Robofest is a “day-long event for youth exploring the creative world of robotics.” Kate, Jenny, Max and I were asked by Vision Ed Inc board president Marianne Petit to be judges for the DANCING WITH THE ROBOSTARS competition at Sony Wonder Technology Lab in New York City. The five competing teams of children each built dancing Lego robots, and choreographed routines for themselves to dance along with their creations. We witnessed a medieval swordplay performance, Hawaiian surfing dance, video game emulation, disco mash-up and a funk collaboration. The young inventors learned their robotics skills under the tutelage of Laura Allen, through Vision Ed Inc in New York. We loved it so much that we gave everyone a prize.
We now have the basic sensing and dimming functions completed, using the MIDI dimmer and three photosensors over TCP/IP. The effect is excellent when the refresh rate is high. If we decide to go with a less frequent refresh rate, then a smoothing algorithm will need to be added so that the light levels do not jump noticeably but instead adjust appropriately over time. Here is the current sensor and display code.
Here’s the two XPorts mounted on a single breadboard. One communicates with the other over TCP/IP and it finally works. Things I learned that I’d like to remember next time: Instead of sending inverted serial like we send to the computer from the PIC, the XPort is expecting non-inverted serial. The code for this is 84 rather than the 16468 that we usually send to setup the SEROUT2. Also, it’s normal for the voltage regulator to get hot when using these. And it’s always worth setting up a DEBUG or normal SEROUT2 for the computer to monitor what the PIC is doing. Otherwise a lot of time can be wasted debugging what you think the PIC is doing, while it’s actually doing something different. Here’s the initial sensor code and the display code. The next step will be to make display happen over MIDI.
Made two XPort TCP/IP to serial adaptor boards using Tom Igoe’s PCB. Preliminary testing indicates that they both function.
A tool for teaching binary finger counting. You hold up the correct finger representation of a binary number and the chip translates to decimal on the 7 segment display. This was a lot of fun to build, and a lot of work too. Very rewarding to have the idea drive development, rather than the last few projects where technology lead the plan. Learned a lot about BASIC, including that it doesn’t have a power function so you have to write your own. My code includes subroutines for running the seven segment counter display, which could come in useful later.
Resistors in the analog circuit seem useful for changing the behavior of the associated sensor. My glove project might need this, but I’m not sure if it will be important. You can also note that I’ve added various capacitors to smooth out the voltage. Had to find out for myself that I needed them, but I guess I have to come to terms with my own learning process. 
The serial circuit goes to the computer screen via 9600 8-N-1. Memories of my “high-speed” modem setup from 1992 flood back to me.
This looks like it will be really useful for debugging. I wonder if doing a USB interface is any more difficult?
The pot’s sensitivity changes with resistors, and differently depending upon whether the resistor is placed before or after the pot.
This is a prototype of the sensor glove for my binary finger counting project. The flex resistors are sewn directly to the glove. I only was able to obtain three working ones this week, so the little finger and thumb still need to be done. The little finger shouldn’t be a problem, but attaching a flex sensor to the thumb is going to be challenging, due to the thumb’s unique physiology and motion dynamics.
This glove will stretch to fit a variety of medium hand sizes. For people with very large or very small hands, a different glove will probably be needed.
This project was to create a binary number game. The red light blinks between one and fifteen times. Then switches (in this case highly economical bare wire contacts) are closed to match the number in binary. So for 5, the first and third switch would need to be closed. When the correct binary number has been entered, the green light blinks rapidly, then the process begins again.
The code for the binary game will be repurposed next week to build an analog glove that replaces the switches.
The big problem I encountered was initially failing to use voltage stabilizing capacitors. Once I lit up more than a single LED, the chip’s behavior became erratic. Adding the capacitors (pictured) immediately solved the problem. Guess that’s why everyone kept telling me to use ‘em. But some things have to be learned independently.
Mostly this was an exercise in getting comfortable with the PIC Basic code. The code uses variables, conditionals and subroutines to get the job done. This should be very helpful for next week, when I repurpose it to work in analog.
