This is a quick tutorial to show how we hooked up a dIMU (An Accelerometer and Gyroscope for the LEGO MINDSTORMS NXT) to a Raspberry Pi. This is a quick How-To, that assumes you have a Raspberry PI with a Raspbian image on it (If you don’t have Raspbian, you can check out this fantastic tutorial on how to get setup.)
Update: You can find a tutorial on how to use a touch sensor with the Raspberry Pi here.

This is also a great example of how to get started with I2C on the Raspberry Pi. We chose the dIMU because it’s easy to access and use, with just four lines to connect (SDA, SCL, GND, and VCC). You may have to make modifications to attach other sensors.
We divide this tutorial up into 4 parts.
First, we get setup on Raspberry Pi to be able to use I2C commands in Python. Then we show how we connect to the Raspberry Pi. Then we show how we connect the Raspberry Pi to the NXT Bread Board Adapter and connect the sensor to the Raspberry Pi. Finally, we show you the Python code for accessing the I2C Sensor.
1). I2C Stuff
First, get I2C setup on your Raspberry Pi. There’s a great tutorial on how to do it here (http://www.instructables.com/id/Raspberry-Pi-I2C-Python/) . But to boil it down, open up your terminal screen and:
- Run the command ”sudo nano /etc/modprobe.d/raspi-blacklist.conf” and add a hash before “blacklist i2c-bcm2708” Save the file.
- Run the command “sudo nano etc/modules” add “i2c-dev” at the end, save the file.
- Run the command “sudo apt-get install i2c-tools” (you will need to be connected to the internet for this to work.
- Run the command “sudo apt-get install python-smbus”
- Run the command “sudo adduser pi i2c” to confugre the last two packages.
And you should now be ready to program I2C in Python.
2). Soldering the Adapter.
You’ll need to solder four jumpers or wires to connect the Raspberry Pi to the breadboard. We used jumper wires we found at Sparkfun that are great for prototyping. We also used a female socket to go over the GPIO pins, so we weren’t soldering directly onto the Raspberry Pi.
We’ve got a picture of our soldering work to the right. Below we’ll show a better breakout diagram to get to the I2C lines.
Those four wires are:
- VCC
- GND
- I2C – SCL
- I2C – SDA
Here’s how we’ll be hooking them up to the Raspberry Pi. Below is a diagram of how the four different lines should be attached to the GPIO pins of the Raspberry Pi.

3). Connecting to the Breadboard Adapter
Connecting to the breadboard adapter is very easy at this point. Now that you’ve soldered the jumpers to the GPIO on the Raspberry Pi, the NXT Breadboard Adapter makes this really easy.
Below is a breakout diagram of the NXT Breadboard Adapter as a Sensor. Connect SCL on the Raspberry Pi to SCL on the NXT Adapter, SDA on the Raspberry Pi to SDA on the NXT, GND to GND and VCC to VCC.

4). Running the Python Program
Finally, the good stuff: getting Python to read data from the sensor. Below is the code for doing reading the LEGO MINDSTORMS sensors over the I2C Line. We used Python 2.7, but this should work in any later version (if it doesn’t please let us know!).
A note about I2C and Python: the majority of the I2C functions use “smbus” in Python. This makes it super-easy. You’re going to use the following commands for most of your low-level transfers:
- bus.write_byte_data
- bus.read_byte_data
The smbus import is really well documented on the web. Doing a quick Google search yields a magnificent amount of information.
The Code: (Download it Here)
# Dexter Industries Example for Connecting the LEGO MINDSTORMS NXT Sensor to the Raspberry Pi # January 2013 # More information at www.dexterindustries.com/howto # # License - Attribute; if you use part of this code, please let us know! # We hope it's useful to you! import smbus import time bus = smbus.SMBus(0) # GYROSCOPE CONSTANTS AND VARIABLES ARE DEFINED address = 105 control_reg_1 = 0x20 control_reg_2 = 0x21 control_reg_3 = 0x22 control_reg_4 = 0x23 control_reg_5 = 0x24 g_x = -1.0000 g_y = -1.0000 g_z = -1.0000 gyro_divisor = 0.0000 # Constant to divide your gyroscope raw return by. # ACCELEROMETER CONSTANTS AND VARIABLES ARE DEFINED accel_address = 0x1D accel_divisor = 0.0000 # Constant to divide accel raw returns by. a_x = -1.0000 a_y = -1.0000 a_z = -1.0000 def setup_gyro(scale): global gyro_divisor bus.write_byte_data(address, control_reg_1, 15) time.sleep(0.1) bus.write_byte_data(address, control_reg_2, 0) time.sleep(0.1) bus.write_byte_data(address, control_reg_3, 8) time.sleep(0.1) if scale == 250: bus.write_byte_data(address, control_reg_4, 0) gyro_divisor = 128 time.sleep(0.1) elif scale == 500: bus.write_byte_data(address, control_reg_4, 16) gyro_divisor = 64 time.sleep(0.1) else: bus.write_byte_data(address, control_reg_4, 48) gyro_divisor = 16 time.sleep(0.1) bus.write_byte_data(address, control_reg_5, 0) time.sleep(0.1) def GetGyroValues(): global g_x # Need to be explicitly declared globals at the beginning global g_y global g_z global gyro_divisor xMSB = bus.read_byte_data(address, 0x29) xLSB = bus.read_byte_data(address, 0x28) g_x = ((xMSB << 8) | xLSB) if(g_x > 32767): g_x = -1*(65536-g_x) g_x = float(g_x/gyro_divisor) yMSB = bus.read_byte_data(address, 0x2B) yLSB = bus.read_byte_data(address, 0x2A) g_y = ((yMSB << 8) | yLSB) if(g_y > 32767): g_y = -1*(65536-g_y) g_y = float(g_y/gyro_divisor) zMSB = bus.read_byte_data(address, 0x2D) zLSB = bus.read_byte_data(address, 0x2C) g_z = ((zMSB << 8) | zLSB) if(g_z > 32767): g_z = -1*(65536-g_z) g_z = float(g_z/gyro_divisor) def setup_accel(range): # incoming "range" can be 2, 4, or 8. Default is 8 if it's not 2 or 4. global accel_divisor if(range == 2): bus.write_byte_data(accel_address, 0x16, 0x05) accel_divisor = 64.0000 elif(range == 4): bus.write_byte_data(accel_address, 0x16, 0x09) accel_divisor = 32.0000 else: bus.write_byte_data(accel_address, 0x16, 0x01) accel_divisor = 16.0000 def accel_axis_reading(register): axis_reading = 0.0000 axis_reading = bus.read_word_data(accel_address, register) if(axis_reading > 511): axis_reading = -1*(1023-axis_reading) # This deals with 2's compliment of raw reading 10 bit number axis_reading = float(axis_reading/accel_divisor) return axis_reading def GetAccelValues(): global a_x global a_y global a_z a_x = float(accel_axis_reading(0x00)) a_y = accel_axis_reading(0x02) a_z = accel_axis_reading(0x04) def main(): # Initialize the gyroscope and accelerometer setup_gyro(500) setup_accel(2) while True: # Update the values GetGyroValues(); GetAccelValues(); # Print the values print "Gyro Values: %d, %d, %d" % (g_x, g_y, g_z) print "Accel Values: %0.4f, %0.4f, %0.4f" % (a_x, a_y, a_z) time.sleep(0.5) # Sleep 1 second print " " #if __name__=='__main__': main()
If python is installed properly, you should be able to copy the code into a new file and type
python Raspberry_Pi_LEGO.py
And see the results!
So if you’ve read this far, you are really interested in Raspberry Pi and the LEGO MINDSTORMS. Drop us a line and let us know what you’d like to see next!

Really good article mate, I plan to try this out but need some Mindstorm cable plug bits for the breadboards. What about running a motor, or getting the PI to control the NXT via I2C
I have done it via USB
http://www.burf.org.uk/2012/12/28/raspberry-pi-and-the-lego-nxt/
Pingback: MAKE | Using Lego Mindstorms Sensors With a Raspberry Pi
Pingback: The Raspberry Pi mini computer gets NXTified