How to design FDM parts that need little or no support
Support material drives cost, print time and manual labour in FDM. Learn which geometry needs it and how to redesign parts to avoid it.
Support material is not a free safety net. Every gram of it costs machine time to print, material to buy, and someone's hands to remove afterwards. On a single prototype that overhead is tolerable. Multiply it across a production batch of 200 or 2,000 parts and support becomes one of the largest line items in the unit cost, often larger than the plastic itself.
This guide sets out why support exists, what it actually costs, and how to think about geometry so that a part prints clean without it. The other guides in this cluster go deeper on overhang angles, bridging, corner treatment and trapped support. This one is the overview and the design mindset.
Why support material is a cost driver, not a detail
FDM builds a part one layer at a time, and each new layer needs something underneath it to fuse to. Where the model overhangs empty space beyond what the material can bridge or self-support, the slicer inserts scaffolding: either the same material as the part (single-material support) or, on machines set up for it, a soluble or breakaway second material. Both approaches add print time because the nozzle has to trace extra paths, and both add material cost because that scaffolding is purged after the print.
The bigger cost is usually not the material or the machine time. It is the manual labour afterwards. Single-material support has to be broken or cut away by hand, often with pliers, a scalpel or a file, and every one of those operations is a chance to scratch, gouge or crack the part. Soluble support avoids the mechanical risk but adds a wash step with its own cycle time, tank maintenance and drying time. Either way, someone has to touch every single part, which is exactly the kind of manual, non-repeatable step a production process should minimise.
- Extra machine time to print the scaffolding itself
- Extra material purchased and then discarded
- Manual removal time per part, which does not shrink with volume
- Risk of surface damage, snapped features or scratched cosmetic faces
- Support marks and witness texture left on the part after removal
- Scrap risk when removal goes wrong on a finished part
Support marks and where you can still reach them
Wherever support touches the part, it leaves a mark: a rougher patch, a slight witness line, sometimes a shallow dimple where the contact points were sparse. On a functional, hidden face that is irrelevant. On a cosmetic face it usually is not acceptable, and the fix is either to reorient the part so that face is not the one under an overhang, or to redesign the geometry so no support lands there in the first place.
Accessibility matters just as much as visibility. Support hidden inside a narrow channel or a small enclosed pocket may be technically removable in the slicer preview but practically impossible to reach with a tool. If you cannot picture a pair of tweezers or a pick reaching every surface that carries support, treat that as a red flag during design review, not something to solve after the first article comes off the printer.
What actually makes support unavoidable
Not every overhang needs support. Angles up to roughly 45 degrees from vertical are generally self-supporting in FDM, and with careful cooling and slow perimeter speeds some materials manage a bit more. Support becomes necessary when a feature goes further than that: a horizontal boss sticking out from a wall, a hole axis running perpendicular to the build direction with a large diameter, a shape that curves back under itself, or a span between two walls that is longer than the material can bridge cleanly.
The honest cases where support genuinely earns its keep are rare in well-designed parts: a true undercut that carries a functional requirement no other geometry can meet, or a one-off prototype where redesign time costs more than the removal work. Everything else on the list below has a support-free alternative that does the same job.
| Feature that needs support | Support-free redesign |
|---|---|
| Horizontal cylindrical boss on a vertical wall | Add a teardrop or chamfered top instead of a full circle |
| 90 degree flat overhang / shelf | Break it into a 45 degree chamfer leading up to the shelf |
| Large horizontal hole through a wall | Reorient so the hole axis is vertical, or use a teardrop hole profile |
| Wide unsupported bridge across an opening | Add a rib, arch or reduce span below the bridging limit |
| Sharp undercut / reverse taper | Split the part along the undercut and join after printing |
Why this matters most in serial production
A single support-heavy prototype costs a few extra minutes of removal work. The same design printed 500 times costs 500 times that removal work, on every batch, forever, unless the design changes. That is the arithmetic that makes support-free design a serial production issue, not a cosmetic preference. In a print farm running many machines in parallel, support also complicates automation: parts that need manual picking through a support structure cannot be handled by simple bed-clearing routines, and they slow down every operator on the line.
We look at support requirements as part of every DFM review, because it is one of the few design changes that pays for itself immediately and keeps paying for itself across the whole production run. The following guides in this cluster cover the specific geometry rules: overhang angles, bridging limits, corner and edge treatment, and how to deal with support that would otherwise be trapped inside the part.
Frequently asked questions
- Does support material always add cost?
- Almost always, yes. Even when the support volume itself is small, the extra print time and the manual removal step add cost that scales with the number of parts, not just with material volume.
- Can soluble support solve all these problems?
- It removes the mechanical damage risk and reaches places tools cannot, but it adds a wash cycle, tank maintenance and drying time, and it is not available on every material and machine combination we run.
- How do I know if my part needs support before I send it in?
- Check every face for overhang angles steeper than about 45 degrees from vertical, look for horizontal holes and bosses, and check spans between walls against typical bridging limits. Send us the STEP file and we will flag it during the review.
- Is it ever worth designing in support on purpose?
- For a true one-off prototype where redesign time exceeds the removal time, yes. For anything that will be reprinted more than a handful of times, redesigning the geometry almost always pays back faster.
- Does support material affect part strength?
- Not directly, but the removal process can. Aggressive cutting or breaking near a functional feature can nick the surface and create a stress riser, which is another reason to avoid support on load-bearing faces.
Have your part reviewed before production
Send us your CAD file together with the application, load and operating conditions. We review geometry, orientation, material and tolerances and come back with concrete change proposals and a quote.
Read next
Overhang angles and self-supporting geometry in FDM
Overhangs up to about 45 degrees print clean without support, but the transition from clean to sagging is gradual, not a hard cutoff.
Bridging in FDM: how far a nozzle can span without sagging
A short, well-anchored bridge can span open air cleanly, but push the length or skip one anchor point and the strand sags before it cools.
Fillets, chamfers, corners and the elephant foot effect
Bottom edge geometry, internal corner radii and lead-in chamfers all interact with print quality and support in ways that are easy to get wrong.
Trapped and internal support material in FDM parts
Support material that ends up sealed inside a finished part cannot be removed at any price, which makes internal geometry a design decision, not a slicer setting.
Part Orientation for FDM: The Most Consequential Production Decision
Orientation is decided once per build, but it touches every other property of the finished part.