4 ms·
> Well, yes, but you can also sense which ones are turning slower and just increase the torque to those. From the link: > If the axles are not physically tied
by almost_usual 3y ago
> Well, yes, but you can also sense which ones are turning slower and just increase the torque to those.
From the link:
> If the axles are not physically tied, there is no way for any of the four wheels to know how much traction it actually has until it breaks free. Then, of course, it knows "too much" torque and it can apply traction control. But the damage is done. It has broken free, and any negative consequences have already happened prior to the computer being able to make any further calculations.
> Again, if physically locked together with large gears and drive shafts, there is no possible way for one wheel to break free in that manner, unless of course something is broken.
Some other interesting tests by an R1T owner.
https://www.rivianforums.com/forum/threads/quad-motor-vs-locking-diff.12010/post-270978 https://www.rivianforums.com/forum/threads/quad-motor-vs-loc...
- CPLX 3y agoI mean, doesn't the set of computer controlled individual motors completely and totally contain all scenarios presented by a locked differential? Specifically, wouldn't one line of code that says "never ever ever turn this wheel faster than this other wheel" be totally and completely indistinguishable from having the wheels on the same axle, as far as the physical world is concerned?
- shkkmo 3y agoThat's not how electric motors work. You can measure the speed and adjust the power going to the motor if you go over or under that speed. The lower latency that this cycle has, the better, but it can never go to zero like with physically connected wheels.
- itishappy 3y agoPedantry: The speed of sound in steel is about 5100m/s, so a solid 2m axel will still take about 4µs for interactions to propagate from one end to the other.
- CPLX 3y agoMaybe so but what does that have to do with anything? A physical axle can have both wheels break free too can't it? Comparing apples to apples the only scenario you're concerned with is where one wheel breaks traction and the other doesn't. Why is a logically driven exact speed match between wheels inferior to a physically linked speed match between the wheels. Explain the difference practically speaking.
- shkkmo 3y agoYou said "totally and completely indistinguishable". Practically, electric motors can't generally transfer torque from one wheel to another like you can with locked hubs. (The only exception would be where your battery is underrated to provide sufficient power to both motor to maximize their torque)