The Central Hub for DSM Community and Information

For 1990-1999 Mitsubishi Eclipse, Eagle Talon, Plymouth Laser, and Galant VR-4 Owners. This is where the DSM platform history is documented and archived. Log in to help us in our mission, and to remove most ads from the browsing experience.

torque splits

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

jtmcinder

DSM Wiseman
5,402
96
Nov 4, 2003
Iowa City, Iowa
In a different thread -- http://dsmtuners.com/forums/showthread.php?p=1478139#post1478139 -- I suggested that a 2G DSM would benefit from a 40/60 center differential. Someone else (ACM) suggested that we want a 30/70. We agree that life sucks (in that only a 35/65 is currently available), but maybe a discussion of this is worthwhile. For autocrossing -- which is what I do -- I think that I am wrong.

My reason for wanting a 40/60 (instead of 35/65) is that I want the split to be as close as possible to the weight distribution (and, therefore, the grip distribution) of a lowered 2G at 1g acceleration. One reason to want a more radical split, instead, is to get the rears to break free first under power so that the car can be rotated using power. In other words, the 40/60 split is to optimize straight-line acceleration while a more radical split might optimize corner-exit behavior.

The best way to understand this issue is in terms of traction circles. At any given time there is so much weight on each tire. The more weight, the larger the traction circle. When sitting still or going straight ahead at a constant speed, the front traction circles are larger than the rear circles, because our cars are front heavy. When cornering at a constant speed, the front circles are bigger than the rear circles (for the same reason) and the outside circles are larger than the inside circles (because weight is transferred to the outside when turning).

Similarly, at any given time, you are using up so much of the available grip at each corner. You use up the available grip by asking the tire to accelerate the car in some direction. Note that the tire doesn't care about the specific direction; it has so much total grip and, at any time, so much of it is being used to make the car turn and/or accelerate or brake. To the extent that the traction circle for a tire is larger than what that tire is currently using, you have grip to spare.

When a car is cornering at the limit, the end of the car that does not have grip to spare is the front. Therefore, if you want to accelerate at the same time, you had better ask the rear tires to do this for you, because if you ask the fronts, at least one of them will exceed its traction circle, which will cause the front end to turn less, which will cause the car to understeer. Therefore, you want a rear-biased center differential, because this is how you get the rears to do the extra work.

Going farther, for corner exit you want a center diff that is very very biased to the rear. Again, if you are at the limit, it is the front end of the car that has nothing to spare, so you need to send the engine torque to the rear if you want to accelerate at the same time. But you don't want a center that is completely biased to the rear because this will be too far from the optimal bias for straight-line acceleration. (Remember that we launch much better than a RWD car because we can use all four tires to do this.) As a compromise, then, you want something a bit more radical than the 40/60 that the launch suggests. How much more depends on which is more important to you.

So there's no simple answer, but I now believe that autocrossers want a lot more bias than 40/60 and maybe even more than 35/65. So I am now very sympathetic to the idea that 30/70 would be better than the currently-available 35/65. And a good discussion of this would be fun.

Nomex on.

- Jtoby
 
jtmcinder said:
but maybe a discussion of this is worthwhile.
I disagree. Unless you're going to re-engineer the entire DSM driveline, and provide for the different scrubbing at both ends (such as a slipping scheme so that a torque split is essentially obviated) you're going to be playing hell on grip and handling all tires. What did Subaru and Audi do with their splits? And are they autoX stars?

But don't go by me. With its weight, I've thought trying to race a DSM was a losing battle to begin with. I have great admiration for those of you who have not only gone ahead and tried it, but have also succeeded at it. :thumb:
 
Defiant said:
I disagree. Unless you're going to re-engineer the entire DSM driveline, and provide for the different scrubbing at both ends (such as a slipping scheme so that a torque split is essentially obviated) you're going to be playing hell on grip and handling all tires. What did Subaru and Audi do with their splits? And are they autoX stars?

Please don't tell me that even you, Defiant, believe that an unequal-split differential turns the wheels at different speeds, because that is not true. If my differentials thread does not make it clear how unequal planetary-type diffs work, I can try again.

As to Subaru and Audi and their splits, Subaru uses a 35/65 planetary in the STi and has been putting a 45/55 in the automatic Legacy for more than 10 years (calling it the VTD). Neither car suffers excess tire-wear. Audi hasn't used a planetary that I know of, so that seems a bit off-point, but Porsche has used an unequal split in the 959/911-AWD.

Yes, if you put different-sized tires on the two ends of an AWD you will end up with an unequal torque-split. But there are are other ways to get an unequal torque-split and these do not ask the diff to differentiate when going straight ahead.

- Jtoby

ps. mis-conception #2 is the idea that a Subaru diff is only 35/65 when locked, when exactly the opposite is true, but I'll save that for another day
 
I believe the answer is in the traction circles of each corner, as you identified. Using the front wheels to turn the car uses up as much traction as is available, particularly in autocross. Feeding power to wheels also use up that same grip. Reducing the power to the front wheels allows more of the grip to be dedicated to turning the car.

The flip side of this is that the rear tyres have grip to spare, typically, so whilst the front tyres are being tortured, the rears are just hanging out. Feed more power to the rear uses up freely available grip, not (necessarily) overpowering the available grip. Pull that additional power from the front and what essentially happens is the front turns better and the rear accelerates better. The beauty of this is that it allows the car to overpower the rear grip before the front - something not achievable previously without reducing the rear grip, even the the front still got tortured under power.

With no other changes, my car, fitted with the 35/65 Cusco, went from a delicate balancing act of push to oversteer on command, with a front end glued in place. I was busy adding rear grip and still the front stuck and I had oversteer on command. Then my rear LSD wore out and the balance went down the toilet !

Even on the street, the difference is quite marked, the car is much livelier, much more responsive - much more FUN. And that's not pushing it, just driving around. I see no downside to this setup, and I believe a 30/70 split may well be even better. Too much of a rear bias will be a bad thing, ultimately, but I do not believe that 65 or 70% is too much.

Such an LSD would be a great improvement - but it needs to be an LSD, not an open diff like the Cusco - and preferably not $2,000 either ! I also believe that a pure Quaife-type 0-100-0 potential torque shift would be a bad thing, no more than 50% should ever be shunted to the front, but up to 90% should be able to be shunted to the rear. I would buy this diff in a heartbeat - even if it did cost as much as the Cusco. After Penskes, this would be the single greatest improvement one could make to a DSM.

IMHO,
Charles
 
I have a question,

I've been thinking of taking full advantage of my awd for handling purposes, instead of just drag racing. I still like to do some drag racing, so what would a 30/70 split do for sombody that would like to still drag race here and there?
 
Since weight transfer under acceleration loads the rears and unloads the fronts, a (significant) rear bias should improve a drag-race launch.

Charles (who's never drag-raced anything, but has read a book or two)


pianoman said:
I have a question,

I've been thinking of taking full advantage of my awd for handling purposes, instead of just drag racing. I still like to do some drag racing, so what would a 30/70 split do for sombody that would like to still drag race here and there?
 
I thought about it for a sec after I posted, and yes, obviously it will help, since the car squats on its rears when it launches. Have you guys thought about that maby the rear-end won't be able to handle the power? Normally its only seeing 50%of the power. ?
 
Do the math.

How much power are drag cars making without breaking rear ends ? Divide by 2, that's a pretty concrete number you can use. Take 65% of your current or target power output and compare the two.

Charles
 
Add Value - Be Respectful - No Trolling - No Misinformation - Participate Often!
Support Vendors who Support the DSM Community

Build Thread Updates

Latest Classifieds

Back
Top