Here to there wrote:

On Tue, 26 Oct 2004 12:32:02 -0700, Peter wrote:

Here to there wrote:


On Tue, 26 Oct 2004 18:51:45 GMT, Pete wrote:


But Molin didn't know he was putting more pressure on the tail than
it could bear. Why he didn't -- and who's to blame for that -- is the
subject of a bitter fight between Airbus and American.

I thought that was one of the main advantages of fly-by-wire systems,
to eliminate truly stupid actions of pilots. Sounds like Airbus shares
a lot of blame for the crash. It's like an auto maker made a car that
sheared off its wheels if the steering wheel was turned too quickly,
and the maker's response was to tell drivers, "Don't do that!"


Ummmmmm... so what exactly do you think will happen to a car if you
turn the wheel rapidly while driving at more than a snail's pace?

If it exceeds the available traction of the tires then I expect the
car to start sliding and possibly spin out. As long as the car
doesn't hit anything then I expect loss of tire rubber to be the
most serious damage. Of course if there is an impact (even with
something like a curb), then there are likely to be much more
severe consequences.

Except that's not the way it frequently happens in real life.
Rapid steering wheel movement at speed is one way that people
manage to flip cars, even when they haven't hit obstacles or
gone off the road. Around here, the tow trucks do a
land office business in the winter when the local students
decide to do donuts in the parking lots, and flip themselves. ;-)

In real life, parking lots unfortunately have many things you can
impact such as curbs, potholes, posts, etc. In the absence of those
there aren't all that many models of cars that can be flipped on a
flat parking lot. That was one of Nader's original complaints
about the Corvair and VW Beetle - due to an unusual rear suspension
design it was possible to flip these. There are also some vehicles
that are relatively narrow with a high center-of-gravity, but most
cars will not flip when driven on a flat surface regardless of the
control inputs.

I'll
give you a hint - you'll get the opportunity to find out either how
expensive it is to replace your suspension, CV joints, etc, or how
well your roof supports the weight of the car after it has flipped.
Probably you'll discover all of those.

BMW had a sales promotion event recently where they had us
try out some of their cars on a large parking lot with a
course laid out with cones. They actively encouraged aggressive
driving and there were frequent incidents where control was lost
resulting in the cars sliding and spinning. As far as I know there
was no serious damage done to any of the vehicles other than loss
of tire rubber (tires were replaced every 2-3 hours during the
event).



Were the drivers turning the wheels rapidly, all the way to
the stops?

Yes, the wheels were turned rapidly and the cars did spin out of
control - but there was no indication that any even came close to
flipping over.

According to the crash report, that seems to
be essentially what the first officer was doing with the rudder
as he attempted to recover from the turbulence.



"Don't do that" is a perfectly reasonable approach. You can't
make everything infinitely strong.

But if there's a clear rule for what 'shouldn't be done' then it
would seem prudent to build it into the firmware for the fly-by-wire
system so that it can't be done.

Well, perhaps, if it was a fly-by-wire system....

Yes, this accident was on the A300 without FBW - my comment was just
agreeing that this should be an advantage of the FBW systems.

My reading of the reports on the accident is that while the co-pilot's
actions may have been the proximate 'cause' of the tail's failure, the
fault was not the co-pilot's but rather with the training which failed
to indicate that such use of the rudder could cause structural failure.
Whether that's the fault of Airbus or American remains to be determined
- sounds like there's still plenty of finger-pointing going on.