Servo Printer Artifacts: Pulse Timing and M569 T?

  • Hello, I've posted elsewhere about the issues I've been having with my servo controlled gantry pritner.

    The advice I got there was definitely helpful (thanks @dc42 and others!), but I'm still experiencing under or over-shoot possibly caused by missed steps.

    Bear with me, I'm a newbie to many things electronic.

    Right now, the Duet sends step/dir from the x/y test points on the Duet to an external servo drive. Along the way, the pulse signals are converted to quadrature signals because our external drive will not accept the signals otherwise.

    Right now, each axes is configured with a timing value of T0. Timings above T0 (...T1, T2, etc.) severely limit the speed each axis will move.

    T0 allows for movement as fast as we want, however, the artifacting appears to get much worse. (Gantry Printer compared to Prusa Mk3 control)

    0_1540336130847_Gantry Printer Artifacting.jpg

    Does anyone out there have experience setting up timings for their servo-controlled machine? Any idea why changing the T parameter to higher values has a negative affect on print speed?

    Could these issues otherwise be caused by sending the step/dir signals from the Duet test points?

    Any advice/direction is appreciated!


  • How do you convert step/dir signals to quadrature signals?

    What you could do is wiring an Arduino to both step/dir pulses signals, and your quadrature signals, and check if they match. You may have an issue there...

  • administrators

    @jpomo10 I am not sure this is caused by.missed steps from the Duet because missed steps should be cumulative (unless you think it's missing them the same in both directions so they cancel out?). You could devise a test where you move slowly in 1 direction and at you max print speed in the other direction and confirm the repeatability of the movement axis. Skipped steps will show up as the start and end points move.

  • @fma

    The step/direction signals are converted using this board:

    Thanks for the advice, I think that indeed there is a discrepancy between the two, but I don't know what's causing it.


    I tested what you suggested and the positioning seems to be fine. That is to say, at slow speeds (10mm/s) or higher speeds (60mm/s) the extruder returns to the same position. I also verified this using the optical encoders on each axis. e.g. Tell duet to move X & Y 100mm --> encoder reads 100mm for each, as well.

    If this is the case, I don't understand why we are seeing artifacts around edges and other geometry. I assumed missed steps, but I understand your reasoning of why that is likely not the problem.

    Our quadrature converter states a Maximum Step input frequency of 5Mhz, surely we are not exceeding this?

  • administrators

    @jpomo10 said in Servo Printer Artifacts: Pulse Timing and M569 T?:

    Our quadrature converter states a Maximum Step input frequency of 5Mhz, surely we are not exceeding this?

    No you wont be exceeding this.

  • Is that PLA? I think the main problem here is poor cooling compared to Prusa i3 MK3.
    This is mostly visiable on the X edges and overhangs

  • @dragonn Yes, this is PLA. I see what you mean, and I agree, proper cooling could likely solve a lot of this. Thanks for the tip!


Looks like your connection to Duet3D was lost, please wait while we try to reconnect.