# Using brushless motors for very high torque

I'm currently building an robotic arm and I want the elbow joint to be comparable to a human one But here's the problem, electric linear actuators are too slow and have low torque, same goes with regular hobby servos and stepper motors

Ideally I would want a brushless motor with no gearbox that could hold a 10kg weight at an arm's distance (let's say 40 cm) and move freely at around 100rpm (about 10rad/s)

Holding force: 40 N.m

I was thinking sensored brushless motors could do the trick but surely that load would be too much for those...

If this cannot be achieved, would it be possible to use a 3D printed 100:1 planetary gearbox and a low kv motor to do it?

Thanks in advance, if you have other types of motors that would be more suitable please comment them, any help is appreciated.

• Hydraulics. Seriously. And no, plastic gears aren't worth anything, and 3D printed plastic gears less so. Commented Nov 8, 2017 at 23:18
• @Janka Mmmh it requires pumps, tanks, it will inevitably have leaks and I'm not sure about efficiency, it's a good option but I have no skill in hydraulics... Since I'm planning to do a full body it sure would save me the size of motors on each joint Do you think something like this would be efficient? youtube.com/watch?v=1BxkvuMxvQU Commented Nov 9, 2017 at 21:07
• Oh yes, air muscles are totally going to work, and being both lightweight and simple. But you have to use a closed-loop regulation to account for heating/cooling of the gas during compression/expansion. That's something not easy to precalculate. Commented Nov 9, 2017 at 21:33

Your expectations are rather aggressive for a DC motor.

First - 40Nm (350in-lbs) is A LOT of torque! Ex: A max rated torque for 1/4-20 bolt is only 75 in-lb (8.5Nm).

Second - The mechanical power of a motor results from RPM*Torque. 40Nm*100rpm is 0.56hp (420W at 100% efficiency). That is A LOT, about 1/2 of what today's high end cordless drills can do. Fortunately it looks like you aren't expecting high torque at the high RPM so actual power will be lower; but, those numbers are likely to affect your motor size (unless you are going to use a transmission). You will probably need to start thinking about a torque curve.

Regarding DC motors: They rely on higher RPM to deliver higher power rather than higher Torque. This has to do with the physics of the motor. High Torque with a DC motor is going to require some sort of a "mechanical advantage".

Here are some design options to consider.

1. (Metal) Gears/Gearbox. Planetary gearbox could do. Plastic gears (at least for the load end) are likely going to fail at high torque. They simply can't handle the sheer. You can get some that are pretty small. Here is gear motor from Motion Dynamics that can to 40Nm. Here is one from Midwest Motion that can also do 40Nm (Note: Both use Planetary gearboxes and are pretty slow and no where close to the power you are looking for).
2. A slight variation of the gearbox (but work mentioning separately) is a worm gear/drive. This can get you a lot of mechanical advantage; but, it will turn the axis or rotation by 90deg.
3. Use a cord-and-spindle and pulleys if needed to gain additional mechanical advantage (this does great for pulling; but, you are going to need a spring for return). Think garage door opener (sort of). Also think human arm (muscles only pull). Other linear-motion leverage options are leadscrews and rack-and-pinion gears.
4. Hydraulics (as mentioned). This is how car brakes, clutches, and power steering do it. It allows you to move the motor off the arm. This is pretty common for large walking robots like Boston Dynamics Atlas.

Note: AC motors are more efficient than DC motors if you don't need positional control. That is likely why Tesla uses them in their cars.

• Thanks a lot for all of that info, I've taken a look at Brushless sensored motors of "Hoverboards" and they appear to be able to give about 10nm if I put a better driver behind. However even those aren't good enough, plus they're heavy I've seen some pretty light gimbal motors and cheap (25USD) able to deliver 1N.m at a speed that would be about 4000 rpm so I'll try to use them to drive either a planetary gearbox or a 2 stage pulley and belt drive with a reduction ratio of about 40:1 I'll prototype this in Plastic, Nylon and aluminium if I get to use my school's CNC router Commented Jan 8, 2018 at 16:28
• Glad I could help. I have done quite a bit with motors and learned a bit along the way. I always want to know the why. Note: Be careful when you are looking at the torque spec to do your gear scaling that you are looking at the stall torque. Stall torque is the maximum torque the motor can deliver when the rotor in not moving. This can be a lower number than the Max Torque. Commented Jan 8, 2018 at 21:09
• Also, if you plan on holding the motor in stall for extended periods of time, you are going to have to be careful you don't overheat the motor. Stall requires a lot of electrical power with no resulting mechanical power output. Conservation of Energy says the power has to go somewhere and it usually goes to heat. Mounting the motor with a good thermal contact to a large block of aluminum (good thermal conductor) will help. Mounting it to steel or plastic (poor thermal conductors) will result in a VERY HOT MOTOR. Commented Jan 8, 2018 at 21:15
• BTW, what school are you at? College? Commented Jan 8, 2018 at 21:28
• Oh thanks well, I'll monitor coil and ESC temperatures carefully when doing tests I'm in a French Engineering School near Paris called Epita, but I'm doing this in my quest to get as close as possible to a human being with the simplest ways possible Commented Jan 9, 2018 at 23:19

R3D34THR4Y, You are on right path. In this video skyentific explains why wide bldc motors are better fit for your problem. He uses OpenTorque Actuator, that was designed and developed by Gabrael Levine.