@fraser: When car is stopped, the magnitude of the offset is so small, about mG's
@Patrick: I agree about studying X, Y, Z acceleration together. I believe that Z component influence may be significant.
I don't know how to correct different offset errors, I think that there are problems with orientation device due to wheels's flexibility or chasis's deformations. (the car's scale is 1:9).
I want to calculate velocity and position from my car by integration of the acceleration vector.