.A. Shimizu, "Single-panel reflective LCD projector," SPIE Projection Displays V, vol. 3634, pp. 197, 1999.
4. C. W. McLaughlin and R. Cooke, "The Impact of DTV on Television and Computer Displays," McLaughlin Consulting Group, http://www.mcgweb.com, 1999.
5. C. W. McLaughlin and R. Cooke, "Display Opportunities in the DTV Era," McLaughlin Consulting Group, http://www.mcgweb.com, 1999.
6. E. Stump and M. Brennesholtz, "Projection Displays," 1999.
7. A.E. Rosenbluth and R.N. Singh, "Projection optics for reflective light valves," SPIE Projection Displays V, vol. 3634, pp. 87, 1999.
8. C. W. McLaughlin, D. Armitage, S. Jurichich and L. Ragle, "Cost Modeling of Display Technologies Utilizing CMOS Backplanes," McLaughlin Consulting Group, http://www.mcgweb.com, 1998.
9. C. W. McLaughlin, F. Schuda and D. Armitage, "Projection Lamps: Technologies, Costs and Markets," McLaughlin Consulting Group, http://www.mcgweb.com, 1999.
10. C. W. McLaughlin and F. Schuda, "Cost Modeling of Projection Systems," McLaughlin Consulting Group, http://www.mcgweb.com, 2000