[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: torsen tech 102
perhaps this is a better way of imaging it. picture a quattro (locker)
accelerating with arnie lifting up the front (i know it couldn't happen, but
this is a movie, so picture arnie on a skateboard). the issue is how much
tractive force the fronts can take. think friction circle. result is not
much. hence most of the engine's available torque must go through the
rears. the amount of tractive force available for the rears will depend
primarily upon how much weight transfer occurs. if arnie can lift the whole
front of the car (this is the movies), then you have 100% tractive force at
the rears, and 100% torque going there.
therefore, as a general rule of thumb tractive force with a locked centre
mirrors weight distribution. this issue was, as i said explored by freeman
in his paper "awd powertrain models" he was modelling a hummer, but in the
paper he quotes koga (1988) with a much more front heavy car, finding much
more torque distribution forward in the static case, and more torque
transfer rearwards upon acceleration. if you think about it, this is
undoubtedly one of the reasons why understeer was such a problem with the
with regards to your other question about the transition of the torsen from
locked mode to differentiating mode, this is dependent upon the rate of
speed difference between the output shafts. if you visualise the graph i
described in "torsen tech 103", this should help. for this torsen model,
the rate was 0.5 rad/sec speed difference before full biasing.
with regards to your understeer/oversteer argument, that is very simplistic.
it all depends upon many and various factors, but you can certainly have
100% rear torque bias with understeer (think bmw rwd), or 100% front bias
with oversteer (think peugeot 205 gti fwd).
however, the fact is that at the point where the torsen hits the bias ratio,
the locker will continue to bias torque, where the torsen will *hold* the
torque bias ratio (i didn't make this clear enough in "torsen tech 102") for
increasing amounts of torque coming in, while allowing output shaft speed
differences. the only situation where the torsen will go from (say) 75%
rear torque to 75% front torque distribution is when traction at the rear is
lost. the locker in the same situation will put 100% torque to the front.
so the torsen exhibits much better behaviour than the locker.
'88 mb 2.3-16
From: Lawson, Dave [mailto:firstname.lastname@example.org]
Sent: Saturday, 23 October 1999 05:23
Cc: quattro list
Subject: RE: torsen tech 102
>2) normal driving. the torsen diff acts as a bevel diff with a 50:50
>distribution. note that the locker can't and doesn't do this but maintains
>a torque distribution which mirrors the weight distribution of the
>i.e. front biased due to the relative lack of tractive force at the rear
>(more weight up front). this is the primary reason for the relative lack
>understeer for the torsen in contrast to the locker on corner entry.
The locker torque distribution doesn't mirror weight distribution until
the distributed torque to any given output exceeds the tractive force
of the output.
>3) cornering. the torsen diff is *locked*. read that again folks. the
>diff does *not* allow *any* speed difference between the front and rear
>axles. why is that? well it's a friction device, and the internal
>(i.e. stiction) prevents speed differences between the shafts. while the
>diff is not at the bias ratio, the diff is locked. however, the diff also
>allows torque to be transferred to the rear axle. how? well because it's
>acting as a locked diff at this point, and it is putting torque to where
>greater tractive force is. based on front and rear (relative) wheel slip.
>in *exactly* the same way as the locker.
Under your scenario, I am interested in the transition from "normal"
driving to "cornering". According to your description it goes from
open differential mode (equalising torque output while allowing speed
differences) to its "locker" mode (equalising shaft speed while allowing
torque differences). So the torsen is cruising along at 50:50 torque
distribution in open mode while the shafts could be at different speeds,
then corner entry and it transitions to "locked", shaft speeds all become
equal. To do this, there has got to be some "range" in which it can
operate in this "open" mode. I am wonder how big that range is, is it
+/- 5% about 50:50, or is it +/- 0.001% about the 50:50 mark?
>4) cornering at the bias ratio. at this point the diff is *unlocked* and
>will allow axle speed differences to occur. when the bias ratio is reached
>in cornering (e.g. 75% torque to the rear), the diff will apportion more
>torque from the fast rotating axle to the slower one. it is back to acting
>as a open diff. how often do you corner at the bias ratio? very rarely in
I wouldn't jump to the "very rarely" conclusion so fast.
>audi's figures show that a 80q accelerating around a 40m circle
>with lateral acceleration of 4m/s/s ends up with a final torque ratio of
>to the rear (torque is progressively moved to the rear). so the diff is
>still locked. according to audi it would take a circuit of 15m at the same
>speed to reach the bias ratio. most often the bias ratio is encountered on
>low cf surfaces, or a low speedon high cf surfaces and tight turns (parking
Given the conditions described above, the bias ratio is acheived when taking
a (15m)50 ft radius corner at 17.3 mph. To transition 62% of the torque to
the rear, all we need to do is take the (40m)130 ft radius corner at 28.3
That is about the radius of an on ramp, and may people take these types
of corners much faster. Add more speed and we approach the bias ratio.
>thats it folks. pretty simple really. a nice little device. does the
>torsen get confused by slip angle differences? no it can't any more than
>the locker can. because it doesn't see them. the torsen and the locker
>identical in behaviour during cornering because they are both locked and
>allow no axle speed differences. in both cases torque will go to where
>relative wheel slip dictates that it should go. once the bias ratio is
>reached (and about when the locker goes into terminal understeer), the
>torsen will limit understeer by diverting torque to the front (i.e.
>axle speed differences). it is at this point seeking to deny wheel spin on
What you have just described here is a behavior of the torsen to send
the bias ratio of torque to the rear of the car which sets up an oversteer
condition and then once the bias ratio is reached to divert torque to the
front which sets up an understeer condition.