3D Printing for Robotics
Which parts are worth printing, which materials survive real loads, and where printed parts fail. For teams building competition robots, drones and research platforms.
Robotics is where 3D printing earns its place most clearly. A robot needs brackets and mounts that exist nowhere in a catalogue, and it needs them changed after every test. Printing turns a two-week machining order into an overnight job, and that iteration speed is usually worth more than the parts themselves.
It also punishes carelessness faster than any other application, because the parts actually get loaded. This page is about the difference.
What is worth printing — and what is not
Print geometry that is expensive or slow to make otherwise. Do not print components that are already cheap and better made another way.
Worth printing
- Chassis plates and frames — geometry changes every iteration
- Motor and sensor mounts — every combination is bespoke
- Gripper fingers and jaws — shape is task-specific and gets tuned constantly
- Cable guides and strain relief — trivial to print, tedious to buy
- Enclosures for electronics — fitted to your board, not a generic box
- Wheels and treads in TPU — grip you cannot buy off the shelf
- Jigs for assembly — makes a repeatable build out of a fiddly one
Not worth printing
- Threads under repeated load — use heat-set brass inserts
- Bearings — a real bearing costs little and lasts far longer
- Parts near sustained heat — most filaments soften below motor temperatures
- High-torque drivetrain parts — printed gears wear; metal does not
- Anything safety-critical — where failure is dangerous, not just annoying
- Standard fasteners — printed bolts are worse than bought ones in every way
Choosing material by part
Most robotics print failures are material choices, not printer problems. PLA is the usual culprit — it is stiff and prints beautifully, which flatters it right up until something snaps.
| Material | Use it for | Watch out for |
|---|---|---|
| PETG | Default for structural parts: chassis, brackets, mounts. | Slightly stringy. Worth tuning retraction once. |
| PLA | Prototypes and fit checks only. | Brittle under impact, softens in summer heat. Not for final parts. |
| ABS / ASA | Parts that get warm — motor mounts, driver enclosures. | Needs an enclosed printer. Warps badly without one. |
| Nylon | Gears, bushings, wear surfaces, living hinges. | Absorbs moisture very fast. Must be dried before printing. |
| TPU (flexible) | Wheels, bumpers, feet, gripper pads, strain relief. | Prints slowly. Direct-drive extruders handle it far better. |
| Carbon-fibre filled | Stiff structural parts where deflection matters. | Abrasive — needs a hardened nozzle or it destroys a brass one. |
Six design rules for parts that take load
A printed part is not isotropic. It is strong across layers and weak between them, and almost every structural failure traces back to ignoring that.
1. Orient the part so load runs across layers, not between them
This is the single most important rule and it is free. A bracket printed flat will peel apart along its layer lines under a load the same bracket printed on edge would shrug off. Decide the orientation before you finish the model, not in the slicer afterwards.
2. Add walls, not infill
For stiffness, four or five perimeters does far more than raising infill from 20% to 50%, and costs less material and less time. Infill supports the top surface; walls carry the load.
3. Fillet every internal corner
A sharp internal corner concentrates stress and is where the part will crack. A small fillet costs nothing to model and makes a large difference to how much abuse the part survives.
4. Use heat-set inserts wherever a screw comes out more than once
Printed threads strip after a handful of cycles. A brass insert melted into the plastic gives a metal thread that outlasts the part, and it is the cheapest reliability upgrade in robotics.
5. Leave clearance on fitting parts
Printers overshoot slightly. Around 0.2 mm of clearance on mating parts is a sensible starting point, but print a small test piece and measure rather than trusting a number — it varies by machine, material and temperature.
6. Design the printed part as the one that fails
When a mechanism jams, something gives. Make it the part you can reprint overnight rather than the gearbox or the motor. Deliberately choosing the weak link is a design decision, not an accident.
Why robotics teams want FDM, not resin
Resin produces beautiful parts and the wrong parts for this job. Standard resin is hard but brittle: it chips and snaps under exactly the impacts a robot takes routinely, and it becomes more brittle with sunlight exposure. Build volumes are also too small for most chassis components.
The exception is small non-structural detail — sensor housings, presentation models, display pieces for a competition stand — where surface finish is the point and nothing is carrying load.
The full comparison is on the FDM vs resin page, and FDM machines in stock are listed with recommendations by budget.
Frequently Asked Questions
Which 3D printing material is best for robot parts?
PETG is the sensible default: strong enough for most structural parts, tolerant of heat, and far less brittle than PLA without needing an enclosed printer. Use PLA only for prototypes and non-load parts — it is stiff but snaps, and softens in heat. ABS and ASA suit parts that get warm, such as motor mounts or anything near a driver board, but need an enclosure. Nylon is the best choice for gears and wear surfaces because it is tough and self-lubricating, though it absorbs moisture badly and must be kept dry. TPU is for anything that should flex: bumpers, feet, cable strain relief and gripper pads.
Can 3D printed parts handle real loads in a robot?
Yes, within limits that are predictable once understood. A printed part is strongest across layers and weakest between them, so it fails by splitting along layer lines rather than by tearing. Orient the part so the main load runs across layers rather than trying to pull them apart, increase wall count rather than infill for stiffness, and add fillets at corners where stress concentrates. Printed brackets, chassis plates and mounts hold up well in practice; printed threads under repeated load do not, and should be replaced with heat-set brass inserts.
Are 3D printed gears strong enough for robotics?
For low-torque and moderate-speed applications, yes, particularly in nylon or a filled filament. They will not match machined metal for high torque or continuous duty, and they wear faster. Two practical rules: print gears flat so the teeth are formed across layers rather than stacked along them, and design the printed gear as the sacrificial part in the train so that when something jams, the cheap component fails rather than the gearbox or motor.
What should you not 3D print for a robot?
Threads that will be repeatedly tightened — use heat-set brass inserts instead. Bearing surfaces under real load, where a proper bearing costs little and lasts far longer. Anything carrying significant sustained heat, since most filaments soften well below what a hot motor or driver reaches. And anything where failure is dangerous rather than merely inconvenient. Printing is for geometry that would be expensive or slow to make otherwise, not for replacing components that are already cheap and better.
Should a robotics team use FDM or resin printing?
FDM, almost without exception. Robotics needs parts that take load, and resin parts are hard but brittle — they chip and snap under the impacts a robot experiences routinely. Resin also degrades under sunlight and build volumes are too small for most chassis parts. The exception is small detailed non-structural components such as sensor housings or presentation models, where resin's surface finish is worth having.