[torsen] Re: Haldex Riddle

Keith Maddock Keith.Maddock at trw.com
Wed Nov 20 09:55:08 EST 2002


Comments interlaced below.

SJ> I believe the exact example would indicate the possibility 
SJ> that the driveshafts are spinning at the same speed, 
SJ> wouldn't that actually cause an Unlock of both devices?

A viscous coupling will decouple when the driveshafts are spinning at the same speed.

The HaldexTM LSC will not decouple unless the ECU commands it.
This simple diagram shows that one-way check valves prevent the hydraulic pressure from being reduced unless it is done with the throttle valve.
This graph proves it.  See when the front and rear axle speeds are equal, the coupling pressure is held up?
*HOWEVER* in this circumstance, the HalexTM LSC will *NOT* be able to INCREASE pressure, unless there is a difference in axle speeds.

Now, what does the NVG/Jeep LSC do in this case?  Not sure!  Do they have the check valves, or is it purely a Coupling Force = f(Delta Axle Speed).  I think they don't have such a thing, because then you'd need some mechancial way of determining when to finally release the pressure.

Perhaps this is another "golden difference" between the Jeep/NVG and the HaldexTM units!!

SJ>  if there is ANY speed differential in driveshafts, 
SJ> BOTH devices should stay locked.  

With any speed differential, the Haldex will be able to increase, hold, or decrease lock.
However with the VC, the amount of lock is purely a function of the amount of speed differential.  There may be some small hysteris due to the temperature of the VC also affecting the lock, so it may stay at an increased level of lock and then gradually release as the fluid cools back down.

SJ> I think we be looking more at the ability for an ecu
SJ> intervention based on wheel speed input (abs sensors) 
SJ> causing the lock/increase lock in 
SJ> the Haldex and a release in the VC.

Haldex can hold a lock based on wheel speed (or other vehicle) sensors, but not increase it.
Conversely, they can also decrease the lock based on the other sensors, and they do this in turns, for example.

SJ> I'm not completly with you on the VC not locking up, 
SJ> but for sake of discussion, I'll accept  the argument. 

Thpppt :-P  Don't give up :)
Consider Haldex's own example:

Viscous coupling: 
Graph shows that the axle speeds never match. 
The coupling never locked.

HaldexTM LSC:
Graph shows that the axle speeds match.
The Coupling locked.

SJ>  I still think you are missing something. 
SJ> More for your discussions with the H-boys.  

I'm still asking you to let me know what I'm missing :)

SJ> I'm starting to really like this toy, TT application included 
SJ> (maybe with some tweeking).

Agreed.  The trick will mining the necessary info to make a simple external Haldex tweaker box.  It seems to me that as soon as they see a throttle release, they probably open the throttle valve and let the pressure out.  I will try to confirm this in Sweden.  If we can somehow override that and just allow the throttle valve to stay closed, we can maintain coupling!  

Problem that I can see, is that all current communication that we do to talk to the Haldex unit is via CAN.  It will be easy enough to find the appropriate CAN messages, I wonder how cheaply a simple CAN-box could be made that would just have the ability to transmit a few basic messages, based on a couple of buttons or a rotary knob...


Keith Maddock, TRW Automotive,  Koblenz, Germany
Slip Control Systems, Systems Design, Traction Control
+49 (0)261/ 895 2474     -    -    keith.maddock at trw.com

More information about the Torsen mailing list