Jump to content

GregValiant

Expert
  • Posts

    5,060
  • Joined

  • Last visited

  • Days Won

    199

Posts posted by GregValiant

  1. It's the ratio of filament in vs extrusion out multiplied by the length of the extrusion

    Given a line width of .4 and a layer height of .2 and filament diameter of 1.75 it's 1mm filament / 30mm extrusion (2.4/.08) and inversely 1mm of extrusion would require .033mm of filament.  That makes a base line for Cura of 100% flow volume.  Simplyfy3d considers that the area under the nozzle will never be an exact rectangle, but will have curved sides.  Simplyfy3d uses a fudge factor of 1.2 when calculating the area under the nozzle (so .08mm^2 becomes .096mm^2 in Simplyfy3d calculations).

    I have no idea where it is on GitHub.

     

  2. "So it best to design a part wall thickness to an exact multiple of the line width (so line width 0,8mm => part thickness 3,2 or 4mm) or just slightly less?"

    YES for thin walls and NO for walls that will require infill.  A thin wall is trapped.  In a thick structure Cura can fudge the inside infill area so the outside of the wall is correct.

    The Z dimension is another story.  To my limited knowledge (someone correct me if I'm wrong please) - Cura will not make partial layer changes just to hit your dimension.  If layer height is .2 and the top of the print is 3.5 then Cura will either make it 3.4 or 3.6.  If that .1mm difference is important then you must make some adjustments to the model or print settings or splice a file together.

     

  3. I think you will have to do some hand coding.  Create the gcode file with the print speed at 60.

    Open the file in your text editor.  Layer numbering in Cura starts with 1 but in Gcode it starts at 0.

    Scroll down to Layer:0 and right under that line add M220 S333

    Use the search function to find Layer:4 and right under that line add M220 S100.

     

    The first M220 will change the feed rate to 333% of 60 (199.8mm/s) for layers 0, 1, 2, 3.  The second M220 will change it back to 100% for layer 4 (Cura layer 5) to the end of the file.  If your max feed settings in config.h are lower they will over-ride and limit the speed.

     

  4. Whether the printer understands the codes or not (dependent on the firmware) the pause code should show up in the gcode file.  The pause plugin named "Pause at Height (BQ printers)" only adds an M24 to the file.  The others add short routines with more flexibility like maintaining the temperature and parking the head.  I'm using 4.6.1 also and the code shows up in the files.

  5.  

    "The first lines around the component" - In Cura under Bed Adhesion check which option you have.  A brim helps hold the part down.  A skirt gets the nozzle flowing before the part starts to print.  Either one has an option for distance to the part.  If your part has enough area touching the build plate you probably don't need either.

    Also in Cura, go to Monitor/Manage Printers/Machine Settings and check that your bed size and height are correct for your printer.  While in there, check the "Start Gcode" window and see if there are any extrusions (lines with Exxx).  That is the area where additional priming lines may be and they would show up in every gcode file (usually as verticle lines on the left side of the bed).

     

    I had to translate to English.  I think the software got it right.

  6. That last M104 may be there as a safety to insure that the hot end is indeed off. 

    In Cura you can use the plugin "Search and Replace".  Have the search term "M104 S0" and the replace term "M104 S190".  That will change those last two shut-offs and keep the hot end on.

    Be aware that as long as the plugin is active, ALL your gcode files would leave the hot end on.

  7. I'm old.

    I started with analog computers, and punch cards.  The g-code went on 1" wide paper tape with 1/8" holes punched in it.  The paper tape was fed into the CNC machines and they moved and worked all by themselves.  It was magic.  (No...they were not steam powered.)

    Over the course of time I came up with some observations that have worked for me.  I call them rules, but they're more like guidelines.

     

    #1.  When you're dumb ya' suffer.

    #2.  Never upgrade an operating system.

    #3.  Everything affects everything.

    #4.  Never blindly trust a Gcode file that was created by someone else.

     

    There are a few more (yes, some involve the opposite sex) but I won't waste anymore of your time with them.

  8. Is it possible to disable the automatic com port connection in Cura?

    An alternative - can I close the port connection from within Cura?

    I have my own program I use for printer communication and I have to close Cura to make the port available for my program and then re-start cura and reload the model.  It's a bit annoying.

  9. It might be time (maybe past time) for some maintenance.  Take apart the nozzle and hot end, check the extruder drive, the belts,  clean the dust and fibers out of the fans and off the wheels and tracks.  If they have them, check the bowden tube fittings and see if the tube is sloppy and can be pushed in and out.

    These machines work pretty hard and it sounds like you've had them running for a while.

    • Thanks 1
  10. It would be odd for the printer to be making things up by itself.  Really odd for multiple printers to making up the same stuff.

    That top image looks under-extruded and the layers don't look bonded together.  Is that what you mean by "not solid"(?) or is it more like vase mode and missing infill?

    You have tried checking with a Gcode file you know is OK (like one of the bench models)?

  11. Those are pretty generic commands.  Your "End Gcode" area should probably have an M107 in it to turn the fan off in case it was on.

    I actually stole some code from the start of one of the Benchmark pieces (the klein bottle).  I liked the fact that it put down some lines before printing actually started.

    A heavier flow at the start makes sure that the plastic is getting pushed into the build surface, and that the lines are welded together.  It's a crutch, but it worked to get me started with successful first layers every time.  Consistent bed leveling and just practice with the machine will get you going.  This FDM stuff is not a fast manufacturing process anyway so going slow at first (20-25mm/sec) isn't a bad idea either.  We all need something to get the confidence up.

    I had problems at first with what I thought were clogs in the nozzle.  It turned out (after a bunch of youtubes and articles) that my hot end had been put together with a slant cut on the end of the bowden tube.  It just wasn't sealing when the nozzle was tightened and plastic was escaping and jamming the hot end.  It was a pain to chase down, but I have to say I learned a lot.  The machine still tries to fool me, but I'm much better now at figuring out its tricks.  Good Luck. 

    Here's my main project.  I think there's 55 separate printed pieces on it.

    DBike 2.JPG

  12. It isn't easy to explain.  I wrote my own program to count characters in gcode files using the "Byte Offset" to find the line where a print stopped.   Cura does not have a window that shows printer responses to Gcode commands so you need to use something like Pronterface.  The code snippet below (for Marlin firmware) is to re-start a file at a point 11500750 bytes into the file.

    Notice that there is a G28 line in the code.   If the print is tall or at the front of the bed this can cause a crash.  If the steppers have timed out and become disabled there must be a "Home" move before a restart can occur.

    Send M27 to return the Byte location of the printing file.  This IS SUPPOSED to be where the printer stopped.

    Send M114 to return the XYZE locations for a restart.

    Send M20 to list the files on the SD card in DOS 8.3 format (as the printer sees them)

     

    M21                                 ;Initialize the SD card
    M23 LWRROD~1.GCO  ;File name in DOS 8.3
    M26 S11500750             ;Byte Offset location to start at (from response to M27)
    M140 S50                       ; Bed temp
    M190 S50                       ;Wait for bed temp
    M104 S215                      ;Hot end temp
    M109 S215                      ;Wait for hot end
    M105                                ;Report temperature
    M220 S90                        ; Feed rate at 90%
    M221 S105                      ; Flow rate at 105%
    G28                                  ;Home all axis
    M106 S255                      ;Fan on 100%
    G0 F1200 Z94.8               ; This is the resume Z +10.  The head has to move up above the print so it doesn't crash.
    G0 F2400 X92.442 Y137.203  ; XY resume location
    G92 E2484.26551            ;E resume location as a retraction
    G0 F300 Z84.8                  ;Drop down to the Z where printing will re-start
    G0 F1200 E2492.26551   ;Un-retract to the real E location to prime the nozzle
    M24                                   ;Print

  13. OK, I will not say "Now before you stay bed leveling, temperature heating or bed adhesive."

    Check your 1st layer height.  I like .2mm but sometimes Cura will default to .28mm.  I think that's a bit too much gap to fill.

     

    Just before the line ";Layer:0" in your gcode file, add a line "M221 S115" to set the flow rate at 115%.  You want to guarantee that the plastic gets smashed into the nice warm plumb build surface that has hair spray on it.

    Search the gcode file for "Layer:1" and just before it, insert "M221 S100" to set the flow back to 100%.

     

    I like to level the bed with parchment paper.  It's only .06mm thick and nothing sticks to it.  It's used to line oven baking pans.  I find that I'm much more consistent than I was using printer paper.

  14. You may need to review bed leveling process.  There are a lot of youtube videos on it.

    When the nozzle is too close to the bed on the first layer, the hole is effectively closed by the bed and the plastic can't get out.  The nozzle should not be leaving marks on the bed (it should be "1st layer height" above the bed.)

    When your second layer starts the head has moved up and flow starts.

  15. Your initial layer height is .2mm and the "Initial Layer Line Width: 150".  Notice that most of the rest of the line widths are the same as the nozzle diameter at .4mm.  (The other odd one is skirt line width at .2.)

    The flow at 100% and E-steps near 100 make good sense.  I think the machine may be trying to compensate for that big line width number on the first layer and putting down too much plastic.  Then with the very small layer height of .08mm, there is nowhere  for the plastic to go and the 2nd (and probably a few more) layers look like they are over-extruding.  What they are really doing is dealing with the heavy first layer.

    I suggest you change the Initial Layer Line Width to .4 and the Skirt Line Width to .4.  Then try printing it again.

     

     

  16. I've been intrigued with saving a print-gone-wrong since I started using the Ender last December.  Right away I came across M27 and M26 that use the Byte Offset of the print position in a file.  I got Micorsoft Excel to talk to the printer and by crunching a gcode file with some Visual Basic macros, I figured it out. 

     

    I started printing Mrs. Rabbit as a conventional slice with infill and wondered what would happen in Vase mode.  So I aborted the print and had Excel send an M27 to get the byte offset and then the XYZE locations, sliced the Jessica file in vase mode and was able to justify the byte location within the new file, used M26 to jump to the byte offset in the new file and used a G92 to synchronize the extruder, and regular G moves to get the XYZ synced in and it worked very well.  Off course she has some.......interesting.......horizontal parts that don't do well in vase mode, but what the heck.  An experiment is an experiment.

     

    She would be impossible to print completely in vase mode since the model is standing on a flat raised dais.  The file with infill handled the platform (and part of her feet) as you can see in the second image.  The last image shows the area of the seam from the other side and not backlit.  Even with the print in my hand I can't tell exactly where one file stops and the other starts.

     

    I am officially pleased with myself for figuring this out (even if I do feel a bit like Igor).  Mrs. Rabbit appears courtesy of Thing-i-verse and she will be on a pedestal that proclaims "I'm not bad.  I'm just drawn that way."

    DSCN2404.JPG

    DSCN2405.JPG

    DSCN2406.JPG

×
×
  • Create New...