Over Filling Inner Perimeter

Seeing extrusion of too much infill filament on thin walls resulting in layer to layer build up and print fails.

Running 1.2.9a stable. config.ini attached. Settings:
LH=0.1mm
Nozzle =0.4mm
default layer widths ext periph=0.40; inner periph=0.68; infill = 1.28mm

Maybe I have something wrong in my config.ini so any help would be apprecated.

Version 1.1.7 stable produces very different results that appear to be correct. (need to revert and kick off a test print)

Are there incompatibilties between the 1.2.9a and 1.1.7 ini files?

Test case attached is a 10mm long x 2.625mm wide x 5mm high rectangular block. Failure is filament packing across the 2.625mm span.

following was extracted from the GCode starting at line 427 (file attached/code snippet below)
spreadsheet attached

                                                                                 extruded vol/len  rect section             
        x0        y0       e0       x1      y1      e1      length (mm) vol mm^3 sect area mm^2     area mm^2    width    layer height    
perim   114.281   150.594  1        105.719 150.594 1.23444 8.56200     0.5638    0.0658              0.068      0.68     0.1    
perim   105.719   149.406  1.26697  114.281 149.406 1.50141 8.56200     0.5638    0.0658              0.068      0.68     0.1    
infill  113.952   150      1.90355  106.048 150     2.34343 7.90400     1.0580    0.1338              0.128      1.28     0.1

Available cross sectional area between centers of two inner perimeters = 0.1 * (150.594-149.406) = 0.1188

compared to extruded filament volume of 2* ½ cross section area of inner perimeter + area of infill = (2 *0.5 *0.0658 + 0.1338) = 0.1997208654

So there are two anomalies
1) infill cross sectional area for 1.28mm wide infill has a cross sectional area of 0.1338 which is greater than retangular cross section of 0.128 for a fully filled 1.28mm x 0.1 infill
2)Trying to stuff filament with a 0.199 cross sectional area into a space of 0.118 mm^2.



Layer GCode extract (starting around line 427)
G1 E1.34343 F1800.00000 ; retract
G92 E0 ; reset extrusion distance
G1 Z1.500 F7800.000 ; move to next layer (13)
G1 X114.281 Y150.594 F7800.000 ; move to first perimeter point
G1 E1.00000 F1800.00000 ; unretract
G1 X105.719 Y150.594 E1.23444 F180.000 ; perimeter

G1 X105.719 Y149.406 E1.26697 ; perimeter
G1 X114.281 Y149.406 E1.50141 ; perimeter

G1 X114.281 Y150.534 E1.53229 ; perimeter
G1 X114.800 Y151.112 F7800.000 ; move to first perimeter point
G1 X105.200 Y151.112 E1.68338 F180.000 ; perimeter
G1 X105.200 Y148.887 E1.71839 ; perimeter
G1 X114.800 Y148.887 E1.86948 ; perimeter
G1 X114.800 Y151.052 E1.90355 ; perimeter
G1 X114.454 Y150.912 F7800.000 ; move inwards before travel

G1 X113.952 Y150.000 F7800.000 ; move to first infill point
G1 X106.048 Y150.000 E2.34343 F180.000 ; infill

This is a photo of the orginal print before extracting the test case.

The long section has a 2.625mm wall thickness and shows of the overflow build up after 35 layers.

The extruder was bouncing from bump to bump. Not happy.

This is a commonly reported issue with the latest version.
There is at least one open issue on github. I can't give a link now because I'm on my phone, but you can google slic3r github.
I have not seen a work around other than to use the previous stable version.
This area also had a recent discussion about it, you can check that. Was a thread about cookie cutters.

---

My printer: Raptosaur - Large Format Delta - [www.paulwanamaker.wordpress.com]
Can you answer questions about Calibration, Printing issues, Mechanics? Write it up and improve the Wiki!

Here are the issues on Github:

Gapfill problem with thin walls #2960 This is the main thread, very good match to what you are seeing.

Infill overextrusion #3082 Later thread.

Multiple users are indicating that 1.2.7 may be a good alternative.

Paul,

Many thanks for digging up the info and 1.2.7 alternative. I'll give it a try!

I tried 1.1.7 stable and it had under flow issues.

I'll check out [github.com] in the future. How did 1.2.9a get to be a stable release with this kind of issue? (hmmmm)

There are a number of GCode viewers out there. Wish there was one that showed colorized filament overlap per layer. Range of colors from not touching, to overfilled. Do you know of any? Would make it easy to view slicing tool issues.

Edited 1 time(s). Last edit at 12/08/2015 09:53PM by to-the-nth.

Quote
to-the-nth
Wish there was one that showed colorized filament overlap per layer. Range of colors from not touching, to overfilled. Do you know of any? Would make it easy to view slicing tool issues.

That's a very interesting idea. I don't know of any viewers like that.

For my research I've been concentrating on showing changes in mm3/sec and certain kinds of new properties, so I've written a viewer that does that in a couple of ways. See the bottom of this lengthy thread.

I can calculate the extrusion widths, but in order to determine overlap there is a tricky part I've not got a handle on: how to identify the nearby segments involved, especially segments that may not be exactly parallel.