I need help resolving some conflicting logic between the nozzle width and shell thickness--parameters in Cura 16.021--and the physical deposited line width (width on x-y plane).


For clarity, let me define a few terms I'm using:

Nozzle size - nominal diameter of the nozzle / Cura parameter

Nozzle diameter - true diameter of the nozzle

Extrusion diameter - diameter of circular cross-section extrusion upon leaving the nozzle

Deposited line width - width of the rectilinear layer deposited on the build surface in the x-y plane (not the z-layer height).

Shell thickness - Cura parameter for part wall thickness

Cura - Cura version 16.021

Extrusion Diagram PROBLEM 1

Physics suggests that for regular continuous flow out of the nozzle the extrusion diameter (cylindrical material extrusion) would be slightly larger than the nozzle diameter, and the deposited line width (rectangular prismatic extrusion) on the x-y plane should be even wider than the extrusion diameter. I'm assuming the deposited line width is equal to the shell thickness as set in Cura. However, I am finding conflicting advice as to whether one should set the shell thickness slightly greater than or slightly less than the nozzle diameter.

In this post on the Ultimaker forum, the moderator seems to suggest that the nozzle size and deposited line width are one and the same.

In this post on StackExchange, the discussion seems to agree with my understanding.


A commonly referenced procedure for calibrating extrusion suggests printing a 25mm cube with an open top and bottom. My understanding is that by printing four walls of a single shell thickness, we should correct the extrusion rate to achieve a deposited line width equal to nozzle diameter (Cura requires the shell thickness to be a multiple of nozzle diameter) . The problem I have with this is, as stated above, I don't understand a physical basis for targeting nozzle size = shell thickness (deposited line width). I'd expect to calibrate the extrusion to something slightly wider than the nozzle diameter.

enter image description here

My printing experience seems to confirm my intuition. If I reduce extrusion rate to achieve shell thickness = nozzle diameter my prints appear visually under-extruded with poor layer adhesion. Calibrating to about 110% of the nozzle size looks about right.


enter image description here

Although Cura constrains shell thickness to be a multiple of nozzle size, it will not slice thin shapes on the x-y plane equal to nozzle size, they need to be slightly wider. I have verified this multiple times by designing a part in SolidWorks with a 0.4mm wall, exporting it to Cura as an STL, and previewing the slices (using a 0.4 mm nozzle size). If I increase the wall thickness in the model to 0.41mm, Cura will slice the wall. This seems to suggest to me that Cura does account for a deposited line width > nozzle diameter. It calls to question whether the shell thickness is actually a physical dimension for the 3D print or just a parameter for the slicer engine. I have taken it to mean the number of passes around the perimeter in units of nozzle size, similar to other slicers that ask for a shell count.

Thanks in advance to anyone who has put in the time to read this post and kindly share some knowledge.

EDIT 1: My interchanging of terms "width" and "thickness" made it sound like I'm referring to z-layer height. Also, Trish pointed out that "line width" is the better term rather than "layer width" so I have updated this. My post is only concerned with the deposited layer width in the x-y plane--"line width".

  • $\begingroup$ Welcome to 3DPrinting.SE! $\endgroup$ Nov 14, 2018 at 22:10
  • 1
    $\begingroup$ Are you referring to the effect called die swell where the extruded polymer diameter is larger than the nozzle diameter in PROBLEM 1? The deposited layer thickness to be even thicker than the extrusion diameter is not true, layer thickness is not dependent upon the nozzle diameter, it is a separate setting called layer height. However, the maximum layer height is dependent on the nozzle diameter, usually you should not print at layer heights over 75 % the nozzle diameter. $\endgroup$
    – 0scar
    Nov 14, 2018 at 23:18
  • $\begingroup$ Thanks for the response. Yes, I am referring to die swell as the expected mechanism for the cylindrical extrusion to have an increased diameter than the nozzle opening. $\endgroup$
    – ASRICK
    Nov 15, 2018 at 14:05
  • $\begingroup$ I will update my post because I created confusion by interchanging extrusion "width" and "thickness." I am only referring to the deposited layer width in the x-y plane, not the z-axis layer height. I'm interested in accurately printing a defined shell thickness. $\endgroup$
    – ASRICK
    Nov 15, 2018 at 14:16
  • $\begingroup$ what is Layer Width? There is layer height, which is the stepping from one layer to the next. Deposited material has a line width $\endgroup$
    – Trish
    Nov 15, 2018 at 14:25

2 Answers 2


So, I think I may have found a satisfactory answer. Cura 3.6 includes separate parameters for line width and shell count:

This seems to decouple the line width from the specified nozzle size and target what I believe is a more optimal width (~110% the nozzle diameter). It was the other version of Cura that was driving a lot of confusion with the line width being defined by the nozzle size. This also removes the weirdness of not being able slice lines equal to the line width.

Thanks to all who responded.

  • $\begingroup$ Self-found solutions are always great! I hope our persistance has helped some $\endgroup$
    – Trish
    Nov 16, 2018 at 11:22
  • $\begingroup$ The old Cura (14.04.x or derivatives) satisfied the user specified wall thickness by adjusting the line width. The new Cura (2, 3, 4) ignores the user specified wall thickness and just rounds to the nearest number of wall lines it can fit without adjusting their widths. $\endgroup$ Aug 8, 2019 at 20:14

There may be die swell, which will depend on the extrusion force, the material type, and the exact temperature within the hot end. This is because the viscosity is highly dependent on those parameters.

As to line width: again extruded temperature, material, ambient temperature, cooling fan coverage all play into how much "sag" takes place before the material hardens. Oh, and X-Y speed of the nozzle, of course.

Further, for the first layer, if the nozzle is nice and close to the bed, the nozzle itself will force the material to spread laterally while limiting the vertical extent.

Personally, I don't think it's worth the effort to create a mathematical model. Spend the 3 cents' worth of electricity, 0.01 cents' worth of material, and 15 minutes of time to calibrate if desired.

  • $\begingroup$ Thanks. What you are describing is exactly what I'd expect. My conflict probably comes down to what the proper calibration method is. There appears to be a popular recommendation across forums to calibrate the line width of a single shell. But in Cura a single shell is just the nozzle size..but why would I want/expect my extrusion line width to be equal to my nozzle size? I have seen the procedure I linked to recommended many times that suggests to adjust extrusion factor so the line width matches the shell size specified in the slicer (the nozzle diameter for a single shell in Cura) $\endgroup$
    – ASRICK
    Nov 15, 2018 at 17:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .