• Torque required is about 17Nm.
  • The speed for rotation is very small about 10degrees/sec.

The load is dynamic and I do not want to apply any torque more then 20Nm, this I want to achieve by using a torque feedback. I want to make an cost effective setup for this purpose. Hence should I use servo motor or a stepper motor with a feedback? I read here to use servo motor when i need feedback anyways. Is it true? wouldn't I will save some money by going for stepper with feedback as the speed is slow?

  • \$\begingroup\$ Have you considered DC motor+gearbox? \$\endgroup\$
    – pjc50
    Feb 29, 2016 at 14:12
  • \$\begingroup\$ Yes, But I thought since the rotation required is really small its better to use servo or stepper. My requirement is to rotate the object by say 60 degrees without applying any torque that is above 20Nm. If at any point it reaches 20Nm I will stop rotating it. For this I just need a torque sensor. \$\endgroup\$ Feb 29, 2016 at 14:14

1 Answer 1


The advantage of steppers is their open-loop simplicity. You just command it to take a step, and without any feedback you know how much it's rotated. Of course you need to do the calculations to be sure the stepper is always operating within its limits: otherwise you miss steps and you either need to add feedback (negating the advantage of simplicity) or accept that you may end up rotating by something other than the desired amount.

If you want torque feedback, then you don't have an open-loop system anymore. The advantage of simplicity is lost, so you might as well use a servomotor.

  • 2
    \$\begingroup\$ It is actually worse with a stepper motor, because the holding torque is quite different than the step torque, and the latter can vary from step to step as well if one uses microstepping. Of course it is not impossible, but torque control will require quite sophisticated adjustment of the winding current. \$\endgroup\$ Feb 29, 2016 at 14:28
  • \$\begingroup\$ @OleksandrR. You mean the complexity of controlling torque with a stepper is more complex, right? \$\endgroup\$
    – Phil Frost
    Feb 29, 2016 at 14:40
  • \$\begingroup\$ Thank you for your answer. I understand that if i make sure my load is within the operating limit and start the motor slowly it will not skip the steps and will rotate to the desired position. But I want to have an encoder just to be 100% sure. So wouldn't I be saving some money by going to stepper motor with a encoder compared to a servo. Or the simplest solution will be to use dc motor with gear heads and encoder, that way I can control the maximum torque applied to the load easily? \$\endgroup\$ Feb 29, 2016 at 14:44
  • \$\begingroup\$ @shekhargupta If you have an encoder and you use that information to drive the motor, that's a servo by definition, even if the motor is a stepper. I think you'll find that after you implement that control loop, the stepper doesn't really make anything simpler for you. And I think you'll also find, as Oleksandr points out, that controlling torque with a stepper is really hard. \$\endgroup\$
    – Phil Frost
    Feb 29, 2016 at 15:02
  • 1
    \$\begingroup\$ The main issue with steppers is that they have no torque when the stator and rotor field are aligned. So if the torque on the rotor is negligible compared to the holding torque, the rotor should be relatively accurately where it should be (about 5% of a step). However if the torque is not negligible, the rotor will turn until the holding torque balances the load torque, up to a step angle. Beyond that it misses a step and usually stalls. When microstepping (for smooth continuous operation, more than 10x is usually useless), the holding torque is substancially decreased. What's the accuracy? \$\endgroup\$ Feb 29, 2016 at 15:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.