There are some semiconductor mems devices that measure acceleration. I have tested two of them and, on both, I have noticed the same problem that bother me.
Imagine that an accelerometer (Ac) has a frequency answer, which can be limited by filtering. Say that that band is limited to 2Hz. Than, the Ac can be used as a tilt sensor, with a pretty good reproducibility of data.
If you have the brilliant idea to use the Ac as a change of tilt sensor, the dc voltage that the Ac outputs will be 2.5V for horizontal position, 3.0V for the tilt of 22° to the right, and 2.0V for the tilt of 22° on the left. You move the Ac about the axis, from extreme left to extreme right position, at a slow speed of 0.5Hz, and you get a signal from 2.0V to 3.0V. On the scope, it looks that the move coincides with the extreme values of the Ac. So the signal is following the move without a noticeable delay or change of the predicted value for the position.
This is in a continuous move (at the above mentioned speed).
If you start from rest, in any position (that means from horizontal, or from extreme right, or from extreme left), if the rest position signal is 2.5V and you move to the right, the signal should increase, gradually to 3.0V, as you tilt to a 22° to the right. Well, for the first 250 millisecond, the signal does not grow more positive. On the contrary, it decreases by a 100mV. In fact imagine a sinusoid starting from 0 and going as the graph of 180° to 360° of a sinusoid goes.
Therefore, the signal decreases when it should increase.
I attributed that to the inertia of the measuring system when starting from rest. It is bothering, because it is not consistent. It happens only when the move starts from rest. If going back and forth, there is not any overlay-ed signal and everything seems smooth.
What is the cause of the different behaviour with respect to inertia (when oscillating, it too changes direction, therefore, inertia is implied)?
2
MEMS Accelerometer
Re: MEMS Accelerometer
Good Answers: