I am working with my phone's accelerometer and I've been getting some strange readings. When I am holding my phone in landscape/portrait/paralel to the ground I am getting acceleration roughly 10 m/s^2 on axis x/y/z respectively (which is fine). However, if I position my phone so that no axis is "dominant", the sum of acceleration on all axes is usually around 15 m/s^2, which seems like a pretty huge error. So my question is: Can measurement error be really that high, or am I simply misunderstanding how to interpret accelerometer's data?
Thanks
Just a guess, but could it be that you simply just added up the 3 values of the axes, instead of doing the maths ;) for reference look up "Parallelogram of force". The wikipedia page for that should help.
Greets from Germany