Grin Magic Drive

[neptronix]

I have to say that hillslayer got me interested in running a geared + DD combo; it can be complimentary.

But then i moved to a place where i have 3 flights of stairs to climb, and i'm back to power/efficiency density maxxing

 

[bananu7]

the chain tensioner means no regen

This has been my thinking as well. I think the tensioner is needed for a typical suspension setup, and obviously that won't work backwards. Which is sort of a pity, since that would also mean no freewheel in the rear hub, making it even simpler.

One way to accomplish this could be to use surron-style central hub and two chains/belts of constant length, but those mechanisms still have considerable backlash. Yet another could be to get rid of the large chainring and use a moto-style small driven sprocket at high RPM, located close to the swingarm pivot, to keep the single-chain backlash manageable. After all, motorcycles do engine braking with no tensioners.

 

[hillslayer@]

The way I understand this motor is it matches the torque input into it. If you added a mid -drive you should have an absolute hill climbing beast. This is probably how those dual motor mid -drives work also . It's kinda weird it took this long to figure this out. It would be great if this ends up being super efficient, so you could power the mid -drive and hub on one not so heavy battery.

 

[TPEHAK]

So it looks like we need to start figuring out a VESC controller for dual motor mid-drive soon. That is going to be quite a challenge. We probably need to reverse engineer that VESC Labs Duet motor controller for this purpose. It looks like the VESC guys use a bigger size MCU STM32F405VGT6 with 100 pins and two amplifiers to control two motors simultaneously. After reverse engineering we need to figure out the algorithm for the firmware about how those two motors should work together.

 

[OlderGiantFan]

I've been spinning Shimano's Biopace rings for decades now (and will continue as long as my stash holds out), For me, they're the 'cat's meow'. FWIW, Sheldon agrees:

Just curious, what rate do you spin at? I spin and I find the faster I go, up to a practical limit, really smooths out the stoke. I'm just down and back too, not the debunked up-stroke. I also have a theory, For the longest time cyclist mouthpieces have equated the human body to a machine, the legs to pistons, that sort of thing. But where biological creations and muscle and bone is a lot different. It seems natural to me that my all my leg muscles should be able to fully relax during half the pedal stroke.

 

[TPEHAK]

Here is an explanation of how two motors and the legs input should work together. One motor produces torque for the rear wheel, another motor works in regenerative brake mode to maintain necessary resistance on the crank arms depending on the support mode selected by user. The sensors are the same, a torque sensor and a cadence sensor.

 

[hillslayer@]

Here is an explanation of how two motors and the legs input should work together. One motor produces torque for the rear wheel, another motor works in regenerative brake mode to maintain necessary resistance on the crank arms depending on the support mode selected by user. The sensors are the same, a torque sensor and a cadence sensor.

So we could run a small IC engine to keep battery charged and help with big hills. How small could this engine be. The first hybrid bikes are almost here. This technology seems to have been around for along time why the delay into ebikes is strange. It's almost like there was a need to sell as much as possible of the soon to be outdated tech before creating a boom of everyone having to get newest tech.

 

[stancecoke]

Here is an explanation of how two motors and the legs input should work together.

The Magic Drive doesn't need a second motor for having an "electronic CVT". The second motor are your legs.
The idea to attach the sprocket to the planet gear carrier is really clever, even if Justin states, that this idea is not new at all.

 

[hillslayer@]

The Magic Drive doesn't need a second motor for having an "electronic CVT". The second motor are your legs.
The idea to attach the sprocket to the planet gear carrier is really clever, even if Justin states, that this idea is not new at all.

The new mid drive referenced above does use 2 motors. This technology is being introduced suddenly into ebikes. I believe this magic drive is going to be old tech pretty quick.

 

[stancecoke]

I believe this magic drive is going to be old tech pretty quick.

What should be the advantage of using two motors, instead of one for providing the same features? The second motor will just cause more costs and more losses.

 

[lnanek]

I guess with the e-CVT there's typically an electric motor directly on the wheels/planetary carrier, a gas motor driving the planets, and another electric motor driving the sun gear. That allows always keeping the gas motor at the same, most efficient speed always no matter what or freewheeling. Then the electric motors can be varied to change the gear ratio or used for regenerative braking.

So even if we drop the gas motor and add a human pedaling, that only removes the need for one of the three motors.

The video didn't seem to be impressed with the efficiency of the alternative, two opposing cones CVT that doesn't use multiple motors, nor simply making a normal mechanical transmission electric.

 

[stancecoke]

So even if we drop the gas motor and add a human pedaling, that only removes the need for one of the three motors.

You don't need the additional motor, that drives the wheel directly, what should be the benefit of this motor on an E-Bike? You need one motor, that is driving the sun wheel, this is the concept of the "traditional" geared hub motor. The second "motor" (= your legs instead of the ICE) is driving the planet carrier. The ring gear is driving the rear wheel of the bike.

 

[hillslayer@]

Really have to wait until there is a tear down video of one of these dual motor mid - drives to understand how it works completely. I would also like more info on the magic drive like torque ratings, watts, top speed. To me it looks like the hub motor may be done with a mid drive you don't have to shift.

 

[stancecoke]

Justin is announcing the two motor solution for #2 of the YouTube videos in a comment to #1:

We will discuss the pros and cons of 2 motor solution in the part 2 video which is the obvious way to decouple pedal torque fully from the terrain of desired. However, in it is remarkable how much functionality you can get from just this single motor solution

On another comment, Alon states, that the second motor will be in the front wheel...

Software was indeed very hard work...
MG2 is in the pipeline, in the front wheel of course!

 

[hillslayer@]

Alright so my understanding of how the ECTV works is kind of like a snowmobile clutch which keeps a 2 stroke motor at its ideal rpm no matter what speed you are at. This should prevent the idiots being able to burn up motors . I understand how a snowmobile clutch does this don't know if this is exactly what the magic hub is capable of.

 

[Daniel James]

what if we were able to lock the planet carrier with a brake, AND drive it with a mid, multiplying the speed and + the torks, excuse me for the noobie explanation, but this seems like what I would want out of something like this, with the option to pedal, or would that just be really unnecessarily redundant?

 

[hillslayer@]

How long before a constant velocity unpowered hub is developed that only needs to be pedaled . Derailleur manufacturers are already panicking about the new ECTV motor.

 

[bananu7]

You don't need the additional motor, that drives the wheel directly, what should be the benefit of this motor on an E-Bike? You need one motor, that is driving the sun wheel, this is the concept of the "traditional" geared hub motor. The second "motor" (= your legs instead of the ICE) is driving the planet carrier. The ring gear is driving the rear wheel of the bike.

After watching the Youtube video linked above, I agree with Inanek that it would require two separate motors. Later in a video an intuitive explanation is presented; one motor is able to turn excess torque into more rpm, the other is able to take excess rpm into more torque, by the virtue of where in the planetary drivetrain they're attached. It looks applicable to the pedal power exactly in the same way it's applied to the ICE.

EDIT: In fact, if your legs were driving the planet carrier directly (blue part in the YT video), you'd get no range adjustment at all - since your legs would be in constant mesh with the wheel. This is indeed the case, and acceptable, for one of the motors.

a constant velocity unpowered hub is developed

That would certainly be much less efficient than any existing non-powered drivetrain, and thus have a major disadvantage. Electric drivetrain can offset those inefficiencies easily, making the whole thing much more practical (and much less complex mechanically).

 

[DanGT86]

what if we were able to lock the planet carrier with a brake, AND drive it with a mid, multiplying the speed and + the torks, excuse me for the noobie explanation, but this seems like what I would want out of something like this, with the option to pedal, or would that just be really unnecessarily redundant?

Seems to me that if you locked the planet carrier with a brake you would not be able to pedal it due to it being locked.

With this system it seems that the rider always needs to resist the backwards torque on the planet carrier. I'm trying to picture how a one way clutch could change that. Seems to me if the planet carrier were only ever allowed to spin forwards then the rider would have the option to add power if they wanted but not be required to resist the counter torque from the motor.

I could be overlooking something very simple though.

 

[bananu7]

I think one key part to think about it is that there are indeed three inputs to the system, and they're all practically limited in what they can do (using names and colors from the video):

* MG2 (blue drivetrain) - directly meshed to the wheels, always turns when the wheels turn in the same proportion (can have a constant factor)
* MG1 (green drivetrain) - acts as a torque adjuster, in some cases will remain stationary or spin in either way for ratio adjustment
* ICE / pedal input (red drivertrain) - this is the only input that gets the full range adjustability from MG1 - i.e. actually uses the whole system as a "gearbox".

If I understand it correctly, in the scenario where ~all of the power is intended to be generated from the "combustion" input, one of the motors acts as a generator while the other takes that energy to create a "backpressure" for the pedals to work. I'm pretty sure it wouldn't work at all in a completely dead system in case of battery running to absolute 0 (unless it could take enough energy from the pedalling to maintain the system operation) or electrical failure.

If it was possible to mechanically lock MG1, i think it would be possible to pedal at a constant ratio, while turning the MG2.

What is not entirely clear to me is whether it would be possible to build/use this system without MG2; my thinking is that you'd then need to have a battery solely to drive the MG1 to provide desirable reduction, but would be unable to generate any driving torque.

 

[stancecoke]

Why do you always talk about the Toyota drive train?!
In his video, Justin explains exactly, that the Magic Drive doesn't need a second motor.

The only disadvantage is, that you have to be the "torque arm" for the motor torque with your force on the pedal all the time...

 

[hillslayer@]

How long before we see a true test and review of the magic hub and one of these ECTV mid -drives . Seems like a bunch of hype and they haven't really built a true working model they trust not to fail.

 

[DanGT86]

I have a pile of geared hubs laying around. I'm thinking about trying this. Still trying to wrap my head around the controller strategy.

Seems to me the controller would need to run speed based throttle and be set to do regen and locked motor when off the throttle.

If I understand this correctly, pedal torque from any steady state (cruise or stopped) would attempt to push the rotor backwards.

This force at a cruise would cause the rotor to slow and a speed based controller would apply more current to maintain that speed.

Im thinking this is why it doesnt require a torque sensor. The simple act of the controller trying to keep a specific commanded rotor speed causes the rider and motor to spilt the torque evenly.

 

[hillslayer@]

That would certainly be much less efficient than any existing non-powered drivetrain, and thus have a major disadvantage. Electric drivetrain can offset those inefficiencies easily, making the whole thing much more practical (and much less complex mechanically).

It has been developed yet just saying probably not impossible

 

[TPEHAK]

BenjaminV said MCU struggless to handle two motors. Do you think there is a chance VESC can handle E-CVT motors? It would be nice to design an INNORACE E-CVT controller, but we need to make sure VESC is capable for this.