Robot motor selection high torque, low speed, small size.

[zenman]

I am in a robot puzzle.

I am writing the software for a robotic leg brace.
The current design for the knee joint has a 7 pole pair, 17 turn, DC Brushless outrunner motor on a 70:1 WORM gear and a 2:1 secondary gear so 140:1 reduction.  the motor of choice is the Hacker B30.

We have to rotate the motor on a rotational load carrying 150 pounds and a radius of about 14 inches.  That's quite a bit of torque.

Under high motor speed (800-1000 RPM) I can do this.  But low speed is an issue. I have to get the motor to rotate at full load under 100 rpms for proper operation. 

The guy who designed this said that outrunners used for model planes are the most power dense and thus best motor for this job and that commutation software is to be the most important part of this.  He quotes the spec that DC Brushless motors have infinite torque as low RPM.  Not sure he knows much about this.  He also picked low voltage, high current motor strategy.

Everything about outrunners say they are the best for this, but in reality they get hot, blow power fets and hard to start under load - without stalling.

I tried to convince the guys to add another planetary gear to get speeds up on the motor, but the system designed said that the blacklash would be too great, Under "ordinary" load the motor has to spin 12000 RPM.

How do most people lift heavy loads, and what DC motors (smaller the better) is suggested.

Why aren't outrunners used for all robot motors?

Thanks,

[jwatte]

It's true that 14 inches times 150 pounds is a lot of torque! This is why prosthetics is not a simple field to make real progress in. In general, the human body is 10-100x more efficient than most equivalent mechanically based designs I'm familiar with.

Why aren't outrunners used in robotics? Outrunners have moving parts on the outside that would get snagged on other parts, I guess?

Also, it is almost true that DC motors, in general, have as much torque at stand-still as you want. There is a limit, where pushing more current into the windings (by increasing the voltage) generates so much heat that they melt/vaporize/seize.

All motors will generate heat and blow MOSFETs that are under-sized and or un-protected, btw. I don't think that's somehow special to outrunner DCBLMs.

Finally, the option (for strong electric motors) are AC Induction motors. That's what Tesla chose for their sports cars, for example.

[zenman]

Thanks good points.

The design call for a small light motor.  If you know of small AC inductions motors, it would be helpful.

The WORM gears add inefficiency but prevent back rotation.

In simple 3 second (fast RPM) moves while moving the just apparatus, the motor gets hot to touch.
To prevent accidental touching of the motor to skin, and snagging, the mech guys built a shroud for the motor, which is supposed to deaden the noise, but it also keeps the heat in, and does nothing for most of the noise.

In situations under load, with low speeds, I am afraid to see what happens because the FETS start blowing and these are 30A FETS (depends on how they are rated)
Most situations like this involve hydraulics.


I used a cordless drill in manual mode and it worked great, but the mech guys frowned on that technique due to it's additional gearing.  Of course they are brushed.

[jwatte]

If a brushed motor will work, there's something to be said for that!
Similarly, there are a number of gearing strategies that prevent back driving in addition to worm drives.
Strain wave (harmonic drive) and cycloidal (used in Dynamixel Pro servos I think) are common in motion control.

[bdeuell]

My understanding is that induction motors and brushless motors are very similar in layout with the main difference being that that the rotor of a brushless motor uses permanent magnets. I believe induction motors are generally used in larger applications where the use of permanent magnets are impractical.

If you have a housing around your outrunner motor is there much of a weight savings over a standard brushless motor? Using a standard brushless motor would also have the advantage of using the housing to cool the windings. Given the intended application of outrunner motors I would expect they are relying on forced air cooling to maintain an  acceptable temperature of the windings.

The harmonic drives I have used can be back-driven, see also http://www.harmonicdrive.net/reference/advantages/.

[jwatte]

Yes, I was mistaken about the back-driving of harmonic drives (strain wave gearing.)
I stand by the cycloidal reduction, though :-)

I also think that another option would be a coreless motor, if the main requirement is light weight. It does use brushes, though.

When it comes to highly efficient brushed DC gearmotors, there are some manufacturers that make very nice motors -- Maxon and Faulhaber, come to mind. They are a step above cheap Chinese imports from what I've seen.

[waltr]

They are a step above cheap Chinese imports from what I've seen.

In my experience these are 10 steps above.
I have a few of each (obtained surplus) and was totally amazed with their output torque for the size and how quiet they are. If you really need a High torque, low speed in a small size that is also quiet and runs cool then do look at  Maxon and Faulhaber gearmotors.
Of course, to have all those characteristics you will pay a substantial price.

[zenman]

All good points,

Of course the "marketers" of this thingy promoted low cost as one of it's differentiators thus the $80 outrunner.  Of the shelf motors were spec-ed.

I also think you are right about the brand. Before motor selection the powers at be like the 18v operation which resounded to an RC car guy in the group, and that mutated to a model airplane motor based on current specs and power density and Hacker was the best of the bunch here.

We were going to add a additional maxon planetary gear reduction of 7:1 (we should have picked 2:1) but that would bring the gear reduction with the previous 70:1 and 2:1 to 7 * 70 * 2 = 1000:1  and the backlash numbers were apparently factors.
I thought this would be easy, but the mech guys had a cow.

The harmonics were mentioned to us, medical robots use those, thus they were eliminated because this had to be different. - and the expense.  I think all roads go here.

I think you guys brought up some good points.

The mindset of the mech guys is that the motor selection is not important, it is the commutation sensing and sequencing algorithm that is the most important.  The motor just has to be spec'ed to accept enough power to turn under load and adjust quickly as load changes. "Just physics"  So software is the key.
The realities is back EMF and efficiencies of a dc motor at high torque and low RPM.

One dude mentioned a 5 inch long 2 inch wide custom motor with several rotors tied together.

I think we need to bring in a variety of motors and try. 

I love the cycloidal drive, will look into that.

Which gets me back to inrunner or outrunner?  The Faulhaber's all look like inrunners?
Are the outrunners good for their fanning ability?  At low speeds, this is moot.

Thanks again,

[jwatte]

Which gets me back to inrunner or outrunner?

Personally, I wouldn't care at all one way or the other. Pick the one that fits the parameters the best.