You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Dialcaliper
Well-known member
Ditto above - the Cusco is an "open" 35/65 differential that replaces the stock 50/50 unit. Any LSD action is still provided by the stock VCU. The Evo 4-9 version works the same way except that instead of the VCU, the ACD clutchpack remains functional.
jepherz
Staff member
Correct me if I'm wrong, but if the center diff is 35/65, doesn't that mean the VC is working all the time to try and keep the ratio 50/50 since the wheels (under normal circumstances) are spinning the same speed? I must not understand their operation!
Dialcaliper
Well-known member
The planetary diff has outputs for the front and rear that are driven at a different radius within the diff, which means that even at the same speed, twice as much leverage is applied to the transfer case output as to the front differential.
The best analogy is to have someone hold a broomstick 2/3 of the way up. Grab both ends of the broomstick, and have the person pull the broomstick away from you. To prevent the stick from twisting, you will have to pull twice as hard on the short end as you do on the long one.
It's not a perfect analogy, as there are a couple weird things going on.
When all four wheels are planted, the planetary diff will "couple" (stop rotating internally), and torque will be applied 50/50 to the wheels as in a normal diff. Torque bias does nothing in straight line acceleration, unless the wheels start to spin.
When the car enters a turn, the front wheels start traveling faster than the rear and the diff begins "differentiating" speeds. (turned front wheels travel in a larger arc than the rear wheels, which sort of "shortcut" the corner - if you've ever made a turn and curbed the rear wheel on the sidewalk, you'll know what I mean).
At this point, the diff starts "preemptively" shifting torque towards the rear, until it very shortly reaches a certain "equilibrium" of relative rotation speeds, at which point it is biasing 35/65. The behavior up until this point is very similar to a Torsen diff, the difference being that a torsen can produce a much larger torque bias (some are up to 5:1 or 6:1, and it can bias *either* way - the planetary will bias towards the rear whenever motion occurs in the diff, regardless of direction - this is why an open planetary would be a bad thing in a front or rear differential. The planetary behaves like a fairly weak 2:1 torsen, but only in one direction. (the differential doesn't care which way it is spinning - as far as the guts are concerned, they are not moving at all until the output speeds change, unlike a torsen which has two "torque channels" - one that pushes each way through the diff)
If the wheels actually start spinning, the planetary will maintain the 2:1 bias, and excess power will go into spinning the wheel up like a flywheel instead of being put to the ground.
What this does is decrease the traction load on the tires going into a turn, which are already heavily loaded trying to turn the car - the end result is that more throttle can be applied before the front tires break loose. This is beneficial on asphalt, where grip is high.
There is a flaw of course. If one of the rear wheels breaks loose in a straight line, the differential will initially start biasing torque the wrong way, adding even more torque to the rear wheels, until the VCU starts resisting the spin, and transfers torque forward again. However, the torque bias is fixed, and the VCU can only transfer a fixed amount of torque forward based on the relative speed in the diff.
You can probably see how this might be detrimental in low traction conditions, hence the "Tarmac Spec" designation. In a highly powered rally car in very low traction, the diff action can overwhelm the VCU's ability to shift torque back forward, reducing the power total power that can be put down to the front wheels compared to a 50/50 LSD setup or a spool, and it will increase the tendency of the rear wheels to lose traction
Contrary to what you might have been told, a rear biased diff won't let you pull mad drifting powerslides like a RWD car will. It will however tend to produce *some* power oversteer, which will allow you to help steer the car with the throttle on asphalt like you can with a RWD car. As it is, a 50/50 AWD car must be almost entirely controlled with steering wheel adjustments, since throttle will produce unpredictable understeer, nothing or oversteer depending on where in the turn you are, and how much traction you have at each wheel.
The primary function is to reduce the load on the front tires, and eliminate the mid corner neutral or "power-understeer" behavior, making the car more predictable.
A stiffer VCU, if you can find one, would probably help out on a car with increased power over stock, especially in this situation.
On the plus side however, it will never be as bad offroad as a 2WD car, which always puts 100% of its torque to the front or rear wheels. /ubbthreads/images/graemlins/idea.gif
The best analogy is to have someone hold a broomstick 2/3 of the way up. Grab both ends of the broomstick, and have the person pull the broomstick away from you. To prevent the stick from twisting, you will have to pull twice as hard on the short end as you do on the long one.
It's not a perfect analogy, as there are a couple weird things going on.
When all four wheels are planted, the planetary diff will "couple" (stop rotating internally), and torque will be applied 50/50 to the wheels as in a normal diff. Torque bias does nothing in straight line acceleration, unless the wheels start to spin.
When the car enters a turn, the front wheels start traveling faster than the rear and the diff begins "differentiating" speeds. (turned front wheels travel in a larger arc than the rear wheels, which sort of "shortcut" the corner - if you've ever made a turn and curbed the rear wheel on the sidewalk, you'll know what I mean).
At this point, the diff starts "preemptively" shifting torque towards the rear, until it very shortly reaches a certain "equilibrium" of relative rotation speeds, at which point it is biasing 35/65. The behavior up until this point is very similar to a Torsen diff, the difference being that a torsen can produce a much larger torque bias (some are up to 5:1 or 6:1, and it can bias *either* way - the planetary will bias towards the rear whenever motion occurs in the diff, regardless of direction - this is why an open planetary would be a bad thing in a front or rear differential. The planetary behaves like a fairly weak 2:1 torsen, but only in one direction. (the differential doesn't care which way it is spinning - as far as the guts are concerned, they are not moving at all until the output speeds change, unlike a torsen which has two "torque channels" - one that pushes each way through the diff)
If the wheels actually start spinning, the planetary will maintain the 2:1 bias, and excess power will go into spinning the wheel up like a flywheel instead of being put to the ground.
What this does is decrease the traction load on the tires going into a turn, which are already heavily loaded trying to turn the car - the end result is that more throttle can be applied before the front tires break loose. This is beneficial on asphalt, where grip is high.
There is a flaw of course. If one of the rear wheels breaks loose in a straight line, the differential will initially start biasing torque the wrong way, adding even more torque to the rear wheels, until the VCU starts resisting the spin, and transfers torque forward again. However, the torque bias is fixed, and the VCU can only transfer a fixed amount of torque forward based on the relative speed in the diff.
You can probably see how this might be detrimental in low traction conditions, hence the "Tarmac Spec" designation. In a highly powered rally car in very low traction, the diff action can overwhelm the VCU's ability to shift torque back forward, reducing the power total power that can be put down to the front wheels compared to a 50/50 LSD setup or a spool, and it will increase the tendency of the rear wheels to lose traction
Contrary to what you might have been told, a rear biased diff won't let you pull mad drifting powerslides like a RWD car will. It will however tend to produce *some* power oversteer, which will allow you to help steer the car with the throttle on asphalt like you can with a RWD car. As it is, a 50/50 AWD car must be almost entirely controlled with steering wheel adjustments, since throttle will produce unpredictable understeer, nothing or oversteer depending on where in the turn you are, and how much traction you have at each wheel.
The primary function is to reduce the load on the front tires, and eliminate the mid corner neutral or "power-understeer" behavior, making the car more predictable.
A stiffer VCU, if you can find one, would probably help out on a car with increased power over stock, especially in this situation.
On the plus side however, it will never be as bad offroad as a 2WD car, which always puts 100% of its torque to the front or rear wheels. /ubbthreads/images/graemlins/idea.gif
Last edited:
Very Interesting post!
Now my question I guess is this: Is it possible to tell at all whether a car has a cusco tarmac gear without dismantling the transaxle? Like is the behaviour so totally different from a stock equipped car? The reason I ask is I'm thinking of the tarmac gear as an upgrade, but I wouldn't want to do it redundantly in case I have one already.
I do have a suspicion that mine is stock 50:50 since during track days and autocrossing, I've noticed that putting down the hammer just results in understeer. I'm having to really anticipate the car a lot since I cannot lay on the hammer in the point and shoot manner that I was used to. Of course when I say "I was used to", I mean compared to RWD. An underpowered RWD at that! (1st get NA MR2) A car that taught me to be on throttle as early as I can as much as I can because it was underpowered enough it couldn't light up the rears to drift nor could it really pick up the fronts to plow.
I've also done donuts and no, the rear end does not try to overtake the fronts either. Although now that I've thought of it, constant radius donuts could see the VCU lock up to the point that it negates whatever biased ratio the centre diff may have.
Thanks!
Randy
Now my question I guess is this: Is it possible to tell at all whether a car has a cusco tarmac gear without dismantling the transaxle? Like is the behaviour so totally different from a stock equipped car? The reason I ask is I'm thinking of the tarmac gear as an upgrade, but I wouldn't want to do it redundantly in case I have one already.
I do have a suspicion that mine is stock 50:50 since during track days and autocrossing, I've noticed that putting down the hammer just results in understeer. I'm having to really anticipate the car a lot since I cannot lay on the hammer in the point and shoot manner that I was used to. Of course when I say "I was used to", I mean compared to RWD. An underpowered RWD at that! (1st get NA MR2) A car that taught me to be on throttle as early as I can as much as I can because it was underpowered enough it couldn't light up the rears to drift nor could it really pick up the fronts to plow.
I've also done donuts and no, the rear end does not try to overtake the fronts either. Although now that I've thought of it, constant radius donuts could see the VCU lock up to the point that it negates whatever biased ratio the centre diff may have.
Thanks!
Randy
Last edited by a moderator:
Dialcaliper
Well-known member
The only thing I can think of is that if you jack all 4 wheels off the ground, and run the car in gear with the wheels completely unloaded, the center diff will try to spin the rear wheels twice as fast as the front (if it's a 33/65 split). (the 50/50 torque split is only stable when the wheels are forcibly regulated to the same speed, ie, on the ground and in a straight line).
The only flaw with this is that the VCU might mask the effect.
A 50/50 split AWD car doesn't really respond predictably to throttle steering like a RWD car since you are affecting grip at both ends of the car. (or a FWD car for that matter). An AWD car with no limited slip and a neutral chassis will typically understeer into the turn, run as the car is set up in the middle, and oversteer out of the turn. You'll need to get used to steering with the steering wheel and applying as much throttle as the car will allow through the turn. Also, AWD lends itself to earlier straight line braking, later apexing and putting the hammer down early out of the turn, instead of trail braking and attempting to carry speed on the fastest line through the corner.
Chassis balance is important, limited slip diffs will add understeer off throttle(even in the rear), and its fundamentally different driving style than a 2WD car.
The only flaw with this is that the VCU might mask the effect.
A 50/50 split AWD car doesn't really respond predictably to throttle steering like a RWD car since you are affecting grip at both ends of the car. (or a FWD car for that matter). An AWD car with no limited slip and a neutral chassis will typically understeer into the turn, run as the car is set up in the middle, and oversteer out of the turn. You'll need to get used to steering with the steering wheel and applying as much throttle as the car will allow through the turn. Also, AWD lends itself to earlier straight line braking, later apexing and putting the hammer down early out of the turn, instead of trail braking and attempting to carry speed on the fastest line through the corner.
Chassis balance is important, limited slip diffs will add understeer off throttle(even in the rear), and its fundamentally different driving style than a 2WD car.
Last edited:
Quote:
The only thing I can think of is that if you jack all 4 wheels off the ground, and run the car in gear with the wheels completely unloaded, the center diff will try to spin the rear wheels twice as fast as the front (if it's a 33/65 split). (the 50/50 torque split is only stable when the wheels are forcibly regulated to the same speed, ie, on the ground and in a straight line).
A 35:65 center diff describes the torque bias of the diff, not the rpm bias, which is determined by the vcu, so infinitely variable with a nominal 50:50 split. If you jacked a car up and ran it in gear, there would be the same resistance (none) so they would spin freely and at the same rpm. However, if you tried to stop one of those wheels from spinning, it would be roughly twice as hard in the rear as in the front, because it would be getting 65% of the engine's torque compared to 35.
The only thing I can think of is that if you jack all 4 wheels off the ground, and run the car in gear with the wheels completely unloaded, the center diff will try to spin the rear wheels twice as fast as the front (if it's a 33/65 split). (the 50/50 torque split is only stable when the wheels are forcibly regulated to the same speed, ie, on the ground and in a straight line).
A 35:65 center diff describes the torque bias of the diff, not the rpm bias, which is determined by the vcu, so infinitely variable with a nominal 50:50 split. If you jacked a car up and ran it in gear, there would be the same resistance (none) so they would spin freely and at the same rpm. However, if you tried to stop one of those wheels from spinning, it would be roughly twice as hard in the rear as in the front, because it would be getting 65% of the engine's torque compared to 35.
Another reason why I'm having serious second thoughts of what centre diff I have is that my car came off a boat. It has many mods and since I don't know stock from not stock, it took me a while to discover them all.
Anyway, I think I have an idea.
Once inclement weather really kicks in here in Canada, I'm going to put a fat line of shoe polish on my sidewalls, like the way dragracers do. Then I'm going to launch the car and videotape the results to see if the rears spin up first with the fronts catching up.
Does that make sense? The snow should really keep it from putting too much stress on the driveline. (and clutch!)
Anyway, I think I have an idea.
Once inclement weather really kicks in here in Canada, I'm going to put a fat line of shoe polish on my sidewalls, like the way dragracers do. Then I'm going to launch the car and videotape the results to see if the rears spin up first with the fronts catching up.
Does that make sense? The snow should really keep it from putting too much stress on the driveline. (and clutch!)
Last edited by a moderator:
Yes it is possible to change the centre diff without removing the gearbox from the car.
jack car up, remove right front wheel, disconnecting the strut from the hub will give you abit more room, but isnt' totally necessary.
drain the oil, remove the end casing, remove 5th gear, the detent springs and the viscous unit.
then you can remove the 'plate' from the gearbox (exposing the rest of the gears and the centre diff)
you can then slide the centre diff out - it's easier if you remove the transfer case and slide the shaft out after you remove the v/c, but you can do it with this shaft still in place.
jack car up, remove right front wheel, disconnecting the strut from the hub will give you abit more room, but isnt' totally necessary.
drain the oil, remove the end casing, remove 5th gear, the detent springs and the viscous unit.
then you can remove the 'plate' from the gearbox (exposing the rest of the gears and the centre diff)
you can then slide the centre diff out - it's easier if you remove the transfer case and slide the shaft out after you remove the v/c, but you can do it with this shaft still in place.
Amazing Thread, I have a question though.
Our TQ split being 50/50 even when we lose traction, doesnt give us much room for increased "oversteer", especially scince our rear Diffs are completely Open, now upgradeing to a a 4 bolt LSD version or sending your Diff off to TRE to have an Evo-3 diff put in, could infact increase our oversteer under massive accelaration through a turn right? or just geeting a huge thick rear sway bar?
Our TQ split being 50/50 even when we lose traction, doesnt give us much room for increased "oversteer", especially scince our rear Diffs are completely Open, now upgradeing to a a 4 bolt LSD version or sending your Diff off to TRE to have an Evo-3 diff put in, could infact increase our oversteer under massive accelaration through a turn right? or just geeting a huge thick rear sway bar?
First of all you have to understand that the cause/source of the understeer is largely the chassis and suspension, not the drivetrain, so in turn the fix will not largely be in the drivetrain. These cars are pretty heavy, most of that is over the front wheels, and being sedans are long and tall, none of these traits do you any favors in the understeer department. Suspension mods, especially a larger rear bar, will go further in reducing understeer by transferring some of the cornering load to the rear tires. Front caster bushings are also good. This doesn't mean that a 4 bolt lsd won't help you put the power down slightly earlier in a corner, it will, but it will have no effect on steady state handling characteristics of the vehicle.
Dialcaliper
Well-known member
Okay, so after second guessing myself, I finally decided to sit down and work out the force balance to figure out what the heck is going on in the planetary differential.
The answer I came up with, surprisingly (to me), is that the differential puts 35% of the torque to the front and 65% of torque to the rear wheels *at all times* regardless of how it is spinning internally, as long as the rotational speeds are *constant* and not in the process of changing speeds due to unbalanced forces. If one side loses traction, and the balance cannot be maintained, the free end will begin to speed up, and the end with grip will slow down (just like a regular differential) until one of the following happens:
a) The end with grip comes to a complete stop, and the free end spins at a speed based on the engine speed and the gear ratio.
b) The driver drops the throttle and reduces the input torque to the point where balance is returned, or
c) Something else comes into play to limit the speed and transfer torque - i.e. the VCU
In other words, the differential does not act as a 50/50 split when the car is moving straight. The torque split is determined strictly by the gear ratios inside it, in this case 35/65.
If the VCU is not rotating, and all wheels have traction, the car will be transferring power to the ground in a true 35/65 ratio.
The answer I came up with, surprisingly (to me), is that the differential puts 35% of the torque to the front and 65% of torque to the rear wheels *at all times* regardless of how it is spinning internally, as long as the rotational speeds are *constant* and not in the process of changing speeds due to unbalanced forces. If one side loses traction, and the balance cannot be maintained, the free end will begin to speed up, and the end with grip will slow down (just like a regular differential) until one of the following happens:
a) The end with grip comes to a complete stop, and the free end spins at a speed based on the engine speed and the gear ratio.
b) The driver drops the throttle and reduces the input torque to the point where balance is returned, or
c) Something else comes into play to limit the speed and transfer torque - i.e. the VCU
In other words, the differential does not act as a 50/50 split when the car is moving straight. The torque split is determined strictly by the gear ratios inside it, in this case 35/65.
If the VCU is not rotating, and all wheels have traction, the car will be transferring power to the ground in a true 35/65 ratio.
