Loading ...
Sorry, an error occurred while loading the content.

Slicing error (Scanline fill is odd)

Expand Messages
  • Rob Jansen
    This error states that there is a problem with the STL model. It is generated in step 7 of the algorithm. Step 5 loops all triangles to see which triangles
    Message 1 of 8 , Oct 9, 2008
    • 0 Attachment
      This error states that there is a problem with the STL model.

      It is generated in step 7 of the algorithm.
      Step 5 loops all triangles to see which triangles intersect with the
      current Z. In that step, the contour of the model (in the current Z)
      is created by collecting all line segments that are formed by the
      intersection-points of the triangles.
      Step 7 then does the scanline fill: it loops all line segments found
      to see which lines intersect with the current Y. This creates a
      collection of points, each point being a point on the contour of the
      object (at the given Y and Z values).

      For a closed (solid) object, at each YZ lines there should be an even
      number of points: one on entering the object and one on leaving (this
      can of course be repeated). If this is not the case, we do not have a
      closed solid model that can be printed.

      STL files that have no true closed objects are possible, (laser)
      scanned objects are a good example: the STL file only contains the
      surfaces that can be seen from one particular viewpoint.
      I agree that STL files should contain valid solid models (how else
      could you do Stereo Lithography with these files) but scanning
      services do deliver non closed STL files from laser scanning.

      I was testing with an STL file (delivered to my be a professional
      laser scanning service) which (as far as I know) is supposed to be a
      closed model. With this file I got this error.
      I have no clue, yet, if this is an problem with the STL file (pinholes
      in the scan?) or a calculation error in the algorithm.

      Debugging involves: checking if the contour is closed, detect where
      the contour is broken, determine which triangles are involved at the
      location of the broken contour, checking the coordinates of these
      triangles and checking the calculated coordinates of the line segments
      (in the current Z-plane).

      The model I have contains 150000 faces and about 600 entries in
      cache_tmp for the slice where the error occurs. This means I need to
      write some extra debugging software to locate the problem.

      I am looking for smaller models with this problem.
      Please email me directly if you could help me to debug this problem
      using these models.

      Rob
    • afogassa
      ... the ... Z) ... found ... the ... even ... (this ... have a ... a ... (pinholes ... the ... segments ... to ... That hapens when the model is not welded or
      Message 2 of 8 , Oct 9, 2008
      • 0 Attachment
        --- In diy_3d_printing_and_fabrication@yahoogroups.com, "Rob Jansen"
        <rob@...> wrote:
        >
        > This error states that there is a problem with the STL model.
        >
        > It is generated in step 7 of the algorithm.
        > Step 5 loops all triangles to see which triangles intersect with
        the
        > current Z. In that step, the contour of the model (in the current
        Z)
        > is created by collecting all line segments that are formed by the
        > intersection-points of the triangles.
        > Step 7 then does the scanline fill: it loops all line segments
        found
        > to see which lines intersect with the current Y. This creates a
        > collection of points, each point being a point on the contour of
        the
        > object (at the given Y and Z values).
        >
        > For a closed (solid) object, at each YZ lines there should be an
        even
        > number of points: one on entering the object and one on leaving
        (this
        > can of course be repeated). If this is not the case, we do not
        have a
        > closed solid model that can be printed.
        >
        > STL files that have no true closed objects are possible, (laser)
        > scanned objects are a good example: the STL file only contains the
        > surfaces that can be seen from one particular viewpoint.
        > I agree that STL files should contain valid solid models (how else
        > could you do Stereo Lithography with these files) but scanning
        > services do deliver non closed STL files from laser scanning.
        >
        > I was testing with an STL file (delivered to my be a professional
        > laser scanning service) which (as far as I know) is supposed to be
        a
        > closed model. With this file I got this error.
        > I have no clue, yet, if this is an problem with the STL file
        (pinholes
        > in the scan?) or a calculation error in the algorithm.
        >
        > Debugging involves: checking if the contour is closed, detect where
        > the contour is broken, determine which triangles are involved at
        the
        > location of the broken contour, checking the coordinates of these
        > triangles and checking the calculated coordinates of the line
        segments
        > (in the current Z-plane).
        >
        > The model I have contains 150000 faces and about 600 entries in
        > cache_tmp for the slice where the error occurs. This means I need
        to
        > write some extra debugging software to locate the problem.
        >
        > I am looking for smaller models with this problem.
        > Please email me directly if you could help me to debug this problem
        > using these models.
        >
        > Rob
        >

        That hapens when the model is not welded or it has open mesh.
        there is tutorial on how to prepare meshes from laser scanners to
        send to 3d printers at Rhino web site.
      • Rob Jansen
        I know, I propose to print then an error message stating that the model is not correct. Something like: Model is not closed, unable to print. ... There are
        Message 3 of 8 , Oct 9, 2008
        • 0 Attachment
          I know,

          I propose to print then an error message stating that the model is not
          correct. Something like: "Model is not closed, unable to print."

          --- In diy_3d_printing_and_fabrication@yahoogroups.com, "afogassa"
          <afogassa@...> wrote:

          > That hapens when the model is not welded or it has open mesh.
          > there is tutorial on how to prepare meshes from laser scanners to
          > send to 3d printers at Rhino web site.

          There are even two tutorials, "Repairing STL files" for Rhino 3 and
          "Scan, Cleanup, Remodel" for Rhino 4. Rhino was on my wish-list, now
          it's in my basket! The nextengine scanner is still on the list.

          Unfortunately there are too many things to buy as a starting company.
          Therefor I am glad with this 3D printer project so I can also play
          around a bit with the stuff that 'real' companies use :-)

          Regards,

          Rob
        • afogassa
          ... not ... to ... now ... company. ... Be prepared to buy lots of ram and the fastest cpu, dealing with meshes takes far more memory then surfaces.I can t
          Message 4 of 8 , Oct 10, 2008
          • 0 Attachment
            --- In diy_3d_printing_and_fabrication@yahoogroups.com, "Rob Jansen"
            <rob@...> wrote:
            >
            > I know,
            >
            > I propose to print then an error message stating that the model is
            not
            > correct. Something like: "Model is not closed, unable to print."
            >
            > --- In diy_3d_printing_and_fabrication@yahoogroups.com, "afogassa"
            > <afogassa@> wrote:
            >
            > > That hapens when the model is not welded or it has open mesh.
            > > there is tutorial on how to prepare meshes from laser scanners
            to
            > > send to 3d printers at Rhino web site.
            >
            > There are even two tutorials, "Repairing STL files" for Rhino 3 and
            > "Scan, Cleanup, Remodel" for Rhino 4. Rhino was on my wish-list,
            now
            > it's in my basket! The nextengine scanner is still on the list.
            >
            > Unfortunately there are too many things to buy as a starting
            company.
            > Therefor I am glad with this 3D printer project so I can also play
            > around a bit with the stuff that 'real' companies use :-)
            >
            > Regards,
            >
            > Rob
            >

            Be prepared to buy lots of ram and the fastest cpu, dealing with
            meshes takes far more memory then surfaces.I can't afford to buy a
            nextengine but I'm trying to go along with a david laser scanner
            setup.
            the problem with laser scanners is that you need to prepare the
            models(paint it or dust cover it)first then get two or more scans
            before you have something usefull, then if you need to modifie the
            model takes a lot of time.
          • John Wasser
            ... Perhaps MeshLab would allow you to fix the meshes without costing anything: http://meshlab.sourceforge.net/ MeshLab is an open source, portable, and
            Message 5 of 8 , Oct 10, 2008
            • 0 Attachment
              --- In diy_3d_printing_and_fabrication@yahoogroups.com, "Rob Jansen" <rob@...> wrote:
              > Unfortunately there are too many things to buy as a starting company.

              Perhaps MeshLab would allow you to fix the meshes without costing anything:
              http://meshlab.sourceforge.net/

              "MeshLab is an open source, portable, and extensible system for the processing and editing
              of unstructured 3D triangular meshes. The system is aimed to help the processing of the
              typical not-so-small unstructured models arising in 3D scanning, providing a set of tools
              for editing, cleaning, healing, inspecting, rendering and converting this kind of meshes.
            • Paulo Cesar de Abreu
              This error can occur in two ways: 1º STL open 2º Error in the calculation and rounding... as the ROB had said: I have no clue, yet, if this is an problem
              Message 6 of 8 , Oct 10, 2008
              • 0 Attachment
                This error can occur in two ways:

                1º STL open
                2º Error in the calculation and rounding...

                as the ROB had said:
                "I have no clue, yet, if this is an problem with the STL file
                (pinholes in the scan?) or a calculation error in the algorithm."


                if the STL is open is possible repair using others programs...

                if rounding error, a way to prevent this, is implementing the
                algorithm that REPRAP uses http://reprap.org/bin/view/Main/EndMatching

                I just added this alerts to my debug while developing the program, but
                I was not in addition to implement a menu that warns you: "there was
                an error, continue, skip all messages..."? at the end of the slice, he
                say how many errors occurred in total etc..

                Translate by Google :-)
              • halebahay
                ... I Just added a file Direct Slicing of a Point Set Model for Rapid Prototyping.pdf which I think is a another way to manipulate slicing issue. I hope it s
                Message 7 of 8 , Oct 10, 2008
                • 0 Attachment
                  --- In diy_3d_printing_and_fabrication@yahoogroups.com, "Paulo Cesar
                  de Abreu" <sid3622@...> wrote:
                  >
                  > This error can occur in two ways:
                  >
                  > 1º STL open
                  > 2º Error in the calculation and rounding...
                  >
                  > as the ROB had said:
                  > "I have no clue, yet, if this is an problem with the STL file
                  > (pinholes in the scan?) or a calculation error in the algorithm."
                  >
                  >
                  > if the STL is open is possible repair using others programs...
                  >
                  > if rounding error, a way to prevent this, is implementing the
                  > algorithm that REPRAP uses http://reprap.org/bin/view/Main/EndMatching
                  >
                  > I just added this alerts to my debug while developing the program, but
                  > I was not in addition to implement a menu that warns you: "there was
                  > an error, continue, skip all messages..."? at the end of the slice, he
                  > say how many errors occurred in total etc..
                  >
                  > Translate by Google :-)
                  >

                  I Just added a file 'Direct Slicing of a Point Set Model for Rapid
                  Prototyping.pdf' which I think is a another way to manipulate slicing
                  issue.
                  I hope it's help.
                • Rob Jansen
                  ... I ve handed quite some models to the program and did not have any problems yet. I even tried scaling and rotating - no problems at all. My model indeed had
                  Message 8 of 8 , Oct 11, 2008
                  • 0 Attachment
                    --- In diy_3d_printing_and_fabrication@yahoogroups.com, "Paulo Cesar
                    de Abreu" <sid3622@...> wrote:
                    >
                    > This error can occur in two ways:
                    >
                    > 1º STL open
                    > 2º Error in the calculation and rounding...
                    >
                    > as the ROB had said:
                    > "I have no clue, yet, if this is an problem with the STL file
                    > (pinholes in the scan?) or a calculation error in the algorithm."

                    I've handed quite some models to the program and did not have any
                    problems yet. I even tried scaling and rotating - no problems at all.
                    My model indeed had pinholes (could not fix those with meshlab) i just
                    had the wrong file. The 500 EUR scanjob also gave me a fixed model
                    (created with Geomagic Studio) which is correct.

                    > I just added this alerts to my debug while developing the program, but
                    > I was not in addition to implement a menu that warns you: "there was
                    > an error, continue, skip all messages..."? at the end of the slice, he
                    > say how many errors occurred in total etc..

                    I changed the code in ButtonImprimirClick to handle the error:

                    if(linhas_slf % 2)
                    {
                    Application->MessageBox("Model is not closed, unable to print",
                    "Error", MB_OK + MB_ICONERROR);
                    free(cache_tmpy);
                    free(cache_tmp);
                    DesabilitaAbilita(3);
                    return;
                    }

                    This stops the printing process when the error occurs. In the original
                    situation the current slice would be skipped and printing just
                    continues with the next slice.

                    What I want to do is to save the stl, cache_tmpy (containing the
                    points on the current scanline) and cache_tmp (the contour of the
                    current slice) together with some other variables like the Z and Y
                    position at which the error occured.
                    Then this file contains all data needed to analyze the problem.

                    Meanwhile I am starting to change the program a bit: Replace "Browse"
                    by "Load File", I like to add a "Save" button to be able to save
                    scaled/rotated models, I added one button to scale XYZ all in one go
                    (using the same scale) and I am changing function and variable names
                    into English to make it more readable (for me that is ...).

                    On my wish list are: making the view full screen, auto zoom to fit the
                    object in the preview pane (ScrollBox1) and saving default settings to
                    be able to store default increment, timing and maybe more parameters
                    as default to be used.

                    Regards,

                    Rob
                  Your message has been successfully submitted and would be delivered to recipients shortly.