waltk88

I just ordered a Quaife LSD for my Cayman S. Well, technically it's not a LSD but rather an ATB (automatic torque biasing) differential. Diff manufacturers differentiate between helical gear-driven units and clutch-type units, restricting the LSD definition to the latter. It's my understanding that ATB advantages are 1) no maintenance required and 2) no tendency towards understeer in low speed corners. True LSDs are able to affect deceleration as well as acceleration and so also benefit braking stability. They also continue to work when one driven wheel is off the ground (bouncing off curbs at the track). However, the LSD clutch plates wear out and need periodic maintenance.

I was curious how OEM car manufacturers choose to equip their high performance vehicles. I know that the S2000 uses an ATB differential. Porsche uses clutch-type LSDs. What about others such as BMW and Corvette?

Mr.E.G.

I'm a suspension and aero guy and I don't pretend to be a drivetrain expert but I think you are mistaken.

As I understand it, an ATB diff is just a specific type of LSD.

For example, Vtec is not the same thing as VVT-i but they are both variable valve timing of some variety.


Just out of curiosity, what does such a diff cost for a Cayman? I hear that the Cayman with an LSD is a totally different car. On the six speed forums I read about a guy that was about to know like 3 or 4 tenths off of his quarter mile time with no other modifications.

rockville

That distinction exist only in the minds of the marketing folks. Here's something to think about, basically all passive LSD difs (and that includes, clutch based, helical based, cone based, viscous, etc) work by creating some sort of torque which tries to keep the left and right wheel spinning at the same speed. There are lots of ways to do it but if you look at the forces inside of the dif they are always such that the faster wheels is trying to be slowed by the slower wheel and of course the slower wheel is trying to be sped up by the faster wheel. The exception to the above are active difs and dif like systems such as the torque vectoring systems Honda has used.

The advantages listed are not really true. They may be true in some cases but they are not universally true and thus are more marketing than fact.

But let's rewind a moment. I think it's important to point out that with any conventional dif the average speed of the left and right wheel is the speed of the dif. That means if the inside tire is spinning at 150rpm and the outside is spinning at 200 rpm the dif MUST spin at 175. This is important because you can see that the relative speed of each wheel vs the dif housing is the same but in the opposite direction.

First, clutch based difs. Clutch pack difs work by having a friction clutch installed in the dif. That friction clutch works between the dif output and the dif housing. That means any time a wheel tries to spin faster or slower than the dif housing (basically any time you turn) you get some amount of friction in the clutch pack. How much friction depends on the design. The simplest clutch pack difs will just have springs. That creates a fixed amount of friction (and thus torque) between the wheels.

This torque between the wheels is called the torque bias ratio (for X Nm of torque on the left wheel, how much torque can the right wheel apply). For an open dif the ratio is 1:1. That means the torque applied to the left and right wheel is always the same. If the left is on ice and can hardly apply any torque, the right will hardly get any torque and you go nowhere. With a LSD you can set it so that if the left wheel sees say 100 Nm the right can apply up to say 300 Nm. That gives you a 3:1 ratio. With "torque sensing" difs the ratio can vary between 1:1 and some max value (say 3:1) depending on how much torque the drive shaft applies to the dif.

Another kind of clutch pack system uses the spread of the spider gears in the diff. That spread applies a force to the clutch pack. With that design the torque is proportional to the torque applied to the dif. This type is commonly called a Salisbury diff. The really trick part is if you design it correctly with some ramps you can make it so the torque bias ratio is different in forward and reverse. Contrary to what your marketing info implied this is often a good thing. The Infinity Pro Series race cars often run a higher bias ratio under engine braking vs acceleration. Under braking they want the extra stability of having lots of bias but when accelerating out of a turn they often don't want the understeer associated with a high bias ratio. When combined with a set of springs this type of dif can have some fixed lower level of bias plus bias that increases with torque applied to the motor. This is common in racing applications.

The type of dif you just bought is a Quafe/Torsen 2/helical dif. It is a friction based dif though it doesn't use clutch packs. Instead the helical gears try to separate from the output gears when they are under load. That creates friction between the dif housing and the gears. That friction is what creates your LSD action.

The clutch pack difs probably got a bad name years ago when they used to wear out rather quickly. These days they can easily last the life of the car. They are also easier to tune than helical units. A variation of the clutch pack is the Tochigi Fuji cone dif used in the later Miatas. Rather than having a stack of flat clutch plates, it uses a cone shaped clutch. The result is basically the same.

So lets look back at the claims you were given.
1. This implies that helical difs need no service and clutch pack difs do. This is false on two counts. First, both will need fluid changes over their operating lives. The helical dif will wear over time and as it does it's locking properties will change. Odds are good the changes will be small enough that you won't really notice them on the road. On a race track they would be the difference between a good dif and scrap metal. Clutch pack difs as well as cone based difs can basically last the life of the car with proper oil changes. So in this regard the helical is no better.

2. This is rather poorly defined but I would assume they are referring to understeer due to a high preload on the clutch pack. This however, can also apply to helical diffs. I've observed exactly what they claim to avoid in some Miatas which were equipped with the helical Torsen 2 diff (basically the same as a Quafe.

3. The claim about working when one wheel is off the ground. That is only true of a helical diff when you have preload in the diff. If you don't have preload the diff will spin like an open diff when one wheel is off the ground.

BTW, when asking about oil for the helical dif in my car an expert I know told me this:
This guy used to design and service diff the Torsens used in the Indy cars in the 1980s. He currently has his own line of differentials.

waltk88

I would call them all LSDs, but companies such as Guard Transmission specifically differentiate between torque-biasing and clutch type differentials and only call their clutch units LSDs, even though they also sell the torque-biasing differentials.

A Quaife ATB is roughly $1200 direct from the Quaife depending on exchange rates. US vendors sell the Quaife unit for $1600-1800.

waltk88

Those damn marketers and their oversimplification!

I think it's directionally correct that maintenance is less of an issue with helical diffs. Even marketing for the companies that sell both types message this. I also question whether clutch pack diffs can last the life of the car without replacing the clutch discs. I know that some Porsche owners have experienced LSD failure. Here's one such owner:
http://www.6speedonline.com/forums/gt3-gt2...lsd-buster.html

rockville

The later second generation Miatas used the cone style clutch LSD. It presumably will last as well as the Torsen helical difs used in the earlier cars. Both diff types wear over time. If done right the clutch pack can last the life of the car. In racing applications my colleague said the Torsens (the original Torsen 1 design, not the parallel helical dif Torsen 2) wore out faster than the clutch pack diffs. The clutch pack diffs are also easier to fine tune.

BTW, that link has some very nice pictures of the ramps which allow the clutch pack dif to be torque sensing thus it wouldn't have understeer at low speed (assuming low engine power).

waltk88

Here's another detailed thread on a Porsche OEM LSD:
http://www.planet-9.com/cayman-boxster-com...-dissected.html

Penforhire

Nice explanation r. IIRC, my Z06 used a clutch-based LSD and I was surprised by how durable and unobtrusive it was (both at auto-x and street-tire 1/4 mile). The S2000's LSD was more of a handful in the wet.

rockville

That second thread has a really good set of pictures showing the difference in ramp angles between two difs.