Rear Diff Options - Power Oversteer Found Here
Nucci, I think you misunderstood. Although, show me a FWD BMW and I guess you have a point?
I was saying for the FRONT diff, I can't imagine too many situations where a clutch pack would be a good thing in an EVO. The rear is a clutch pack, and is desirable, basically what this whole thread is about.
As for the torsion diffs, I could be off, but I beleive the majority of the "limited slip" comes from the circumference of the gear, not the ends? The gears get pushed out radially, as well as axially, due to the helical cut, but the majority of the contact area is the circumference. I have seen some modifications on this idea like what you are talking about, but I have never looked into it. To me, you really don't want "lock up" in the wheels that steer. They need to be able to rotate at different speeds while cornering and if you want to minimize slip, that speed differential HAS to happen. A torsion diff does just that, splits up the torque and speed to where ever can use it best.
I was saying for the FRONT diff, I can't imagine too many situations where a clutch pack would be a good thing in an EVO. The rear is a clutch pack, and is desirable, basically what this whole thread is about.
As for the torsion diffs, I could be off, but I beleive the majority of the "limited slip" comes from the circumference of the gear, not the ends? The gears get pushed out radially, as well as axially, due to the helical cut, but the majority of the contact area is the circumference. I have seen some modifications on this idea like what you are talking about, but I have never looked into it. To me, you really don't want "lock up" in the wheels that steer. They need to be able to rotate at different speeds while cornering and if you want to minimize slip, that speed differential HAS to happen. A torsion diff does just that, splits up the torque and speed to where ever can use it best.
That is most definitely not true. (1, 1.5, 2 way only describes the relative lockup in accel and decel and has nothing to do with the method of torque biasing) Torsen type 2 units apply their friction through the end faces of the gears, where they meet the case. One end of the gears is the friction surface for accel, and the other end of the gears is the friction surface for decel. Making the surface finish smoother, or adding needle bearings to these interface surfaces, would reduce the locking effect at the same input torque (as compared to OE). Making the surface finish rougher, or possibly machining away part of the gear end surface (to increase unit frictional loading) would increase the locking effect as compared to OE. Since there are different surfaces for accel and decel, you could make a 1.5 way or 1 way if you wanted (stock Torsens should act like a 2 way).
. . .
It would depend if it was a Eaton-style clutch diff where springs provide static preload which never changes and is the only method of torque biasing, or a Salisbury/ZF-style where separate angled ramps for accel and decel allow the diff's own input/output torque balance to adjust the lockup. Salisburys can be made to have very very little lockup on decel with 90 degree ramps (the decel ramp is vertical - I guess that makes it a wall, not a ramp...) BMW road race cars love these diffs.
Matt
. . .
It would depend if it was a Eaton-style clutch diff where springs provide static preload which never changes and is the only method of torque biasing, or a Salisbury/ZF-style where separate angled ramps for accel and decel allow the diff's own input/output torque balance to adjust the lockup. Salisburys can be made to have very very little lockup on decel with 90 degree ramps (the decel ramp is vertical - I guess that makes it a wall, not a ramp...) BMW road race cars love these diffs.
Matt
probably no one? i was under the impression the helical was good enough for the front. i would assume going from the helical to aftermarket the improvements would be marginal at best? in any case if anything should go first its the rear
The more and more I see people making Evos into fast autocross cars, I'm convinced that the end goal is to make it throttle steer like an American RWD car. (I don't necessarily think this idea transfers to road racing or the streets though, due to differences in corner radius) Since the loaded front tire is so super-saturated by cornering force in comparison to the other tires, we need to get drive torque off the loaded front tire and onto the unloaded front tire and the loaded rear tire (this thread covers the rear diff in depth already). Whoever figures out how to do that first, or better than everybody else, will have an advantage.
The only reason I brought tech regarding clutch pack/ZF diffs is because as of now, nobody has (publicly anyways) shown how to make modifications to the front half (diff + ACD) of an Evo driveline in order to move drive torque from the loaded tire to the unloaded tire. Since nobody yet has the answer, they might as well have good information regarding the operation of all the available mechanical limited-slip diff styles, in order to be able to think up new crazy ideas to test - until there is a solution.
Last edited by nucci; Sep 26, 2010 at 12:49 PM.
Depends upon what you do... My concern is making to the car too much RWD biased for a course with short radius corners...and then messing it up for when I run (someday) run on the big track.
A Torsen diff can also be tuned simply by changing the helical angle.
As this thread is lengthy and supposed to be about the rear diff, maybe we should continue front and center diff talk elsewhere?
As this thread is lengthy and supposed to be about the rear diff, maybe we should continue front and center diff talk elsewhere?
i thought the point was to make the loaded tire get as much of the torque as possible or something like that, but im no expert
im just not sure if the 60/40 is necessarily faster than an acd, or at least one with the acd tuned - otherwise wouldn't all the overseas competition evos switch to a rear biased center diff as opposed to tuning the acd? the 60/40 may very well be faster, i'd just like to see it in direct comparison to the acd system







I didn't mention it because its simply...out of my budget.