XenonJohn

There is a big circular magnet in the base of the globe.The saucer shaped base repels it using electromagnets and a lot of cleverness to stop it just "falling off" the repelling electro-magnets (which it does do if you do not carefully allow it to learn where the globe is in space when you first set it up).

The code for the MPU6050 would be different. There is another instructable by someone who built a segway with an MPU6050 and you could probably look at their code, just get your machine to balance like a segway first then add other features back in one at a time. If one wheel motor could not physically turn for any reason (jammed chain for example), it will draw a huge current and probably would burn the motor controller. It is a very good idea to put a fuse or circuit breaker into the +ve feed from your battery pack.

Have you noticed this has now been copied (too) in China?It is one of those hoverboards with a seat stuck on top essentially. Described as self-balancing so the nose wheel is just for starting and "landing" I assume. The levers twist each half of the hoverboard instead of doing it with your feet.

Hi Ira, great project. I bought an MPU6050 to try out as much cheaper than the IMU's I use. I notice from Jeff Rowberg's site that the Angle_Y value is derived from on-board sensor fusion between the gyro and accel within the 6050 (Kalman filter or something like that). Therefore you do not need to do the sensor fusion with a complementary filter on the Arduino at all, as you have at the moment. I have changed the balancing part of your code so it back calculates a rate of rotational tilting for the D term of the PID, I use Angle_Y for the P term and have an I term too. You then play with the P I and D constants listed at the top plus the overallgain value until it balances. It might work more smoothly as a result as I use a very similar method in my current one wheelers. I have sent yo...

Hi Ira, great project. I bought an MPU6050 to try out as much cheaper than the IMU's I use. I notice from Jeff Rowberg's site that the Angle_Y value is derived from on-board sensor fusion between the gyro and accel within the 6050 (Kalman filter or something like that). Therefore you do not need to do the sensor fusion with a complementary filter on the Arduino at all, as you have at the moment. I have changed the balancing part of your code so it back calculates a rate of rotational tilting for the D term of the PID, I use Angle_Y for the P term and have an I term too. You then play with the P I and D constants listed at the top plus the overallgain value until it balances. It might work more smoothly as a result as I use a very similar method in my current one wheelers. I have sent you a PM as can email you the code to try out.Best wishesJohn

I have built a few self balancers using the Sparkfun Part No. SEN10121 IMU which also connects via I2C. I see some people here having trouble with their IMU not communicating well with the Arduino. Here is just my 2 cents worth but what worked for me was (i) to obtain some flexible thin lengths of coloured insulated wire by extracting them from some 9 core shielded cable (much easier to work with and keep everything neat between IMU and Arduino), (ii) keep these wires between Arduino and IMU really short, (iii) SOLDER the IMU wires to the Arduino rather than using push in pins. As my wires are thin I tin a 2mm uninsulated part at the end then solder that to the underside of the Arduino board where the relevant header pins just slightly poke through. (iv) I choose a common ground pin rat...

I have built a few self balancers using the Sparkfun Part No. SEN10121 IMU which also connects via I2C. I see some people here having trouble with their IMU not communicating well with the Arduino. Here is just my 2 cents worth but what worked for me was (i) to obtain some flexible thin lengths of coloured insulated wire by extracting them from some 9 core shielded cable (much easier to work with and keep everything neat between IMU and Arduino), (ii) keep these wires between Arduino and IMU really short, (iii) SOLDER the IMU wires to the Arduino rather than using push in pins. As my wires are thin I tin a 2mm uninsulated part at the end then solder that to the underside of the Arduino board where the relevant header pins just slightly poke through. (iv) I choose a common ground pin rather than using various GND pins on the Arduino (I have several to choose from as I use a Mega). Hope this may help somebody. I am not an engineer but have a huge amount of trial and error / falling-off experience!

The code is not written for the MPU6050. It was written for the specific Sparkfun one described in the Instructable.All the IMU's out there are not the same.

Someone on Hackaday has recently worked out how to use an Arduino to control the motor controller within those hoverboard toys. They use an unusual serial protocol. Therefore if I were to make this board now I would use two hoverboard wheels, the motor controller, their battery pack and an arduino with my own IMU.