Interesting ideas here, Colin!
First, as you know, the friction circle limits the amount of
braking when can put into our 1.2g turn so again something
around 0.30 is max, assuming a friction circle limit of
1.25g. However, as you also allude, things might be more
dramatic if we are pulling (as an extreme example) 1.2g's
BRAKING with 0.30g's turn. Now the error in the turn is
much more significant. But remember that our symptom was
that the reported Peaks (both Absolute and Sustained)
indicated a 0.15 difference. For lateral g levels much
lower (where this heavy brake/light turn would be skewed
much more seriously by a bad alignment) we have no symptom
that shows the problem and nothing to measure to locate it.
(I hope that made sense...)
I like your idea that a bad alignment should result in a
perfectly straight line producing a S turn. This is
correct, but the effect would not be noticable until the
alignment is really serious. Couple with this that there is
a small amount of "crosstalk" inherent in the accelerometer
package, and I doubt that you could use this as a valid test
of alignment.
The values we are looking for in alignment are generally
pretty easy to discern with the eye. An error of 6 degrees,
for instance, is equal to the difference on the minute hand
between 12:00 and 12:01. This is not too hard for most to
see, but if there are elements in your interior which are at
a kilter it can trick your eye. But as we saw yesterday, it
takes something like 10 degrees (about 2 minutes on our
minute hand example) to make a significant difference.
--Byron
Colin Watson wrote:
>
> Byron,
>
> Your calculations for a 5 degree misaligned cube assumed the maximum
> sustained G's were measured on corner exit where the car is accelerating at
> 0.30g. On corner entry I would expect the braking component could be much
> greater than 0.30g.
>
> What is a reasonable value for the braking component? I suspect that the
> difference between how fast the car accelerates vs. decelerates combined
> with the alignment angle is what causes the skew.
>
> How far off does the cube have to be misaligned before it makes an
> acceleration/deceleration run down a flat runway into an S curve?
>
> -Colin Watson
|