PussyMakerBack to the toolGuide

3D printing a sleeve mould: walls, layers, the core and inhibition

Short answer: print the two halves and the core with three or four walls, low infill (the Mould 3MF asks for 5 %), the flat parting face up, and expect the layer lines to show in the silicone. The core needs a finer finish than the halves, because its surface becomes the canal's. Keep PLA away from heat, and know what stops platinum silicone from curing. Most of this is general FDM advice, not casting advice, and the page says which is which.

What is sourced and what is not. The wall, infill, layer and seam figures are Prusa's general guidance for printing anything. Nobody we read gives a tested setting for a silicone mould or a cast core, so the figures are a starting point and a real test print is the proof. For adults (18+).

Walls and infill

A part's strength comes mostly from its perimeters, the outer walls, and not from what fills the inside. Prusa's rule: two perimeters is the minimum, three is usual, four is for strong parts. [Prusa] Most models print at 10 to 15 % infill and rarely need more than 30 %, because infill mostly holds up the top layers. [Prusa infill]

A mould half is a thick block being squeezed by two rubber bands and filled from the top, so it is a case for the thick end of that range: three or four walls, low infill. Print time drops with infill, and the block is already thick. The core is thinner and hangs from its bar rather than being squeezed, so give it enough walls to hold its own weight and the pull of demoulding: three walls is a reasonable floor. The Mould 3MF from the designer asks Bambu Studio for 5 % sparse infill, below Prusa's usual range: the author's choice, resting on the walls doing the work, and not tested against a broken mould.

Layer height

Layer height sets both print time and how fine the surface is. Thin layers give more detail; taller ones show ridges. Prusa advises against going below 0.10 mm, and its default first layer is 0.20 mm. [Prusa]

We found no source that names a best layer height for a silicone mould, so this page does not invent one. Print a small test piece at two heights, cast into both, and keep whichever surface you can live with. That test costs an afternoon and answers the question for your printer.

The core's finish

The outer mould's surface ends up against the outside of the sleeve. The core's surface ends up against the canal, the part a body actually touches. Print the core fine: 0.12 mm layers, or a resin printer if you have one, so the canal doesn't carry coarse layer lines. That is not only cosmetic. Layer grooves running round a canal are somewhere residue can sit after a wash, which is a hygiene problem, not just a look. Washing a textured canal.

A self-levelling coat, such as XTC-3D, fills fine texture under about 0.3 mm and can smooth an FDM core the same way it smooths the mould halves; keep the coat thin over deeper texture or it blurs the pattern. Cure and test as below before you cast into it.

Layer lines in the cast

The silicone copies the surface it cures against, so the ridges of an FDM print come out on the sleeve, inside and out. Sinterit's case study found the layer structure shows in the cast and that a PVA coat only partly compensates; it concludes FDM suits prototype casts rather than fine detail. [Sinterit] Formlabs, which sells resin printers, says the same and says resin is smoother. [Formlabs] Both sell competing methods, so read the size of the effect as theirs.

If the lines bother you, a filler coat fills them: Smooth-On sells XTC-3D for exactly that, applied to the print and sanded smooth after it cures. [Smooth-On] Test that it does not inhibit your silicone before you use it, because platinum cure is fussy about what it touches; see cure inhibition below.

The author's method: sand, then UV resin. Sand the inside of each half until the layer lines are gone. Then spread a thin, flat coat of UV-hardening resin over the sanded surface and cure it under UV light. The coat fills what the sanding left, and the inside comes out very smooth and waterproof. Keep the coat thin so it does not blur the shape or the pin fit. This comes from the author's own casting, not from a listed source. One caution that is sourced: an undercured resin surface can stop platinum silicone from curing, so cure it fully and run the test patch on the finished coat. [FAQ 210] In the guide.

Cure inhibition

Platinum silicone refuses to cure, and stays sticky, against some things it touches while curing. The chemistry side is on the silicone page; this is the printing side of it.

Ironing the parting faces

Ironing is a second, slow pass over a flat top surface with the nozzle barely extruding, which smooths it. It adds print time, can cause heat creep with PLA, and does little on curved or sloped surfaces. [Prusa]

That makes it a fit for one job here: the parting face. It is the flat surface where the two halves meet, and a seal is only as good as the two faces. The halves print with that face up, so ironing reaches it. The designer's 3MF does not ask for it; switch ironing on for top surfaces in your slicer if you want it, and check it before you print.

The seam

Every perimeter has to start and end somewhere, so every FDM print has a vertical seam. On FDM it cannot be removed, only moved and hidden. Prusa's options are Nearest, Aligned, Rear and Random; Aligned and Rear line the seam up in one place. [Prusa] On a mould, put it away from the cavity: a seam inside the hollow is a ridge on the sleeve. On the core, put it away from the canal surface for the same reason.

PLA and heat

PLA softens above 60 °C. [Prusa] A PLA mould, or a PLA core, left in a hot car, on a radiator or in a warm oven will soften and warp, and a warped mould does not close. We found no source that says a PLA mould survives a heat post-cure of the silicone, so keep PLA out of the oven, and let the cure run at room temperature. Some silicones are slower and some refuse to cure when it is cold, so read the datasheet.

We opened no page on PETG, so this page says nothing about it beyond that it is another material to test. The guide keeps the same line.

Pins, bands and fit

The two halves are held by conical pins on one face and matching sockets in the other. The pins are 4 mm in radius at the root, tapering to 2.5 mm at the tip, and each has a 1 mm hole down its middle, so the slicer gives the core its own walls; thin pins snapped off in the author's prints. The halves are also held by rubber bands in the grooves round the outside. In this tool the sockets are cut 0.3 mm looser than the pins, and 0.5 mm deeper, so a print that comes out a little fat still closes. Those are working defaults, not tested figures: a real print-and-cast run will replace them. If the halves will not close on your printer, that clearance is the number to change. The core's feet and its tip spigot sit in their own sockets the same way, cut loose enough to seat without forcing. See what the mould looks like.

Sources

Read on 2026-09-20. Pages change; check the current one. Manufacturer pages: Prusa Research, Smooth-On, Dow, Polytek. Case studies: Sinterit and Formlabs, both sellers of competing print methods. The full list, with what each page says and where sources disagree, is in docs/GUIDE-SOURCES.md in the project, section 7, and the claims we could not confirm are listed there too.