S.E. Chocholeck talks Torsens...

[audidudi@mindspring.com (Jeffrey J. Goggin)]

The address and telephone number I tracked down for Mr. Chocholek (thanks to
the difficulty in pronouncing his last name -- it's "Sho-Sho-leck" -- he's
Stan to most people) turned out to be that of his summer cottage.  He said
he hasn't been spending much time there until now and I was fortunate to
have caught him in; he also said that he'd gotten the message I left on his
answering machine way back when but it had cut me off before I left my
telephone number.  He tried to track me down through the Gleason people but
no one there -- apparently, the department is down to only six people at
this point -- recognized my name (this isn't surprsing since I didn't
actually call them until earlier this week ... haven't received a call back
yet, either).  He confirmed he was indeed the author of the paper I've
posted at my web site [http://www.mindspring.com/~audidudi/Torsen.htm] and
seemed surprised that I was familiar with it [if he only knew!].  Although
he didn't say as much, I also think he was rather pleased that somebody had
enough interest in his work to track him down and discuss it with him.
Overall, he struck me as very friendly and outgoing, and quickly agreed to
answer a few questions for me...

I asked him where the "Torsen" idea came from in the first place.  He said
he knew the two people who invented the Torsen (my notes aren't clear on
this but I seem to recall him suggesting they were related somehow; maybe
father-and-son?) while he was working at Rockwell International as a test
engineer.  He said he blew up one of the early designs after driving just
five miles and didn't think it had much of a future ... he also said he's
always been known as an "anti-locker" kind of guy since he doesn't really
care much for limited-slip diffs for most people/in most cars/in most
situations.  When I suggested that I thought they were sometimes a
"band-aid" for poor chassis design and/or tuning, he quickly agreed ... he
also said he feels there are many other (i.e., better and cheaper) ways to
enhance traction for normal conditions.  I suggested accountants must love
him for this -- it's rare that an engineer prefers nothing over something --
and he laughed. (He also pointed out that cost was one of the reasons why
the viscous center diff is so popular these days; the fact that they also
tend to preserve the original car's basic character -- i.e., FWD cars still
behave like FWD cars and RWD cars still behave like RWD cars, only with more
traction when necessary -- hasn't exactly hindered its success, either.)

Despite his prediction to the contrary, the inventors convinced Gleason
their diff had a future and sold it to them, lock, stock and barrel.  They
tried developing it themselves but didn't have much luck.  Eventually, the
need for some outside help became obvious and when Gleason approached him
with an offer he couldn't refuse, he agreed to join them as a contract engineer.

After quite a few changes -- one of which apparently included convincing the
Torsen's inventors to go on to do other things -- he finally refined the
design enough so that it would survive as well as work.  The Torsen, he
said, is really quite "dumb" -- his *exact* words, I kid you not! -- and is
nothing more than a bleepity-bleep "gear-jamming friction device."  He said
the biggest problem they had to overcome was figuring out how to make gears
that were strong enough to survive the load yet also had the right degree of
friction necessary to allow the diff to operate properly. I'm not a
metallurgist but I gather this is one of those balancing acts that can be
very difficult to accomplish: If you make the gears hard enough to survive,
then they're too slippery; make them soft enough to provide the correct
amount of friction and they either break or wear out prematurely! 

In any event, they finally got things figured out and started marketing
Torsens as aftermarket replacements to dirt-track racers, where it met with
some success.  However, this stopped within a year or so because they had
some concerns about the company's liability exposure -- I didn't ask him to
explain this any further but think he was implying there were some
reliability problems in high-powered applications -- and the decision was
made to concentrate on serving the OEM market instead.  This is how they
became involved with Audi as well as other auto/truck manufacturers around
the world.
 
He said he remembers the only requirement Audi had for the diff was that it
allow the car to climb a hill with two wheels on the road and two wheels
off; they were also excited by the fact that the Torsen would allow them to
use ABS all of the time.  I explained to him that many Audi enthusiasts,
such as yours truly, are using the cars in ways that Audi probably never
imagined -- autocross, open-track days, some racing, etc. -- and some of us
with Torsen-equipped cars have noticed a few handling quirks that we
eventually concluded were being caused by the Torsen as a result of the way
it functions.

He seemed rather surprised by this and we began to talk about how the
differential works.  He said it works by sensing the amount of torque that
can be supported by each wheel (or in the case of an AWD car, the diff's two
output shafts) and tries to maximize this by allocating the torque
accordingly, up to the bias-ratio limit.  I suggested that a simplified
description might be that it attempts to equalize the rotational speeds of
the two output (drive)shafts and he said that while this isn't technically
accurate -- it's a result, not the cause -- he finds this a useful way to
characterize the Torsen's behavior this way for discussion purposes.

I asked him if it was possible for the Torsen to be "fooled" and allocate
torque incorrectly and he thought it was unlikely this would happen under
normal circumstances but agreed that it was definitely possible where one
end of the car -- i.e., both front wheels or both rear wheels -- was sliding
instead of rolling.  He said that in this case, the Torsen *will*
incorrectly allocate torque to the wrong end of the car.  He again reminded
me that it's a "dumb" device -- there's that word again! -- and distributes
torque based only upon the resistance offered by the driveshafts, which it
interprets as an indication of how much traction's available: The greater
the resistance it encounters, the more traction it assumes is available.
Thus it transfers MORE torque to the "slower-turning driveshaft" and LESS
torque to the "faster-turning driveshaft."  
 
Well, that being the case, I asked him how it deals with the slip-angle
differences between the front and rear wheels when the car is turning.  He
admitted this will fool it as well and more so as the corner radii becomes
smaller [which certainly correlates with my autocrossing experience] and
said this was one of the reasons that Land Rover rejected the Torsen.  For
most normal driving, though, he didn't think this was a problem since the
typical driver slows down for corners and/or backs off on the throttle,
which reduces the amount of torque the Torsen has to allocate ... for
racing, though, he definitely agreed it was a problem even though racecars
tend to have smaller slip angles.  In fact, he said, there was a factory
driver -- he couldn't remember who -- that absolutely refused to have a
Torsen in his car because he found it made the car too difficult to drive.
There was also another driver who didn't particularly like it but agreed to
race with it on occasion, depending upon the event.  [Could the fact that
one driver didn't use a Torsen be the reason why there is so much
conflicting evidence about whether Audi ever raced with one or not?]
    
At this point, he asked me more about how the car behaved during those times
the Torsen was suspected of being fooled and I explained to him the typical
"understeer/oversteer/understeer" scenario.  He speculated that it's
definitely possible for the Torsen to cause this, depending upon how much
torque's present and/or how well the chassis copes with the torque shift,
and said it may even be possible that it can get caught in a vicious loop
where correcting for oversteer by dialing-in some opposite lock with the
steering wheel changes the slip angle relationship between front and rear
wheels enough to cause the Torsen to incorrectly allocate the torque for a
*second* time, which in turn calls for the driver to make another correction
... however, he was very quick to add that he thought this sort of thing was
likely to happen only in exceptional cases and not everyday driving.

I asked him if perhaps reducing the bias-ratio (aka "softening" the Torsen
in industry jargon) might make this less of a problem for those of us who
take our Audis on the track and he thought it probably would ... he said he
also knew someone he thought could tell me exactly how to do this and
promised to get a name and number for me shortly.  I asked if he was
refering to shimming the side gears and he said, Yes, but that it was as
much a matter of using the correct material for the shims as getting the
dimensions right.  He said the pins the side gears rotate around can be also
changed and this might be necessary in some applications.

Unfortunately, I was running out of time at this point and didn't get to ask
him the myriad of other questions that were on my mind.  I asked if he would
mind if I called again with more questions and he said he'd be more than
happy to answer them if he can ... he also suggested that I talk to the Audi
racing department but when I told him they hadn't been very receptive to my
previous inquiries, he laughed and said that was typical of ALL the racers
he's worked with over the years.

* * * * * * * * * * * * * * * * * * * * * * * *

I have summarized my discussion with Mr. Chocholek as accurately and
impartially as I can ... I also know that some people who are reading this
will assume I've put a favorable spin things, intentionally or otherwise,
and consider his answers only as relevant as the questions I put to him.
While it's now clear (to me, anyway) the Torsen definitely has the
*potential* to "bite" in a center diff application, this doesn't necessarily
mean it actually does so ... that Dave Eaton, Phil Payne and others haven't
been able to find any evidence of it means there may well be other factors
involved here, something Mr. Chocholek said shouldn't be ruled out.  I've
been doing some thinking about this myself recently and think the answer
might be more obvious than we suspect... 

Anyway, my research into the Torsen diff and/or the "spider bite" phenomenon
is nowhere near finished and will continue ... stay tuned!  :^)
________________________________________________________________________
    _                _
   / |      _| o    | \       _| o  Jeffrey Goggin
  /__| | | / | | __ |  | | | / | |  audidudi@mindspring.com
 /   | |_| \_| |    |_/  |_| \_| |  http://www.mindspring.com/~audidudi/
________________________________________________________________________