For a science project, I'm 3D-printing some custom pipes and tubes to regulate the flow of gas (a combination of ethyl alcohol and water vapor) through an apparatus. They need to be pretty small, as the entire experiment is designed on a small scale. I'd also like the use a little filament as possible.

How thin can I make the walls of these pipes and tubes before either they collapse or gas leaks out? I know that's possibly an engineering issue, but I'd also need to take resolution into consideration. Ideally, the pipes would be about two centimeters in diameter, possibly a little larger or a little smaller in some parts. I'd most likely use ABS, but PLA is my backup in case there's some unforeseen reaction between the gas and the pipes.

The printer I'm using is an FDM printer, a version of the MakerBot Replicator.

  • $\begingroup$ What type of printer are we talking about? I suspect from the context (ABS and PLA) it's an FDM machine, but your question doesn't rule out a hypothetical PLA laser sintering machine. $\endgroup$ Commented Jan 13, 2016 at 21:30
  • $\begingroup$ @TomvanderZanden Ah, yes, it is an FDM machine. I'll specify that. $\endgroup$
    – HDE 226868
    Commented Jan 13, 2016 at 21:32
  • $\begingroup$ This might be a question that'd be better answered on the Engineering SO site, where you're more likely to find users with the specific expertise needed to answer a question like this. $\endgroup$
    – Jeff
    Commented Jan 13, 2016 at 21:36
  • 1
    $\begingroup$ @Jeff Good point. I made it a bit clearer that I'm also concerned about printer resolution, but primarily the strength of the filament. $\endgroup$
    – HDE 226868
    Commented Jan 13, 2016 at 21:46
  • $\begingroup$ @Jeff I don't think that's the case. Thin walled cylinders present some unique issues with 3D printing. $\endgroup$ Commented Jan 13, 2016 at 21:57

4 Answers 4


The thinnest wall your printer can print is determined by its nozzle size, and will be a little thicker than that nozzle size.

A great challenge when dealing with thin, hollow cylinders is that the cross-section has very little surface area and it can delaminate easily, especially if the tube is long.

You could try printing the tube with a very thick extrusion with, and using only a single perimeter. That would give better gas-tightness and layer adhesion than two, thinner perimeters, but it may turn out too fragile for your application. In that case, you'll need to print additional perimeters. Sticking to thicker extrusion widths would still be beneficial.

At a two centimeter diameter I'd say the single perimeter has a decent chance of working if you handle them gingerly.


The thinnest possible will be a single outline, so the best you can possibly do is slightly larger than your nozzle size.

The biggest issues I see in terms of making a thin airtight tubes are:

  • Complex geometry: anywhere you have overhangs you may need to increase the wall thickness to ensure a sealed overlap. Bridging will need extra layers to ensure good layer adhesion and any supports will have to be removed carefully to not damage the surface.
  • Perimeter start/end: The joint where the perimeter starts/ends not only has to seal to the layer below but it has to seal with the now cooling start point. This could be especially problematic if all of your layers are started at the same point (which visually produces a vertical seam) because you don't have a smooth perimeter on the previous layer and any imperfections in the joint are compounded.


  • Print on end where possible: You'll get a higher quality outer perimeter if your layers are full rings instead of printing up both sides of a pipe on its side.
  • If you can manage a single perimeter when printing on end try using the "spiral vase" setting, which will smoothly raise the z position and print as a spiral instead of printing discrete layers. This removes any issues related to sealing the start/end of a perimeter.
  • If you're printing with multiple perimeters increment the wall thickness in multiples of the nozzle size to ensure the perimeters are sealed together. Having the perimeters not bound together increases the chance of leaks and reduces stiffness.
  • Do the pipes have to have the same thickness all along them? You can probably reduce thickness to save material for simple geometry (straight sections) if you change thickness gently.

When printing thin cylinders vertically,

  • Slow print speed, to avoid rocking the model.
  • no shells, just one nozzle thick, so you don't need fill. Fill can easily rock the print.
  • Print at least 2 at a time at the smallest layer thickness you can do. This causes the hot end to have to move away from one cylinder to get to the next. This gives each cylinder layer some extra time too cool and harden before the next layer goes on. I generally print .1 or .05 for cylinders.
  • Your slicer may have a setting that tells it to wait between layers so the plastic has more time to cool. I think the setting is called ORBIT on mine.

Consider following up with an acetone vapor bath to help seal the pores in the print.


This is difficult to give a definitive answer, but I have some thoughts.

For a vertical wall, the minimum thickness is determined by your nozzle size, typically 0.4mm. If you are able to print entirely in a vertical orientation (a pipe standing on end), you may consider using "vase mode" with no top or bottom to print. This can help simplify the CAD modeling, as you don't have to model the hollow interior and wall thickness.

For a horizontal wall, in theory this could be a single layer height, about 0.1mm depending on the printer and settings. Realistically, you will probably need 2 to 4 layers to get something "solid". However, this is not likely to have good results for a hollow pipe due to performance of the printer in bridging and overhangs. A thin, round pipe laying on its side is not likely to be successfully printable.

For printing in a horizontal orientation, I'd recommend a non-circular shape; the specific shape would depend on your overall design and goals. A "teardrop" or "flat teardrop shape can help with horizontal holes through parts.

Also, rather than individual pipes, perhaps you could design a block with hollow passages. This could allow thicker walls, make it much stronger, reduce assembly needs, allow for extremely complex routing and internal connections, and may even allow for functional parts like valves or sensors to be installed.

Finally, I would not expect an FDM-printed part to be airtight. There will most likely be some small gaps between layers and where layers start and stop. As an example, I tried making an ice mold, but almost all the water seeped through before it froze.

Another poster recommended an acetone vapor bath, which should help seal any pores. This only works with ABS. Another idea might be to coat the inner surfaces of the pipes with a resin that can seal everything up.


You must log in to answer this question.

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