We used to say that it was included in the calculations, but
in reality, it would be more accurate to say that the
calculations make the difference negligible. Here's why...
After setting a good zero setting with the cube flat on a
table, what you are doing as you measure on each side is
setting a good one-G setting in each direction. We assume
that the accelerometer is sitting at 90 degrees to the
table, although as you correctly pointed out, it's actually
89 degrees. Since we know that earth g's are always 1 g,
what you are asking about is the slight loss in g's when the
angle measured is actually 89 degrees instead of 90 as our
math assumes. The difference between the two is the
difference between the sin of 90 degrees and the sin of 89
degrees. The sine of 90 is 1.00, hence 1.00 g's. The sine
of 89 is 0.9998. Since our accelerometer rounds to the
nearest 100th of a g, this is still 1g. Of course, if we
had more resolution and accuracy we could claim that we know
the difference in that "missing" 0.0002g, but, hey, we
don't. :-)
On the other side of the equation, the zero g setting is
really, really important to get right. One degree there
makes a sizable difference, about 0.02g.
--Byron
John Moody wrote:
>
> Question: should I actually put the cube on its side, or would it be more
> accurate to rest the base against a square block during calibration?
>
> I guess what I'm asking is does the calibration routine account for the fact
> that the sides of the cube are not square to the base. Unless otherwise
> known, I would calibrate the cube on a known level surface, and use a 90
> degree square block against the base for all the side, front, and back
> positions.
>
> Best regards,
> John Moody
/// Archives at http://www.team.net/archive
/// Check http://www.team.net/mailman/listinfo for other lists of interest.
|