acceleration sensors: which one do you use ?

[Al Johnson]

On Friday 26 June 2009, Marcelo wrote:
> On Fri, Jun 26, 2009 at 14:28, Sebastian
>
> Krzyszkowiak<seba.dos1 at gmail.com> wrote:
> > But that would be nice if someone will explain, how that second accel
> > can be used, where the first couldn't...
>
> Since you have acceleration for two different points in the body and
> you know the positions of those points, you can compute the direction
> in which it's rotating wrt to some axis.  You can't do that with a
> single acceleration vector without resorting to collecting samples
> over time.

In theory you can calculate angular acceleration about two axes, but in 
practise the small separation of the accelerometers means this will only work 
for large angular accelerations. If this is what you want you would be better 
replacing one of the accelerometers with a 3 axis gyro, though a 3 axis 
magnetometer may actually be more useful.