Skip to content
Spool & Slice

Specialty & Engineering

Best filament for cosplay armor, helmets and props

Print the bulk of the armor in cheap PLA+, the helmet in a foaming lightweight PLA, and anything that has to bend in TPU. The three decisions that actually matter are weight on the body, the surface you are going to paint, and heat.

How this page is funded. Buy links go to Amazon and we earn a commission on qualifying purchases at no extra cost to you. Commission never decides an order or a score, we accept no free products or sponsored placements, and every roundup names something to skip. Full disclosure.

The picks, ranked

The picks, ranked
#ProductBest forScorePrice
1
Spools of PLA filament stacked on a workshop shelf (illustrative photograph of the product type)SUNLU PLA+A chest plate is several kilograms of plastic before it is a costume. Buy the cheapest spool that is tough enough to survive being sanded and handled.185-235 C nozzle, no enclosure
The bulk of the armor8.9/10Check price on Amazon#ad disclosure
2
Spools of PLA filament stacked on a workshop shelf (illustrative photograph of the product type)colorFabb LW-PLAThe only foaming PLA here that publishes what its density becomes, which is the number that decides whether a helmet is wearable for eight hours.0.403-0.476 g/cm3 fully foamed
Helmets and anything worn on the head8.6/10Check price on Amazon#ad disclosure
3
A partly used spool of translucent filament beside a 3D printer (illustrative photograph of the product type)Overture PETGThe hinges, buckles and mounting tabs are what break at a convention. PETG takes the abuse PLA snaps under.215-270 C nozzle, noted for impact resistance
Straps, mounts and parts that get handled8.4/10Check price on Amazon#ad disclosure
4
Spools of PLA filament stacked on a workshop shelf (illustrative photograph of the product type)Polymaker PolyTerra PLAA matte surface hides layer lines under primer, so the same finish takes fewer coats and less sanding than a glossy spool.Matte finish, standard PLA price
Surfaces you are going to paint8.1/10Check price on Amazon#ad disclosure
5
Spools of PLA filament stacked on a workshop shelf (illustrative photograph of the product type)SUNLU LW-PLAPublishes a weight saving but no density, no foaming temperature and no tolerance. Fine for a first lightweight print, thin on data.About 30% lighter, 0.8 kg spool
A cheaper way into foaming PLA8.0/10Check price on Amazon#ad disclosure
6
A functional 3D printed part in a dark engineering filament (illustrative photograph of the product type)SUNLU TPU 95AThe parts that flex where the costume moves. Budget several times the print time of the same part in PLA.95A, roughly 20-40 mm/s
Joints, straps and soft edges7.8/10Check price on Amazon#ad disclosure
7
A functional 3D printed part in a dark engineering filament (illustrative photograph of the product type)Bambu Lab PLA-CFCarbon fill buys stiffness you do not need, wears your nozzle, and leaves a fibrous surface that fights primer. It adds no heat resistance either.Hardened nozzle required, still PLA underneathSkip this
Not for a piece you plan to finish3.2/10Check price on Amazon#ad disclosure

No live price is currently verified for these products, so every button reads "Check price" rather than a number we did not fetch. Scores are our own fit rating for the stated use, not a measurement and not a customer rating.

Painted miniature figures in close-up, showing a hand-finished paint surface
Painted miniature figures in close-up, showing a hand-finished paint surface.

What a cosplay build asks of a filament

Cosplay is not one printing problem. A pauldron, a helmet, a hinge and a soft under-strap all fail differently, and the material that is right for one is wrong for the next. Five criteria decide every choice on this page, in this order.

  1. Weight on the body. Anything worn above the shoulders is a weight problem before it is a strength problem. This is the one criterion where a specialty material earns its price.
  2. The surface you are going to paint. Almost every printed costume gets primed and painted, so the finish that matters is the one under the primer, not the one on the spool.
  3. Heat it will sit in. Not the convention hall — the car it rides in, the trunk it waits in, and the stage lights.
  4. Whether the part has to flex. Straps, joints and anything that clips on repeatedly are a different material entirely.
  5. How much of it you need. A full armor set is measured in kilograms. A material that is fine for one bracket can be indefensible across six spools.

Weight is the constraint people meet last

A helmet printed in ordinary PLA is comfortable in the living room and intolerable by the afternoon. Lightweight foaming PLA — sold as LW-PLA — exists for exactly this, and it is the one material choice on this page that changes what a costume can be.

The mechanism is real and colorFabb, who developed the material, publish the numbers for it. Their LW-PLA prints across a range of 195 to 260 C and starts foaming at around 230 C, which expands the extruded volume by nearly three times colorFabb. Density falls from a published 1.210 to 1.430 g/cm3 unfoamed to 0.403 to 0.476 g/cm3 fully activated colorFabb. That is not a marketing figure; that is roughly a third of the mass for the same shell.

SUNLU sell a cheaper foaming PLA and publish a coarser claim: about 30% lighter for the same model, on a 0.8 kg spool, with a matte surface they describe as easy to paint SUNLU. They do not publish a density, a foaming temperature or a diameter tolerance, so the tuning is yours to find.

Two things worth knowing before you commit a helmet to it. Foaming is controlled by nozzle temperature, so the flow rate that worked at 200 C is wrong at 240 C and the same spool behaves like two different materials. And a foamed wall contains less plastic than a solid one of the same thickness, so it dents where solid PLA would scuff. Print the shell in the foaming spool and the mounting points as separate solid pieces, then join the two.

Heat: the car, the trunk and the stage lights

The most common way a printed costume dies has nothing to do with the convention. Prusa Research are explicit that PLA gets soft and deforms at temperatures over 60 C Prusa Research, and a closed car in summer passes that comfortably. Lightweight PLA is no better and arguably worse: colorFabb publish a glass transition temperature of 55 C for LW-PLA colorFabb.

This is why the pieces that carry load — pauldron mounts, helmet hinges, belt hardware — are worth printing in PETG even when the shell around them is PLA. Prusa list PETG at 215 to 270 C nozzle with a 70 to 90 C bed and note it for impact resistance Prusa Research, and it does not need an enclosure to get there. For the full comparison, see PLA versus PETG.

If the costume is for an outdoor shoot that will sit in direct sun for hours, the material logic changes again and ASA becomes the honest answer — the same reasoning as outdoor and UV parts. For a convention floor, PLA+ and PETG are enough.

The surface you are going to paint

Costume parts are judged at arm's length under bad lighting, which means layer lines matter more than dimensional accuracy. The finishing sequence — sand, fill, prime, sand again — is where the hours go, so the useful question is which spool shortens it.

The parts that have to bend

Every armored costume has soft components: strap ends, joint gussets, the edge trim that stops a plate from cutting into a forearm. Printing those in PLA is why costumes fail on the second wearing.

TPU 95A is the standard answer and it is slow. Prusa's flexible materials guidance puts printing at roughly 20 mm/s with 30 to 40 mm/s as the ceiling Prusa Research, so a part that takes two hours in PLA can take six. Budget the time before you design the part, not after. The full material page is best TPU filament.

Printing pieces bigger than the build plate

Nearly every meaningful cosplay part is larger than the plate, so the build is a splitting-and-joining exercise. Two consequences follow.

  1. Warping scales with the footprint. A large flat piece is the geometry warping punishes hardest Prusa Research, which is a large part of why PLA remains the right answer here: Prusa's material guide lists low warping as its noted property Prusa Research. Splitting a piece into two shorter sections reduces the contraction force fighting your bed.
  2. First-layer adhesion decides the day. A twelve-hour armor plate that lifts at hour nine costs a spool and a day. Wash the plate, then read bed adhesion and build plates before you buy anything else.

Design the joins as mechanical joints — dovetails, pin holes, screw bosses — rather than trusting glue alone on a piece that will be worn, bumped and packed into a suitcase.

What we would not buy for this

How much to buy, and in what order

A full armor set consumes several kilograms, and buying it one spool at a time is the most expensive way to do it. Work out the cost per kilogram rather than the price per spool — the arithmetic is on price per kilogram — and check buying in bulk before a large build, because color consistency across a set is a real reason to buy the whole run at once.

One caution on storage: a build that runs for weeks means spools that sit open for weeks. PETG and TPU both take up moisture from room air, and a spool that printed cleanly in week one can string and pop in week four. Sealed storage costs less than one ruined armor plate; the numbers are on how to store filament.

Every pick, in detail

01

SUNLU PLA+

SUNLU

The bulk of the armor · 8.9/10 our fit score

Spools of PLA filament stacked on a workshop shelf (illustrative photograph of the product type)

Illustrative photo of the product type, not this exact item.

The cheapest filament we would load without flinching, and the reason most people never need to spend more.

Published specifications for SUNLU PLA+
Diameter1.75 mm, published tolerance +/- 0.02 mm
Spool weight1 kg
Nozzle185-235 C (PLA range) Prusa Research
Drying50 C / 7 h (PLA guidance) Overture

What it does well

  • Consistently the lowest cost per kilogram among brands that publish a tolerance figure
  • The PLA+ formulation is noticeably less brittle than plain PLA on thin parts
  • Stocked widely enough that a reorder rarely means a different formulation

What it does not

  • Color-to-color variation is real; a temperature that works for black may not work for white
  • "PLA+" is a marketing term, not a standard — nothing about it is defined across brands
02

colorFabb LW-PLA

colorFabb

Helmets and anything worn on the head · 8.6/10 our fit score

Spools of PLA filament stacked on a workshop shelf (illustrative photograph of the product type)

Illustrative photo of the product type, not this exact item.

The foaming PLA that publishes its own density figures, which is the only way to know what the weight saving actually is before you print a helmet in it.

Published specifications for colorFabb LW-PLA
Printing temperature195-260 C, foaming starts near 230 C colorFabb
Density unfoamed1.210-1.430 g/cm3 colorFabb
Density fully foamed0.403-0.476 g/cm3 colorFabb
Glass transition55 C colorFabb
Bed (optional)50-60 C colorFabb
Diameter tolerance+/- 0.1 mm colorFabb

What it does well

  • Publishes both the unfoamed and the fully foamed density, so the weight saving is a number rather than a claim
  • Foaming is controlled by nozzle temperature, so one spool covers a stiff base layer and a light shell
  • colorFabb state the volume can expand nearly three times, which stretches a spool across a large wearable

What it does not

  • Priced as a specialty material — several times the cost per kilogram of budget PLA+, and armor is bought by the kilogram
  • The published tolerance is +/- 0.1 mm, looser than any ordinary PLA on this site
  • Still PLA underneath: colorFabb publish a 55 C glass transition, so a hot car ruins it exactly like any other PLA
03

Overture PETG

Overture

Straps, mounts and parts that get handled · 8.4/10 our fit score

A partly used spool of translucent filament beside a 3D printer (illustrative photograph of the product type)

Illustrative photo of the product type, not this exact item.

The PETG most people should start with: forgiving on temperature, widely stocked, and cheap enough to learn on.

Published specifications for Overture PETG
Diameter1.75 mm, published tolerance +/- 0.02 mm
Nozzle215-270 C (PETG range) Prusa Research
Bed70-90 C Prusa Research
Drying65 C / 7 h Overture

What it does well

  • Overture publish an explicit drying temperature and time for their own PETG
  • Tolerant of a wide temperature window, which matters because PETG punishes a bad guess with stringing

What it does not

  • Strings more than the premium spools at the same settings
  • Bonds hard to smooth PEI — Prusa's PETG page says "Do not print on the smooth PEI sheet as the adhesion may be too strong"
04

Polymaker PolyTerra PLA

Polymaker

Surfaces you are going to paint · 8.1/10 our fit score

Spools of PLA filament stacked on a workshop shelf (illustrative photograph of the product type)

Illustrative photo of the product type, not this exact item.

The matte finish that actually hides layer lines, on a spool priced with ordinary PLA rather than as a specialty.

Published specifications for Polymaker PolyTerra PLA
Diameter1.75 mm, published tolerance +/- 0.02 mm
FinishMatte
Nozzle185-235 C (PLA range) Prusa Research

What it does well

  • The matte surface diffuses light and genuinely reduces visible layer lines on curved parts
  • Priced alongside standard PLA rather than as a premium finish

What it does not

  • The matting agent is mildly abrasive over long runs; a hardened nozzle is cheap insurance
  • Slightly more brittle than the equivalent glossy PLA
05

SUNLU LW-PLA

SUNLU

A cheaper way into foaming PLA · 8.0/10 our fit score

Spools of PLA filament stacked on a workshop shelf (illustrative photograph of the product type)

Illustrative photo of the product type, not this exact item.

The cheap way into foaming PLA, sold in a smaller spool than it looks and without the published numbers colorFabb give you.

Published specifications for SUNLU LW-PLA
Spool weight0.8 kg SUNLU
Weight savingAbout 30% lighter than the same model in standard PLA SUNLU
FinishMatte, described as easy to paint SUNLU
Nozzle185-235 C (PLA range) Prusa Research

What it does well

  • Publishes a weight-saving figure, which most budget brands do not commit to at all
  • The matte surface is the one you want under primer, so a helmet needs less filling
  • Cheaper per spool than the original foaming PLA, which matters on a part measured in kilograms

What it does not

  • SUNLU publish no density, no foaming temperature and no tolerance for it — you tune the foam by trial
  • The spool is 0.8 kg, not the usual 1 kg, so the cost per kilogram is higher than the sticker suggests
  • Sold in white only at the time of writing, so plan on painting it
06

SUNLU TPU 95A

SUNLU

Joints, straps and soft edges · 7.8/10 our fit score

A functional 3D printed part in a dark engineering filament (illustrative photograph of the product type)

Illustrative photo of the product type, not this exact item.

The cheapest way to find out whether your extruder can handle flexible filament at all. Print it slowly.

Published specifications for SUNLU TPU 95A
Shore hardness95A
Nozzle230-245 C (Flex range) Prusa Research
Bed60-75 C Prusa Research
SpeedUsually 20 mm/s; 30-40 mm/s maximum Prusa Research

What it does well

  • 95A is the stiffest common TPU and therefore the most forgiving on a Bowden setup
  • Cheap enough that a first failed roll is not a real loss

What it does not

  • Hygroscopic — Prusa's flexible-materials page lists moisture absorption as an inherent property
  • The speed ceiling is low; a part that takes two hours in PLA can take six
07

Bambu Lab PLA-CF

Bambu LabSkip this

Not for a piece you plan to finish · 3.2/10 our fit score

A functional 3D printed part in a dark engineering filament (illustrative photograph of the product type)

Illustrative photo of the product type, not this exact item.

A stiffer, matte PLA with a carbon fiber filler. It buys rigidity and a finish, not heat resistance — and it eats brass nozzles.

Published specifications for Bambu Lab PLA-CF
Nozzle materialHardened required E3D
Base polymerPLA — softens above about 60 C Prusa Research
NozzleComposite range 225-290 C Prusa Research

What it does well

  • Meaningfully stiffer than unfilled PLA, with a matte finish that hides layer lines
  • Prints at PLA-like temperatures, so no enclosure is needed

What it does not

  • It is still PLA underneath: Prusa are explicit that PLA softens and deforms above about 60 C
  • Requires a hardened nozzle; Prusa's material guide flags one as required for composites

Common questions

What filament do most people use for cosplay armor?

PLA, and usually a budget PLA+ grade. It is the cheapest material that prints large flat pieces without an enclosure, and Prusa list low warping as its noted property. Armor is bought by the kilogram, so cost per kilogram decides more than any datasheet figure.

Is LW-PLA worth it for a cosplay helmet?

For anything worn above the shoulders, yes. colorFabb publish a density of 0.403 to 0.476 g/cm3 for fully foamed LW-PLA against 1.210 to 1.430 g/cm3 unfoamed, so a helmet shell can lose roughly two thirds of its mass. It costs several times as much per kilogram as ordinary PLA, which is why it belongs on the helmet and not on the leg armor.

Can I leave a 3D printed costume in a hot car?

No. Prusa Research state that PLA gets soft and deforms above about 60 C, and colorFabb publish a 55 C glass transition for lightweight PLA. A closed car in summer reaches both. The costume rides in the cabin with the air conditioning on.

PLA or PETG for cosplay?

Both, in different places. PLA+ for the shell, because it is cheap, prints flat and sands well. PETG for hinges, buckles, mounting tabs and anything that takes repeated stress, because Prusa note it for impact resistance. PETG is harder to sand smooth, so it is the wrong choice for a large visible surface.

Sources

Scooter M. · Enthusiast

I'm Scooter, an enthusiast who's genuinely into this. I read the manuals, compile the published specs, and do the math. No lab coat.

About · How we choose · Last reviewed