Top layer rough in corners. Overextrusion?

Hello all,

Before I get to my question, I'd like to give some details on what I have going on. I'm printing on a Makerfront i3 Pro, with a genuine E3D V6 hotend with a Titan extruder in a bowden configuration. Bowden tube is about 400mm long. I'm using Hatchbox black PLA at 195C. Speeds are 30mm/s for all features except first layer, which is 15mm/s. XY acceleration and jerk settings are 1000 and 20, respectively. Print surface is 0.25" cast aluminum plate (mic6) with PEI coating. Using Slic3r for slicing and Pronterface for communication.

My problem, in short, is that the top infill layers seem to be overextruding in the corners of the top layers, as the hotend path gets shorter and shorter. I calibrated my e-steps multiple times, and I'm as close as I can get using a dial caliper. If I print a single wall box with a 0.45 width, I get a measurements of 0.44-0.45mm on all walls. So, in my mind, my extruded is calibrated. I do also notice that some of the overlap areas are slightly rough, possibly reinforcing an overextrusion issue. I have overlap set to 15%. I've measured the filament ad nauseam; it comes out to 1.74mm every time. Extrusion multiplier is currently set to 1.0.

The picture attached shows my part (black) next to another (grey) that was printed on a Prusa i3 MK2.

What am I missing here? Is this overextrusion, or something else?

Forget the single wall box for calibrating E-Steps. Never will get you to your goal.
Do it like this!

Then your perimeters of the hole will be touching each other, and then you can start to hunt the top layer issue.

Hi Dancer,

Thanks for your reply!

I followed your advise and did the calibration according to the link you posted. My e-steps were 417 when I started, and I finished at 365. That gave tiny gaps on the 95% infill layers and no gaps on teh solid. Seems like quite a bit of difference from what E3D recommended for the Titian extruder. How could the difference be that great?

I printed another object as before, and the top layer surface quality is better, but there is still rough parts in the tight corners. Any suggestions for next steps?

EDIT: forgot to add that the holes perimeters look much better now. I changed the "extra length on retraction" setting from -0.5 to 0. That seemed to help greatly.

Thanks for the help!

Edited 1 time(s). Last edit at 10/19/2017 09:25PM by letthewookiewin7.

hmmm....


which version of slic3r are you using? - I think I remember having a similar issue long ago. now I use prusa-version of slic3r: Prusa-Slic3r

EDIT: Maybe with newer Version of Slic3r you have to redo extruder-calibration. There was an calculation-error in an older version of slic3r

Edited 3 time(s). Last edit at 10/23/2017 07:45AM by Dancer.

I'm using 1.2.9. All of the above calibration steps were down while using this version.

I think there is a clue in the difference between what E3D recommends as a starting point and what I'm getting good print quality & surface finish at. 365 is about 13 percent less than the 418.5 that they recommend for the Titan extruder.

How does slic3r relate acceleration & deceleration to the extruder steps? Could the bowden be introducing too much slope in the system, causing the pressure in the head to be greater than expected duing deceleration moves? Seems like a stretch, but I am out of ideas.

Sorry to say it very directly: Trash Slic3r 1.2.9

move on to the Slic3r-Versions from Prusa - it's a fork of the original Slic3r, but much more advanced than original Slic3r will be with in the next years.
1.2.9 had a bad calculation error with infill espacially at gap-fill and thin wall.

I've never beeen able to print really good with 1.2.9 - with Prusa-Versions it's much much better - and even has some really cool features 1.2.9 is still missing.

Well, after a very labor-some move and the holiday season, I finally have my printer set back up and running.

I starting using Slic3r PE. I also upgraded my Marlin firmware to v1.1.8 so I could take advantage of the Linear Advance feature. I am using a (pretty high) K value of 150, but since I have a bowden extruder apparently that's normal. There was an obvious difference during the K factor pattern printing with values from 0 to 200, so at least I know the logic is turned on a working.

Alas, after all that, still plenty of roughness on the top layer.

Any other ideas?

Set the overlap to 0.
What nozle size are you using? I would expect this gets worse as nozle size increases.

I'll try that and report back.

I'm printing with a 0.4mm nozzle. Unfortunately I dont have another size to try.

One other thing that I noticed is that it seems like the infill gets more rough (more over-extrusion) as the z height increases. I printed an object with 6 perimeters so that I could get a thickness measurement on the perimeter itself and it was dimensionally accurate (3.21 to an expected 3.20). But when I measured the same print on the infill, it was 0.2mm larger than it should be (3.4). To me, that suggests it is not a z axis backlash issue, but I don't understand the progressivley worsening over-extrusion.

0.4mm Nozle is typical so don't worry about trying a different one.
In slic3r there is a top layer extusion multiplier. So you can try triming that if its your only issue.
Checking the Z calibration is easy so confirm that is OK. Just measure height of hot end to a reference point then request a z height of say 100mm and re measure. Thie diference should be 100mm.
It is possible that the bowden has an influence. 400mm doesn't seem an unusual length and the speeds seem slow enough to not be an issue. You need to rout the bowden so the bends stay large and consistent at all positions, including Z changes. You dont want an S shape.
Hopefully someone can comment on the accel and jerk settings.