An XO Oscilloscope to Measure Children's OLPC Learning


When designing the XO computer, One Laptop Per Child designed the sound card to be able to measure DC voltages. Now why would the talented team behind the "$100 laptop" go to the trouble of hacking an audio jack microphone into a data port? Maybe they had the same revelation as Martin Visser did:

olpc tam tam
OLPC XO's TamTam in action
A colleague of mine who has quite a few OLPCs, mainly for testing the wireless stuff out, told of the absolutely clever analog input port the OLPC has. Not content with a regular AC microphone input, it can be configured in two other modes.

One is a straight DC input that can measure between 0 and 3VDC. I immediately thought of kids in Africa being able to hack electronics together from old radios and the like and using the OLPC as a simple oscilloscope or voltmeter.

The other analog input puts 2.5V and allows you to measure across this. This means a simple potentiometer can be read. Great for all sorts of science experiments, but also a good way of providing another interface to control the Tamtam musical instruments.

Taking this idea a step further in his comment on Building One Laptop Per Socialized Child, Nick explains how children could use cheap sensors to learn about electricity:
One of the many possible IPL activities would focus on electricity. The student would be guided to build his own battery (using fruits for examples). Then, by questioning, he would be guided to understand how to connect more of them to obtain more voltage or current. The student would both simulate it with e-toys, and measure it physically with the sound card.
To capitalize on these learning opportunities, One Laptop Per Child developers like Jaya Kumar, are creating dedicated voltmeter to oscilloscope activities, transforming the Children's Machine into a full science learning laboratory. Just check out Walter Bender's update on their progress:
Arjun Sarwal reports that the Measure Activity now features a frequency-domain representation in addition to a time-domain representation. Journal integration is complete. He also built a $1 temperature-sensing peripheral and a $1.50 intrusion alarm system; both have been tested using the measure activity.

"The great thing about the XOs is that they are inherently networked, so by simply connecting a sensor to each XO, and using a combination of such sensors and the cameras, a highly powerful, flexible and robust sensor network for surveillance can be built."

But don't just take Arjun's or my word for how cool an oscilloscoped XO laptop could be. Check out the measure activity movie:
See, with their simple sound card hack, the OLPC developers will allow children to fulfill Walter Bender dreams of children using XO technology to not just play Dance Dance Revolution but to build their own Dance Dance Revolution machines, learning learning in the process.

Related Entries


I clicked on the video link and got: "The url contained a malformed video id." But clicking on the video itself worked find.

This is truly a great tool, with lots of potential. Being able to directly measure voltages and related variables (temperatures for example), provides the ability to interface the XO with the physical world. It may sound trivial, but no current laptop can actually do that without external cards, special hardware is required.

I can see many other uses for the input: by connecting to some switches, it can be used as a controlled stopwatch. As Galileo used to say, the most difficult thing to measure in a Mechanics experiment is time. By accurately measuring times, kids can learn the laws of Mechanics using the XO, friction, motion, acceleration, the pendulum, gravity, etc. If you add the ability to measure temperature, you can do thermodynamic experiments too, and maybe a little chemistry.

Properly designed activity would pretty much be able to cover a all lot of the curriculum for Mechanics and Electromagnetism, giving kids a chance to practice and learn how the scientific method works.

Does anyone know what the bandwidth limitations of the audio card might be? I'm just wondering whether there are the makings of a radio receiver in the XO?

I have been working on a $1 weather station.

The hardware uses a few simple components and attaches to the microphone inputs. One channel is used for Wind speed and Wind direction (pulsed input) and the other for measuring temperature (thermistor).

Allen, the audio limitations are quite low frequency. Better off with a crystal set.
Good lateral thinking though.

I and Etoys team tested another sensor device called World-Stethoscope with XO at OLPC office on May.

World-Stethoscope converts from the DC voltages (0-5V) to the sounds (0-5KHz). Etoys handles these sounds from microphone inputs as numeric values by the built-in FFT. So kids can control the sketch that is drawn by them with voltage, temperature, luminance and etc.
But the World-Stethoscope device is a little expensive ($40 per unit). So I'm interested in the XO sound card input. I think we should handle it via Etoys.

"Does anyone know what the bandwidth limitations of the audio card might be? I'm just wondering whether there are the makings of a radio receiver in the XO?"

Hi Allen . . .

Good luck trying to find out! I have attempted for a full week now to find out the specifics on the power input connector for the XO. Yes . . . a little $2 item that was probably purchased at a local Radio Shack store!

I have received polite boilerplate correspondence from both Potenco and the OLPC development team (Quanta never replied) but no actual concrete answers. It's rather odd that detailed information about almost every other aspect of the machine is freely available but that they (in Potenco's case) won't tell which connector it is or (in OPLC's case) presumably don't really know (their web posted XO specifications page says that they are using a 2 pin connector while all photos clearly show a single pin barrel connector of some sort.)

If I can't extract this seemingly innocuous informational tidbit from these engineers of such impressive pedigree, then heaven help the poor kid in Nigeria who wants to hook up a car battery to the XO when no other power source is available. :-/

The correct schematic representation of a "single-pin barrel connector" is a two-pin symbol - the shell of the barrel is the other pin.

Sometimes oddball (or obscure, custom or single-sourced) connectors are a tactic for locking out competition or third-party sourcing. More likely the connector decision for the production version is being pushed out and left to the procurement people to decide.

That's a pity because it's an important decision - probably more failures will occur at that connector than anywhere else. However, "hot, clock-stopping technology" does not apparently include lowly connectors, crucial though they may be to the success of the product.

Hi Lee . . .

Thanks for the clarification in "two pin" as also being a barrel with the external shell charged with opposite polarity in relation to the internal one. I suppose all "single pin" connectors are double pin when that is taken into account.

I just find it comical that a clear and simple answer from the parties I was in communication with was not forthcoming. An attempt to lock out "third party gate crashers" will only work until there are millions of machines out there (they could have chosen something other than standard USB ports for data I/O if they REALLY wanted to throw a monkey wrench into the plans of uninvited developers!) If they make their power supply connector too obscure then they will be doing a certain amount of harm to their third world target audience who might not be able to locate suitable replacements in their own locales. Yes . . . more is the pity.