Callibration of New Machine

Home Forums TinyG TinyG Support Callibration of New Machine

Viewing 3 posts - 16 through 18 (of 18 total)
  • Author
    Posts
  • #11113
    Zootalaws
    Member

    I just json it from a stock file.

    #11114
    melticus
    Member

    sorry guys. I’m changing job at the moment so have not had time to get back to my machine today. thanks for your replies. I am indeed using an external stepper motor driver to drive my axis. My 8ntension is to publish my machine design when I get it working.

    What I mean by torque reduction is I observe that while it is stepping, it is easy to stop the x carriage with the slow g1 command. with the fast g0, the movement nearly takes my finger off. Not very scientific but that’s how I did it…

    #11115
    cmcgrath5035
    Moderator

    What I mean by torque reduction is I observe that while it is stepping, it is easy to stop the x carriage with the slow g1 command. with the fast g0, the movement nearly takes my finger off. Not very scientific but that’s how I did it…

    OK, understand your methodology.
    Are you using microstepping? If so, do you need it for precision?
    Typical stepper motors are ‘native’ 200 steps per rotation (some are 400).
    By native, I mean that the pole/winding structure of the motor has 200 combinations of the two poles that will result in discrete movement.
    Microstepping is achieved by having the driver modify the phases of the winding pulses to divide each of the 200 steps into 2,4,8,16,etc pseudo steps. These pseudo steps do not exhibit the same holding torque (force to hold position) as native steps.
    When you run at very low speeds, e.g. f40, an increased percentage of time is actually holding torque, not moving torque.

    This is a very readable reference for the TI 8825 drivers that are native to tinyG http://www.ti.com/lit/ds/symlink/drv8825.pdf.
    The discussion on microstepping can be informative, if interested.

    Looking back at your posted pictures, I have to wonder if the .2 mm offset you were seeing is the result of a long standing tinyG computation precision bug. The best released version of tinyG FW is $fb=440.20. If that is not what you are running, update.
    What your pictures depict is often referred to as ‘drift’, i.e. after running a Gcode job, the physical location of (0,0,0) is not the same as logical (0,0,0).
    tinyG (and therefore cnc.js) report logical position, not physical. If step pulses issued by tinyG do not result in proper movement, due perhaps to inadequate torque or binding mechanics, or if tinyG skips a move due to a computational error (the bug), drift can occur.

    drift is almost always the result of G2 or G3 arc commands.
    Testing first with G1 commands is good to get started, but running a job such as depicted in your pictures, which I’ll assume uses G2 and G3 moves, is also necessary.

Viewing 3 posts - 16 through 18 (of 18 total)
  • You must be logged in to reply to this topic.