Understanding the Traction-LOK Diff

[Vlad Soare]

Hi guys,

I'm trying to understand how the clutch-type diff we have in our Mustangs works. I've read a lot and seen a lot of videos, and I think I now have a pretty good idea of how such a diff is supposed to work, but there's still something I don't get.

As far as I understand, there are three ways of getting pressure in the plates.

1. The preload spring. This pushes the side gears slightly apart, thus ensuring there's always a bit of friction in the plates to begin with.
2. When one of the wheels tries to rotate much faster than the other, its corresponding side gear will tend to move inwards, which will cause the other side gear to move outwards, which in turn will increase the pressure in the friction plates on the other side - in effect locking the other wheel to the diff casing and thus forcing it to rotate as well.
3. Unlike an open diff, in a clutch-type LSD the spider gears aren't connected directly to the differential case. Instead, the case is connected to two pressure rings, and the cross pin sits in a beveled opening between those rings. As the differential case rotates, the pressure rings will rotate as well, pulling the cross pin in the process. But the cross pin, due to the beveled nature of the opening in which it sits, tends to push the pressure rings apart. This puts even more pressure in the pressure plates, further increasing the lock effect.

So far, so good. But here's what puzzles me. I've been looking at pictures of the Traction-LOK and have been watching several videos of people taking them apart and/or rebuilding them. And I'll be damned if I can see anything related to point no. 3. I can't see any pressure rings, and the cross pin appears to be firmly connected to the differential casing.

Am I missing anything, or is no. 3 really not implemented in this particular diff? Is the amount of lock we can get simply limited to the tension in the preload spring, plus whatever lateral force might be induced in the side gears? Is there no extra mechanism for progressively increasing the locking effect as you apply more and more input torque?

Thank you.

Sponsored

 

[TeeLew]

Hi guys,

I'm trying to understand how the clutch-type diff we have in our Mustangs works....

As far as I understand, there are three ways of getting pressure in the plates.

1. The preload spring. This pushes the side gears slightly apart, thus ensuring there's always a bit of friction in the plates to begin with.
2. When one of the wheels tries to rotate much faster than the other, its corresponding side gear will tend to move inwards, which will cause the other side gear to move outwards, which in turn will increase the pressure in the friction plates on the other side - in effect locking the other wheel to the diff casing and thus forcing it to rotate as well.
3. Unlike an open diff, in a clutch-type LSD the spider gears aren't connected directly to the differential case. Instead, the case is connected to two pressure rings, and the cross pin sits in a beveled opening between those rings. As the differential case rotates, the pressure rings will rotate as well, pulling the cross pin in the process. But the cross pin, due to the beveled nature of the opening in which it sits, tends to push the pressure rings apart. This puts even more pressure in the pressure plates, further increasing the lock effect.

Hi Vlad,

Let's get 'into the weeds' as we say.

1. You've got this part correct. The preload spring provides a continuous spreading force which provides a certain amount of clutch pack engagement regardless of drive torque.

2. A Trak-Loc locking has nothing to do with one wheel attempting to rotate faster. It is strictly a function of torque on the Crownwheel/Ring gear. As that torque is increased, the cross shaft transmits the drive torque to the pinion gears (the little ones on the cross-shaft). They then apply that force on the side gears. Because this is a bevel gear, the transmission of this force also produces a spreading force which is resisted by the clutch packs on both sides. 1/2 of the clutch plates have tabs on their outer diameter which locks them to the differential housing. The other half have internal splines on the inner diameter to fix them to the side gear. The greater the torque on the Crownwheel, the greater the spreading load on the gears which is laterally supported by the clutch packs. This spreading load provides the friction necessary to resist the independent rotational motion of the right and left wheels. This is what locks the differential. Because the main job of the pinion/side gears is to allow relative rotational movement as lower torque, there is a relatively low locking potential to this differential overall. When using this type of differential in a racing environment (Emco is the racing manufacturer which uses this style of differential), the common approach is to use _a lot_ of preload, because the additional locking load provided by the gears is somewhat limited.

3. There is another type of clutch-pack differential which uses the cross-shaft as a cam to work against side 'ramps'. In this scenario, the spreading force is not put through the pinion gears, but through these 'ramp' plates. These are used in racing because you're able to vary the locking forces between the drive and coast independently and the amount of locking force availabe to the clutches can be very high. By changing these ramp (it's actually a cam, but it's called a ramp) angles, you can also vary the overall locking forces independent (mostly) from crownwheel torque. The only differential that I know of available to a Mustang which uses clutches/ramps is the OS Giken. I suspect it can be made to work quite well.

 

[Superspirit]

Very nice explanation!

 

[Vlad Soare]

Hi Vlad,

Let's get 'into the weeds' as we say.

1. You've got this part correct. The preload spring provides a continuous spreading force which provides a certain amount of clutch pack engagement regardless of drive torque.

2. A Trak-Loc locking has nothing to do with one wheel attempting to rotate faster. It is strictly a function of torque on the Crownwheel/Ring gear. As that torque is increased, the cross shaft transmits the drive torque to the pinion gears (the little ones on the cross-shaft). They then apply that force on the side gears. Because this is a bevel gear, the transmission of this force also produces a spreading force which is resisted by the clutch packs on both sides. 1/2 of the clutch plates have tabs on their outer diameter which locks them to the differential housing. The other half have internal splines on the inner diameter to fix them to the side gear. The greater the torque on the Crownwheel, the greater the spreading load on the gears which is laterally supported by the clutch packs. This spreading load provides the friction necessary to resist the independent rotational motion of the right and left wheels. This is what locks the differential. Because the main job of the pinion/side gears is to allow relative rotational movement as lower torque, there is a relatively low locking potential to this differential overall. When using this type of differential in a racing environment (Emco is the racing manufacturer which uses this style of differential), the common approach is to use _a lot_ of preload, because the additional locking load provided by the gears is somewhat limited.

3. There is another type of clutch-pack differential which uses the cross-shaft as a cam to work against side 'ramps'. In this scenario, the spreading force is not put through the pinion gears, but through these 'ramp' plates. These are used in racing because you're able to vary the locking forces between the drive and coast independently and the amount of locking force availabe to the clutches can be very high. By changing these ramp (it's actually a cam, but it's called a ramp) angles, you can also vary the overall locking forces independent (mostly) from crownwheel torque. The only differential that I know of available to a Mustang which uses clutches/ramps is the OS Giken. I suspect it can be made to work quite well.

Great explanation. Thank you!

So, if the amount of locking load provided by the pinion gears depends solely on the input torque, would it be correct to say that it's a 1-way LSD? While braking or coasting, the only load in the friction plates is provided by the preload - which isn't much. Right?
Or does that preload, little as it may be, make it technically a 1.5-way?

Also, if the amount of lock increases with the input torque, independently of what the wheels are doing at the time, then what happens if you keep your foot down while going round a corner? Will it still be able to do its basic job as a differential, i.e. to allow the wheels to rotate at different speeds round a corner? Can, given enough input torque, the load in the pressure plates ever become high enough to eliminate any slip at all? I have a hunch that the answer is no, because if you could get so much lock with just the no. 2 method, then there would have been no need to invent no. 3, would there?

 

[TeeLew]

Great explanation. Thank you!

So, if the amount of locking load provided by the pinion gears depends solely on the input torque, would it be correct to say that it's a 1-way LSD? While braking or coasting, the only load in the friction plates is provided by the preload - which isn't much. Right?
Or does that preload, little as it may be, make it technically a 1.5-way?

Also, if the amount of lock increases with the input torque, independently of what the wheels are doing at the time, then what happens if you keep your foot down while going round a corner? Will it still be able to do its basic job as a differential, i.e. to allow the wheels to rotate at different speeds round a corner? Can, given enough input torque, the load in the pressure plates ever become high enough to eliminate any slip at all? I have a hunch that the answer is no, because if you could get so much lock with just the no. 2 method, then there would have been no need to invent no. 3, would there?

My pleasure. I like having these more in-depth discussions.

I don't really like the 1, 1.5, 2-way nomenclature that Giken uses. It just doesn't describe the action in a precise enough manner. I'll try to explain what's going on.

• The Trac-Lok (or Torsen for that matter) works similarly both on coast and drive. The locking provided by the pinion/side gear interaction when transmitting engine torque will be the same in either direction. Obviously, the drive lock will be proportional to the force driving the car forward. The coast lock will be applied at the same proportion, but is only a function of the rotational inertia of the engine braking on decel, so quite limited.
• The preload is present at all times (regardless of differential type, for our application). The locking characteristics of a Trac-Lok with (black) and without(blue) the preload spring might look something like this:
  
• Remember when I said a clutch/ramp-style differential could be varied between coast and drive? That's when the 1, 1.5, 2 nomenclature comes into play. Using these designations, the Trac-Lok is considered a 2-way diff, because the coast/drive locking proportion is symmetrical.
• In the case of a 1-way differential, the locking would only be in one (generally drive) direction and the coast side locking would remain at whatever the preload value regardless of engine braking inertial load on the car. If we chose the Giken ramp to mimic the Trac-Lok drive locking, It would look something like this:
  
• The 1.5 setting provides locking in both directions, but they are not symmetrical between drive and coast. This is commonly done to address the handling characteristics of a given car. On Mustangs, Giken will often provide less locking on coast versus the drive side to allow the wheels to rotate individually on corner entry and the a greater locking on power to effectively lock the drive wheels with respect to each other once accelerating forward. A 1.5 differential might look like this:
  
• Many cars have an issue with the rear end being unstable on entry. By increasing the lock of the differential and not allowing the rear tires to rotate independently, we add yaw stability (yaw damping if you prefer) to the car. In this scenario, your 1.5 way differential might be set to look like this:
  
• Now, to answer your last question about what happens if you're going around a corner with the throttle flat, let's back up. A differential which provides equal wheel torques across the axle and allows a rotational difference is an open. A differential which provides equal wheel rotational speeds across the axle and allows a wheel torque difference is some version of limited-slip or locker. No differential can do both. Many, like the Trac-Lok, really don't have the torque capacity to maintain absolute locking, so you get a little bit of slip on the friction plates. If the torque split is great enough, you can still get inside rear wheelspin, although it would be reduced compared to an open differential.
• My personal optimum is a differential which locks in a predicable, linear manner from about 0 to 20% throttle. Over about 20%, the rear tires should be locked with respect to each other rotationally and provide maximum forward traction. This will allow the outside rear tire to maintain a higher drive torque & contact patch longitudinal force than the inside. Keeping the inside rear front spinning given it some capacity to provide a lateral load component to grip. This long. contact patch force difference across the rear axle provides a turning moment to the car as a whole and helps turn the car while on throttle.

I'll let you chew on that for a bit, but don't hesitate to ask more questions. I've spent a bit of time playing with these types of things.






----------------------------------------------------------------------
For some reason, I can't delete these at the end. They're not here when I edit!

 

[Vlad Soare]

Wow, thank you so much for your elaborate explanations. That's really more than I could have hoped for. I'll probably come back with more questions, but now I need a bit of time to assimilate all of this.

 

[Vlad Soare]

Actually, I have no more questions regarding its inner workings. Your explanations are perfectly clear. It's all clear to me now.

Now, it's obvious that the pressure plates will eventually wear out. It's not a matter of if, but when. There's nothing you can do to prevent that. They won't just wear out when you abuse them, i.e. by doing donuts all day long. Due to the preload, even a casual drive to the shops will cause a bit of wear. Every time you turn the steering wheel a little, you're killing the clutches in the diff a little.
But how long it takes for the plates to become so worn out as to turn the diff into an open one... there seems to be no straight answer to that. Or at least I couldn't find it. Some say they will wear out quickly - though what "quickly" means is anyone's guess. Others say theirs are still strong after more than a hundred thousand miles.
Which is why I'm not going to ask how long I should expect mine to last. I guess nobody could answer that.
So, I'm going to look at it from a different perspective. How will I know when they're worn out and need to be replaced?
Of course, the day I put my foot down hard and I'm just sitting there spinning one wheel, I will know. But I'd like not to get to that point in the first place.
If I lift one rear wheel and try to turn it by hand while the other wheel is on the ground, can I assume that the diff is fine as long as turning the wheel takes some noticeable effort?

 

[ice445]

Actually, I have no more questions regarding its inner workings. Your explanations are perfectly clear. It's all clear to me now.

Now, it's obvious that the pressure plates will eventually wear out. It's not a matter of if, but when. There's nothing you can do to prevent that. They won't just wear out when you abuse them, i.e. by doing donuts all day long. Due to the preload, even a casual drive to the shops will cause a bit of wear. Every time you turn the steering wheel a little, you're killing the clutches in the diff a little.
But how long it takes for the plates to become so worn out as to turn the diff into an open one... there seems to be no straight answer to that. Or at least I couldn't find it. Some say they will wear out quickly - though what "quickly" means is anyone's guess. Others say theirs are still strong after more than a hundred thousand miles.
Which is why I'm not going to ask how long I should expect mine to last. I guess nobody could answer that.
So, I'm going to look at it from a different perspective. How will I know when they're worn out and need to be replaced?
Of course, the day I put my foot down hard and I'm just sitting there spinning one wheel, I will know. But I'd like not to get to that point in the first place.
If I lift one rear wheel and try to turn it by hand while the other wheel is on the ground, can I assume that the diff is fine as long as turning the wheel takes some noticeable effort?

In theory they will last a lot longer on a street driven car versus one that is used in autocross or on track. The more torque you put into the unit, the more the clutch packs will slip trying to maintain the lock. Supercharging will also probably delete the unit's clutches a lot quicker.

The one thing I don't get (and it's because I can't really put together in my head how all the parts fit together) is if actual wheel spin is easier on the clutch packs versus tires that are always gripping hard. Like, is a car running 255's that blows both rear tires off on a launch or a street pull harder on the diff, or is a car running meaty 305's with a sticky compound easier on it? I would think that with the tires spinning it's much easier to get a higher torque difference between each wheel which makes the diff work harder, correct? I know technically the contact patch doesn't increase as much with a tire diameter increase, but for the sake of theory let's just assume that the bigger tires in this case have 2x the friction as the smaller tires. Is that extra friction making the diff's job easier or more difficult? Is there more shock load on the clutches as one tire grabs and the other doesn't?

I'm also curious as to what aftermarket differentials will fit in the stock super 8.8 housing in lieu of the Trak Lok, assuming you do manage to wear it out.

 

[Vlad Soare]

In theory they will last a lot longer on a street driven car versus one that is used in autocross or on track.

I'm sure they will. But how much is a lot? Will they last twenty thousand miles? Fifty thousand? A hundred thousand? I don't think there's a straight answer to that. That's probably like asking how long your brake pads will last - impossible to answer. Which is why I'm trying to approach this from the opposite angle: I don't care how long they will last, I just want to know how I will be able to tell when they're on their last legs.

To make matters worse, I can't for the life of me find any information regarding any scheduled maintenance of the diff. There's nothing in my service booklet, nothing in my user's manual, nothing on Etis (which is actually Ford's official source of information for licensed independent operators, so it must be right). The service manual describes how to change the oil, but not how often you should do it. And here's what the owner's manual says:

They seem to imply that no maintenance is necessary, and that changing the oil will never be needed under normal circumstances.
This seems a bit odd. I know a Torsen is fit-and-forget, but a Trac-LOK isn't supposed to be, is it?

 

[ice445]

I'm sure they will. But how much is a lot? Will they last twenty thousand miles? Fifty thousand? A hundred thousand? I don't think there's a straight answer to that. That's probably like asking how long your brake pads will last - impossible to answer. Which is why I'm trying to approach this from the opposite angle: I don't care how long they will last, I just want to know how I will be able to tell when they're on their last legs.

To make matters worse, I can't for the life of me find any information regarding any scheduled maintenance of the diff. There's nothing in my service booklet, nothing in my user's manual, nothing on Etis (which is actually Ford's official source of information for licensed independent operators, so it must be right). The service manual describes how to change the oil, but not how often you should do it. And here's what the owner's manual says:

They seem to imply that no maintenance is necessary, and that changing the oil will never be needed under normal circumstances.
This seems a bit odd. I know a Torsen is fit-and-forget, but a Trac-LOK isn't supposed to be, is it?

Just like the gearbox, the official answer is that no maintenance is required for the life of the unit. Of course that's just what manufacturers do now, especially with splash lubricated stuff. "Let it break".

As for how to tell? I remember reading on here that someone was saying their car kept spinning the inside wheel really bad in autocross, I think he was at 50,000 miles or so?

 

[TeeLew]

Now, it's obvious that the pressure plates will eventually wear out. It's not a matter of if, but when. There's nothing you can do to prevent that. They won't just wear out when you abuse them, i.e. by doing donuts all day long. Due to the preload, even a casual drive to the shops will cause a bit of wear. Every time you turn the steering wheel a little, you're killing the clutches in the diff a little.
But how long it takes for the plates to become so worn out as to turn the diff into an open one... there seems to be no straight answer to that. Or at least I couldn't find it. Some say they will wear out quickly - though what "quickly" means is anyone's guess. Others say theirs are still strong after more than a hundred thousand miles.
Which is why I'm not going to ask how long I should expect mine to last. I guess nobody could answer that.
So, I'm going to look at it from a different perspective. How will I know when they're worn out and need to be replaced?
Of course, the day I put my foot down hard and I'm just sitting there spinning one wheel, I will know. But I'd like not to get to that point in the first place.
If I lift one rear wheel and try to turn it by hand while the other wheel is on the ground, can I assume that the diff is fine as long as turning the wheel takes some noticeable effort?

Differential clutch wear is similar to tire wear. Does it happen through everyday driving? Yes, but it's at a relatively low rate compared to competitive use. The wear rate will be reflected by increased differential temperatures. If you have a lot of slip at high torque loads, you're going to see it in terms of temperature and wear.

In the dark ages of the Fox, a trick was to sand all the friction material off the face of the discs. This would allow you to stack more total discs in the clutch pack and it reduces the wear by making the clutches steel-on-steel, which is how a racing differential is built. More friction faces increase the locking ability as well, so there's that. The downside is you start getting a lot of noise at low speeds when you're on the street.

Ford Performance sells clutch packs with carbon fiber friction faces. Those might be an interesting thing to try both for wear and performance.

A good check to determine wear is to jack up one corner of the rear of the car and, with a torque wrench, determine how much torque is necessary to overcome the preload in the differential. As wear increases, this value will drop. At some point, you'll have to make the call that the wear is great enough to warrant a rebuild.

 

[stanger1]

Wow,
Some very sharp guys in the group. Most impressive.

 

[dx2]

Here are some images that show the main components of the traction-lok differential to accompany Tim's explanation. Images are taken from the workshop manual

Some helpful generic animations:

 

[Elp_jc]

My personal optimum is a differential which locks in a predicable, linear manner from about 0 to 20% throttle. Over about 20%, the rear tires should be locked with respect to each other rotationally and provide maximum forward traction. This will allow the outside rear tire to maintain a higher drive torque & contact patch longitudinal force than the inside. Keeping the inside rear front spinning given it some capacity to provide a lateral load component to grip. This long. contact patch force difference across the rear axle provides a turning moment to the car as a whole and helps turn the car while on throttle.

The closest LSD to that is the Torsen, no?

 

[TeeLew]

The closest LSD to that is the Torsen, no?

Not really. Torsens are good for what they are, but I'm not aware of anyone racing them professionally, at least on a rear drive car. They're a very good set-it-and-forget-it option. They're tough (also heavy), so you don't have to worry about wearing them out. Since you can't really change them much, you also can't get yourself into trouble with them. They provide good limited slip characteristics for the street without inducing maintenance or driveability issues found in similar options. On track, their weaknesses are more apparent.

One issue is the bias ratio. This is the ratio of wheel torque on the loaded side / wheel torque on the unloaded side. A solid axle has a infinite bias ratio and an open is basically 1. The maximum bias ratio of a torsen is determined by the angle of the helical gears inside the diff. For a street application, we have to take low-speed driving into account. If the bias ratio is too high, it will scrub the rear tires in parking lots & similar situations. If it's too low, we get single-wheel wheelspin on corner exit. The happy zone seems to be with a bias ratio of 2.5-3.0. (This is another point in the discussion where I wish our friend BMac was around. He's much more in tune with these particular cars than I am.)

Anyway, I question whether the bias ratio we have is high enough to actually lock the rear tires together in high speed corners with a lot of throttle. In the past, I've seen torsens never really lock, so they end up consuming a lot of energy differentiating across the axle when you really don't want them to. This energy then has to be dissipated as excess final drive heat, which is a common problem with Mustangs on the track. I wonder if those two things might be related?

Another issue with the torsen diff is that it lacks the capacity for us to adjust the most influential differential tool, preload. Differential preload is a very powerful adjustment which we can use to vary the yaw stability and on/off throttle characteristics. If you want to play with a differential, this is the variable you should make an effort to understand first.

My last gripe with the torsen is that they are symmetrical on coast/drive. Generally speaking, we'll want to bias the differential one way or another. I will usually bias a differential to have a higher bias ratio on coast than drive to add corner entry stability, but it's a horses-for-courses type a thing.

This is some data from a reasonable amateur driver going through the esses at Sears Point. The particulars are not all that important, but look at the individual wheelspeeds and get an idea of what's happening. The front speeds are independent and vary in the expected manner every time he turns the wheels from side to side. The rears are almost perfect line-on-line when he's full throttle, but they do allow some differentiation when he is off throttle. Notice that even when the driver is off throttle, the rear speed difference is smaller than the front. This is the effect produced by the preload and coast-side locking. It allows some relative motion, but it doesn't freewheel. I figured I could try to describe it and type forever or just show the picture. I hope it helps!

Sponsored