(This post makes no reference to the OLPC project. Please see OLPC News for such commentary -LF)
In the summer of 2000, I was presented with a challenge to design a system that would serve to bring telecommunications and computer functions to remote villages without power or telephones. Lee Thorn, chairman of the Jhai Foundation, sought me out during a 30-year reunion of the place where we had met before. The villagers in Laos with whom he was working had requested telecommunications as thier priority need.
Details of the resulting design can be found at the Jhai Foundation website. I am no longer affiliated except as an advisor to the Jhai Foundation - In 2003 I turned over the design work to a crew of volunteers and focused my attention on issues of power generation. The Jhai PC, as it became known, has gone through some changes as might be expected with an emergent design.
I would like to exercise my prerogative as the originating designer to specify what characteristics make this design qualify for the designation.
Most salient, the Jhai PC is intended for use in rural or semi-rural areas and must therefore be well protected from the kinds of harsh environments and minimal protections available in rural locations. Ideally, the device should withstand ambient tempereatures of 40 Celsius in still air. It should be sealed against dust and moisture ( to 100 percent relative humidity) as well as fungus (a significant problem for electronics in tropical climates).
Keyboards and displays must of course be accessible for use and replaceable, though not easily detached. I recommend the use of standard connectors located inside the protective housing of the Jhai PC, which itself should have a key lock and an alarm to detect unauthorized opening. Once opened, the device should have its components readily available for service, without interrupting its operation. My original mechanical design mounted its stack of PC-104 cards on a frame which pivoted up from the baseplate in the box after the box was opened. Cables were designed to run down the pivot arms to allow operation when the board stack was swung up for service. Other components were mounted along the inside of the box.
For the original installation a telephone was provided wired into the box and the VOIP interface inside. The number dialed could be a local telephone number, in which case the call was routed through the local phone syustem where such a connection was made at the terminating end of the system (a town with phone service). If the number was an IP number (the familiar four groups of three-digit numbers punctuated by periods) the call would be routed through an external ISP to the Internet.
It may be necessary to eliminate the telephone set and use an audio headset, due to telecommunications regulations in some countries. A dial pad could probably be provided under these regulations.
Reliability requires that no moving parts be used in the components of the Jhai PC. I set the service lifetime of the original Jhai PC at ten years minimum. Hard disc drives cannot be expected to last more than a few years in such environments, at the best. Solid-state memories of the "flash" type such as "Compact Flash" cards familiar to digital camera users would be used instead of hard discs. The maximum size practical at this time for such memories is 2 GByte. These cards place significant requirements on the operating system due to the fact that they have a limited number of rewrite actions while retaining reliability.
The operating system (almost certainly a widely-supported open system such as GNU-Linux) should therefore be designed to run resident in RAM (random access memory, probably 128 MByte in size). Upon removal of power, the relevant files should be rewritten to the flash memory (files are "relevant" if they have been changed and must be saved to allow the system to continue - temporary files are not relevant by this definition).
But how, one may ask, can any files be written if power is removed? The answer is that there must be a power reservoir to allow this function, and a means of detecting when external power is removed. Circuits to perform this detection and notification function are as old as computers, but the technology for the reservoir has only recently been perfected. "Double-layer electric" capacitors, sometimes called "supercapacitors", now offer capacitance values in the region of 1 - 10 Farads at low voltages. One Farad will store 1 Joule at 1 volt, though as the voltage falls during discharge the usability of the energy decreases. Such a reservoir and power regulators which make use of its characteristics are a necessary part of the Jhai PC.
The Jhai PC must also operate unattended and require no specialized operator. It must work as a treetop relay station as well as in the village. For this purpose the Jhai PC needs its signature element - a supervisory processor module. This will be a specially-programmed microcontroller - the kind of microprocessor used widely in control applications with its own internal memory. Several types currently available will suffice to perform its tasks.
These tasks are:
1. Control of shutdown operations during power removal.
2. Control of power to external peripheral devices under control of the main processor. Saving power will be important where that power is precious.
3. Compilation of reports on the status of battery charge, battery health (profiles of recent charge-discharge cycles), security breaches (as indicated by the tamper alarm system) and other indicators of system health.
4. Diagnositc tests of the main computer and re-booting when restarted.
5. Diagnostic "console" for maintenance. This console would be available both locally (by attachment of an external terminal) or remote, over the network.
The primary function of the supervisory processor will be to keep the system running with a minimum of skilled attention on site. Many Jhai PCs can be monitored from remote locations and it will be possible to construct a staged support system that serves to keep the devices maintained efficiently. Without such a support system any Jhai PC will soon enough stop working and will lose the confidence of the people it is intended to serve.
The cost of operation of the Jhai PC must be quite low, and the need for replacement elements must be kept minimal. "Consumables" such as paper and ink for printers must be accounted for in this cost of operation. It was for this reason that dot matrix printers were specified for the original Jhai PC installation. Such printers, which are quite rugged, require only paper and an ink ribbon for operation, and there are various ways of re-inking the ribbon to minimize cost.
Ink-jet printers with expensive cartridges are much less desirable, even though they will print graphics and elaborate character faces. While ink cartridges can be refilled, the technology is rather more complex than that used for re-inking cloth ribbons, and the number of re-use cycles is smaller. Neither solution is ideal, and more work will have to be done to optimize the printers for this type of system.
Most importantly, the Jhai PCmust be well integrated into the social fabric of the community where it is in use. All persons in the village must be able to recognize that the continued operation of the Jhai PC is essential to the economic progress of the whole village. That is, indeed, the purpose of providing telecommunications and the other functions which the Jahi PC performs. This means that the system cannot simply be delivered and installed, but that a social and educational process must be followed to introduce the villagers to the system and to allow them to "take ownership" of it. I understand that he Jhai Foundation is currently working to develop this process and to develop the layers of training (viz., training the trainers, etc.) necessary to implement this process.
Earlier I mentioned an alarm system to warn of tampering. No alarm system can protect a valuble resource from being stolen or damaged - that is a social issue for the village residents. The alarm can only alert them that response is necessary.
It is my hope that these basic design criteria will be of help to others in implementing Jhai PCs. The Jhai Foundation owns only the name and has made the design open. I have attempted here to set down the notes of the original designer in that spirit of openness, confident that there will be more things to design that will move forward the effort to bring modern communications to all.