Any ideas what causes this detachment of filament at intersections? Steps/mm seems ok, filament diameter ok. Infill lines look a little thin too. I tried over extruding by 20% but no improvement. Could it be a blocked nozzle?
Thanks in advance.
adavidm
The leading edge of the nozzle hits the previous line of filament, and it extrudes at the back end. When crossing over, after it’s passed the line of infill, it’s effectively been wiped off. This causes the filament to be pulled off and then continues a slight distance away from the crossover.
I’ve seen this mostly with kisslicer, and I suspect that in general, it’s a miscalculation in filament flow that causes this.
You’ll notice that the break always happens in the direction of nozzle travel.
Hmm, makes sense. It would also account for the under-size infill. I’m using Cura (latest version), with a completely fresh install, but I’ve also tried the Slic3r engine in RepetierHost with similar results(on rectangular infill). Slic3r’s honeycomb infill doesn’t cross over like Cura so the problem isn’t is bad.
I’m confused as to why over-extruding by 20% doesn’t make a huge difference (good or bad), hence the question about the nozzle. I’d have expected such a big ‘error’ would produce dramatically worse results, but it doesn’t seem to.
I’ll go and double-check steps/mm, filament dia, etc. Any other variables that controls extrusion volume? I’m 100% sure the nozzle is 0.5mm.
thanks for the insight. I’ll report back…
adavidm
Just had another thought. If the extruder is over-extruding, rather than under, might the buildup of excess material exacerbate the issue?
Didn’t really think of that. I’ll try a “under” extruded run this evening and compare.
Honestly, as more people chime in, I expect I’ll learn why this happens as well. (I certainly don’t know everything!) I’ve been noticing it in some prints but haven’t had the time to research why it’s happening. Doesn’t happen on Slic3r for me, but does on KISS.
The gap between perimeters seems a tad large. Perhaps you have set an incorrect nozzle diameter compared to the what extrusion width you are aiming for?
I have had little luck printing these kind of structures with only single perimeter bonding. The surface area seems to me to be too small to get a reliable result as far as my experience goes.
A slic3r config with no top or bottom layers and rectilinear infill with a high extrusion width gives fine results for me. Of course it depends on what you need the print for.
Also, this seems to happen only in one orientation, not perpendicular. So I’d also check for backlash, see if your belts are all tightened and the pulley doesn’t slip on the motor shaft. Things like that. (It might sound strange, but it fixed a similar problem in my case)
The gap between perimeters is another problem I’m facing. I meant to mention it earlier. In cura I believe the default extrusion width is 0.5mm which should be ok. Slic3r produces similar perimeter problems.
I also had the thin infill lines problem at the same time, just as you. In my case it turned out the filament was only 1.68mm instead of 1.75mm. Maybe your case is similar.
@David_Martin Look here http://richrap.blogspot.com/2015/01/slic3r-advanced-perimeter-tuning-3d.html . It is @Richard_Horne 's excellent tutorial on perimeter extrusion width optimization.
I’ve not seen that article before, thanks!
Those do not look like they are straight runs but the start and stop points. If that is the case then it looks like retraction is too high. For those lines running left to right in the picture if it is printing each segment between the vertical lines as separate parts and the retraction is too much then it will not be able to recover after such a short distance (.4 mm in this case?) hence the gap in fill after the retraction and move. Try turning down the retraction by half and test it out again or see if your slicer has a setting for extra extrusion after retraction.
@Brigham_Valdez I never measured my filament either, until I encountered strange problems and finally somebody told me “really, measure your filament” - “ok I will but it won’t… oh 1.68” 
@Brigham_Valdez When you see some fantastic finish prints ask the printer if he does measure. I have found when you are trying for that last 1% it does matter.
That’s great if it works for you. My own experience is that filament from different sources DOES vary as much as 10% and I’ll see it in the finished print.
Jeez, is this guy for real? Turn on your fans man, you need to cool down.
Didn’t mean to offend, or question your experience. I was only offering mine. However, I prefer questions that we have no answer for than answers I can’t question …
Wow, this escalated quickly!
There’s an 80mm radial fan blowing on the part.
Have tried everything from cheap to faberdashery pla, no particular difference. This particular batch is from Ooznest in the UK.
The gaps ARE in straight runs, retraction is not in play here.
I’ve just taken the nozzle apart and cleaned it thoroughly.
Once the silicone on that is set I’ll give it a go. I’m leaning towards a backlash/flexing problem now. Its a delta printer and the arms are not as rigid as I’d like. I’m suspecting this is introducing a vertical movement that causes the nozzle to be ‘wiped’ as described above.
Thanks for all the advice. I’ll be so glad to get this issue nailed.
adavidm
@David_Martin If the detail from your last post was in your first … BV probably has the answer for you.
Yeah, I realise now I could have been more helpful in my first post, apologies to all and all I can say in my defense is that I was writing in a hurry during a break. Mea culpa
Sorry to be dense but which part of BV’s advice do you mean, the filament? I’d be happy if it was but I have tried quite a few different reels from different vendors. Is there a single gold standard I should try to elimate that? I was under the impression faberdashery was very good stuff. Temperature-wise I’ve tried from 185-210.
I’ll look into his other point about steps/mm as I said earlier. After that I’ll recheck the mechanicals and do some more methodical testing.
Cheers
adavidm
