Building Carbon Fibre Chainplates: Part One

Building Carbon Fibre Chainplates: Part One

Some jobs start small and grow into something else entirely. This one started with a bolt swap in Brittany. By the time it caught up with us properly, it had turned into a full chainplate rebuild — and Shayne was on the other side of the Atlantic delivering Gunboats when it happened.

While Shayne was away, the plan had been to sail Paikea south toward Grenada, out ahead of hurricane season. That plan ended abruptly with a discovery that couldn’t be worked around: the chainplates were pulling out of the deck. Not moving slightly, not something to monitor — actually pulling out. That’s the kind of problem that ends a passage plan on the spot, and it’s what set the carbon chainplate project in motion.

The approach this time wasn’t a repair — it was a proper rebuild, engineered from scratch. Rather than just replacing bolts and hoping the existing bulkhead held, the new plan was to laminate carbon fibre chainplates directly into a brand new section of bulkhead, built specifically to receive them, then install that whole assembly as one integrated unit.

Getting there started with templating. Shayne dry-fitted a paper template against the actual bulkhead, matching triangles and circles drawn on the template to reference points on the boat itself — not a quick job, and worth getting right, because that template becomes the reference for exactly where the new chainplate needs to sit once it’s built. Reference lines and heights were carefully marked and cross-checked against the old position, so the new bulkhead section would go in exactly where it needed to be, not just close enough.

The laminate itself came together as a proper vacuum-bagged stack — peel ply, foam core, breather felt, perforated release film, the standard sequence — with small holes worked into the layup specifically to let trapped air escape from the foam during the first layer of glass. A pin sleeve went into the layup too: a stainless steel sleeve grooved and slotted to stop it rotating or working its way out under load, wrapped first in a full layer of glass fibre to insulate it from the carbon and reduce the risk of galvanic corrosion between dissimilar materials over time.

Getting the laminate to full thickness — this piece was heading toward roughly 10mm — meant splitting the job into two stages rather than one continuous layup. Two things drove that decision. First, the gel time on the resin was short enough that laminating the full stack and getting the vacuum bag on in one go wasn’t realistic — and as it turned out, even half the stack was a tight enough time window on its own. Second, and more importantly, a laminate this thick risks a genuine exothermic reaction if it’s all catalyzed and curing at once. Thick resin sections generate their own heat as they cure, and if that heat builds faster than it can dissipate, the reaction runs away — degrading the resin and potentially damaging the part outright. Splitting the layup into two stages, with a cure in between, keeps that heat under control.

Between the two stages, the first half got sanded before the second layer went on — and this is worth dwelling on, because it’s a point of genuine disagreement in the industry. Peel ply leaves a surface that looks ready to bond to, but it isn’t, not without sanding first. Nylon peel plies carry a silicone release agent, and while that agent lets the peel ply come away cleanly, it also leaves a resin-rich surface that isn’t mechanically prepared for a strong secondary bond. Sanding that surface breaks up the resin matrix right at the surface, creating the texture a genuine structural bond actually needs. Some people will tell you a fresh peel-ply surface is good enough to bond straight to without sanding. It isn’t — and that’s not an opinion, it’s something borne out repeatedly in composite R&D lab testing over the years.

Pulling the old hardware out told its own story. The backing plate and bolts came out dramatically heavier than expected — somewhere in the order of four kilograms of stainless steel and hardware that had spent years slowly working itself loose. The bolt holes themselves had migrated significantly further than anyone had estimated from the outside.

What the old bulkhead revealed once opened up explained why the problem had accelerated recently rather than staying stable. The boat had spent a long stretch somewhere dry, during which the plywood core, already degraded, had effectively stopped deteriorating further simply because it had dried out. Arriving in the Caribbean changed that completely — regular rain meant the wood was wet again, and decay that had stalled picked right back up, accelerating fast. By the time it was opened up, what remained holding the chainplate location together wasn’t structural wood at all — it was just the thin glass skins on either side, with essentially nothing solid left in between.

With the new bulkhead section templated, laminated, and fitted into position — pin height set flush with the deck rather than proud of it — the final step was a proper sea trial rather than just taking the repair on faith. The approach was straightforward and practical: sail on it gently in good conditions, load it up progressively, and watch closely for any sign of movement or anything behaving the way it shouldn’t. Good settled weather made that possible, and with everything holding solid under real sailing load, the new port-side chainplate had proven itself.

This is the first of a three-part build. Part Two picks up with the engineering reasoning behind the finished system — why bonded surface area beats bolted compression for a load path like this, and how the chainplate and bulkhead work together as one integrated structure rather than two separate parts bolted together.

In This Series

This post is part of our chainplates deep dive. Read the full story here: Paikea’s Chainplates – The Full Story

Or explore the rest of the series:

In the Members Library

We’ve put together the complete photo build log for this one — from the first bolt that told us something was wrong through to the finished, painted chainplate. See the full build on the members page.