RE: yaw'ing

["Dave Eaton" <Dave.Eaton@clear.net.nz>]

actually it's seems pretty straight forward to me.  tcs systems have been
doing yaw control for years.  witness audi/bmw/mb et. al.  the downside of
locking the appropriate brakes on detection of yaw is those that i've noted.
cycle time and the requirement for torque amongst other things.  for
example, consider the task of asking your tcs-controlled (particularly 2wd)
car up a slippery slope.  no amount of tcs intervention is going to help you
if you can't get the torque down.

the use of (rear) differential technology to do the same sorts of things
certainly makes sense to me.  witness porsche's efforts on the original c4
system, suitably developed in the latest incarnation.  ditto mitsibushi.
audi/subaru next???  who was the 1st?  probably lancia actually with the
torsen in the rear in the 16v integrale, but what about the audi v8?

to say that the technology isn't going to help, and can only go wrong (that
sees to be what you're saying), seems to fly in the face of the facts that
every serious awd performance car is doing something similar.  not to
mention what the wrc rally boys have been doing for years now.  as for your
case of what yaw control will do if the car is sliding sideways makes me
smile.  in this case active rear diff control can certainly help, if you
figure it out.  mitsibushi, for example, delevoped their yaw control
algorithms to deal presisely with this case.  rally's would be much fun of
the cars aren't sideways.  so the system is fully capable of "the next
step" - controlling a car going sideways and making it quick.  this is the
big win over a simple tcs system which is the cotton-wool thing.  mitsibushi
even wrote an sae paper about it.

as i said before, you either get a passive diff to shift torque (speed or
torque sensing), but better to get a computer to do it for you, to make the
feedback loop tighter, and the control more precise.  we can all go back to
the understeering locked centre technology because it's safe and it works,
but we could use the same argument to dispense with the motronic, or the abs
computer...

what will audi do?  no-one has any idea yet.  certainly none of us do.  the
s6 sounds like it's either the next generation, or simply that the getrag
box requires a different solution.  time will tell.  i'm perfectly confident
that when it debuts, it will be good, and it will be better than what has
gone before.  just as generation 3 (torsen with edl and esp) was better than
generation 2 (torsen) was better than generation 1.  progress through
technology.  bring it on.

dave
'95 rs2
'90 ur-q
'61 mb fintail

-----Original Message-----

Date: Fri, 21 May 1999 08:54:20 EDT
From: QSHIPQ@aol.com
Subject: overmoment in yaw

Dave E. writes:
>yaw control is certainly now the state of the art, building on the advances
>made in coupling an awd system to tcs and abs systems.  for this to work
you
>need your usual compliment of abs sensors, a yaw rate sensor (usually
>mounted behind the c of g, a steering angle sensor, and a lateral "g"
>sensor).  the system works by comparing yaw with the steering angle and
>figuring out if understeer or oversteer is happening.

Can it make a mistake?  Like when sliding sideways?  What is it comparing
there?  Or does it shut down.  My thinking is it's just fooled.  It can't
"figure it out" all or most of the performance driving.  Most likely just
raises the comfort limit a little higher than a torque sensing diff.

 >the previous
>generation systems would then brake the left rear to correct understeer and
>the right front to correct oversteer (some systems apply all the brakes
>depending on the exact circumstances).  the major disadvantages with tcs
are
>basically the lag in the system, the "cycle time" for the brakes, and the
>requirement for torque to make it all continue to happen.  with an active
>rear diff, the locking rate of the diff is changed to produce the right
>"moment" across the rear (fully locked rear = understeer, unlocked rear =
>oversteer).

Again, the system is designed to avoid the "next step".  Once that next step
has "happened" the system has been fooled by definition.  Is it better than
the torque sensing, or just different?  I argue the latter.  If you can use
braking (which you already have) to accomplish the same thing (and probably
better, see archives) without the hardware and obvious disadvantages, the
numbers boys are happier.  Steamboat would be an interesting comparo of
these
compromised systems in a controlled "next" step environment.  Again, we are
looking at active systems, designed to "help" you, that may be doing the
exact opposite.  Hard to predict the nut behind the wheel.