GCode Tuning



  • Hi -

    Is there a recommended order to tuning the configuration gcode parameters for things like jerk, acceleration and speed changes?
    They are all sort of interrelated and codependent in my mind.

    I tried searching but could not find anything relevant so far.

    I am trying to use KISS slicer to tune my gcode parameters. I believe that Slic3r can do this too, but you have to go into the gcode to tweak it.

    Here is a discussion about it -

    http://wiki.snolabs.com/index.php?title=Tuning_Wizards#.3CTUNINGVAL.3E_G-code_token_only

    Here for example are my settings and results. (Ignore the M204 in the gif, pretend its M201).

    KISS Settings

    Here is how it varies the M201 for example in the gcode- right before a layer change it increments the parameter. This was a random location -
    alt text

    I ended up with a value of 1300 for accelerations. Picking the 'best' spot of surface finish.

    Test Cube Results

    Test Cube Calc


  • administrators

    If you want to play with accelerations in GCode files, I suggest you use M204, not M201.

    You appear to be trying to adjust acceleration to try to reduce ringing. Have you tried using the dynamic acceleration adjustment feature of RRF? See https://duet3d.dozuki.com/Wiki/Gcode#Section_M593_Configure_Dynamic_Acceleration_Adjustment.



  • @sinned6915 said in GCode Tuning:

    Is there a recommended order to tuning the configuration gcode parameters for things like jerk, acceleration and speed changes?

    Stay tuned!



  • dc42: I tried playing with M204 but I could not get the entire parameter set figured out. which parameter to vary- printing or travel accelerations.
    no matter what I tried, it seemed that something was limiting / capping the movement and the ripples were 'straight' and did not converge or vary. M201 worked out 'easier' I could tell a difference in results.

    I was not trying to tune out ringing specifically, i was using it as an artifact to gauge the 'sweet spot'.

    these are TMC2224 drivers. I chose them on purposed to try and mimic the native Duet drivers as much as I could. that and the donor switched his board to 5160's and i got these 'cheap'.

    i think that tuning in general is a multi variable problem with many interlinked and interdependent constraints.

    for example, accelerations are affected by jerk values too.

    I guess following the KISS example of tuning filament printing temp, extruder flow, preload viscoelasticity and then finally destring , there should be some rational to do velocity, acceleration, jerk, cooling.

    there has to be some orderly method of establishing upper bound and lower bound limits.

    i know some people are happy with 'stock' settings, but i am trying to learn more about how these parameters interrelate and affect print quality.



  • M204 only has two parameters. P and T, for print and travel.

    https://duet3d.dozuki.com/Wiki/Gcode#Section_M204_Set_printing_and_travel_accelerations

    If you set M201 at your maximum acceleration value, M204 acts as a lower limit applied in either print moves or travel moves. So in your case, M204 P should have worked.

    @sinned6915 said in GCode Tuning:

    i think that tuning in general is a multi variable problem with many interlinked and interdependent constraints.

    Tell me about it!



  • @Phaedrux

    I am trying to learn how these different parameters work together so that I can better optimize my settings to get better prints.

    That might be exactly the thing I could not figure out. If M204 caps BOTH print and travel speeds, then which one do you tune first?

    Is there some proportion, or golden ratio between the 2?

    M204 P<TUNINGVAL> T2000

    OR

    M204 P2000 <TUNINGVAL>

    what / how do people dial in their printers? does anyone do this? i am surprised that there is no discussion about how to do this.


  • administrators

    The main effect of tuning acceleration is to control ringing. Looking at your photos, that's exactly what you have done. So IMO you should try tuning DAA instead.



  • dc42:

    use the KISS <TUNINGVAL> procedure to vary the Hertz over the height or measure the peaks and speeds?

    thanks

    Dennis


  • administrators

    @sinned6915 said in GCode Tuning:

    dc42:

    use the KISS <TUNINGVAL> procedure to vary the Hertz over the height or measure the peaks and speeds?

    thanks

    Dennis

    I suggest you measure the peaks and speed to get an estimate, then vary the Hz from perhaps 30% below to 30% above that value.


Log in to reply