What Makes a 3D Printed Part Functional

A part is functional when it survives its job for as long as the job lasts, and when the next hundred you print do the same. That second half is the one that costs money. Most shops can make one good bracket. Making the four hundredth bracket identical to the first is a different problem, and it is solved before the slicer, not inside it.

The order of what decides the outcome is not the order most people work in. Material choice first, then part orientation, then wall count and geometry, then temperature, then everything else. Infill is near the bottom. The spool brand only matters at the point where the filament stops being a constant, which is later in this article and less romantic than it sounds.

What follows is the sequence we see working in production shops, plus the parts of it that get skipped.

Write down the duty cycle before you pick anything

Functional is a word that means nothing until you say what the part has to survive. Before opening CAD, get five answers on paper:

• The load, and whether it is constant, intermittent, or impact
• The highest temperature the part will see, including a van in July and the heat coming off whatever it is bolted to
• Chemicals it touches, including coolant, IPA, hand oils and cleaning agents
• Whether it lives in sunlight
• The tolerance the fit actually needs, and what it mates against

Half the failures we get asked about are not print failures. They are a part designed for a load nobody wrote down, printed in a material that was never going to hold it. No slicer profile fixes a material selection error, and a shop that treats every failure as a settings problem will chase its own tail for a year.

Matching material to the failure mode

Pick for how the part will fail, not for the biggest number on the spec sheet. Published tensile figures come from a bar pulled slowly along one axis in a lab. Your part gets dropped, over-torqued, loaded at an angle across layer lines, and left somewhere warm.

The tensile figures below are the ones we publish for our own grades. Every producer's formulation differs, so treat them as the shape of the comparison rather than a universal constant, and check the sheet for whatever you actually run.

Material	Tensile	Behaviour at failure	Heat	Good for
PLA-Pro	62 MPa	Snaps with little warning	Softens near 55–60 °C	Jigs, fixtures, indoor enclosures, anything dimensional
PETG	53 MPa	Bends, then tears	Usable to about 70 °C	Brackets, fluid-adjacent parts, handled hardware
ABS / ASA	46 MPa	Bends, then cracks	Good past 90 °C	Outdoor housings, automotive clips, hot environments
TPU	Varies with shore	Deforms, recovers	Grade dependent	Gaskets, bumpers, vibration mounts, living hinges

PLA is underrated for functional work and oversold at the same time. It is stiff, it prints to size better than anything else on the list, and for a jig that lives on a bench at 22 °C it is often the correct answer. It is also the first material to quietly creep and let go when a part is clamped, warm, or both. A PLA fixture in a machine shop near a heat source is a maintenance ticket waiting to be written.

PETG is where most functional work lands. It gives up peak tensile strength and buys toughness, which is the trade you want on anything that gets handled. ASA earns its place outdoors, where PETG yellows and loses impact strength over a season or two, and where PLA is simply not a candidate. Run ABS and ASA in the 240–260 °C range and give them an enclosure, because a warped part is a rejected part regardless of how strong the polymer is.

If the answer to the duty cycle sheet is sustained load at 100 °C, none of these are it. That is a filled nylon or a different manufacturing process, and it is better to find that out on paper than after two weeks of profile tuning.

Orientation is the largest single lever

A printed part is anisotropic. It is strong along the extruded bead and weak between layers, and the difference is not small. The same geometry in the same material can be several times tougher in one orientation than another, and the only cost of getting it right is print time and support material.

The rule is simple to state and easy to forget under deadline: orient so the main load runs along the layers, not across them. For a bracket carrying a hanging weight, that usually means printing it on its side or standing it up rather than flat, even when flat is faster and needs no support.

Things that follow from this:

• Any thin protruding feature printed flat will snap at its root. Rotate it or thicken it.
• Bolt holes and bosses want their axis parallel to the layers where possible, so clamping load does not try to peel layers apart.
• If two load paths conflict, split the part in two and fasten them together. A two-piece part in the right orientation beats a one-piece part in the wrong one.
• Fillet every internal corner. Printed parts crack at sharp internal corners in exactly the way machined ones do, and a 2 mm fillet costs nothing.

Walls do the work, infill mostly does not

When a part breaks, the reflex is to raise infill. It is usually the wrong lever. Under bending or torsion, most of the stress sits in the outer shell, so going from three walls to five at the same infill buys more than doubling infill at three walls, and costs less time.

A sane default for functional parts is four to six perimeters, 20 to 30 percent gyroid or cubic infill, and five top and bottom layers. Then stop adjusting sliders and look at the part. Keep wall thickness a whole multiple of your extrusion width so the slicer does not leave a void down the middle of a thin wall. Use heat-set inserts or captive nuts anywhere a fastener gets real torque, because threads cut into plastic will strip and nobody in the field will admit to over-tightening them.

Temperature, cooling and the finish trap

The most common self-inflicted weakness in functional printing is a profile tuned for looks. Lower nozzle temperature and more part cooling give crisper corners, fewer strings and worse layer adhesion. For a display piece that trade is fine. For a load-bearing part it is backwards.

For functional work, bias toward the top of the material's working window and cut part cooling back rather than running it at full blast. You will accept slightly softer overhangs and get a stronger Z axis, which is the axis the part is going to fail in. Confirm the window on your own hardware with a temperature tower instead of trusting a number from a forum, since hotends, flow rates and grades all sit in different places inside the same nominal band.

Dry the filament first. This is not a settings issue, but every temperature conclusion you draw from a wet spool is wrong, and PETG and nylon absorb moisture fast enough to invalidate a test between Monday and Thursday.

Where filament consistency actually bites

Here is the part that only shows up once you are printing in volume. A qualified profile is a set of assumptions, and one of them is that the filament diameter is what you typed into the slicer. Extrusion is calculated volumetrically. If the diameter moves, the amount of plastic laid down moves with it, and no sensor on the machine notices.

The failed part is not the expensive bit. The expensive bit is the investigation. When Thursday's parts fit and Friday's bind, the shop stops production and audits everything: nozzle wear, belt tension, ambient humidity, the new operator, the slicer update. That can eat two days. If the variable that moved was the filament, none of those checks find it, and the process gets re-tuned around a fault that will come back with the next spool.

That is the argument for buying against a published figure rather than a description of quality. We publish a dimensional tolerance of ±0.02 mm because it is a term you can hold a supplier to and check with a caliper in five minutes, which "high quality" is not. We also supply an SDS, a RoHS statement and a REACH statement per material for shops that keep documentation on file. Ask whoever you buy from now for their number. A supplier who will not state one is an uncontrolled variable in your process, and you will meet it eventually.

None of this makes a badly oriented part strong. Tolerance removes one line from the list of things that can move under a working profile. That is all it does, and on a production floor that is worth paying for.

Qualify the part by breaking one

A functional part is qualified when you have destroyed one on purpose. Print two extra, load them the way the real part gets loaded, and take one to failure.

Then look at the fracture. A clean flat break along a layer line means adhesion, which points at temperature, cooling or moisture. Plastic that stretched and tore before letting go means the material was working properly and you need more material, a better orientation, or a different polymer. Those two conclusions send you to opposite ends of the shop, which is why guessing is expensive.

Do this once per profile change and once per new batch on parts that matter. It costs an hour. It is the only method that tells you whether this week's output matches last month's, and it is the difference between a shop that makes functional parts and a shop that hopes.

Quick reference

• Write the duty cycle down before choosing a material
• Choose for failure behaviour, not headline tensile strength
• Orient so load runs along layers, even if it costs support material
• Four to six perimeters, 20 to 30 percent infill, five top and bottom layers
• Fillet internal corners, use inserts for anything torqued
• Run hot and cool less than a cosmetic profile would
• Dry the spool before drawing any conclusion from a test print
• Break one part on purpose per profile change and per new batch

Frequently asked

What is the best filament for functional parts?

PETG for most handled or load-bearing parts indoors, ASA or ABS where heat and sunlight are involved, PLA where stiffness and dimensional accuracy matter more than toughness, TPU where the part needs to flex. There is no single answer, which is why the duty cycle sheet comes first.

Is PLA strong enough for functional parts?

Often yes, for jigs, fixtures and enclosures that live indoors at room temperature. It is stiff and prints accurately. It creeps under sustained load and softens around 55–60 °C, so keep it away from clamped assemblies, engine bays and sunny windows.

Does higher infill make a stronger part?

Less than you expect. Stress concentrates in the outer walls under bending and torsion, so perimeters buy more strength per minute of print time. Raise infill after you are already at five or six walls, not before.

Why do my parts fail along layer lines?

Layer adhesion, which comes down to nozzle temperature too low, part cooling too high, wet filament, or the part being oriented so the load pulls layers apart. Check orientation first, then dry the spool, then raise temperature and cut the fan.

How do I know a printed part will last?

Test it to failure under the real load, not a bench guess. Repeat it after any profile change or new material batch. Accelerated heat and load soaks are worth running on anything that will be permanently clamped, since creep does not show up in a five-minute pull test.

Browse VANYO materials, or get in touch if you are qualifying a second supplier.