Gyroscope Sensor Comparison (MPU6050 & BNO055)

[DiscountBatman]

My mate and I are building competing rocket control designs. He's using the BNO055 and I'm using the MPU6050 with and Arduino Nano. Our code tends to max out at 100Hz, so we compared the raw sensor data at ~100Hz in the +-1000dps (degrees per second) range. We're reading the data straight from the registers on with the Wire library.

MPU6050_vs_BNO055_rawdata

The plot shows the measured data points over ~10 seconds, with the gyro bias removed. We calculated a standard deviation of ~3 LSB for the MPU6050 and ~6 for the BNO055. This suggests that the MPU6050 has less noise than the BNO055 when reading the raw gyro sensor data @ 100Hz 1000dps.

While this is only comparing one sensor of each type, the results are intriguing. Might even suggest the Bosch has reused the same MEMS architecture between the two sensors?

I'd recommend if you don't need the sensor fusion data of the BNO, just use the MPU and save yourself some $$$.

Fly straight!

Fero.

[OverTheTop]

If you can get the data out faster you can average a number of points and improve the noise by sqrt(n) if you average n samples, assuming they are available of course.