The thing that no longer exists: making replacement parts for obsolete kit

A steady share of the one-off work that comes into the workshop isn't decorative at all. It's a single small part for something otherwise perfectly good — a knob, a bracket, a gear, a clip, a bezel — where the manufacturer has gone, the model is discontinued, or the spare exists but costs more than the machine is worth.

These are some of the most satisfying jobs we do, and also the ones where expectations most often need managing. Here's an honest account of what's feasible.

What reproduces well

• Plastic parts that broke because they were plastic. Clips, tabs, housings, handles, battery doors. These usually failed at a known weak point and we can print a replacement that's stronger than the original, because we're not constrained by injection-moulding economics or draft angles.
• Brackets, mounts and spacers. Simple geometry, forgiving tolerances, easy wins.
• Knobs, dials and levers. Provided the shaft interface can be measured or the original still exists in part.
• Flat components. Gaskets, shims, panels, escutcheons, faceplates — laser cut from the right sheet material, often better than the original.
• Badges, plates and legends. Fibre marking on metal reproduces worn or missing labelling on machinery, and it's permanent.

What's difficult or not worth it

• Load-bearing metal parts. If a component holds a real structural load, a printed or laser-cut substitute is usually the wrong answer. We'll say so rather than make something that fails in service.
• Precision-fit mechanical parts. Gears meshing under load, bearing surfaces, threads carrying torque. Sometimes possible, often better machined properly, and always more expensive than people expect.
• Anything safety-critical. Brakes, gas, electrical isolation, anything that holds a person up, anything in a vehicle's safety systems. We won't quote these.
• Parts with no reference at all. If the original is missing entirely and there's no photo, no drawing and nothing to measure against, we're guessing. Sometimes an owners' forum has what's needed — worth a look before you write it off.
• Anything still in production and cheap. If the genuine part is fifteen pounds and available, buy the genuine part. We'll tell you that too.

How to give us the best chance

The quality of what you send determines the quality of what comes back. In rough order of usefulness:

• Send the broken part itself. Even in pieces. Even badly. A physical original — however mangled — beats any description, because we can measure it directly.
• Send the mating part too, if you can. The thing it bolts to, plugs into or sits against. Fit is defined by the interface, and we can only guarantee an interface we can hold.
• Photograph it against a ruler. Multiple angles, straight on rather than at a slant, with a steel rule in frame. Phone photos are fine; perspective distortion is the enemy.
• Give us the key measurements. Especially anything that has to fit — hole centres, shaft diameter, overall thickness. Digital callipers are worth the twelve quid if you do this more than once.
• Tell us what it does and how it failed. A part that snapped under vibration needs a different design from one that wore away. Knowing the failure mode lets us fix the original's weakness rather than reproduce it.
• Say what it's near. Heat, oil, solvents, sunlight, food, water. This picks the material.

Improving on the original

This is the part people don't anticipate. A reproduction doesn't have to be a copy. Because we're making one, we're free of the constraints that shaped the original:

• Thicken the failure point. If it broke at a thin section, that section can simply be thicker. There's no mould to respect.
• Add a fillet. Most plastic parts crack at sharp internal corners. A radius there costs nothing and dramatically improves fatigue life.
• Change material. The original may have been the cheapest plastic that worked at scale. We can use something tougher or more heat-resistant.
• Split it in two. A part that's awkward to fit as one piece can sometimes be made as two that bolt together — easier to install without dismantling the whole assembly.
• Design it for next time. We keep the file. If it fails again, or a friend with the same machine needs one, it's a reprint rather than a fresh project.

Cost and lead time, honestly

The measuring and modelling is the bulk of the cost, not the making. A simple bracket might be an hour of modelling and two hours of unattended print time. Something with awkward organic curves or a fussy interface can be half a day at the screen.

This means the economics only work above a certain threshold. If the whole appliance is worth eighty pounds, a bespoke part probably isn't sensible. If it's a machine worth thousands, or something with sentimental value, or a piece of equipment your business depends on, it usually is.

We'll give you a straight answer on which side of that line you're on before you spend anything.

A note on rights

We reproduce parts for owners repairing their own equipment. We don't produce copies of parts that are currently manufactured and protected, and we don't reproduce branded components for resale. If you're unsure whether your request falls the right side of that, ask and we'll talk it through.

Got a broken part and no supplier? Photograph it against a ruler and email enquiries@cobblestonecreative.co.uk — we'll tell you inside 24 hours whether it's worth doing. Related: six questions that pick the material.

Ask about a replacement part