You measure Acceleration with an Accelerometer (which ... measures acceleration). As I'm sure you know, acceleration is the second derivative of position, and in our three-dimensional world, position is a 3-D vector, usually expressed in terms of three vectors in the X, Y, and Z directions, with Z conventionally "up".
The sensor you have chosen, the ADXL 335, is a triaxial accelerometer with three output voltages proportional to acceleration in the X, Y, and Z axes (as defined by the chip). The chip specification tells you how to, for example, transform the voltage measured on the X output into acceleration along the X direction. The relationship between voltage and acceleration is (fortunately) linear, and can be expressed as "X-Voltage = Offset + Gain * X-Acceleration". The only problem is that Offset = 1.5 v ± 0.3 v, and Gain = 300 mv/g ± 30 mv/g. Do you see a problem here? What value do you use?
If you knew the actual value for Offset and Gain, you could invert the above equation and express Acceleration accurately in terms of the measured Voltage (which I assume you are measuring somehow with your Arduino). So how do you determine the Offset and Gain? Aha, the "Calibration of your Measurement System" problem. Think about how you'd solve this.
Here's a hint -- under what conditions do you know the acceleration of the Accelerometer in world 3D coordinates (which may or may not be the same as the Accelerometer 3D coordinates)?
Here's a further digression to help(??) with your thinking. I'm pretty sure that ADXL can tell you which are the X, Y, and Z axes of the Accelerometer (almost surely Z is normal to the plane of the chip, X is along one "side" of the chip, Y is the other side, and X, Y, and Z form a right-hand coordinate system). What are the "world" X, Y, and Z coordinates? Well, let's say that Z is "up", so -Z points to the center of the Earth. Where are X and Y? Who knows?, meaning it is more-or-less arbitrary, so try to ignore it for now, and work with the coordinate that you do know.
For what it is worth, some Masters-level BME students came to me with this exact question (a similar chip, similar accuracy) and didn't realize there was a problem. They had the same question, "How to measure acceleration". I tried to get them to figure out (for themselves) how to begin to calibrate their accelerometer, without "telling them the answer". What would you think to do?
One more hint -- you need no special equipment, but do need to understand something about vectors and the cartesian coordinate system.
Bob Schor
