Display Engineering Leads the Way for $100 Laptop Design
by Mary Lou Jepsen
Early in 2005, I helped found One Laptop per Child (OLPC) with Nicholas Negroponte, founder and chairman emeritus of the MIT Media Laboratory and became the first OLPC employee. The effort emerged as a way to capture the endless momentum of Moore's Law – which holds that the number of transistors on an integrated circuit for minimum component cost doubles every 24 months – and create a laptop for those far on the other side of the digital divide; namely, the poor children of the world and their families. In fact, the vast majority of the world lives without so many of the things we consider essential, not least of which is access to information. In 2007, we intend to launch 5–10 million laptops in Brazil, Argentina, Libya, Nigeria, and Thailand. The children themselves will own these laptops, which will be given to them by the respective national Ministries of Education – the computers will last for 5 years and are cheaper than the cost of 5 years worth of textbooks in the average developing country.
Why would a display engineer such as myself devise the architecture of a laptop? Because the display is the most expensive and power-hungry component of a conventional laptop, at an average cost of more than $100 per display – it is a true barrier to the $100 laptop. Because it consumes 7 W of power on average, it makes human-powered battery recharging infeasible. One-half of the world's children live with little or no access to electricity, making the issue of power consumption critical to the development of the $100 laptop.
In fact, the display is the current bottleneck in portable consumer electronics: it is expensive, low in resolution, hard to read, fragile, and power hungry. I believe that is why the hardware-design decisions for consumer-electronics products need to be made by display engineers.
I ended up designing the laptop from the display backwards. We added self-refresh to the display-timing controller chip, giving the software the option to turn the CPU off the vast majority of time. In typical use scenarios, 80% of the time the laptop is on, the CPU's primary functions are to refresh the display and keep the WiFi on; the machine itself is idling. The WiFi chip in our machine has a tiny ARM core in it, so it functions without the CPU. We can turn off – or wake up – the CPU in about 10 msec. With RedHat, we re-wrote the low-level OS and high-level application to take advantage of rapid CPU sleep and wake. The CPU at full bore runs at 4 W – this time-averages to less than 1 W in our typical-use scenario; by comparison, my laptop runs at 25–45 W; the CPU alone accounts for ~15 W.
The display itself is 7.5 in. on the diagonal with 1200 x 900 pixels, which is 200 pixels per inch (ppi) – higher resolution than more than 95% of the laptops that ship today. The display offers a sunlight-readable mode in black and white with e-paper-like quality. When the backlight is turned on, the display becomes color with comparable resolution. This is because of the pixel layout and composition: Each pixel has luminance and limited chrominance information. By having a color arranged in diagonal stripes, we are able to achieve approximately XGA resolution in color. The power consumption of the display is about 100 mW with the backlight off; this is about 2% of the average laptop display power consumption. With the backlight on, the power consumption rises to about 1 W, or 14% of the average power consumption of a laptop display. All this costs one-third of an average laptop display. In addition, in a very rare achievement for a new display technology, these displays leverage the exact mature manufacturing infrastructure of LCD fabs, so that it is fully manufacturable at multi-million units per month by early 2007.
The Chief Strategy Officer at Advanced Micro Devices, Billy Edwards, describes the design of our $100 laptop as the first fundamental revisit of personal-computer architecture since IBM launched the PC in 1981. We have not created a cost-reduced version of today's laptop; we have created an entirely new approach to the idea of a laptop. Here are some things we have in our laptop that you would want in yours:
• Instant on.
• Flash memory instead of a moving hard disk.
• Display self-refresh (while CPU is asleep).
• CPU fast-sleep and fast wake-up (~10 msec).
• Massive mesh networking via WiFi.
• Three-to-four times the range of typi-cal laptop WiFi antennae (up to about 1 km).
• At 2 W, one-tenth the power consump-tion of a typical laptop.
• Human-power input for battery recharge.
• Tolerance of multiple power-chargingsources such as car batteries.
• E-book mode, in a form factor a child can take to bed and curl up with a good read.
Of course, many of our decisions had to take into account the rugged environments in which the machines will be used – they can be dropped, and they seal when closed to resist water and dust incursion. We are working on making the design increasingly eco-friendly. We have also been developing a new battery chemistry that extends NiMH battery lifetime from today's typical 500–2000 charge/ recharge cycles.
Despite disbelief last year that the $100 laptop was even possible, individuals and corporations would attend our meetings. Both we, and they, learned at every meeting. We persevered, and we have gotten the world to help. The children of the world are going to go online with our machines. They are our future, our most valuable resource. This is real; it's happening now. By all means, join us. •
Mary Lou Jepsen is Chief Technology Officer at One Laptop per Child, P.O. Box 425087, Cambridge, MA 02142; e-mail: mlj@laptop. org.