90° overhangs without supports: wave toolpaths (wave overhangs)
A 2026 paper lays lines as waves spreading from the edge, so flat overhangs print with no supports: 30–48% less sag than arc overhangs on hard shapes, at about 2 mm/s.
- Wave overhangs bond each new line to the side of a line that has already set, not to the layer below, so they can reach out a full 90°.
- On the paper's hardest shape, maximum sag was 30–48% lower than arc overhangs, with no coverage gaps.
- The price is speed: the demo used 2 mm/s and took 72% longer than with supports, but used 39% less material.
- As of October 2026 it lives in experimental forks of OrcaSlicer and PrusaSlicer. Use PLA with full fan.
A flat 90° overhang normally needs supports, because there is nothing under the plastic. Wave overhangs get around this by bonding each new line to the side of a line that has already set, spreading out from the edge one wavefront at a time. A 2026 paper found that on the hardest shape, overhangs sagged 30–48% less than with arc overhangs, but the overhangs had to print very slowly, around 2 mm/s.
As of October 2026 you can try it through experimental forks of OrcaSlicer and PrusaSlicer. It is not a button in the mainline slicers yet.
The old problem: the 45° rule and supports
The overhangs and bridging lesson explains why normal extrusion manages about 45° before you need supports, which waste material and time, have to be removed, and mark the underside. Design-side workarounds (chamfers, teardrop holes) are in design for no supports.
In 2022 arc overhangs appeared: concentric arcs laid out from the edge, each new arc gripping the last, so a 90° overhang can print. The inventor’s README says to print PLA as cold as possible (he used 195 °C), fans at full, and slow, about 5 mm/s. The weak spots are corners, holes and narrow areas, where arcs don’t fit.
How wave overhangs work
The authors borrow Huygens’ principle, the same idea used to describe water waves.
- Each new line sits a fixed distance from the previous one, so they overlap sideways by a set amount.
- Wavefronts are long and continuous, so earlier lines have time to set before the next one leans on them.
- The waves bend around corners and holes and pass through narrow necks, where arc overhangs fail.
- Extrusion is computed from the volume of each line, not tied to the nominal layer height.
The work was published in Additive Manufacturing Letters (June 2026), tested on an unmodified desktop FDM printer with PLA, with sag measured by a 3D scanner.
Test results
Wave versus arc, same settings, on single-layer overhangs at three difficulty levels:
| Measured | Result |
|---|---|
| Easy and medium shapes | wave and arc similar |
| Hard shape (with a concavity): maximum sag | wave 30–48% lower |
| Hard shape: spread of surface deviation | wave 49–66% lower |
| Uncovered gaps | systematic with arc, none with wave |
| Multi-layer demo vs supports | 39% less material, 72% longer print |
The 72% comes from printing the overhangs at a conservative 2 mm/s.
Wave overhangs don’t make prints faster. They trade time for no support removal and a clean underside.
Working out the time for one overhang layer
A rough estimate: a flat flange 20 × 30 mm = 600 mm², assuming a line spacing of about 0.4 mm.
- Approximate total line length = 600 ÷ 0.4 = 1,500 mm.
- At 2 mm/s (the paper’s speed) = 1,500 ÷ 2 = 750 s ≈ 12.5 minutes.
- At 5 mm/s (the arc overhang README’s suggestion) = 300 s = 5 minutes.
- A normal layer on supports at 40 mm/s = 1,500 ÷ 40 ≈ 38 seconds, plus printing the supports and removing them.
These are estimates to show the proportions; they ignore travel and acceleration. Use the slicer’s preview time for the real number.
The point: a single first overhang layer can take ten-plus minutes. A part with overhangs at several heights gets much longer.
Ways to handle a 90° overhang, compared
| Method | Underside | Time | Status, Oct 2026 |
|---|---|---|---|
| Normal supports | marked, needs cleanup | moderate + removal | in every slicer |
| Redesign (chamfers etc.) | good | no extra | needs a model change |
| Arc overhangs | good, some gaps | slow (~5 mm/s) | fork/script; inventor stopped work in Feb 2023 |
| Wave overhangs | good, less sag on hard shapes | very slow (~2 mm/s in the paper) | experimental Orca and Prusa forks |
How to try it
There are two forks:
- PrusaSlicer-WaveOverhangs, by the arc overhang inventor, with settings such as
Wave overhang pattern(Monotonic / Zig Zag / Smart),Wave overhang line spacing,Wave overhang print speedandWave overhang fan speed. - OrcaSlicer-WaveOverhangs, which detects overhangs and generates the waves automatically, also handles angled overhangs, and has about 40 settings. It is labelled Experimental / Alpha.
Steps in the Orca fork: open a model with an overhang → Print Settings → Wave overhangs → turn on “Use wave overhangs (Experimental)” → slice and check the preview. The overhang should show lines spreading out in rings from the edge.
Safe starting point: PLA, part fan at full, low overhang speed (start in single-digit mm/s). Try a short flange first; the README itself says most setting combinations have not been print-tested.
Common mistakes
- Using PETG, ABS or PC. They cool slowly or warp easily; the fork says they are likely to fail.
- Turning the fan down. Each ring must set before the next one grips it. Weak airflow means sag.
- Going long on the first try. Larger spans curl up at the edges from heat build-up. Start small.
- Printing real jobs straight from a test fork. Install it separately and keep your production profiles in the mainline slicer.
- Skipping the preview. Where the waves can’t cover an area, the Orca fork adds supports only there. Check before you print.
Previous: Z anti-aliasing for smooth shallow slopes · Next: conical slicing for overhangs on a 3-axis printer
Sources
FAQ
Does it work with PETG or ABS?
The paper only tested PLA, and the OrcaSlicer fork's README says PETG, ABS and PC are likely to fail; PETG cools too slowly and delaminates under heavy fan. The authors expect higher-expansion plastics to curl up more at the edges.
How far can it reach?
Nobody has measured that yet. The authors say maximum span, top speed and overhang strength are uncharacterised, and large overhangs may still need supports.
How is it different from arc overhangs?
Arc overhangs stack circular arcs from the edge and struggle at corners, holes and narrow necks. Wave overhangs send a wavefront that bends around corners and holes, and the long continuous lines give earlier ones more time to cool.
Circuits and code were tested on sample boards. Boards differ in pin layout, so check your board's pinout before you power anything.