Technical details
“That's Alcohol” installation is a prototype for innovative ethylometers to be proposed to italian police in order to prevent drunk-driving. The system is based on a TGS-822 Alcohol Sensor (made by Figaro) that detects the alcoholometric level into the user breathe. The data are collected in real-time by an Arduino board that drives analogically a 24.7 Kg-force servo-motor (Hitec HS 805 BB+) depending on the drunkness-level result achieved. Finally, the servo-motor is connected to a plastic table that pushes through some pistons the string-puppets to let them leaning more or less.
For a stronger feedback, Arduino board dialogues through serproxy with a Flash-based animation that displays step-by-step the results and some suggestions on how to reach a better drunkness-level.
More info
Made by David Boardman, Roberto Pansolli, Pete Knocke (Domus Academy Milan + Interaction Design Institute Ivrea)
It's this one the same sensor used by italian police? As I am italian I could install one in my car where I already have an arduino to be aware of the consequences of my night parties... F
@fubby: unfortunately that script is somewhere lost on an old computer - the project is dated 2007. However, I recall that the code was nothing really complicated.
For the project we used a Figaro sensor that was extremely sensible and accurate. For That's Alcohol project we didn't need to have a precise calibration - the interaction design of the project didn't require it. Anyhow, you can find on the Web different explanations regarding how to calibrate an alcohol sensor. Basically, you just need a glass, water and pure alcohol and some patience. Have fun!
yes the script was really simple. However, the pins on the 822 are confusing. I found tutorials for the similar MQ-3 but I have difficulties translating those to the 822.
The sensor runs optimally at a certain temperature. In fact, it needs a Vh (heating) to work properly. I remember it was heating up quite a lot in the installation. However, the sensor stayed for 3 days in an warm environment with a poor air flow - and nothing really happened, besides some unexpected values that sometimes popped out (abnormal high or low values). Here some programming effort would help you to normalize the received data.