Tonight, at the Bayview Bash, Carrie tested out the GyroBelt to live Salsa music. As you may remember from previous posts, the GyroBelt is a prototype device for what will eventually become an eFashion skirt. The prototype and the skirt both try to avoid the Christmas tree effect by only illuminating the LEDs under certain conditions. Namely a high speed dance spin as detected by a MEMs gyroscope.
Pictured above is the driver board of the GyroBelt. In the lower right you can see the red SparkFun Gyro breakout board. Immediatly above that are the four transistors that lead out to the bus for the LEDs. (In addition to mono chromatic LEDs the driver board also supports RGB leds.) To the left of the Gyro you can see an ATmega168 running an Arduino sketch. Above that are a couple trim pots that I use to adjust variables within the sketch.
The board has a pretty goofy power arrangement. In the upper left you see a chip and some support circuitry that actually steps a single lithium ion cell up to 5 volts. I wanted the MCU running at 5V for the FET transistors’ gates. However, the gyro needs 3.3V, so there is also a linear regulator that steps the Li-ion cell down to 3.3V. The power going to the LEDs is the raw 3.7Vs coming from the cell. So that’s three total power rails for this relatively simple circuit. All the awkwardness stems from the fact that I built the LED belt first and sized all the resistors for 3.7 V. I should have used two or three cells, run the MCU at 3.3V and used an intermediate stage of NPN transistors to get a good FET gate voltage. Then I’d just have one step down power supply.
The LED belt itself is just kapton film with resistors and connectors soldered on to etched power bus lines. Its the last circuit of any complexity I’ll make with kapton. Kapton is just too fiddly. I thought it would be cool to have flexible circuit boards, but its really more of a pain in the butt during assembly than anything else. When I make the skirt I’ll run normal wires and have some sort of connection points for the LEDs.
Anyway, my next step is to try to get RGB color shifting to function. The sketch needs a little more work for that to happen. Once I have the sketch working pretty well on the GyroBelt and a few dancers have provided feedback, I’ll attempt to assemble an actual skirt. I’m ditching the tassles, the GyroBelt has shown me that they wind up being too stiff and just get tangled all the time. So right now I think the final skirt will have the LEDs mounted to the surface of the skirt.
Stay tuned to the GyroBelt tag for more updates!