We are working with the LSM9DS1 and I would like to know if it is normal to obtain very large offsets for the zero-gauss level, and completely different offsets on different sensors (same PCB design).
I previously worked on the LSM303D (which is now NRND), and the magnetic offsets obtained with our calibration procedure were very small, and allowed us to use the LSM303D without magnetometer calibration.
On the LSM9DS1, I obtain offset values of up to 0.38 gauss (about the intensity of the earth magnetic field at my location). In addition, the offset values are very different from one sensor to another. For instance (for a magnetic scale +-8 gauss):
- PCB n°1
- offset x: -196 units (= -0.0479 gauss),
- offset y: 719 units (= 0.1755 gauss),
- offset z: -45 units (= -0.0110 gauss).
- PCB n°2
- offset x: 161 units (= 0.0393 gauss),
- offset y: 1591 units (= 0.3884 gauss),
- offset z: -721 units (= 0.1760 gauss).
I know that the datasheet indicates +-1 gauss for the zero-gauss level, but the effective differences between sensors are surprising.
- We tested 8 PCBs with the LSM9DS1.
- We followed the design guidelines for minimizing hard/soft-iron effects, and sensors are on “extra PCBs” with only the capacitors and I2C connectors.
- The calibration procedure has been done in an environment with little/no noticeable magnetic disturbance.
- We also check the calibration data with a 3D plot. Attached is a picture of the calibration data for the PCB 2.
- We used the self-test procedure to check the sensors.
Thanks in advance for your support.