[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: Torsen Tunnel Visions
I originally sent this message to Dave and Scott, but I think there
are at least a few others out here who are hanging with us. I invite
comments from anyone. If you're bored with the torsen thread (I
think most of you are), <delete> is just a keystroke away. The
quotes below are from Scott.
>What does that mean to the torsen? That the rears are "spinning"
>(since they are travelling faster), so torque to front. So initially the
>fronts get the torque. As Slip Angle increases, you get the rears
>slowing in relation to the fronts. Why? Because you can steer.
But in your scenario, you're talking about on-throttle oversteer. Even
if the torsen has shifted torque to the front, the rears are still getting
at least 30% of the total torque. If the rears are already in a slide (slip
angle), it shouldn't take much torque to get them spinning (forward
rotation), right? If this is the case, then the torque shouldn't be shifted
back to the rear until everything is lined up right again (barring some
really stupid steering inputs).
>Torsen: Rears slower than fronts, fronts "spinning", rears get torque.
>Torsen Centers don't care what the actual chassis dynamics are Dave.
>That's not it's job. It's dumb, it controls ALL chassis dynamics in an
>absolute traction world. Either the front and rear driveshaft spin at
>the same speed, or distribute torque. Don't make the mistake of
>making it smarter than it is.
I think YOU might be giving the torsen more credit than it is due. I've
got a blown-up printout of fig. 1 sitting here on my desk that I've
mangled with lots of vectors after staring at it for hours. I still don't
completely understand the forces at work (mainly because there are
several unknowns in our application -- the frictional interfaces mu 1
through 4, and the tooth angles theta and beta) but I think I've got a
grasp on the general workings.
Your interpretation sounds more applicable to VC. You say that as the
rears slow, torsen interprets this to be fronts spinning, so torque is
sent rearward. On the contrary, it appears to me that there is no
"spin-up" (speed difference) required to shift the torque, only a
difference in _supportable_ torque. The transfer is instantaneous.
However, it IS true that if a loss of traction on either end occurs, then
that end will spin faster than the end that can support torque as a
consequence of the torque shifting mechanisms.
Therefore, if the fronts are getting the torque, the rears must be (or
soon will be) spinning faster. If, as you say, the rears start to slow
down because of slip angle, the torque is instantly and gradually
transferred back towards equilibrium (50:50) according to how much
torque the rear can support. The torsen does not look at the shaft
speeds and at some point say "OK, rears are moving slower now, let's
dump 70% torque back there in one single event (shades of VC). It's
much too dumb for that.
We're looking at this from opposite viewpoints. You say that torsen
"reads" shaft speed to determine torque split. I say shaft speed is a
RESULT of torque split. What it "reads" is the shafts' resitance to
I could very well be wrong on this, but that's the conclusion I draw
from visualizing the forces at work in fig. 1.
'90 CQ 20V - my 'racing-iron' ;^)