viscous coupling

[donald girod]

I have a 2001 Impreza standard trans., and I understand that the rear
wheels are driven thru a "viscous coupling". I would like to know just
how this works. What I have read indicates that the coupling contains
two sets of interleaved disks, one set driven by the transmission and
the other set connected to the rear drive shaft, hence driving the rear
wheels; the coupling is filled with a heavy silicone fluid which
transmits torque from the transmission to the rear axle when the
relative rotation between the sets of disks exceeds a certain rpm.

When you are driving in a straight line on dry pavement, is there in
fact any torque transmitted to the rear wheels? That is, is the normal
relative rotation between the disks significantly greater than zero?

At what relative rotation rate does the transferred torque become
noticeable? Something I read indicated that at a relative rpm of 100,
the transferred torque is about 70 ft-lb.

Obviously if the front wheels lose traction and spin, torque will
transfer to the rear. The reverse situation seems to be impossible,
since at least one front wheel would seem to have to be rotating faster
than the rear ones in order for any transfer to take place. Although in
a turn, the rear wheels turn faster than the front wheels and I guess
would feel some drag. Do I have this right?

 

[Jim Stewart]

I have a 2001 Impreza standard trans., and I understand that the rear
wheels are driven thru a "viscous coupling". I would like to know just
how this works. What I have read indicates that the coupling contains
two sets of interleaved disks, one set driven by the transmission and
the other set connected to the rear drive shaft, hence driving the rear
wheels; the coupling is filled with a heavy silicone fluid which
transmits torque from the transmission to the rear axle when the
relative rotation between the sets of disks exceeds a certain rpm.

There is a center differential. The engine
drives it and one output goes to the front
and the other to the rear. The VC is inside
this center diff.

When you are driving in a straight line on dry pavement, is there in
fact any torque transmitted to the rear wheels? That is, is the normal
relative rotation between the disks significantly greater than zero?

Definitely. The front and rear are driven at
all times through the center differential. As
long as the fronts and rears are turning the
same speed, there's a 50/50 torque split and
the VC is not doing a thing.

When there is a speed difference, the fluid
in the VC heats up. That forces the plates
inside the VC closer together and causes the
front and rear drive to be locked together,
"short circuiting" the center dif.

At what relative rotation rate does the transferred torque become
noticeable? Something I read indicated that at a relative rpm of 100,
the transferred torque is about 70 ft-lb.

Obviously if the front wheels lose traction and spin, torque will
transfer to the rear. The reverse situation seems to be impossible,
since at least one front wheel would seem to have to be rotating faster
than the rear ones in order for any transfer to take place. Although in
a turn, the rear wheels turn faster than the front wheels and I guess
would feel some drag. Do I have this right?

No. The VC will lock any time there is a large
*difference* between the front and rear. It doesn't
matter which end is spinning and which is standing
still.

I don't think that you could feel the drag on a
turn if the VC was working right. It takes some
finite time for the VC to heat up and lock.

My wife claims she can smell the VC locking when
she drives in the snow.

 

[Rob Munach]

There is a center differential. The engine
drives it and one output goes to the front
and the other to the rear. The VC is inside
this center diff.



Definitely. The front and rear are driven at
all times through the center differential. As
long as the fronts and rears are turning the
same speed, there's a 50/50 torque split and
the VC is not doing a thing.

When there is a speed difference, the fluid
in the VC heats up. That forces the plates
inside the VC closer together and causes the
front and rear drive to be locked together,
"short circuiting" the center dif.



No. The VC will lock any time there is a large
*difference* between the front and rear. It doesn't
matter which end is spinning and which is standing
still.

I don't think that you could feel the drag on a
turn if the VC was working right. It takes some
finite time for the VC to heat up and lock.

My wife claims she can smell the VC locking when
she drives in the snow.

That is an excellent response. FWIW, I thought that when the VC heat up,
the fluid thickens which gives it its "lock-up' characteristics. I don't
believe there are any plates in there.

Regards,

 

[gpatmac]

Without looking it up (I'm not absolutely sure) I thought that the front and
rear diffs were open on the 01 auto.

 

[Danny Russell]

That is an excellent response. FWIW, I thought that when the VC heat up,
the fluid thickens which gives it its "lock-up' characteristics. I don't
believe there are any plates in there.

Rob, yeah there's no friction plates but rather a series of interleaved slotted plates (vanes).
Half are coupled to the front while the remaining half are coupled to the rear. They don't contact
each other, but they're spaced very close to each other. -Danny

 

[Jim Stewart]

Rob Munach wrote:

That is an excellent response. FWIW, I thought that when the VC heat up,
the fluid thickens which gives it its "lock-up' characteristics. I don't
believe there are any plates in there.

Thanks. You know, we really need a VC/AWD FAQ
somewhere. There's info scattered on the web,
but nothing that's real specific to Subaru. I
spent about 20 minutes going through the service
manual so I could back up what I said, but it's
really no help.

 

[ed]

Rob

There definitely are plates in the VC, I have one I cut an opening in to
show people what the inside parts look like. The fluid fills ripples stamped
out of the disks to give drive when it thickens. And regarding Jim's post
that his wife smells the VC locking, the fluid is sealed in the VC, so it
doesn't comtaminate or get contaminated from the gear oil in the
transmission, and it's covered in gear oil since it spins in the
transmission, so you could not smell anything from it when it's working. The
auto trans uses a computer controlled clutch pack to give drive to the rear
wheels, and the torque split is variable depending on what the computer
detects the front wheels are doing.
Ed B

 

[Jim Stewart]

Rob

There definitely are plates in the VC, I have one I cut an opening in to
show people what the inside parts look like. The fluid fills ripples stamped
out of the disks to give drive when it thickens. And regarding Jim's post
that his wife smells the VC locking, the fluid is sealed in the VC, so it
doesn't comtaminate or get contaminated from the gear oil in the
transmission, and it's covered in gear oil since it spins in the
transmission, so you could not smell anything from it when it's working. The
auto trans uses a computer controlled clutch pack to give drive to the rear
wheels, and the torque split is variable depending on what the computer
detects the front wheels are doing.
Ed B

I didn't mean to imply that she smelled anything
coming from the inside of the housing. I'm
sure she could smell a generalized "hot oil"
smell if the case was heating up though.

She can smell WD-40 on me hours later if I've
used it for a cutting fluid so I don't doubt
her veracity.

 

[CompUser]

Rob

There definitely are plates in the VC, I have one I cut an opening in to
show people what the inside parts look like.

Got the pics posted anywhere?

 

[donald girod]

I read all the responses, and I still don't understand how there can be
any torque transfer to the rear wheels on dry pavement driving in a
straight line, or indeed any torque transfer at all unless the front
wheels are turning faster than the rear wheels. I think the front
differential is hard-coupled to the engine (thru the clutch) and is not
driven thru the viscous coupling; the rear wheels are driven thru the
coupling. It is impossible for the front wheels to turn at a speed
different from what is determined by the engine speed and gearing. But
the rear wheels can turn slower (or faster, in a turn) than the speed
determined by the engine and gears. If the front wheels are turning
faster than the rear wheels (due to slippage), then the torque gets
divided 50-50 front and rear, but if everything is turning the same
speed, then the rear wheels are turned by the pavement going by under
them as the front wheels pull the car forward, there is no relative
motion in between the disks in the viscous coupling, and no torque is
transmitted.

If this is wrong, please explain how. I'm not insisting that I am
right,but I just don't understand. To say there is a "center
differential" is not really correct in the sense of a standard
differential with spider gears etc. If there were a differential like
the one on the front or rear axle, then the front and rear wheels could
turn at any speed whatsoever relative to each other, and all torque
would transfer to the slipping wheel as happens left/right in a normal
2-wheel differential (not limited-slip). Clearly this is undesirable
for traction. The viscous coupling, on the other hand, transfers
torque to the rear wheels that are not slipping (yet), but ONLY if a
front wheel is slipping.

 

[Jim Stewart]

I read all the responses, and I still don't understand how there can be
any torque transfer to the rear wheels on dry pavement driving in a
straight line, or indeed any torque transfer at all unless the front
wheels are turning faster than the rear wheels. I think the front
differential is hard-coupled to the engine (thru the clutch) and is not
driven thru the viscous coupling;

You are wrong. Simple as that. The center dif
input is hard-coupled to the engine.


the rear wheels are driven thru the

coupling. It is impossible for the front wheels to turn at a speed
different from what is determined by the engine speed and gearing. But
the rear wheels can turn slower (or faster, in a turn) than the speed
determined by the engine and gears. If the front wheels are turning
faster than the rear wheels (due to slippage), then the torque gets
divided 50-50 front and rear, but if everything is turning the same
speed, then the rear wheels are turned by the pavement going by under
them as the front wheels pull the car forward, there is no relative
motion in between the disks in the viscous coupling, and no torque is
transmitted.

If this is wrong, please explain how. I'm not insisting that I am
right,but I just don't understand. To say there is a "center
differential" is not really correct in the sense of a standard
differential with spider gears etc.

Yes it is. It's a dif just like on a rear-wheel
drive car.

If there were a differential like

the one on the front or rear axle, then the front and rear wheels could
turn at any speed whatsoever relative to each other, and all torque
would transfer to the slipping wheel as happens left/right in a normal
2-wheel differential (not limited-slip). Clearly this is undesirable
for traction. The viscous coupling, on the other hand, transfers
torque to the rear wheels that are not slipping (yet), but ONLY if a
front wheel is slipping.

Let's go back to first principles.

For the sake of discussion, think of a rear wheel
drive car. You have a dif with power in and 2
wheels out. Spyder gears and all. The car is going
straight, what's the torque split between the wheels.
50/50 right?

Now take a front wheel drive car. Put power into
the dif and you get 50/50 again, right?

Ok, feed the front dif and the rear dif with the
outputs of a center dif. 50/50 front and rear,
25% to each wheel. We haven't broken any rules
and we haven't said a word about the VC. That's
the way your Subie goes down a straight and dry
road. Exactly like that. You don't have to make
it any more complicated and I have *not* over-
simplified it.

Now, assume one or both wheels on the same dif
start to spin. The VC in the center dif heats
up and starts to hard couple the front dif drive
to the rear dif drive. Since now they both
*have* to turn at the same speed, most of the
torque goes to the unspinning end of the car.

That's all there is to it.

A minor correction to my earlier post. After
doing some research, it appears that a properly
operating VC does not ever 'lock', as in creating
a solid linkage between the two output shafts
of the center dir. It appears to be more like
a slipping clutch, causing torque to be transmitted
along with some slippage. This does not affect
my explaination in any significant way.

 

[Mike]

Ok... that makes good sense vis-a-vis front/rear slip.... What about
right/left? Is there a lateral diff. on the car so that all that new
traction is evenly split or sent specifically to the wheel that needs it the
most?

 

[Jim Stewart]

Ok... that makes good sense vis-a-vis front/rear slip.... What about
right/left? Is there a lateral diff. on the car so that all that new
traction is evenly split or sent specifically to the wheel that needs it the
most?

I'm not sure I understand the question.

Front and rear difs are not limited slip. Spin
a front wheel and a rear wheel at the same time
and you're outa luck. Doesn't seem to be a big
problem in practice.

 

[Ross]

I read all the responses, and I still don't understand how there can be any
torque transfer to the rear wheels on dry pavement driving in a straight
line, or indeed any torque transfer at all unless the front wheels are
turning faster than the rear wheels. I think the front differential is
hard-coupled to the engine (thru the clutch) and is not driven thru the
viscous coupling; the rear wheels are driven thru the coupling. It is
impossible for the front wheels to turn at a speed different from what is
determined by the engine speed and gearing. But the rear wheels can turn
slower (or faster, in a turn) than the speed determined by the engine and
gears. If the front wheels are turning faster than the rear wheels (due to
slippage), then the torque gets divided 50-50 front and rear, but if
everything is turning the same speed, then the rear wheels are turned by
the pavement going by under them as the front wheels pull the car forward,
there is no relative motion in between the disks in the viscous coupling,
and no torque is transmitted.

If this is wrong, please explain how. I'm not insisting that I am
right,but I just don't understand. To say there is a "center
differential" is not really correct in the sense of a standard
differential with spider gears etc. If there were a differential like the
one on the front or rear axle, then the front and rear wheels could turn
at any speed whatsoever relative to each other, and all torque would
transfer to the slipping wheel as happens left/right in a normal 2-wheel
differential (not limited-slip). Clearly this is undesirable for
traction. The viscous coupling, on the other hand, transfers torque to
the rear wheels that are not slipping (yet), but ONLY if a front wheel is
slipping.

The center diff in manual Subarus is in essence exactly the same as and open
diff as found in most cars *except* that the two drive shafts are connected
by a viscous coupling. A viscous compling on its own would serve no purpose
what so ever. it needs to be combined with the open diff to work.

An open diff applies exactly the same amount of torque to each wheel, If
the front wheels slip and can only recieve 1Nm of torque this is how much
the rear wheels will get regardless of how much traction they can take. In
this scenario the two drive shafts have a significantly different rotational
velocity. Now imagine that the diff suddenly becomes a VC Center diff
(essentially an open diff with the two drive shafts connected by a VC) What
the VC does is try to make the two shafts spin at close to the same speed.
The front wheels' drive shaft is spinning faster than the rear, this causes
the vanes or plates to stir the thick fluid in the VC and it begins to flow
round in the housing, this flow then drags the the vanes or plates for the
other drive shaft round with it thus applying more torque to the rear
wheels. One way to imagine how the VC works is to stir a cake mixture in a
bowl. you need to hold the bowl to stop it turning with the spoon. this is
a viscous coupling in a very simpe sense.

An explanation on how they work can be found here...
http://auto.howstuffworks.com/differential9.htm

 

[Rob Munach]

Rob

There definitely are plates in the VC, I have one I cut an opening in to
show people what the inside parts look like. The fluid fills ripples stamped
out of the disks to give drive when it thickens. And regarding Jim's post
that his wife smells the VC locking, the fluid is sealed in the VC, so it
doesn't comtaminate or get contaminated from the gear oil in the
transmission, and it's covered in gear oil since it spins in the
transmission, so you could not smell anything from it when it's working. The
auto trans uses a computer controlled clutch pack to give drive to the rear
wheels, and the torque split is variable depending on what the computer
detects the front wheels are doing.
Ed B

Thanks Ed. I guess what I meant to say is that there are no plates that
get forced together when the fluid heats up.

-rm

 

[Rob Munach]

The center diff in manual Subarus is in essence exactly the same as and open
diff as found in most cars *except* that the two drive shafts are connected
by a viscous coupling. A viscous compling on its own would serve no purpose
what so ever. it needs to be combined with the open diff to work.

Actually, this is not true. A viscous coupling w/o a center diff will
lock the front and rear together whenever there is slip. This is
basically what the clutch pack does in the auto box except the
technology is different. I'd prefer to have the VC w/ center diff,
however, many *AWD* vehicles nowadays do not use a center diff.

-rm

 

[donald girod]

Ok, I accept that. You are saying that what we basically have in the
transfer case is a limited-slip center differential, similar to what one
might find in a limited slip rear axle. In the absence of the VC in the
mix, a car with a center diff. would never experience any front-rear
windup or binding, because the front and rear axles are free to turn at
whatever independent speeds they want, just as right and left wheels on
an axle with a differential are free to turn. However, without the VC,
torque would be limited by the axle with the least traction, while with
the VC, torque is limited (more or less) by the axle with the most
traction.

And, with the VC, any difference in rotation between the front and rear
outputs is "abnormal" in the sense that it will cause relative rotation
in the VC, and drag and heating and so on. Very slow and/or brief
relative rotation, as in cornering, causes little drag and little
heating. Rapid rotation (as in wheelspin) or long-term rotation (as in
mismatched tires) will cause drag and heating etc., and in the case of
tire mismatch, possibly long-term heating and damage to the VC.

If the VC lost some of its ability to transfer torque (wear, damage,
whatever), the front and rear wheels would still both be driven but you
would lose some of the advantage of transferring torque to the axle with
traction.

And according to some other posts, some other cars have the VC without a
center differential; this was the impression I had about the Subaru.

Now do I have it right?

 

[Ross]

The center diff in manual Subarus is in essence exactly the same as and

Actually, this is not true. A viscous coupling w/o a center diff will lock
the front and rear together whenever there is slip. This is basically what
the clutch pack does in the auto box except the technology is different.
I'd prefer to have the VC w/ center diff, however, many *AWD* vehicles
nowadays do not use a center diff.

If this is the case then you'd have a FWD car until you had slip this is not
the case on modern manual Subarus. The rear wheels would be rolling stock.
So on a loose surface with all 4 corners having equal grip if you dumped the
clutch only the front wheels would spin getting 100% of the torque until the
VC has enough difference in rotational velocity to transfer torque to rear
and even then the front wheels would need to contunue to spin faster than
the rears to maintain torque to the rears (More difference = more torque
transfer), a very messy situation which would produce insane amounts of
understeer. Subaru AWD is symetrical and this is not what happens.

I know from fact that my STi Type UK will spin all 4 wheels instantly and at
the same rate on a hard launch from loose surface suggesting power being
supplied to the transmission via an open diff with a VC (No DCCD fitted).

If you drove the housing of the VC from the engine then your VC would last
about an hour and your fuel economy would be even worse than it already
is!!!

Ross

 

[Jim Stewart]

Ok, I accept that. You are saying that what we basically have in the
transfer case is a limited-slip center differential, similar to what one
might find in a limited slip rear axle. In the absence of the VC in the
mix, a car with a center diff. would never experience any front-rear
windup or binding, because the front and rear axles are free to turn at
whatever independent speeds they want, just as right and left wheels on
an axle with a differential are free to turn. However, without the VC,
torque would be limited by the axle with the least traction, while with
the VC, torque is limited (more or less) by the axle with the most
traction.

And, with the VC, any difference in rotation between the front and rear
outputs is "abnormal" in the sense that it will cause relative rotation
in the VC, and drag and heating and so on. Very slow and/or brief
relative rotation, as in cornering, causes little drag and little
heating. Rapid rotation (as in wheelspin) or long-term rotation (as in
mismatched tires) will cause drag and heating etc., and in the case of
tire mismatch, possibly long-term heating and damage to the VC.

If the VC lost some of its ability to transfer torque (wear, damage,
whatever), the front and rear wheels would still both be driven but you
would lose some of the advantage of transferring torque to the axle with
traction.

And according to some other posts, some other cars have the VC without a
center differential; this was the impression I had about the Subaru.

Now do I have it right?

You got it.

 

[YKhan]

Actually, this is not true. A viscous coupling w/o a center diff will

lock the front and rear together whenever there is slip. This is
basically what the clutch pack does in the auto box except the
technology is different. I'd prefer to have the VC w/ center diff,
however, many *AWD* vehicles nowadays do not use a center diff.

-rm

Actually, there is a VC in both the automatic and manual versions of
Subies. However, they use the VC in completely different ways in each
case. In the case of manual tranny Subies, the VC acts as a
slip-limiter for the centre diff. In the case of auto Subies, the VC
activates a progressive clutch pack that locks the front and rears
together: there is no centre diff in an auto Subie.

In the case of a manual Subie, the VC starts causing drag when the
front and rear axels have a big enough difference. In the case of an
auto Subie, the front gets the majority of the power and the rear axel
follows along like in a standard FWD car, until it looks like the front
axel starts spinning much faster than the rear axel, then that causes
the VC to start causing drag which locks up the front and rear axels
together. In a manual, the front and rear are getting 50:50 power at
all times, even during slippage. In an auto, the front gets 90% of the
power when there's no slippage, and when there is slippage, they get
locked together at 50:50.

Yousuf Khan