Quad motor vs locking diff

[R.I.P.]

I think you may have missed this since I didn't see anything regarding the difference between Kyle's initial attempt and then off-handedly mentioning that he is is going to switch to what he determined to be the most off-roady settings.
: https://www.rivianforums.com/forum/threads/quad-motor-vs-locking-diff.12010/post-271334

I think that some issues are related to additional development needed and some issues are related to people not using the tech as intended. When Land Rover came out with Terrain Response (drive modes) a lot of hard-core people thought turning a little plastic knob connected to an electrical control couldn't possibly do anything like throwing a beefy metal lever connected to linkages... so they didn't use the right drive mode and thus "proved" the little plastic knob didn't do anything. Because they didn't use it.

No, what Kyle demonstrated is reproducible, and like he says in the next video, has improved with updates. The truck is still prone to "flailing", but it for sure is more controlled than when I got it.

Sponsored

 

[ryanpei]

Anyone have videos of Rivian doing a roller test? Or the controlled slope ice tests? I'm surprised we have not seen any yet. TFL and many others have some on other AWD systems.

 

[SoCal Rob]

No, what Kyle demonstrated is reproducible, and like he says in the next video, has improved with updates. The truck is still prone to "flailing", but it for sure is more controlled than when I got it.

I get that it's reproducible. My skepticism is that it is reproducible in the same way that people using a chainsaw in exactly the same way as an axe will get reproducible results. I'd like to see someone who knows how the system works and wants to see it succeed do their absolute best with the available settings.

I could make a traditional 4x4 look utterly incompetent if I drove it the same way I drive our LR3... and I wanted to prove a point by not utilizing the strengths of the hardware effectively. Imagine this is a click-bait video voiceover and you'll get the idea, "This solid axle system with differentials hanging down is TERRIBLE! I placed the centerline of the vehicle directly over the rock as I'm used to doing in off-road height with an independent suspension and now I'm stranded because my differential broke when I bashed it on the rock! (Click to subscribe)"

 

[R.I.P.]

R.I.P: "A traditional four-wheel drive with fully locked diffs turns all four wheels at exactly the same speed. One wheel can never, and this is the point that is so very important NEVER slip. Any slip can mean falling off a rock and impaling your truck. Any slip starts to create holes that you have to dig yourself out of. When it is critical that you don't slip any wheel, four wheels locked into identical rotation speeds can never be approximated by a quad motor or software."

I would suggest that software and a motor at each wheel should be able to emulate any type of differential. Not suggesting it currently does, just that it should have that capability.

Yeah... I used to be with you on that one. Thing is, on my trip up to the intake today, there is a place in the road (generous to call it that) where it banks sharply to the right before turning left and going up another grade. Driving the old Ford, I unlocked the diff before I got there because I don't want to end up at the bottom of the ravine. I modulated my speed through the off-camber portion, knowing that if it did happen to spin a wheel it would be on the uphill side while my limp (open diff) right wheel will hold me on the road. Physics, it always works that way, I made it across the sketchy part then hit the rear locker to help me up the steep grade.

How to make a computer know how to emulate this? Use a tilt sensor? What if the uphill wheel catches on a root and all of a sudden it is no longer the wheel with the least traction? The differential just reacts, what kind of incredible sensor array would it require to emulate this, and what kind of processing speed would you need to make these adjustments fast enough to keep me from spinning off the road and into the abyss?

To clarify, I think it is possible to create such a system, but my R1T is already looking like a Rube Goldberg machine; working way too hard to try to do something so simple. If there were overriding advantages to my quad motors, I could see myself being a bit more enthusiastic about a fix, but in the end I feel like I am carrying around the weight of two extra motors to do something almost as well as a dual-motor setup could do using two beautifully simple differentials.

Make sense? I really am not trying to upset anybody here, I am just trying to lay out the reasons that;

1. The truck is not working the way I had hoped, and in fact is worse on the snow/ice than my old F250 and
2. Trying to make it work seems like spending effort going in the wrong direction.

Rivian is making a dual-motor version of the truck that should be lighter, more efficient, and solve the problem. Vouila! Hooray for tech! Sure, it might not be new tech, but I cannot wait to get back to the good ol tech of a differential.

 

[Zybane]

Of course it can. This should be obvious with a little common sense. I could build a quad motor EV in my garage that puts the same exact power to all four wheels at the same time. To suggest otherwise is just silly.

There's nothing magical about mechanically locking wheels on an axle with a locking diff or front and rear axles with a locked transfer case. Its entirely possible to replicate that same behavior with a motor controlling each wheel.

Prove me wrong.

No you cannot. Exact same power to each wheel would make each wheel spin the same speed if the truck was on a lift. As soon as one tire touches the ground, the power to continue rotating (and hence moving the vehicle) DRASTICALLY changes, instantly. Now for example you have three free spinning wheels each drawing 10 Amps of power, and one spinning wheel drawing 200 amps of power. And the power required to propel the vehicle, to each wheel, while the vehicle is moving, is CONSTANTLY changing.

I think it’ll come down to the resolution of the wheel speed sensors and how quickly the wheel speed can be controlled. I don’t know if perfection is possible, but my hope is that Rivian could do this closely enough that it makes no real-world difference.

It can never be instantaneous. In a "electronic" version of locking diffs, there will ALWAYS be mis-matched power being applied to all wheels. Hence you will see some wheels slip/brake traction and speed up, while other wheels remain stationary and lug/stall as higher current is applied.

@Zybane, I’m looking forward to getting our R1S since I’m used to difficult trails in a vehicle with independent suspension, different drive modes, and adjustable suspension heights. I know how important it is to set things up correctly and use just enough throttle to allow the vehicle to do what it needs to do. I’m used to the old-school, “you need solid axles, 3 lockers, and coil springs,” crowd telling me that I can’t possibly get to places I’ve already gotten to.

Most people don't know their ass from a hole in the wall.

Keep in mind that torque is what causes rotation, and higher torque is needed to move heavier loads. If all 4 wheels are free in the air, they will all spin at the same speed immediately. If all 4 have 100% traction, the vehicle will not move at all until enough torque is applied to move 7000 lbs. Hopefully we all agree on this.

In our slippery slope case, the speed control algorithm very slowly applies equal power and therefore torque, to all four motors with the goal of speed control by ramping up all four wheels from 0 mph to 1 mph max in 3 seconds. But .... the rear left is not touching the ground, the front left is very slippery mud, both on the right side have 100% traction in dirt.

Time=0 sec
No power applied, no motion on any wheels

T=0+ sec
Apply initial low torque=X on all wheels. If you apply way too much they will all spin, right? At low applied torque, the free wheel must start turning immediately because it has zero traction resistance to the ground. But the other 3 wheels have enough traction resistance on the ground that torque=X is insufficient to rotate those wheels. Right out of the gate, we've already failed at true locking diff emulation. How much power do we apply to the wheel that is now spinning? Do we stop it because the others are stopped, or do we let it spin? We can't speed control it to the others, they are not turning and there is no vehicle speed.....

T = 1 sec
Torque has ramped on the other 3 wheels to torque=5X. The front left with low traction now also starts to spin - it needs less torque because there is less resistance. The right side still has good traction but torque=5X is still not sufficient to cause rotation because it is not enough power to move the 7000 lb truck. Do we let the front left spin (again breaking our locking diff goal), or reduce power until it stops? OK, just let it spin and continue the full ramp up to 1 mph.

T= 2 sec
Torque is ramped up to torque=10X and only now the right side has enough torque to starting moving 7000 lbs of dead weight. Now that we know the speed of the right side wheels is starting to ramp from 0 to 1 mph, we can slave speed control the left wheels with low/no traction to match the right side.

T= 3 sec
From 2 to 3 seconds we attempt to speed match the left to the right side.

This is the challenge of software speed matching between differing dynamic loads. As the load changes, appled power must change to maintain the same speed in variable speed applictions.

Exactly. Why three mechanically locked diffs will always be superior. Whether those diffs are connected to any prime mover.

I'm confused, can 100% of torque not be applied at ~0rpm in an electric motor? In that case, you apply maximum torque but control the maximum rotational velocity of each tire to an arbitrarily small number and in this case you've basically simulated a locked diff. If 100% of torque isn't enough to move the vehicle then nothing else really matters

100% torque (motor electrical current) cannot be applied to all four wheels at the same time off-road. That is what happens when you floor the vehicle to get a 3 second 0-60 time. Current has to be applied slowly/staggered to an electric motor to balance traction requirements. The same thing happens in diesel-electric locomotives. The only way you could apply maximum motor torque and control wheel velocity is the brakes. That is some serious forces involved.

The bottom line is that an un-locked driveline, like the Rivian has will ALWAYS be reactionary. Having three locked differentials is a physical means to overcome constantly changing traction variables that no computer controlled independent motors can sense, react to, and provide the real time ascending or descending stepped current response.

This is precisely the reason why Jeep is designing an electric Wrangler replacement that has one electric motor connected to a transfer case then out to two lock-able differentials.

 

[Inkedsphynx]

No you cannot. Exact same power to each wheel would make each wheel spin the same speed if the truck was on a lift. As soon as one tire touches the ground, the power to continue rotating (and hence moving the vehicle) DRASTICALLY changes, instantly. Now for example you have three free spinning wheels each drawing 10 Amps of power, and one spinning wheel drawing 200 amps of power. And the power required to propel the vehicle, to each wheel, while the vehicle is moving, is CONSTANTLY changing.



Most people don't know their ass from a hole in the wall.



Exactly. Why three mechanically locked diffs will always be superior. Whether those diffs are connected to any prime mover.



100% torque (motor electrical current) cannot be applied to all four wheels at the same time off-road. That is what happens when you floor the vehicle to get a 3 second 0-60 time. Current has to be applied slowly/staggered to an electric motor to balance traction requirements. The same thing happens in diesel-electric locomotives. The only way you could apply maximum motor torque and control wheel velocity is the brakes. That is some serious forces involved.

The bottom line is that an un-locked driveline, like the Rivian has will ALWAYS be reactionary. Having three locked differentials is a physical means to overcome constantly changing traction variables that no computer controlled independent motors can sense, react to, and provide the real time ascending or descending stepped current response.

This is precisely the reason why Jeep is designing an electric Wrangler replacement that has one electric motor connected to a transfer case then out to two locked differentials.

Isn't that what the recommended launch does? Sport mode, apply full brakes, mash accelerator, burn? What's the issue with doing that in a special mode offroad?

I guess this is where my lack of mechanical/electric engineering knowledge comes into play. I'd have thought that there's more controlling the speed of the motor than simply current input.

 

[R.I.P.]

I get that it's reproducible. My skepticism is that it is reproducible in the same way that people using a chainsaw in exactly the same way as an axe will get reproducible results. I'd like to see someone who knows how the system works and wants to see it succeed do their absolute best with the available settings.

I could make a traditional 4x4 look utterly incompetent if I drove it the same way I drive our LR3... and I wanted to prove a point by not utilizing the strengths of the hardware effectively. Imagine this is a click-bait video voiceover and you'll get the idea, "This solid axle system with differentials hanging down is TERRIBLE! I placed the centerline of the vehicle directly over the rock as I'm used to doing in off-road height with an independent suspension and now I'm stranded because my differential broke when I bashed it on the rock! (Click to subscribe)"

I am truly sorry if I come off incompetent, but there is really no way to effectively address that on an internet forum. All I can tell you is that i am an avid off-roader, have built many vehicles, many of which were electric (as per my bio) and I have tried to explain things to the best of my ability so that the broadest audience possible can understand what I have observed.

I purchased the R1T to replace an old F250 that I often use to get to the top of the mountain. The Ford is not lifted, and whereas the tires are quite aggressive, they are stock sized. Based on all of the evidence I could scrounge, I was convinced the R1T would quite handily replace it. The Rivian has failed, spectacularly, because the truck is pure evil on the terrain I need it to work on.

I am not sad I bought the truck, and I will get good use out of it until I find one that can replace the old gasser more effectively. I hope that truck will be a Rivian. There is a ton I really like about it. The quad motors is just not one of them.

P.S.
Along those lines, I am no longer considering the Cybertruk Tri-motor either. With what I have learned, I expect it to have similar trouble on slippery surfaces.

 

[quartz]

Just like auto tranny made it easy to drive, Rivian made it easy to drive off-road. Sure, in some situations lockers and diffs are better suited, but for most off-road excursions, the R1 is vastly superior from UX and ease-of-use perspective. If you are hardcore enthusiast looking for an engaging drive, Rivian is not for you.

Also, consider that Quad R1 may perform best with different driving techniques than what you are used to. Try taking different lines, left-foot breaking, etc.

 

[zefram47]

Anyone have videos of Rivian doing a roller test? Or the controlled slope ice tests? I'm surprised we have not seen any yet. TFL and many others have some on other AWD systems.

I know one of the guys at TFL. I'm kinda surprised they didn't do it...wonder if the rollers couldn't handle the weight. I'll reach out and see if they respond.

 

[Zybane]

Isn't that what the recommended launch does? Sport mode, apply full brakes, mash accelerator, burn? What's the issue with doing that in a special mode offroad?

I guess this is where my lack of mechanical/electric engineering knowledge comes into play. I'd have thought that there's more controlling the speed of the motor than simply current input.

It's extremely stressful on components, wastes a ton of energy, generates a massive amount of heat, is dangerous, and brakes typically don't have the fidelity at near release friction coefficient which is different on every wheels brake disc.

 

[R.I.P.]

I know one of the guys at TFL. I'm kinda surprised they didn't do it...wonder if the rollers couldn't handle the weight. I'll reach out and see if they respond.

It is actually easy to predict the result of a roller test, the vehicle puts torque to all 4 wheels as a rule and will pull itself off the rollers with ease. Which I feel compelled to add, is a huge part of the problem I have been trying to convey. Torque to all wheels on a very slippery surface will send the vehicle sideways, unless it is pointed directly downhill. An open diff will only spin two wheels, leaving the other two as lateral support for the vehicle. If the vehicle can move on the slippery surface, it is more likely to be in the direction the spinning wheel(s) are turning.

Likewise, a differential powered vehicle with traction control will spin first one side, then the other trying to gain a grip; always leaving one side for lateral support. The very last thing you want is to spin all your wheels at once which is what my R1T is always trying to do.

Is there a different way I should be trying to explain this? I feel like I am not using the right words to convey why this is a problem. Is there somebody in this thread that can explain better than I am about the importance of rolling friction, the mechanism that the differential uses to maintain that, and the reason that not having it causes the sideways motion of the vehicle that can be so innerving?

I would really welcome some help here trying to get this point across lol.

 

[Zybane]

I forgot to mention one thing, the Hummer EV that has a rear "E-locker" with two motors on the rear axle also suffers from this issue. It's not just Rivian, but will be all EV's that don't have a mechanically locked differential (the Hummer EV does have a FRONT mechanically locked diff).

In the below video, with the Hummer EV's "E-diff" set to on, at 9:50 and then at 11:40 in the video, you can clearly see the rear wheels spinning at different speeds:



That is BAD for off-road. Independent electric motors will NEVER be as good as locked axles off-road.

 

[Zoidz]

I'm confused, can 100% of torque not be applied at ~0rpm in an electric motor? In that case, you apply maximum torque but control the maximum rotational velocity of each tire to an arbitrarily small number and in this case you've basically simulated a locked diff. If 100% of torque isn't enough to move the vehicle then nothing else really matters

You are correct that a DC motor develops maximum torque at 0 rpm or stall speed. As mentioned by others, that would result in essentially an 835 hp launch being unleashed all at once from a standstill in the mud. Not advised, lol. You would need a way to keep the rotational speed down while still developing maximum torque.

The problem is that speed control of the motor and the power applied to the motor are tightly related. If you apply the power needed to develop maximum torque, it's going to inherently make the motor spin fast. This is not a stepper type of motor that can turn a degree or two and stop. So if you want maximum torque and still have low rotational speed, you need some other way to limit the speed. This could be done by applying the brakes, but that's wasting a ton of power as heat and would put alot of stress on the suspension since the excess braked torque will be transferred throught the caliper to the suspension. Ouch.

In simple terms, speed in EVs is controlled by pulsing DC power to the motor. The amplitude and the duration of the pulses determine how fast the motor spins and how much torque is developed.

 

[R.I.P.]

I forgot to mention one thing, the Hummer EV that has a rear "E-locker" with two motors on the rear axle also suffers from this issue. It's not just Rivian, but will be all EV's that don't have a mechanically locked differential (the Hummer EV does have a FRONT mechanically locked diff).

In the below video, with the Hummer EV's "E-diff" set to on, at 9:50 in the video, you can clearly see the rear wheels spinning at different speeds:



That is BAD for off-road.

CORRECT. No, it is not just a Rivian issue.

 

[Dark-Fx]

I forgot to mention one thing, the Hummer EV that has a rear "E-locker" with two motors on the rear axle also suffers from this issue. It's not just Rivian, but will be all EV's that don't have a mechanically locked differential (the Hummer EV does have a FRONT mechanically locked diff).

In the below video, with the Hummer EV's "E-diff" set to on, at 9:50 and then at 11:40 in the video, you can clearly see the rear wheels spinning at different speeds:



That is BAD for off-road. Independent electric motors will NEVER be as good as locked axles off-road.

GM's regular on-road traction control in the Hummer is horrible compared to what Rivian does. I don't trust them to have their shit together for their virtual locker either.

Sponsored