In the beginning...

In our first iteration of the project we mapped TJ's motor and sensor connections and removed his main circuit board. It was replaced entirely with an Adafruit DC motor controller, MCP23017 GPIO Expander, and an MCP2221A to link them over USB to our computer using I2C. Power was supplied by an external DC power supply.

We did this because TJ's CPU lives under a black blob on the PCB. We had little hope of interacting with it or using it in any way, so we elected to remove the whole board to better explore its workings and connections to TJ's motors and positional sensors.

Since then, we have learned a lot.

Our new system is much simpler. While the CPU may remain useless to us, the original circuit board also carries useful stuff like motor controllers, power supplies, push buttons, a power switch and more. We thought, why not use the stuff we already have? It just requires a bit of hot-wiring. 

In our second iteration of the project, we are leaving the external motor controllers and I2C stuff behind. We have another TJ, and this time we're leaving his circuit board intact. We'll be bypassing his original CPU to connect directly to the existing onboard motor controllers, thereby saving greatly on both cost and complexity. 

Now, we make better use of the existing parts.

In this simplified arrangement, we have spliced his existing motor controllers into GPIO pins on his new replacement CPU, a Raspberry Pi 4. We are using TJ's existing D battery bank as the power supply for his existing motors and controllers and the pi. A python script running locally on the Pi controls its GPIO. We've chosen a Pi 4 because we'll be doing speech recognition, but for simple motor control any MCU will do!