Yes it really is Torque split (on both cars)
Perhaps you've never read this Car and Driver article ?
http://www.caranddriver.com/features...different.html

:P

I wish they had gone back to the track and tried some more testing to make the STI better...

You mean playing with the DCCD mid-turn?

No, read the last paragraph of the article...

Yes just maybe if you turn the dial Enough while cornering, in each corner... Er wait a minute ... :P

lmao

i have found if you adjust it and try different settings you prove Dean very correct in stating the subaru engineers really know their stuff. That doesn't mean it can't add fun to your fun runs though :)

Also the owners manual does state it won't harm anything to adjust it as you drive as long as there is no wheel spin. And during wheel spin if you are in a manually selected locking rate it is Okay to select Auto. Which i've also done from time to time when playing around in the snow with the dccd :)

Yeah, 'cause C&D is an engineering bible. :rolleyes:

The torque split ratio is defined by the mechanical properties of the open center diff. The DCCD clutch simply limits wheel spin. You cannot, under any circumstances (except perhaps if you change the laws of physics) get different than 35f/65r torque split out of the STi's center-diff ('04-'05, the split I hear is different on the '06, but it's still *fixed*).

Here are those pages:

http://velozt.com/elements/tech/sti/.../DCCD_FAQ.html

Read it carefully, and you'll see that no matter what you're doing with that knob, the center diff ratio is *always* fixed at 35/65. The clutches allow you to send power to the wheels that aren't slipping, but that's go nothing to do with the diff's ratio.

you do know in that link you sent me it has a Chart that show Actual Front and rear Torque being different At different locking rates? which says in a nutshell that YES the torque gets moved around.

Did you even read that or did you just post it thinking 'ya i'm right, and this proves it' ? cause it proves you wrong.

It also has a lot of wrong stuff in there things like 'if your traction will only permit 100 ft lbs of torque to be put down with out wheel spin the engine never produces more than 100 ft lbs of torque. Wow that's so wrong i don't know where to begin. If that's how physics worked no one would ever spin their wheels because as in that explation your motor can't produce more power than your wheels can put down (WRONG!!!!)

maybe they meant you can't transmit more torque to the ground than your wheels will permit traction wise and when your motor produces Over that amount your tires spin.

If you Re-read that article Carefully it says when a center diff is fully open the torque split never changes, But if you can Lock it some how... than it does.

Reading comprehension is our friend :)

i'm trying to find a PDF Produces by the subaru engineers that explains it. it had it on my old work machine. it even explains the conditions that envoke various amounts of lock.

such as as you get on the brakes it uses less lock, as you turn the steering wheel , less lock, as you get back on the gas and straighten out it dials up the lock. just can't find the damn thing. its got a tag like Driver-controllable center differential .pdf and since it came from engineers and not 1 guy piecing together his understandings from way things work and nasioc, its bound to be a lil better.

This from the article you linked sums it up the best

So what should you remember from Part 1 of this article?

1. With an open diff the torque split is always fixed by the gearing of the planetary gearset. With the STi, it is 65R/35F.

2. When the diff is locked the answer to “what is the torque split” is: It depends. It is not always 50:50.

IT DEPENDS ON THE LOCKING RATE :)

http://www.purepwnage.com/media/Kyle...ng%20noobs.mp3
good track

1) You are a crackhead... three rambling, blubbering posts in a row. :roll:

2) You may have read the article, but you certainly don't understand it. I specifically said, "Read it carefully". Your comments indicate you blew over it, and didn't bother to comprehend what's written in there. This especially shows your incompetents regarding the topic:

Of course the motor makes over 100ftlbs, that's what causes wheelspin. If you actually read the article, you'd understand that the example cases are to demonstrate torque distribution (which had nothing to do with changing the ratio of the torque split) without the wheels slipping. If the motor were making more than 100ftlbs, the wheels would slip, and the torque at each end would instantly revert to the fixed 35/65 ratio instead of applying torque to the wheels with grip.

3) THE CENTER DIFF USES A PLANETARY GEAR. THE RATIO IS FIXED. NO AMOUNT OF LOCKING OF A CLUTCH CAN CHANGE THE NUMBER OF TEETH IN THE GEAR. This of it this way: if the motor makes a maximum of 300ftlbs of torque, you will *never* get more than 105ftlbs to the front wheels and never more than 195ftlbs to the rear wheels. When traction is limited, you can distribute the power to the wheels with more traction via the DCCD clutches, but not in a manner that exceeds the fixed ratio of the diff. That DCCD knob does not make the car into a rear-wheel drive car, the ratio is FIXED at 35/65. Period.

4) Go read the article again. And again. And 5 more times. Until you understand it. Then come back here and talk shit to me again.

Ahhh... this is where it gets good! Where's the smilie with the little dude eating popcorn?

I need an adult! I need an adult! Wow. i should forward this thread to Allon so we can get some feedback from that side.

Scott, I realize Alex's posts were rambling and while your above statement may be true in ideal conditions on paper as was my "Pin locking brake system" :), perhaps you should read your referenced article again...

A differential allows for differing speeds on each of the output shafts. The ratio of the gears in a planetary differential related to output torque is only absolute when it is fully open.

As the article clearly points out, under partial to full locking, the actual amount of torque at each end is a function of a number of different factors.

Dean, part of the problem is that the article uses the term "torque split" inconsistently. The "torque split" of the differential is not the same thing as the "torque split" seen at the wheels.

People have this misconception that adjusting the DCCD changes the way the differential splits torque between the wheels, which is patently not the case. The differential cannot send more than its fixed ratios of torque to the wheels. What the clutches can do however, is allow the differential to send power to wheels that have traction.

Think of it this way: if the rear wheels have little traction, the clutch can engage and act as resistance in place of the resistance that should be provided by the ground, allowing the front wheels to see up to their full 35% of the engine's output. The torque distribution in that situation will appear like there's more torque to the front half of the differential, but that's not the case, because the "missing" torque at the rear is actually being gobbled up by the clutch to prevent wheel spin.

The article is a pain to read and understand, but if you've read it 100 times over the last 2 years like I have, it will start to make sense.

I agree, with you, but by definition, a twisting motion that is not applied is not torque...

So, the gotcha becomes are we talking about torque at the wheels which is after the clutches on the output shafts of the whole "DCCD" differential unit, or on the output shafts of the internal planetary diff gears within the case which is a subset of the whole "DCCD"?

Without clearly identifying which, you can argue this forever with many parties being correct...

Just because an engine is capable of producing 300 ft/lbs of torque doesn't mean it is. If the maximum torque that can be absorbed by the tires and assorted drive line losses including DCCD clutches is 200ft/lbs, that is all the maximum measurable torque that the overall system will attain...

Can you point me to any diagrams of the actual internals of the DCCD? It would be interesting to see if the clutches are strictly on the output shafts or internal to the planetary diff itself.

I have a file called "DCCD technicians reference booklet" that has diagrams of everything, but I can't seem to attach it. I can email it to you if you want.

It's a messy subject. But the easy way to think about it is: the torque split is 35f/65r, so that's what's going to the ground when you have traction. The clutch can help prevent loss of traction by effectively taking power away from the wheels that are slipping allowing the wheels with grip to get their full share of the engine's power. The DCCD allows the driver to adjust the sensitivity of the clutch's engagement, from no engagement at all (the wheels are free to turn at any speed relative to each other) to full lock (the wheels must turn the same speed).

I get the feeling that I'm going to see the internals of a DCCD center diff sooner rather than later... mines not working. :(

Send it to me too.

I sent it to you and Dean. Can you put it on the site somewhere? I can't attach it to a post for some reason.

Zip it first and then upload it maybe Mike?

I'm about to add it to the tech section.

edit: http://www.seccs.org/tech

A quick look at Mikes' doc of the internals make it appear that it does change from 35/65 to 50/50 assuming sufficient traction is available to reach those levels by effectively locking the planetary gears. It appears there is only a single clutch that appears to control the locking between the input and output shaft in effect overriding the mechanical gear ratios of the diff.

This makes complete sense IMHO. A locked set of planetary gears results in the equivalent of a solid shaft which has a 50/50 split. The clutch slips between the geared ratio of 35/65 and the locked up 50/50.

I am continuing to read and look at pictures. The reduction of clutch engagement on braking may be why I am finding left foot braking to be less effective than I would expect while cornering...

I'm having a bitch of a time visualizing what's going on in there, but it doesn't sound to me like there's anything changing the ratios of the gears.

It seems like the clutch engages, adding drag between the planetary carrier and the differential's case, thus providing drag torque allowing an opposite force for the front wheel side of the diff to push against in order to raise the torque seen at the front wheels.

I'd kill for a cutaway 3D animation of this beast in action.

I'll make you one if you kill Dean.

(that sweet SM trophy will be mine!)

So would I. What I am seeing is that the clutch acts in parallel with the planetaries. The clutch effectively provides friction between the same two surfaces the planetaries rotate between. That friction allows the torque split to be varied from 35/65 when the clutch is disengaged to 50/50 when fully engaged/locked.

There is another interesting statements in the last two paragraphs on page "81" that seams to imply the maximum split in auto is 45/55...

OK, I'll jump into the middle of this, which looks like it may not be a good idea! :lol:

If the diff fully locks and makes it one rigid transfer case a'la 4WD truck (which the drawings make it look like it does), then you do get 50/50 torque when all four wheels on the ground, and 0/100 if the front axle is completely off the ground. I'll make a better explanation, and if you guys are really nice maybe even an animation. :P

Nope.

It's not a transfer case, nor does it act like one. At full lock, you may not have a speed differential between the front and rear wheels, but there is more torque at the rear due to the gearing in the diff. On dry pavement, with tons of traction, in a straight line, at 100% lock, the wheelspeeds will be the same, but the rears are pushing the car harder than the fronts are.

With the front wheels off the ground, there is zero torque on the them because there's no ground pushing back as they turn, but the rears are not getting 100% of the engine's torque... they're getting 65%, the max allowed by the center diff. The rest of the power is being wasted spinning the front wheels. Similarly, with the rears in the air, the fronts can get only 35% of the engine's power, even though they're generating 100% of the torque that's reaching the ground, since the rears are spinning freely.

It's the difference between the torque split of the differential, and the effective torque reaching the wheels. Sure there are situations where the torque at the wheels vary greatly from 100%/0% to 0%/100%, but never is the torque passing through the diff split any other why than 35%/65%. To put it another way, the torque distribution may be 100/0, but that means the torque getting to the ground is only 35% of what the motor's generating. Which is exactly why you don't want an AWD car lifting wheels, doing so wastes power in the differentials ...which brings us full circle to what started this thread.

Fixed??? From what I am seeing, it is pretty clear that the clutch allows the potential torque split to the output shafts of the entire DCCD to go from 35/65 to 50/50 at the same time as it reduces the ability for the two output shafts to rotate at different speeds.

I finally found a mostly useful diagram and formulas...
http://gemini.tntech.edu/~slc3675/me...ure/planet.jpg
If you swap the words input and output, you have the outside ring as the input, one of the outputs is the shaft driven by the arm on the planetary and the last output the shaft on the sun gear. by varying the diameter of the sun and planet gear I believe is how you set the torque split in an open condition. Subaru then effectively adds a clutch between the two output shafts which as it is engaged, the two shafts are forced to rotate at more and more similar speed until lock up at which point all gears are fixed and shafts are rotating at the same speed and effective torque split is 50/50.

The clutch could just as easily be between the input and either of the outputs and have the same effect I believe, but it is more difficult to visualize I think.

Well, you are close. At 0% lock is when the gears are in effect and the torque is distributed via the gear ratio, so with tons of traction, in a straight line the rears are pushing harder than the fronts. And, if I understand the drawing correctly the clutches completely lock out the gears. So then in the same situation when we 100% lock the diff, the gears are taken out of the equation and we have 50/50 torque to the front and rear. I'll draw a load path on the cross section tomorrow. I promise.

Again, if I understand the drawing correctly, this is not the case -- the clutches lock the input with the two outputs. So then with either axle off of the ground the axle still on the groung would _have_ to get 100% of the torque -- there would be no where else it could go.

I'll try to get a drawing done of this thing. It should explain everything. I am not sure at this point, but this looks very similar to something I have worked with before.

Dean, yes, from what I can tell so far, you are right in that the ring gear is the input and the clutches lock all three together...

But if everything is locked together, the torques can actually vary from 100/0 <=> 50/50 <=> 0/100.

Please note that I used the words torque split which I believe we are using to talk about the theoretical split assuming both output shafts have sufficient traction/friction for that level of torque to be achieved... With everything locked, the theoretical torque split as I just described it can only be 50/50 IMHO...

My understanding of this is that the DCCD regulates it from 35/65 to 50/50 when locked/fully engaged at least that is what the Subaru guy I worked with while working on the Speed GT car a couple of years ago told me and makes sense from our own testing.

Mike

i agree with scott. with my very limited knowledge of mechical things torque split is determined by the mechanical gearing which cannot be changed.
how i understand it is to get that "50/50'' the DCCD clutches slip to reduce power to the rear wheels while the front wheels get the max power the mechanical gears allow which gives you the "50/50" feel if that makes any sense.

Other forces such as a clutch acting between the same two axis as the planetary gears most certainly can effect the split.

In the extreme case, consider if we weld the planetaries to the sun and ring gear. We have in no way changed the number of teeth or diameter of any of the gears, but I guarantee the split is 50/50.

The clutch is just a variable weld. :)

Let me ask you this: if the clutch on your motor is slipping, can you get all the power/torque the motor is generating to the transmission? Of course not.

If the clutch in the DCCD is slipping, it too is taking power from the wheels. But what we gain is the resistance of the clutch acting as something for the diff to push against when the tires on one half of the car are slipping. That way you can generate torque through the open diff. But that doesn't change the gearing of the diff, 35% is still the max that can go to the front, 65% is the max at the rear. Sure the observed f/r torque split may be 50/50 with diff in "locked" mode, but that's only because, at best, the torque is really 35% front / 35% rear / 30% diff clutches. The physical split at the differential doesn't change.

Welding the the diff shut *does* effectively change the "number of teeth" in diff, it makes the ration 1:1, not the normal 35:65. The clutch is not like a "variable weld". Even with the clutch completely locked up, the gearing in the diff still splits the torque 35/65... remember, even though the wheels are turning at the same speed w/ the DCCD clutch locked, there is still more torque being applied to the rear wheels than the fronts.

No it doesn't. If the clutch acts between the same two axis as the planetaries and effectively locks those two axis, the planetaries are no longer rotating and therefore not transferring any torque, the clutch is! The same as if it they were welded. 0% X35/65 = 0; 100% X 50/50 = 50/50.

Depending on the level of clutch engagement, you are somewhere between the two.

Anything that limits the rotation of any of the axis relative to any other axis inside the diff is going to affect there ability of the gear ratios to split the torque...

My dccd switch got reinstalled sideways, now my car has a 35/65 left-right split

On an otherwise crappy work day, I literally laughed out loud at this. Kudos to you Mike!

Imagine the open planetary differential. The power comes into the diff on the planet gears. The planetary carrier is attached to the rear wheels, the sun gear is attached to the front wheels. When there are no wheels slipping, the tooth ratio between the planetary carrier and the sun gear determines that the torque coming into the system on the planet gears goes out 65% to the planetary carrier, and 35% to the sun gear.

Now, let's say the rear wheels are slipping. In this case, the power takes the path of least resistance, and the planet gears easily accelerate the planetary carrier, effectively robbing the sun gear of it's power. i.e. in order for the front wheels to receive power, the rear wheels must be pushing against something. And of course the opposite is true, if there is no traction up front, the sun gear will spin freely and prevent power from reaching the planetary carrier.

Now we toss in the clutch. The clutch engages when there is a speed difference between the planetary carrier and the sun gear. So, in our example case, where the rear wheels have lost traction, as the planetary carrier gets faster than the sun gear, it is effectively braked by the clutch, allowing the planet gears to transfer power to the sun carrier. The clutch acts as a replacement for the lost traction at the wheels. The settings of the DCCD controller simply change how much speed difference is allowed between the planetary carrier and the sun gear before the clutch engages.

So, let's take the extreme example of 100% DCCD lock, and no rear traction: The power comes in on the planet gears, which attempt to push the planetary carrier at 65% of the input torque and the sun gear at 35% of the input torque. There is no resistance at the rear wheels, so all the power attempts to go through the planetary carrier, but because there is no speed difference between the wheels "allowed" due to the DCCD setting, the clutch engages to prevent the planetary carrier from spinning freely, thus allowing power to reach the sun gear. So, in this case, the front wheels are seeing 35% of the input, and the rear wheels 0%, and the clutch 65% of the power which is being converted into heat/clutch wear in order to act like ground traction.

In practice, this interaction between the clutch and gears is very dynamic, and ever changing, but due to the gear ratios in the gearset, you're never splitting the maximum torque differently than 35/65. At least that's the best I can understand from the documents Mike dug up.

I think the input is actually the ring gear, but it matters not. What I think you are missing is that the clutch is not like a classic engine clutch on a single shaft varying between 0=totally independent speeds and 100%=identical speeds, it is a clutch between the two shafts continuously being driven by the differential at nearly identical speeds. The gears of the differential and the DCCD clutch are acting in parallel, not series.

It goes from allowing the gears of the diff to spin freely at whatever rate the gearing calls for to not allowing any of the gears to spin as if all three axis/shafts were welded together.

Again, I am still talking about theoretical split, not actual based on traction. If the clutch locks any two axis, gear ratios are negated since none of the gears are turning relative to each other. And if it acts as resistance between any two axis, the split will tend towards 50/50 which is how they accomplish the variable.

Please reread Mike's document, specifically, page 7/64 of the document which clearly states:That is pretty clear IMHO.

I don't understand your use of the word "axis" in terms of the gearing. Everything is spinning about a single axis in the differential.

The documents explicitly mention that the power comes in on the planet gears, the ring gear (aka planetary carrier) drives the rear wheels, and the sun gear drives the fronts. When there is a speed difference between the sun and ring gears the faster gear is pressed into the clutch pack, slowing it and preventing a "run-away" use of the engine's power. The bigger the speed difference, the harder the clutch is used, with the DCCD setting acting like a gain value to amplify the clutch use (at 100% lock, any speed difference is translated into clutch use, at 0% lock the clutch is never used).

If the clutch really works the way you're implying, then anytime the you're driving, even w/ plenty of traction, if the DCCD is locking at all, then you're slipping the DCCD clutch. That thing would wear out in a year!

As far as the quote from the document, at 100% lock, and slippery condition, the observed torque at the wheels would be 50/50, but it would be less than the normal total torque at the wheels... since it really would be 35%/35% with the remainder of the power being wasted heating the clutch.

OK, input = planetary, but it doesn't matter all shafts are rotating in the same direction at almost identical speeds.

axis = shaft(input or output)

It does work the way I am describing, the speed differentials are tiny compared to what a normal clutch sees, so wear is minimal.

There is no heating the clutch at 100%, it is locked!

The other side of the clutch that the gears are "being pressed into" is on the the input shaft, not empty space, but again, it doesn't matter which shafts the clutch is generating friction between as long as it is between shafts, not into space. The friction between shafts shifts the bias towards lock up which is 50/50 theoretical distribution.

Yes, while it is slipping, it is generating some heat, and power loss, but if it was really the percentages you are talking about, then you would definitely burn it up in no time, and you would need an oil cooler on your DCCD the size of a radiator.

At 100%, the clutch is not "locked". The clutch only engages if there is a speed differential between front and rear wheels. At 100%, the clutch engages if there is *any* speed differential. At lesser values, the clutch does not engage until the speed differential is more significant, at 0% the clutch never engages. What this means is that the clutch *never* engages without slipping, and as the output shaft speeds start to match each other, the clutch disengages.

The "percentages [you're implying] I'm talking about" never occur because you never have a long period of time where the output shafts are at drastically different speeds.

Whereas, in your scenario, you've basically got some electric servo that slaps the clutch up against all three shafts attempting to "weld" them together, which is exactly like the clutch on the engine attempting to "weld" the flywheel of the motor to the input shaft of the transmission. So, in your setup, anytime the car is moving and >0% lock dialed in, you've got the clutch attempting to engage.

Then why did you get your DCCD controller fixed if all it can do is 35/65? :rolleyes:

Go reread pages 15/72-18/75.

If you still don't believe it can lock up the entire differential to make a 50/50 split then so be it.

Yes, it does not engage when it is not required, and it slips some or much of the time depending on all the inputs, so what, that's what clutches do...

The clutch only has to act between two sets of gears, in this case, sun and planetary carrier. (P74) which are traveling within a a few maybe 10s of RPMs and only has to handle 15% of the engine torque to get from 35 to 50% front as the gears themselves are already distributing it at 35/65.

A car clutch has to handle 1000-2000-4000RPM and 100% of the engine torque. Completely different.

Not that it matters much, but I'm pretty sure the ring gear (aka differential case) is the input, the sun gear is output to the front, and the planet carrier is the output to the rear.

There are two clutch packs - the pilot clutch and the LSD clutch.

The pilot clutch is engaged by an electromagnet. This clutch basically clamps the input (diff. case) to the rear wheel output (planet carrier). The clamping force on that clutch pack varies with the current that the ECU sends to the magnet's coil. This current determines the variable 0-100% locked. *EDIT TO ADD: With any planetary gearset is you lock (or clamp with clutches) any two of the elements together (the ring gear and the planet carrier in this case), the gearset all tends to turn as one element.*

The LSD clutch is in between the planet carrier (rear output) and a plate splined to the sun gear (front output). There are three balls on the other side of the planer carrier in cam pockets between the armature and the planet carrier. When a front/back speed differential occurs these balls force the clutches together with a cam action (the planet carrier is actually forced up against the sun gear with the LSD clutches in between them). This is sort of a hybrid of a locking differential and a clutch differential.

If the front and rear are turning at the same speed there is no slippage in the clutches, as everything is rotating together. Only when the speed varies between the front and back do the clutches slip, and the relative speeds are usually small, so the slippage is small and the heat generated is small. If you are doing donuts, driving on icy patches, drag racing, etc. then things will start to heat up.

When the electromagnet is turned completely off this creates an open planetary differential. The 35/65 front/rear "torque split" is determined by the gearing in the planetary gearset. As an example if you were to lock the rear wheels with the parking brake, the front wheels would still put out 35% of the engine torque and eventually start to drag the back wheels. This has nothing to do with clutches and everything to do with planetary gearing which I studied the shit out of in 1986 thinking I was going to become some hotshot autmotive engineer. :rolleyes:

I will add that if these clutch packs had teeth, and I've used similar toothed clutches in the past on manufacturing equipment, then when engaged the differential would become a pure transfer case. I am not sure how much torque it would take to overcome these regular slip clutches at the 100% lock setting, so I do not know how completely locked the diff. is, but I would guess it's pretty "locked" at 50/50. Dunno for sure though...

That's what I said originally but Scott tricked me into re-misreading P17/74. :)I agree with that, when I was talking about clutches, I was focusing on the electric/electronically controlled one, not the LSD functions.

If the electromagnet is turned all the way on locking 2 shafts, would you agree that the third shaft is also effectively locked and therefore the planetary gearing is no longer in play and the theoretical torque split is 50/50?

Yes, until the clutches slip, which they will at some unknown torque. Whenever you lock any two elements together in a planetary gearset, the whole thing turns at once - this is the kinda thing that happens for the 1:1 ratio in just about any automatic transmission.

This center diff. is like a regular positraction clutch-type diff, but instead of just springs preloading the clutch packs, there is a variable electromagnet and that ball cam-action too.