RE: torsen naderism

[audidudi@mindspring.com (Jeffrey J. Goggin)]

>2) even if it does occur on a type 44, you are completely wrong to
>attest that an occurance on a particular chassis means an occurance on
>all chassis.  overlaping design parameters of all sorts of things
>(tyres, roll stiffness, torque charactereistics etc. etc) will mean
>totally different results.  by definition.  the paper says this.  it's
>also obvious.

Once again, I disagree.  You need to separate how the device itself operates
from how a car -- ANY car -- with said device installed operates.  The
operational properties of a Torsen diff are the same regardless of whether
it's installed in a FWD, RWD or AWD car; that various cars may or may not
respond differently to its presence isn't a reflection of the device but the
whole system of which it's but one small part.

The Torsen allocates torque based on the input signals it receives from the
driveshafts ... nothing more, nothing less.  Assuming identical design (and
this appears to be true of every Torsen Audi used in a center diff
application), they'll respond identically to the same set of input signals
from the driveshafts.  It doesn't matter whether it's set up on a test rig,
in a FWD car, RWD car or AWD car ... the results are ALWAYS the same,
provided the input signals are the same.

To me, this is a non-negotiable point.  The Torsen doesn't -- CAN'T! --
operate one way in your RS2 and another way in my 200q because the device
itself hasn't changed.  Put a bullet in a gun and it'll function exactly the
same way every time you pull the trigger ... it doesn't matter whether
you're shooting at a target or a person.  The end results may be different
in each instance but manner in which the bullet itself operated didn't change.

>for example, you talk about the understeer characteristic of your torsen
>type 44 chassis as though that is part and parcel of the torsen.  of
>course it's not.  for example, the torsen ur-q exhibits totally
>different handling characteristics to the non-torsen one (much less
>understeer), and the type 44 i attest.  my old s2 understeered *more*
>than my ur-q.  i've also read test drive articles of recent quattro
>audis where the tester says the chassis only exhibits understeer at the
>limits of adhesion (ref: a6q test in latest car mag).

I'm lost here ... what exactly is it that you're trying to say?  Understeer
is part and parcel of ALL Audi designs -- heck, it's practically part and
parcel of ALL steet cars, especially those that carry more than 60 percent
of their weight on the front wheels -- and stuffing a Torsen diff inside a
Type 016 transaxle doesn't automatically change it.    
>i simply cannot understand the statement that "since it occurs on one
>chassis, it must occur on all".  this is just pain wrong/ignorant.  if
>you continue to accept that this statement is true, then turn it around
>(it doesn't occur on one chasis, so it can't occur on any), and the
>opposite must be true as well.

Hmmm ... blue and yellow paint mixed together make green paint; therefore,
it must also be true that all green paint is made from blue and yellow paint
mixed together.  Right.

Again, Dave, you have to separate the operation of the Torsen from the
operation of the car that is equipped with a Torsen.  These are two
different things entirely and you're not being careful enough to allow for
this.    

>fact is that audo had *quite* different design ideals for the type 44
>than for the ur-q, the s2 and the rs2 (the chassis which i've owned).
>markedly different characteristics.  even between these 3 cars.

Granted.  But Audi used, in essence, the same diff ... what does that tell you?

>as the paper makes quite clear in the preface, the art of engineering
>design is to optimise the various *overlapping* systems to produce
>desired effects, supress undesired defects, and reduce cost.  all at the
>same time.  it also makes quite clear (and its obvious) that these
>factors in an awd application include a long list of chassis dynamics.

Agreed ... but the Torsen itself operates exactly the same way every time.
Feed it the same input signals and you'll get exactly the same torque split,
every time.  Alas, the input signals themselves are fairly crude and there
are several different situations in which the same signals can be generated
and the Torsen has no way of telling them apart ... there's no context for
it to determine which "their," "there" or "they're" is meant and when this
happens, it can respond inappopriately.  Why is this so hard to grasp?

>so, is it possible to have an occurance on one chassis, and not on the
>other?  damn right.

No, it's possible that one chassis copes with its occurence better than
another.  If it happens in one, it happens in all ... it simply can't be any
other way.

>this is one of the reasons i object very strongly to your "science".  it
>isn't [science].  the thing i learned very early in my chemistry was
>that you had to be very careful with your null hypothesis so that you
>isolated just the factor you were after *without* any overlapping
>condition ruining your results.

Then I'm afraid it appears you didn't learn this lesson very well ... the
Torsen has no brain.  It can only respond in a preprogrammed way.  Sometimes
it gets things right (as your experience suggests) and sometimes it gets
things wrong (as mine and Scott's would suggest).  Either way, the Torsen
itself is always operating exactly the same as it has no other alternative.

Again, I find this whole subject tiresome.  My goal was to stimulate a
discussion about how to modify its operational characteristics and instead,
I've been bogged down having to explain them.  This isn't fun any more.
Either you grasp these concepts and we move on toether or you don't and I
move on alone.  (Okay, Scott can come along, too.)  :^)
 
________________________________________________________________________
    _                _
   / |      _| o    | \       _| o  Jeffrey Goggin
  /__| | | / | | __ |  | | | / | |  audidudi@mindspring.com
 /   | |_| \_| |    |_/  |_| \_| |  http://www.mindspring.com/~audidudi/
________________________________________________________________________