Having trouble figuring out raw accelerometer values

[Gixxerman1980]

I'm pretty new to all of this, but I'm learning. I'm currently trying to build a box where my tinyduino does three things.

1. Using a pot, it will act as a pulse wave modulator to control the speed of a vibration motor.
2. The on board accelerometer will monitor G total to track vibration from motor.
3. G total will be sent to LCD display for vibration level tracking

I've gone through the data sheet to the best of my knowledge and from what I can tell, the raw counts @ the 2g range should be +/- 256. Sitting still and relatively flat, I'm getting X: 9, Y: -9, Z: 261, with some slight variances over time. Inverted, I'm getting X: 13, Y: -13, Z: -250, again, with slight variances. I assumed that I shouldn't see anything higher than 256, so is there some calibration that I need to do? Also, I want to have the G total calculated and from what I've seen, I need to use the Pythagorean Theorem from my XYZ to figure that. Now, at rest, my G total should be 1.000. To figure that correctly, do I need to map my axis from +/- 256 to +/- 1.000?

[Ben Rose]

Looks like you're on the right track. As far as the values you're seeing, those look pretty good as well- you're unlikely to see exactly +256 to -256 out of the box. Luckily this sensor is likely to have an offset but otherwise be linear- so, for example, I think you could simply subtract 5 from the Z and get -255 to 256. If you find the min/max for the other axis and adjust those similarly, you should be pretty close to seeing 1g at rest when you divide the value you calculate by 256.