first 15 or so layers not good

I am printing a bowl. I have 4 layers on the bottom and the perimeters are 2 layers at about 45-55°.
The first 15 or so layers print rough and with holes in them. After that they are perfect. nothing changes except the height from the bed.
no bed heat
PLA
Temp is set to 205 1st layer 200 rest.
Speed is slow - about 40mm/sec
0.4mm nozzle
0.2mm layer height
0.3mm first layer
200% layer width
slic3r ver 1.3.0

the designed horiz thickness of the wall is 0.8mm. the radial thickness is less than that.
the sliced model display shows some lighter shading in those layers.


So why does it go from bad to good?
Better still, how do I make it all good?

Edited 2 time(s). Last edit at 12/14/2017 04:10PM by ruggb.

I suspect that the overhang angle is too great on the bottom 15 layers, and the overhang angle reduces after that. Overhang angles steeper than about 45 degrees generally need to be printed with support.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

good thought if it were the case.
here is a side view
the angle is the same - bottom 1/2 is bad, top 1/2 is good.

It looks like the bottom is warped, so there is over extrusion in the lower layers?
Did you try using a brim to eliminate warping?

Edited 1 time(s). Last edit at 12/15/2017 04:22AM by o_lampe.

For the holes, how much skirt are you printing to purge and charge the nozzle?

For the roughness, it kind of looks like the pattern I get when my Z=0 is too low and the bottom solid layer perimeters get squeezed almost exclusively outward.

on those layers it sounds like the printer is doing strange moves instead of just a smooth rotation.
I can actually feel it stepping around what should be a smooth curve.
It actually sounds like the printer is following a bad slicer output. I don't hear it on the good layers.


When I first printed this with thicker walls I had the same anomaly and these layers were rough - but since it was thicker, there were no holes in it.
I have printed a couple with smooth walls all the way.

@IMBoring25
the bottom is 4 solid layers, so it is pretty well purged by the time it gets there. the bottom is smooth as silk, so Z0 is set pretty well I think.
It really appears to be a bad slicer output, but I have no idea what to do about it or how to figure it.

@o lampe
With the size of the base there shouldn't be any issue with warping if it sticks - and it isn't warped and is definitely stuck.

I have a couple of thoughts - the slicer is set to print external perimeters first - that probably isn't good since there is only 2 and it is on an angle.
But there isn't visible slope change between bad and good.
I also have a different accel for 1st layer, but that is far from a first layer. It just increases the commands and probably isn't needed.

You happen to have bed mesh leveling taper off in 15 layers?

I increased the temp, as upon very close inspection it appeared that the bonding was not very good in that section.
It looks a lot better, BUT, 15 layers are still being printed with stepping like head movement which produces horiz lines. At 20 layers it becomes smooth.
You can barely see the lines in this pic at about 1 - 2 o'clock. They are more obvious with local vision. At 3,6,9,&12 o'clock for the X & Y tangents it is smooth for about 15° - IE when moving mostly X or Y. This is a CoreXY so that would be both motors moving about equally.

This stepping movement AND the lower temp combined to produce the holes and very rough surface. Now it is just the movement issue. OBVIOUSLY, this movement is in the GCode from the slicer but why??
What about the layers 5-20 is causing the slicer to produce movement like that? The bottom is 4 layers thick and runs smooth as silk.

Try a different slicer?

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
o_lampe
You happen to have bed mesh leveling taper off in 15 layers?

I don't use any bed leveling

Why do you suspect the slicer? Might be the .stl file is rendered in low resolution? Or the firmware is not segmentation free?

Quote
o_lampe
Why do you suspect the slicer? Might be the .stl file is rendered in low resolution? Or the firmware is not segmentation free?

It sounds to me that either the slicer or the program that exported the STL model used a lower resolution when approximating the circles at the bottom than it did with the slightly larger circles 15mm higher up. The OP could look at the GCode to confirm that the problem lies in the GCode (or refute it).

However, it could also be caused by the XY jerk values being set too low in your firmware. The OP hasn't said which firmware his printer uses.

Edited 1 time(s). Last edit at 12/17/2017 05:12AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

It was not clear from the original picture it's hollow. New theory... Your cooling settings are slowing down the layers starting where the base "disc" goes away and as the diameter increases the larger diameter naturally takes longer and speeds the print back up. It might be that something in moving more slowly is causing you problems, either minimum step rate or the resolution at which you can change step rate.

@o lampe
The stl is generated from Sketchup. Why would it affect one section? But then I would expect to see it in the gcode
I am using Marlin 1.1.x bug fix build - don't know what is meant by "segmentation free"

@dc42
the gcode looks the same on layer 12 and layer 25 so I don't know how it could be the slicer either. I have had Marlin do things so strange I had to power down reset everything to fix it.

@IMBoring25

my cooling settings are off for layer 1&2 and auto always on for the rest. It has to be less than 30 sec layer time to slow down. I didn't time it but one layer is 2 revolutions of about 250-300mm and it does it at 40 or 60mm/sec

The anomaly is also not consistently all the way around. Most always it is on the back side. With the increase in temp, the stepping went all round except for the x & y tangent areas.

I guess I will create a test print to see if I can make it happen consistently and change Marlin versions.

Thanks for your inputs

An .stl file always cuts circles into segments. The more segments, the smoother it will come out.

You could try to scale the part to eg. 150%. Then you see:
If the segments are the same length, it's firmware or slicer.
If there's same amount of segments, it's the .stl file.

that whole structure is symmetrical and all the segments are the same proportion. They can't be all the same size as the diameter changes and there are 120 segments all the way around. I created it at x1000.
I printed a single layer (2 perimeters) model - ie without the wall being a 2 sided entity with fill, and it printed perfectly.
Slic3r definitely has issues if the model is not to its liking. I had another problem where instead of slicing with long bridging it created 3 sections (or more) of bridging, one being at right angles to what it should have done and parallel to the supports. Turns out the supports I built in were not evenly spaced and it upset Slic3r. When I redid it to space it out evenly Slic3r was happy and bridged it all in one direction.