[torsen] Re: Haldex differential

QSHIPQ at aol.com QSHIPQ at aol.com
Tue Nov 19 10:53:54 EST 2002

In a message dated 11/19/02 7:27:14 AM Central Standard Time, Kei
th.Maddock at trw.com writes:


>I see where you are going with "torque shift" in terms of moving torque, as >
opposed to "torque capacity".  Though by your own rules, since the rear axle 
has >0% torque in the no-slip, and is then capable of receiving 100% torque 
at full lock >(assuming strong enough clutch), wouldn't you call that a 
"shift" of 100% torque?

Keith, you have to envision that in terms of what's happening.  In the case 
of 100/0, that's a capacity argument that lasts exactly the length of the 
wheelbase.  In the case of a torsen, it would be 0/0 for the length of the 
wheelbase. Is either relevent or reflected in "torque shift"?  If so, we need 
to be consistent with all limited slip devices in addressing shift AND 
capacity, currently it's not common practice.  I'll leave that up to the SAE 
boys to figure out whilst defining awd.

>I understand if this was a locking or torsen diff, that since the no-slip 
torque >transfer characteristic is 50/50, you're only "shifting" up to 50% 
more to either >end, but the LSC isn't a differential based system like the 

Exactly, it is a 95f/5r or a 50f/50r diff.  The real gray area comes into 
play here because with any ls/locked diff/coupling, capacity is hugely 
different than torque shift.

>To-mayto, To-mahto? :)
Maybe,  I think the vision is that you can shift torque from one axle to the 
other and anywhere in between.  The coupling can only "slip" in one direction 
as you said with 1 input and 1 output.  A LSD can slip in either direction 1 
input/2ouputs.  A nice marketing arena that audi uses fully to their 

>I'll see what I can find out from the Haldex boys this winter :)   We'll be 
working >with a Golf R32 and a few Saabs with Haldex Gen I's, and a S3 with a 
Haldex Gen >II transplant, so I'm sure we'll have more than a few 
conversations with them!

Looking forward to it.  Keith, right now, the system is deactivated and 
reactivated on the fly based on a variety of control inputs from the ecu.  I 
really don't see any difference between that intervention and the ability for 
the throttle valve to regulate the system more "aggressively".  Remember, as 
with ANY center coupling device, a turn is a traction argument.  With <15% of 
rotational differences, I would argue that the  haldex system is active (but 
throttled open) as soon as you move the car from straight ahead.  This would 
support a more complete electronic control of the device.  

The hackers will figure this out if the engineers don't.  No question in my 
mind that the TT can be run in "locked" mode with the simple override of the 
throttle valve (the "switch" they describe in the application section).  We 
know that the throttle valve can completely unlock the coupling, which the 
device can't do by definition, unless it is the locking device as well.  One 
could argue we need some sort of slip to activate the lock, but once locked, 
it certainly could stay that way, pick and choose your other included ECU 
standard interventions.  At steamboat for example, you default the diff to be 
locked all the time, except under braking (or including braking), add in EDL, 
ESC, or ABS algorithums based on track times.  

Get those boys talking man.  Right now the TT is one of the worst performers 
at steamboat, btinstructed that.  I'm thinking it can be better than the 
lockers (or be one)  with a simple understanding of the pinouts to the 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.audifans.com/pipermail/torsen/attachments/20021119/eab69d94/attachment.htm

More information about the Torsen mailing list