Plastic spanning a gap — the top of a hole, the underside of a horizontal opening, the ceiling of an internal cavity — droops in the middle instead of running straight across. Badly enough, and the strands break and fall.
What is holding it up
A bridge is a strand of molten polymer stretched between two anchors with nothing underneath. Only two things resist gravity: the tension the moving nozzle puts into the strand as it draws it out, and how quickly the strand solidifies.
Both are worth stating explicitly because both are counter-intuitive in their consequences.
Tension comes from speed. The nozzle is effectively drawing the strand taut as it travels. Move slowly and the strand is laid down slack; move quickly and it is pulled straight before it has time to sag. This is why the correct instruction is to print bridges faster, and why slowing down — the universal remedy for most print problems — makes bridging worse.
Solidification comes from airflow. Every second the strand spends soft is a second gravity is working on it. Maximum cooling over a bridge is right on any material that tolerates cooling at all.
The settings, and what each does
- Put maximum airflow over the bridge layers. Most slicers have a separate bridge cooling setting; this is the first thing to turn up.
- Raise bridge speed rather than lowering it. Somewhere around 40 mm/s is a common starting figure, and machines with good cooling manage more.
- Reduce bridge flow slightly, to around 90 or 95% of normal. A slightly thinner strand cools faster and weighs less, and both help. Too little and the strand breaks mid-span.
- Check the bridge direction. Slicers choose a direction to lay bridge lines, and it is not always the shortest crossing. Forcing the direction so the strands take the short route across an opening can halve the effective span.
- Consider one extra solid layer above the bridge. The first layer over a gap is never perfect; the second lands on something and tidies it up.
Where the span becomes the problem
Cooling and speed have limits. Beyond about 40 mm of unsupported span, most machines and materials produce a visible sag no matter how well tuned they are, and past that the strand can break entirely.
At that point the honest options are supports, a redesign, or a change of orientation. A long horizontal opening can often be rotated to print vertically, or given a chamfered or pointed top that the printer builds as a steep overhang instead of a flat bridge — a technique borrowed from designing parts for printability rather than from tuning the printer, and by far the most reliable of the three.
Bridging is not the same as an overhang
The two get conflated and they behave differently.
A bridge is anchored at both ends, so it can be held in tension, and it fails by sagging in the middle. Fast and cold is right.
An overhang is anchored at one end. There is no tension to be had, and it fails by drooping or curling upward at the free edge. Slow and cold is right there, because each bead needs time to set before the next lands on it.
Applying bridge advice to an overhang, or the reverse, is a common reason a change makes one region better and another worse on the same part.
Test it deliberately
A bridging test print — a series of increasing spans between two towers — takes minutes and gives you your machine's real limit in your usual material. That number is worth more than any general guidance, because it depends on the cooling duct, the fan, the ambient temperature and the filament, and it turns a design decision into a measured one.
The materials that bridge well and badly
PLA bridges better than anything else in common use: it sets fast and it holds tension well. PETG is noticeably worse, staying soft longer and sagging further. The polyamides are worse still. ABS and ASA suffer a double penalty, because the cooling that would help them bridge is the same cooling that causes them to crack and delaminate — so on those materials, bridging is a design problem to be avoided rather than a settings problem to be solved.