101

[QSHIPQ <QSHIPQ@aol.com>]

>yes scott, and, to use your argument, the *physics* of the gen 1 locked
centre
>is a torque transfer of +/- 50%....
All four wheels same cf, not correct.  You don't want to go here Dave.  That
matrix is impossible, not just improbable for either Gen I or Gen II>
The bite happens in a constant cf environment, that's an easy one.  In a non
constant cf environment, all sorts of events become significant.  Don't make
this too complicated, you can't even address the model before adding in 4
different cf's to the equation.
>*****************************************************************************
>my question remains, forget the bs, and tell me *how* i can reproduce this
>fabled spider bite in either of my cars...
>*****************************************************************************
Haven't driven your car, can't give you the non bs answer.  Come to the states
or fly me there, I have a high confidence that I can make spiders fly in any
torsen.  By physics I can show it CAN happen.  By driving your car, I can
identify the baseline for the model.  Your logic is that physics can't happen
cuz 'I can't get it to happen'.  No, that's not physics, that's the model.
Does it affect chassis dynamics?  Your logic and non-event says NO.  My logic,
physics of the switch, and my BTDT, explains your non event and my event.
That isn't a two way street.

>btw, in your ramble about statistical non-events, you forget that *every*
event
>has statistical significance.  if i have 10 events, and 10,000 non-events,
then
>i can derive a distribution, a standard deviation, and figure that,
>statistically, your *event*, is actually a *non* event (ie. outside the 3rd
>s.d. of a normal curve).  enough said...
NOT TRUE.  For statitical nonevents to be true you have to have a correlation
that your nonevent is statistically significant TO the model and the physical
properties of the switch.  You have shown neither, the most basic being
because there is no model.  Doesn't change the physical properties of the
switch,  nor does it address my EVENT.  For a given set of variables, and the
hypothesis, and the physical model of the switch, my event is not 3 standard
deviations from the mean.  A non significant event, with no correlation to the
model or the physics of the hypothesis, is just a non-significant event.  It
says nothing of the event, the nonevent or the physics of the switch.  In fact
we could question your methodology given Tshift CAN happen, with numbers that
indicate that the deviation may not be so far skewed.
Null Hypothesis:  A torsen center car will exibit traction behavior with all
chassis variables.  This will have no effect on chassis dynamics thru a turn. 

Model:  Given cf (constant), suspension (variable), tires (variable), throttle
position (variable), BR (variable), Slip Angle (variable) front, slip angle
rear (variable), Wheelbase (constant, but variable to different apps),
Relative slip angle (variable), Trg (variable), and chassis (constant, but
variable to different apps).  We found the following numbers.  
Let's get the cray.....

Methodology:  Phil and Dave write:  "I did this and there was no effect on
chassis dynamics"  Er, ah, ok.  I did this bite and found regardless of the
variables, but given Tshift properties, that the absolute traction device is
ONLY an absolute traction device.  Significant to the physics, the null, the
hypothesis, and the summary of the methodology.

Summary:  The Null is not correct, chassis dynamics indicate that given the
right set of variables, Tshift occurs, and in fact, the chassis dynamics can
upset the turning characteristics of the vehicle.

Implications for further research:  Oversimplified model based on btdt.  The
Physics indicates this can happen.  More research needed to address the
modeling of this phenomenon.  Given U-O-U is in the design of the switch in
reference to audi chassis dynamics, the switch does effect chassis dynamics by
definition.

Phil and Dave:  No event, wrong physics and wrong conclusion.

The model is there, you are looking for easy answers, I don't.  Combining a
bunch of variables Tshift happens.  What they are?  A lot of em.  Get em
right, the torsen treats you exactly wrong.  Why?  Cuz a torsen is an absolute
traction device, in a world of variables that are "visioned" as constants.

I'm dumb, it's dumb, and the bite happens.  So does the physics of the switch.
You agree with that, you have to.  I'm not sure I need more beef.  That would
be a noneventsters job.  A statistical nonevent is hardly the answer, simple
or complicated.  

Scott