Everything you need to know about #Filaween filament testing!


Filaments need testing. Many, many manufacturers claim to be “the best” or “twice as strong as ABS” – but how do you verify those claims? Sure, there are the full-on scientific approaches like EN ISO 527-1 or ASTM DD638, but not only are these quite challenging to perform correctly, but also only have limited use for characterizing the vast properties of polymers and specifically 3D-printed plastics. So i’ve compiled an assortment of tests that look for the performance specifically of actual, 3D printed parts, with all their complexities in geometry and polymer weirdnesses. Yes, they don’t yield a full set (not even close) of parameters you’d use for designing injection molded parts, but should characterize an actual 3D printed part fairly well when it comes to the strength of a particular filament.

Of course, print quality needs verification, too, but thankfully Make: Magazine have already published a set of test parts for testing the accuracy and performance of a particular setup. Since there already is very little difference in the results of “normal” filaments, these test and the respective print settings were chosen to make for a challenging set of tests. Every filament will still have some room for tuning out any particular shortcomings we ran into, but this will usually result in a degradation of the print elsewhere.



Out of the Make: magazine 2015 printer shootout test files, i’ve decided to use the “Fine details”, “Overhangs” and “Bridges” test parts, as these are most influenced by the filament used. I’m also printing a 3DBenchy to judge real-world print quality. Each test part is printed in sequence and then rated according to the Make: guidelines, or, in the 3DBenchy’s case, by the following scale:

1: Unusable print
2: Most features printed with significant degradation
3: Usable print with flaws
4: Mostly flawless print without major issues
5: Flawless, as much as FFF 3D printing allows

Scores from the three Make: tests and the individual 3DBenchy rating are summed up for a total score from 4 to 20.

Ease of use

Additionally, an “ease of use score” is given based on the following parameters:

1: Not printable with commonly available hardware
2: Specialized hardware for best results, e.g. heated build chamber
3: Some hardware upgrades required, e.g. all metal hotend, enclosed build chamber, specialized build surface
4: Some tuning necessary, reasonable requirements for use or storage
5: Profiles commonly available, good prints easy to achieve with most setups

This score is not included in Filaween videos, but if there are any special challenges with a material, they will be mentioned.


3D printed test part is being bent by hand
3D printed test part is being bent by hand

The test parts are custom designed and modeled in Onshape.

Each specimen is printed once vertically and once vertically to test for either layer strength or layer adhesion.  For the bend test parts, the chuck is clamped to a solid surface, the test specimen placed all the way inside and loaded at the handle until it breaks. For the vertically printed part, the deflection at 1kg (9.81N) is recorded.

Pull test rings are suspended from a suitable solid support beam and loaded at the opposite end until they break. For the vertically printed version, the top and bottom corners should be aligned as printed. These tests are helpful where the tested polymer is too soft or too tough to fail the bend tests.

A combined strength score is calculated to factor in deviations in performance on each test evenly. The factors are as follows:

  • Bend test, Layer strength: Factor 2.5
  • Bend test, Layer adhesion: Factor 5.0
  • Pull test, Layer strength: Factor 1.5
  • Pull test, Layer adhesion: Factor 2

Each result can contribute a maximum of 32kg to the total score, which caps the rating for samples i simply can’t break without additional machinery.

Temperature resistance

3DBenchy temperature test under boiling water
3DBenchy temperature test under boiling water

For getting an approximate feeling for a filament’s temperature stability, boiling (100°C) water is poured over the previously printed 3DBenchy and a score is assigned by the following criteria:

1: Melted or deformed by just the water itself
2: Soft, “chewing gum” like consistency, easy to deform
3: Some rigidity retained, still realtively easy to deform
4: Mostly full rigidity retained, can be permanently deformed with some force
5: No impact

Print parameters

All prints are done on an Original Prusa i3 MK2 using an E3D v6 hotend. For abrasive materials, a nickel alloy coated nozzle from Mapa refining is used.

Where applicable, the default settings from the Prusa Research build of Slic3r are used. The Slic3r build can be downloaded and used for any open-source 3D printer, even using the default print settings. Print temperatures are listed in the full datawall below!

The most important print settings are:

  • 0.2mm layer height
  • 2 perimeters
  • 4 top solid layers
  • 3 bottom solid layers
  • 10% grid infill
  • 20mm/s to 60mm/s depending on the area of the print
  • 120mm/s travel moves
  • 0.45mm extrusion width
  • 15% infill-perimeter overlap
  • 95% bridge flow rate
  • 0.8mm retracts

All videos

The full list of #Filaween videos can be found on Youtube.

Full filament list and results

All tested filaments and filaments in testing as well as their test results are listed live on on Google Docs.

You can support me without spending a single penny!