Skin line distance

[cakeforcat 2]

Hi, I've recently calibrated my flow rate on my Ender 3 so it's extruding at a 0.44mm line width accurately but realized that the spacing between the skin lines (bottom layers at least) are spaced away from each other creating a mesh rather than a surface. You can see it on the image, the tiny holes are big enough to let water pass through. I've checked the generated gcode and if I understood it correctly(a move in X between the end and start of a new line) the distance between each line is something around 0.62mm, which would explain the the size of the gaps. Is there a setting to get rid of this effect or is it a bug in Cura? Note that there is no gap between outer shell lines. Also I've seen as it was printing that the gaps were in every bottom layer so it's not a bed adhesion problem.

Edited September 24, 2019 by cakeforcat

[gr5 2,271]

I have two answers for you.

 

1) See "initial layer line width" parameter in cura.  Make sure that is 100% (not the default 120%).  Also look for line width everywhere in cura and make it match your nozzle (typically 0.4mm).

 

2) Regardless - you have underextrusion - at least on the bottom layer.  Answer 1 will not help you much.  If you use default line width (I'm guessing it's 0.35) and you have initial line width at 120% then .35 * 1.2 is .42mm line width.  Anyway regardless of what the actual line width is, the printer commands the extruder to extrude the correct amount to just perfectly fill in the lines.

 

Your underextrusion could simply be a leveling issue (assuming it's only happening on the bottom layer) or it could be that your bottom layer is just printed to fast and/or too thick or too cold.

 

I guess the big question is this: is it doing this on the third layer (or higher) as well?  If so then you are printing too fast for your printer (it could be there is something wrong with the printer but at this speed and temp, the extrusion portion of the printer can't keep up).

 

 

[burtoogle 516]

Hello @cakeforcat, if you could please save the project file for your example (File -> Save) and attach the .3mf file to this thread, I will check the spacing of the generated lines WRT the layer height and line widths used in the profile. Thanks.

[cakeforcat 2]

My nozzle diameter is indeed 0.4mm but by a suggestion(made by CHEP on YouTube if I remember correctly) of making the line width a bit larger than that I have set 0.44 in every setting in Cura and calibrated my flow rate to match that (Using a 1 line thick spirally printed cube and measuring the wall thickness). After that the line width is 0.44 as I've measured and this is working just fine when printing the outer lines on the bottom layers and the spiraling wall and it's not a problem of just the first layer as I've mentioned, but is happening on all 5 bottom layers I specified. I'll try to slow my speed a little but It's weird that it's only happening on the diagonal inner lines and not on the outer lines.

 

[cakeforcat 2]

@burtoogle I'm 90% sure I didn't change anything in the settings since then but to be safe I also attached the .gcode I printed with and the .stl model of the cube(this I don't know why, I think I'm just throwing as much information as I can for no reason 😛 )

CE3_cube.3mf cube.stl CE3_cube.gcode

[burtoogle 516]

Thanks for the files.

 

This is the first layer of your project without any changes by me...

 

 

This is the first layer after resetting the flow to 100%...

 

 

The gaps between the lines look smaller with the 100% flow.

 

[cakeforcat 2]

I have a different cube printed that shows that's the case too. Right now I'm mostly thinking it actually is a bug in Cura, because going back to 100% flow indeed closes the gaps but at the expense of line width accuracy(I'm measuring 0.49 at 100% flow and it shows over-extrusion on outer shell lines.). The thing is, That just changing the flow affects the whole print wherein here the only affected area is the inner lines, they shouldn't be this far apart when I have a line width set to 0.44 and flow rate calibrated to make 0.44 lines.

[burtoogle 516]

Here's your first layer being printed (100% flow) visualised in a 3rd party gcode viewer. The numbers you see are calculated from the gcode.

 

 

You can see that the gcode viewer is reporting an extrusion rate of 0.020 mm of filament per mm of line length. Assuming you are using 1.75mm dia filament, that equates to (2.405 * 0.020) = 0.048 mm^3 per mm of line length. The layer height is 0.12 mm and so the line width will be 0.048 / 0.12 = 0.401 mm. That corresponds nicely with the line width in the profile.

 

So it looks like cura has got its sums right here.

 

 

Edited September 24, 2019 by burtoogle
Correct layer -> line

[cakeforcat 2]

Yeah, apart from 0.401mm not being 0.44 I have set(But I' guess you mean it's not a big difference which I agree), You can clearly see the literal gap between these lines(or maybe they are just this thin because of this gcode viewer, correct me if I'm wrong). I'm not saying cura is miscalculating extrusion, It's ok as I've calibrated it (Hence the 90% flow rate). What I'm saying is those gaps between those lines exist at all. Now I don't know if the gaps between them are because of spacing in the gcode(the slicing is wrong, so either a wrong setting or a bug) or just under-extrusion(which only affects those skin lines as the rest of the print is fine, which is really weird and means my settings are wrong, but then If i know the lines are 0.44(I've measured) and you calculated them to be around 0.4 then why is cura with a set line width of 0.44 slicing with such big gaps?).

[burtoogle 516]

17 minutes ago, cakeforcat said:

Yeah, apart from 0.401mm not being 0.44 I have set(But I' guess you mean it's not a big difference which I agree),

 

No no, the line width in the profile was set to 0.4mm so the difference is weeny (technical term).

 

As for the line spacing, I'm too lazy to work it out from those diagonal lines so let's look at some vertical lines.

 

Here's the end of one of the wall lines...

 

 

And here's the end of the next wall line...

 

 

One finishes at an X coordinate of 120.7, the other at X coordinate 121.1. So the line spacing is 0.4. That matches the line width used in the profile.

 

So the gcode confirms that the extrusion rate is correct and the line spacing is correct WRT the profile.

 

 

[cakeforcat 2]

Hmmmm, that's weird. Here's my profile i definitely sent you in cura, and the one with which i printed.

[cakeforcat 2]

Just to be clear with my approach(not that I'm saying yours is wrong, just want make it clear so someone can correct me).

Here is a part of the gcode:

;TYPE:SKIN
G1 F1350 X119.65 Y100.35 E99.2533
G1 X119.65 Y119.65 E99.8888
G1 X100.35 Y119.65 E100.5243
G1 X100.35 Y100.35 E101.1598
G0 F7500 X100.525 Y101.084
G1 F1350 X101.084 Y100.525 E101.18584
G0 F7500 X101.706 Y100.525
G1 F1350 X100.525 Y101.706 E101.24083
G0 F7500 X100.525 Y102.328
G1 F1350 X102.328 Y100.525 E101.32479
G0 F7500 X102.951 Y100.525
G1 F1350 X100.525 Y102.951 E101.43776
G0 F7500 X100.525 Y103.573
G1 F1350 X103.573 Y100.525 E101.5797
G0 F7500 X104.195 Y100.525
G1 F1350 X100.525 Y104.195 E101.7506
G0 F7500 X100.525 Y104.817
G1 F1350 X104.817 Y100.525 E101.95046
G0 F7500 X105.44 Y100.525
G1 F1350 X100.525 Y105.44 E102.17933
G0 F7500 X100.525 Y106.062
G1 F1350 X106.062 Y100.525 E102.43717
G0 F7500 X106.684 Y100.525
G1 F1350 X100.525 Y106.684 E102.72398
G0 F7500 X100.525 Y107.306
G1 F1350 X107.306 Y100.525 E103.03974

You can take one of the lines and a move between the lines

G1 F1350 X104.817 Y100.525 E101.95046
G0 F7500 X105.44 Y100.525

and by looking at the difference in the X direction it comes to a move of 0.623mm which is larger than the line width of 0.44 you can see in my screenshot(and even more than the 0.4 you say is in the profile I sent). Is it intentional or am I reading it incorrectly?

[tinkergnome 927]

12 minutes ago, cakeforcat said:

by looking at the difference in the X direction it comes to a move of 0.623mm which is larger than the line width of 0.44 you can see in my screenshot(and even more than the 0.4 you say is in the profile I sent). Is it intentional or am I reading it incorrectly?

 

dunno, but you're looking at extrusion lines at an angle of 45 deg and at travel moves in 0 respective 90 deg.

The travel moves are 0.622mm or 0.623 and Pythagoras says:  the corresponding line distance is 0,4398mm or 0.4405mm.

Looks perfect.

The material flow is probably wrong (either the material diameter, the flow rating or the extruder steps/mm). That's all.

 

[burtoogle 516]

32 minutes ago, cakeforcat said:

and by looking at the difference in the X direction it comes to a move of 0.623mm which is larger than the line width of 0.44 you can see in my screenshot(and even more than the 0.4 you say is in the profile I sent). Is it intentional or am I reading it incorrectly?

 

Yes, but the skin lines are being printed on the diagonal so the X coordinates of the line ends will be spaced sqrt(2) * line spacing, e.g. 1.414 * 0.44 = 0.622

[burtoogle 516]

Ha! @tinkergnome, we think alike!

[cakeforcat 2]

Thanks now I understand it. So now I know it's a problem with extrusion but only in the skin lines so it might be something with speed, I'll try to lower it and see what happens. (I've also realized the walls of the cube i measured were printed really slowly so that may be the cause they are fine.)

[burtoogle 516]

Even if you have calibrated your extruder using the typical low speed extrude method, the actual achieved extrusion volume can be reduced quite markedly depending on lots of factors such as nozzle size, material, temperature, extrusion rate, extruder mechanical properties, etc. The bottom line is that for a small nozzle size, the achieved extrusion amount reduces markedly with increased extrusion rate. The printer controller I use (Duet) implements a non-linear extrusion function that can compensate for the reduced extrusion amounts at higher extrusion rates. So then if you ask for 10 mm of filament to be extruded that's what you get across a fairly reasonable range of extrusion rates.

[cakeforcat 2]

So now that we came to this, any idea how to fix this if reducing speed won't work(it's already at 25mm/sec)?

[johnse 31]

Unless you are truly needing single walls to be .44mm thick, I think maybe your calibration method is perhaps not useful. Instead, slice an actual object, and check calibration of the macro dimensions and let the slicer and firmware work together to get well printed objects.

 

if you need better accuracy for given parts, settings like a small negative Horizontal expansion might help.