I recently became curious about the Line Width setting in Cura and why one might change it if they aren't using different size nozzle.

Since I've gotten my Ender 3, I've always kept the line width equal to my nozzle size (0.4 mm). I've looked around a bit, and it seems like most people actually set their line widths to be higher, depending upon who you ask anywhere from 120 - 150 % nozzle diameter.

Why is this? They mention that it helps with print adhesion, but why? Shouldn't a 0.4 mm nozzle create a line of plastic 0.4 mm wide, necessitating a line spacing of 0.4 mm?


4 Answers 4


There are several things at play that can make a wider line nice to have:

First layer adhesion

Due to some filaments having serious struggle to get the first line or layer stuck to the bed, it can be an easy fix to just increase the line width, generating a bigger Adhesive Force $F_a\propto A(l,w)$, where A is the area covered by the line, and thus simply $A=l*w$ with length l and width w of the line. So, a wider line means better initial adhesion and can lead to less failed prints in layer 1.

Plastic Goo

Plastics under heat behave in certain ways: they turn into a gooey substance that expands. This is also the reason why prints shrink a little as they cool. Now, if we press the plastic onto the bed with more force (as we force more plastic through than before to go from 0.4 mm to 0.5 mm) for the first time, we have a roughly flat area. The extra filament will make a wider line. The slicher can account for that, and does.

Now, next layer up: Where does the extra material go now? Plastic goo has one property that is very interesting: it tries to shrink its surface as much as possible. Heat a short piece with an airgun and it gets a little beady. But on the other hand, it comes hot enough from the nozzle to melt a tiny surface area of the already built layers, which is how layer bonding works in the first place. But our goopy plastic finds the layer below not exactly flat like the first layer found its lower surface, it finds a shape of ridges and valley. Taking into account that it wants to have the least surface to non-plastic (=air) and slightly cross bonds with the print, it will fill these nooks and crevices inside the print a tiny little better, as the increased force we use to push it out also increased the speed at which it expands to them: we reduce the time a tiny bit to reach there. How does it matter?

a thermal image of a 3D print

Well, heat transfer bases, roughly speaking, on a formula like this: $Q = mc\Delta T$ Q is the thermal energy of the object, m the mass of the object, c its specific heat capacity and T the temperature, ΔT the temperature change. But we don't have a homogenous object, we got pretty much a heat distribution with touching zones of different heat. The actual formula for the heat transfer inside the object is a long mess containing stuff like the gradient $\text{grad}T$, thermal conductivities, and integrals, but what matters is the result: The faster-expanding line of filament loses a little less thermal energy to its surroundings than the less forceful extruded line, which can increase the bonding between the two as the temperature on several fronts:

  • it enters the crevices further before reverting from goo to solid, leading to better adhesion for more surface.
  • it contains more thermal energy that can and will get transmitted to the layer below and has a bigger surface area, so it can increase the zone thickness that gets remelted a tiny bit, increasing the layer bonding strength a little.

This can result in a problem though: if you don't give the printed lines enough time to cool, it can lead to the material to accumulate heat more and more, leading to the whole thing to melt and turn into goop. An easy fix to this side problem is minimum layer time. But that would be only tangential to the original question, so look for example at the question here or the video the thermal picture above is taken from here.

  • 1
    $\begingroup$ Thank you for the incredibly detailed answer, @Trish. This is more explanation than I could have ever hoped for. $\endgroup$
    – ifconfig
    Sep 21, 2018 at 14:29
  • $\begingroup$ Very nice answer. I always set extrusion width to 110% of nozzle diameter, even when printing clays and gels. $\endgroup$
    – Davo
    Sep 25, 2018 at 11:23

Shouldn't a 0.4 mm nozzle create a line of plastic 0.4 mm wide

Not necessarily. Due to a phenomenon known as die swell extruding plastic through a 0.4 mm nozzle, the line of plastic that is created is actually slightly wider. Pressure inside the extruder compresses the plastic slightly, and it expands again as it exists the nozzle.

They mention that it helps with print adhesion, but why?

When you extrude a thicker line of plastic than your nozzle diameter, the "excess" plastic is compressed by the nozzle and forced out to the side. This pushes the plastic into the layer below, increasing adhesion. You can compare this to taking a hot glue gun, pressing the tip into the surface and squeezing the trigger, versus lifting the glue gun above the surface and letting the glue drip onto the surface. Doing the former creates much stronger adhesion.

As a side effect, using thicker lines makes it easier to get the first layer to stick, since the thicker line has more surface area to adhere to.

  • 1
    $\begingroup$ @TomvanderZanden A question that pops up is what would be better, increase the line width or the extrusion multiplier, but that would be a whole new question I guess? $\endgroup$
    – 0scar
    Sep 21, 2018 at 9:47
  • $\begingroup$ @0scar yeah, that does come to mind too. $\endgroup$
    – ifconfig
    Sep 21, 2018 at 14:26
  • 2
    $\begingroup$ @ifconfig For that reason I have asked: "Slicer line width vs. extrusion multiplier for layer adhesion?" $\endgroup$
    – 0scar
    Sep 21, 2018 at 14:30
  • $\begingroup$ Ultimaker Cura defaults the extrusion with of my UM3E to 0.35 mm for a 0.4 mm nozzle. So, although people tend to increase line width, the software decreases it. $\endgroup$
    – 0scar
    Aug 9, 2019 at 5:05
  • $\begingroup$ @0scar That is news to me. Simplify3D does the opposite. It would be interesting to learn why Ultimaker does this. $\endgroup$ Aug 9, 2019 at 6:15

There are several yet unmentioned considerations:

Given the same movement speed thicker lines fill up a layer quicker, because more volume is extruded per second. In some systems the extrusion flow is the limiting factor for speed, but around corners the print head needs to slow down. Thicker lines = less lines = less corners = less slow down = higher print speed.

Thicker lines have less detail, though. A line of 0.6 mm cannot represent details smaller than that, so smaller line widths capture the input geometry better. Also corners will get rounded by the same distance, so thicker lines = rounder corners.

Thicker lines create worse overhang. Thicker lines require more pressure from the nozzle and if the layer below is (partly) missing the back pressure from the previous layer is less, which results in overextension, which will then also more likely go downward instead of to the sides.

The higher pressure can force lines into small crevices of the layer below, though. This was highlighted by Trish already.

The model Cura uses for a single line is rectangular, while in actuality the printed lines are rounded on the sides. This makes the width of the full extend from side to side larger than computed, at the cost of the corners of the rectangular model. This means that the line width setting should be set slightly smaller than what you would want the lines to end up like.

  • $\begingroup$ Is there a reference for Ultimaker Cura using a rectangular line model? This seems highly unusual, e.g. Slic3r uses a rectangle and 2 half circles, for Ultimaker Cura a similar model is expected. $\endgroup$
    – 0scar
    Aug 9, 2019 at 4:58
  • 3
    $\begingroup$ You can quote me on it. I wrote the larger part of CuraEngine. $\endgroup$ Aug 9, 2019 at 8:55
  • $\begingroup$ Nice, I'll take your word for it, but please link to the CuraEngine manual if you can. Thanks for the information, much appreciated! $\endgroup$
    – 0scar
    Aug 9, 2019 at 16:06
  • $\begingroup$ It's not in the Cura manual (yet). Is it in the slic3r manual? $\endgroup$ Aug 9, 2019 at 16:30

I'll give a short answer here: It's the volume. The the nozzle redistributes the volume of the plastic into a different shape. i.e. the nozzle is turning a cylinder of 0.4 mm diameter into a rectangle of the same volume, which a function of the layer height / volume = line width.

  • $\begingroup$ it isn't really a rectangle but a flattened object consisting of 2 half circles of diameter h and a rectangle of short length h in between them. The line width is this things width. But you do not explain why it is conventional at all $\endgroup$
    – Trish
    Feb 26, 2019 at 16:09
  • $\begingroup$ Why is what conventional at all? $\endgroup$
    – user77232
    Feb 26, 2019 at 18:15
  • $\begingroup$ OP said: "I've looked around a bit, and it seems like most people actually set their line widths to be higher, depending upon who you ask anywhere from 120 - 150 % nozzle diameter." $\endgroup$
    – Trish
    Feb 26, 2019 at 19:17
  • $\begingroup$ Yes, and then the OP asked why that is so. Previous answers didn't address the volumetric concern. Ever poke a hole in a paint can and then run a line with it? how come the width of the painted line isn't equal to the diameter of the hole? $\endgroup$
    – user77232
    Feb 27, 2019 at 1:16
  • 1
    $\begingroup$ Hmmm, it seems like you could be on to something here as you are approaching the question from a different angle from everyone else... if you could tidy up your answer and include the comment about the paint can [along with an explanation/diagram of what happens - no I have never done that but I'd be interested in seeing what happens :-) (I guess the line is width than the diameter?)], I'd give you an up vote. $\endgroup$
    – Greenonline
    Mar 3, 2019 at 16:41

You must log in to answer this question.

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