RE: VW's VAQ 'diff' explained

[anarki]

EricE said:

So what about the Golf R? Does it have a VAQ up front? LSD? None of these and just two open diffs with simulated locking via brakes? Or maybe a combination of some of these?

The golf R is four-wheel drive. This is controlled via a haldex clutch based system.

[xRIEx]

TooMany2cvs said:

Basically, AIUI, there's no diff at all in the gearbox - just a single output shaft.

I don't think so:

article said:

but is actually entirely independent of the diff cage

I read it as it transmits torque from one drive shaft to the other, completely external from the diff (so like a torsen effect, but with greater torque bias (and not contained in the diff)).

Edited by xRIEx on Wednesday 30th July 12:18

[anonymous-user]

There is the normal "open" diff in the gearbox. This additional system just bolts on the Offside of the transmission casing (where the 4wd cars bolt on their 4wd "spur" gearbox, and in fact using the same drive outputs (the 4wd version takes rwd torque from the diff crownwheel/cage))

[StottyEvo]

It sounds pretty handy to me, on a right hander more power will be transferred to the front left so aid cornering and traction. Win win. An LSD won't do this as effectively and an open diff will spin the front left up.

I can't really knock it.. Unless it breaks and is expensive to fix

[AJFord]

Great explanation above from Max Torque. Effectively all VW are doing is using the normal mode of a differential which assigns equal torque (not speed) to each wheel by virtue of locking or partially locking one (right) driveshaft to the crown wheel. This means that when the right wheel spins and the clutch engages then the bevel gears in the diff have a torque acting on them that transfers torque to the left wheel. If the left hand wheel spins then the torque is directly transmitted through the clutch to the right wheel and the effect is to reduce the spin speed of the left wheel as well as provide drive to the right wheel. I think the above is basically correct but the action of a diff is not entirely intuitive - at least to me.

[Clivey]

Max_Torque said:

With out the ability to "Overspeed" a wheel this system is not true "torque vectoring". it can only modify the torque split between the front wheels within the limits set by the wheel speeds as a result of the wheel paths (difference in radii of the driven wheels when cornering).

Just as their Haldex AWD (FWD transverse-based) systems cannot* "overspeed" the rear wheels, meaning that cars fitted with the system have always felt like front drivers.

• Though apparently the new systems (Audi TT) can...we shall see whether that translates into a wider dynamic repertoire or whether it's just marketing guff.

[RedSwede]

I don't think there's enough information on this system in the article to fully understand how it works - like for a start, is this the only "differential"? are the front wheels otherwise locked together?

But that hasn't stopped a dozen people chiming in about how the system is useless and they could design a better system on the back of a fag packet. Hell, given a decent lunch our and a reasonably sized envelope, they could design a better car.

I do agree though that VW should stop designing compromised cars, though. It is about time they produced something with the feel of an Exige, the weight of a Caterham, the all-weather usability of a G-Wagen, luxury of a Jag XJ, performance of an MTM RS6 and the commodiousness of a Transit. I think they could realistically sell that for a solid £22-£23k without it being overpriced. As long as they included all the essential stuff like Nav, audio, Xenons, heated seats, active suspension, cruise, head up display, etc as standard. And it looked no worse than a Miura, at least.

[Tickle]

From reviews I have read between cars offering elec diffs and mech diffs, the reviewer prefers the mech one as there is no guessing or pre-empting its effectiveness.

What advantage does this give over a good old mechanical LSD (as per Intgera r's, Focus RS, Megane R26, 275 etc)?
Is it cheaper?
More effective?

[aka_kerrly]

voicey said:

It sounds similar to the E-Diff fitted to the F430 - lovely bit of kit.

My interpretation to since the Ferrari system is also a separate 'box' which has a hydraulically actuated clutch system that is electronically controlled.

Ferrari have raved about the E-Diff and how it has improved the traction and handling balance by linking various systems so the diff knows how the car is behaving in terms of acceleration/braking/steering/wheel speed. I read somewhere that a E-diff equipped car was circa 3 seconds quicker around a track than the previous model sans diff.

[StottyEvo]

Clivey said:

Max_Torque said:

With out the ability to "Overspeed" a wheel this system is not true "torque vectoring". it can only modify the torque split between the front wheels within the limits set by the wheel speeds as a result of the wheel paths (difference in radii of the driven wheels when cornering).

Just as their Haldex AWD (FWD transverse-based) systems cannot* "overspeed" the rear wheels, meaning that cars fitted with the system have always felt like front drivers.

• Though apparently the new systems (Audi TT) can...we shall see whether that translates into a wider dynamic repertoire or whether it's just marketing guff.

I'm genuinely wondering, in this application why would you want to overspeed a wheel?

[RacerMike]

Clivey said:

Just as their Haldex AWD (FWD transverse-based) systems cannot* "overspeed" the rear wheels, meaning that cars fitted with the system have always felt like front drivers.

• Though apparently the new systems (Audi TT) can...we shall see whether that translates into a wider dynamic repertoire or whether it's just marketing guff.

Yeah....I'm not so sure whether it's just clever marketing speak or not. It's been widely reported that the GenV Haldex in the Golf R can 'distribute 100% of the torque to the rear wheels', but having read carefully through Haldex's own technical brief, this is clearly tripe.

The actual explanation is that, due to electronic control they can apply 100% locking torque to the rear diff (which I get the impression was not possible before with GenIV due to it being a passive system). There is no way a Golf R or a TT can ever be 100% RWD as the marketing quote suggests as it would require disengagement of the front diff which is no possible due to the way the power is taken off for the prop shaft.

So in conclusion, the Golf R and TT can only ever be 50:50 split....which is still better than the old system, but it certainly isn't 0:100 split...

[RacerMike]

StottyEvo said:

I'm genuinely wondering, in this application why would you want to overspeed a wheel?

That's the very definition of Torque Vectoring in an additive form. A basic torque vectoring 4x4 car with Yaw control can achieve Yaw control by over speeding the rear axle.

[aka_kerrly]

RedSwede said:

I do agree though that VW should stop designing compromised cars, though. It is about time they produced something with the feel of an Exige, the weight of a Caterham, the all-weather usability of a G-Wagen, luxury of a Jag XJ, performance of an MTM RS6 and the commodiousness of a Transit. I think they could realistically sell that for a solid £22-£23k without it being overpriced. As long as they included all the essential stuff like Nav, audio, Xenons, heated seats, active suspension, cruise, head up display, etc as standard. And it looked no worse than a Miura, at least.

haha, you know what, as sad as it is to say if VW designed exactly that and got it to market tomorrow the average mx5 loving PH would still whine and find a way to be negative.

If that exact car was built by a British company PH would still moan and say it's likely to be unreliable and wouldn't buy one.

If it were German it would be moaned at for not having character

if it were Italian millions would say they would buy it but only three actually would.

If it were American it wouldn't be able to corner.


And so on..

[Mr2Mike]

RedSwede said:

I don't think there's enough information on this system in the article to fully understand how it works - like for a start, is this the only "differential"? are the front wheels otherwise locked together?

There's plenty of information, and others have already summarised it.

The car has a standard, open differential within the transaxle. This device is simply a clutch which links one of the driveshafts to the outside cage of the differential. With the clutch open the system behaves exactly like a standard open diff, because it is one.

With the clutch fully engaged, the diff is effectively locked. If that's all it did it would simply be an electronic diff lock, but the advantage of this system is that the clutch can be progressively actuated to provide a greater or lesser degree of limited slip.

It's neither complicated nor especially clever, it's a reasonably simple and cheap (though not to the end user I'm sure) way of providing the benefits of a mechanical LSD without the downsides.

[Clivey]

RacerMike said:

Yeah....I'm not so sure whether it's just clever marketing speak or not. It's been widely reported that the GenV Haldex in the Golf R can 'distribute 100% of the torque to the rear wheels', but having read carefully through Haldex's own technical brief, this is clearly tripe.

The actual explanation is that, due to electronic control they can apply 100% locking torque to the rear diff (which I get the impression was not possible before with GenIV due to it being a passive system). There is no way a Golf R or a TT can ever be 100% RWD as the marketing quote suggests as it would require disengagement of the front diff which is no possible due to the way the power is taken off for the prop shaft.

So in conclusion, the Golf R and TT can only ever be 50:50 split....which is still better than the old system, but it certainly isn't 0:100 split...

I have seen claims by some manufacturers that they've achieved a rear-biased torque split with Haldex AWD. IIRC Vauxhall claimed the Insignia SRI 4x4's default torque split was 95/5 F/R in "normal" mode but could be set to 30/70 F/R in "sport". I'll have to see if I can find a copy of the brochure.

EDIT: Found it. - The claim was 40% Front, 60% rear. Here's the page from the brochure:

Edited by Clivey on Wednesday 30th July 12:55

[anonymous-user]

RedSwede said:

I don't think there's enough information on this system in the article to fully understand how it works - like for a start, is this the only "differential"? are the front wheels otherwise locked together?

But that hasn't stopped a dozen people chiming in about how the system is useless and they could design a better system on the back of a fag packet. Hell, given a decent lunch our and a reasonably sized envelope, they could design a better car.

Er, there is plenty of info to understand how the system works!

The car has a normal differential in the gearbox casing.

Because 4wd versions are availible, which use the Haldex system, space exists "behind" the engine block on the offside (RHS) of the transmission casing (this is where the "spur" or "bevel" gears live to send drive torque backwards in the 4wd versions). Also, because the 4wd versions need to take power directly from the front cross axle crownwheel rather than one of the front halfshafts, convienently there is a nice easy location for the new smart front LSD to also do that.

So, we have a system of variable "friction" clutches, modulated by hydraulic pressure from an electric pump, that can modify the "locking" action (the Slip) between one of the halfshafts and the crownwheel / diffcage. (exactly as in a conventional LSD)

But, we have no epicyclic overspeed capability, so all we can do is to either have both front wheels turning at the same speed (fully locked) or allow them to turn at different speeds (fully open).

In a turn, the different wheel paths result in a wheel speed varriance, and the lateral G results in a tyre normal load difference. As such, for a conventional differential, which MUST maintain the same torque for both wheels (set by the lowest denominator), if the driver applies a drive torque higher than that set by the inside wheel (the lightly normally loaded one) that wheel will overspeed, resulting in that excess torque simply being used to speed up that wheel/tyre and not to drive the car forward.

With a locked diff, the limit becomes that of the outside wheel/tyre, which will be significantly higher due to the weight transfer onto that outside tyre (latG x CofG height x front mass). This means the car can longitudinally accelerate harder than one without an LSD.

With the locking system, the diff will act to EQUALISE the speed differential across the front axle, and as a result the inside wheel will overspeed. However, this will not result in a high longitudunal acceleration (assuming the tyre is full loaded and is at the "peak" of its slip/force curve). In fact, this will result in less longitudinal drive force on the inside wheel. However, the outside tyre can accept this higher drive torque, and produce a high longitudinal drive force, and as such the inside wheel drive force falls and the outside one rises. This is the "effective" Torque Vectoring that they are talking about.

It does however rely on the driver applying more throttle, and hence more drive torque, themselves (unlike a proper torque vectoring system that can do this with the torque it is currently supplied with). The result of that is probably what Dan is talking about as a "slighty none intuitive" need to apply more throttle that seems normal to get the front diff to really "work" in his Gti/Meganne piece


As i said before, the biggest benefit of the electronic control of slip is that your input parameter matrix become very wide indeed. For a conventional LSD using ramps or sprung plates etc the input parameter is primarily drive torque (both positive (drive) and negative (overun). With electronic control, you can apply modifications to the wheel speed split (and hence vehicle yaw authority) based on things such as vehicle speed, driving mode (Normal, sport, eco etc) and of course multiple inputs from the dynamic stability control system (for example, keeping an open diff with no handwheel feel corruption most times, but locking it hard when yaw stability (and phase lag) become critical, such as in an emergency lane change etc

[P4GNO]

It sounds like a developed version of the electro-hydrolic diff system from the Porsche 959 which was then used on the 928 from around 1989. That was based on electrical inputs from wheel sensors and an accelerometer that in turn used a pump and slave on the diff operating clutches to achieve from 0-100% lock.
This system appears to do something very similar but applying it after the diff.

[CrutyRammers]

A great explanation from Max_Torque as usual

[neil1jnr]

anarki said:

EricE said:

So what about the Golf R? Does it have a VAQ up front? LSD? None of these and just two open diffs with simulated locking via brakes? Or maybe a combination of some of these?

The golf R is four-wheel drive. This is controlled via a haldex clutch based system.

I haven't driven the GTI PP but I have the Leon 280 (same diff) and Golf R.

Put it this way, if I were buying a golf it, would be the GTI just because of the diff. It just work's brilliantly.

[RacerMike]

Clivey said:

I have seen claims by some manufacturers that they've achieved a rear-biased torque split with Haldex AWD. IIRC Vauxhall claimed the Insignia SRI 4x4's default torque split was 95/5 F/R in "normal" mode but could be set to 30/70 F/R in "sport". I'll have to see if I can find a copy of the brochure.

EDIT: Found it. - The claim was 40% Front, 60% rear. Here's the page from the brochure:

Edited by Clivey on Wednesday 30th July 12:55

I don't believe that's Haldex though. Certainly not if it's got a separate rear diff. The Haldex system has no center diff and hence cannot split torque between front and rear. It can only add torque to the rear as a proportion of what's coming from the front.