Page 1 of 2

Extruder calibration and fill problem

Posted: Thu Jul 26, 2018 6:43 pm
by pandelume
Hi everyone,

I'm running into a problem which I haven't experienced before and which has me a bit baffled. After measuring actual filament diameter and calibrating the extruder by printing a single outline box and calculating the extrusion multiplier (rinse and repeat until specified width matches extruded wall thickness), I get the correct single wall thickness, but significant underfill on top and bottom layers. The pictures below should illustrate my situation (bottom fill only).


These are the results after calibration of single wall width. I usually calibrate single wall width and then print the box with a single bottom (and sometimes top, with ~30% honeycomb infill) layer to make sure the fill is behaving.

I'm using simplify3d 4.0.1. So far I have only come close to closing these gaps by increasing the extrusion multiplier to ~1.25, which gives me a single wall thickness of ~.65mm using a .35mm nozzle. Picture below shows the bottom fill of a test box - most of the bottom layer broke off since it adhered so well to the bed, but you can see the fill density with the higher extrusion multiplier.


I'm aiming for .45mm to .5mm single wall thickness, but so far haven't been able to get there and get the top and bottom fill to cooperate. Honeycomb infill doesn't seem to be affected.

I seem to be doing something wrong, but I can't put my finger on it. Maybe I should note that I'm assuming the extrusion multiplier or other slicer settings are the root of this problem, but I'm open to other suggestions.

In any case, any help is greatly appreciated!

Re: Extruder calibration and fill problem

Posted: Thu Jul 26, 2018 7:39 pm
by ednisley
pandelume wrote:
Thu Jul 26, 2018 6:43 pm
I'm assuming the extrusion multiplier or ...
Verily, it is written: Hell hath no fury like that of an unjustified assumption.

How long has it been since you've measured / set the initial Z Offset? Most likely, the nozzle starts 0.2-ish mm too far from the platform, thereby starving the first layer.

The cube's measured height should be exactly equal to the model's height, assuming the height is a multiple of the layer thickness. If it's off, tweak the Z Offset to get the correct height before doing anything else.

When the cube height is OK, then measure & adjust the filament diameter and Extrusion Multiplier to make the thread / wall width come out right.

Now the printer and slicer agree about:
  • Where they start
  • How much plastic they extrude
Which should solve the bottom infill problem without doing anything else.

I'd be astonished to get a solid top infill with a single layer: the threads emerge in mid-air with nothing to compress their bottom side, so they're round and too skinny. I've found three top layers to be a workable minimum: one provides a floor, two evens out the slumps, and the third comes out Good Enough.

Measure and adjust!

Re: Extruder calibration and fill problem

Posted: Thu Jul 26, 2018 8:02 pm
by airscapes
Ed taught me!
Check the depth of the groves created by the filament drive. Retract you filament and examining the tooth marks with a magnifier glass or loop.
Ed, do you have a link to thread with the pictures of correct depth. should be about 15% of the thickness. If too loose you will get what you are experiencing.
Nozzle could be clogged as well.. if you have an extra, replace it before you set starting height.

In S3D set extrusion width to .4 ,Layer height to .2 and first layer to 90%, then both thin wall options to allow thin wall (advance tab) then print the attached stl file.
It should measure 2mm tall and .4mm thick. Too tall your bed is too far away, too short your are too close, to fat less multiplier too thin more.
I know you said you did this but the numbers you used seemed wrong for a single wall.
This method works perfectly on the M2.. Had the drive tension too loose the other day and got similar to your problem.. supports falling over.. under extrusion.. etc.
If you need the process to set starting height without using the app let me know and I will post the instructions MG support gave me.

NOTE the one attachment is showing up 2 times for some reason
Single wall 2mm tall .4 wall thickness
(1.64 KiB) Downloaded 57 times
Once you get the above correct test first and second layer with this file, should have no spaces between extrusions.
(1.64 KiB) Downloaded 57 times

Re: Extruder calibration and fill problem

Posted: Thu Jul 26, 2018 8:11 pm
by pandelume
Hi Ed,

Thanks for the reply. I just checked the cube heights I have here with me, and the sparse filled one is coming in at 5.2mm and the high extrusion multiplier one is coming in at 5.1mm for a nominal 5mm height. They're coming in high, so you're probably right about starting too far from the bed - I'll adjust the z-offset today and see if I get a better bottom fill layer.

I normally use at least 3 top layers for just the reasons you've stated. What prompted me to recalibrate extrusion in this case was sparse fill on the top (3rd) layer of something I had printed. For speed and diagnostic purposes I only print 1 top and bottom layer on the first test cubes, but I'm not expecting perfect single layer top fill over ~30-35% infill. By looking at the gaps between filaments as they pass over the infill I get a good idea of how well the top fill is filling.

What I'll do tonight is print this cube again with 1 bottom layer and 3 top fill layers over 30% infill, which will better show any differences in top and bottom filament widths. I will then adjust z offset as you suggest until the printed height is correct, recalibrate extrusion multiplier and see where we are.


Hi airscapes,

Thanks for the suggestions. I'll check the drive gear tension and the nozzle before doing anything else. Those are two good points which I haven't looked at since I last changed filament spools.

I'll also give your .stl a try and report back my results.


Thanks again,

Re: Extruder calibration and fill problem

Posted: Thu Jul 26, 2018 8:39 pm
by ednisley
airscapes wrote:
Thu Jul 26, 2018 8:02 pm
pictures of correct depth
The old (coarse, pre-2016-ish) and new (fine, post-2016-ish) gear patterns: ... culations/


I'm going to be an outlier on this, too, because the stock M2 extruder adjustment sets a fixed distance between the idler roller and the drive gear. IMO, it should have a spring to apply a fixed force on the filament, so (small) diameter variations don't make any difference.

So I added a spring to the (more or less) stock setup: ... feed-gear/


And ever since then, I haven't changed a thing. Admittedly, eSun PETG suffices for my widgets, so there's not much variation going on.

An indentation 15% of 1.75 mm is only 0.26 mm deep and typical filament has a +/-0.05 mm tolerance, so normal variation soaks up nearly half of the indentation. The side wall area varies dramatically with the indentation depth, which means skinnier filament strips out much much more easily than you'd expect.

Re: Extruder calibration and fill problem

Posted: Thu Jul 26, 2018 9:13 pm
by airscapes
I tried the hinged spring drive but didn't have much luck.. Made a dial indicator for the OEM screw setup which once set correctly is nice as you can back it off for soft and the go back to where you were without issues. The virtues of spring loaded make sense but the constant pressure for hard firmament is not going to be good for soft.. like you say, if you don't change filament no worries..

Re: Extruder calibration and fill problem

Posted: Thu Jul 26, 2018 10:00 pm
by ednisley
airscapes wrote:
Thu Jul 26, 2018 9:13 pm
the constant pressure for hard firmament is not going to be good for soft
A constant idler-to-gear distance produces an equal indentation for a given diameter of any filament, but it can't compensate for expected variations around that.

With a spring-loaded idler, you must figure out what pressure produces the proper indentation / grip on each filament type, then the spring applies (nearly) that force regardless of diameter variations.

It ain't perfect, but it's a step in the right direction ... [grin]

Re: Extruder calibration and fill problem

Posted: Mon Jul 30, 2018 7:04 pm
by pandelume
Hi everyone! Sorry for the delay in getting back to this - I didn't get a chance to really sit down and run through all of the suggestions, but I was able to do a couple of them. Unfortunately, they did not completely resolve the issues I'm having.

First, drive tension: I checked the indentations from the drive gear, and they measure something like 20-25% of total filament diameter. This is a little tricky to measure since the filament gets deformed during extrusion, but I think it's safe to say it's not too loose.

Second, z-offset calibration: I tweaked this until the thin-wall box comes out right at nominal height, and I then recalculated the extrusion multiplier, etc. Bottom layer fill improved, but isn't perfect (in other words, the bottom fill strands still have small gaps between them). Top layer (3rd top solid layer) is still not complete - I'd say it's worse than the bottom fill layer for a box printed with 35% hex infill and 3 solid top layers.

Unfortunately, I can't upload any pictures at the moment; I'll add them later tonight.

I'll also update once I have a chance to try a few other things, such as airscapes' STL and procedure. In the meantime, any suggestions are welcome.


Re: Extruder calibration and fill problem

Posted: Thu Aug 02, 2018 8:02 pm
by pandelume
Time for some updates to this and some clarifying pictures.

I was able to run through all the suggestions so far. These have been helpful and have resolved a lot of the issue - but not quite all of it.

This is the part that caused me to stop and recalibrate. It has sparse fill on the top and bottom, as previously discussed. After initial recalibration I got the test prints in the OP.


After running through Ed's suggestions I got some improvement in both top and bottom surfaces. I didn't take pictures of the thin-wall box because there's not much to see, but height and extrusion thickness were on the marks of 5mm height and .45mm thickness with first layer height of .97% and .93 extrusion multiplier (I think - I didn't write down the settings or measurements, although I should have). This is with 3 bottom layers and 3 top layers over 35% hex infill.

In the picture above, a bit of the bottom surface adhered so well to the PEI bed that it broke off the internal layer (right hand corner of the box). This suggests to me that the internal layers are too sparse.

After running airscapes' test print and process I got height and thickness for the 26mm x 26mm x 2mm thin box of 2.06mm height and .406mm thickness with first layer height of 90% and extrusion multiplier of (I think - I don't have it in front of me) .91. No pictures of that since there's no fill. By this point I was tired of boxes, so I printed a spare part to see how the fill turned out:


[bottom picture in next post due to attachment limit]

Re: Extruder calibration and fill problem

Posted: Thu Aug 02, 2018 8:11 pm
by pandelume

Bottom fill is further improved - probably because of the reduction in first layer height - but not perfect. You can feel the discontinuities between extruded bottom filaments with your fingernail. Top fill is still not so great, with clear gaps between filaments.

At this point I'm not sure what I'm missing. It seems like the fill lines are just spaced too far apart to form a continuous surface. In general this probably isn't a big deal, but it definitely weakens the top and bottom surfaces to the point where surface failures (like the adhesion issue in the test box in the previous post) become a concern.