Why 3D Prints Have Gaps Between the Walls

Hold the part up to a window and there it is: a hairline of daylight running the length of a 2mm wall, or a dark seam between the outer perimeter and the one behind it, or a dull line under the top surface where the infill never quite reached the edge. Gaps between the walls are the one FDM defect most people attack with the wrong tool, because the reflex is to raise flow or add infill and the cause is usually neither. It is arithmetic. A slicer fills a wall with a whole number of extruded lines, and when your wall thickness is not a whole number of lines wide, the remainder has to go somewhere. Work out what your lines are actually worth and most of these gaps close in one reslice. For the parts where they will not close, the provider directory lists shops running processes that have no walls to divide in the first place.

This post contains affiliate links. If you purchase through these links, 3D Prototyping Hub may earn a small commission at no extra cost to you.

This post also contains Amazon affiliate links. As an Amazon Associate we earn from qualifying purchases.

Find Out Which Gap You Have Before Changing Anything

Four distinct faults get reported as "gaps in the walls", and the fixes have nothing in common. Where the void sits tells you which one you are looking at.

The quickest discriminator takes one glance at the first layer. If the solid areas are dense and only certain features show voids, it is geometry. If the solid areas are sparse too, it is flow. Geometry faults are repeatable in the same place on every single layer, which is the other tell — a flow fault wanders.

The Arithmetic That Decides It

A slicer does not draw a wall. It draws lines of a fixed width, side by side, until the next one will not fit.

That width is set by your profile, not by the nozzle, although it is usually close to it: a 0.4mm nozzle is typically run somewhere between 0.42mm and 0.45mm. Everything below assumes 0.42mm, and you should substitute your own number — measure a single line on a printed part with a pair of digital calipers rather than trusting the setting, because a worn nozzle prints wider than the profile claims.

Now divide. A wall of thickness T takes as many whole lines of width W as fit, and whatever is left over is the problem:

Look at which numbers fail. 1.0mm, 1.5mm and 2.0mm — the three wall thicknesses every CAD user types by habit, because they are round in millimetres. None of them is round in extrusion widths. That is the whole reason this defect is so common on otherwise well-tuned machines.

The One Setting That Fixes Most of This

Before redesigning anything, check which perimeter generator your slicer is using.

The older approach — called Classic in PrusaSlicer and OrcaSlicer — places every perimeter at the same fixed width and abandons whatever is left. The newer one, Arachne, varies the width of individual lines so they fill the available thickness exactly: that 1.5mm wall becomes three lines of 0.5mm instead of three of 0.42mm plus a void. It has been the default in Cura since version 5 and is available in current PrusaSlicer, OrcaSlicer and Bambu Studio, but an inherited profile or an older install can easily still be on Classic.

Switching it on costs nothing and closes the majority of centre-of-wall gaps on the spot. Two things to know about it. It can only stretch a line so far before the result looks over-extruded, so a wall at 1.2 times a line width still has to compromise. And it interacts with the minimum-feature-size and minimum-wall-width settings nearby, which decide whether a very thin rib is printed as one widened line or dropped from the model entirely — if small details have been vanishing, that pair is where to look.

Where the Wall Meets the Infill

The second gap is a different interface, and infill density is not the lever.

Sparse infill lines meet the innermost perimeter at an angle, touch it at a point, and bond over a tiny area. Raising density from 15 to 40 percent brings the lines closer together but does not make any one of them grip the wall better — you pay double the print time for the same dark seam. The two settings that matter are infill overlap, expressed as a percentage of extrusion width and usually worth pushing to 15–25 percent, and the option called connect infill lines to perimeters or infill anchor length, which turns each infill line to run along the wall for a few millimetres before heading back. That second one is what converts a point contact into a weld.

Two consequences worth knowing. Overlap pushed much past 30 percent starts bulging the outer wall outward, which shows up as a dimensional error rather than a gap — if your parts have gone slightly oversized since you turned it up, see why prints come out the wrong size. And a wall-to-infill gap directly beneath a top surface is also what leaves the top layers with nothing to bridge onto, which is the pillowing failure arriving from the other direction.

Gap Fill Is a Patch, and It Has a Cost

Every modern slicer has a gap-fill function that squeezes a thin bead of plastic into leftover space. Leave it on — but understand that it is covering for the arithmetic rather than solving it.

Gap fill moves at a different speed and volume from the lines around it, so on a long wall it often reads as a slightly rough, beaded stripe and can leave blobs and zits where it starts and stops. There is usually a threshold setting that filters out gap fill below some width; raising it trades a visible void for a cleaner surface, and lowering it does the reverse. Neither is a fix. A wall designed as a whole number of lines needs no gap fill at all, prints faster, and comes out stronger than one that is 80 percent perimeters and 20 percent improvised bead.

When the Nozzle Is the Right Thing to Change

If your parts have a house style — enclosures with 2.4mm walls, brackets with 3mm ribs — it is often easier to change the nozzle than every model.

A 0.6mm nozzle run at 0.6mm lines divides exactly into 1.2mm, 1.8mm, 2.4mm and 3.0mm, which covers most functional geometry, and it prints those walls in half as many passes. The trade is resolution: it cannot place two lines into anything under about 1.2mm, so fine ribs and lettering either print as a single fat line or disappear. In the other direction, a 0.2mm nozzle makes genuinely thin walls printable at the cost of roughly double the time per unit area. The full comparison by diameter is in our guide to nozzle sizes.

Wear belongs in this section too, because it is invisible and it moves the number everything else depends on. A brass nozzle printing carbon-fibre or glow-in-the-dark filament opens out measurably inside a couple of kilograms, so walls that were exact multiples quietly stop being exact multiples. Hardened steel nozzles solve it for abrasives, and for everything else a cheap assortment pack of brass 0.4mm nozzles treated as consumables is less effort than diagnosing the drift. Machines that make this easy are the ones with a standard nozzle thread and published spares — ELEGOO's FDM range and Anycubic's US store both list hotend and nozzle spares next to the printers, which is the detail worth checking before you buy rather than after.

If It Is Flow, None of the Above Applies

One honest caveat on everything so far: if the voids are everywhere, stop reading this page.

Short flow produces gaps that look similar and respond to none of these settings. The usual causes are a partial clog, moisture in the filament, a worn nozzle or an extruder that slips under load, and the giveaway is that solid bottom layers are affected along with the walls. Cleaning filament is the cheapest thing to rule out first, and the full diagnostic order is in why 3D prints under-extrude.

Design So the Slicer Has Nothing to Improvise

Four habits remove this defect from future parts rather than fixing it on this one.

• Dimension walls in extrusion widths, not round millimetres. Work out your width once, then use its multiples as your standard thicknesses. Rounding up beats rounding down: a wall slightly over a whole number prints as a mildly squished solid, while one slightly under loses an entire perimeter.
• Give anything structural at least two full lines. A single-line wall has no interior bond at all, which is why it both leaks and snaps — see why printed parts leak.
• Do not taper a wall to an edge. A draft angle that runs the wall from 2mm down to 0.3mm passes through every bad value on the way, so the gap opens and closes up the height of the part.
• Keep ribs and bosses on the same grid. The broader set of these rules is collected in our design for 3D printing guidelines.

Fix It in This Order

1. Check whether the solid layers are clean. If not, it is flow — go to the under-extrusion guide instead.
2. Switch the perimeter generator to Arachne. Reslice and look at the preview.
3. Measure one real extruded line with calipers and write the number down.
4. Round the model's wall thicknesses to multiples of that number.
5. Raise infill overlap to 15–25 percent and turn on connect-infill-to-perimeters.
6. Only then consider a different nozzle diameter — and recheck step 4 afterwards, because the multiple has changed.

When to Hand It Over

Three cases are not worth another evening.

A pressure-tight or watertight part is the clearest: a hairline void running the full height of a wall is a leak path, and no amount of surface finishing reaches it. A load-bearing part with thin walls is the second, because the gap sits exactly where bending stress peaks and acts as a stress riser on top of the lost cross-section — the mechanism is the same one behind brackets that sag and parts that snap along layer lines. The third is a batch, where tuning one part is fine and reproducing it twenty times is not. SLS, MJF and resin processes build solid cross-sections with no perimeters to divide, so the fault does not exist for them; the providers in our directory are listed by location and process, and sending the same file out for a quote costs nothing but tells you quickly whether the evening is worth spending.

Most wall gaps, though, are one setting and one piece of arithmetic. Find out what your lines are really worth, make your walls a whole number of them, and the slicer stops having anything to improvise with.

Hero photo by Inés Álvarez Fdez via Unsplash.