Odd infill issue

[jens3 11]

I have run into an issue where infill lines in Cura 4.3 do not always go back to the wall surfaces.

I have tried different infill patterns and this issue is present on all of them.

I am printing 1 mm wide and 0.4 mm high at up to 60 mm/sec with a Volcano hot end. I am using a Duet wifi controller and have tried a variety of settings for pressure advance as there is a very long bowden tube.

From all my testing, I am pretty certain that the issue is not related to pressure advance which was my first guess.

What happens is that the infill line is started and there is a distinct and consistent delay between the physical move starting and extrusion happening.

I can confirm that on non-infill lines, extrusion happens the instant that the printing move starts.

The attached picture shows the issue although it might be a bit hard to see. The infill at the right edge of the model has the bottom line meeting the wall but the next line up doesn't meet the wall (overlap set at 50%)

 

I suspect this issue may not be obvious with higher infill (currently 10%) or when line width is a more common 0.4 mm.

Has anyone seen this issue ? Any solutions ?

 

Edited December 22, 2019 by jens3

[burtoogle 516]

Could you please attach some sample gcode that exhibits this issue. Thanks.

[jens3 11]

The gcode (as per cura preview) does not show any issues.

I have done a bunch more testing and I need to follow up on two items - I have a considerable amount of play between the bowden tube and the bowden coupler with the tube sliding in and out during retraction moves and I also need to re-calibrate the pressure advance setting. I will run another test print after this is done and post a follow-up tomorrow.

[jens3 11]

Bowden tube has been replaced, couplers have been replaced, pressure advance has been re-calibrated and speeds have been reduced in case I got into volumetric limits of the Volcano hot end.

On things like walls, I can see extrusion of material the instant the nozzle starts moving. While the 'starting' gap with infill has reduced with all the improvements, there is still a noticeable delay when doing infill between when the nozzle starts moving and when material is extruded.

The nozzle is always moving as it should starting from the wall with the core issue being a delay between movement starting and filament coming out the nozzle.

I am currently at 50% infill overlap and will go to 100% on my next test.

[jens3 11]

Gone to 100% infill overlap, added 1 mm3 of extra prime just for the heck of it and this is the result:

 

In this print, all but one of the infill lines meet the wall. The joint that doesn't meet the wall appears to show the extra prime amount as a bit of a ball at the beginning of the infill line which is well away from the wall. The wall/infill joint right next to it shows the infill barely touching the wall in spite of the setting of 100% overlap

[jens3 11]

The gcode that does layers 4 and 5 ....

layer4and5.gcode

[burtoogle 516]

Thanks for the gcode. Not really enough there to be sure but I'm guessing you are using combing. If so, you need to set the max comb distance with no retract to a value something like 10 or 20. That will stop underextrusion at the start of infill.

[jens3 11]

combing is off. Would you like me to post the entire file?

[burtoogle 516]

What would be most useful is the project file so that it includes the model and settings. Do File -> Save and attach the .3mf file. Thanks.

[jens3 11]

 ... as requested

CFFFP_duet roof molding sample.3mf

[burtoogle 516]

I'm guessing that the infill line that is affected is the first line of the infill to be printed following the outer wall. The outer wall is printed slower than the infill so what could be happening is that the extruder can't accelerate quickly enough to supply the required amount of filament. Does this problem still occur if you use the same print speed for infill as you do for the wall lines that immediately precede it? In your example project that would be the outer wall line.

[jens3 11]

The speeds have already been drastically reduced to avoid just that possibility. I have reduced the infill speed to 10 mm/sec and will do a test print.

[burtoogle 516]

You want to aim to keep the extrusion rate similar between walls and infill. i.e. the product of print speed and line width (aka flow rate) should be the same for all of the different line types (walls, infill, skin).

[jens3 11]

Turns out it is indeed the first move after the wall but reducing the speed by a factor of 3 did not change the outcome by a bit.

[burtoogle 516]

Yes, but is the speed of that infill > the speed of the last wall printed before it?

[jens3 11]

Outer wall speed is 20, inner wall speed is 40, infill speed is 10

[burtoogle 516]

OK, please just try printing with all of the speeds set to 30. Thanks

[jens3 11]

Printed with 30 mm/sec all around.

Just to clarify, the sequence is :

 

Print inner wall on previous layer

increase layer

print 'defective' infill line on current layer

print a bunch of perfect infill lines

print one infill line intersecting the defective line that just barely touches the wall layer (once it gets printed)

Print outer wall

print inner wall

repeat.

 

With the new settings, the 'defective' line now juuuust touches the wall in one out of 3 layers

Only the infill line immediately next to the defective infill line shows a reduced overlap (maybe 10%), all other infill lines show the requested 100% overlap between infill and wall.

 

Running the file on a simulator show the end points of the infill lines to be between the inner and outer wall, just as would be expected.  There is a definite lag on the defective inlill line for the filament to flow but there is also a noticeable (but not as big) lag on the infill/wall joint next to the dud joint.