OLPC / Freeplay
OLPC is a MIT project – One Laptop Per Child. The objective is to create a $100 laptop for children in the developing world, with pretty interesting specifications. These laptops will only be available for mass purchases, either by NGOs or governments.
Freeplay is a company (and foundation) that produces devices based on wind-up and solar energy, such as radios and flashlights. Their goal is that populations in developing countries have access to information (through radio broadcasts) without the need of having a energy infrastructure in place (be it regular electricity or batteries).
This week it was announced that the OLPC would see a price hike, and that it wouldn’t be able to reach the $100 price target until at least 6 million units are sold, and that it would start to be sold around the $130-140 price range.
One of the interesting business/social practices of Freeplay is that they help fund their social works in the developing world through the sale of their products in the developed world.
I think the specifications of the OLPC are very interesting also for children in the developed world – think of it as a Fisher Price “My first laptop”, for children 5 and up, simple, sturdy, pre-packaged with reading and math software for first graders. If it was sold for approximately $199/€159 in the US and in Europe, and the difference between the cost and the retail price applied towards financing the OLPCs to the developing world, it would help both increase the amount of OLPCs produced and lower both the overall cost and purchase price for the developing world.
Feeling the heat
Laptop computers today pack more power than supercomputers of 30 years ago, with batteries that last at least three or more hours, in packages weighing less than 3kg. All of this is wonderful, except for one problem: the extreme temperatures reached on the surfaces in contact with your body (The new MacBooks / Pro from Apple don’t have any mention of being “laptops” anywhere). The temperature problem can be divided into three areas:
- Main processor / Video processor / Chipset (including RAM)
- Non-volatile storage (Hard disk / Flash memory)
- Battery / voltage converters
Modern CPUs boast a better ratio between wattage and computing power than CPUs from the past, but the increase in processing power has assured that the amount of electricity used has increased (and subsequently the amount of heat).
Hard disks have more capacity, and are faster every passing year, consequently producing more heat. Flash storage devices, which produce less heat, are still behind in terms of capacity, cost per MB, and read/write cycles compared to hard disks.
Finally, the batteries to power everything must have much higher capacities, and the increased chemical reactions generate even more heat, as the conversion and regulation, which also creates even more excess waste heat (although of the three it’s perhaps the part that is the least of our problems).
The possible solution (besides either saying that laptop computers are not “laptops”, or reducing their power, which are both not the solutions people are looking for) is a redesign in laptop architecture. The main area which is never in contact with the user’s body is the screen part. It would be ideal if most waste heat could be routed some way to the back of the screen. I can imagine three solutions, each with their own drawbacks:
- Move to the back of the screen the main and video processors, along with most of the chipset, and perhaps also the main fixed storage. The main drawback is in terms of balance – the screen part of the laptop computer should be as light as possible as compared to the “main” body. With the miniaturization of motherboards and CPUs, and with the possible switch to flash memory fixed storage this should be feasible. More space could be used in the main body for batteries, perhaps, for longer autonomy and to offset the weight.
- Use of a liquid cooling system, as available for desktop computers. Easier to implement than solution #1, but not very elegant on a laptop. There would be problems with routing the liquid tubing from the main body to the screen part, and it would also increase the overall weight of the system.
- Using peltier junctions to transfer heat from the end of the main body chassis to the botton of the screen lid. Mechanically difficult to implement, and it would also increase the power usage of the whole system a lot. Probably the worst of these solutions, even if theoretically more elegant in terms of engineering.
These are just a few ideas, I would like to know your opinion on this problem.
