Jump to content

Near perfect print/Z-Seam - Unable to reproduce (?)


NX3DP
Go to solution Solved by GregValiant,

Recommended Posts

Posted · Near perfect print/Z-Seam - Unable to reproduce (?)

Links and information at bottom of post.

 

Quick background. I bought my first 3D printer a little over 2 years ago, a Qidi Tech i-Mate S (Core XY). Due to me having moderate/severe OCD/perfectionism I am pretty much still tuning and calibrating my printer 😛

 

As a part of this, I printed out a series of Ghosting/Ringing/XY Cubes with different iterations of configuration settings in Cura. This is because I want to know what each setting does, and how the print changes with each setting, as well as trying to find the "perfect" print settings.

 

I have repeated this process twice now, because the first time I did not document the process well enough, and so when I had my "perfect" print, and could not reproduce it, I thought it was because I might have missed to document a setting.

 

As I did my second round of XY Cube iterations and again got a "perfect" print, and having saved the Cura profile, project file, G-Code output and the exact profile settings, and still unable to reproduce the results (except for re-printing the originally generated G-Code file), I am at a loss as to what is going on here.

 

Both of these near perfect prints have occurred when changing settings under Travel and in relation to Retract At Layer Change and Retract Before Outer Wall (Enabled/Disabled in various combinations).

 

The profile is based off of a Duplicate of the Default Profile Print Core 0.4mm - 0.2mm.

 

So here is a comparison photo.

VhLeFrU.png

 

X7COGzV.png

 

83l8uRB.png

LEFT: Original "CUDE_006" (CUra DEfault profile, iteration 006) XY Cube print.
RIGHT: CUDE_006 Cura project file opened and re-sliced and printed...

 

Attempts to recreate the print results I have tried; 

1. Opening the CUDE_006 .3mf Project File and re-slicing. Not a match.
2. New project in Cura, selecting the CUDE_006 profile, importing the XY cube, positioning it as per the CUDE_006 settings and re-slicing. Not a match.
3. New project in Cura, duplicate the default profile, configure settings as per document (Excel sheet, nx_qt_ims_cu_profile_settings_table_v1_chrjon_240223.xlsx), import XY cube mode, position it correctly, re-slice. Not a match.

 

Reprinting the original CUDE_006 cube G-Code file produces pretty much identical final results.

 

I also had thoughts about the method of sending the file to the printer, if that could impact the G-Code somehow, i.e. Wireless to USB drive (in printer) vs.  file save/copy to USB drive in PC and then plugging the USB drive in to the printer. But nothing observed here.

 

I have concluded that this is happening in slicer upon generating and saving the output G-Code file because I have used Notepad++ to compare the G-Code files with one another (Compare plug-in) and I can see the variations right away in the G-Code. I have also verified these deviations by printing non-matching G-Code files and they do not produce the same result (no surprise there).

 

Comparing the two files in Prusa G-Code Viewer just confirms what Notepad++ Compare is telling me.

 

There is no relation to physical state of the printer or filament, since I can reproduce the "perfect" cube print at any time simply by reprinting the original G-Code, despite having had a number of "bad" cubes printed just before it. If the produced G-Code was identical, but the resulting print differed, then I would accept that it is something happening outside of the slicer. But that is not the case.

 

Comparing all of my attempts to re-create the output results by looking at the G-Code files, none of my attempts match the original one.

 

Why is this so important to me?
1. The resulting print and Z-Seam is in my opinion near perfect.
2. Why would the Cura slicer software produce different G-Code output files based on the source Project file source?
3. Why am I unable to recreate the results when setting things up EXACTLY as they were when it produced the CUDE_006 XY cube print?

 

I would expect, that if I open a project file, or re-create the exact settings as per my original print/G-Code, that the output would be identical. Is that too much to ask?


YouTube video:


IMGUR photos: https://imgur.com/a/cura-xy-ghosting-ringing-cube-comparison-wcA6rZf


OneDrive Files: https://1drv.ms/f/s!Au48vW16rQ29jo1TV-CVVtl-L6exKg?e=JlN8rN

 

Printer: Qidi Tech i-Mate S (Core XY)
Filament: Polymaker PolyTerra PLA, Lava Red
Filament Box: Creality Space Pi (PLA setting, 50C)
Slicer: UltiMaker Cura v5.7.1 (attempted to re-create with v5.7.2 also, no match)

CU_XY_Ghost_Cube_CUDE_006.3mf CU_XY_Ghost_Cube_CUDE_006.gcode CU_XY_Ghost_Cube_CUDE_006_RECREATE.gcode CU_XY_Ghost_Cube_CUDE_006_RESAVE.gcode Ringing_Ghosting_Cube.stl

  • Link to post
    Share on other sites

    • Solution
    Posted · Near perfect print/Z-Seam - Unable to reproduce (?)

    In Cura, go to "Manage Profiles" and then "Import" that perfect gcode file.  The base pofile will import along with all the changes listed in the "Settings" section that is at the end of any Cura gcode file.

    The new profile will have the name of the Gcode file.  Use that profile and generate a new gcode.  See if that shows an improvement.

  • Link to post
    Share on other sites

    Posted · Near perfect print/Z-Seam - Unable to reproduce (?)
    1 hour ago, NX3DP said:

    2. Why would the Cura slicer software produce different G-Code output files based on the source Project file source?

    Different versions of Cura? CuraEngine does not stand still, they're always trying to improve it, although obviously that doesn't always end up improving every single slice.

     

    2 hours ago, NX3DP said:

    I printed out a series of Ghosting/Ringing/XY Cubes with different iterations of configuration settings in Cura. This is because I want to know what each setting does, and how the print changes with each setting, as well as trying to find the "perfect" print settings.

    People are way too obsessed with calibration cubes. The quality of your calibration cubes doesn't matter. The quality of the prints you produce does. And almost always they'll require different settings to the cubes. And to each other.

     

    1 hour ago, NX3DP said:

    Due to me having moderate/severe OCD/perfectionism I am pretty much still tuning and calibrating my printer 😛

    As someone who actually has OCD (diagnosed by my psychiatrist), I find it offensive when people say things like this. It's belittling a serious problem which affects not just my quality of life, but also has deleterious effects on my physical health.

  • Link to post
    Share on other sites

    Posted · Near perfect print/Z-Seam - Unable to reproduce (?)
    7 hours ago, GregValiant said:

    In Cura, go to "Manage Profiles" and then "Import" that perfect gcode file.  The base pofile will import along with all the changes listed in the "Settings" section that is at the end of any Cura gcode file.

    The new profile will have the name of the Gcode file.  Use that profile and generate a new gcode.  See if that shows an improvement.

     

    I will try this and see what happens.

     

    6 hours ago, Slashee_the_Cow said:

    Different versions of Cura? CuraEngine does not stand still, they're always trying to improve it, although obviously that doesn't always end up improving every single slice.

    Nope. Same version of Cura. 5.7.1. Interestingly it produces different G-Code if I slice the file from a "cold start", i.e. Cura closed, vs. slicing something else first and then opening my project file and slicing it.

     

    6 hours ago, Slashee_the_Cow said:

    People are way too obsessed with calibration cubes. The quality of your calibration cubes doesn't matter. The quality of the prints you produce does. And almost always they'll require different settings to the cubes. And to each other.

    I would tend to agree, but based on what I am doing, i.e. testing and tuning each setting in the slicer to see what it does, a calibration cube is a pretty good model to print. And it doesn't negate the fact that the slicer is unable to reproduce the same G-Code when doing the exact same thing in it.

     

    6 hours ago, Slashee_the_Cow said:

    As someone who actually has OCD (diagnosed by my psychiatrist), I find it offensive when people say things like this. It's belittling a serious problem which affects not just my quality of life, but also has deleterious effects on my physical health.

    I find it offensive when someone who has OCD accuses me of not having OCD (which is also diagnosed)... You don't think calibrating a 3D printer and printing "cubes" and what not for two years is not a pretty obvious sign?

    How many times do you think I've thought about throwing my 3D printer out the window and doing something else. And then I realize that whatever that something else would be, I would be approaching it in a similar matter and still driving myself to "insanity".

  • Link to post
    Share on other sites

    Posted · Near perfect print/Z-Seam - Unable to reproduce (?)

    I just tried to look at your project file (didn't have time earlier) and it seems you're using a custom machine definition - you'll need to post that for us to be able to load the project.

     

    I did compare your three gcode files and there's no differences which should make a salient difference in the finished product. The "RESAVE" file has it printing skin lines in a different order (probably due to monotonic order) but it's basically taking one chunk of code and moving it to another place (and you don't say there's a problem with skin).

     

    The other difference is that the extrusion distances on many of the lines are different - but they're all within about 10 micrometres of each other. This is well within the bounds of rounding errors (which is why Cura resets the extruder position every so often) and in almost all cases are under one E step on your average printer (scrolling through the code, about the biggest difference I could find would have been two E steps on my printer, and that was very much an outlier).

  • Link to post
    Share on other sites

    Posted (edited) · Near perfect print/Z-Seam - Unable to reproduce (?)

    I've uploaded the contents of my Cura 5.7.1 Configuration Folder "C:\Users\nonex\AppData\Roaming\cura\5.7" to my OneDrive, in the folder "_AppData". A ZIP file with the entire folder structure intact, and also all the individual files. I think that is what you are asking for.

     

    The Cura profile/definitions were provided to me by Qidi Tech Support.

     

    Yes, I did notice that the G-Code files are very similar in some cases, but never identical. And since G-Code is processed top to bottom in sequential order, then re-ordering of lines or chunks of lines of the G-Code obviously has a noticeable impact. And why would the slicer do that if I am doing the setup exactly the same before slicing and saving the file?

     

    The closest files are (guesstimate) 99,9% identical, with only what appears to be, as you say, some kind of rounding error in difference.

     

    But what a cruel act of faith it would be for a random rounding error to produce this "perfect" result, TWICE (CUBE_24 was my original winner in the first round of testing, and now CUDE_006), when I am testing different settings.

     

    I also realize that even if I am able to consistently replicate this G-Code and print result, as soon as I tweak my next setting in the profile it will go "out of wack" again and not be "perfect". But it just bugs me to no end how this is possible... 😐 

     

    I mean... computers, binary 1's and 0's should be perfect, and this is making me think it is not so, which makes me feel very... hopeless (?)😞 

     

    @GregValiant I did just now recreate the profile by importing it from the G-Code, awesome tip/trick that I did not know was possible! Unfortunately the raw G-Code of the resulting sliced file is not an exact match but I will print it out and verify!

     

    UPDATE: Well smack me sideways and call me Sally! 😛 

    Re-printed cube from recreated print using the imported profile from the G-Code of "CU_XY_Ghost_Cube_CUDE_006.gcode" produces pretty much exactly the same result! Even though the resulting G-Code differ significantly...

    G-Code uploaded to OneDrive, file name: CU_XY_Ghost_Cube_CUDE_006_RECREATE_FROM_GCODE.gcode

     

    The question then becomes, will this be a viable way to move forward with my iterations of settings to keep this "perfect" base print to continue to improve. Probably not 😐

     

    I am very interested in how this "works". How and what makes it behave this way?

    Edited by NX3DP
  • Link to post
    Share on other sites

    Posted · Near perfect print/Z-Seam - Unable to reproduce (?)

    I just realized what this is really about for me.

     

    Consistency.

     

    No wonder I am still trying to calibrate and tune the printer and my profile settings when the underlying workflow/process is incapable of producing the same results when repeated exactly the same.

     

    How am I supposed to find the right settings when it is inconsistent like this.

     

    Based on my experience up until this point, I might as well just run the default profile and #YOLO my prints, or maybe even better, just randomly enable/disable settings in the slicer and hope for the best? 😛 

     

    You might think I am exaggerating but in my opinion the difference between my two cubes in the original post is "average overall finish/seam" vs "almost perfect overall print/seam".

     

    This can't be normal and something everyone has to deal with that 3D prints stuff?

     

    Let's say someone prints a model (single) and it turns out really nicely.

    They want to print a batch of that model. They put 4 of the models on the plate and print. Only now the 4 newly printed ones look "like shit" compared to the original one, despite using the exact same settings, profile, filament etc.

     

    Perhaps a bad example because I know first hand what printing multiple objects vs. one at a time (multiple on plate but sequentially) vs. printing a single model/item at a time can do with your prints and lines/quality etc.

     

    If the definition of insanity is doing the same thing over and over again and expecting different results, then this would be the exact opposite?

     

    😐 

  • Link to post
    Share on other sites

    Posted · Near perfect print/Z-Seam - Unable to reproduce (?)

    Well alrighty then. After some deliberation I went ahead and ordered a Bambu Lab X1C Combo printer since they are having an anniversary sale.

     

    Adios and thanks for the replies!

  • Link to post
    Share on other sites

    Create an account or sign in to comment

    You need to be a member in order to leave a comment

    Create an account

    Sign up for a new account in our community. It's easy!

    Register a new account

    Sign in

    Already have an account? Sign in here.

    Sign In Now
    • Our picks

      • UltiMaker Cura 5.8 beta released
        Another Cura release has arrived and in this 5.8 beta release, the focus is on improving Z seams, as well as completing support for the full Method series of printers by introducing a profile for the UltiMaker Method.
          • Like
        • 1 reply
      • Introducing the UltiMaker Factor 4
        We are happy to announce the next evolution in the UltiMaker 3D printer lineup: the UltiMaker Factor 4 industrial-grade 3D printer, designed to take manufacturing to new levels of efficiency and reliability. Factor 4 is an end-to-end 3D printing solution for light industrial applications
          • Thanks
          • Like
        • 3 replies
    ×
    ×
    • Create New...