Causes for Understeer???

[MrAWD]

Here is the second part of the essay!

Ah, but grasshopper, weight distribution under power is NOT 60/40. 60/40 is the upper bound; the actual distribution is going to be somewhat more rear biased. The power and grip (and the higher the CG) the more rear bias you have.

I think that I already went over this part and still waiting for the answer

If the car is perfectly balanced mid-corner, and the torque split is 50/50, then adding power MUST cause understeer, because rearward weight transfer is moving grip rearward.

OK, so we give grip back to the fronts by reducing steering lock - but our centre diff, which is always (for lack of a better term) "seeking" to restablish a 50/50 torque split, will steal that grip to put power down with it. That means more understeer. The only way to get away from the understeer is to reduce power.

Yes, you would get more understeer if you have a mechanical diff of any kind (even Quaife). The problem for them is that they can only fix one of the problems, which are either to go straight or to turn.

With ACD, you get that problem solved for both cases, so you can still apply extra power and still be fine. What happens there is that the ACD gets help from the steering angle and throttle opening sensor! When you apply more power (to the already “perfectly” balanced car) in the middle of the turn, ACD locks a bit more and that evens things out.

The best part here is if your car does goes off the “perfect” line, a small correction on the steering wheel adjust both front and rear ends. Front gets more or less of the steering angle, while rear gets more or less power depending on the steering angle. Basically, more turn opens up the center diff and more power goes to the rear which helps rotation. On the other hand, less steering angle locks the diff a bit more and cuts some of the power that was going to the rear and send it back to the front, which is also turned less, so car is back to stable in either case!

Fedja

 

[jtmcinder]

One thing that needs to be said before this gets completely out of control is that the ACD of an Evo is very different from the DCCD of an STi. My previous comments on Dennis' whacko ideas on the STi need to be read separately from any comments that Fedja has made concerning the Evo. One reason why these two systems are different comes from the differing philosophies of Scooby and Mitsu. The other reason is the layout of the engines which places some constraints on the options.

- Jtoby

 

[CowPimp]

I'm not going to be getting into what the best method of teaching someone to autocross is, or the differences between the STi and EVO AWD systems, but merely responding to the original question.

Upon installing swaybays into my ride (Front and rear from Extreme Motorsports), I definitely noticed a dramatic reduction in understeer. I swear the car does not bias toward over or understeer and balances quite nicely unless I am totally drifting, which I have only done on a couple of occasions. In such a case, I notice a slight amount of oversteer, if anything. Otherwise, it plows through the turn with an even keel.

However, it must be noted that I am a firm believer in the phrase "slower in, faster out." Therefore, I rarely throw myself into a turn so wrecklessly that either type of steering bias would be induced. Although, I really mash the pedal on the way out of the turn.

It should also be noted that the most prominent display of oversteer I have experienced with my current set of mods occured when releasing the throttle too abruptly on the way out of a turn. The car held, but I did have to correct a significant, although not excessive, amount of oversteer.

 

[Greg Collier]

Hey i'm looking to finally upgrade my suspention from stock, and as i've gotten more comfortable pushing my car to it's own limits i've noticed it gets ridiculous understeer, which usually compensates from giving it more gas to correct, but that results in the whole car sliding in the direction of entry into a turn which is not a good thing, i'm wondering that asides from getting 17" rims with eagle f-1 tires thats on my first list of things to get if i should go with a mild shock and strut layout, or go with coilovers???

I'm also planing on getting swaybars for the front and rear, strut tower bar, and a better rear one too.

In all honesty, I prefer oversteer since it's easier to correct. Anyone who can shed some light on the best way to go without spending ridiculous amounts of money let me know. thanks in advance!

It sounds to me like you're getting a handle on what you perfer in your driving technique. Seat time is obviously one of the best ways and not always the cheapest ways of learning the craft. A local guy named Gary Thomason is a national SCCA auto-cross champion and has proved his ability year after year. I watched him run a full day of practice one Saturday at Qual Com stadium. He ran his first lap in his Corvette and set the top time of the day, then he drove his buddy's WRX and set the 3rd fastest time. Some guy was there running a new lowered Ford F-150 and Gary jumped in and by his 3rd run did like the 6th fastest run of the day. Everyone was amazed. It just shows that when you have a handle on your driving technique you can pretty much drive anything.
It sounds like you're serious about your cars modifications and if I might be so bold to interject my 2-cents, "take baby steps". Do your 1st mod, and see how it changes the cars handling, and so-on and so-forth. I have just about every suspension modification there is on my race car, and every time we change or adjust something I have to re-learn the car. My car's FWD so I prefer understeer. I constantly have to let the car "set" before I can balls out accelerate. I'm a firm believer in the concept that, Slower is Faster, and it ususally shows up in my lap times by a second or more.
I've thoroughly enjoyed this thread and the amount of solid information that has been given! Keep up the good work and most of all have fun.....................................
Greg

 

[jtmcinder]

At fully locked (100% locking) the torque split is 50% front.

At fully unlocked (0% locking) the torque split is 100% front ASSUMING FRONT WHEELSPIN (which is the more likely case)

So at 50% locked, and again, assuming front wheelspin, the bias will be 75% front, 25% rear.

I've been trying to make heads or tails of this for a few hours (while also watching Spino Spino Spino [my daughter's name for Jurassic Park III] and feed a 15-month-old). I do not know if Dennis means this to be a description of an Evo or an STi, but it seems to be a mixture of two types of system that are not compatible.

First, the only type of LSD combination that produces a 0/100 or 100/0 split when unlocked is one where power to one end of the car goes through a VC or hydraulic clutch. Examples of such a system are the 4-Matic (Mercedes) and the 4-Motion (VW). Such a system is also used in the A8 and in automatic trans version of Subaru Legacy.

Second, the systems I refer to above also produce a 50/50 split when fully locked, so both the unlocked and fully-locked descriptions fit the same story.

The problem is these systems do not produce a 25/75 or 75/25 split when 50% locked. Interestingly, the only kind of LSD that I know of that produces what can be described as a 3:1 torque split when 50% locked is a helical (i.e., a Type-2 Torsen, such as a Quaife). But you cannot combine a Type-2 Torsen with 4-Matic-type system.

This is the point I keep coming back to and then dump it on the list to help me with. Most of all, Dennis needs to say whether this is his understanding of an Evo or an STi (or both, if he thinks that they are the same) and then he needs to explain exactly what sort of system it is. Are the front and rear wheels both connected to the engine through gears (as in a DSM) or is one end connected through a VC or clutch (as in a 4-Matic)? If there is a mechanical differential, is it a spider-type (like a DSM) or a planetary-type (like in a Cusco Tarmac diff for a DSM or the stock center in a 3000GT VR4 [which uses a planetary to get a 45/55 split])?

In short: Slow down and be clear. You don't have to explain how any of these things work if you don't want to (because I can do that for you), but you have to say what you think is in these cars or you are just causing grief.

- Jtoby

 

[DG-FNR]

For Fedja: Toby is a lost cause.

Let's take the case of a purely open diff. No locking mechanism.

In the case of adaquate traction on both ends of the car, the power split is 50/50.

Place one end of the car on ice, and apply power. Front wheels spin up, rear wheels don't move. Power split is 100/50.

This is a direct consequence of the fact that the power transmission across an open diff is a function of the speeds of the output shafts, with the faster shaft getting the larger share of the power in direct proportion to the speeds of the shafts.

Now when we have the car on the limit, the faster shaft normally indicates slip at that end of the car - wheelspin, or incipiant wheelspin. If wheelspin develops, the open diff sends power to the faster turning shaft, making the wheelspin worse. Loss of acceleration.

So an open diff has a torque split of 50/50 (best case) and 100/0 (worst case) where the first number represents the end of the car with the least grip. At no time can an open diff send power towards the grippy end of the car.

In order to get around this problem, there needs to be some sort of mechanism to slow down the shaft that is spinning too quickly. If that braking mechanism is connected to the opposite side diff output shaft, we get an effective power transfer from the fast shaft to the slow shaft.

But the best this braking mechanism can do - no matter what the mechanism might be - is to lock the two output shfts together. At this point, the torque split is 50/50. A locking mechanism can only go to 50/50. Note that I'm not talking front/rear split here, but less grip/more grip.

A sufficiantly "intelligent" locking mechanism, be it electronic or mechanical, "seeks" a 50/50 split. When it "detects" power "leaking" out a slipping output shaft, it attempts to apply a locking force to send power to the other output, up to the limit of 50/50.

A less sophisticated locking mechanism cannot react fast enough, or reacts in discrete steps (Detroit Locker), or cannot generate enough locking force to actually reach 100% locked, or otherwise underperforms our "perfect" locker. Such a diff might wind up being able to generate (say) a 60/40 torque split, but the high number is always towards the end with the LEAST grip.

So to review: front end of the car on ice, rear on grippy pavement. Open diff, 100/0 torque split. "Perfect" locker, 50/50 split. Imperfect locker, somewhere between 100/0 and 50/50.

The EVO's centre diff is a locker - a computer-controlled locker, but a locker none the less. That means the best that it can do is a 50/50 split - that's physics; there's no way around that.

It has the potential to outperform mechanical lockers, but that it will is by no means a given. There is a dependancy on the calibration of the control system, but also a dependancy of the precision and authority of the actuators. WRC transmissions actually step up the speed across their control clutches to twice input speed, so as to reduce the amount of actuator torque required to result in a change in shaft speed (doubling the speed halves the torque across the shaft) making their actuators more sensitive. And yes, sufficiantly advanced diff controllers can take in info from other sensors and act on that information to provide handling changes simply not possible on a mechanical diff. for example, an WRC controller looking at steering angle and "hand brake" position might decide that a tight turn was being called for, and route power to the outside wheels only, pivoting the car around like an APC.

Given the necessity to work correctly during street driving, I rather doubt the EVO diff controller is this sophisticated or aggressive. I also wonder if the actuators are fast enough/sensitive enough... but that's another topic for another day. The important part to understand about the EVO's system is that it can generate torque splits between 100/0 to 50/50, where the largest portion of the split is ALWAYS towards the wheel with the least grip.

Because weight transfers rearward under acceleration, this means that there will always be a movement towards understeer with this kind of diff. If we assume the system works perfectly, the amount of understeer will be dramatially less than with an open diff, and some degree less when compared to a mechanical diff, depending on how good the locking mechanism on the mechanical diff is - where the Quaife is probably the best. (and is very, very good)

Now the mechanism on the STi is different. It is NOT just a regular plain diff with a locking mechanism; it has a mechanism whereby the power split can be biased "over" 50/50. Where any open diff with a locker must fall between 100/0 and 50/50, the STi diff can go 35/65. It can send more than 50% of the power to the end of the car with the most grip, where the EVO cannot. This means that the STi is capable of true power-on oversteer, because while weight transfer is sending grip rearward under power, the power is also being sent rearward. If it works perfectly, the power transfer rate equals the grip gained rate, and the car stays neutral through the turn. If power is being sent rearward FASTER than it is gaining grip from weight transfer, oversteer will result.

Is that clear?

DG

 

[jtmcinder]

Is that clear?

It is clear that you do not know what you're talking about.

#1) Types of center differential

There are three totally different kinds of center differential; Dennis is only describing one of them. The type Dennis seems to understand is the spider-type center that comes stock in a DSM. It is 50/50 when locked (regardless of how it gets locked).

The second kind of center diff only transmits power to one end of the car via gears. The other end gets power through some sort of clutch. Sometimes the clutch is just a VC. VW uses this and calls it the "Syncro." Other times the clutch is hydraulic as in the 4-Matic and 4-Motion systems of Mercedes-Benz and VW, respectively. These diffs all have a 100/0 or 0/100 split when the center is open and become closer and closer to 50/50 as they lock. [edit: for reasons that I explain in a later post, it is probably better to think of these cars as always having a 0/100 or 100/0 torque split, but also having the ability to bias torque, but I will leave it this way for now.]

The third kind of center diff is the planetary-type that allows for a pre-set, non-50/50 split when open and a 50/50 split when locked. These are amazingly cool and understanding them is crucial if you really want to know where drivetrain technology is going. This is the kind of diff that ACM has in his ESP car; the kind that I have had to repeatedly remind Dennis of since he keeps making the false claim that it is impossible to get rid of power-on push in a DSM. These have actually existed for quite some time since they are used in the center of the 3000GT VR4 (to get a 45/55 split) as well as some Subaru Legacys with automatic transmissions. If an explanation of how these work would be useful, I can try to provide one in a separate thread. For now, the key issue to keep in mind is that you can have a non-50/50 split when open and a 50/50 split when locked while still having both ends of the car connected to the engine via gears. This is important because it allows you to have a pre-set unequal split most of the time (such as 35/65, as in the STi) and still not trash the tires if you run it locked and hit some dry pavement.

#2) Methods of limited slip

Some diffs have an limited-slip device built into them and some have them piggybacked. This is another basic point that needs to be clear. A Quaife (Torsen Type-2), for example, is not only a diff, but is also a limited-slip device. The helical gears that provide the differential action are also the locking devices. Very cool and no plates to wear out, but it also means that you cannot make a Quaife with a split that is biased in a fixed direction. You can change how much bias a Quaife can produce (and there is a pretty graph that converts percent locked to bias and back again), but you can't set the bias in advance. It locks as a function of a difference in reflected torque; it does not "care" if the end of the car with more torque is the front or the rear and it can't be made to care (as far as I know).

Other diffs have no limited-slip function built in but allow you to add one on. Some of these "piggybacks" are hard to separate from the diff since they share space and some are physically distant, but they all share the fact that you could break or disable the limited-slip device without preventing the car from being driven. In fact, The Great Unwashed might not even notice that the limited-slip device is gone.

One type of piggyback limited-slip device is a VC on the output shafts. When the outputs turn at different speeds (signalling slip [on the assumption that drive ratios after the diff are the same and the tires are the same diameter]), the VC heats up and resists the difference. This is what DSMs have.

Another type of limited-slip device is the Detroit Locker. These racket into place when power is applied and lock the diff solid; they release when power is not applied. They are the ultimate one-way diff. Great for drag-racing. Lousy for anything else.

A third type of limited-slip device is the Kaaz-like clutch-pack. (It has another name [Salisbury?] but I can't remember it right now.) These work by resisting a difference in rotation speeds between the carrier and one or both outputs. They start off with relatively little pressure between the plates that do this resisting (so it is effective open when coasting) but gain pressure (and therefore locking effect) as a function of how much torque is being transmitted through the diff. They do this by having the spider float in V-shaped slots so that transmitted torque forces the spider to wedge one side of the casing holding the plates outwards, adding pressure. One neat thing about these is that the V-shaped slots do not need to be the same shape on the front and rear sides, so you can create a limited-slip device that works more in more direction (e.g., accelerating) and the other (braking). A Kaaz-like clutch-pack that has symmetrical slots is a 2-way; if there is a V-shape on both sides but one is steeper, it's a 1.5-way; if one side is actually flat, then it's a 1-way.

A fourth type of limited-slip device also uses a clutch-pack but the pressure is controlled by a computer, instead of transmitted torque. This is what you have in an STi. If you have been reading carefully so far, it may have just occured to you how similar a hydraulic-clutch limited-slip device is to a hydraulic-clutch center differential. I agre with you and think that this causes many people to become confused. The key thing to focus on when trying to keep this straight is whether the clutch in question transmits power to one end of the car (4-Matic differential) or only acts to limit slip (STi). As I'll explain in a minute, it will never be the case that a car has both of these.

#3) Combinations

If the center diff is either a VC or a hydraulic clutch, then there will not and can not be a limited-slip device piggybacked on. In a way, the type of center diff that is 0/100 or 100/0 when open and 50/50 when fully locked already is a limited-slip device. [edit: but see a later post for a better way to think about this.]

A Quaife is not only a diff but is also a limited-slip device. It is truly an LSD. But that doesn't prevent you from adding a piggyback on top of it. Many people, it seems, leave the VC in the car when they install a Quaife center. If the center Quaife for our cars is a Type-2 Torsen, then this makes a lot of sense because Torsens create a torque bias or ratio, so if one output is reflecting zero torque it goes open. The Type-2R Torsen has a pre-loaded center to avoid this issue. But if our centers are just Type-2s, then it makes sense to leave the VC in there.

Spider-type diffs can have VCs or clutch-packs added to them and need to have something added or else they are not limited-slip. The stock center of a DSM is a spider-type diff with a VC slapped on. When you buy a Kaaz, you get a spider-type diff with a clutch-pack.

Planetary diffs can also have a limited-slip device added. ACM had been running a Cusco 35/65 planetary in the center with a (half-dead?) VC. The center diff in an STi is also a 35/65 planetary, but it has a computer-controlled hydraulic clutch-pack on top of it. In automatic mode, it adjusts the pressure between the plates on the fly. In a nutshell, it defaults to open and adds pressure when wheel-slip during acceleration is detected. In manual mode, you get to set the pressure. You can set it to fully locked, if you wish, which creates a 50/50 (and non-differentiating) center. It short, it's the center-diff equivalent of a remote cut-out exhaust: hit the switch and you have a spool. Note, however, that it is not advised that you put it in 50/50 and then try to use a 2-wheel dyno. That is asking too much of the plates in the clutch.

#4) That's nice, now can you answer the question about the Evo?

Not right now.

- Jtoby

 

[jtmcinder]

#4) What's in an Evo? (More precisely: what is an ACD?)

It's just a hydraulic limited-slip device, very similar to what's in an STi. In fact, and maybe contradicting something I said yesterday (if you're keeping score at home), the only real difference between the Evo's ACD and the STi's DCCD are these: (a) the diff in an Evo is the same old 50/50 spider-type as always while the STi has a 35/65 planetary-type, and (b) the Evo has only three discrete settings which mostly alter the unlocking rate when you add steering input while the STi (in manual mode) allows you set the clamping pressure from always zero to always max (which is something crazy, like 15 bar).

There is an incredible amount of mis-information out there if you look around. You have people claiming that the torque split in an Evo changes as a function of which button you press and then others coming back and saying that the first guy posted the wrong numbers. It's funny if you're in the right mood. Really annoying if you're not. It gives you this image of these little Japanese guys inside the diff switching the gears around when a different button is pressed. Cute and all, but not really accurate. (They are, after all, little Mexican guys, instead, since nobody else would take the job.)

The key mistake that seems to be made over and over again (see almost any recent post by Dennis for an example) is this idea that a partially-active limited-slip device piggybacked on a 50/50 spider-type diff can usefully be thought of as producing such and such a torque split (other than 50/50). You are just going to mess yourself up if you follow this trend. It is much better to keep torque split separate from the torque-biasing effects of a limited-slip device. Torque split -- assuming that we are not talking about a 4-Matic-type differential -- is determined by the gearing. That's it. Don't think about whether you are on ice or whether some wheels are on ice and others are not. That will just mess with your head. All you need to look at are the gears. Our cars and Evos have equal gearing to the front and rear (within rounding, OK, Fedja?), so we have a 50/50 torque split. A 3000GT VR4 has a 60-tooth input ring ultimately driving a 27-tooth sun (with equal teeth on the two planets, so they can be ignored). So, since the sun is connected to the front, only 45% of the torque goes to the front. The (effective) output gears in an STi have almost a 2:1 ratio, instead, so it has a 35/65 split. Same goes for the Cusco center for our cars.

Separately, you can think about the torque-biasing ability of whatever limited-slip device is tacked onto the system. A VC can create such and such a bias, but a hydraulic clutch can produce a heck of a lot more. Plus, a hydraulic system can be turned on and off almost instantly, so, as long as you have a good computer with lots of good data to control it, you can bias torque in a much more efficient manner. And you can do it even when one output is reflecting zero torque, so it works when a Type-2 Torsen does not.

One company does seem to do a nice job of not confusing torque split with the torque-biasing abilities of a limited-slip device: Quaife. If you find a copy of the pretty graph I mentioned earlier, you will see that it does not give you a torque split as a function of locking percent. Instead, Quaife is always careful to only talk in terms of torque bias. Grok that and you are more than half-way home.

#5) How does automatic yaw control work?

Oh, please, GMFB! I have a day job, you know.

- Jtoby

 

[MrAWD]

For Fedja: Toby is a lost cause.

It comes very interesting that you see a need to explain to me (not to Toby) how those diffs are working. I gave you a comments on the stuff that you wrote earlier and wherever I though that you where off, I pointed that out. Where I agreed with you, I simply skipped in my reply. But, what I get from you is telling me that nothing particular of the things that I wrote where worth mentioning and you felt to explain how everything works again!! Or you actually agreed with everything that I said, but just felt to say some of it again! Very interesting!

Let see now where I don’t agree with the stuff that you wrote up there!

Power split is 100/50.

I take this just like a typo, so no biggy!

Now when we have the car on the limit, the faster shaft normally indicates slip at that end of the car - wheelspin, or incipiant wheelspin. If wheelspin develops, the open diff sends power to the faster turning shaft, making the wheelspin worse. Loss of acceleration.

Loss of acceleration only if the goal was to propel the car forward. If the goal was to rotate the car, spinning of the rear wheels only would be perfect if one has a wheel locked fully to one side and trying to do donuts!

So an open diff has a torque split of 50/50 (best case) and 100/0 (worst case) where the first number represents the end of the car with the least grip. At no time can an open diff send power towards the grippy end of the car.

This is a first spot where you are very wrong! 50:50 is not the best case! 100:0 or 0:100 is not the worst case!! The thing is that car needs to do what is has been told to do. If I am trying to turn with the car, 50:50 is pretty bad thing to have (assuming decent grip conditions) and in most cases will make the car to either understeer or just drift wide. You should try to see things in more then just one way, in order to avoid claims like this one above! The best way scenario you mentioned above is correct only if you are trying to go straight line only!

In order to get around this problem, there needs to be some sort of mechanism to slow down the shaft that is spinning too quickly. If that braking mechanism is connected to the opposite side diff output shaft, we get an effective power transfer from the fast shaft to the slow shaft.

Correct!

But the best this braking mechanism can do - no matter what the mechanism might be - is to lock the two output shfts together. At this point, the torque split is 50/50. A locking mechanism can only go to 50/50. Note that I'm not talking front/rear split here, but less grip/more grip.

Now you are back on the “best case scenario and 50:50”. The things are more dynamic in there that you might be seeing them. Lets assume, that you have a situation where one end of the car is spinning, basically 100:0 situation. The fact is that you don’t need 50:50 to solve this problem and here is why. When you locking mechanism starts with its job, more of the power is shifted to the side what was at zero(0) and you are taking power (spinning if you like) from the 100% side. Let say as a one of the transient state would be 10:90 distribution. Now let say the car was standing and with new distribution, it starts moving with that 10% side, which is getting all the torque. At that speed, spinning end actually slow down a lot relative to the surface. Partially due to the 10% less power being sent to the faster side (accompanied by the engagement of the slower side) and partially due to speed of the car and relative to the surface your spinning wheels are slower as well.

At the next moment in time, you get to the 25:75 split, which will make car to move even faster. I believe that somewhere around that spot speed difference between the surface and the 75% spinning side will be minimal and thanks to the grip in those tires, that side would hock up and there will be no need for 50:50 any more. Now take those numbers just as proofing concept and not what will exactly happen!

A sufficiantly "intelligent" locking mechanism, be it electronic or mechanical, "seeks" a 50/50 split. When it "detects" power "leaking" out a slipping output shaft, it attempts to apply a locking force to send power to the other output, up to the limit of 50/50.

No it doesn’t seek 50:50. You are mixing things here, mainly the purpose and the way to get there. Sufficiently "intelligent" locking mechanism is not there to lock the differential. It is there to make sure the power is distributed in the most efficient way for the current conditions.

A less sophisticated locking mechanism cannot react fast enough, or reacts in discrete steps (Detroit Locker), or cannot generate enough locking force to actually reach 100% locked, or otherwise underperforms our "perfect" locker. Such a diff might wind up being able to generate (say) a 60/40 torque split, but the high number is always towards the end with the LEAST grip.

You are correct here. A less sophisticated locking mechanism can not react fast enough, but also very important is the fact that they have no control of the locking. Being able to go “only” up to 40:60 is actually not that bad, as long as they could have get there sooner!

So to review: front end of the car on ice, rear on grippy pavement. Open diff, 100/0 torque split. "Perfect" locker, 50/50 split. Imperfect locker, somewhere between 100/0 and 50/50.

Wrong as explained above…

And yes, sufficiantly advanced diff controllers can take in info from other sensors and act on that information to provide handling changes simply not possible on a mechanical diff. for example, an WRC controller looking at steering angle and "hand brake" position might decide that a tight turn was being called for, and route power to the outside wheels only, pivoting the car around like an APC.

Actually, all of the ACD equipped EVOs have that hand brake feature. Except, it doesn’t work the way you explained above. When you pull the hand brake, ACD goes into the full open mode and allows locking of the rear wheels with the power still delivered to the front. If the power would go to the outside wheel, then hand brake would lock only one of the rear wheels, which is not correct.

The important part to understand about the EVO's system is that it can generate torque splits between 100/0 to 50/50, where the largest portion of the split is ALWAYS towards the wheel with the least grip.

Correct! But the good thing about this is that least grip side is in most cases the rear end!!

Because weight transfers rearward under acceleration, this means that there will always be a movement towards understeer with this kind of diff. If we assume the system works perfectly, the amount of understeer will be dramatially less than with an open diff, and some degree less when compared to a mechanical diff, depending on how good the locking mechanism on the mechanical diff is - where the Quaife is probably the best. (and is very, very good)

Again, I asked you were do you think that weight gets transferred on your car while cornering, and you never answered. ACD on EVO goes into more locking mode when car is doing straight line acceleration so all of the wheels get their share of pavement. For the cornering phase, diff is more open and allows rotation of the rear end, because that side still gets much less of the weight then the outside front.

Your Quaife will do a great thing for the straight line and it will keep those wheels with the pavement. Cornering phase is where is fails to be that good and fails to send the power where you really need it!


Fedja

 

[DG-FNR]

Dennis of since he keeps making the false claim that it is impossible to get rid of power-on push in a DSM.

JToby selective editing strikes again.

It is impossible to eliminate power-on understeer in a DSM USING A STANDARD LOCKING DIFF.

Charles' unit is NOT a standard locking diff, so that statement DOES NOT APPLY TO IT.

The stock diff, the Speed Design 4-spider (both which use the OEM VC as the locking mechanism) the Quaife, and the various clutchpack diffs are all lockers, of varying efficiancy. They ALL cause power-on understeer, with the degree of understeer depending on the locking mechanism's sophistication and configuration - Quaife being the best, and a poorly configured KAAZ probably being the worst.

Now the issue with Charles' unit is that it has NO integral locking mechanism; it instead relying on the OEM VC to provide locking function. That means it is going to suffer from the same problems as the OEM VC - a delay in how quickly the locking force develops, followed by a sudden ramp-up as the VC enters "hump mode". In a nutshell, the VC needs to slip for a certain amount of time before it starts developing a locking force, and then it hits hard all at once - not quite as bad as a Detroit locker, but a similar (if less harsh) effect.

I actually proved this out (unintentionally) on a chassis dyno. An unwelded VC was fitted (actually, a very poorly welded VC that immediately fractured the weld and resumed normal operation) When the pull started , we got what sounded like clutch slip that slowly started to go away, and then suddenly was (almost) completely gone. And yes, it trashed the center diff bushings.

What practical effect this locking characteristic has on the handling of a car equipped with a biased diff I don't know, as I haven't driven a car so equipped. I HAVE driven powerful DSMs that use the OEM centre diff though, and I know what they do - nasty power-on understeer that suddenly jerks the nose had toward the centre of the turn when the VC locks and finally sends signifigant power rearward. It is unsettling and unpredictable, very difficult to drive, and gets worse as power levels increase. One of the primary benefits of the Quaife centre diff is that the car responds very predictably and can really be leaned on.

If a biased centre diff could be mated to a locking mechanism as progressive as a Quaife, we'd have a real winner on our hands.

DG

 

[jtmcinder]

Fedja just made two very strong assertions about the ACD in an Evo. He claimed that the ACD unlocks when the handbrake is pulled and he claimed that the ACD unlocks (at least to some extent) when steering input is added.

These are both correct.

- Jtoby

 

[jtmcinder]

It is impossible to eliminate power-on understeer in a DSM USING A STANDARD LOCKING DIFF.

Charles' unit is NOT a standard locking diff, so that statement DOES NOT APPLY TO IT.

The problem is not selective editing (although I'd be happy to post links to several places where you have made the claim in question). The problem is your complete inability to listen and/or understand what I say.

How you get a limitation of slip and how torque is split are best thought of as separate questions. Charles does have a standard device for limiting slip. It's an OE VC. What Charles has changed is his differential. It is not a 50/50 spider-type; it is a 35/65 planetary-type. Neither a spider- nor a planetary-type differential has any ability at all to limit slip. Read that and understand it. Neither of them is an LSD. You have spent too much time eating, sleeping, and drinking Quaife. What makes a stock DSM center an LSD is the piggybacked VC. What makes Charles' center an LSD is the piggybacked VC.

- Jtoby

 

[jtmcinder]

If a biased centre diff could be mated to a locking mechanism as progressive as a Quaife, we'd have a real winner on our hands.

As I've tried to explain, a Quaife cannot be anything other than a 50/50 torque split. However, if you open up to the idea that a computer-controlled hydraulic clutch can be at least as good as a Quaife in terms of limiting slip at the right times, then you just wrote an ad for the STi, because that is what it has: a rear-biased split with limited slip.

But guess what? You be wrong and just another shill for Scooby. Ask anyone who has driven both an STi and an Evo which works better.

- Jtoby

 

[MrAWD]

It is impossible to eliminate power-on understeer in a DSM USING A STANDARD LOCKING DIFF.

Hmm, I read somewhere that if you think it is impossible, it is possible that you are not thinking!

Now the issue with Charles' unit is that it has NO integral locking mechanism; it instead relying on the OEM VC to provide locking function. That means it is going to suffer from the same problems as the OEM VC - a delay in how quickly the locking force develops, followed by a sudden ramp-up as the VC enters "hump mode". In a nutshell, the VC needs to slip for a certain amount of time before it starts developing a locking force, and then it hits hard all at once - not quite as bad as a Detroit locker, but a similar (if less harsh) effect.

Actually the fact that OEM VC takes a bit time to get into more locking as a great thing!!

What practical effect this locking characteristic has on the handling of a car equipped with a biased diff I don't know, as I haven't driven a car so equipped. I HAVE driven powerful DSMs that use the OEM centre diff though, and I know what they do - nasty power-on understeer that suddenly jerks the nose had toward the centre of the turn when the VC locks and finally sends signifigant power rearward. It is unsettling and unpredictable, very difficult to drive, and gets worse as power levels increase. One of the primary benefits of the Quaife centre diff is that the car responds very predictably and can really be leaned on.

Now I understand where your errors are coming from! You think that center diff needs to start locking in order to stop oversteer. But, the reality is quite opposite in here! With welded diff, car will try to go straight all the time and it would oversteer like crazy. Open center diff would give you something that would not like to put the power down, but it would rotate like crazy!

So, that “nasty power-on understeer that suddenly jerks the nose had toward the centre of the turn when the VC locks and finally sends signifigant power rearward” actually sends your nose outside of the turn and car starts to push!! All this time you had it backward, old man!


Fedja

 

[DG-FNR]

This is a first spot where you are very wrong! 50:50 is not the best case! 100:0 or 0:100 is not the worst case!!

OK Fedja, I see where we're talking at cross-purposes here.

When I say "50/50 is the best case" and "100/0 is the worst case" I'm talking about a POWER TRANSFER perspective, not a vehicle dynamics perspective.

Perhaps I could substitute "maximum" for "best" and "minimum" for "worst".

As in "when considering a loss of traction at one end of the car, the MAXIMUM a locking diff mechanism can do in terms of sending power to the non-slipping end is to send 50% of the power to the end that grips, and the MINIMUM it can do is to send 0% to the gripping end. At NO time can it exceed 50% power transfer to the gripping end of the car"

Is that clearer?

Otherwise, it sounds like we're pretty much in agreement everywhere, with the exception that you're talking about a theoretical ideal power distribution while cornering, and I have not discussed that.

We also agree that, if you have an electronic diff controller, and the program has enough inputs from various sensors, a well-done calibration, and actuators with enough speed, strength, and finesse, that you can make the power distribution do anything you want, which includes crazy things categorically impossible with purely mechanical systems, eg, the WRC car drivelines.

OK, so now let's talk cornering dynamics. These are artificial situations.

Let's assume a steady-state corner taken at half the car's total cornering capacity. Let's also assume a nose-heavy weight bias. Given that we are steady-state, we have no longnitudnal weight transfer, and so (all else being equal) we have unused grip capacity on both ends of the car, with some amount more on the front than on the rear.

Agreed?

We gently roll in the throttle, using up half of the reserve grip capacity on each end of the car (still well below the limits) Because we are accelerating, we get longnitudnal weight transfer; the fronts lose some grip; the rears gain it. Because neither end of the car is slipping, the diff splits power 50/50, but the locking mechanism is not in play - it's just being a normal diff.

Agreed?

We continue to add throttle. As we do, we use up more of the remaining reserve grip capacity in the tires, and we continue to transfer weight (and thus grip) rearward.

Think of a pair of buckets, each bucket representing the grip capacity of that end of the car. We are filling each bucket with a hose (flowing at the same rate) that represents power being applied - and note that this is true in both a straight-line and a cornering case. The "front" bucket is getting smaller, and the "rear" bucket is getting bigger - this is the effect of longnitudnal weight transfer.

Agreed?

At some point, one of the buckets is going to overflow (representing a tire/end of the car that can take no more power) Because the front bucket is shrinking and the rear bucket is growing, it is almost (but not quite) certain that the front bucket will overflow first, causing understeer.

Agreed?

OK, so let's assume that the front end is the end that runs out of grip first. We have power-on understeer. In a car with an open centre diff, the front side output shaft starts rotating faster than the rear side, power starts going forward (as it must, we have no locking mechanism) the fronts spin up, and we have 100/0 power split and massive terminal power-on understeer.

Agreed?

Same conditions, but now with a viscous coupling locking mechanism. Fronts run out of grip. Power starts routing forward, but the slip across the plates in the VC starts to produce locking force - a little bit at first, then a whole lot as the VC humps and really locks. We get understeer plus a loss of thrust as power is sent forward where it does no good - bad at first, but improving with time as the VC starts locking.

Agreed?

Same conditions, Quaife centre. We assume that the Quaife works as perfectly as Quaife claims (which I don't think is completely true, but it's probably pretty close) Fronts run out of grip, Quaife reacts instantly, and is able to keep 50% of the power going to the rears. We still get understeer, but at least we never lose rear thrust .

Agreed?

Same conditions, computer-controlled locker (in the centre slot only) Because it is a locker, it can only do the same sort of thing as a Quaife. If we assume that the Quaife is perfect at sensing slip and rerouting, then the electronic locker can only be as good as the Quaife. If we assume that the Quaife is somewhat less than perfect at sensing slip and rerouting, the the electronic locker can beat the Quaife - but the Quaife is very good, so there's not a huge amount of room for improvement. And the electronic version needs to have calibration, actuators etc fast enough - which is possible, but not a given.

So depending on how good the electronic locker is and how good the Quaife is, the electronic locker may be able to keep thrust from falling off better than the Quaife - but it still understeers.

See?

Charles does have a standard device for limiting slip. It's an OE VC. What Charles has changed is his differential. It is not a 50/50 spider-type; it is a 35/65 planetary-type. Neither a spider- nor a planetary-type differential has any ability at all to limit slip. Read that and understand it. Neither of them is an LSD. You have spent too much time eating, sleeping, and drinking Quaife. What makes a stock DSM center an LSD is the piggybacked VC. What makes Charles' center an LSD is the piggybacked VC.

That is what I have been saying all along. I have never claimed otherwise. If you think otherwise, you are (as usual) in error.

DG

 

[JiveMasterT]

i dunno about you guys, but i had junky tires and my car tended to oversteer a lot. now i have new all season tires and it still oversteers a little bit sometimes, but mostly when my gas tank is somewhat low.

it is rare that the thing doesnt drive like it is on rails though. usually when i push it through turns i end up oversteering... it might just be how i drive though. i used to be able to get my awd astro van sideways as well, and im probably just used to driving a certain way or something.

 

[jtmcinder]

As in "when considering a loss of traction at one end of the car, the MAXIMUM a locking diff mechanism can do in terms of sending power to the non-slipping end is to send 50% of the power to the end that grips, and the MINIMUM it can do is to send 0% to the gripping end. At NO time can it exceed 50% power transfer to the gripping end of the car"

Wrong. A fully-locked center sends 100% of the power to the rear if the fronts suddenly have zero grip.

Step One to understand why is to see that the fronts having zero grip is the same thing as breaking both front axles at the same time.

Step Two to understanding why is to see that breaking both front axles at the same time is the same as suddenly having the front output from the center diff strip all its splines.

Step Three is to see that a fully-locked center diff that happens not to be connected to the front wheels is the same as rear-wheel drive.

Step Four is to remember that rear-wheel drive is 0/100.

Drink, Horsey.

- Jtoby

 

[MrAWD]

OK Fedja, I see where we're talking at cross-purposes here.

When I say "50/50 is the best case" and "100/0 is the worst case" I'm talking about a POWER TRANSFER perspective, not a vehicle dynamics perspective.

Perhaps I could substitute "maximum" for "best" and "minimum" for "worst".

As in "when considering a loss of traction at one end of the car, the MAXIMUM a locking diff mechanism can do in terms of sending power to the non-slipping end is to send 50% of the power to the end that grips, and the MINIMUM it can do is to send 0% to the gripping end. At NO time can it exceed 50% power transfer to the gripping end of the car"

Is that clearer?

I do understand what are you trying to say here and I gave you already that you are correct with the stuff above. What you don’t get is the effect that diff has to the handling of the AWD DSM (or EVO due to similarities), which is what we are discussing here!

And since we are on the handling subject, your affection for the power delivery goes ahead of the turning ability of the car. The point here is to go the fastest way from the point A to the point B, which are at the beginning and the end of the corner respectively! But, you are using a straight line analogy to explain what is happening in the corner, and you are missing a few things here and there!

Otherwise, it sounds like we're pretty much in agreement everywhere, with the exception that you're talking about a theoretical ideal power distribution while cornering, and I have not discussed that.

We are agreeing on a few things here, but I am not talking about the theoretical ideal case, but rather the way things actually are!

We also agree that, if you have an electronic diff controller, and the program has enough inputs from various sensors, a well-done calibration, and actuators with enough speed, strength, and finesse, that you can make the power distribution do anything you want, which includes crazy things categorically impossible with purely mechanical systems, eg, the WRC car drivelines.

We agree on the benefits of the electronic control of the center diff, but I was not talking about the WRC cars at all. My ACD example was about what is going on with the EVO that we have here in the US!

Let's assume a steady-state corner taken at half the car's total cornering capacity. Let's also assume a nose-heavy weight bias. Given that we are steady-state, we have no longnitudnal weight transfer, and so (all else being equal) we have unused grip capacity on both ends of the car, with some amount more on the front than on the rear.

No, you don’t have the same amount of grip front vs. rear. There is much more weight in the front and that gives you more grip!

We continue to add throttle. As we do, we use up more of the remaining reserve grip capacity in the tires, and we continue to transfer weight (and thus grip) rearward.

Some of the weight goes toward the back, but the majority of the weight still stays at the outside front wheel. Draw some vectors and include centrifugal and longitudinal sides into the picture and you will see that centrifugal side in the front is still the most important one!!

On the same subject, did you ever spin outside front while cornering? Which is the first wheel to starts to spin in the middle of the corner? Mind you that things are different at the exit phase, but even then you don’t spin the outside, but rather inside front!

Think of a pair of buckets, each bucket representing the grip capacity of that end of the car. We are filling each bucket with a hose (flowing at the same rate) that represents power being applied - and note that this is true in both a straight-line and a cornering case. The "front" bucket is getting smaller, and the "rear" bucket is getting bigger - this is the effect of longnitudnal weight transfer.

Well, you used buckets and I used vectors. Fine! If the amount of water in your bucket is the grip that you have on each of the wheels, you would see that front bucket has much more water since the beginning. A little bit of the longitudinal weight transfer will not change that amount as much as you think it will!

At some point, one of the buckets is going to overflow (representing a tire/end of the car that can take no more power) Because the front bucket is shrinking and the rear bucket is growing, it is almost (but not quite) certain that the front bucket will overflow first, causing understeer.

That is where you are mistaken! Rear end is the one that will loose the grip first and during the heavy cornering you will never be able to create that much longitudinal force to even those buckets. What you are explaining is the straight line acceleration, and you are completely forgetting the effect of the centrifugal components!!

OK, so let's assume that the front end is the end that runs out of grip first. We have power-on understeer. In a car with an open centre diff, the front side output shaft starts rotating faster than the rear side, power starts going forward (as it must, we have no locking mechanism) the fronts spin up, and we have 100/0 power split and massive terminal power-on understeer.

Charles’s car equipped with Cusco unit at the beginning without the VC unit never spin the front wheels!! That is basically an open diff with different ratios. Again, you don’t loose the grip on the outside front!!

Same conditions, but now with a viscous coupling locking mechanism. Fronts run out of grip. Power starts routing forward, but the slip across the plates in the VC starts to produce locking force - a little bit at first, then a whole lot as the VC humps and really locks. We get understeer plus a loss of thrust as power is sent forward where it does no good - bad at first, but improving with time as the VC starts locking.

Here you go backward again! When VC starts to lock, thing actually get worse and car starts to push more.

Same conditions, Quaife centre. We assume that the Quaife works as perfectly as Quaife claims (which I don't think is completely true, but it's probably pretty close) Fronts run out of grip, Quaife reacts instantly, and is able to keep 50% of the power going to the rears. We still get understeer, but at least we never lose rear thrust .

You are right about the Quaife and getting understeer and better power delivery. The reasons on why that happen are wrong tough. That instant reaction of Quaife gives you that understeer, because it is doing only thing that it knows how. But that action overdrives your front wheels and you get understeer. If more of that power would stay in the back, car would have a chance to oversteer, but that is a dream story for the cars with mechanical diffs (even Quaife is mechanical)!

Same conditions, computer-controlled locker (in the centre slot only) Because it is a locker, it can only do the same sort of thing as a Quaife. If we assume that the Quaife is perfect at sensing slip and rerouting, then the electronic locker can only be as good as the Quaife. If we assume that the Quaife is somewhat less than perfect at sensing slip and rerouting, the the electronic locker can beat the Quaife - but the Quaife is very good, so there's not a huge amount of room for improvement. And the electronic version needs to have calibration, actuators etc fast enough - which is possible, but not a given.

So depending on how good the electronic locker is and how good the Quaife is, the electronic locker may be able to keep thrust from falling off better than the Quaife - but it still understeers.

The main point that you are missing here is that locking of the center diff in the middle of the turn will create understeer, not oversteer as you believe. Once you accept that, and part of the corner weight transfer, we would agree on more things! Active diff can do both things well, while mechanical diffs can do only one thing right, and that one hurts the other unfortunately!


Fedja

 

[MrAWD]

Wrong. A fully-locked center sends 100% of the power to the rear if the fronts suddenly have zero grip.

Actually, you are wrong here!! With 100% locking power goes equally to the both ends all the time. The torque on the zero grip side is zero. 100% of what is possible is on the other side, which is much less from the 100% of the torque that engine is producing at the time! If 100% power would go to the rear, front wheels would not spin, but they do!

Step One to understand why is to see that the fronts having zero grip is the same thing as breaking both front axles at the same time.

No, wrong again!! The power still goes to the both ends and has to spin all four wheels equally. Breaking the front axles would just give you a RWD car, and that is not what you have!


Fedja

 

[jtmcinder]

Hey, Folks, have you ever disconnected your transfer case, added a spool, and used a 2-wheel dyno? Well, if you have, I have some great news for you. According to Dennis Grant, since a locked center can only send 50% of the power to the wheels with grip, you only were measuring 50% of your power. You actually have twice the HP and torque that the dyno told you. Therefore, I must say, you guys with 1400 WHP DSMs totally rock! Although I have to wonder why your traps are only in the 120s. Can any of you drive?

- Jtoby

ps. if citing Dennis Grant as your authority doesn't help, it looks like you can add in Fedja

pps. you guys really need to get a grip (as it were) on the fact that sending power to an end of the car that isn't using it doesn't make it evaporate. A locked center tries to send half the power in each direction (because it turns the tires the same speeds), but if the tires at one end have no grip at all, then they use none of the power (ignoring drivetrain losses), giving you an infinite amount of torque bias.

 

[MrAWD]

Hey, Folks, have you ever disconnected your transfer case, added a spool, and used a 2-wheel dyno? Well, if you have, I have some great news for you. According to Dennis Grant, since a locked center can only send 50% of the power to the wheels with grip, you only were measuring 50% of your power. You actually have twice the HP and torque that the dyno told you. Therefore, I must say, you guys with 1400 WHP DSMs totally rock!

Hmm, I read all of the stuff that was said up there, and somehow I missed where did Dennis say things that you claim up there?? This looks like that you took some of the stuff out of what he said and converted it into something else, but still managed to blame Dennis for it.

ps. if citing Dennis Grant as your authority doesn't help, it looks like you can add in Fedja

You definitely have strange ways of putting things together…

pps. you guys really need to get a grip (as it were) on the fact that sending power to an end of the car that isn't using it doesn't make it evaporate. A locked center tries to send half the power in each direction (because it turns the tires the same speeds), but if the tires at one end have no grip at all, then they use none of the power (ignoring drivetrain losses), giving you an infinite amount of torque bias.

Yes, they use none of the power if there is no grip, but the power is still sent that way! And because they use none of the power they have zero torque, but power still exist!! And that is different from no having the power there at all!


Fedja

 

[jtmcinder]

If you think that a fully-locked diff can't send 100% of the power to the rear (or front), then you don't get it. But you probably will be happy to hear that I will only try one more time to explain it before giving up.

Before getting started, however, let's be clear that it does not matter whether it's a 35/65 or a 50/50 torque-split diff. We're talking about situations where the limited-slip device is completely locking the outputs together, so the gearing inside the diff is irrelevant. Whether or not even an Evo or STi hydraulic clutch has the ability ti fully lock the diff at WOT is also irrelevant; we are assuming that it can to make a theoretical point.

A fully-locked diff forces the front and rear wheels to turn at the same speed. So if there is no wheel slip (or, at least, if wheel slip is equal front and rear), the amounts of torque going to front and rear are the same. Remember the trick that Quaife taught us with regards to determining the torque distribution: It is the same as how much torque is being reflected. So if the fronts suddenly have zero grip (for whatever reason), only the rear will be reflecting torque, so only the rear is getting any torque. This is probably best thought of as an infinite torque ratio or bias, but if you want to call it a 0/100 split, knock yourself out.

I am not going to get involved in your discussion of how the differential and limited-slip device interact with weight transfer to produce handling. We are clearly not ready for such a conversation. I am only concerned with correcting the misconceptions that many people have with regard to how a differential and limited-slip device work together to create a torque distribution. And as long as anyone is claiming that the most that a fully-locked diff can do is send 50% of the available engine torque to the end that has grip, we have work to do.

- Jtoby

 

[MrAWD]

If you think that a fully-locked diff can't send 100% of the power to the rear (or front), then you don't get it. But you probably will be happy to hear that I will only try one more time to explain it before giving up.

I just can’t believe that you are so stuck with the wording that you use and can’t see anything else in here
It is more then obvious that solid connection that you get with fully-locked diff sends power equally everywhere (read both ends). You in the way get 100% of power that you have on both sides.

But, that 50:50 comes from the difference in the speeds of the axles (or two sides of the diff). With a full lock, you always get the same speed on both axles. Now if you are trying to play smart and mix those terms for whatever reason you have, the only thing that you will get is unnecessary confusion!

A fully-locked diff forces the front and rear wheels to turn at the same speed. So if there is no wheel slip (or, at least, if wheel slip is equal front and rear), the amounts of torque going to front and rear are the same.

You see, mixing things again and even got your self confused. Torque is something that actually does the work, while power is just ability for something to do the work. If the wheels on one side have no grip there is no work done on that side (lets ignore all of the losses along the way), so your torque amount is zero! But even though there is no torque present, the power is still available on that end. If diff is fully locked all of the power is available on that end too! So, be careful when using different terms since they have quite a different meaning to things said here!

I am not going to get involved in your discussion of how the differential and limited-slip device interact with weight transfer to produce handling. We are clearly not ready for such a conversation. I am only concerned with correcting the misconceptions that many people have with regard to how a differential and limited-slip device work together to create a torque distribution. And as long as anyone is claiming that the most that a fully-locked diff can do is send 50% of the available engine torque to the end that has grip, we have work to do.

Well, if you are here for correcting the misconceptions that gets created, try to stay on the same side of things. Your first sentence is talking about the percentages of the power, while the last one is using torque instead. Unfortunately, you are using the same logic in both cases and that just creates even more of the confusion and misconceptions.


Fedja

 

[Greg Collier]

The Infinite Thread.....................round And Round It Goes, Where It Stops No Body Knows.....................................
So Is Space-time Or Time-space?

 

[Revolution]

OK so in the end I want to upgrade to a planetary type LSD with an evo like piggy back supercomputer bolted onto it ,right? This is what will make my dsm handle amazing right? After all that reading this is what I want right? Some one just tell me is theirs something outtheir like it or do we only have the quafies and the kaaz's.


Ok on a not so goofy note, couldnt we make OUR cars the dsm's react or act a certain way in the three variances of a corner by playing with the lsd's? Say a 1.5 front a 1.5 center and a 2 way in the rear ? Im not actually stating to use it in this exact order but by utilizing a different lsd's couldnt we minimize understeer or induce over steer.
Im laying this question out their in the hopes some one could give a rational answer.