A bridge is a length of plastic printed horizontally across open air, anchored only at its two ends. It is one of the most useful tricks in FDM design: a well-placed bridge lets you span the top of a doorway, cap a hollow box, or roof an internal channel with no supports to remove and no scarred surface underneath. But bridges only work within limits, and understanding those limits lets you design parts that print clean the first time.
What actually happens on a bridge
When the nozzle reaches the edge of a gap it keeps extruding into thin air. The freshly laid line has no layer beneath it, so it relies on two things to survive: tension between its anchored ends, and cooling fast enough to solidify before gravity pulls it down. Get either wrong and the strand sags into a droop, or breaks and drags as loose filament. Because a bridge is held taut, it also shrinks and tightens as it cools, which is why the underside of a good bridge looks like a set of straight, slightly taut ropes.
A bridge must be anchored at both ends on the same layer. A strand supported at only one end is an overhang, not a bridge, and it will curl up or droop almost immediately. If only one side is solid, you need a support or a redesign, not bridging.
How far can you span?
With good part cooling, most machines bridge 5–10 mm effortlessly and can reach roughly 20–50 mm with tuned settings, decreasing sag as the span grows. Beyond that, the strand’s own weight wins and the middle droops noticeably. The achievable span depends heavily on cooling power, material, and speed – PLA with a strong fan bridges far better than PETG or ABS, which stay soft longer.
| Span | Expectation | Approach |
|---|---|---|
| ≤ 5 mm | Prints clean on nearly any setup | No special action needed |
| 5–20 mm | Reliable with good cooling | Enable bridge settings; full fan |
| 20–50 mm | Possible with tuning; some sag | Slow bridge speed, max cooling, dry filament |
| > 50 mm | Sag usually unacceptable | Add a support, split into shorter spans, or redesign |
These are guidelines, not guarantees. A cheap way to learn your own machine’s ceiling is to print a bridging test with progressively longer unsupported spans and note where the underside starts to droop.
Cooling and speed for clean bridges
Bridges are where part cooling earns its keep. Run the fan at or near 100% so each strand freezes the instant it lands. Speed is a balance: too fast and the thin, unsupported line stretches and snaps; too slow and it stays molten long enough to sag. Many slicers expose a dedicated bridge speed and flow, and a moderate bridging speed with slightly increased flow – so the strand is a touch fatter and stronger – often gives the cleanest result.
Dry filament matters more on bridges than anywhere else. Moisture flashes to steam as the plastic exits the nozzle, leaving weak, foamy strands that sag and break exactly where they have no support to hide the flaw.
Two more slicer habits help: make sure bridge infill runs as a single straight pass across the shortest direction of the gap, and let the slicer detect bridge perimeters too, not just infill, so the outline over the gap is treated with the same care.
Designing bridges instead of supports
The best support is the one you designed out of existence. When you control the model, you can turn a support-hungry feature into a clean bridge. Orient a part so a horizontal hole becomes a bridged span rather than a steep overhang. Cap an enclosure across its narrowest dimension so the bridge is as short as possible. Break one long span into two shorter ones by adding an internal pillar or rib where the design allows. Every gap you keep under your machine’s comfortable span is a support you never have to print or remove.
Bridging is the close cousin of overhang handling: below a certain angle a wall is a self-supporting overhang, and only when it goes fully horizontal does it become a bridge. Knowing where that transition sits on your printer lets you choose the better tool for each feature – see our overhang angles guide for where self-support ends and bridging or supports begin.
When to chamfer instead of bridge
Not every gap should be bridged. The classic case is the top of a horizontal hole or an internal channel. Instead of leaving a flat roof that must bridge – and often sags at its apex – reshape the top of the hole into a teardrop or add a 45° chamfer so the roof becomes a self-supporting slope rather than an open span. A teardrop-profiled hole prints cleanly with no support and no bridge sag, at the cost of a slightly non-round top that is usually irrelevant for a bolt or a cable. Reserve true flat bridges for surfaces that must stay flat, like the ceiling of an enclosure, and use chamfers or teardrops everywhere the exact top profile does not matter.
Want to design parts that print support-free by default? Our courses cover bridging, overhangs, and orientation as one connected skill.
Take the FDM course See Pro pricingBridging checklist
- Confirm both ends of every span land on solid material on the same layer.
- Keep spans under your tested comfortable limit – often ~20 mm, more with strong cooling.
- Run part cooling near 100% and dry the filament before printing bridges.
- Use the slicer’s bridge speed and flow settings; tune speed so strands neither snap nor sag.
- Orient the part so gaps span their shortest direction.
- Split long spans with an internal rib or pillar where the design allows.
- Chamfer or teardrop the tops of holes and channels instead of bridging a flat roof.
- Cross-check borderline features against the overhang angles guide, and start with Foundations if the basics are new.