wuchuk
- we make stuff -
paper circuits
who needs to etch a circuit board when you can cut it?
We had the idea to make a custom circuit board by cutting a thin sheet of copper with an computer-connected paper die cutter. We designed the circuit layout in EAGLE (CadSoft), converted the image of the layout to lines in Inkscape, and cut 3M Copper Foil Shielding Tape using Make the Cut! software that came with the KNK Zing. Then we removed excess copper by hand very carefully (but still ripped the paper a bit and lost small pads) to reveal the circuit. We soldered on our components like with any regulary circuit, tinning the copper first. We also had to be careful not to heat it too much or else it burned through the paper backing.
We made one circuit run off an Arduino Uno with Sparkfun's PWM Shield board and another with the PIC32MX120F032B. We programmed the Arduino with the Arduino IDE and the PIC with MPLab X.
The trickiest part was getting high-quality cuts with the paper cutter that cut all the way through the copper without tearing it. We had to play around with the cut settings of the Zing and the adhesive bond between the copper and paper. The 3M copper tape worked best because it came with an adhesive backing, which we affixed to regular printer paper.
In the future, we'll play with different paper to get easier removal of the sticky copper from the backing.
We had the idea to make a custom circuit board by cutting a thin sheet of copper with an computer-connected paper die cutter. We designed the circuit layout in EAGLE (CadSoft), converted the image of the layout to lines in Inkscape, and cut 3M Copper Foil Shielding Tape using Make the Cut! software that came with the KNK Zing. Then we removed excess copper by hand very carefully (but still ripped the paper a bit and lost small pads) to reveal the circuit. We soldered on our components like with any regulary circuit, tinning the copper first. We also had to be careful not to heat it too much or else it burned through the paper backing.
We made one circuit run off an Arduino Uno with Sparkfun's PWM Shield board and another with the PIC32MX120F032B. We programmed the Arduino with the Arduino IDE and the PIC with MPLab X.
The trickiest part was getting high-quality cuts with the paper cutter that cut all the way through the copper without tearing it. We had to play around with the cut settings of the Zing and the adhesive bond between the copper and paper. The 3M copper tape worked best because it came with an adhesive backing, which we affixed to regular printer paper.
In the future, we'll play with different paper to get easier removal of the sticky copper from the backing.
LED circuit costumes
we complete the circuit!
For Halloween 2012, Nick and I made an LED circuit. Nick was the LED, a hat made out of a plastic cake dome, RGB LED strips from Sparkfun, and a simple switch circuit. Mary was a 1K resistor, making her first dress on a sewing machine! Mary also made a battery by sewing up a cloth shell for a foam roller. All together, we complete the circuit. Our costumes won first place at a friends' Halloween party, earning us the prestigious prize of an inflatable stegosaurus.
Keepon Harlem Shake
dancing robots
Keepons were meant to do this.
We opened up our Keepons, drilled a hole in the base for wires, and soldered to the I2C and power connections specified in the very helpful Beatbots Arduino code: https://github.com/beatbots/mykeepon
The Arduino code is meant for you to command the Keepon through serial, so to make our own dance routines, we wrote some Processing code to send out commands to the Arduino's (3 Arduino's for 3 unique Keepon routines).
Mary designed the all-important helmet in CAD, printed it in ABS, and glued on the visor. She also made the Cookie Monster costume from an old sock.
Nick used his PIC code to run the RGB LED's he used for his Halloween costume earlier. The PIC circuit is made out of copper sheet and paper like our paper circuits project.
The Accelecolorator
Redbull Create 2013 Qualifier Entry
Redbull sent us a kit in the mail with RGB LEDs and the Encabullator microcontroller. We used everything, including the box, to make an interactive wearable device that changes color with limb orientation and movement. We also added 3-axis accelerometers and 12 A/D converters to read in the acceleration data.
Data from the accelerometers (sewn into each cuff) was mapped to an RGB color displayed on the cuff. In "dance" mode, all the cuffs are lit and the colors change with acceleration.
In "game" mode the wearer has to match the lit cuff's color to the color on his/her chest.
Our video shows the Accelecolorator in action at Chicago's famous landmarks.
Pressure reaction LED shoes
Lightup shoes aren't just for kids. We saw many variations on RGB LED shoes out there, but we wanted our shoes to give us information about how we walk. By placing a softpot position sensor on the sole of the shoe, we measured where the foot was making contact with the ground along the long axis. Then we lit up the corresponding LED along the edge of the shoe. We created several color and brightness patterns, but settled on a simple fade.
Again we used paper circuits to create a flexible and light circuit board and customized it to work with the PIC32MX150 chip. We chose the PIC because the LEDs need a fast data transfer rate, so an Arduino would not suffice.
We tried several flexible sensors, but the softpot gave the best data. "Flex" sensors were not bent enough by normal walking, FSR's gave only on-off signals, and accelerometers/gyros gave noisy data that did not easily differentiate between moving and still. The biggest challenge was figuring out the precise timing for the LEDs to show the behaviors we wanted, because the spec sheets weren't always accurate.
After the first prototype (above) ripped apart while Mary walked around the block outside, we decided to remake the shoes to be sturdier and with a smaller circuit. We ended up using LED strips encased in plastic and epoxying (Miller-Stephenson Epoxy 907 is strong and ductile) them around Converse shoes, which have convenient laces to hold the battery.
We didn't have much luck using velcro, which other groups have used, since the sticky back of the velcro wasn't very sticky.
Mary wore these on Halloween 2013, and we're finally calling this project completed after several years. However, it is a battery eater, and we need to make sleep modes for future PIC circuits.
LED Goggles
For Halloween 2013, Nick was inspired by the Adafruit project: http://learn.adafruit.com/kaleidoscope-eyes-neopixel-led-goggles-trinket-gemma/. We made a paper circuit and programmed a PIC to run the different behaviors, which could be selected by tapping on an accelerometer on the side of the goggles. The programming of the two LED arrays was harder then expected, but then again, maybe we're just bad at programming with rollovers.
Nitinol flower
Mary's vision was to make something that moves like petals of a flower - silent and nonlinear. Nitinol seemed like the perfect actuator, but we spent years figuring out how to control it so that it would create a quick, dramatic movement without burning out. After several experiments that ended in small fires, the breakthrough was to build a current-controller circuit - based on op-amps and a MOSFET - for each wire. We also decided to use thin flexinol (0.004" diameter) that exhibits high force and can be left on indefinitely at a certain max current.
Another challenging component of the project was figuring out how to attach nitinol wires to the petals. The wires often melted through tape and even burned through thread at times. We searched for thread material that could tolerate high heat, but they were expensive and had multiple strands. We ended up using conventional sewing thread, which we put in a sewing machine, and sewed a loop path to weave the wire through. This worked quite well, but the loop had to be sewed precisely and the wire had to go through each stitch or else there were weak spots that bent too much when the nitinol was actuated. We received great advice from Jie Qi, formerly of MIT's High-Low Lab, on how to crimp to the end of the wires (you can't solder to them). It works best to double the wire back and forth through the crimp and use regular pliers to crimp the ferrule or crimp bead. Then we soldered the crimp to a piece of copper tape attached to the paper flower petal, and soldered wire leads to the copper tape.
We chose paper for the petals because it has just the right amount of stiffness to bring the petals up and in after the nitinol is turned off. We saw several videos online of cloth flowers actuated by nitinol, but they seemed to all require elastic to close the petals, which we felt ruined the magic of the experience.
To make the flower responsive to user input, we placed a distance sensor (Maxbotics EZ4) in the center so when someone approaches and blocks the beam, the nitinol wires open the petals. As long as the sensor sees an object, the nitinol will have a set current to keep the petals open. When the person moves away, the nitinol turns off and the stiffness of the paper brings the petals back up. This creates a very seamless experience of interacting with a life-like flower since there is no visible or audible indication of an actuator or sensor.
The final prototype uses an Arduino lilypad as the flower center and has six layers of custom round PCB's (each PCB has components for one current-controller circuit to power one nitinol wire) attached beneath it. The boards are all held together with header pins that create a mechanical as well as electrical connection. The sensor, microcontroller, and PCB's are all concealed inside the flower body.
Current-controller circuit