Death to Diffs! Separate motor for each wheel
 

Given my PTSD from spending most of the last 15 years driving an open diff, and now driving a Colorado with one... I'm really intrigued by the idea of doing away with diffs forever and having one motor per rear wheel.

The Tesla model S "Plaid" has two rear motors. So does the big Cybertruck. I assume this means there's no diff. I believe the Rivian also has one motor per wheel.

It does put you at a disadvantage compared to locking diffs which can send all the power to a single wheel. Now you're at 1/N (N = number of motors) per wheel.

But it means there's no need for an lsd and the tradeoffs there. I suspect for street driving it'll be just awesome.

I'm still intrigued by this.

Pros of separate motors for each wheel
- No LSD required to put power to wheels with grip
- No diff needed at all!
- Traction control simpler: No need to brake the spinning wheel to transfer torque to the other. Just cut power to spinning wheels
- Torque vectoring algorithms can now be developed in software and rolled out as they improve. They could develop user-selectable drive modes.

Cons of one motor per wheel
- Max torque per wheel = 1/NUM_MOTORS. I'm guessing this is not an issue unless you're doing extreme rock crawling. Massive torque to a single wheel is what breaks things anyway
- If one motor dies things might get weird. But I assume the software could disable the other motor on the same axle and the failure mode is no different than one motor per axle.
- Even the slightest difference in torque will be felt in the vehicle under acceleration / deceleration. How difficult is it to ensure identical torque to each wheel? Sending equal power won't cut it if there's any efficiency variation in the motors. Is some kind of adaptive loop needed to compensate for this over time?


Assuming the last con is a non-issue (or easily solveable), overall this seems like a huge win. I'm not sure why more people aren't excited about this. This is a big step forward in drivetrain function and capability.

Also - when I first heard about the tri-motor Model S Plaid, I thought they might hook them both up to the existing rear axle, but geared differently. One geared lower for low end torque, and the other higher so it pulls harder at high speed. This way you can avoid needing a 2-speed gearbox like the Taycan.

However, I'm thinking this is not worth it because
- It's not worth losing out on the other advantages (torque vectoring, diffs suck ass, etc)-
- With two motors you have so damn much torque you can gear them both higher (no point in spinning wheels) so the usable torque may be shifted high enough that a gearbox is truly not needed.

You’re overthinking this. You’d never want more torque to the wheel - the electric motors already exceed the traction limit. Differentials never provide equal torque, so why would micro-management of torque be necessary?

Also, the big issue with a dead motor is not disabling the opposite motor (I strongly suspect that you would not want to do that). It’s the drag caused by the dead one.

That seems more complicated than a gearbox.

Huh? The car has a gearbox. It's a 9:1 ratio. The had to pick a number to balance low vs high end torque.

After putting the LSD on my E46 if the tires are even slightly different in diameter the car is miserable to drive on a highway. There's obviously some range where you wouldn't feel it, and maybe the motor tolerances are well within this so it's a non-issue.

The neat thing about an LSD thought is that it's putting more torque to the larger diameter tire, so it will wear faster and over time they will equalize. It's a self-stabilizing system. Will the opposite happen with one-motor per wheel (since the smaller wheel will put more torque to the ground)? If so this would create an unstable system - ie if you tires aren't perfectly equal they will continue to get worse over time.

Hmmm. If the dead one died because of an electrical malfunction and the car switches it off, I'd assume it's the same as being manually switched off, no? :dunno: But if there's some kind of physical malfunction, then yeah you're fcuked.

Having two motors input into a fixed (single gear) transmission is complicated? It would mean they always have to be powered together and turn together (although at different speeds). Maybe this is infeasible for some reason?

Shifting gears while in motion sucks. It's either a) slow and jerky, or b) requires heroics to prevent it from being slow and jerky. No thanks, do not want.

I wonder if you were at a cruising speed, what could go wrong if you tried to just "power" the rear motors, and let the front wheels regen to the batteries? :dunno:

The reason why different size tires are a problem for an LSD is because the wheels are linked together. This is obviated by decoupling the wheels.

And yes, the mechanism for switching between the two motors would either resemble a gearbox, OR you’d have two motors attached to the axle at all times.

Furthermore, I don’t understand how two motors would be less complex or expensive than a gearbox?

Ah, I was operating under the assumption that you already decided to have two motors, for some reason. For example, because you already have them and can just stick them in. Rather than design a whole new motor, which might be an awkward shape to fit in. Which is, I assume, why they are doing this.

Exactly. That doesn't seem complicated at all. There's already a transmission and gears in there already. Just modify it to be 2 sources in, 1 out.

The problem is that the torque to the ground is uneven. Torque vectoring works independently of why the uneven torque is occurring. In the LSD case it's because they are linked together, but that's not the only way it can happen. If the torque is uneven enough, for any reason, the car will want to not drive straight.

Torque vectoring is one of the reasons people are excited about Plaid - so it can obviously occur without an LSD.