4 ms·

Worth noting, placement at the wheel is somewhat arbitrary. There's nothing to prevent this contraption from working the other way around and being mounted at t

Worth noting, placement at the wheel is somewhat arbitrary. There's nothing to prevent this contraption from working the other way around and being mounted at the chassis instead, aside from squeezing out every bit of space optimization. Axle support would be a concern either way. Controlling deflection of the drive axle adds a structural requirement, certainly a well supported bearing close to this device. The gears will be turning at a large multiple of wheel speed, so will require roller bearings except for low power and speed use: a radial and thrust bearing for each end of each gear, or 22 of each per wheel. Increasing torque also increases the level of thrust support required for the moving gear carriers handling all of those helical gears. It's a neat idea for highly space constrained applications that have to package a drivetrain very close to a wheel, which does end up being a rather niche use case. Mars rovers, perhaps. Trying to compete against CV joints in passenger vehicles with this would be quite difficult.

benj111 3y ago
It's trying to compete against a CV joint and gear box. That seems like an easier sell. I'm not sure if we're talking past each other but you wouldn't need an axle to deflect. The motor drive could attach directly to the 'uni wheel' which is in turn connected directly to the wheel. If you're talking about the motor drive shaft, then surely you'd just be moving thrust bearings from the gearbox to the motor (if motors don't already have them)
- nabilhat 3y ago
  If the drive source is attached directly to the uniwheel, then handling radial drive inputs isn't required and the uniwheel isn't necessary. It's effectively a hub motor with reduction gears at that point. Once the motor is mounted remotely from the uniwheel and the uniwheel is handling radial offsets, then we need to prevent the axle from getting pushed around radially. Rotating shafts that deflect introduce off-balance vibration and destructive mechanical loads. If the motor is located very near the uniwheel, keeping the shaft's free length very short, the motor's output bearing does effectively provide that support. Then we're only constrained by relocating the suspension elements. Thrust bearings are for the gears. The depictions show helical gears, which would surely be necessary when operating at transportation speeds for noise reasons. Helical gears experience axial forces, which would need to be handled for each gear. Mainly on one side for ICE drivetrains, since coasting loads are small, so one side could get away with a ground and hardened rub surface. EV's using regenerative braking would need robust thrust handling on both sides. Herringbone gears would eliminate the axial thrust issue, at substantially more expense. The easiest method is to separate the gear areas with a central runout groove so you'd have a hope of grinding in a decent surface on the faces, then it's only a little more than double the work; effectively cutting two gear shapes per gear plus the runout. https://en.wikipedia.org/wiki/Herringbone_gear https://en.wikipedia.org/wiki/Herringbone_gear
- governmentdude 3y ago
  > you wouldn't need an axle to deflect This is really cool for rear wheels. I don't know why their demo shows it on all 4, since it has to be able to steer (note that the demo does show a CV joint in the axle on the front, but that seems like it would limit steering angle severely, since it is just 1 joint. They're probably planning to apply torque to steer).
aidenn0 3y ago
> Worth noting, placement at the wheel is somewhat arbitrary You're suggesting placing the ring-gear in the chassis and running the axle to the wheel? That would involve two reduction gearsets in the chassis instead of the current "one plus a CV gear," which kind of defeats the purpose of reducing chassis space usage. > The gears will be turning at a large multiple of wheel speed, so will require roller bearings except for low power and speed use: a radial and thrust bearing for each end of each gear, or 22 of each per wheel. Increasing torque also increases the level of thrust support required for the moving gear carriers handling all of those helical gears. I imagine the final product would use herringbone gears, since the large number of thrust bearings would all add to unsprung weight.

benj111 3y ago

It's trying to compete against a CV joint and gear box. That seems like an easier sell. I'm not sure if we're talking past each other but you wouldn't need an axle to deflect. The motor drive could attach directly to the 'uni wheel' which is in turn connected directly to the wheel. If you're talking about the motor drive shaft, then surely you'd just be moving thrust bearings from the gearbox to the motor (if motors don't already have them)

nabilhat 3y ago
If the drive source is attached directly to the uniwheel, then handling radial drive inputs isn't required and the uniwheel isn't necessary. It's effectively a hub motor with reduction gears at that point. Once the motor is mounted remotely from the uniwheel and the uniwheel is handling radial offsets, then we need to prevent the axle from getting pushed around radially. Rotating shafts that deflect introduce off-balance vibration and destructive mechanical loads. If the motor is located very near the uniwheel, keeping the shaft's free length very short, the motor's output bearing does effectively provide that support. Then we're only constrained by relocating the suspension elements. Thrust bearings are for the gears. The depictions show helical gears, which would surely be necessary when operating at transportation speeds for noise reasons. Helical gears experience axial forces, which would need to be handled for each gear. Mainly on one side for ICE drivetrains, since coasting loads are small, so one side could get away with a ground and hardened rub surface. EV's using regenerative braking would need robust thrust handling on both sides. Herringbone gears would eliminate the axial thrust issue, at substantially more expense. The easiest method is to separate the gear areas with a central runout groove so you'd have a hope of grinding in a decent surface on the faces, then it's only a little more than double the work; effectively cutting two gear shapes per gear plus the runout. https://en.wikipedia.org/wiki/Herringbone_gear https://en.wikipedia.org/wiki/Herringbone_gear
governmentdude 3y ago
> you wouldn't need an axle to deflect This is really cool for rear wheels. I don't know why their demo shows it on all 4, since it has to be able to steer (note that the demo does show a CV joint in the axle on the front, but that seems like it would limit steering angle severely, since it is just 1 joint. They're probably planning to apply torque to steer).

aidenn0 3y ago

> Worth noting, placement at the wheel is somewhat arbitrary You're suggesting placing the ring-gear in the chassis and running the axle to the wheel? That would involve two reduction gearsets in the chassis instead of the current "one plus a CV gear," which kind of defeats the purpose of reducing chassis space usage. > The gears will be turning at a large multiple of wheel speed, so will require roller bearings except for low power and speed use: a radial and thrust bearing for each end of each gear, or 22 of each per wheel. Increasing torque also increases the level of thrust support required for the moving gear carriers handling all of those helical gears. I imagine the final product would use herringbone gears, since the large number of thrust bearings would all add to unsprung weight.