3D Scan a Broken Boat Part With Your Phone: A Practical Guide
Something on board breaks, the part number leads nowhere, and the yard quotes six weeks. That is the moment most boat owners start thinking about 3D printing — and the first question is always the same: how do I get this broken thing into the computer?
Smartphone 3D scanning has quietly become good enough to answer that. Not perfect, not a substitute for a caliper, but genuinely useful for the awkward organic shapes that are painful to measure by hand. Here is what actually works, where it falls apart, and how to decide whether to scan at all.
First: decide whether scanning is the right tool
Scanning is seductive and often unnecessary. Be honest about the geometry in front of you:
- Prismatic and round parts — spacers, bushings, hose tails, threaded adapters, flat brackets. Do not scan these. A £15 digital caliper gives you dimensions accurate to 0.02 mm; a phone scan gives you a wobbly surface accurate to maybe 0.5 mm. Measure and model, or start from a parametric file and type in your numbers.
- Free-form and sculpted parts — a moulded handle, a cowl vent, a shaped trim piece, a fairing that follows the hull. Scan these. There is no sensible way to measure a compound curve with a ruler.
- Parts that must mate with a curved surface — a mount that sits on a coachroof camber, a pad that follows a toe rail. Scan the surface, not the part, and build your bracket onto it.
Most marine repairs fall in the first category, which is why our catalogue is built around sized-to-fit files rather than fixed models. We made the case for that approach in Parametric vs Modelling Your Own, and it holds here: if you can describe the part with six numbers, scanning is the slow road.
What your phone can actually do
There are two different technologies hiding behind the words "phone 3D scanning", and they behave very differently.
| Method | How it works | Realistic accuracy | Best for |
|---|---|---|---|
| LiDAR / depth sensor | Time-of-flight sensor on Pro-tier iPhones and iPads | ~5–10 mm | Rooms, lockers, hull sections, large mounting surfaces |
| Photogrammetry | Dozens of ordinary photos processed into a mesh | ~0.3–1 mm on small objects | Hand-sized parts, handles, brackets, castings |
The counter-intuitive part: for a broken hatch handle, the plain camera beats the LiDAR sensor comfortably. LiDAR is built for spatial mapping, not for fine detail. Use it when you need the shape of a locker or the curve of a deck; use photogrammetry when you need the part itself.
A photogrammetry workflow that works on a boat
1. Prepare the object
Photogrammetry works by matching features between photos. Shiny, black, or featureless surfaces have no features to match, and stainless or glossy gelcoat will simply fail. Fix it with a light dusting of chalk spray, athlete's foot powder, or dry shampoo — anything that leaves a matte, slightly speckled film and washes off. A few dots of pencil or a sticker or two also help the software lock on.
2. Set up
Put the part on a plain, non-reflective surface — a cockpit cushion, a sheet of card, a folded towel. Work in bright, even shade, never in direct sun; hard shadows get baked into the mesh as if they were geometry. Overcast days are ideal, which is convenient, because those are the days you are not sailing anyway.
3. Shoot
Walk around the part in three orbits — low, level and high — taking a photo every 10–15 degrees. That is roughly 60–80 images. Keep about 70% overlap between consecutive shots, hold the same distance, and move yourself rather than the object. Then flip the part over and repeat, so the underside is captured too. Lock focus and exposure if your app allows it.
4. Include a scale reference
This is the step people skip and regret. Photogrammetry produces a shape with no absolute size. Lay a steel rule or a coin next to the part, and — more importantly — take one caliper measurement of a clean, flat feature on the real object. You will use that number to scale the mesh in your CAD or slicer. Without it, your printed part will be confidently the wrong size.
5. Process and clean up
Free or low-cost apps such as Polycam, RealityScan, KIRI Engine or Scaniverse will hand you a mesh in a few minutes. Expect it to be messy: floating fragments, a fused-in chunk of the towel, holes where the part touched the table. Import into Meshmixer, Blender or Fusion, delete the background, close the holes, and check the scale against your caliper number.
The step that separates a scan from a printable part
A raw scan is a lumpy triangle soup. It will print, and it will look approximately right, but the flat faces will not be flat and the holes will not be round — which matters enormously the moment a bolt has to pass through one.
The reliable approach is to treat the scan as a reference, not as the model. Bring the mesh into CAD, align it to the origin, then rebuild the part on top of it with proper sketches: real circles for holes, real planes for faces, real dimensions taken with the caliper wherever a surface is measurable. Keep the scanned geometry only for the parts that genuinely are organic. The result is a clean, editable, correctly sized model — and one you can adjust later when the fit is 0.3 mm tight.
If the broken part has threads, do not trust the scan for them at all. Identify the thread properly and generate it. Our guides on measuring a hose and thread and on replacing discontinued parts cover that in detail.
When the part is broken in half
A common case: you have the pieces, but none of them is complete. Two tactics help. If the part is symmetrical, scan the intact half and mirror it — this rescues a surprising number of handles and brackets. If it is not, dry-fit the fragments back together with masking tape or a blob of hot glue, scan the assembly, and repair the seam digitally. Either way, capture the mating features first: the bolt spacing and the mounting footprint are what must be right, and those are usually still measurable on the surviving wreckage.
Print it, then check the fit before you finish
Print the first attempt fast and cheap — 0.3 mm layers, 15% infill, PLA if you have it — and take it to the boat. You are testing fit, not strength. Only once the mating surfaces are right do you reprint in the material the job actually needs, in the right orientation. Material choice is covered in PETG vs ASA vs Nylon, and if the part lives on deck, UV resistance is not optional.
You will also want printed jigs and measuring aids for the fit-checking stage, and it is worth browsing the free files before you scan anything — a good number of "obsolete" boat parts turn out to be generic shapes that someone has already parameterised.
A note on safety
3D-printed parts are not certified marine equipment. A scanned replica reproduces a shape, not a material specification or a load rating — the original may have been glass-filled nylon, cast bronze, or moulded to a standard your printer cannot approach. Verify that the material and the part are suitable for the job before use, and do not use printed parts for standing rigging, steering, gas systems, or anything below the waterline unless you have genuinely engineered and tested the replacement.
The short version
Measure what can be measured. Scan only what cannot. Always include a scale reference, always rebuild the functional features in CAD, and always test-fit a rough print before committing to the real material. Done that way, a phone and twenty minutes of photographs will get you a part that no chandlery still stocks.
Questions about a specific part, or a shape you are not sure how to capture? Write to info@marinelab3d.com — we are always interested in the awkward ones.