I'd like to print a cylinder, 50 mm diameter, 200 mm long, with 1.5 mm diameter holes tightly fit like this:

3D rendering of a cylinder with holes

The holes go all the way through from top to bottom.

I am using a Prusa i3 MK3S. With 0.1 mm detail and 20 % infill, printing one cylinder is going to take 5 days and 5 hours. I need at least five cylinders.

  • Is the above setting appropriate for this job?
  • Is there any way I can reduce the printing time?
  • 1
    $\begingroup$ What's it going to be used for? It might be easier to get some (glass?) tube with a bore diameter of 1.5 mm and 3-d print caps for both ends to hold them. $\endgroup$ Commented Dec 20, 2021 at 20:51
  • $\begingroup$ Why not print it on its side? $\endgroup$ Commented Dec 21, 2021 at 2:09
  • $\begingroup$ @R..GitHubSTOPHELPINGICE, would it be possible to print it on its side and still get good circular holes? $\endgroup$
    – adipro
    Commented Jan 6, 2022 at 1:54
  • $\begingroup$ Try a smaller test print and see. There will be discrete stepping in the Z direction but it should be consistent & smooth if that works for you. And you could print the holes slightly undersized and clean them up with a tool to make them more circular. $\endgroup$ Commented Jan 6, 2022 at 14:17

3 Answers 3


That's a hell of a print!

You are printing a model that has a highly complex structure there, with about 650ish holes, assuming there is space for about 2 perimeters between each hole.

imitated design of a cylinder with ca 650 thin holes

Taking my standard 0.3 mm layer height and 0.4 mm nozzle using a 0.45 mm wide line, I sliced a 10mm high slice of the model for a first estimate of the expected print time - and came out with 2:21 hours. That means the expected print time with 0.3 mm layer height is in the area of 47 hours - or just about 2 days.

As a result, 5 days and 5 hours are in the order I'd expect from a 0.1 mm layer height print for the same nozzle, in fact, your settings seem to have a faster print speed than I do work with.

In general, I don't think this model is good for FDM printing at all, due to many non-fully formed lines inside the model (yellow) and the red perimeters being a very dense pattern.

Sliced cross-section: in yellow: many extra thin lines!


Print faster

You might get a faster speed with a high-flow solution, for example, using a long melt zone (volcano-style) or an even higher flow core-heating 3DSolex nozzle. The latter originally only comes in 0.6 mm and up, also mandating fewer holes, but in late 2021 CHT nozzles in 0.4 mm came to the market. This could drop print time some, but it'd still be a several days print job.

Reduced pattern

Besides increasing layer height to drop printing time by the same factor, reducing the number of holes and as a result, spacing them further apart not only can increase the print speed but also make the print form more reliable.

Another alternative would be to alter the pattern from a circle to a hexagonal pattern: by using hexagons, the resulting pattern does not contain thin walls and might print much faster - depending on hole size, you might experience a drop by a factor of 2!

enter image description here enter image description here

Printed differently

Going from FDM printing to a system such as SLS might be faster and more reliable while SLA/DLP would make this print not only fast and reliable but also trivial - if one can get a 200 mm high SLA/DLP printer, all three models should be able to be printed in one go at the same time!

  • $\begingroup$ Changing to hex pattern is a good idea. One way to get efficient tool paths from the slicer is to model the cylinder solid, then in the slicer, set infill to hexagon pattern, and set top and bottom shell/solid layers to zero. The infill is exposed. $\endgroup$ Commented Dec 20, 2021 at 20:32
  • $\begingroup$ The complexity of the design suggests that movement speed, not extrusion speed, is the limiting factor. I wouldn't expect any speedup from a high-flow printhead, and possibly even a slowdown because of the need to drive the increased mass. $\endgroup$
    – Mark
    Commented Dec 20, 2021 at 22:25
  • $\begingroup$ @Mark High flow allows higher movement speeds if you have better paths. A core heating nozzle weighs about 2 grams more than a standard nozzle. $\endgroup$
    – Trish
    Commented Dec 20, 2021 at 22:44
  • $\begingroup$ BondTech now sells CHT nozzles in a wider variety than just 0.6mm; I've got a 0.8mm and 1.4mm on hand. $\endgroup$
    – William
    Commented Dec 20, 2021 at 23:01
  • 1
    $\begingroup$ @Trish: I meant if the original design produced high travel. Though using the experimental Arachne engine of Cura (upcoming Cura 5) could avoid that with its variable line widths. $\endgroup$ Commented Dec 21, 2021 at 14:31

Late to the party, but for other people that stumble here with a similar problem i've seen a lot of people use a 0 top layers, 0 bottom layers settings to expose infill structure. That with a suitable infill pattern might give you a different slicing that might be faster to print.


Holes, circles, and curves are slow to print, and get slower as the number of wall segments increase.

Ideally you'd use square holes, but hexagonal-sided ones would be a fair compromise too.

Your slicer should be able to give a breakdown on the times for each kind of line in the part. Check and see how much of the time is spent on "moves".

Is there a reason for using all these little holes? Fewer, larger holes will provide better airflow and require less printing, up to the ultimate of one single large hole, where you're essentially printing a pipe/tube.

Do you need the holes to go right-through? Printing a big "cup" with holes across the floor, and separately a "lid" then gluing them together would be quicker and you have one large hole in the middle. If this idea works, you might prefer to just buy plastic pipe of a suitable size, and print only endcaps.

Additionally, if you're printing 5 cylinders because you need a metre total length, then printing two endcaps for a 1 metre piece of plastic PVC pipe should take a couple of hours at most. And it will save you a lot of filament.

  • $\begingroup$ that pattern actually is pretty common for filters or catalysts, so I assume there is a reason for it. $\endgroup$
    – Trish
    Commented Dec 20, 2021 at 14:22
  • $\begingroup$ @Trish plausibly there is a reason for the design as-is, but OP wanted ideas to reduce print time, and not printing 90% of it would help that immensely. $\endgroup$
    – Criggie
    Commented Dec 20, 2021 at 18:19
  • 1
    $\begingroup$ The design screams "laminar flow" and is very likely intentional and necessary. $\endgroup$ Commented Dec 21, 2021 at 14:32
  • $\begingroup$ @R..GitHubSTOPHELPINGICE then I would explore printing two end caps, sized to fit on tubes that are already available, like straws. At a ballpark, 1.5mm holes represent under 25% of the cross section. There will be very-limited air flow through such small holes. This kinda feels like an XY question. where the question is about the intended answer, not the underlying goal. $\endgroup$
    – Criggie
    Commented Dec 21, 2021 at 21:32

You must log in to answer this question.

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