front/rear differential?

front/rear differential?

Unlike a traditional 4WD/AWD vehicle, which have a drive shaft and differential between the front and rear, the Model X has front & rear motors, presumably with no mechanical linkage between the two. How will the two motors be synchronized, and how will the differential function be accomplished? Via a SW algorithm?

ian | 7 août 2014

Yup. Software.

Iowa92x | 7 août 2014

No differential front to back, but a mechanical differential for side to side wheel speed differences on turns and such, correct?

Boukman | 7 août 2014

From all the reviews I have seen and read about the MS, the traction control system is excellent. Even in icy and snowy conditions. So I think that Tesla has developed the expertise to tackle synchronizing the the two power train quite easily. As it has been already pointed out, it will be all in the software allowing updating as they get even better at this...

Remnant | 10 août 2014

@ kmmcdonald (OP, August 7, 2014)

<< How will the two motors be synchronized, and how will the differential function be accomplished? Via a SW algorithm? >>

The electronic management of differential functions is known as "torque vectoring" (TV).

There are several specific implementations of TV, such as Porsche, Audi, Mercedes-Benz, Acura, etc.

I have no info regarding Tesla's TV. It may be a distinct implementation or a mere iteration of one of those listed above, but TV is a must in order to achieve a precise AWD.

balabanshik | 11 août 2014

I would argue that with electric motors the construction is simpler, and more efficient at the same time.

Consider a normal differential: it makes it so that two wheels can rotate with different speeds (essential in cornering), but it also, as an unwanted side effect, puts more torque to the wheel with the least resistance, so that if one wheel spins (through having little grip), the other gets no torque, and the car stays in place. Now, in a simple 4WD car, you have three differentials, and therefore even if one wheel spins, the car doesn't move. There are several ways around that problem, these include limiting the differential slip (by making the diff locking or otherwise complicating it), and traction control (basically, reducing the amount of torque coming from the engine), as well as the aforementioned "torque vectoring", which is usually achieved by gently applying brakes on the slipping wheel. Complicating the diff makes it bigger, heavier, more expensive, and more prone to break, and traction control systems have to overcome the inertia of the ICE (and there's a lot of that).

Now consider the electric drive system of a Model S (2WD). The motor pretty much delivers torque proportional to the amount of current you send its way, and that can be controlled with very short response times, instantly for most practical purposes. That's why Tesla's traction control system is so efficient: it doesn't have to wait for the ICE to spin up/slow down, and the wheel slip can be eliminated as soon as it is detected (which is pretty fast these days).

The amazing thing about this arrangement is that if we simply duplicate the system for the front wheels, it will work way better than any 4WD ICE solution, since the front wheels can do their own traction control, independently of the back wheels. In an ICE car, you have only one engine, so you have to dance around that.

Now ideally, the best possible solution would be to put an electric motor for every one of the four wheels. That would create a truly slip-free car: each wheel would apply just the right amount of torque to the road, and I have no doubt we shall see that relatively soon. In the meanwhile, we have to do with just two independent axles (that would still give the car way better traction control than any ICE-based system, at pretty much no extra cost).

bikeandsail | 11 août 2014

I own a Lexus RX400h hybrid 4wd. It has the front wheels driven by the ICE and electric motors and the rear driven only by a separate electric motor. IE only wires go to the rear no mechanical link.It works seamlessly and makes way more sense than a mechanical link.

Iowa92x | 11 août 2014

Does the Model S have a differential? Wish someone would answer yes or no.


ian | 11 août 2014
Iowa92x | 12 août 2014

Thank you.

Remnant | 22 octobre 2014

@ balabanshik (August 11, 2014)

<< ... ideally, the best possible solution would be to put an electric motor for every one of the four wheels. That would create a truly slip-free car: ... >>

I concur.

Here are my speculations about the MX Power train:

As apparent in the D options for the MS, Tesla has diversified the power and size of its electric motors.

It could conceivably consider redistributing the MX power through four semi-axle motors, like Mercedes does in its SLS, which could result in sizable weight reduction benefits by eliminating the two differentials and distributing the torque strictly electronically, via some torque vectoring algorithm.

Arguably, such an innovation could delay production significantly.

David Trushin | 22 octobre 2014

Don't believe there are differentials in the model S. the motors are mounted transverse and have rotation axis parallel to the axles.

Brian H | 22 octobre 2014

No way; the MX motors will be the same as the D's. On the KISS principle.

ian | 22 octobre 2014

David - There are differentials. Open differentials.

Brian - Yes, for now. But possibly in the future they could change to a motor for each wheel.

timf2001 | 22 octobre 2014

Tesla has patented having separate motors for each of the front wheels while keeping a single motor in the rear, so it's possible we could see such an arrangement on the Model X. I wouldn't expect four individual motors at this time.

Remnant | 22 octobre 2014

@ Ian T. (October 22, 2014)

<< There are differentials. Open differentials. >>

Doesn't this mean that the differential functions are actually fulfilled by the Traction-Control/Stability-Control algorithms of the MS?

What happens when the driver turns the TC/SC off? Isn't that like having no differential at all?

Won't this trash the drive tires in a hurry? I think I've seen such a tire issue on one of our threads

ian | 24 octobre 2014

I think there is some confusion here as to what we're calling a differential.

There are open, limited slip and locking differentials.

Open diffs are the simplest and allow the wheels on the same axle to rotate at different speeds while cornering (hence the name "differentials"). The problem here is that when one wheel has no traction all the torque goes to that wheel and you don't go anywhere.

Limited slip diffs allow some of the torque to be transferred to the wheel with traction through various mechanical or electronic means. These can get pretty complicated mechanically, require regular maintenance and can get pretty heavy.

Locking diffs are just that, diffs that you can lock. These are popular on serious 4X4's because they are the best at transferring torque across an axle especially when one wheel has no traction such as when it's in the air. Of course then you have to unlock them when you want to to drive around a corner with both wheels on the ground.

There are also open differentials that are made to operate like limited slip differentials through the use of traction control systems like what Tesla uses (and many other brands) where the wheel with no traction is braked automatically so the torque coming from the motor is sent to the wheel with traction.

I hope that gets us on the same page regarding the use of the word "differential". ;-)


ian | 24 octobre 2014

Sorry. To answer your questions. No, the traction control systems perform the job of limiting slip.

When the driver turns the TSC off ( I don't think Tesla allows for this completely though), the diffs act like regular open differentials.

Not necessarily. Only if you are turning the TSC off so you can do burnouts! Haha! ;-)

It's the tires slipping and sliding on the pavement that wears them out. Even in the cases where the alignment may be slightly out (which has to do with camber and toe).


ian | 24 octobre 2014

Sorry for the repetition of balabanshik's fine description of how differentials work. Seems there is still some confusion though and I was nearly responding to Remnant's questions.

I swear I'd read the whole thread! Just forgot what was posted previously. ;-)


Remnant | 24 octobre 2014

@ Ian T. (October 24, 2014)

<< ... the traction control systems perform the job of limiting slip. >>

I got it. The open differential is a regular differential and helps cornering because it mechanically distributes the torque between the wheels on that axle. If you add traction control, or electronically controlled braking, like Tesla, you have a limited slip differential, that is, a partial locker.

And that's what puzzles me, the fact that the super-hitech Tesla relies on gears and friction to achieve something that can be done more reliably and with less strain on the power train components by electrical means.

Indeed, I've learned that it is possible to control the torque itself electronically, separately for each of the four wheels, which would be a faster, more efficient, lighter, and potentially more economical procedure, by using one smaller motor on each semi-axle.

Here is how Mercedes does this in its AMG-SLS electric coupe: | 24 octobre 2014

@Remnant - The starting point for Tesla AWD was a car that already had a working rear axle motor in place. Do they tear the whole thing up and start developing a quad system, or just add a front unit similar to the rear? In those terms I think it's a no-brainer why the D system is what it is. There is a lot more to Tesla than AWD capability.

Red Sage ca us | 24 octobre 2014

I'm pretty sure that the Tesla Model X will not use a drive shaft connected to a pair of universal joints to share power between front and rear live axles supported by leaf springs attached to a box frame.

Remnant | 25 octobre 2014

@ sg021 (October 24, 2014)

<< The starting point for Tesla AWD was a car that already had a working rear axle motor in place. >>

Working, yes. Properly, no!

In many cases, the MS rear (motor) axle trashes tires with a frenzy that calls for serious review. I, for one, although I have completed a P85D configuration today, I will neither order that, nor complete the MX reservation I have, unless Tesla addresses the tire trashing issue.

I'm not an engineer, but I made an honest effort to understand the engineering issues involved.

Yes, the ideal fix, IMHO, would be to abandon the mechanical, open differential now in place, place one smaller motor on each semi-axle, and design a torque vectoring program that controls the torque distribution between the motors.

I'm not saying this is the only fix. Tesla might prefer to patch up the rogue traction control program that has been trashing tires, if it can tell what went wrong in it.

However, you need to be aware of the size of this issue.

Here are some links:

And here are a dozen or so of quotes from the latter thread above:

mferrazano | 16 March 2013
I just got back from the Tesla SC in Fremont. The tires were so bad that one actually deflated in a parking lot and I had the car towed to Fremont. BOTH rear tires were so severely damaged that both were junked. The guys at the SC were great, Andy, Kyle, Eric spent a good part of their Sat nite replacing the tires and aligning the car. BTW the car was totally out of alignment on all four wheels to a degree not seen before. When I saw the tires off the car, the wear pattern was amazing. 80% of the rear tire tread was in pristine condition, the remaining 20% or less was worn down to the core. The tires were replaced under warranty. The now handles great. I repeat check your tires carefully, the damage on my tires was hidden since it was on the inside edge and could only be seen by getting on your hands and knees and looking under the car at the tires.

qcarl | 18 March 2013
P85 with 21" wheels. Picked up on Dec.30 and now with 6,000 miles. After seeing this thread I checked, and the inside of both rear tires completely bald!

John38 | 2 June 2013
We have a Model S. I did not notice that our 21" tires were wearing on the far inside. At 9596 miles I got a warning indicator of low air pressure so I stopped to check the tires. I found the wear was so bad it had wore all the way through the tire and the air escaped. Got the car towed home. Called Tesla Service center and they said this is normal. Paying $0.15/mile for just tires seems crazy! I am thinking about replacing the 21" performace wheels with the standard 19" wheels or some other wheel . . . any suggestions?

rodneynelson | 21 June 2013
I'm at 6,000 miles. I had a pull off the road because my tire went flat. Soon as it was on the flatbed truck, you could see that both rear tires tires inside were worn to shining metal. One tire just blown up. … I do not believe I was ever informed that I should be prepared to replace rear tires every 3,000-6,000 miles. I would have never have bought a car with kind of excessive expense. … Consider some drivers are getting 10,000 miles and other only 2,000 on a pair of tires, is just not right. Considering that one side of the tire has 80% thread and the other metal speaks of a manufactures problem. At this point Tesla does not appear willing [to admit] that they have a problem.

tranhv68 | 21 June 2013
I have a P85. The rear tires need to be replaced with only 6000 miles on the odometer. Like everyone else it is due to the negative camber set at the factory. The inside of the tire is worn down to the cords and the outside of the tire is like new. I had the same problem with my Nissan GTR. It was originally shipped from the factory with a track setting and I had them readjust the camber for what they called a street setting which just means less negative camber. Tesla needs to give customers the option for a neutral or less positive camber with the understanding that performance may be reduced. Also, although hard acceleration would wear out the tires faster, a neutral camber would allow for that wear to be evenly distributed over the entire tread.

rodneynelson | 21 June 2013
I have found that there are a number us owners who believe that the premature wear is do to a defective design and construction of the car. We were certainly never communicated to by Tesla reps that we should consider changing or rear tires between 3,000 and 6,000 miles. Several customers have had to change the rear tires because of inside improper wear at 1,600 miles. I am not alone in my feelings that this is very disappointing for us. What is also disappointing is Tesla belief that this problem of improper inside wear is the result of our aggressive driving. I drive totally normal.

doctorbyday | 24 June 2013
I had severe wear at 9300 miles. Only reason I noticed it was having the car at eye level on a flat bed after getting a flat after hitting a pothole on I-26 near Orangeburg in SC. I have a P85 with 21 inch Centennial tires. As I type this I am getting the alignment done. The toe-in is the only possible adjustment and the tires are currently toed out per the mechanic. I have to stay in the care to make sure it stays in neutral as they needed to spin the wheels to scribe the tires in order to get accurate measurements. My nearest Tesla service center also said it is best to sit in the car during alignment to simulate actual weight distribution, so i will stay in it while they check the front alignment.

ben | 20 Mai 2014
Like everyone else, my S is eating tires, especially in the rear. I've found a distributor where if I commit to buying at least 24 tires (6 cars' worth) I can get us each 4 original equipment 19" 245/45R19 Goodyear Eagle RS-A2 (rated UTQG: 440 A A) for a total of $650 ALL IN (tires, taxes and shipping to an installer of your choice).

rick | 26 June 2014
My P85+ rear tires were replaced at 9500 miles with exactly the wear as stated here. Within a week after that, Tesla replaced the camber bolts on the rear wheels and the tech indicated this mod will enable the wheels to sit more squarely on the road and spread the load more evenly over the whole tire. With any luck this will allow the rear tires to last a bit longer. Because the tires on the front are a different size than the rear I cannot rotate my tires, side to side "rotation" would provide no benefit.

Doubleup | 9 July 2014
I have had my car at Tesla Service and they old me camber bolts weren't needed as the car was within spec... However here I am after intentionally driving a little gentler still needing tires after 10500 miles. Major inside wear. … My tire guy doesn’t understand why it is wearing so drastically in the inner "one inch". We can’t even see a performance benefit to that extreme of a camber... My fronts are still holding up, I am thinking it will be 3:1 rear/front replacement.

cdjsp11 | 9 July 2014
I have a p85+, replaced the original michelin pilot sport tires with 11000 miles. Tesla Service Center did a fix under warranty. I believe the control arms and bushings. Alignment was also checked. Replaced the rear tires with perelli Nero 265 21s. I've put 12000 miles on the new tires now, and they still look good. Hoping to get at least another 5000 + miles on them. From here on out its an improvement from the original tires. The perelli's were also a lot cheaper than the michelins per tire. The original tires are still on the front with 23000 miles and still have some good life left.

TeslaDude69 | 11 July 2014
I had the same issue on all 4 of my 21" tires. At 8200 miles I got a flat, when the wheel came off, was shocked at the wear on the inside corner on all 4. Worn down to fiber, looked like it was rubbing on the inside of fender.

AmpedRealtor | 11 July 2014
Correct me if I'm wrong, but isn't the reason for adding camber to the rear wheels of performance vehicles to improve cornering performance? If so, removing camber shouldn't have any impact on straight-line performance or stability. When driving over grooved pavement there should be no difference with or without rear camber.
I've had the camber bolts installed to reduce rear camber, yet I have not noticed any difference in the car's stability on grooved pavement or in other circumstances. I don't drive it on the track and I don't corner like a race car driver. Service also confirms that reducing rear camber will have no impact on stability unless you are flinging the car fast into corners, in which case you will sacrifice some stability by reducing camber.

ian | 25 octobre 2014

I just addressed your insinuation that it's the use of an open diff and TC/SC by wire in the other thread where I asked where you're getting these ideas. Stop reading the tire wear threads and read more about how these systems work. They do not cause the premature tire wear seen in the S.

ian | 25 octobre 2014

None of those quotes has anything to do with the traction control system "trashing" tires.

ian | 25 octobre 2014

In fact, it's the alignment that is the culprit as admitted in many of those quotes.

Red Sage ca us | 25 octobre 2014

Remnant: Tires that were not properly rotated; a loose suspension alignment bolt, or a heavy lead foot were found to be the cause of tire issues. It had nothing to do with either traction control or regeneration. This was all resolved months ago. Why bring it up now, as if it is an active, or recurring issue?

DTsea | 25 octobre 2014

My car with 30k miles has the original tires in good shape. I think remnant doesn't know what he is talking about.

jjs | 25 octobre 2014

+1 Red Sage
+1 DTsea
+1 Ian T
@Remnant - Camber and alignment do, and have caused premature tire wear. Get them fixed and the issue is solved.

There is no evidence of the lack of torque vectoring or method of regen contributing to premature tire wear. As you have stated, you are not an engineer. There ARE many engineers that frequent this forum. If after 2 years if torque vectoring/regen were contributors to this issue we would have heard about it.

However, if you still believe you are correct, I think it prudent that you do not purchase a Tesla.

Remnant | 26 octobre 2014

@ jjs (October 25, 2014)

<< I think it prudent that you do not purchase a Tesla. >>

Good advice. Thank you.

Remnant | 30 octobre 2014

@ jjs | October 25, 2014

<< There is no evidence of the lack of torque vectoring or method of regen contributing to premature tire wear. >>

It would be useful, IMHO, to Distinguish Electric from Brake-based Torque Vectoring (ETV vs. BTV)

BTV means getting a digital program to "ride the brakes" in order to assist an open differential functionally to become a limited slip differential. The torque is thus redistributed, according to certain algorithms, to prevent/limit slippage and/or assist cornering.

ETV, OTOH, does control the driving forces (torque) directly, in the electric motor itself, in addition to the braking forces, and thus reduces the use of friction, increases efficiency, and reduces the wear of the drive train.

Tesla will have to develop a quad motor design, if it wants to advance its TV from BTV to ETV.

Since Mercedes SLS uses an ETV implementation already, it might be able to sell its digital control unit to Tesla.

But here is what Nelson Ireson, Senior Editor at High Gear Media, had to say about Brake-based Torque Vectoring:

A TV system is an alternative to a limited-slip differential, leveraging existing wheel speed sensors and electronic brake control borrowed from the stability control system. How does the TV system mimic the presence of a limited-slip diff? By controlling the amount of wheelspin across the open differential through brake application, preventing the "one wheel peel" open diffs are known for.

All well and good so far, right? Not really. In our experience, TV systems' interferences with driver inputs are not only not imperceptible, they're downright annoying. In practice, TV systems aren't predictable in their engagement, often over-limit wheel spin, and generally serve to slow the car and make it harder to drive at the limit--especially when compared to the system they're replacing: limited-slip differentials. Even BMW doesn't get this right, so we don't expect anything special here.

Limited-slip diffs have their downsides, too, particularly in front-wheel drive cars. The first-generation Mazdaspeed3 is an excellent illustration: torque steer galore. But unlike computer-driven electronic nannies, the action of a limited-slip diff is mechanical, and therefore predictable. Don't hammer the gas until the wheel is straight, and expect some tug if you do--particularly over bumpier asphalt. Plan for it. Use it.

You can't do that with torque vectoring. You can't plan for that dab of brake on the inside wheel, which, depending on the surface, can actually reduce front-axle traction even further, particularly if you're already at the lateral limit, leading to understeer where an open diff wouldn't. Further, it taxes the brake system harder, applying the already-overworked front brakes during acceleration as well as braking, leading to reduced times before brake fade begins.

Red Sage ca us | 30 octobre 2014

Remnant stated, "Tesla will have to develop a quad motor design, if it wants to advance its TV from BTV to ETV."

Methinks they don't want to. It is likely an application of the KISS Principle (Keep It Simple Stupid). That's why Tesla Motors now has fully electric brakes instead of hydraulic. Elon Musk pointed out how the fully digital system was more effective in all situations. That would include its application for traction control and torque vectoring. The Mercedes-Benz SLS AMG Electric Drive is a truly awesome system. But that car also costs four times as much as the Tesla Model S P85D. If Tesla was a luxury car company, if they wanted to offer a $500,000 car, they would likely explore having a separate motor at each corner of the car. But since their goal is to go the opposite direction, to eventually offer all this technology at affordable prices to the masses, it is better to keep the system as cost effective as possible, and that means to use as few parts as possible, making it as robust as possible, and finally, as simple as possible.

Remnant | 31 octobre 2014

@ Red Sage (October 30, 2014)

<< The Mercedes-Benz SLS AMG Electric Drive is a truly awesome system. But that car also costs four times as much as the Tesla Model S P85D. >>

That's prototype price, Sage. Prototype makers often sell licenses to (or even full-fledged modules of) the inventions included in their prototypes for prices way lower than their own costs or prototype pricing. Besides, quite a few MS features are already Mercedes-sourced.

My point though has been that Tesla is likely to want to minimize the friction losses in its power linkage. Replacing the friction involved in BVT with the direct electromagnetic control of torque offered by EVT looks very attractive from that angle, IMHO.

vandacca | 31 octobre 2014

@Remnant, I guess it depends on what those "friction losses in it's power linkage" equate to in added range. If it's going to cost another $50,000 (just a guess) to put in 4 motors and all you get is another 5km of range (sorry, I live in Canada: 3 miles), then it may not be the best use of their resources.

I admit that the 4-motor solution sounds like the perfect solution and that Mercedes SLS AMG looks like an awesome car (to drive). Maybe we'll have to wait for Tesla to re-introduce the Roadster and come up with a new chasis to support 4 motors.

Red Sage ca us | 31 octobre 2014

Remnant: All anyone has to do is watch Chris Harris drive the Mercedes-Benz SLS AMG Electric Drive to understand just how awesome it is to behold its torque vectoring abilities. If Tesla Motors was interested in making cars as limited edition, small volume, 'toys for the rich' they certainly could manage the feat. Friction 'losses' are the least concern with electric vehicles. What matters is the application of power to the pavement, and in that regard the Tesla Model S P85D surpasses the 'prototype' for performance in every regard except top speed.

vandacca | 31 octobre 2014

@Red Sage, I'm sure Tesla can apply their magic (aka physics) to get both great torque vectoring abilities and power to the pavement without sacrificing range! But not right now. Once Tesla is firing on all cylinders (pardon the ICE pun), they'll have more time and resources to explore new improvements.

David Trushin | 31 octobre 2014

I kept getting headaches from this thread. My doctor said, take two aspiring and stop reading remnant's posts. Feel much better now.

Brian H | 1 novembre 2014

What does an aspiring taste like? Is it more like the red pill, or the blue?

vandacca | 1 novembre 2014

I think it's more like the red pill, but with an artificial sweetener.

David Trushin | 1 novembre 2014

My bad. Doctor really said 2 tylenol. But I was aspiring to greater humor.

vandacca | 1 novembre 2014

FYI @David Trushin, @Brian H is head of the grammar police, very little gets pasted/past him.

David Trushin | 1 novembre 2014

Gimme a brake! Grammer police?

Remnant | 1 novembre 2014

@ vandacca (October 31, 2014)

<< Maybe we'll have to wait for Tesla to re-introduce the Roadster and come up with a new chasis to support 4 motors. >>

I agree.

The Forums are just speculation round tables or even soap boxes of self-promotion, not attempts to browbeat Tesla.

I trust the Forum participants understand this and appreciate the opportunity to broadcast their opinions on these threads.

Red Sage ca us | 1 novembre 2014

Graham Kops... never crumble, bumble, or stumble.

vandacca | 2 novembre 2014

Well played @David Trushin.
Well said @Remnant.

BTW, I noticed that some members now how state/country appended to their names. Does that only apply to US residents? Are there some US residents that don't have it?

Brian H | 2 novembre 2014

Anyone can have it; change your username to suit yourself.

Remnant | 7 novembre 2014

@ vandacca (October 31, 2014)

<< If it's going to cost another $50,000 (just a guess) to put in 4 motors ... >>

In the Design Studio, the addition of the 188-hp front motor is only billed $5,000. So, the addition of yet two more, in order to obtain a Quad Motor design would probably get billed some $10,000. I have no idea how much an Electric Torque Vectoring module might be, but let's guess $1,200. It does not look prohibitive, does it?

It is more likely the Quad-EVT promotion will be a production & marketing issue rather than a cost & pricing issue.

Red Sage ca us | 7 novembre 2014

As electric motors multiply, their individual power typically divides to preserve efficiency.

Remnant | 8 novembre 2014

Now heralded by the MS-D series, the only way for MX to amaze, which Tesla has promised, in the Drive Train department is to go Quad-Motor + EVT, rather than present just another Dual Motor iteration.

The exact power per motor is not critical as long as the total satisfies to the goals for which the vehicle is built, such as towing so many specified lbs. My guess is that the MX will be able to send some 1,000 hp total to the 4 wheels.

vandacca | 8 novembre 2014

@Remnant, I would love to see quad-motors, but it isn't going to happen for the MX. At least not for the next 3-4 years.

Remnant | 8 novembre 2014

@ vandacca (November 8, 2014)

<< @Remnant, I would love to see quad-motors, but it isn't going to happen for the MX. At least not for the next 3-4 years. >>

You might be right, but I was not trying to forecast the advent of MXQ.

I was only trying to convey "the only way for MX to amaze [ME]", in the Drive Train department, since a Dual Motor would not.