A Hall sensored motor is not very appropriate for an exoskeleton (or any robot) due to the torque ripple you will get. Can you find a motor with a high resolution encoder? Or you could easily mount a MA732 sensor to that motor with a 3D printed adapter, probably, since it looks like it has a sub shaft on the back that you could glue a magnet to.
The max current of 5A is a bit too low for ODrive - it will work, but you may need to change the current shunt resistors (and the firmware) to improve current sensing noise.
Also, are you sure 269 mNm * 4k rpm is going to be useful for an exoskeleton?
With very efficient gearing, I suppose it could be. But be careful to make it backdrivable.
Thanks for the reply. That’s actually the idea I have in mind. I could mount an external hall sensor to the shaft. I plan on 3D printing the actuator housing, so it should not be a problem.
As for the current, torque and speed, maybe I could share some more detail. I actually want to generate a peak torque of ~30Nm from my actuator. So definitely I will be using gear reduction for that. I plan on using a synchronous belt and pulleys. Based on my design calculations, I would require a motor with a peak torque of ~2Nm.
Considering this Maxon motor, it has a stall torque of 4.3 Nm at 82A. Of course, that’s the limit, with no rotation, but it can work within that range.
And for the rotation speed and ‘backdrivability’, can these not be varied with the voltage? Speed control should be possible with the ODrive.
I actually have been looking for a compact motor with a decent torque for quite some time. I’m currently in Germany, if by any chance you have something to help, that would be greatly appreciated! I contacted Maxon, that’s the only one they have with a proper lead time.
At 82A you will burn this motor in a fraction of a second. The motor has 0.3 Ohm resistance, so I^2 R = 2kW of heating power at that current. Also, you will probably saturate the iron core at well below that, so the torque will not extend linearly up to that current.
These have about 0.1 Nm/A up to 30A continuous, i.e. 3 Nm, probably 10Nm peak by your crazy standards
You cannot gain backdrivability at the controller end (not without force/torque sensors) - maybe you misunderstand what I mean by backdrivability.
The gearing mechanism should be efficient in both directions, i.e. worm drives, harmonic drives etc have (almost) zero backdrivability whereas a good low-friction belt system has good backdrivability.
The motor’s cogging will also have an impact on this - even if you use a belt drive, there is a minimum torque required to make a motor start moving due to the detent torque caused by the permanent magnets and iron cores.
Wow, that motor sounds good, although can’t find anywhere where I could buy it here. I will keep looking for similar
I understand what you mean now with back drivable. Well, for belt and pulley system, it should allow that. That’s the reason I did not go with high reduction harmonic drives. Plus, it just adds to the costs.
Hmmm, you gave me an insight into the current and heating. Is there a way to calculate that really? They do provide a torque constant for the motor. If it’s operating at at most 2Nm, shouldn’t it be fine?
I will keep looking for similar 
