Hi Everyone,

I'm testing Madgwick's quaternion DCM C code with a Sparkfun 9DOF sensor stick. It's oscillating a lot due to the accel integral term (like 5-10 degrees peak-to-peak, lasting for minutes).

I know the idea is to keep the P gain low, so that the IMU is not as sensitive to dynamic acceleration, but other than increasing P, massively reducing I, or maybe adding a D term (which nobody else seems to need), I'm not sure how else to dampen the osc..

Any tips?

Thanks.