A problem with cube corners... but not the usual one
Posted by animoose
|
A problem with cube corners... but not the usual one January 04, 2017 12:07AM |
Registered: 8 years ago Posts: 367 |
I've been having a problem with the corners on a 10mm test cube. My printer (an Eclips3D coreXY) has worked reliably for a while and this problem has appeared in just the last few days. The common problem people see with cube corners is that they stick out a bit. I have two other problems: on two of the corners, there is a slight inset in the side of the wall just before the corner (where "before" means the part that was printed just before reaching the corner). And one of the other corners is rounded rather than sharp. The two pictures below show this. Sorry they are so out of focus, my phone camera was having a bad day.
Some things that I have tried which do no affect the result:
- material. Happens with both nGen and PLA.
- layer time. Happens at both 5s min time per layer and 30s.
- layer height. Happens at both 0.3mm and 0.2mm.
- position on the print bed.
- angle of the object. Same result if I turn it by 45 degrees.
- same with Cura and Slic3r.
- same with retraction on or off.
- same with increased and decreased extrusion multiplier.
- same after increasing motor current, increasing and decreasing jerk (actually junction deviation, as I use Smoothieware).
The position and size of the inset part are consistent, and the overall cube dimension measured across the middle of the sides is 10mm to within 0.1mm.
I don't think I've made any significant mechanical changes recently. I did tighten up a few screws, but none of them were really loose. The fact that the pattern doesn't change with the angle of the object makes me thing a mechanical issue is unlikely.
If anyone has any ideas on what the problem count be, or ways I can diagnose it, I'd love to hear them.
[Edit: images were missing, now added.]
Edited 1 time(s). Last edit at 01/04/2017 10:32AM by animoose.
See my blog at [moosteria.blogspot.com].
Some things that I have tried which do no affect the result:
- material. Happens with both nGen and PLA.
- layer time. Happens at both 5s min time per layer and 30s.
- layer height. Happens at both 0.3mm and 0.2mm.
- position on the print bed.
- angle of the object. Same result if I turn it by 45 degrees.
- same with Cura and Slic3r.
- same with retraction on or off.
- same with increased and decreased extrusion multiplier.
- same after increasing motor current, increasing and decreasing jerk (actually junction deviation, as I use Smoothieware).
The position and size of the inset part are consistent, and the overall cube dimension measured across the middle of the sides is 10mm to within 0.1mm.
I don't think I've made any significant mechanical changes recently. I did tighten up a few screws, but none of them were really loose. The fact that the pattern doesn't change with the angle of the object makes me thing a mechanical issue is unlikely.
If anyone has any ideas on what the problem count be, or ways I can diagnose it, I'd love to hear them.
[Edit: images were missing, now added.]
Edited 1 time(s). Last edit at 01/04/2017 10:32AM by animoose.
See my blog at [moosteria.blogspot.com].
|
Re: A problem with cube corners... but not the usual one January 04, 2017 03:10PM |
Registered: 8 years ago Posts: 367 |
In case it makes it clearer, here is a sketch of what the prints look like from above.
See my blog at [moosteria.blogspot.com].
See my blog at [moosteria.blogspot.com].
|
Re: A problem with cube corners... but not the usual one January 05, 2017 07:46AM |
Registered: 8 years ago Posts: 413 |
Lets see if we can get some interest in this one.
It does seem an odd issue and you have done some good checks.
As you say it doesn't seem to be mechanical.
Have you looked at the Gcode to check that is correct?
Since the issue seems to be related to corners, which is actually motor speed changes then any acceleration control parameters may have an effect. Might be worth stating these.
Might be good to show all the slicer settings as the picture seems to show a small outer perimeter and a massive infill. So how many perimeters are you printing and what is the infill width multiplier.
Can you print a 20mm cube to see if the problem scales up or even goes away?
It does seem an odd issue and you have done some good checks.
As you say it doesn't seem to be mechanical.
Have you looked at the Gcode to check that is correct?
Since the issue seems to be related to corners, which is actually motor speed changes then any acceleration control parameters may have an effect. Might be worth stating these.
Might be good to show all the slicer settings as the picture seems to show a small outer perimeter and a massive infill. So how many perimeters are you printing and what is the infill width multiplier.
Can you print a 20mm cube to see if the problem scales up or even goes away?
|
Re: A problem with cube corners... but not the usual one January 05, 2017 04:01PM |
Registered: 8 years ago Posts: 367 |
The Gcode looks good when examined in a viewer. I was thinking it might be acceleration, but the setting I use for this are the same ones I have used for a long time. I did upgrade to a new version of Smoothieware about a month back, and the new config had a lower value for the motor current than I was using before, but I've put this back to how it was. Besides, I am 99% sure I've had good prints of similar models since moving to this config. Similarly, the slicer settings are the same ones I've used for a long time. When I get a change, I'll add them to this topic.
I have one other idea about something that might be happening. One some of the test cubes, the top layer is weak or has holes. So I wonder if there could be an extrusion issue. This might cause the problems on the sides if not enough filament gets through for a segment, and so it get stretch round the corner, causing the rounding, or causing it to cave in a bit for the indented parts. I am not convinced by this for several reasons. I checked the extruder calibration with the usual trick of marking 50mm on the filament and making sure 50mm is extruded, and also by getting the value reported in Repetier for how much filament is needed and confirming that it's the amount actually consumed (it was, to within about 2mm in 228mm). The rounding and indenting is very consistent from layer to layer as well.
The suggestion of a 20mm cube is a good one. I also though of printing something which is say 50mm x 10mm, to see if there are differences between the longer side (where the printer has more time to get up to speed) and the shorter one.
See my blog at [moosteria.blogspot.com].
I have one other idea about something that might be happening. One some of the test cubes, the top layer is weak or has holes. So I wonder if there could be an extrusion issue. This might cause the problems on the sides if not enough filament gets through for a segment, and so it get stretch round the corner, causing the rounding, or causing it to cave in a bit for the indented parts. I am not convinced by this for several reasons. I checked the extruder calibration with the usual trick of marking 50mm on the filament and making sure 50mm is extruded, and also by getting the value reported in Repetier for how much filament is needed and confirming that it's the amount actually consumed (it was, to within about 2mm in 228mm). The rounding and indenting is very consistent from layer to layer as well.
The suggestion of a 20mm cube is a good one. I also though of printing something which is say 50mm x 10mm, to see if there are differences between the longer side (where the printer has more time to get up to speed) and the shorter one.
See my blog at [moosteria.blogspot.com].
Sorry, only registered users may post in this forum.