How Center Diffs Work

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

Heath this was afair request on your part. I was cc'd on that PM and he made a mod based on a later PM from me, so you have given him a reasonable amount of time. I will do it for you and then he can come back later and rephrase his posts with out your qoutes.
 
I understand that the VC in an FZJ80 is a limited slip device, and not really a coupler--that is precisely my question. When the maximum torque supported by the ring gear is limited by low traction at one of the axles, the differential motion between the front and rear drive shafts cause the VC to generate friction on the faster-spinning shaft (braking it), raising the maximum supported torque on that axle--as well as the other axle (50/50 split). Isn't this essentially the same thing that happens in electronic traction control systems where the brakes are applied to the spinning wheel?
I don't quite understand where the 80/20 split comes in. I realize that there is a limit to the amount of braking force the VC can apply (not counting "humping"), but I don't understand how the distribution of torque could be anything other than 50/50.

Hayes
 
sumo. when you get around to it, please fill the blanks on your opening posts with "acceptable" paraphaseing. its like im catching half a phone conversation, so, i dont know what you are disproving, to know if i agree. call me edith bunker if that doesnt make any sence.
 
Ummm...


It work because Mr. T say it do.


I poosh de buton an de lite com on, she all locked up. I poosh it agin and she not locked up, but de front stil pull me out de slide. work jus fine ifn y'ask me.
 
I understand that the VC in an FZJ80 is a limited slip device, and not really a coupler--that is precisely my question. When the maximum torque supported by the ring gear is limited by low traction at one of the axles, the differential motion between the front and rear drive shafts cause the VC to generate friction on the faster-spinning shaft (braking it), raising the maximum supported torque on that axle--as well as the other axle (50/50 split). Isn't this essentially the same thing that happens in electronic traction control systems where the brakes are applied to the spinning wheel?
I don't quite understand where the 80/20 split comes in. I realize that there is a limit to the amount of braking force the VC can apply (not counting "humping"), but I don't understand how the distribution of torque could be anything other than 50/50.

Hayes

Distribution of axle torque in an open diff is always 50/50. Distribution of torque in a locked diff can be 100/0. In a locked center diff in a straight line, F/R axle torque distribution follows exatly weight distribution up to wheel spin. These are all knowns.

In LSD land, we refer something less than 100% locked as Torque Bias Ratio. Torque Bias Ratio is a ratio of how much torque each axle can support, based on the lowest axle torque.

Hayes, I'm not sure the confusion. If what you say is true, then a locked center diff with 2 front wheels in the air, can't deliver 100% of Teng (*gear multiplier) to the 2 rear wheels. Maybe we all should make a correction and call the HF2AV a Viscous Limited Slip Differential (VLSD) instead of Viscous Coupler (VC). Technically, it does "couple", but it's coupling two already coupled axles.

HTH

ST
 
Last edited:
Distribution of axle torque in an open diff is always 50/50. Distribution of torque in a locked diff can be 100/0. In a locked center diff in a straight line, F/R axle torque distribution follows exatly weight distribution up to wheel spin. These are all knowns.

In LSD land, we refer something less than 100% locked as Torque Bias Ratio. Torque Bias Ratio is a ratio of how much torque each axle can support, based on the lowest axle torque.

Hayes, I'm not sure the confusion. If what you say is true, then a locked center diff with 2 front wheels in the air, can't deliver 100% of Teng (*gear multiplier) to the 2 rear wheels.

HTH

ST


SUMO,
I thought through exactly what you are saying above yesterday while trying to check my confusion.

It seems to me that comparing a VC-damped differential like the HF2AV (or an ATRAC-like system where the slip is damped by computer-controlled braking) to a locked diff is erroneous.

A locked diff is not a diff. Hence the ideal torque distribution up to 100/0.

1. When a locked diff sees one wheel in the air (zero traction, zero resistance, zero supportable torque), NO torque is sent to that axle.

2. When an open diff with traction control (VC or ATRAC) sees one wheel in the air (zero traction, zero resistance, zero supportable torque), the diff encounters no resistance to rotation and splits that zero 50/50--no go. The VC merely adds resistance (through friction) to the rotation of the spinning axle, effectively "fooling" the diff into behaving like the the wheel in the air is on the ground. Now there is actually some torque at the diff. The differential does EXACTLY the same thing diffs always do--they split that torque 50/50.

***The difference between #1 and#2 above is that in the case of the locked diff, the tire in the air is still free and clear with no resistance to rotation--no supportable torque, and no torque expended by the tranny trying to turn it.
With an open diff and VC, the tire in the air IS supporting torque through the friction supplied by the VC--the tranny IS expending some (exactly 50%) of the total torque against this friction. Same thing with ATRAC braking. Half of the torque coming down from the tranny is pushing against the VC/brake.

To see how comparing a locked diff and an open diff with some kind of damped traction control (VC or ATRAC) is wrong, think of how the torque would be distributed if the VC was providing enough friction to completely keep the traction-less wheel from spinning at all--the brakes are applied fully on the spinning wheel. This is not the same as a locked diff. The tranny HAS to torque against the braked wheel because of the open diff.


I realize that LSD's are discussed in terms of "torque bias ratio." But the center diff in the HF2AV is not a LSD. It is an open diff with a device that increases supportable torque AFTER that open diff.

I realize that I may be totally confused, but I have yet to see how the torque in an HF2AV is biased any other way than 50/50.

Ryan
 
SUMO,
I thought through exactly what you are saying above yesterday while trying to check my confusion.
It seems to me that comparing a VC-damped differential like the HF2AV (or an ATRAC-like system where the slip is damped by computer-controlled braking) to a locked diff is erroneous.
A locked diff is not a diff. Hence the ideal torque distribution up to 100/0.

Ryan,
A locked diff has a torque bias ratio as well, it's infinity. You still have a ring gear and two axles with supported torque. I think you are trying to speak to wheel speed differences, that's not a component of alxe torque necessarily.

I realize that LSD's are discussed in terms of "torque bias ratio." But the center diff in the HF2AV is not a LSD. It is an open diff with a device that increases supportable torque AFTER that open diff.

The VC in the 80 is absolutely a LSD! Most LSD operate AFTER an open diff, or it's a coupler Ryan. To really confsue you, try looking at how a torsen LSD works. It is also AFTER the diff and allocates torque like a locked diff up to it's max Torque Bias Ratio.

I realize that I may be totally confused, but I have yet to see how the torque in an HF2AV is biased any other way than 50/50.
Ryan

A LSD is defined quite clearly in terms of Torque Bias Ratio Ryan. You might try searching TBR or try the Torsen website, they put out a lof of information on TBR.

We need to compare supported axle torque between different devices, from locked to VLSD, to clutchpack, to Torsen, to open. Torque Bias Ratio does that very well. The accepted TBR of an open diff is 1:1. Other LSD devices can be compared quite easily. However, we should note that few EDL systems are 'over the center axle', meaning exactly that EDL rarely defines the center axle, usually only front and/or rear axles.

Sorry I can't be more help in your confusion. Ryan, the first read I would recommend is the "Torsen White Paper" presented by it's primary engineer Chocholek. He does a good job of definiing TBR and different differentials operation in terms of torque. It's hardly too nerdy, hope it helps
http://zhome.com/ZCMnL/tech/Torsen/Torsen.htm

For more (and Chocholek's paper can be found here as well)
www.torsen.com



ST
 
Last edited:
Example in plain English...

Driving up steep hill in snow covered grass (just tried this)-High Range, wheel slip as soon as back tires enter slope. No forward movement without ripping up the grass.

Put into Low Range, up to the top of the hill with limited slipping, and very little "Go" pedal, then start to spin, no traction.

Push in the Center Diff Lock, forward movement in L-Range with absolutely no wheel spin, over the hill I go. Works very well.
 
I don’t think that reading a technical write-up on Torsen diffs and torque bias is necessary to discuss the distribution of torque in a straight-forward open diff backed up with what amounts to be a sort of viscous damper. I think you misunderstand my point. I really think that upon close examination, you can’t treat the VC in the HF2AV analogously to a locked diff with varying degrees of “locked” (as you suggest in post #26). Understanding this will demonstrate my point that the HF2AV should ALWAYS split torque 50/50 between the front and rear drive shafts—never 80/20 or any other ratio. Bear with me and address the following questions/descriptions:


1. By itself, a classic differential gear with ring, pinion, and spider gears splits the incoming torque evenly (50/50) between the two output shafts/axles.

2. A locked diff isn’t a “diff.” The torque ratio is always ideal, up to 100% to one shaft because the two output shafts are effectively one shaft driven by the pinion.

3. The “viscous coupler” in the HF2AV doesn’t actually couple anything—it just creates friction when there is differential motion between the front and rear drive shafts. “Viscous damper” is a more accurate description.

Now consider these three scenarios:

1. An open diff with no viscous damper. One output shaft effectively cannot support any torque because there is no traction at the wheels on that shaft (floating in the air or on a surface with zero mu, ignoring torque supported by rotational inertia and any friction inherent in the drive system). The net torque delivered to the two output shafts is effectively zero.

2. A locked diff under the same circumstances. Saying that the output shaft with zero traction cannot support any torque is not really accurate because that shaft is locked to the—and effectively the same—shaft that has traction and can support 100% of the available torque. The tranny is applying no force to the shaft with no traction. Hence the 100/0 distribution of torque between the two shafts.

3. An open diff with a viscous damper (HF2AV) under the same circumstances. The differential wants to behave just like scenario #1. Fortunate for us, the VC provides resistance—a force that will support torque—to the axle with zero traction. Lets say the VC requires z ftlbs of torque to rotate—the axle can now support z ftlbs of torque (as opposed to zero). The other axle could support an infinite amount of torque, but because the available torque is transferred through the open diff, the total torque is limited to 2z...z ftlbs to one axle z ftlbs to the other.

3A. Scenario #3 is not the same as a “partially locked” diff. Assume the VC is “humped” and is effectively supporting infinite torque on that axle—the axle is totally “braked” and will not turn. The tranny HAS to apply force to that axle. It ALSO has to apply force to the axle with traction. It’s not applying zero torque to the “humped” axle and 100% to the other, it’s applying 50% to the humped side and 50% to the other side. Reconsider scenario #1. There is no force resisting rotation on the “tractionless” axle.


The VC merely increases the supportable torque of the axle with the least supportable torque. It’s increasing the total torque that has to be split 50/50 through the diff.

50/50.

TBR of an open diff is 1:1. You said it yourself.

Hayes
 
I don’t think that reading a technical write-up on Torsen diffs and torque bias is necessary to discuss the distribution of torque in a straight-forward open diff backed up with what amounts to be a sort of viscous damper. I think you misunderstand my point. I really think that upon close examination, you can’t treat the VC in the HF2AV analogously to a locked diff with varying degrees of “locked” (as you suggest in post #26). Understanding this will demonstrate my point that the HF2AV should ALWAYS split torque 50/50 between the front and rear drive shafts—never 80/20 or any other ratio. Bear with me and address the following questions/descriptions:

OK


1. By itself, a classic differential gear with ring, pinion, and spider gears splits the incoming torque evenly (50/50) between the two output shafts/axles.
Defines an "open" differential

2. A locked diff isn’t a “diff.” The torque ratio is always ideal, up to 100% to one shaft because the two output shafts are effectively one shaft driven by the pinion.

Ideal only in a straight line. Can be much less than Ideal in a turn. I agree a locked diff and no diff are the same thing.

3. The “viscous coupler” in the HF2AV doesn’t actually couple anything—it just creates friction when there is differential motion between the front and rear drive shafts. “Viscous damper” is a more accurate description.

Actually, it does couple the axles, beyond 50/50, that defines a LSD device, and is also attributed a Torque Bias Ratio. ALL LSD's have a Torque Bias Ratio, it defines the limit of locking, even Quaifes and Torsen gear type LSD's. A plate type LSD is a viscous damper too, so is a electronic plate type. All have TBR's.

Now consider these three scenarios:

1. An open diff with no viscous damper. One output shaft effectively cannot support any torque because there is no traction at the wheels on that shaft (floating in the air or on a surface with zero mu, ignoring torque supported by rotational inertia and any friction inherent in the drive system). The net torque delivered to the two output shafts is effectively zero.

Open Differential
TBR = 1
1 x 0 = 0% of least axle torque to other axle

2. A locked diff under the same circumstances. Saying that the output shaft with zero traction cannot support any torque is not really accurate because that shaft is locked to the—and effectively the same—shaft that has traction and can support 100% of the available torque. The tranny is applying no force to the shaft with no traction. Hence the 100/0 distribution of torque between the two shafts.

Locked or no differential
TBR = infinity or (100% x potential Torque Ring Gear)
100% of least alxe torque to other axle

3. An open diff with a viscous damper (HF2AV) under the same circumstances. The differential wants to behave just like scenario #1. Fortunate for us, the VC provides resistance—a force that will support torque—to the axle with zero traction. Lets say the VC requires z ftlbs of torque to rotate—the axle can now support z ftlbs of torque (as opposed to zero). The other axle could support an infinite amount of torque, but because the available torque is transferred through the open diff, the total torque is limited to 2z...z ftlbs to one axle z ftlbs to the other.

Limited Slip Differential (torsen, Quaife, wet plate, electronic plate, haldex, VC)
TBR = 4:1
4 x least axle torque to other axle


3A. Scenario #3 is not the same as a “partially locked” diff. Assume the VC is “humped” and is effectively supporting infinite torque on that axle—the axle is totally “braked” and will not turn. The tranny HAS to apply force to that axle. It ALSO has to apply force to the axle with traction. It’s not applying zero torque to the “humped” axle and 100% to the other, it’s applying 50% to the humped side and 50% to the other side. Reconsider scenario #1. There is no force resisting rotation on the “tractionless” axle.

Hayes, I think you are hung up on VC heat. Heat only causes the locking effect, it's not the same as TBR. What is inside the HF2A could be wet plate, a torsen or a VC. With a "80%" locker, that LSD will have a 4:1 TBR regardless of LSD type. The torsen article I referenced should help. A VC LSD is no different than any other type of LSD in terms of how it works, and it's ascribed TBR.


The VC merely increases the supportable torque of the axle with the least supportable torque. It’s increasing the total torque that has to be split 50/50 through the diff.

I don't agree with these statements. The VC merely increases the supportable torque of the axle with the *most* supportable torque, which in turn increases Total Supported Torque of both axles. The amount of torque of the higher supported torque axle is defined by the Torque Bias Ratio of the LSD device, which is a torque multiplier of the axle with the least supported torque. The second sentence defines what EDL does and that's a different animal. Let's take a look at the EDL, since that's really what you are speaking to. EDL is an an open diff with a TBR = 1. Let's say the front axle can support 50lb/ft of torque.

OPEN DIFF
TBR = 1
1(TBR) x 50lb/ft Front = 50lb/ft Rear
TOTAL SUPPORTED TORQUE OF BOTH AXLES = 100lb/ft
TORQUE SPLIT = 50lb/ft Front and 50lb/ft Rear

OPEN DIFF (EDL)
TBR = 1
Electronically apply brakes to front axle (EDL) to generate 50lb/ft of braking force so front Axle now supports 100lb/ft (50lb/ft tractive + 50lb/ft Braking torque)
1(TBR) x 100lb/ft Front = 100lb/ft Rear
TOTAL SUPPORTED TORQUE OF BOTH AXLES = 200lb/ft
TORQUE SPLIT = 100lb/ft Front and 100lb/ft Rear

VC or LSD (80% locker)
Front Axle supports 50lb/ft of torque
TBR = 4:1
Rear Axle supports 200lb/ft of torque
TOTAL SUPPORTED TORQUE OF BOTH AXLES = 250lb/ft
TORQUE SPLIT = 50lb/ft Front and 200lb/ft Rear


Hayes, that's how TBR is defined with EDL vs a mechanical LSD device. With EDL torque to the axles is always 50/50 regardless of weight distribution and turning radius. With a 'locker' torque follows axle load or turning radius. With a LSD, torque follows axle load or turning radius up to the TBR.

This is all defined in the torsen white paper. VC, torsen, Clutchpack, electronic clutchpack, Haldex. All better than EDL for total supported torque IMO. The new technology combines EDL with a gear type (torsen/Quaife) LSD device, which I have quite successfully retrofitted (install rear torsen into stock rear EDL) in the Audis for several years now.

ST
 
Last edited:
SUMO.
You really didn't address the scenarios.

I'm asking you to actually explain--in a physical sense--where your TBR comes from. Please point out where my understanding is incorrect, and your TBR is correct.

Repeating the "TBR = 4:1" us just restating your insistence that the HF2AV HAS a TBR. Your formulas explain nothing. I understand them, and I understand how TBR works. I still don't see any reason to insist that the HF2AV functions this way. As far as I see it, the HF2AV works the same as your "EDL," albeit less intelligently.

I've said nothing about heat.
Frictional heat created in the VC by differential speeds between the axles is merely what causes the VC to thicken (effectively applying a brake!).

Hayes


Actually, just find me somewhere where Toyota describes the HF2AV in terms of TBR.
 
As far as I see it, the HF2AV works the same as your "EDL," albeit less intelligently.

Hayes - here's a quote from a PM Sumo sent me:

SUMOTOY said:
EDL replaced the LSD in many applications. Isn't that because it's doing the same thing as the LSD?

but hey, he also PM'd me

SUMOTOY said:
Coupled diff don't necessarily work at ALL like a LSD.

and posts in this thread

SUMOTOY said:
A VC LSD is no different than any other type of LSD in terms of how it works, and it's ascribed TBR.

I'd say you have the right idea here

Actually, just find me somewhere where Toyota describes the HF2AV in terms of TBR.

but I think you'll have to find it yourself. If you do, please share.
 
SUMO.
You really didn't address the scenarios.

I'm asking you to actually explain--in a physical sense--where your TBR comes from. Please point out where my understanding is incorrect, and your TBR is correct.

Repeating the "TBR = 4:1" us just restating your insistence that the HF2AV HAS a TBR. Your formulas explain nothing. I understand them, and I understand how TBR works. I still don't see any reason to insist that the HF2AV functions this way. As far as I see it, the HF2AV works the same as your "EDL," albeit less intelligently.

I've said nothing about heat.
Frictional heat created in the VC by differential speeds between the axles is merely what causes the VC to thicken (effectively applying a brake!).

Hayes


Actually, just find me somewhere where Toyota describes the HF2AV in terms of TBR.

Hayes:
I tried to point out how EDL differs from LSD in terms of torque split to help you grasp the concept. IYO, there is no such thing as a fixed rear bias torque. Found in the audi quattro Gen II torsens, the Mitsu 3000VR4, BMW X5, etc.

Look carefully at the Total Supported Axle Torque in the formulas above. A VC is a LSD device, period. It differs from a Torsen LSD in that it requires wheel spin before TBR, where a torsen has reached the TBR before a wheel spins. Otherwise they are the same in terms of TBR application.

Toyota doesn't have to describe the VC in TBR terms, ALL LSD devices have a TBR, read the Torsen white paper I referenced. TBR doesn't support your understanding of LSD devices.

Hayes, maybe it would help if you correct the TBR formulas I presented above to support and demonstrate your opinion. From what I understand you to believe, the Total Supported Torque of a LSD device is no different than an open diff or open diff with EDL. Which means you believe in the VC example above the VC will have 200lb/ft of torque, 100lb/ft to each axle? If we have the given that the front axle can only support 50lb/ft of torque, how can that be? A LSD allows more torque to be supported by the other axle by locking the other axle. EDL allows more total torque to be supported by increasing the torque capacity of the lower supported torque axle, increasing the torque supported by both axles.

A VC isn't a brake type LSD. it locks the two driveshafts together allowing the *transfer* of torque from one axle to the other. A locker allows 100% torque transfer, a LSD allows something less than that. = TBR. EDL and open diffs can't transfer torque between axles, EDL only increases the total torque supported by both axles equally, always 50/50

Are you saying you believe that the VLSD can't transfer torque between axles?

ST
 
Last edited:
From what I understand you to believe, the Total Supported Torque of a LSD device is no different than an open diff or open diff with EDL. Which means you believe in the VC example above the VC will have 200lb/ft of torque, 100lb/ft to each axle? If we have the given that the front axle can only support 50lb/ft of torque, how can that be?

No, I'm saying that the VC in the HF2AV is no different than an EDL in as much as it "allows more total torque to be supported by increasing the torque capacity of the lower supported torque axle, increasing the torque supported by both axles." It's not activated by the ABS, or applied through the braking system, but it IS "activated" by wheel spin, and it does brake the spinning axle through the friction generated in the VC itself.

A LSD allows more torque to be supported by the other axle by locking the other axle. EDL allows more total torque to be supported by increasing the torque capacity of the lower supported torque axle, increasing the torque supported by both axles.
A VC isn't a brake type LSD. it locks the two driveshafts together allowing the *transfer* of torque from one axle to the other. A locker allows 100% torque transfer, a LSD allows something less than that. = TBR. EDL and open diffs can't transfer torque between axles, EDL only increases the total torque supported by both axles equally, always 50/50

Yes, locked IS different than the braking in EDL.
As I've said, the HF2AV behaves like an dumb EDL, not a partial-locker that can be described with TBR.

Just like in an EDL system, because it's all transferred through the traditional open diff, the engine has to apply torque to the braked axle as well as the unbraked one. In a locker, there is no braking of any axle, there is no diff, it's one axle. In the HF2AV, there is always an open diff.

I think your equation for the HF2AV is wrong because it shouldn't have a TBR in it at all. It should look just like EDL equation.

Are you saying you believe that the VLSD can't transfer torque between axles?

No, I'm saying that the HF2AV is not a VLSD. It's an open diff with a viscous damper.

Toyota doesn't have to describe the VC in TBR terms, ALL LSD devices have a TBR, read the Torsen white paper I referenced. TBR doesn't support your understanding of LSD devices.

OK. The TBR of the HF2AV is 1:1.

I think it could be argued that Toyota sees things the same way.


Hayes
 
So in the end..

What do all these equations mean to the world?

Or more importantly, what are you trying to tell us?
 
Try this one SUMO:

Consider a VLSD like what is found in many Subarus.
There is no traditional geared differential--just a viscous coupler.
YES, the VC is only able to transfer a fraction of the torque supplied by the transmission, hence a TBR other than 1:1.

The diff in the HF2AV IS a geared differential that ALWAYS biases torque 1:1--as all traditional geared diffs do.
The VC in the HF2AV is just a crude braking device that increases the minimum supportable torque through the friction created when there is differential rotation between the axles.

Geared diff TBR = 1:1 (FZJ80)
VC difff TBR not = 1:1
 
So in the end..

What do all these equations mean to the world?

Or more importantly, what are you trying to tell us?

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.
 
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.


You aren't having this conversation... :D .





Truth for its own sake.

Granted, it's running on to impressive length, but it IS tech.
Hopefully, somebody will come away with a better understanding of something :doh: .
 
Back
Top Bottom