Quaife vs. equal length halfshafts

[dsm-onster]

Swapping in awd ELHSs is not a big deal to me since I have the smaller halfshaft width. and I'm a 1G. Will it help if I have and ATB differential?, is my question.

ELHS supposedly discourages torque steer. I understand the theory jsut fine. . . The "type-2" automatic torque biasing differentials (quaife, karking, etc.) "lock as a function of the difference in the amounts of torque being transmitted to each of the outputs". Which negates torque steer also.

First,
1. I'm getting terrible torque steer. FWD, 60-1, 30 psi, cams, go figure . Does the onset of torquesteer decrease my traction? I notice that I'll pull to the left then I'll correct and BOOM the tires go up in smoke!
2. If I get ELHS, it should negate this regardless. If I get an ATB diff, then I should negate this. BUT, will the ELHS still assist in traction even w/ and ATB diff? Or does the ATB diff just take over for unequal halfshafts and make it like they wern't there no matter what the torque on each shaft is?
3. If this is the case, wouldn't this cause a clutch diff like the KAAZ to wear earlier? ELHS would negate some torque differentiation from even occuring.

Thanx for your clarification. . .

 

[jtmcinder]

A Torsen locks as a function of the torque that is reflected all the way back to the diff. This suggests (to me) that you'd want the the ELHSs, even if you had a Torsen.

With that said, for drag-racing, you might be better off with a modified clutch-pack, such as a Kaaz 1.5-way. Better to have the thing locked in advance than react to a difference in reflected torques.

- Jtoby

ps. interesting question; I hope this turns into a useful thread

 

[dsm-onster]

: I was really hoping you'd reply first. Your stickies on diffs and LSD are excellent. . .

A Torsen locks as a function of the torque that is reflected all the way back to the diff. This suggests (to me) that you'd want the the ELHSs, even if you had a Torsen.

Would going w/ a clutch style LSD then mean that it doesn't matter what the length of my shafts are or what weight is on each tire? Because, both wheels are GOING to be engaged anyway, and 50/50. . .

 

[dsm-onster]

I'm interested in a KarKing or Quaife as a consideration for a DD. It seams a Quaife is much better for the street (no different fluids, no 1 million figure 8s for a break in, no 80K rebuild, i want to keep my car forever).

A quaife locks based on the amount of TRANSMITTED torque to the spider gear, right? I'm assuming that a quaife doesn't react instantly to suddenly different tranferals of torque. When it does, Does it really lock? Both wheels move at the same rate w/ the same torque, right? These are probably dumb-dumb questions. But you said torsen locks as funtion of the reflected torque. To me, this means that when there's suddenly a loss in torque on one side, the torsen doesn't let any more be lost than when it senses it, yet torque is still devided unevenly. Or should I assume that it LOCKS when a sudden loss is sensed?

 

[jtmcinder]

Please read my stickies again. Helicals don't have spider gears. Even more: saying that a locked diff produces a 50/50 split is one of my biggest pet peeves. A locked diff does not have a meaningful torque split: it merely turns the two outputs at the same speed such that the torque split depends entirely on the amounts of grip enjoyed by the two outputs.

By my understanding of the causes of torque-steer, you would still have this problem (albeit with greatly reduced side-effects) after swapping in a front LSD (of any sort). Personally, if I were in your position, I'd do this: get ELHS and do all the other, simple stuff, such as motor mounts and (front) suspension bushings. Then see where you are. I say this because I'm very cautious about spending money -(after all, I have a DSM!)- and I like to see what each mod does before adding the next.

[edit: WRT a question in your second post: helicals are limited to a certain ratio of output torques. If one wheel suddenly has no grip and reflects almost nothing (e.g., over-powering at corner exit when autocrossing a wrong-wheel-drive), then the diff can effectively snap to open (since zero times anything is still zero) and all you get is a lot of smoke from the wheel with no grip. This is why I suggested a modified clutch-pack for you, instead.]

- Jtoby

 

[dsm-onster]

Pardon my stupidity. . .I'm just trying to understand the effect of a type-2s (helical) ATB differential has on both of the output speeds.

EDIT: DING [light bulb]! Have you been trying to say the whole time for ATB LSDs that as long as both sides show equal resistance then they are free to rotate at different speeds, like when going around a turn?

Another edit: If that is the case then, I can rap on the brakes (while on the gas) and still eliver torque to the ouside wheel if I'm in a hard turn caousing the inside wheel to rise, correct?

 

[jtmcinder]

Pardon my stupidity. . .I'm just trying to understand the effect of a type-2s (helical) ATB differential has on both of the output speeds.

No problem. Most well-known autocrossers who own them don't know how they work.

EDIT: DING [light bulb]! Have you been trying to say the whole time for ATB LSDs that as long as both sides show equal resistance then they are free to rotate at different speeds, like when going around a turn?

Yes! That's the beauty of a helical: it does not lock until it needs to.

Another edit: If that is the case then, I can rap on the brakes (while on the gas) and still eliver torque to the ouside wheel if I'm in a hard turn caousing the inside wheel to rise, correct?

Yeppers. Another good reason to learn to left-foot brake, although this all works much better in an AWD, since LFBing when powering out of a turn in a FWD will cause understeer as the tail holds you back.

- Jtoby

.

 

[dsm-onster]

Hmm, doesn't my current un-equal length half shafts provide unequal resistance between the two shafts? Thus the torque steer? Of course SEVERAL other things contribute to torque steer. BUT, since the torsen (torque sensing) differential can react to this difference, as it seams, instantaneously, then all imbalances in "reflected torque" are resisted by the ATB diff regardless.

As long as the tires convey a resistance, then no matter how the axles react to the torque, there will still be load. Just because the halfshafts are unequal, my weight over each tire is unequal, & etc. doesn't mean that there is zero resistance on one side. Therefore, the quaife will act. I still don't see how it could NOT act instantaneously, so does it matter what imbalance there is in resistances between the wheels caused by unequal length halfshafts? It's going to bind as soon as you start applying enough torque to see an inbalance in resistance between the axles due to differing lengths.

 

[jtmcinder]

I think what's happening is that you are buying into the name "automatic torque biasing" too much (which isn't your fault; it's the one thing about Quaife I don't like). The ATB label makes it sound like the thing is almost magical: sending torque where it can be used, taking it away from outputs that don't have grip. (It's as bad as the Subaru ads, in fact.)

That isn't what a Quaife does. Like all LSDs, it locks. That's it. What differs between LSDs is why and when they lock and, also, how and how hard they lock; but all LSDs have their effects by locking. End of story.

I guess you can say that a locked diff (i.e., a non-diff; a spool) is automatically biasing torque to the output with more grip, but please note that it is doing this because it has ceased to be a diff, so it's still dishonest to say that the diff is automatically biasing torque.

Example application of this to your issue: assuming that the diff is locked (but the amounts of grip are equal [or, at least, sufficient]), the difference between the half-shafts will cause more torque to be delivered to the ground on the shorter side. You know those torque-limiting gizmos that tire places use to keep employees from over-torquing lug-nuts? Well, you got one of those on your driver's side. I'd get rid of it. (I've already got a major torque-limiting device known as "Toby's mad tuning skillz"; if you have anything like this, I wouldn't also have a second such device hooked up to one - and only one - axle.)

- Jtoby

 

[dsm-onster]

Sir, thank you for your discusion w/ me so far. I REALLY want to understand this. The method is in the details and I know there IS a method to the madness I see. If I don't get the details, I'm just displaying madness.

I think what's happening is that you are buying into the name "automatic torque biasing" too much (which isn't your fault; it's the one thing about Quaife I don't like). The ATB label makes it sound like the thing is almost magical: sending torque where it can be used, taking it away from outputs that don't have grip. (It's as bad as the Subaru ads, in fact.)

I think so to. I was straying from the concept that a quaife locks and was wondering why Mitsubishi would run an AYC diff if a torsen does the same thing and is completely mechanical. A torsen simply locks so that which ever wheel has the most traction can "enjoy" the most torque.

That isn't what a Quaife does. Like all LSDs, it locks. That's it. What differs between LSDs is why and when they lock and, also, how and how hard they lock; but all LSDs have their effects by locking. End of story.

I'm not getting any meat on when and how hard the different systems lock.

I guess you can say that a locked diff (i.e., a non-diff; a spool) is automatically biasing torque to the output with more grip, but please note that it is doing this because it has ceased to be a diff, so it's still dishonest to say that the diff is automatically biasing torque.

The only way a torsen system delivers torque to the wheel w/ more traction is by temporarily becoming a locked "non-diff", and there's nothing more, correct? It just locks so that the tire w/ the most traction CAN use all of the torque. As soon as the "resistance torque" equalizes then it returns to a state of being an open differential.

Example application of this to your issue: assuming that the diff is locked (but the amounts of grip are equal [or, at least, sufficient]), the difference between the half-shafts will cause more torque to be delivered to the ground on the shorter side. You know those torque-limiting gizmos that tire places use to keep employees from over-torquing lug-nuts? Well, you got one of those on your driver's side. I'd get rid of it. (I've already got a major torque-limiting device known as "Toby's mad tuning skillz"; if you have anything like this, I wouldn't also have a second such device hooked up to one - and only one - axle.)

Believe me, I'll take your word for it and do away w/ the long driverside torquewrench . But I'll refer to my first paragraph: it's just madness unless you understand .

So, I know a way to insure that one tire does not break loose before another is by adding ELHS. A short half shaft can deliver more torque than a long halfshaft before twisting.

I'll add what you've taught me:

SO, though a torsen is "locked", if the shafts are unequal in length (thus one twists during a launch), THAT one doesn't deliver as much accelerating force as much as the other wheel/axle. Though there is a lock at the diff, there is not an equal amount of torque delivered THROUGH the wheels to the ground. Torque steer remains and the car shoots off to an angle from the line of the track.

Or the other consequence is that the short shaft wheel breaks loose earlier as the shaft is transmitting more torque during the time the other shaft is twisting. And then when the long shaft wheel "catches up", it has a buddy w/ no traction and has to do all the work itself. I know this would happen during a fraction of a second. So to the rest of the world it would seam that the PAIR of tires are breaking loose at the same time. But I'll bet looking at the burnout marks, one will begin before the other. . .

 

[dsm-onster]

DING! [another light bulb] Why then is a KAAZ (salisbury) limited slip device best for straight line acceleration?

Is this why?:

A torsen just doesn't start to lock before any wheel slip occurs. Period. A torsen locks when the amounts of torque used by the outputs differ. If there is a miniscule difference in the traction, that would lead to one slipping before the other. Considering doing all you can to a FWD for traction would mean that all that's left is: will the tires be enough considering the weight on them (after transfer)? This is the reason why you suggested doing all the FWD traction mods I can do first [insert "duh!" smiley].

Now, uncontrollable circumstances (minute differences in the compound between the two tires, small surface elevations on the track surface, tiny amounts of foreign material on the track surface, and on and on) would cause one tire to slip. Since one wheel just starts to slip before lock, then what is the other wheel doing? Isn't the quaife acting like an open diff just as slip on one wheel is occuring? So the wheel that is not slightly slipping is loosing torque, until the quaife locks, correct? However quick it is, it's still not instantaneous, right?

Conversely, an "always on" KAAZ, doesn't have that brief moment where the tire WITH traction looses torque. It's totally different. A wheel rolling through less traction moves at the same speed as the other wheel; and that other wheel, in effect, "gets" more torque. The sum of the two torques at the outputs equal the torque of the input. There is no reaction time. As soon as the wheel w/ less traction rolls away from that condition, the torques on the two wheels are equal again.

Kind of like my lsd insert but much more reliable (not using excessive and improper force on the poor stock diff internals) and at least a bit more predictable in the curves.

See! I could have just taken your advice. But, you would have never gotten the satisfaction of seeing me understand . Nor would I be able to share the advice w/out raising alot of questions. . .

But does a torsen's reaction time really make that much difference? A viscous LSD would for sure.

 

[pneumo]

You seem to be asking a lot of good questions!

I used to have a Quaife in my fwd Laser. I'm the curious type, so I played around with it to see how it worked. One day I jacked up one front wheel and spun it to observe the Quaife in action. Since the other wheel was on the ground the lsd action kicked in. The raised wheel turned about 1/4 turn before locking. The tranny has a little bit of backlash, so not all of that 1/4 turn was from the diffs action. I pushed on the wheel to turn it more, but had no luck with my meager output.

1/4 turn isn't much. It's not like the diff will allow 1/4 turn to occur all at once since both wheels will be turning when slip occurs. So one wheel might turn 10 rotations while the other turns 10 1/4 during a slip event. I'm sure you've seen a slip vs grip graph for tire traction, so you know that a tire that's slipping slightly more than optimum is still providing nearly just as much traction.

Since the Quaife acts like an open diff until a torque difference appears, what's the power distribution of an open diff when both wheels are not slipping? Wouldn't it be 50/50?

I'm having trouble reconciling the theory that an ATB diff simply locks when slip occurs. How about cornering? I've felt the Quaife send power to the outside wheel under power in a corner and the car didn't drive like it had a spool. Or are you saying the lock occurs between the diff case and the outside shaft, not between the two shafts?

 

[D_Eclipse9916]

I hate to butt in, but this is a very good diff questionaire.

With the Kaaz, saying its always locked in, then whats the difference between a welded. I had my local dsm shop take a look and drive it, and he said that it wasnt a welded diff, but that it could be a kaaz 1.5 or something like it.

When I put one rear tire in the air, all the rest of the tires on the ground, and try turning the wheel, it goes about 1/8 of a turn before locking. Also when reversing or turning in parking lots or at low speeds the rear inside tire or maybe outside likes to click a lot and if in low traction (gravel or rain) itll spin the tire.

At first I thought it was the center diff but I remember that you (JTOBY) said that to check every single wheel, one up in the air at a time, and see if they all lock. Well every single one locks. I have no idea what is going on.....

 

[jtmcinder]

Man, lots of questions.

Why would I prefer a modified clutch-pack (aka Salisbury; aka Kaaz) for straight-line? Partly for the reason Matt gave: it locks in advance (based on input torque) while a helical waits. But also because you can sometimes light a tire up so much that almost no torque is being reflected; under those conditions, a helical goes open, while the clutch-pack stays locked.

What's the torque split of an open diff with no tire slip? What-ever the internal gears say. So, if you have a spider (as we do), it's 50/50. If you have a Cusco Tarmac center, it's 35/65. If you have a new STi, it's 41/59. Etc.

A Kaaz is not always locked. Yes, most have a preload on the clutches, but not enough to transfer even 100 ft-lbs (WARNING: that number will need to be washed before being used, given from whence I pulled it!). What makes them lock hard is the wedging action of the spider-legs on the case (see my tech article). This clamps the clutches harder.

- Jtoby

 

[JrCRXHF]

Just a little info on the Cusco Tarmac center it cost around $1300.00 and from what i can see it is for the evo 1-3 which means it should go in a DSM fine i think.

 

[dsm-onster]

Man, lots of questions.

Why would I prefer a modified clutch-pack (aka Salisbury; aka Kaaz) for straight-line? Partly for the reason Matt gave: it locks in advance (based on input torque) while a helical waits. But also because you can sometimes light a tire up so much that almost no torque is being reflected; under those conditions, a helical goes open, while the clutch-pack stays locked.

I didn't want to mention that this was the conclusion I was drawing becasue it just seams so hard to believe that a smoking, burning tire has no reflecting torque. I would think that it would at least have a little. But, I have heard of torsens sometimes breaking loose one wheel. . . I just really would want the daily driver ease of a quaife. But for the sake of performance (and every little bit counts for a FWD), I'll have to stick w/ the decision to get a KAAZ.

A Kaaz is not always locked. Yes, most have a preload on the clutches, but not enough to transfer even 100 ft-lbs (WARNING: that number will need to be washed before being used, given from whence I pulled it!). What makes them lock hard is the wedging action of the spider-legs on the case (see my tech article). This clamps the clutches harder.

- Jtoby

This is another piece to the KAAZ puzzle I didn't realize. . .

I was going to ask about the Cusco, but I really wanted to get the basic facts straight in my head first.

It seams the Cusco 1 way (or maybe 1.5) would be best for FWD. But more expensive than a Quaife!?

 

[daren_p]

Example application of this to your issue: assuming that the diff is locked (but the amounts of grip are equal [or, at least, sufficient]), the difference between the half-shafts will cause more torque to be delivered to the ground on the shorter side. You know those torque-limiting gizmos that tire places use to keep employees from over-torquing lug-nuts? Well, you got one of those on your driver's side. I'd get rid of it. (I've already got a major torque-limiting device known as "Toby's mad tuning skillz"; if you have anything like this, I wouldn't also have a second such device hooked up to one - and only one - axle.)

- Jtoby

I can comment that the tires I just replaced had 3 summers use on them, 2 summers with the stock diff & one summer with the Quaife. They were directional so I could only rotate front to back. Going by what you said with the shorter side getting more torque, this appears to be the case as I noticed a good amount more wear on the passenger side compared to the drivers.

I would imagine most of this additional wear came from when I was running the stock diff but it sounds like even with the Quaife the shorter passenger side will see more torque? Before the Quaife I had thought about ELHS but after the install the torque steer was basically gone, so I put no more thought into it. So even though I have a Quaife would swapping to ELHS provide more traction? Another reason I passed on the idea is I believe the awd's have a smaller diameter shaft compared to the fwds, but I guess an upgraded awd shaft is always an option.

 

[jtmcinder]

I don't think that there are any pure 1-ways for our cars, but I could be wrong. In any event, a Cusco 1.5-way works on the same method as a Kaaz.

As to a helical going fully open when one tire is smoking: I was hyperbolizing. But remember the basic rule that determines how much torque a helical can reroute: it is always a ratio. If one wheel is down to 25 ft-lbs, then even the steepest Quaife is only going to let you put down 100 ft-lbs on the other side. At corner exit, in first or second gear (i.e., autocrossing), that's close enough to count as worthlessly open in my book.

- Jtoby

 

[underradar92]

Apart from comparing diff.'s even with a locked/welded diff, wouldn't an unequal length driveshaft situation still torque-steer, simply due to geometry and chassis load differences?

If so, a differential of any style cannot make up for that, true?

 

[jtmcinder]

That's my understanding, yes, although I don't think it has much to do with geometry and/or chassis loading.

- Jtoby