posted 09 January 2003 12:51               

I brake in a straightline and turn in whilst coming on the power... oversteer city..
but this seems to be the recommended method..
ie. brake, turn in and power through and out of bend..

If I brake, turn in and then wait a split second and then add the power it goes round with no problem.. and no oversteer..
but it is more like brake, turn in and power out of the bend then, rather than powering through.

Just had new tyres fitted and four wheel alignment and haven't had much of a chance to play since so might have been the alignment and tyres that were the problem..?

How should I be doing it?

posted 09 January 2003 13:59               

sorry for the confusion.. it goes straight on rather than around the bend...

doh doh doh..

Sorry

posted 09 January 2003 14:56               

been tempted to try some driving courses but never got the time or the money to do it..

posted 09 January 2003 21:44               

Go deep into the corner bury the stop pedal to get the front end to bite as the front goes down I then plant foot on loud pedal to catch the unweighted rear end and get it up to speed and drift through.

Lots of fun this weather as breakaway is much easier

N.B. I always cadence brake as that is my reflex braking style, this gets tail gaters to back off and allows you to feel the grip of the road befor you get to the turn in point.

posted 10 January 2003 13:45            

the rear diff on the uk car has a 3.54:1 final drive, the front diff is a 3.9.

posted 10 January 2003 14:18            

is not like they dont make the bits, the sport has a rear diff with a 3.9 crown wheel and pinion bolted to it and so doesnt need a 1.1:1 transfer case, which is why it has a 1:1.

posted 10 January 2003 23:49            

Possibly (probably ? Engriesh is not my first language you know) because I'm a pretty crap driver, I'll grant you that

posted 11 January 2003 11:40               

It is possible to boot the tail out coming out of tight junctions and hairpins

On wet roads on P zero it was quite easy to get the tail out.

With S02 in the wet it was like driving with P zeros in the dry!

With my engine modifications the car is now too powerful. The front end just scrabbles away the power so I have had a Lateral performance ATB dif fitted at the front to tie the front end down

posted 11 January 2003 22:13               

EU cars will have trouble getting the back end out because of the 50:50 torque split (pretty much anything that doesn't have the electronic centre diff), I remember it used to take shedloads of welly to unsettle the back end out of tight corners with the stock STi transmission.

As to understeer, it's difficult to overcome with the 50:50 split, I found that it's down to commitment and available torque, the more of each available, the more likely it is to break away

Cheers,

posted 12 January 2003 13:30            

I'm hoping to to something about this, and wondered (Eur. MY99) if :

. I could find a fitting Impreza DCCD
. I could find the Cusco 35/65 diff
. I would benefit from the STi 20kg/m diff

posted 12 January 2003 18:32               

it's the 50:50 torque split that causes the problem, along with the centre VC LSD. To start with there's an equal torque distribution, and should either end decide to break away, it will be "calmed" by the VC LSD.

When you're under power, there is weight transfer to the rear of the car, increasing the coefficient of friction between the rear tyres and the road, while reducing it at the front. Given that both front and back receive the same amount of torque, the front is going to break away sooner than the rear.

At this point obviously the torque distribution swings to the back (the front isn't capable of taking any more, so it has to go to the back!). The reason you can get more than 50% to the back is because the VC LSD is trying to maintain equal axle speeds, the only way it can do this is to make the front torque less than the rear. The 50:50 ratio is *ONLY* the natural tendancy of the diff, in operation it is hardly ever 50:50, it varies greatly, and the outer limits will depend much on the diff type. A VC type will be limited by the fluid, both its film shear strength and its temperature tolerance.

The DCCD on the other hand is a planetary differential with an electromagnetically controlled viscous coupling. Its natural tendency is to put more torque to the back; this is why 22Bs must have the diff locked on rolling roads, otherwise the rollers will spin up at different rates! In addition to the fact that the DCCD naturally puts more torque to the back, when the control is in the "open" position the VC plates are far apart, allowing different axle speeds with ease, so when an end does break away, it won't be "caught" by the other end, as in the case of the normal diff. With the control wound up a little, the VC plates are pulled closer together and so the over-steery tendency is reduced.

People often have trouble understanding the dynamic torque split. The following illustrations should help....

Put two cars on dry tarmac with warm slicks, and accept that neither car has enough torque to spin any of the wheels, nor to compress the tyres and thus reduce the rolling diameter. Now install strain gauges on each of the driveshafts to measure the torque. Now let them do a 1/4 mile run. The car with the DCCD will register more strain on the rear shafts than the front, it's literally pushing harder at the back than it is pulling at the front. The car with a normal VC diff will show the same strain on both front and rear.

Next assume it's got pretty darn cold, and there's a big patch of ice under the front wheels, get both cars to do the same thing again. The car with a fully open DCCD probably won't even get off the line, as the front wheels have no traction, they will spin. As there is no traction, there is no strain on either the front or rear axles (OK, in reality drivetrain losses and the VC not being 100% open would make it move but it's just to make a point). Now assume that the driver knows about the ice, and he fully locks the DCCD. Now the front and rear axles MUST turn at the same rate. Since the front tyres have no grip at all, 100% of the torque goes to the back, then as the front wheels get onto dry tarmac, we have a 36:64 split as before, then the rear tyres cross the ice, and we have 100% to the front and nothing to the rear. Finally the car clears the ice and shoots off down the strip.

Now consider the car with the normal VC centre diff. As that driver begins to drive off he too has no grip at the front, so 100% of the torque goes to the back also, but in his case the front wheels are turning faster than the rear ones; this relative speed is necessary for the VC to work, it needs the film shear resistance from the fluid to lock the diff up; given the same gear ratios as the DCCD equipped car (with diff locked), this car would be travelling slower off the line. It manages to clear the ice with its front tyres, and hits dry tarmac. THe front tyres are spinning faster than the car is moving, but since the tyres now have perfect traction, they immediately slow down to the car speed. Given that the engine speed is still the same (from the infinitessimally small amount of time between clearing the ice and hitting the dry tarmac), the only way for the front wheels to slow down is speed up the rears, or slow down the engine. There will be a large shock, which will be detected by the strain gauge. After the driveline shunt the car continues to move, with all four wheels travelling at the same speed as the car. Now the rear wheels lose traction as they hit the ice. We get 100% of the torque to the front, and none to the rear as the tyres cannot grip. There will be a "blip" where there is torque to the back which will spin the rears up a little, creating the difference in speed needed for the VC to work. As the car clears the ice there is another driveline shunt and it's back to 50:50 split.

These are extreme cases, in reality the torque split would never extend all the way to 100% because of frictional losses etc, but just for illustration purposes it will do. Now it gets more complicated...

With a VC diff obviously the fluid will determine the eventual torque split. If you put a million footpounds into the diff in the case of the ice, you can be assured that you won't get one million foot pounds to one end and nothing to the other, instead what will happen is that the diff will allow greater and greater difference in axle speeds, with greater and greater torque going to one side, but sooner or later, the difference will be so great that the engine has hit the rev limiter and can't turn any faster. So in reality there will be a limit to the amount of torque one can get to either end; this will be down to the number of plates and the type of fluid. For example STi centre diffs come in 12 and 20 kg versions; if memroy serves correctly they can deliver either 12 or 20 kgfm (117 or 196 Nm) more torque to one axle than the other. If this doesn't sound like much then please remember that an EU car has just over 200 lbft, and its fifth gear has a ratio of 0.738:1 so that's 147 lbft going into the centre diff. 196 Nm is 144 lbft, so that diff could cope with the entire torque output from the engine going to one axle, in fifth at least. In first you've got about 4 times as much torque at the input to the diff, and so it couldn't contain all the available torque, so what you get is massive wheelspin at the front while the back end grips, unless of course you can light up the rear, then it would revert to pretty much 50:50 split.

A quaife ATB (based on the Ford TORSEN [TORque SENsing]) needs a difference in axle torques rather than speeds to begin to lock; it can distribute more torque to one end of the car than the other even before any wheel breaks away, but it does need some torque on both output shafts in order to work. They are available with different torque ratios, but figures round 5:1 are common. This means that it can only load 5 times more torque into one output than the other, if one output has no grip whatsoever then it will behave like an open diff. So it would be quite useless on ice, but great on tarmac.

TORSENS are installed on some off road 4 wheel drive vehicles. It is interesting to read the owner's manual... obviously with one wheel off the ground, it can't take any torque (ie the strain gauge would always read zero), so the 5:1 ratio cannot be satisfied. End result is that the vehicle cannot move, which obviously isn't much good for an off-roader. The solution is to lightly apply the brakes, this gives the axle feeding the wheel in the air something to fight (think of what the strain gauge would show), and of course the diff will then allow 5 times that much torque to the other wheel, but obviously it needs as much torque to rotate at all because of the brakes, end result is that it has 4 times as much torque available to move the vehicle as is being absorbed by the brake on the airborne wheel.

Now we could go on to electrohydraulic active differentials but I think I'll leave that one for another day

Cheers,

posted 13 January 2003 08:43               

I go home for the weekend and come back and there is a novil in my thread!!

Interesting stuff.. all of it..

Never thought about applying the brake when one wheel spinning etc..

My car is actually much better now.. set og Pzeros and tracking set correctly and it is far more confident than with the Falkens I have and tracking out.. home much is the tracking and how much is the tyres orbiously I don;t know but now I can actually steer in the wet without understeer at anything over running speed now!

posted 13 January 2003 10:05               

I assume a viscous diff is fitted because it's cheaper than a sun 'n' planet type diff.

posted 13 January 2003 10:39            

It now makes sense why under cornering - slow or fast - as soon as the fronts break out under power, there isn't really as much torque transfer to the rears as I would have expected it. Giving it more welly in that case only makes the feeling worse and seems to slow the car down instead of accelerate.

If I understand it correctly, even though being 50:50 as well, the STi 20 kg CD would actually allow more torque differences and partly help in this case. Is this correct ? If not, off I go to read it again But it would still not turn the car into an oversteery monster, as that is - I understand that now - a function of the planetary CD.

posted 13 January 2003 14:14               

You have mistaken the actual differential's workings (D in LSD) and the limited slip coupling used with the differential (LS in LSD).

The DCCD unit is an epicyclic or planetary gear differential unit (the D bit) with an electronically controlled viscous coupling providing the LS bit.

Euro cars have regular 50:50 differential ("D") units coupled with viscous coupling "LS" bits.

The D bit dictates the natural torque split and caters for different axle speeds, which is nice when you need to turn corners etc . The LS bit steps in and throws torque about when differences occur between axles in terms or rotation speed or torque, depending on the type of LS unit used.

Evil,

On another note: personally, I have never found the ATB to be limited by slippy stuff... if there is a little grip available on both wheels, it does the job... if you have no grip, you have no grip, at which point putting down more power should really be the last of your priorities. If you have trouble moving off on really slippy stuff, you can always use some gentle left foot braking to provide the diff with an anchor to push the better traction wheel against, as they recommend with the "HumVee" or "Hummer" (The silly US army jeep type thingy that makes range rover drivers feel under-endowed), which uses, IIRC, torsens all around.

posted 13 January 2003 14:27               

The different ratios on the front and rear diffs (sorted to same final drive by centre diff) must give slightly difference characteristics.. if you have ever riden a push bike with lots of gears you can put the bike in the same drive equivlent gears in two ways, large front gear and large rear gear and small front gear and small rear gear (obviously selected to match) and the effort needed to pedal is slightly different and the effect on the rear wheel when traction is on the limit is different too...
Perhaps this is the difference between torque and power??

Also the spin characterists of the wheels momentem on the axle must have an effect?

posted 13 January 2003 14:31               

Getting closer. Probably spot on with regards to ATB design.

DCCD allows locking to be varied between bugger all and significantly more lock than even an uprated 20kg sti unit could provide. The more lock the viscous diff has dialed in, the faster the transfer should occur, iirc.

Moray

posted 13 January 2003 14:40            

I know because my company was responsible for defending the ford torsen diff in the first place.

it is the viscous coupling in the euro diff which may be slow to respond.

the vc in the dccd cannot be slow to respond because it doesnt act passively.

If there is any coupling it will have been set in advance of created speed difference and affect them accordingly.

The greater the lock, the more restricted the speed difference between them.

The actual transfer of torque from 50:50 when gripping on a straight to 65:35 when an axle is free to spin, is via a mechanical interaction of gears and will be too fast to be detectable by any mere mortal.

Of course 65:35 is based on the gearing in the diff, and only represents the torque transmitted to the axles, when either of therse axles slips, the fact that a percentage torque is being applied is irrelevant, as all the torque (minus frictional losses) can only be transmitted via wheels with grip. No reaction force, no torque trasnmitted to the ground.

iff the diff is locked up, regardless of the gearing of the planetry diff, the front and rear axles will receive 50% of the torque each. again as stated above, if one axle cant grip enough to transmit that to the road, what it cant transmit will automatically go to the rear axle until that one slips too.

This is why the rear wheels spin on type rs.

they get 65% of the dynamic torque when attempting to pull away, once they slip, they speed up relative to the front wheels, because there is no limited slip nature unless you apply it using the dccd they will just spin whilst the front wheel remain stationery.

posted 14 January 2003 08:14               

However, also got a chance to test the Quaife (my00 Euro car) in these conditions a little - it exhibits exactly the same tendencies as I've found on track.

Once the front is sliding because of too high corner entry speed, it's all over - more power doesn't help, the rear starts sliding too late to help out. It becomes a front biased (and uncontrollable) drift.

However, with correct entry speed & then gas (after the weight transfer has occured) the rear breaks traction first....and a beautiful slide

What I'm unsure of is whether the 20kg/cm centre will significantly help things - it's not changing the torque distribution front/rear until one end has lost grip. I guess it depends on just how much quicker it reacts?

It feels to me that the ideal solution would be to actually remove some of the power from the front wheels.....this would give them less to do when nailing it on the apex with too high entry speed.

posted 14 January 2003 15:33            

didnt know you get dynos with the torque from the fron and rear wheels separately measured.

This begs the question what are you measuring when the front and rear wheels are allowed to act independently, ie by unlocking the dccdiff on the type r/ra. Does the rolling road assume that of the front and rear wheel act separately then it should be in 2wd mode? If so, I would love to lock the front wheels down, and open it as completely as possible to see what it would produce!

The only way I know of alterning front to rear torque distribution is by a hydraulic (like a torque converter in an automatic) or an epicyclic (planetry) centre diff. I dont recall the legacy having either, so I am not sure how you would get a measured torque spread between the front and rear axles.

can they not produce 4wd rolling roads with the front and rear rollers geared to eachother? or belted? too much drag I guess!

posted 14 January 2003 22:30               

I was surprised when I saw it - I thought all manuals had 50:50 split - then later on Snet I noticed someone else stating that wagons (he was talking about Imps) had rear bias to cope better with the extra rear weight. From memory the rear torque is nearly twice that of the front.

Dyno was AVA Turbos, Glasgow. They also showed me the plots from a 4 door Imp and the front & rear were near enough 50:50.