Troubleshooting Accelerometers - Calibration Issues

[czechman45]

Our lab just got these fancy new accelerometers that I'm trying to figure out. They are triaxial, but are not yet calibrated. As I was looking into how to calibrate them, I realized that, for a given axis, I would get a different voltage reading depending on whether the axis was oriented vertically (positive down), horizontally, or vertically (positive up). This is in fact how they recommend calibrating them.

Here's my problem. Doesn't this go against the whole idea of acceleration?! In each case, the sensor is not accelerating, it is just oriented differently, so why am I reading different values for the acceleration.

Also, does this mean that once the sensors are calibrated, that all of the readings I get will be orientation specific? If so, who is that dealt with? I would have a constantly changing orientation and no way to find euler angles.

Thanks for any help that you can give.

 

[MATLABdude]

Are you accounting for gravity, and downward acceleration of 9.81 m/s^2?

 

[MRFMengineer]

Yes, each axis will give different readings depending on orientation, that is the whole purpose of a triaxial accelerometer. You can compute the 3D acceleration vector relative to the calibrated origin. If you only care about total magnitude of acceleration, then you can can to compute the length of the vector by sum of squares. total^2 = x^2 + y^2 + z^2

 

[Unrest]

Here's my problem. Doesn't this go against the whole idea of acceleration?! In each case, the sensor is not accelerating, it is just oriented differently, so why am I reading different values for the acceleration.

Depending on your application, there may be fundamentally no solution to your problem. It's the equivalence principle that says acceleration and gravity are indistinguishable.