Adam M

Following on from a long argument on this thread over on 22b.com (forgive the plug it is the best way to give the history) I asked simon if he would ask Damian Harty to comment.

To explain whats going on in brief summary.

I am of the opinion that in the absence of a dccd, the torque split front to rear of the impreza is 50:50. However, I have heard from several respected dealerships that via second hand knowledge of development carried out by prodrive among others, that they have altered the cars torque distribution and as such handling, by changing the drop gear ratios in the diff.

This is the reason why the uk turbo cars have a 3.54 rear crown wheel and pinion but a 3.9 at the front. The final drive ratio is the same because the pinion driving the rear diff is turning slower than the pinion driving the front diff as a result of the drop gears being in the ratio of 11:10.

jap cars have the same drop gears and the same crown wheel and pinion ratios front and rear.

The thing is, the car must have the same final drive ratio front and rear else the centre diff has to do a lot of work to allow the wheels to drive without slipping, this normally causes it to fail (ask Harvey after his was repaired wrongly!).

I am happy to accept that if the front and rear overall final drive ratios were different then there would be a torque difference, but overall the final drives are always the same.

This begs the question, why do IM (I think) convert the uk cars (not sure about euro cars) so that they have 1.1:1 drop gears and convert the front and rear final drive ratios back to equal by having differentl crown wheels and pinions.

It has been suggested that it is to lower the speed of rotation of the prop shaft, but most people dismiss this as nonsense.

Damian's reply was......


"Simon, Adam,

This is a good question and one that took me a while to get to grips with when I first encountered it. It's all to do with where the drop gears are in the overall scheme of things. For my Ascii diagram to make sense you need to make sure it's viewed in Courier font:


Engine
|
Centre Differential
| |
Front Diff Drop Gears
|
Rear Diff

The overall ratio at the rear is the cumulative effect of the drop gears and rear diff. Normally it's set to be pretty much the same as the front diff reduction ratio. I say "pretty much" because there are always differences in tyre laden radius with the uneven weight distribution in the car that produce effects of a few % floating around. In other words, although the rear diff has the same ratio, the overall ratio has changed. Some people call Drop gear+diff ratio "final drive ratio" and others call the diff ratio alone the "final drive ratio" - it's all a matter of convenience.

While a different reduction ratio in the crownwheel & pinion (or drop-gear + c&p combination) does redistribute torque, it only does it in a relatively small way. An epicyclic centre can deliver twice as much torque to one shaft as another; to acheive that you'd need double (or halve) the final drive ratio at one end of the car or the other. It would give an impractical difference in drive ratios to attempt to use them as the sole means of torque distribution, hence the invention of epicyclic diffs.

To the best of my knowledge there are at least three torque splits in Imprezas: a 50:50 bevel centre differential, a 45:55 epicyclic and a 30:70 epicyclic. Again to the best of my knowledge, the latter comes with the latest high capacity DCCD-A units, but I have no idea which markets they are for. Mike Wood is the man to quiz on that sort of detail...

So the answer to the question is that there's no reason in principle you couldn't redistribute torque with the drop gears, but attempting it in practice would wind up the centre visco (if fitted) or DCCD to a fairly torturous extent and render the whole thing a bust - not to mention the gear sizes you'd need might not fit in the casings any more.

Note that these comments apply to current road-going Subaru Imprezas and are not by any means a general principle applicable to all 4x4 drivelines. The WRC has drop gears in a different place so that changing them changes the final drive ratio of the whole car, for example.

Not sure if that makes sufficient sense - let me know if it doesn't.



Regards,



Damian"

Damian seems to have confirmed what i thought that the overall drive ratio is the same giving no torque split in the absence of a dccd but it doesn't explain why UK cars have their drop gears changed from standard, Damian is going to ask the people who do it to find out.

But he did say that if anyone wants to have a follow up question and answer, then he would be willing to oblige.

johnfelstead

Hi Damian, thanks for taking the time to answer the questions.

I have a query about you stating you would need an epicyclic centre diff unit to achieve a torque split diference. I understand fully how the DCCD achieves its torque split and how that varies as you increase clamping load on the clutch plates that make up the LSD section of that unit. You have a 33.3:66.6 F:R split on the DCCD unit used in an Impreza when the clutch plates are fully unlocked which is acheived via the 2:1 gearing of the epicyclic diff assembly being fed to the same ratio final drive units via a 1:1 drop gear on the rear prop drive.

Mike Wood has stated to me on a number of ocasions that the 55:45 F:R torque split is used on the UK Viscous Coupling equipped European spec (and some US spec) cars and is achieved by using the drop gear and rear final drive ratio change.

What you are sugesting, if i understand you correctly, is that the drop gear and rear diff ratio change isnt being used to achieve this. This obviously contradicts what Mike has always told me, but you are also contradicting what i know about Impreza drive assemblies, which is that they only come with 2 configurations of centre diff unit. You either have a VC or a DCCD, there isnt a 3rd option (epicyclic that provides 55:45).

Just to make things clear and hopefully aid the debate.

The non-DCCD JDM spec (and some US spec) transmition systems uses a VC centre diff and a 1:1 drop gear with same ratio front and rear diffs. (i believe this is 50:50)
The Euro spec (and some US spec) transmition systems uses a VC centre diff and a 1.1:1 drop gear with diferent ratio front and rear diffs. (i believe this is 55:45)
The DCCD spec (all markets) transmition systems uses a epicyclic centre diff and a 1:1 drop gear with same ratio front and rear diffs. (i believe this is 33.3:66.6)

Damian Harty

I'm still working on a definitive answer to all this, but I do note my "Ascii diagram" has gone astray... here is what I meant it to look like:

johnfelstead

Hi Damien, the diagram makes sense now.

RT

Great discussion - let me add an observation.

I know of two Bugeye STI7 type UK's (265hp) that have had trouble after installing an aftermarket rear diff. One had put on a Cusco type RS, the other had installed a STI mechanical diff (as seen on a Spec C). No alterations to the center diff area were made.

When road speeds reached around 110-120kmh in a straight line, the rear of the car started bouncing. Began as a minor oscillation that rattled the coins, and quickly progressed into a violent banging until the drivers slowed down. Sounds awfully like center diff "wind-up".

I am puzzled as to why or how a rear diff change ONLY, can result in the above. Somehow removing the suretrac rear, has resulted in the center diff winding-up. Any ideas behind the mechanics of this?

Cheers.

EMS

That would explain the terrible understeer accelerating out of a corner.......

Mark.

mutant_matt

Well, one of the reasons anyway Mark

Nice discussion guys, I must have missed that one on 22b.....

Matt (unable to contribute to the debate at this level )

johnfelstead

sounds like they have the wrong diff ratio in the back axle RT.

But you should be thanking Subaru for keeping you safe with a nice understeer based handling charicteristic Mark.

RT

John, only the rear LSD was replaced. The crown wheel (is that what its called?) was transferred over from the old unit. Could a wrong rear diff ratio be the case? On two separate cars?
:O

johnfelstead

are the drive shafts interchangable and locking into the LSD units properly?

EMS

LOL @ John

carlos_hiraoka

Here is my input for this thread. A while ago Subaru Peru sponsored the Impreza Turbo GT class for rallying, basically they were stock euro spec MY98 WRX's with 220bhp. As the class became more popular some mods were permited. Drivers complained about the understeer behaviour of the car and the following mods were done to make it better: replace the drop gear for the 1:1 ratio ones, change final drive of rear diff for 3.9, and finally use a plated diff (STi unit for R160). This improved the car a lot, only problem was that most of this cars evolved into group N machinery which lead to the end of the this class (a shame really). And because of the stock of drop gears and 3.9 rear diffs imported is that it also became a common upgrade for many road cars here.

So in the long term if anyone has the time / $$$ to do it, it may be interesting if he could report this, cause according to D. Harty a car fitted with a 1:1 drop gears and a 3.9 final drive should be the same as one with a 1.1:1 and a 3.54 final drive .....

David_Wallis

interesting thought carlos..

I wonder if I can get a 1:1 from a jap 6spd to shove in my uk 6spd. and replace the rear with the r180 3.9:1 with better hubs and calipers?

carlos_hiraoka

David, u should be able to get the 1:1 drop gears (stock or straight cut/Modena) easily, and a 3.9:1 for the R160 ..... no need to replace the whole rear end as the R180 rear end (at least on the older GC8) uses bigger axles, hubs, etc.

Any aditional news on why would the 1.1:1 drop gears change the torque distribution ?????

Carlos H.

johnfelstead

Any update on this Damian?

greasemonkey

Thought a fair bit on this lately, and my current opinion is as follows:

The Impreza's bevel gear centre differential has a 50/50 torque split.

The 1.1:1 drop gears and unequal diff ratios cause a rotational speed difference between the front and rear centre diff outputs under normal "straight line" driving conditions.

This rotation speed disparity will also preload the viscous coupling, which will then try and equalise the rotation speed of the outputs. It is this effect that will result in a slight FWD bias when traction is broken.

Comments anyone?