torque vectoring in AWD Model S

[Todd Burch]

I doubt the 4-motor drivetrain is coming soon. Possibly never. Why?

1. It would very likely require a complete redesign of the drivetrain skateboard, which is obviously being reused (to good effect) between the Model S/X and in (perhaps) slightly modified form the Gen3. Tesla's ability to reuse the skateboard platform on different cars is a huge benefit. They anticipated dual-motor AWD when designing the skateboard, and certainly went through a 4-motor design before deciding it wasn't worth it.

2. Torque vectoring, as I believe JB referred to it, is probably good enough to surpass any existing systems out there--and probably good enough for just about any situation. Sure, you could have 4 independent motors to introduce very fast responding, completely independent torque...and it certainly would give better traction than a 2-motor system...but...

A 4-motor system would likely mean:
-Higher cost
-More weight
-Higher complexity
-Lower reliability (although granted electric drivetrains can be made to far surpass mechanical drivetrain reliability)

Given that the benefits of a 4-motor system over a 2-motor system might be useful in 0.0001% of the driving...it's just not worth the value proposition.

It's like driving an SUV around 100% of the time when you only need all that seating/cargo capacity 0.01% of the time. Or driving a specialized offroad vehicle when you never offroad. You just don't need all that stuff.

 

[Raffy.Roma]

That sounds like snake oil to me. Providing a 'basic' torque and then braking individual wheels to reduce the torque on those wheels, which is what stability control does, is equivalent to actively controlling the torque on each wheel. No matter what the approach is, both ESP and torque vectoring lead to different torques on different wheels.

Don't agree. ESP (or ESC is the same) only allows the brakes to be activated not the torque on each wheel. On the contrary Torque Vectoring also allows the torque to be applied on each wheel IMO.

 

[Todd Burch]

Regardless, let's see how the AWD performs. If JB's demeanor was any hint a few months ago when he briefly mentioned it, it sounds like the Model X AWD system will clearly surpass Audi's system, and there will likely be a new market leader for AWD performance.

 

[Raffy.Roma]

Regardless, let's see how the AWD performs. If JB's demeanor was any hint a few months ago when he briefly mentioned it, it sounds like the Model X AWD system will clearly surpass Audi's system, and there will likely be a new market leader for AWD performance.

Then it means that Tesla is going to perform Torque Vectoring on the Model X as I thought. Can't wait to have the confirmation of this from Tesla.

 

[Johan]

Then it means that Tesla is going to perform Torque Vectoring on the Model X as I thought. Can't wait to have the confirmation of this from Tesla.

You just made a definition of true torque vectoring in a post just above and now you believe the X will have this feature. Again: How with two wheels connected to one motor with one fixed gear???

To pose the question even simpler: how can one of the motors put torque on only one wheel (i.e. Without "cheating" by breaking on the other)?

 

[Todd Burch]

Then it means that Tesla is going to perform Torque Vectoring on the Model X as I thought. Can't wait to have the confirmation of this from Tesla.

JB Straubel has already spoken publicly about torque vectoring on the Model X, so we know it's there. That doesn't necessarily mean it's 4 independent motors though (and of course in the case of the X, we know it doesn't).

Remember an advantage you get with electric motors is the ability to adjust torque not only very precisely, but also at very high frequencies.

 

[Raffy.Roma]

You just made a definition of true torque vectoring in a post just above and now you believe the X will have this feature. Again: How with two wheels connected to one motor with one fixed gear???

To pose the question even simpler: how can one of the motors put torque on only one wheel (i.e. Without "cheating" by breaking on the other)?

I have more an electronic background rather than a mechanical background, so maybe that a mechanical engineer could answer to you better than me. Anyway I suppose that the wheels (both in the front side and in the rear side) are connected to the motor in an independent way thanks to a differential. This way the motor could apply the torque on each wheel in a completely independent way.
Of course this is not compatible with your view of the wheels connected to one motor with one fixed gear.
Hope that Tesla wil give further details on this matter soon.

I would like to add that in case Tesla managed to perform Torque Vectoring on the Model X (and the Model S of course) we would have an AWD System that would be much much better than for instance Audi, because the commands concerning the torque on each single wheel would be executed much faster. In fact electric engines don't have the inertia of ICE engines and have the torque available immediately.

- - - Updated - - -

JB Straubel has already spoken publicly about torque vectoring on the Model X, so we know it's there. That doesn't necessarily mean it's 4 independent motors though (and of course in the case of the X, we know it doesn't).

Remember an advantage you get with electric motors is the ability to adjust torque not only very precisely, but also at very high frequencies.

Agree. At every speed electric motors are able to adjust the torque to keep the wished path very rapidly.

 

[vgrinshpun]

Just to clarify this subject, as it was discussed some time ago in another thread.

Since Tesla AWD system has only two motors - one per axle - the speed advantage over any AWD system in an ICE is in much faster and more precise control of total torque delivered to open differential that is in turn connected to both left and right wheels.

The motor can not control torque to each wheel, however. The torque control to an individual wheel is achieved by applying brake to a wheel, reducing torque on that wheel and forcing open differential to augment normal 50%-50% torque split between left and right wheel, transferring torque to the wheel on the opposite side of the car. This part of the system is similar to conventional AWD systems based on open differentials. I had similar system in Toyota Sienna minivan a while back.

The fact that Tesla Model S has open differential is mentioned by Edmund.com in the following link: 2012 Tesla Model S Road Test

"The 2012 Tesla Model S is powered by a transverse rear-mounted liquid-cooled AC induction motor that is capable of 16,000 rpm. It drives the rear wheels through a single-speed transaxle with a 9.73-to-1 reduction that feeds an open differential."

 

[Raffy.Roma]

Since Tesla AWD system has only two motors - one per axle - the speed advantage over any AWD system in an ICE is in much faster and more precise control of total torque delivered to open differential that is in turn connected to both left and right wheels.

The motor can not control torque to each wheel, however. The torque control to an individual wheel is achieved by applying brake to a wheel, reducing torque on that wheel and forcing open differential to augment normal 50%-50% torque split between left and right wheel, transferring torque to the wheel on the opposite side of the car.

Interesting! So the Torque Vectoring in the Tesla AWD System is achieved not only by increasing the torque on the wished wheels but also by decreasing the torque on the wished wheels using the brakes. Main thing is that the differential torques between left and right are those wished. Then the speed advantage over any AWD system in an ICE is given by the fact that there are 2 motors and that these motors are also electrical (faster in their answer with respect to an ICE).

I would call it Passive Torque Vectoring in order to differentiate this system with respect to the Torque Vectoring with 4 electric engines.

 

[vgrinshpun]

Interesting! So the Torque Vectoring in the Tesla AWD System is achieved not by increasing the torque on the wished wheels but by decreasing the torque on the wished wheels. Main thing is that the differential torques between left and right are those wished. Then the speed advantage over any AWD system in an ICE is given by the fact that there are 2 motors and these motors are also electrical (faster in their answer with respect to an ICE).

I would call it Passive Torque Vectoring in order to differentiate this system with respect to the Torque Vectoring with 4 electric engines.

There is still an active element in Tesla AWD, since two motors have completely independent and fast control of torque to each pair of wheels, or, more precisely, to each open differential (front and rear). If I would rate robustness of the AWD systems best to worst: 4 motor electric drive, 2 motor electric drive, ICE based AWD (distant third).

For a fascinating look at 4 motor Torque Vectoring system see this 18-minute video, if you did not have a chance to see it earlier (this is my favorite, so I linked it quite a few times before :biggrin: Mercedes SLS Electric Drive. Can Volts Ever Match Pistons? - /CHRIS HARRIS ON CARS - YouTube

 

[Raffy.Roma]

There is still an active element in Tesla AWD, since two motors have completely independent and fast control of torque to each pair of wheels, or, more precisely to each open differential (front and rear). If I would rate robustness of the AWD systems best to worst: 4 motor electric drive, 2 motor electric drive, ICE based AWD (distant third).

Agree. In fact I added "only" when I wrote "not by increasing". I did it 1 minute before of your reply (actually while you were writing your answer).
I also agree with your rating. I would add that if it were possible to control the differential torques on each axis precisely by using the brakes the 4 motor electric drive AWD System and the 2 motor electric drive AWD System could be rated as equal.

 

[catbauer24]

Here's the correct answer:
A locked wheel (brake: disc, differential, or regenerative) transfers torque to the other side. Negative torque can be applied only with a reverse motor on one side. It is impossible to accomplish this in a diff, individual motors are needed.

 

[sorka]

And how would that be better than the usual ABS + stability control, which also does wheel-specific braking?

1) That only kicks when the inside wheel starts to lose traction and it's not as precise.
2) It wastes lots of energy due to friction braking.

You need an electronic diff that can direct torque left or right depending on needs. But even that is not as efficient as an open diff which is what the Tesla has. A limited slip diff is better for handling than an open diff but the viscous coupling results in friction and lose of energy when going around corners.

If it was for the TC, an open diff would be dangerous in car as powerful as the PD.

 

[sorka]

Here's the correct answer:
A locked wheel (brake: disc, differential, or regenerative) transfers torque to the other side. Negative torque can be applied only with a reverse motor on one side. It is impossible to accomplish this in a diff, individual motors are needed.

What the heck? That doesn't make any sense.

 

[sorka]

I would call it Passive Torque Vectoring in order to differentiate this system with respect to the Torque Vectoring with 4 electric engines.

...or an active diff.

 

[DoctorJJ]

What the heck? That doesn't make any sense.

Sure it does. If the motor puts out 1000 ft lbs of torque and it goes to a differential which then powers 2 wheels, 500ft lbs of counter force applied to each wheel would stop the motor. If it is an open diff or limited slip diff, then applying enough torque to stop one wheel would then transfer all the torque to the other wheel. You ever seen "My Cousin Vinny"???

I don't know how they do it on the Porsche Cayenne but it works. Our new Cayenne Turbo has PDCC (Porsche Dynamic Chassis Control) and Porsche Torque vectoring and that big heavy SUV handles well enough to scare you. Maybe it's a magic Porsche unicorn in the center diff, but I really don't care. It works. If they can do it that well on an ICE car, then the near instantaneous feedback and control of torque output on a Tesla should be even better.

 

[sorka]

Sure it does. If the motor puts out 1000 ft lbs of torque and it goes to a differential which then powers 2 wheels, 500ft lbs of counter force applied to each wheel would stop the motor. If it is an open diff or limited slip diff, then applying enough torque to stop one wheel would then transfer all the torque to the other wheel. You ever seen "My Cousin Vinny"???

Only true in a mechanical open diff. Not true an an electronic diff that does torque vectoring like the one in the Honda SH-AWD and SH-AWD Hybrids.

 

[Stiction]

A problem with 4 motors: If you lose a motor under high acceleration you need a system to detect and IMMEDIATELY drop the opposing torque...else you are off-roading in a very short amount of time .

 

[Dithermaster]

A problem with 4 motors: If you lose a motor under high acceleration you need a system to detect and IMMEDIATELY drop the opposing torque...else you are off-roading in a very short amount of time .

Perhaps stability control would notice that and drop opposing torque. Having the tire go slower than expected (motor loss) is a bit different than having it go faster than expected (spin), but stability control just looks at yaw and decides if it needs to correct.

 

[voyager]

Torque vectoring using only one electric motor:
Charged EVs | GKN’s new torque-vectoring axle combines an electric motor and dual-clutch system