Limited Slip Differential for Belt Drive ESkates

[JRP3]

I came up with one possible solution to low traction wheel slip since I haven’t been able to find a dual shaft motor. I made a “limited slip differential” for the board so I could have the 2 front motors turning together when taking off in low traction situations. Seems to be working pretty well off road so far, I even get a bit of slip when carving on pavement, I put some strips of tape on the coupler and the motors to see if they ever rotate independently and they do to some degree. It’s a 1.5 inch flexible pipe coupler that was able to stretch around the slightly more than 2 inch diameter of the motors with some work. The only down side is I have to remove a motor to get it on and off, though I could slit it and use zip ties or hose clamps, which would allow some adjustment of pressure and slippage. I did find after riding about 5 miles on pavement that the rubber sleeve started to migrate to one side so I need to put zip ties or something on the motors to prevent slippage. Just had a thought that I could adjust the amount of slip by biasing the rubber sleeve more to one motor which would provide greater grip on that motor while reducing contact and grip on the other to allow more slip. Or slide it completely off of one motor for road use.

 

[amberwolf]

If you're using FOC controllers (like VESC) that modulate motor current instead of motor speed, you should get a form of this electronically, assuming identical control signal to the controllers and identical controllers/motors.

 

[JRP3]

It's an Exway Atlas board running Hobbywing components, not sure what their controllers are, but wouldn't it be the opposite, i.e. speed control would have all motors turning at the same time and speed regardless of torque, while torque control would cause the wheel with the least traction to spin first? In any case the motors are definitely not perfectly matched, I had one motor that lagged so far behind the others I got Exway to send me a new one.

 

[amberwolf]

I guess it depends on the situation, and the controller options and setup.

Speed control just feeds a voltage to the motor; it's all "open-loop" with no feedback to the controller / no controller monitoring of what the actual speed is to cause it to feed more current to make it go faster...it just feeds a specific average voltage (PWM percentage) based on the control input.

If current caused by that amoutn of voltage is sufficient, the motor with least traction would spin out and reach it's speed first, and so less current would flow just because of that (as BEMF increases and reduces the ability for current to flow). But since it's feeding a specific amount of voltage, it's just "trying" to cause a certain speed from the motor.

In a current-control system, the control input "directly" modulates the phase currents, so if a motor has less traction than the current provided can push against and breaks loose, then the speed jumps and phase current in that motor drops, so the controller would try to keep feeding more current till the motor reached maximum speed.

But an FOC controller could have the option that if speed is greater than some amount but current is lower than expected that it would reduce the current until the speed drops below some amount (I don't know if any of them actually do have this). Or if the two controllers could cross-communicate so that if one's RPM was more than a certain amount faster than the other, the faster one would cut current until it was just *below* the RPM of the other and then maintain current to keep that RPM, etc.

I'm not sure if the Lebowski or VESC do this directly, but both are opensource so code could be changed so that they do.

Other FOC controllers designed to be used in multiple-motor-systems might be capable of it (Sevcon probably is, but they're apparently insanely complicated to setup).


Speed controllers could do this too, but I don't know if any do.




FOC could also monitor the current and reduce it even more if it suddenly dropped, since that should only happen if the motor spun up really fast (and it could also monitor the RPM, to know when that happens for sure), and actually have single-controller traction control.

 

[JRP3]

Apparently VESC does have some traction control option, don't know how it's implemented, some people say it works others say not so much.

 

[JRP3]

Here's a video of the speed difference before they sent me the new motor. The delay isn't as bad in lower "gears" or if you throttle up quickly, you can see it get progressively worse as I click up through the "gears".

 

[amberwolf]

From the description, best guess is it was wound differently from teh ohters (different kV).

 

[JRP3]

Tried another version using 2 rubber pipe caps with springs underneath them to keep them pressed together. Probably could have gone with stiffer springs but this does work. The downside is the rubber on rubber interface doesn't release smoothly, it grabs until it doesn't and then slips. I think the concept is good I just need something other than rubber mating surfaces.

 

[amberwolf]

I'm not sure if this would help, but you might look things up based on this:

plastic friction coefficients - Google Search

I don't know what the release terminology would be, though.

I suspect that the more flexible the surface is the more likely it is to release suddenly rather than gradually.