Hi,
Occasionally I get a small value less than zero (e.g. -131) from the accelerometer Z axis i.e. negative gravity during ascent or hover.
If I have the accelerometer dlpf set to 0 it happens a lot (0.1% best guess based on code I have to detect it and abort a flight); by setting it to 2 it's a lot less if at all. It only happens when the props are rotating. The props all rotate the correct direction and are therefore only capable of producing positive values for the Z axis acceleration.
I'm using the internal FIFO, with ADC sampling divisor set to 1 (500Hz) over I2C @ 400kbps.
It is not physically possible for a Quad hovering to experience negative G - even a dead weight drops at > 0g due to friction.
Has anyone else seen something similar; I've seen this on 3 different sensors.
Thanks,
Andy