Ringing/echo after migration to duet3d, rootcause lookup



  • Hello Everyone,
    I've been following the board for a while and from the very begining I'd like to thank You very much for the support/knowdlege I was able to "absorb" 🙂

    I have duet3d with corexy printer that runs smooth as silk, however I've just swapped to the duet board on my Anycubic delta linear plus as well.

    After initial setup and some internet research and implementing "Lesson learnt" steps from my previous setup I got stuck with the echo/ringing that I cannot resolve.
    Here is my calibration cube that within "worst" lighting condition looks as follows:
    PET-G (0.2) V=20mm/s Jerk 5, ACC 500
    alt text
    alt text
    Not the best one, isn't it?

    Strange echo/ringing is visible.

    The steps I've checked (thinks I've gone through):

    1. Mechanical check:
      Belts are set with correct tension, no spring belt tensioner installed.
      Manual inspection with trying to move effector does not reveal any wobbling.
      I have a PET-G printed dampers that holds both axis arms together.
    1. Parameter "tunning"/"tests":
      I've printed dozen of those cubes with different parameters varying Jerk 15,10,5 and ACC (3000,1500,2000,500!) with slow speed.
      To my surprise unfortunately results with low ACC & Jerk did not improved the situation with the ringing/echo.
      Even on Jerk 5, Acc 500 ringing is highly visible.

    Then I moved to extrusion tuning as I thought maybe with the Preassure Advance would hide the problem.
    It did not ... 😞

    1. Further extruder tuning:
      I have a titan extruder so I tried to change motor current and tune Jerk for extruder motor.
      I re-calibrated extrusion width/steps for extruder to be precise.
      Also I've replaced the bowden tube to the capricorn as I thought the stock one might be to flexible. It did not boost quality as well.

    So far I was able to compensate the results a bit changing microsteps 1/16 with interpolation to 1/128 fixed,
    and the results are now:

    alt text
    alt text

    I'm pretty sure this is not the best quality I could achieve, maybe there is a paremeter that I did not setup correctly e.g something like steps per second or similar?

    Any help appreciated!
    Thank You!


  • administrators

    Most of the artefacts don't look like ringing to me, they look like uneven steps. What is the specification of the stepper motors in your printer, and what current have you set in M906?



  • Hello @dc42, thank You for your answer

    @dc42 said in Ringing/echo after migration to duet3d, rootcause lookup:

    What is the specification of the stepper motors in your printer, and what current have you set in M906?

    I have stock anycubic motors, that are 200step Nea17 motors, I have no exact specs but I assume 1,5A-1,7A?
    My "current" config

    M906 X1300 Y1300 Z1300 E1100 I30


  • administrators

    We really need a little more info than that. Do the motors have a part number label on them? Can you measure the resistance of one phase, and the length of the motor excluding shaft? Are you using 12V or 24V power?



  • Some time ago a Friend of mine found this specs,
    unfortunatelly no label/sticker is present on motors itself.
    alt text

    It is most probably motors that are in anycubics Kossel Delta.

    I use 12V power supply



  • I've just replaced titan with stock extruder but same pattern is visible 😞 So I guess it has nothing to do with extrusion



  • What does it actually look like while printing? Those vibrations seem extreme.

    Can you post your config file as well just in case something is really off.



  • @phaedrux said in Ringing/echo after migration to duet3d, rootcause lookup:

    What does it actually look like while printing? Those vibrations seem extreme.

    Can you post your config file as well just in case something is really off.

    Hmm it looks kinda normal no wobbling, oscilations etc.
    I attach the config.g

    I really hope I missed something 🙂0_1546643162195_config.g



  • After a print if you run M122 do you see anything for hiccups or steperrors in the move section?



  • What slicer? What G-code flavor are you running? This looks exactly what happened to my stock Anycubic delta, the cause was a bug in cura, (3.5?) 3.4 and 3.6 worked fine, one thing I noted is the stock G code flavor was different than the others, not sure if that was the cause.

    You should run Reprap flavor.
    https://imgur.com/a/eRt4M3E


  • administrators

    What are your E steps/mm? If they are less than about 400, try increasing extruder microstepping to 64 or 128 in the M350 command. If the M350 command comes after the M92 E command then you won't need to adjust the steps/mm, it will be done automatically.



  • @dc42 said in Ringing/echo after migration to duet3d, rootcause lookup:

    What are your E steps/mm? If they are less than about 400, try increasing extruder microstepping to 64 or 128 in the M350 command. If the M350 command comes after the M92 E command then you won't need to adjust the steps/mm, it will be done automatically.

    M350 X128 Y128 Z128 E32 I0                            ;  NO interpolation
    M92 X640 Y640 Z640 E812                              ; Set steps per mm
    

    Regarding slicer i use simplify3d with Reprap firmware setup:
    0_1546684717781_Screenshot_20190105_113805.png

    This is YT video during the print
    https://youtu.be/h0db7zNWiL4



  • Not too sure what this does in the grand scheme but you have relative extruder moves enabled in config.g but not checked in S3D. I always assumed that these needed to track together.



  • S3D includes at the start of g-code proper command to switch to relative/absolute dependent on the checkbox. So I don't think this should cause any problem.



  • OK, if the checkbox in S3D is unchecked is it not going to produce absolute extruder moves and if there is an M83 in config,g is the printer not going to expect relative moves?
    Maybe this doesn't work as I thought.

    Answered my own question. I didn't realize that S3D would output the M82 as a result of the checkbox. Learned something new.


  • administrators

    @morf said in Ringing/echo after migration to duet3d, rootcause lookup:

    M350 X128 Y128 Z128 E32 I0                            ;  NO interpolation
    M92 X640 Y640 Z640 E812                              ; Set steps per mm
    

    Using XYZ microstepping of 128 may lead to processor overload (although probably only in faster travel moves), so I suggest you reduce it to either 16 with interpolation, 32 or possibly 64. Your extruder microstepping and steps/mm look good.


 

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