Strange coefficient in offset formulae (mpu6050 calibration)

[vipnoob]

Hello there everyone, I got stuck with the following equation during the calibration IMU sensor, mpu6050
So there is the following library that needs for calibration the sensor: DIYino/MPU6050_calibration.ino at master · Protonerd/DIYino (github.com)

Code where I got stuck

bool calibration(){
ax_offset=-mean_ax/accel_offset_divisor;
ay_offset=-mean_ay/accel_offset_divisor;
//az_offset=-mean_az/accel_offset_divisor;
az_offset=(16384-mean_az)/8;

gx_offset=-mean_gx/gyro_offset_divisor;
gy_offset=-mean_gy/gyro_offset_divisor;
gz_offset=-mean_gz/gyro_offset_divisor;

so what does accel_offset_divisor mean ? Why does for acceleration it equal 8, but for gyro it equal 4 ?

[J-M-L]

when you have such questions, best is to go to the documentation .

Here is my naive understanding

if you look at this document "MPU Hardware Offset Registers Application Note Revision: 1.0"

you'll see (page 5) that the Gyro register offset (the bias) is measured in the +-1000dps sensitivity range

and (page 7) that the accelerometer offset is expressed for +-8G

they also mention that the acceleration offset is only on 15 bits, the 16th bit (LSb) should not be touched

---

Now In the code you listed for calibration, they set them at the highest sensitivity level of 250deg/s and 2g

          // set the fll scale range of the gyro- and accelerometer respectively
        accelgyro.setFullScaleGyroRange(0); //0: 250deg/s | 1: 500deg/s | 2: 1000deg/s | 3: 2000deg/s
        accelgyro.setFullScaleAccelRange(0); //0: 2g | 1: 4g | 2: 8g | 3: 16g

So since the offset values obtained are not in the expected range, they need to be adjusted.

from 1000dps to 250dps you go /4
from 8g to 2g you go /4 but the LSb is not part of the offset (it's stored on bit 1 to 15) so you need to get rid of that one too, so an extra shift right which is the same thing as dividing by 2 ➜ /4/2 <==> /8

So that's my guess

[vipnoob]

Thank you! Frankly speaking, I didn't know about this doc
So if got right for my settings (+- 2000 deg/s I need to value * 2) and (+- 16g I need to (value >> 1) * 2 ), right ?

[J-M-L]

The offset You obtained with the script as it is (max sensitivity and divide by 4 and 8) are what you want to use in your real code irrespective of the future settings you choose.