I’ve been designing clothing for 10 years and for about 9 of them I have been obsessed with putting the creation of one’s garment into the wearer’s hands. This project is a demonstration of that interest as a dynamic textile. With fiber optic fabric and wearable LCD screens on the verge of breaking into the market, I decided to take a crack at a completely soft display screen. This also was my first project where I fully explored the challenge of designing a soft circuit.
My concept was to come up with a textile that was able to receive an image from an html form sent over the internet. Once received, the image would seemingly bloom into existence over the surface of the textile.
It was important to me that the output was not a luminous one, I wanted the change to happen within the nature of the material, or rather the fibers themselves. This led me to use thermochromic ink, which is an ink that turns from color to colorless with applied heat. Making a display screen, I needed to figure out what resolution to make it. I would need to treat each pixel as it’s own entity, I would need to handcraft each pixel by putting multiple layers of electrical components together using the softest materials that could be found. Once it was figured out how to make one, I would need to replicate the process how ever many times my resolution called for.
With this in mind, I decided a small grid of 6 x 6 would be a good starting point.
Knowing what kind of output I wanted to achieve, I could gather my materials and appropriate hardware.
Here is a comprehensive list of what I used in the end.
36 NPN 2N2222 transistors – found at any radioshack
Hardware:
Arduino Diecimila
WIZnet module
ethernet shield
Power:
3.7 v lithium ion battery
Software:
the software was a simple html form of check boxes and php script to store and recall data from a MySQL database. I then wrote an arduino program that pulled the data from the server and switched on the pixels i had created in correspondence to the checkboxes that were submitted.
Not only was this project meant to be an exploration of a dynamic electronic textile, but was also meant to create a bridge between two people that were remote of each other, but wanted to collaborate creatively.
There were several material tests that I conducted before I knew which will ones would work best.
The first material I needed to figure out was what I was going to use for a heating element.
Steel wool is conductive and also highly resistive, enough so, that I have used it before in a project using thermochromic inks and once given power heated up very nicely. However, it is also extremely flammable because of this nature, so although I knew it to work, it was still a material that would need to be handled thoughtfully.
I tested it in two ways, I used it raw by needle felting it into the sheep’s wool into squares and second, I mixed it with the wool and cut that into squares.
The raw steel wool got too hot, on the verge of catching fire, and the mixed squares were too slow. Also the pixels themselves were not insulated enough causing the heat to dissipate through out the grid so no discernible pixel was evident.
What was then suggested to me was nichrome, the heating element within toasters and hair dryers. This sounded extremely promising, so I ordered some from United Nuclear and while I was waiting for arrival took the opportunity to crack open some light bulbs in order to snatch the filament for some preliminary tests. I would like to add that you can find spools of nichrome elsewhere that would be more economical if you are looking for more than 10 feet.
Cracking open the lightbulbs was easier than I thought, but still challenging.
What had occurred to me through this process was that I would need a piece of low resistant metal to sandwich the nichrome in-between. I went to metalliferous and searched for the thinnest, smallest piece of metal I could find. I found 18 ga. .5″ squares that I immediately vetoed because of the thickness. I went for the 24 ga. nickel/silver and copper circles with a .375″ diameter. I also chose to use thermal adhesive to fasten the very thin piece of nichrome and help dissipate the heat. There are two kinds of adhesive that Arctic Silver produce, one of the same name primarily made of silver and Arctic Alumina made of aluminum oxide and boron nitride (ceramics) . I chose the latter because of it’s insulative qualities and it is neither electrically conductive nor capacitive.
Feeling really confident of this combination of elements, I set out to make a pixel prototype. The one that I would build the other 35 after.
What’s pictured above is a pixel made with the copper disks, it worked really well, but was really heavy. I eventually tried copper foils, which was too fragile and finally tried copper based fabric I purchased from lessemf.com. After all the trials , the fabric turned to be superior through feel and durability.
After building my one pixel and being satisfied with the outcome, I went on to create another prototype focusing on the circuit build. I chose to make a 3×3 grid. Along with the circuit, I also wanted to try a thermochromic pigment, the pigment differs from the ink in that it dyes the fibers themselves, so it doesn’t change the texture and looks a lot slicker than a coating (the ink) laid on top of the textile.
Without a the proper machinery to put the fibers in proper dye bath, I was left to mixing the pigment with warm water and painting it on the surface. This was more of a “just cause I want it to be so!” moment. Hopefully in the future I can have the time or resources to really dye some fiber so I can felt my own hypercolor items! For this experiment though, it was fun to pretend
I designed my soft circuit in three layers.
Display Layer:
This consists of my handcrafted pixels sandwiched in-between two layers of felt. I screened the thermochromic ink on the top felt layer, the bottom layer I poked the two ends of the nichrome through to the other side. One end of the nichrome is hooked up to power, the other is connected to the collector pin of my transistor. While my circuit was still on my breadboard I used a TIP 120 NPN transistor. For my soft circuit, I opted for the the smaller 2N2222 transistors I found at radioshack.
Transistor and Ground Layer:
On this layer I poked the collector pin through the felt and bent it preparing it for the connection to the waiting end of my nichrome. On the other side of the textile, I curled the emitter pin and sewed each one to the conductive fabric traces I had laid out that were to go to ground. With the remaining base pin, I bent this up too and poked it through the third layer where I curled that as well and connected it to a 1K resistor.
Output Pin Layer:
This layer was permanently connected to the Transistor Layer by the base pin coming through and being sewn to the resistors. The other end of the resistor is then connected to soft circuit traces that I made out of conductive fabric. Each trace will lead to an output pin on the microcontroller.
I teamed up with Sonaar and Ozge for the project. We were racking our brains for items that we could hack into a pong controller. I then received a most excited email from the both of them stating that a huge wooden spool had been found, and I would love it. Already having several mischievously warm encounters with giant wooden spools, I was looking forward to this particular meeting
Other than having a week to come up with and execute the project, the above choices were the constraits.
I chose: shaking/dancing + monkey + sound/color
I’ll tell you that a monkey devouring your head while possessing you to control the tempo of music by furiously banging your head IS metal.
The monkey was formed by my two hands as a felt hat. In the back, is a pouch I constructed to store the hardware, an arduino, breadboard, and an accelerometer.
I have not worked with an XBee yet, so there was a USB cord supplying power and sending data to my computer through arduino to processing to produce a sketch. The sketch being one song, one phrase and one shade of red that corresponded with each of the 3 phases of brutal metalness you are able to achieve.
As you shake your head, the sensor takes readings on the z-axis, which translates as an up to down, chin to chest shaking of the head aka *head-banging*
level 1 of metalness (slow head-bobbing)
song: U2 – with or without you
color: light red
text: ” easy listenin’ “
level 2
song: Poison – nothin’ but a good time
color: medium red
text: ” more metal than puppies, but not by much “
level 3
song: dethklok – themesong
color: blood red
text: ” a fuckin’ monkey is eating your head, that’s metal! “
In processing I used the sonia library, and found it quite easy, though I was just loading and calling songs. The most challenging part of the project was mapping the values of the accelerometer to play the songs in a linear fashion. The values I received started at a central point of the full range of values you could get, it then either went up or down from there. I first applied some capacitive coupling to my circuit to smooth out my readings, which were pretty jumpy. which was done by bridging my power and ground with a capacitor.
Second, I wrote my code so it took the last byte read and the first read, took the difference from those, then using a relational operator put it in one of the three phases of metalness. Included is a short video that shows how jumpy and difficult it was reading the actual head-banging movement. hopefully, a higher quality video is to come.
