# RE: Dave E. - you're making the wrong assumptions.

```now scott, we're really getting into the 'what ifs' now aren't we?

so, we're really overcooked the corner, so much that the rears have *no*
forward motion, only sidewards slip, and the fronts are still going

>-----Original Message-----
>
>Date: Tue, 3 Mar 1998 11:54:32 EST
>From: QSHIPQ <QSHIPQ@aol.com>
>Subject: Dave E. - you're making the wrong assumptions.
>
>Dave you are missing the biggest point.  Look at it this way, take a car any
>car go into a turn at 50mph, lay it out sideways, steering with the skid,
>what
>happens?  The rear tires will stop ALL forward rotation, the fronts, since
>they have a lower slip angle, will still have forward rotation.  REREAD this.
>Now, go at that same turn with your Torsen q.  The physical properties of
>that
>track and slip didn't change, YOU have to change them with power or lift
>throttle.
>
>What does the torsen think of this disparity, accel on?  IT assumes the
>differential between the slower rear driveshaft and the faster front to mean
>that the faster front is spinning.  Takes Tshift and puts it to the rear, YOU
>don't have to lift a wheel for this to happen.  AND you haven't equalized T1
>to T2, because you haven't done anything to reduce forward traction (which
>reduces transmitted Trg).  As soon as Tshift goes forward, the rear wheel
>slows again due to track NOT traction, and you hunt Tshift again, btdt.  NO
>lifted wheel.
>
>not it doesn't.  i accept that at the worst, in this scenario, the torsen
>will supply t(min) to the front axle, and t(max) to the rear axle (given *no*
>rear axle rotation).  in this case the torsen will pull you through the
>corner with t(min) to the front, progressively dialling in more torque to the
>rear as the rears gain longitudional traction.  in reality, the torsen will
>apportion as much torque to the rear axle as it will support, the rest going
>to the front axle.  this is what you want....
>
>the generation 1 locker will do the same thing, except that t(min) will be
>greater for the fronts (at 50%), than the 30% available to the torsen.  a win
>to the generation 1 manually locked diff.
>
>but then we've seen that the wheel lift scenario is a win to the torsen over
>the gen 1 locked diff.
>
>however, most oversteer that i've ever had experience of involves sidewards
>slip due to the increase in longitudional traction being demanded of the
>wheel (classic definiton), overcoming cornering and traction adhesion.  in
>this case the effect of which you speak doesn't occur due to the higher
>rotation of the rear drive shaft than the fronts.  in this case torque is
>removed from the rear axle and sent to the front, which is exactly what you
>want.
>
>so where's the hunt?
>
>Dave, frustration is high, I understand.  FORGET wheel lift, you don't have
>it
>in my scenario, so Trg stays at maximum, so does the Tshift front to rear.
>Understand, you can have a rear wheel spinning, and not have Tshift OR a
>reduction in Trg.  By definition of the center torsen in a turn.  Reread page
>10, the unlikely has to do with a torsen on a common axle, NOT a Center.
>This
>is the physics of the bite.  You don't understand it, that is different from
>the reality of maximum traction in a straight line, and a torsen center in a
>turn.
>
>no you can't scott.  thats not what it says on page 10.  you have a wheel
>spinning and you have maximum t(shift) with the torsen and also with the
>manually locked generation 1 diff (not shift in this case as much as
>modulation).  the determinant here with the torsen is t(min).  if t(min)
>torque causes wheel spin on one virtual axle, then t(min) *stays* to the axle
>(and disappears as wheel spin).  this is a *good* thing.  the generation 1
>manually locked centre diff will *also* do this.  torque will 'disappear' to
>the axle in this case (ie. 50% of  torque to the spinning wheel).
>
>so with the torsen you have maximum torque loss of t(min) (30% with quattro),
>while the manually locked generation 1 cars have a maxiumum t(min) loss of
>50%.  win to the torsen.
>
>dave
>'95 rs2
>'90 ur-q
>
>

```