/ Hatches, Paikea Refit, Working with Composites / By

Building Composite Hatches: From 3D-Printed Mould to Finished Part

Every hatch on this boat starts the same way — not with fibreglass, but with a mould nobody will ever see once the part comes out of it.

Why 3D-printed moulds, and why not always

Building tooling for a job like this runs against a genuine principle: unless a mould gets used for multiple parts, it’s effectively a single-use piece of plastic — wasteful, avoidable where possible, and worth building around rather than accepting as a default. Composite work that can be done without dedicated tooling gets done that way whenever it genuinely can be.

Hatches, though, are exactly the case where a mould earns its place — eight hatches across this boat alone, with complicated, repeated geometry that would be nearly impossible to reproduce consistently by hand each time.

Two moulds, printed in pieces

Building the hatch system meant creating two separate 3D-printed moulds — one for the gutter and frame, one for the lid itself, which carries far more geometry than it looks like it should from the outside. The seal gutter built into the lid mould exists for exactly the reason covered in the design post: containing and improving the seal’s performance, rather than leaving it as an unsupported strip of rubber.

Printed on a Bamboo Lab printer small enough to live aboard, the mould’s full size exceeded the printer’s bed, meaning it had to be broken into sections and printed separately — jigsaw-puzzle-jointed together so each piece locates precisely during assembly. That same sectioning solved a second real problem: two different hatch sizes exist on this boat, with the aft cabin hatches needing to run 60mm shorter in one dimension due to a tighter deck opening. Rather than building two entirely separate moulds, one mould section simply drops out to shrink the overall length — one piece of tooling doing the work of two.

Where 3D-printed tooling shows its limits

By the fourth hatch pulled from this mould, some real wear had started showing — a section of the mould’s internal shell had partially collapsed under repeated vacuum loading, a genuine limitation of 3D-printed tooling that solid machined moulds don’t face in the same way. Getting wall thickness and infill density right for a mould that needs to survive multiple vacuum cycles is part experience, part trial and error — there’s no single fixed rule, and it depends heavily on the specific geometry: a large flat span needs meaningfully more internal support than a small detail near an edge.

Surface finish carries its own trade-off. PETG sands well, and one mould here went through a full finishing pass — 150 grit, then 220, with a final 400-grit pass specifically around difficult draw angles — while a second mould was left entirely unsanded, relying instead on generous draft angles to release cleanly without any surface prep at all. Neither approach is universally correct; it depends on how many pulls a given mould needs to survive and how much post-processing the finished part is getting anyway.

Wax, applied deliberately imperfectly

Release strategy varies by mould geometry. Simple shapes get wrapped directly in Teflon film — no wax at all, genuinely simple and reliable for straightforward geometry. More complex shapes rely on traditional release wax instead, applied here in a way that would horrify anyone chasing a true Class A finish: thick, heavy-handed, worked into every seam and dent with a rag or a finger rather than buffed to a thin, even film.

That’s a deliberate choice, not carelessness. Every part coming off these moulds gets sanded, primed, and painted afterward regardless — drain holes cut, hinges fitted, surfaces reworked — so a showroom-perfect release finish would be wasted effort. The priority here is a safe, reliable release every time, not a mirror finish nobody’s going to see once the part is finished and installed.

Plasticine fills the gap between mould and table, creating a dam that stops resin migrating underneath during infusion, and doubles as a way to smooth over the jigsaw-puzzle joints between mould sections — pressed in, roughly levelled, then finished by wiping with the same wax, whose solvent content softens the plasticine just enough to fair it into the surrounding surface without pressing so hard it scoops the fill back out of the joint.

Gel coat: not for looks, for practicality

Gel coat goes onto the more geometrically complex mould for two specific, practical reasons rather than any pursuit of a glossy finish. It provides a white, primed base that makes the eventual painting process considerably easier around complicated geometry, and it tends to release from 3D-printed tooling more reliably than the structural resin underneath — likely down to a small amount of internal shrinkage as the gel coat cures, though that’s observed experience rather than confirmed theory.

Polyester gel coat is compatible only with polyester, vinylester, and other styrenated resin systems — it can technically be forced to work over epoxy, but it’s genuinely troublesome and not recommended without real experience. For an epoxy equivalent, a thin epoxy primer applied to the wax, lightly scuffed once cured, gives similar adhesion for a subsequent epoxy laminate.

Masking tape defines the gel coat’s outer edge before it goes on — peeled away before the gel coat fully cures, which leaves a crisp, precise trim line to work to later. Get the timing wrong, particularly in hot weather when gel coat kicks off fast, and there’s barely enough working time to pull the tape before it sets permanently in place.

Gel coat’s tacky, uncured surface (distinct from flow coat, which contains wax and cures fully) is used deliberately, too — the first layer of fibreglass cloth sticks directly to that tackiness without needing spray adhesive to hold it in position during layup.

Trimming without cooking the resin

Freeing a finished part from the mould reveals genuinely accurate, repeatable geometry — not CNC-machined precision, but close, since the mould itself came off a computer-controlled 3D printer rather than being shaped by hand. Small reference features, printed directly into the mould at just 0.2mm proud of the surface, leave faint trim lines and hinge-location marks on the finished part — a detail only achievable because the tooling itself is digitally designed, something no traditional hand-built mould could replicate as easily.

Trimming the excess laminate uses a jigsaw rather than an angle grinder with a cutoff wheel — a deliberate material choice. Grinding generates real heat, and since this resin has a relatively low heat deflection temperature, that heat softens it quickly enough to smear and burn rather than cut cleanly. A jigsaw keeps the cut cool and controlled.

The sealing blade edge — the critical surface that presses into the foam seal — gets trimmed flush and tapered by hand, a genuinely tedious step that pays for itself the moment the finished hatch actually gets fitted to the boat. A tapered, consistent edge here is what makes the seal work properly later; skipping this step to save time would undermine the entire point of the design covered in the previous post.

In This Series

This post is part of our composite hatches deep dive. Read the full story here: Paikea’s Composite Hatches: The Full Story

Or explore the rest of the series:

Scroll to Top