How Center Diffs Work

This site may earn a commission from merchant affiliate
links, including eBay, Amazon, Skimlinks, and others.

A thought just occurred to me.

SUMO, are you working under the impression that the front and rear drive shafts in HF2AV are only coupled via the viscous coupler and are NOT physically coupled through actual side gears in the differential?

If so, that's where your misunderstanding is.

The front and rear drive shafts ARE physically coupled by gears in the differential. You can completely remove your viscous coupler from the transfer case and still have torque transferred to both drive shafts. This is what the HF2A is. The viscous coupler just sits on top of this connection.

The viscous coupler in a Subaru-like system IS the differential--and (being viscous) is limited in its ability to transfer torque--hence the torque bias ratio other than 1:1.

The viscous coupler in the HF2AV is never the link through which torque is transmitted from axle to axle. It is precisely as I have described it:

It is a device that dampens the differential motion between the front and rear drive shafts.
It's a brake of sorts.

When you realize that the HF2AV is a straight-forward geared open differential IN EVERY WAY, you'll see that its TBR is only and always 1:1.


From the full service manual, here's a diagram of the transfer case:

HF2AV%20diagram.jpg



A description of the center diff:

HF2AV%20cdl.jpg



And finally, a description of the viscous coupler, and its role in the transfer case:

HF2AV%20vc%20general.jpg



I sincerely hope this clears up your confusion.

Hayes
 
Last edited:
The purpose of this thread is damage control!

It won't work, unless we rename the title to, "How Center Diffs IN AUDIS Work"! Now quick, someone save this thing before it is deleted!!! Hayes, I love your work, you are a gentleman and a scholar! :cheers: :cheers: :cheers:
 
I am a little lost on to what the point here is also. Is anything broken? Can it be improved? The rest of the world want to know why we are having this conversation.

Nothing is broken carry on.:D

The purpose of this thread is damage control!

Nice thought but unfortunately to late.:whoops:

It won't work, unless we rename the title to, "How Center Diffs IN AUDIS Work"! Now quick, someone save this thing before it is deleted!!! Hayes, I love your work, you are a gentleman and a scholar! :cheers: :cheers: :cheers:

In this some truth may lie.

On another note can anybody help me with my flaring problem on my 93?

Oh yaa nevermind we already covered that.:doh:
 
A thought just occurred to me.

SUMO, are you working under the impression that the front and rear drive shafts in HF2AV are only coupled via the viscous coupler and are NOT physically coupled through actual side gears in the differential?

If so, that's where your misunderstanding is.

The front and rear drive shafts ARE physically coupled by gears in the differential. You can completely remove your viscous coupler from the transfer case and still have torque transferred to both drive shafts. This is what the HF2A is. The viscous coupler just sits on top of this connection.

Ryan
The F/R axles are coupled via the center diff AND the VC. You can remove the VC OR the locker mechanism, and the truck still have torque transferred to both driveshafts, thru the open diff 50/50.

Once locked, the center diff does not allow differentiation. Once semi locked, the cener diff does not allow differentiation (ok 20%). You can lock the axles together wtih a dog clutch (over the center diff) or lock the axles together with a VC (over the center diff). It's the same principle, think of the VC as a slipping locker....

How the rear axle mounted VLSD differs from the EDL, is that rear EDL *can't lock over the center diff*! IOW, the open diff character of the vehicle is the same, no increase in understeer, no increase in steering effort, no change in torque allocation. You are just raising the amount of torque supported by both axles.

When you lock or semi lock axles together, it's 'over the center diff' which means that the center diff can't apply torque to both axles equally unless the weight distribution is 50/50.

Ryan, I have a good pic of the VLSD system power transmission in a pic from Toyota that might help you. The operation of a locked center diff or a VC LSD locked diff is the same. If you could make a 100% locking VC (humping) it is the same as a locked center diff. That doesn't at all mean at 99% or 80% it's the same as EDL.

In Toyota's description of the VLSD, along with a pic of the VC "over" the center diff... "In an idealized center viscous coupling, power is taken off front and rear driveshafts on either side of an open center differential. When a speed difference "locks" the viscous coupling, the front and rear shafts are tied together"

It's locking axles, not generating more torque at the lower supported axle. Quite the opposite. Using your idea, in the equation above, you need to generate 50lb/ft of torque in the front axle. That's not what VLSD's do, that's what brakes do (EDL).

ST
 
It is not locking axles..

It is locking axles together..



I still see this as an argument in semantics.



A viscous coupler never "locks" the front axle to the rear axle. Any slip means it is not locked up.

No matter what, if you only have traction on one tire, you won't go anywhere...
 
It is not locking axles..

It is locking axles together..

I still see this as an argument in semantics.

A viscous coupler never "locks" the front axle to the rear axle. Any slip means it is not locked up.

No matter what, if you only have traction on one tire, you won't go anywhere...

It's more physics really. If you believe that the VLSD only raises torque of the axle with the least amount of supported torque, then the TBR is 1:1. If you believe that the VLSD allows transfer of torque from the axle with the least amount of supported torque to the axle with the most amount of supported torque, then the TBR is above 1:1.

The difference in how that physics applies to Total Supported Torque to Both Axles is significant (see post 31).

The HF2AV allows up to a 80/20 to 20/80 F/R axle torque allocation when VLSD engaged/slipping. When Locked (or VC 'humping') the HF2AV allows up to 100/0 to 0/100 F/R axle torque allocation. When not slipping or not locked the open diff allocates torque 50/50 always. This defines a 4:1 TBR VLSD as applied to the 80. And just like all LSD's, 1 drive wheel with no supported torque (0) x (any)TBR = 0

EDL can't 'transfer' any torque to an apposing axle, it can only raise Total Supported Torque to Both Axles. As defines EDL operation F/R torque allocation will be
(1 Drive wheel with no supported torque + 1 drive wheel with brake torque) = 2 x torque of drive wheel with brake torque.

LSD's allow transfer of axle torque to an apposing axle as a ratio of
TBR = Highest supported axle torque / lowest supported axle torque

ST
 
Last edited:
Suprisingly everything mechanical has a physics component to it.



But you ignored every word I said.

There is no mechanical "locking" of a tcase if anything viscous is stuck in there.

If you put enough traction on one tire and none on the other 3 you will simply spin away..

I never brought up the concept of a computer controling the brakes on the axles. Just the fact that you keep referring to the transfercase locking up the axles..
 
Alright.
I have honestly and actively been questioning my assumptions, but I still have some problems with yours.


IN AN OPEN DIFF Trg is exactly allocated 50% T1 and 50% T2 ALWAYS
------------------------------
ADD VISCOUS COUPLER Torque Bias Ratio
Let's add the HF2AV Viscous Coupler to the center diff, and see what happens.... From my research the VC is considered a 80% locker (Humping it will go 100%, but that's a short term failure mode we will ignore for now). What this means is that 4 times the amount of torque supported by the axle with the least supported torque can be supported by the axle with the higher supported torque. The driveline boys refer to this as a Torque Bias Ratio (TBR) which is Defined as maximum /minimum supported torque of a given axle (80/20 = 4:1 TBR)ST


The first problem is with the following statement about the viscous coupler:

"What this means is that 4 times the amount of torque supported by the axle with the least supported torque can be supported by the axle with the higher supported torque. "

I don't see how any torque transferred through the viscous coupler has anything to do with the quantity of the least supported torque at an axle. The viscous coupler acts through rotational speed differentials created by torque directly on the VC itself. The least supported torque at either axle could vary widely due to traction conditions, but the amount of torque transferred via the VC depends only upon the rotational speed differences.

Example:

1. Tire spinning at 500 rpm in the air. Little to no supported torque.

2. Tire spinning at 500 rpm in thick mud. More supported torque than the previous scenario.


In both cases the torque transferred through the VC will be the same, but the minimum supported torque will be different.

I think this may be at the root of what I believe is your misapplication of TBR principles to the HF2AV.


The second problem I see is where you state:

"IN AN OPEN DIFF Trg is exactly allocated 50% T1 and 50% T2 ALWAYS

I have yet to understand how adding a VC on top of an open diff gives you the ability to contradict your statement.

Although I can see the desire to describe it as such, I don't see how a VC on top of an open diff is truly a "partial locker." In a truly locked differential, there is no differential, it's effectively a single axle. Linking the axles with a viscous coupler over the differential just introduces more friction to the system--the OPEN DIFF (see above) is still there.

Unless you can show otherwise, your original statement holds true, an OPEN DIFF always allocates torque 50/50, ALWAYS.

This is my central assumption.
I will humbly admit my error if you can show it to me.

If my assumption is correct, then my description of the viscous coupler raising the minimum supported torque is the correct one.


Here is a quick and crude diagram of how I see torque transfer in the HF2AV. No, it's not a perfect free-body diagram, I have no CAD, but hopefully it will get my idea across, and possibly even allow you to point out my error.

Assume all the numbers represent generic units of torque, and the final output from the transmission is able to supply a maximum torque of 100.


The first drawing shows an open diff with one axle supporting very little torque (5). The two outer arrows represent the maximum torque supported at each axle. The rest should be clear.

open.gif



The second drawing shows my conception of how the viscous coupler acts on the center differential. The VC requires (supports) a certain amount of torque to rotate--and this varies with the rotational speed differential--but I have simplified the torque required to rotate the VC as "20."

As above, the "-5" represents the torque required to spin the right-side axle (supported torque) and the "-20" represents torque required to rotate the VC. Together, it requires a torque of 25 to rotate the right side axle--this is the supported torque on the right axle.

This torque is transferred through the open diff (the output from the trany goes to the diff, not the VC!) which ALWAYS SPLITS THE TORQUE 50/50. Therefore the torque transferred to the left-side axle is also 25!

vc.gif



I realize that the few masochistic MUDders still reading this thread may think this is all wasted forum space, but at this point I either need confirmation or refutation--I'm almost to the point where I don't care which--and so far I haven't seen any explanation that gets me either.
So thanks for all your patience and I hope you're enjoying the show. :D :popcorn:

Hayes
 
Personally, I wish there was more boobies but hell. You take what you can get :D


I still want to know too :D
 
If I wasn't lazy I would peruse the UF online databases again for extensive information on differentials, etc...

Unfortunately I am lazy :flipoff2:
 
Suprisingly everything mechanical has a physics component to it.

But you ignored every word I said.

Don't let it bother you mace. Eventually you get used to it.

OK damage control was probably not the best label, how about containment?

Yaa now if only it was working.
 
The purpose of this thread is damage control!


As you can see from my post count, I read and listen far more than post. I did that for about a year before even registering....

Not sure what you mean by "damage control", but let me say this for whatever it is worth.

I remember SumoToy from an Audi Quattro board we both frequented for several years. Have never met the guy, but have learned a good bit from him over the years.

Find exception with his blunt techie-talk if you want, but if you can get past giving grades for presentation you can learn something from him...assuming the goal is to learn. The guy has forgotten more about the internal workings of the Audi system and turbo stuff than most folks I have met who work on the stuff for a living know...and knowing what a tech-nerd he is...I am banking on the fact that since he owns and wrenches on LC's, that system is well understood too.


Bruce...back to lurking:)
 
As you can see from my post count, I read and listen far more than post. I did that for about a year before even registering....

Not sure what you mean by "damage control", but let me say this for whatever it is worth.

I remember SumoToy from an Audi Quattro board we both frequented for several years. Have never met the guy, but have learned a good bit from him over the years.

Find exception with his blunt techie-talk if you want, but if you can get past giving grades for presentation you can learn something from him...assuming the goal is to learn. The guy has forgotten more about the internal workings of the Audi system and turbo stuff than most folks I have met who work on the stuff for a living know...and knowing what a tech-nerd he is...I am banking on the fact that since he owns and wrenches on LC's, that system is well understood too.


Bruce...back to lurking:)

I guess one's perspective on his wealth of knowledge would reflect their own on the subject.
 
I realize that the few masochistic MUDders still reading this thread may think this is all wasted forum space, but at this point I either need confirmation or refutation--I'm almost to the point where I don't care which--and so far I haven't seen any explanation that gets me either.
So thanks for all your patience and I hope you're enjoying the show. :D :popcorn:

Hayes

I must be masochistic for posting to this thread. What I was having trouble with was where the front and rear outputs where actually both connected to the VC. I knew where the rear was connected from doing the rear output seal replacement, but couldn't rember good enough - fortunately one of your photos showed me (I think).

I modified your dwg a bit. Took the VC pic. from here

http://members18.clubphoto.com/steven666314/1144038/guest.phtml

The rear drive slips, it spins faster than the front, this shering action shering action between the two plates heating up the silicon fluid, and the power starts being transmitted between the plates with the fluid as a medium. The 80/20 split comes from the limits of the fluid's ability to transfer, and of course, any transfer is going to have some losses. Does this help? Or make it worse?
Output.webp
vc.webp
 
Last edited:
Alright.
I have honestly and actively been questioning my assumptions, but I still have some problems with yours.



vc.gif



I realize that the few masochistic MUDders still reading this thread may think this is all wasted forum space, but at this point I either need confirmation or refutation--I'm almost to the point where I don't care which--and so far I haven't seen any explanation that gets me either.
So thanks for all your patience and I hope you're enjoying the show. :D :popcorn:

Hayes

Just when I was going to fire up Visio. Hayes your second drawing is good (VC coupling over the center axle), the torque allocation isn't. Hayes, what needs to be corrected in your drawing is the torque allocation will follow the highest resistance up to the torque bias ratio. IOW, the torque supported by axle two will be 4 times the torque supported by axle one. In your example, the VC will couple the axles together, so torque is transfered to the other axle at (4TBR x 5lb/ft) = 20lb/ft. Total supported torque by both axles in your example is 25lb/ft.

You have shown a LSD generating friction (more torque load on the faster spinning axle). It should be thought as a coupler, so it's more torque load on the SLOWER spinning axle. I'm not sure how in your example you could draw a humping VLSD, which has locked both axles over the center diff, allowing 100% Trg to be supported by either axle. If it's not fully locked at 80% (4:1 TBR), up to 80% of the Trg can be supported by either axle.

In a locked diff over the center diff, the TBR = 'up to' 100% x Trg. So again in your example, 5lb/ft will be supproted by axle 1 and 45lb/ft will be supported by axle 2. Total Supported Torque by Both Axles is 50lb/ft (=Trg)

Maybe it would help if you drew a locker over the center diff, or a torsen over the center diff. I think you are not grasping how a center diff works under locked or LSD conditions.

In EDL, you don't have any locking/coupling of the axles over the center diff, *SO NO TORQUE TRANSFER CAN TAKE PLACE BETWEEN AXLES* the diff allocates the torque 50/50, axle 1 will have 5lb/ft of torque (traction) and add, lets say 5lb/ft of torque from braking. So axle 1 has 10ft/lbs of supported torque at one side of the center diff. By definition then, axle 2 has 10lbs of supported torque at the other side of the center diff and Total Torque to Both Axles = 20lb/ft

In the VLSD the torque is transfered from one axle to the other thru the coupled axles at a 4:1 TBR. In EDL NO torque can be transfered since the axles are only coupled thru the center diff, defining a TBR 1:1

HTH

ST
 
Back
Top Bottom