Posts: 2,179
Threads: 20
Joined: Jun 2016
Reputation:
64
This is not supported and seems it has mpu9250 imu as well as a pni magnetometer?
I have no experience with this device, but if it is a PNI magnetic sensor it would have a much higher resolution than a hall sensor, but this is not the limiting factor in pypilot by any means. Most of the steering is based off of the gyros, and making the compass more precise would have limited if any improvement.
Basically, at least half the mpu9250/mpu9255 I tested several hundreds were thrown away. The quality control very bad from the cheapest source. I have since switched to icm20948 which has sensors which are no better, but all the ones I have used from reputable sources have proven reliable.
As for "half of 360 is 189" this is still good enough to hold a course. It may have something to do with the specific location the sensors are mounted, poor compass calibration or simply need to calibrate the accelerometers. The mpu9250 often did not have factory calibrated accelerometers, and sometimes they could not be properly calibrated which leads to a lot of errors. Further, often one or more other sensors did not work correctly, for example one of the gyros or even compass axes did not work. I did not see such errors with icm20948.
I suggest you simply use icm20948 and compare the results since it is already supported. This sensor you link would not provide an advantage in precision without also greatly improved gyros as well.
It tells me, for example, half of 360° is 189°
