Welcome! Log In Create A New Profile

Advanced

Strange outstanding perimeters

Posted by smoothcriminal 
Strange outstanding perimeters
March 30, 2015 05:07PM
Hi!

Since a few weeks I own an Ultimaker 2.
After some relatively good prints I encountered some problems with Cura on negative angles.
After some reading in the Internet I recently switched to Slic3r.
The prints are much better but I got some new problems.
In the attached image you can see the marked outstanding perimeters.
I figured out, that they only appear if the printer prints some infill, before and after each hole of the die.
I played with all parameters in Slic3r but didn't got it to work properly.

Do you have some advices?

Thanks and best regards
Falk
Attachments:
open | download - IMG_4820a.jpg (497.3 KB)
Re: Strange outstanding perimeters
March 31, 2015 03:37AM
Hi,

can you attach the original .stl model ?

Frank
Re: Strange outstanding perimeters
March 31, 2015 03:59AM
Hi,

Here it is.

Falk
Attachments:
open | download - Wuerfel.stl (162.5 KB)
Re: Strange outstanding perimeters
March 31, 2015 09:19AM
The marks correspond to the layers where horizontal holes start or end, and are probably where Slic3r has changed the direction in which the perimeter was printed. The mis-registration is most likely due to backlash in the printer's mechanics, which will cause the position of the nozzle to differ depending on what direction it travelled to reach a particular point. The usual culprit is loose belts.

Dave
Re: Strange outstanding perimeters
March 31, 2015 10:33AM
Hi Dave,

thank you for your explanation, but I can confirm, that the outstanding layer is not depending on the direction.
It comes after each fast solid infill.
But what speed I ever set for the different options, the result stays the same.

Falk

Edit:

See attachment.
Slic3r makes huge "solid" areas (2) for printing only little "caves" (1) in the perimeters.
Which setting affects this areas?
After each of this areas the perimeters are a little bigger.

Falk

Edited 1 time(s). Last edit at 03/31/2015 12:12PM by smoothcriminal.
Attachments:
open | download - IMG_4822a.jpg (500.8 KB)
Re: Strange outstanding perimeters
March 31, 2015 02:56PM
I suggest you look at the G code, locate one of the layers in question, and see whether the perimeter is printed in the same place as the previous layer (X & Y positions for the edges the same). I would be very surprised if it were not, and I still believe the fault is likely to be with mechanical positioning error initiated by different directions or speeds. Another possibility is that Bowden extruders in particular exhibit lag, and so the flow can be too fast or too slow at any given point while the nozzle is accelerating or decelerating.

The solid infill will be done after the perimeters (unless you have set infill to print first), so it is difficult to see how it could affect the perimeter. The area of solid infill may look excessive, but when it is building up to a horizontal hole, the infill must ensure that there are the correct number of solid bottom layers for the hole channel as it will be in a few layers time (when it is wider), plus extra distance to cover the perimeters of the hole sidewall.

Dave
Re: Strange outstanding perimeters
April 01, 2015 06:23AM
Hi Falk,

indeed it is a interaction of many slicing and firmware settings.

For this model CURA is the better choice, because the higher slicing resolution and more precise extrusion calculation, also depending on model resolution.

A worse model/slicing resolution causes acceleration/deceleration peaks.
For a smooth surface it is important to have a steady trajectory generation.

You can counteract with lowering your acceleration settings.
This causes less pressure peaks in the extruder.
Of course this extends printing time, but gives much "softer" moves.

Frank.
Re: Strange outstanding perimeters
April 01, 2015 07:33AM
Quote
NitroXpress
Hi Falk,

indeed it is a interaction of many slicing and firmware settings.

For this model CURA is the better choice, because the higher slicing resolution and more precise extrusion calculation, also depending on model resolution.

A worse model/slicing resolution causes acceleration/deceleration peaks.
For a smooth surface it is important to have a steady trajectory generation.

You can counteract with lowering your acceleration settings.
This causes less pressure peaks in the extruder.
Of course this extends printing time, but gives much "softer" moves.

Frank.

From my experience, Cura has a lower slicing resolution. Which is what you need if your printer cannot handle lots of direction changes, because the lower resolution slices (e.g.) a curved perimeter into fewer segments and so the printer has fewer direction changes. It won't however help when the perimeter is a straight line that changes in length from one layer to the next.

Decreasing the acceleration is the preferred method, but the easiest cure-all for such printing artefacts when they are only problematic on a small percentage of prints is to print more slowly. You can usually get away with slowing down just the outer perimeter, but may need to slow all the perimeters.

Dave

Edited 1 time(s). Last edit at 04/01/2015 07:35AM by dmould.
Re: Strange outstanding perimeters
April 01, 2015 07:52AM
The outstanding perimeters are an example of overextrusion. After the prinder foes a "fast solid infill" certain pressure is built in the melting chamber of the nozzle. This is slowly released, along with the predetermined extrusion for the perimeters of the next layers and hence a bit more plastic is extruded. The phenomenon is more visible in case of Bowden printers. Some measures to reduce the effect:
- reduce the infill speed - the best (at least from this perspective) would be to maintain the same speed over the entire print
- make sure the external perimeters are not printed first
- increase the number o perimeters (to let the overextrusion inside the object, though the extra amount of plastic will anyway push the external perimeter out)
Re: Strange outstanding perimeters
April 01, 2015 09:45AM
Yes, decreasing speed is often a solution for many printing problems...
But it's not the "cure"...

It's important to have the perfect interaction between printhead moves and extruder flow.

For example look at the attached picture.
I made a thinwalled tube 25mm diameter, 100mm high.
Wall thickness is 0,2mm.
Every 10mm of Z-high i increased the velocity about 10mm/sec.
Starting with the "standard velocity" of 60mm/sec at the left side, and ending with 150mm/sec at the right side.

As you can see, 150mm/sec gives perfect surface quality.
With 60mm/sec you can see the created segments from the slicer.

Frank
Attachments:
open | download - tube (399.6 KB)
Re: Strange outstanding perimeters
April 01, 2015 10:40AM
Quote
NitroXpress
As you can see, 150mm/sec gives perfect surface quality.
With 60mm/sec you can see the created segments from the slicer.
Frank

What you are seeing is the effect of mechanical smoothing that effectively rounds off any corners due to inertia not allowing instant changes in direction. The surface that was printed fast is smoother which in this particular case is what you want, but it is in fact less accurate because it does not faithfully reproduce the input - which contained segments. Do the same when printing a polygon (e.g a hexagonal nut) and you will see the same effect - except in that case the segmented section is what you want (a polygon) and the smooth section (more circular) is what you do not want! Incidentally, the segments are more likely to have been created by the CAD application than the slicer - an STL contains only straight lines and so must represent circles as multiple straight segments, so even if the CAD application works with real circles (most do not), those circles will have to be converted to polygons when exporting as an STL.

Dave
Re: Strange outstanding perimeters
April 01, 2015 10:54AM
Yes, you are right.

This is mainly a effect of mechanic and trajectory generation.
At a lower speed the mechanic and the trajectory generation can move the head more accurate at the original path.

Therefore i asked about the minimum segment length in my other thread...

Frank

Edited 1 time(s). Last edit at 04/01/2015 10:55AM by NitroXpress.
Re: Strange outstanding perimeters
April 01, 2015 11:03AM
Hi!

Thank you for the many suggestions, I will go on trying winking smiley

Best regards
Falk
Sorry, only registered users may post in this forum.

Click here to login