Results 1 to 6 of 6

Thread: Can we get accurate torque and current information from the dynamixels now?

  1. Can we get accurate torque and current information from the dynamixels now?

    Hi, I recall that when the MX series came out few years ago, users were excited about the possibility to use the motors for torque control but at the end, they found that the motors could not provide torque information but current information. It was kind of like false advertisement. Even so, the current information could not be used in a useful way because there were lots of noisy signals and other issues. How is the situation now? Is the X-series any better?

  2. #2
    Join Date
    Sep 2010
    Location
    ಠ_ಠ
    Posts
    2,296
    Images
    27
    Rep Power
    278

    Re: Can we get accurate torque and current information from the dynamixels now?

    I have not used Robotis stuff in a few years, but I specifically recall the MX-28 having neither current-limiting motor drivers nor current sensors. It had only a magnetic encoder and an H-bridge made from two SOIC-8 N-FET half-bridge chips, so it could easily kill the motor and/or H-bridge by stalling the motor and the firmware not cutting the PWM quickly enough to limit the stall current. I think the next step up (MX-64) and beyond did have current sensors to improve stall survival and enable actual current control modes, but never tested that hardware to see how well they worked. Torque is proportional to current, so with accurate current control you can get fairly accurate torque control. Although I'm thinking jwatte lost multiple MX-64 by using 4S LiPo that allowed current to hit dangerous levels quicker than the sensor and firmware could respond...

    Torque sensing can be kinda expensive and/or bulky and a bit difficult to achieve on a continuously rotating shaft, so I doubt most of them actually implement it. The easiest way to implement it is having two absolute position rotary encoders along a slightly flexible shaft (or two ends of a flexible shaft coupling): the difference in the two position values will increase with increasing torque on the shaft giving you magnitude and direction of the torque.
    Please pardon the pedantry... and the profanity... and the convoluted speech pattern...
    "You have failed me, Brain!"
    bleh
    more bleh

  3. Re: Can we get accurate torque and current information from the dynamixels now?

    Thanks tician. Could you please explain what you mean by "flexible shaft" or two ends of a flexible shaft coupling? Do you have a photo? What LiPo capacity would be safe to use if I use high-end high torque motors? In general, is it better to stay away from 4S LiPo?

  4. #4
    Join Date
    Sep 2010
    Location
    ಠ_ಠ
    Posts
    2,296
    Images
    27
    Rep Power
    278

    Re: Can we get accurate torque and current information from the dynamixels now?

    "A low-cost compliant 7-DOF robotic manipulator" (pdf here)
    Involves two disks linked with rubber tube stuck in cutouts in the disks to make compliant linkages (flex couplings) that allow fairly accurate measurement of torque by comparing radial offset of the two disks.

    "Optical torque sensors for implementation of local impedance control of the arm of humanoid robot" (pdf here)
    Involves a custom aluminum hub/linkage with cheap optical slot detectors to measure the offset of two ends of a shaft. Not the exact paper I was looking for, but same concept and nice images and explanation.

    Plenty of other work out there around low cost torque sensing.



    As for LiPo batteries: use only 3S for MX series servos. Not sure about XM/XH servos.

    The motors and H-bridges in the MX series are rated for lower voltages than the older RX/EX components because so few people were actually using anything other than 12V or 3S LiPo. A nominally charged 4S LiPo battery is barely within the voltage limits and a fully charged 4S LiPo exceeds the voltage limits. Look through some of jwatte's threads for what not to do. I've use 4S LiFePO4 on occasion since it has lower per cell voltage and the chemistry is much harder to damage, but the higher voltage is still much riskier than a 12V supply or 3S LiPo.
    Please pardon the pedantry... and the profanity... and the convoluted speech pattern...
    "You have failed me, Brain!"
    bleh
    more bleh

  5. Re: Can we get accurate torque and current information from the dynamixels now?

    Thanks tician.

  6. #6

    Re: Can we get accurate torque and current information from the dynamixels now?

    I would assume that most servo motors that provide "torque control" actually measure the relationship between "current" and "torque" and actually implement it as "current control."

Thread Information

Users Browsing this Thread

There are currently 3 users browsing this thread. (0 members and 3 guests)

Similar Threads

  1. Question(s) Dynamixel torque control using current feedback
    By Gray Fox in forum DYNAMIXEL & Robot Actuators
    Replies: 4
    Last Post: 11-22-2017, 02:04 AM
  2. PID Controller for Dynamixels work in Torque
    By gimmegmailemail in forum DYNAMIXEL & Robot Actuators
    Replies: 0
    Last Post: 10-30-2014, 04:24 PM
  3. Development of new automation systems and very accurate encoders
    By RoMoKo in forum Robotics General Discussion
    Replies: 1
    Last Post: 07-25-2013, 02:36 PM
  4. How to decide run current and hold current in stepper motor
    By KevinQ in forum DYNAMIXEL & Robot Actuators
    Replies: 1
    Last Post: 05-01-2012, 05:44 PM
  5. Question(s) Bioloid information
    By bugbaal in forum Humanoids, Walkers & Crawlers
    Replies: 6
    Last Post: 10-30-2011, 07:57 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •