LinuxCNC putting out double steps instead of single steps

More
25 Oct 2024 11:12 - 25 Oct 2024 11:13 #313193 by PeteTech
Hello all, 

First time poster so I hope I give all the necessary information. I do work with Linux systems but only have a basic understanding.

We have a relatively simple 2 axis system running with LinuxCNC and our own front end that communicates with LinuxCNC in the background. We have been making these systems for years but have recently discovered a new issue. 

It has happened twice now, (with a week or 2 in between) that once the machine has been started it runs with a clear and steady double speed / double distance. We have not yet been able to scope the machine yet, but it would seem we are getting two steps for each one that is expected. On second occurence, shutting down linuxcnc ( not shutting the OS though) resolved the issue.  The installation is a direct clone from another machine that has not displayed this issue.  
  • The system is setup on a standard PC using a MSI A520M Pro motherboard with a AMD Ryzen Socket AM4 processor
  • "uname -a" command gives us:  5.10.0-0.deb10.22-rt-amd64 #1 SMP PREEMPT_RT Debian 5.10.17 8-3~deb10u1 (2023-04-25) x86_64 GNU/Linux
  • Debian version is Buster, release 10
  • We connect via parallel port to a DB25-1205 board.
  • We run the motor(5.6A Wantai Servo motor) with a Wantai DQ860HA Servo Drive
  • "top" command shows that the "rtapi_a+" command is using about 22% of the CPU when LinuxCnc is running, and drops to about 3% when LinuxCNC is closed. 
  • No unsteadiness in the motion. Accuracy still consistent but twice the distance instructed. Motor obviously stalls once the max drive/motor speed is exceeded.

    I am wondering along the lines of whether the real time kernel could for some reason throw a servo thread at twice expected rate?  
    I have not been able to recreate it again to test further. Are there any suggestions as to what might be a cause or which direction to  look in when we do recreate the condition?
Last edit: 25 Oct 2024 11:13 by PeteTech.

Please Log in or Create an account to join the conversation.

More
25 Oct 2024 11:53 #313198 by tommylight
Had the same issue a while back, turned out the DIP switches on the drive were bad and caused changing microstepping during machine motion from vibrations. Soldered them or cut them as needed, no issues since.

Please Log in or Create an account to join the conversation.

More
29 Oct 2024 13:04 #313427 by PeteTech
Thank you for letting us know.

We do have hardware on our suspect list but not likely as restarting LinuxCNC without removing power to the hardware fixes the issue.

Please Log in or Create an account to join the conversation.

Time to create page: 0.069 seconds
Powered by Kunena Forum