depends what you value


a 1.5 way clutch type would appear to streamline the car's behavior during the various corner transitions


to me this is a confidence adding device, regardless of how quicker it makes the car, if it makes it easier to drive smoother and consistent, then that is indeed money very well spent for any weekend warrior

The problem for alot of people will be that no one can see their new diff they've just spent $2k on. So there's no point in getting it.

Better to get a set of super sik 18x10 xxrs, and get some rear LCA's to pump the camber out to -6; then everyone will know you are a race car driver and respect you.

The benefit, how I use it, is that it can help achieve/enhance turn-in rotation under trail braking. This allows me to enter a turn under a state of positive slip angle, rather than negative slip angle (understeer). Normally, once I enter the corner, I have to use a combination of steering and throttle modulation to get the car into a state of slip for maximum cornering speed and a set up for a good exit. The 1.5/2 way LSDs enhances the car's rotational response to trail braking, without having to play with brake bias, allowing me to spend less time balancing the car's angle and weight transfer, and more time accelerating.

It can potentially decrease braking distances, but on a FR layout car, that benefit is marginal to none. With a FF or AWD with a front clutch type, it'll help prevent single wheel lockup, similar to what you see in the data in the 1st post.

The major differences:

Engagement - Cusco RS is positive engagement, while OS Giken is negative engagement. In other words, the Cusco is trying to engage all the time, while the OSG is trying to disengage all the time. Cusco RS engagement is controlled by springs in the center gear, while the Cusco MZ engagement is controlled by cone springs on the ends. OSG engagement is controlled by cone springs on the ends, and disengagement is controlled by springs in the center gear. This sounds complicated, but we'll be more than happy to show you in person what is going on.

Maintenance - Clutch packs all eventually wear out. How quickly they wear out depends on your setup, how you drive, and if you're using the proper fluid for your application. **USE THE RIGHT FLUID** Typically speaking, the OSG clutch disks should last longer than the Cusco RS clutch disks.

Maintenance cost - OSG diff fluid is more expensive than Cusco diff fluid. Our blanket recommendation is to use the fluids by the respective manufacturers, but they may vary on a case-by-case basis. Clutch disk cost is approximately the same, but the Cusco will require more frequent clutch disk changes. That being said, you will NOT be replacing the clutch disks frequently. Most Cusco LSD owners have never replaced their clutches, and I've yet to run into a single OSG owner who has replaced their clutch disks.

Initial cost - OSG costs ~200 more than Cusco. Custom configuration of the OSG also costs more than custom configuration of the Cusco.

Break in - Cusco requires a 30 minute break-in. OSG requires no break-in.


Both LSDs come in a 1/1.5 and 1.5/2.0 configuration. Meaning, if you buy a 1.5 way LSD, you can reconfigure it into a 1.0 or 2.0 way LSD later on, if you choose to do so, depending on the version you buy.

a few more months and ill be contacting u guys at csg

Hmmmm a little birdie tells me that we got another LSD for testing...

Thanks again @CSG Mike and @CSG David, the OS Giken I got from you was installed by EFI Logics today, great shop. So far quieter than stock with Whiteline diff and other reinforcements. Can't wait to try it on track and AX. Last planned big mod on my car. Thumbs up

Torsen work quite differently from other types of mechanical lsd. In reality they are not limited slip differentials at all as raising one wheel off the ground (or reducing traction torque close to zero as if on ice) will defeat them.

For normal spirited driving they are close to ideal unless you get the car hard over onto the bump stops, because they torque bias: i.e. the torque delivered to the pinion gets divided equally as long as equal tractive torque is available but as one wheel, usually the inside wheel, loses its grip on the road the torque available for transfer to the other wheel rises as a multiple of the torque utilized by the wheel starting to slip. This multiple is fixed by the internal bias ratio which differs from Torsen to Torsen.

Other types just brake the overspeeding wheel against the slower turning wheel and no torque biasing occurs. Eventually, the mechanical type will simply overspeed both wheels (locked) even though one of the wheels won't be delivering much drive due to inadequate grip, but the Torsen cannot do that, it reverts to a more equal torque split if both wheels begin to spin up.

The ideal is the new electronically modulated mechanical lsd which vary the degree of lock not by relative wheel speed but by computer map. Pioneered in racing this is now available for mundane cars. The latest clever awd systems use this technology to eliminate gearing altogether in the transfer case, replacing them with basically the guts of an electronic lsd, just the oil bathed plates and a solenoid pressure control. In theory a geared differential is no longer actually required, just a standard crown and pinion with the axles connected only through the lsd clutches but there are heat dissipation and wear issues still with using such systems full time. They work for awd because they are part time awd. BMW, Jaguar and others use gearless transfer cases. Torque vectoring can be achieved by these electronic lsd. Haldex is a major manufacturer of this type of transfer case. The new European Golf R awd looks very much like it is using these gearless clutchpacks between the front axles as well, for lsd effect without gears.

For hard track work, especially with high roll stiffness, the Torsen is much less useful because loss of traction on one wheel deprives the other of traction (torque drive) also. You can get a weird feeling of cycling going on as the Torsen reacts to incipient wheelspin by shuffling torque back and forth across the helical "spider gears". In fact, I think somewhere on this site somebody actually measured this happening by graphing the wheel speed sensor traces.

Finally, an lsd isn't the be all and end all. Lotus refused to fit one, and still does so for at least the Elise and Exige, (not sure about the Evora (edit, sure enough the Evora uses only an electronic lsd)), as did Jaguar for a long time until very recently. So did Porsche for its excellent Cayman, again until demand from the racing set "required" Porsche to offer one. A road car is still nicer to drive without one and all the traction control a driver actually needs nowadays can be supplied by the ABS system and software. Purely mechanical lsd are of decreasing interest to most drivers even if they drive aggressively.

A lsd will push for the same reason that stability control works by braking a wheel. Torsen or mechanical will drive the inside rear wheel (mechanical will drag the outside wheel) which induces understeer, i.e. counter rotation forces, making turn in more difficult (which is where the electronics work some magic by backing off the locking) and then on exit the locking effect will tend to straighten the car which is when the electronics usually maximize locking, but if you have high rear roll stiffness relative to the front axle this is when the Torsen will wimp out but the clutch type will shine. Mechanical lsd do rob power though and you benefit less from them if you have a low powered car, such as an Elise or a BRZ. Torsen do not rob power. Mechanical lsd work well only if you have enough drive torque to easily and consistently overspeed one of the rear wheels. The Torsen is completely adequate for our cars most of the time.

A reactive system will *never* be as good as a proactive system.

If a driver chooses to drive with electronic aids, that is their choice, and in all reality, the majority of drivers out there will be faster with the aids on, whether they want to admit it or not.

For those that want to truly be in control of their cars actions, instead of letting a computer decide what it thinks the driver wants, a mechanical LSD cannot be overlooked.

Miatas are faster with a LSD, and this is 100% substantiated by evidence. They don't have very much power. I'm talking NA and NB, not the current NC, which has more power.

The Miata will be faster with more power though, lol. All I am saying is that a mechanical lsd absorbs power (which shows up as heat in the diff oil).

I'd be interested to know if a Miata is faster with a locker or a Torsen. I'd be surprised if a locker worked better but it is of course possible. I'd also be interested to know if this is a Miata with modified suspension or stock suspension because roll stiffness front to rear makes a big difference to traction out of corners with and without a locker versus open or Torsen.

A Torsen is actually a special type of open diff, not a locker. It never locks and is unable to resist differentiation unless both wheels have some traction. A Torsen reveals the diff(erence) between traction and torque delivery which needs to be understood when selecting the type of lsd or differentiation limiting device you need.

A locker works no differently to a brake based system until you want both wheels to overspeed available traction because a locker doesn't transfer torque (which of course a Torsen does) it merely brakes the overspeeding wheel against the other wheel (whereas the brake based systems brake the overspeeding wheel against the chassis). To benefit, the engine must be able to deliver enough torque to at least be capable of spinning up the inside wheel.

These are all reasons why Torsen works better for the front axle than the rear and for the centre diff in awd systems. The new Haldex type systems work even better as lsd than lockers or Torsens.

A clutch-type LSD is not a "locker". A locker is something else entirely, and not really appropriate for road course work.

Also, a Torsen is most definitely a limited slip.

A locker is either locked or open; it can't dynamically change.

An open diff is open. A limited slip diff can limit the amount of wheel slip based on how it is adjusted/set, and can range all the way from open to fully locked, although not all diffs use the full range. A locker in locked position is 100% locked whether you want it to be or not.

Any diff that locks to 100%, no slip, is technically a locker. The fact that some lock immediately with a pawl (rare now) or lock progressively using clutch packs makes no difference. It still locks.

A Torsen is by definition an open differential. It never locks and cannot limit slip. It only transfers torque from a wheel with grip turning faster to a wheel with more grip turning slower. It always differentiates the wheels and never limits slip.

Tor Sen stands for Torque sensing, by which is meant "sensing" (technically "reacting to") the torque the driven wheel can deliver to the ground. The Torsen can then deliver that torque multiplied by the bias ratio to the other wheel if that wheel is turning more slowly and has the capability to use the torque (i.e. enough grip on the road).Failure to understand this essential difference to an lsd creates a mystery where there is none.

OK, if you prefer that difference. It is true that there are lsd that lock essentially immediately and those that do so progressively, though a clutch type usually locks up pretty dramatically. If you want to call one type of lsd a locker and the other just an lsd that is fine with me. A Torsen isn't either type and never locks nor even limits slip. A Torsen is an open differential as you can easily test by raising one driven wheel off the ground. It limits the rate at which differentiation progresses but it doesn't limit slip between the drive wheels or drive shafts nor can it.

Precisely what it is doing, is limiting slip by forcing torque to go to the other wheel.

You can do the same test with a clutch type LSD, and get the same result, provided it is tuned to do that.

However, the very example you bring up, is why a clutch type LSD is such a substantial upgrade over a torsen unit for those that are cornering hard.

@Ubersuber, I don't understand why you're trying to redefine this discussion in semantics which only you recognize. it's only going to confuse readers.

A welded differential or spool is locked 100% of the time.

A "locker" is a differential gear with a mechanism to change between 100% lock and 100% open. They may lock automatically or by cable or pneumatic shifter, but they cannot lock any less than fully.

A "limited slip differential" is "a type of automotive differential gear arrangement that allows for some difference in angular velocity of the output shafts, but imposes a mechanical bound on the disparity." (Wikipedia)

An open differential is a differential gear which is fully open all of the time. Four spiders in a carrier with no extra mechanisms. We all know what this looks like.

Inside the category of limited slip differential there are various types. Due to design limitations, relatively aggressive clutch-type differentials are the only differentials which can deliver power to a wheel which is lifted off the ground or on ice. Viscous and helical differentials will generally still allow you to get stuck in the extreme case of zero traction on one wheel. However, this does not mean that clutch types are "lockers" and other types of LSD are "open", it's just a finer point of the characteristics of each diff. Under normal conditions, all types of LSD bias torque, by definition.

My Kaaz locker works way better than the Torsen unlimited slip

So based on your definitions of it either being a locker or an open diff, what exactly would you define a limited slip diff as?

Lots of great info and insights in this thread; thx!

To balance the most recent highly technical/theoretical posts, I'd like reiterate that changes/upgrades to the diff need to be qualified, otherwise it is likely to be a waste of money, time, and aggravation.

Generally speaking, a clutch type LSD is almost always a performance increase. This is the rule, with very few exceptions.

The tradeoff is that you have to maintain them with a little more vigilance. You don't want to overheat a LSD, but this shouldn't ever happen under 99% of what most users will do, as long as you use the right fluid, and that includes any stress testing that we do.

That being said, the rear end grip gain under corner exit (powering out) is so massive that we're actually debating doing a test on stock suspension with stock tires, just to demonstrate it.

If you ever make it down here, please ask for a ride!

Clutch diffs may act locked under some circumstances, but they are still not "lockers". A locker is either 100% locked or totally open. A very tight clutch-type will act more like a spool than a locker. But you still wouldn't call it a spool! It's still a clutch-type limited slip.

No, it is a limited slip.

"Never locks" is not the definition of an open diff. An open diff cannot bias torque, it can only ever apply EQUAL torque to both rear wheels. A Torsen biases torque to the outside wheel during cornering.
Of course it can and does. If it couldn't, my 500+hp FD would always spin the inside wheel during cornering. But in fact it never does that, it *always* spins both rears when too much power is applied and rear grip is overcome.
It *only* transfers torque from one wheel to the other? That's kind of the whole point of a limited slip...

Like a clutch-type, it uses friction to connect the two drive wheels, and the greater the input torque the greater the friction and the greater the effective "lockup".

Preloaded clutch type has the advantage of being able to apply the breakaway torque at one wheel even when the other is off the ground. T2R also does this to some extent, but with a clutch-type you can have as much as you want/need. In my experience with clutch-types, though, I definitely prefer MINIMAL preload and to let the ramps give increased clamping under power.


TL/DR: A Torsen is a limited slip differential. A clutch-type is not a "locker".

That would be very interesting to see; but I think you'd have to do it using both FRS and BRZ springs.

Ideally we'd love to have a loaner stock suspension for both, so we can just hot swap at the track.

Our dampers have some springs we're testing on them, as well as top mounts and cascam plates...

Because there is a lot of misunderstanding about how these devices work, much of which comes from the standard terminology. That misunderstanding translates to owners buying the wrong type. The standard Torsen fitted to these cars is adequate for most people. With a car this modestly powered I don't know why anyone would bother with a mechanical lsd and a two way or 1.5 way makes no sense at all. Spend the money on a supercharger or a turbocharger first, but then you might not want to even bother with upgrading the diff.

All the mechanical type lock at 100% quite quickly which makes them indistinguishable from true lockers under those conditions. By "lock" I mean the drive axles turn at the same speed which occurs whenever sufficient torque is applied to the pinion input to activate the clutches (preload aside), they are "progressive" only in an extremely limited sense. Most drivers will experience these devices as locked most of the time the car is being accelerated (and of course with 2 way or 1.5 way, the same will occur on overrun). The clutches only slip when torque inputs are relatively low relative to tire grip. Torsen never lock and cannot lock, they can always differentiate regardless of traction levels. A viscous coupling should provide drive to the axle with grip even if the other axle is broken. A viscous coupling replaces a gearbox and works like a torque converter with special fluid that becomes more viscous when heated than when cooler. Any differential action heats the fluid and locks the coupling.

A mechanical clutch type lsd with no spring preload is an open differential until torque is applied. I don't know why anyone would bother to fit such an lsd nowadays because a Torsen is superior to that design.

With spring preload the mechanical type is a genuine lsd and will limit wheel slip always, which a Torsen will not. This is pretty much only desirable if the roll stiffness is altered significantly from stock, which you should of course figure out first. Even then, for anything but track work a mechanical lsd on this car is either going to be useless or worse than useless.

That's why.

Not unless they have high preload. Ideally, they won't.
Acting like a "true locker" is not what you want. Big time understeer, for no good reason.

This is a good thing. Locked or welded diff is not what you want for good handling!

Viscous doesn't "replace a gearbox", and viscous doesn't LOCK. Viscous LSD requires differential wheel speeds. Unlike a clutch-type or a gear-type (torsen, quaife), a viscous type requires one wheel to start spinning before it biases torque. Definitely inferior for performance applications.

Low preload is the way to go IMO. As low as you can get away with. A rwd car with front-biased weight distribution like the FR-S/BRZ will require more preload than one with 50/50 or more rearward weight distribution.

"Mechanical type" doesn't specify much. Gear-types and clutch-types and lockers are all "mechanical". A clutch type with zero preload is still a "genuine lsd", and of course so are Torsens (T1, T2, and T2R). On a car with decent weight distribution, on most tracks lifting a rear wheel never happens, limiting wheelspin with a wheel in the air is usually not a requirement for track work.

Assuming you mean clutch-type here, I agree it's unnecessary, but it won't be *worse* than useless. An appropriately set up clutch-type shouldn't have a ton of preload. A lightly-preloaded clutch type is not going to have undesirable effects on handling and street driveability.

This is worth one more attempt at clarification.

A Torsen limits slip at the tire without limiting differentiation by actually pushing torque across the differential to the wheel with better grip. It does not absorb any torque and it is not wheel speed sensitive, only torque sensitive. The Torsen never reaches any form of lockup. A Torsen never hinders any required differentiation, these are very cleverly designed devices and very difficult to understand if you are not an engineer.

The other mechanical lsd only limit slip inside the differential. They do nothing about the real problem which is limited grip at the tire. They do not and cannot transfer the excess torque as a Torsen does. Very rapidly after wheelspin begins the clutch type lsd locks up 100% and both axles turn at exactly the same speed, at least one tire is abandoned to traction robbing wheelspin (it has to because these diffs cease differentiation and lock up) and often both begin to spin. These devices are axle speed sensistive, not torque differential sensitive. All input torque is immediately divided 50/50 and the only difference in axle speeds tolerated results from tire grip. If the less loaded tire can develop enough grip some differentiation can occur in these lsd. They can only do so by physically dragging the tire that is to turn more slowly against the friction in the clutch, the opposite of what is usually meant by "progressive". In other words, the lsd locks up unless the inside tire can be dragged down to a slower speed by forcing the clutch on that side of the diff to slip...not often in practice.

This is a fundamental difference you need to consider, as the engineers at Toyota and Subaru must have done, when choosing which type of diff to fit. Nobody is suggesting an open diff would be preferable to the Torsen. What is being suggested is a mechanical clutch type diff will be superior to the Torsen under the conditions discussed in this thread. My purpose was to underline what you will be giving up if you decide to fit such a diff to your car for the benefit you seek on the track. My idea is that in return for that benefit you will sacrifice a huge benefit the Torsen gives you everywhere else, which is why Subaru fitted the Torsen type.

Just BTW, the technical product support article on the Cusco website supports my use of the terminology. I recommend reading what they have to say before making this decision.

You do realize that a clutch LSD can (and does) lock less than 100%, right?

You also realize that every word you say just confuses people worse because you try to sound smart, but in reality just go on paragraph long diatribes where a single sentence would've been more than sufficient?

Both torsens and clutch-types limit the differential in wheel speeds, transferring torque from the wheel trying to spin to the slower wheel via friction (between clutch plates in the clutch-type, in the worm gears in the Torsen).

Torsen does "absorb torque" via friction, and they do get quite hot.

Every day I feel and hear the Torsen in my s2000 hinder differentiation as I make the 90 degree corner out of my driveway.

Of course the clutch type DOES transfer torque, same as the Torsen.

No, it binds the two output shafts BEFORE wheelspin, and can bias torque prior to one wheel spinning. It does not necessarily lock up 100%, and for normal tarmac track driving it shouldn't have to to ensure getting all the power to both wheels. The can be set up to give 100% lock under maximum acceleration, but IMO for a well set up and well-balanced car this should not be necessary.

Again, they don't always fully lock, and they don't need to to achieve both drive wheels spinning.

No, they aren't speed sensitive, and they ARE torque sensitive! They lock up more with more applied torque via engagement ramps.

WRONG! That is in fact what an OPEN diff does! A limited slip works because it allows the wheel with more grip to take MORE torque. In a 4:1 Torsen, the bias ratio limits the torque difference to 80/20. Even a 2.5:1 Torsen can deliver 71.5/28.5 bias. A tight enough clutch-type can go 100/0, but again, IMO this isn't usually necessary.

This does happen often in practice, most of the time in fact. Both the Torsen and the clutch-type allow differentiation at the same time they are providing more torque to the outside wheel with more grip. The clutches slip most of the time in a clutch type, and this is a GOOD thing. Only on a tightish clutch type under high torque load will the outputs be "locked".

I would dispute the idea that a clutch-type is always superior. There's a reason the Boss 302 LS came with the T2R Torsen while lesser V8 Mustangs got the clutch-type, and it's not because the T2R was inferior.
That said, there are aftermarket clutch-types that are better for performance than an OEM Torsen, but there are also plenty of aftermarket clutch-types that are no better.

Can we split this thread into "why" and "how they work"?

Oxymoron

Ok, now you're being pedantic just for the sake of it.

Here's comparison done in the past with GRM:

http://grassrootsmotorsports.com/art...hats-the-diff/

Might help most understand the differences between types of LSDs.

Oxymoronic would have been pedantic, I resisted.

Technically, use of "oxymoron" was colloquial rather than correct useage.

Now I'm being pedantic.

Working your way towards another ban I see. Carry on.

Havent taken my power mechanics class yet so im still fairly unsure how these work. Never the less...

:popcorn:

No, you're the moron.

Clutch LSD's aren't on/off switches like you're claiming, they only lock as much as needed, and do so progressively. How aggressive they lock can be tuned to suit the driver. You're mixing behaviours between diff types. A true locker fully locks under any throttle, regardless of differential wheel speeds. A clutch LSD only reacts when the wheel speeds are different, and are tunable to control how great the difference in speeds can be before it starts to lock up. It doesn't go fully locked as soon as the speed is slightly different between the wheels.

I hear/feel this ALL the time in my FRS as well, in tight parking lot turns and even in the parking garage at my office. Any throttle application in a tight turn at very low speeds has the inside tire scratching at the pavement.

Well didn't this thread turn into a bag of d1cks

There is benefit in putting a mechanical lsd in even a stock gt86. Even an average driver can feel it. Drive one with one and one without and you'll see for yourself. Even just on some spirited road driving.

Several of us in NZ had the TRD mechancial 2 way put in as a dealer fitted option. I was one that didn't, and I saw the difference at the track. I then drove one with the mechanical and I felt for myself how much better the car could "spring" out of the corner. (Where as the reality is that it was my car getting held back out of the corner.)

The trd units have the preload set a little high IMO. Great for drifting, but I think it needs to be adjusted, which I plan to do. I actually wished I'd got a 1.5 as I don't mind the accel setting, but wished it was softer on decel as I get push into some corners.

If I didn't already have one, I'd be getting this csg spec'd one, because they've already done the leg work with respect to the settings.