These are kind of inks that change from a colored state to a whitish/transparent state when heated. We had a sample of the liquid crystal ink and the leuco dyes. Of both we made a little 2-pixel display. Both pixels could be controlled separately (yes, with an arduino).
It turned out that the crystal ink changes it’s color quite rapidly (compared with the leuco dye) and with a minimal change in temperature. It only needs about 0.80mA.
The leuco dye on the other hand changes quite slowly and needs more power, about 400mA to get a noticeable change in reasonable time and 200mA to maintain the temperature. However, the look of the liquid crystal is very glossy. And not completely in line with Aram’s ideas.
A third option we had is a color changing T-shirt from American apparel. This T-shirt is one big leuco dye surface and quite cheap. Aram has experimented with sewing in conductive threads as heating wire. The T-shirt showed surprisingly good color changing properties and relatively sharp edges. Still around 200mA is needed to produce enough heat in the wire for a good and quick change in color.
One of the major challenges that we need to overcome is the constructional complexity scaling up significantly with the amount of segments.
- In the period before Aram’s arrival we have experimented with photochromic ink as well. These inks change from transparent to color when exposed to UV light. The effect is good and sharp pixels can be drawn using LEDs with a narrow light angle but some major drawbacks make the approach less attractive.
- Making a matrix display with UV LEDs would be quite simple however the UV LEDs are expensive, 1 LED would cost around 5 euros. Just a 10×10 pixel display would already add up to 500 euro solely for UV LEDs.
- Safety measures have to be taken into account. The UV light comming from the LEDs is very concentrated and harmfull to the eyes. A UV absorber would have to be placed in front of the display. We tried a color protection spray which has a UV block but the factor is too low to block any significant amount. Blocking foil used for windows etc would work but gives a plastic look.
Another test incorporating the twitter communication has been hacking a name badge we still had lying around from eArts. We hacked up the badge so it could be controled using again an arduino. A python script fetches the most recent tweets of one or more persons and sends them to the arduino. The arduino emulates the user presses and ‘writes’ the tweets to the badge.
How We Will Proceed From Here
- The T shirt is very attractive to experiment with, but the full Twitter concept is too complicated to implement with a dot matrix. since experiments with the heating wire are promising, we keep a simple version of the concept by displaying several pre-stitched texts. No real-time connection to the virtual environment, but keeping the ‘twitter-like’ message on a shirt. The unusual messages will still provoke reaction, thus relating to the original concept. Users can switch between texts.
- Using a Label printer. This is a low-tech version of the concept and Aram’s favorite. Twitter feeds are sent to a label printer. Visitors at events can go to a labelprinter and print their latest tweet. The first event would be PICNIC where users probably have an RFID in their badge. This badge can be scanned and the tweet printed.
- Display a speech bubble with latest tweet on a mobile phone or PDA. Not necessarily iPhone, a mobile java application will probably do.
- Badge hack we will extend with a GSM connection and have it as a DIY project