How We Tested 3D Printers

By:
David Wise and Austin Palmer

Last Updated:
Monday
October 2, 2017

Share:
There are infinite possibilities of items that can be made with a 3D printer.
There are infinite possibilities of items that can be made with a 3D printer.

We bought the top 3D printers available on the consumer market and put them through a grueling series of tests to crown the winners. We printed over 150 different objects comparing the print quality side-by-side, as well as conducting a handful of different tests and evaluations to rank each model's ease of use, printing capabilities, and their reliability, as well as how helpful the manufacturer's customer support was. Check out our complete review here to see which printers came out on top and which ones were the best bang for the buck.

Print Quality


The first and more important rating metric we used to evaluate these products was Print Quality. We printed a set of test model on each printer, using both ABS and PLA if possible, as these are the two most commonly used 3D printer filaments, or 3D ink. These models were selected to test a specific attribute of the printer or as an all-around print. A panel — made of people with varied experience levels, from novice to expert — then rated and scored each print without knowing which printer each model came from. The full set of test models for one printer is shown below.

The set of prints from the winner of our Print Quality metric  the Ultimaker 2+.
The set of prints from the winner of our Print Quality metric, the Ultimaker 2+.

We will go through each test model one-by-one, identifying a few of the key features that we looked for in the following sections.

Our fleet of jolly 3D Benchy tugboats.
Our fleet of jolly 3D Benchy tugboats.

3DBenchy


3D Benchy is a commonly known torture test for 3D printers and is a good starting point to compare different models. This little boat, designed by Creative-Tools.com, is designed specifically to push your 3D printer to its limits. This model has overhanging sections, fine details, bridging, as well as thin cylinders. There is much more detail about how to evaluate a Benchy at the above site but in essence, we were looking for clean prints — no glaring defects, solid bridges over unsupported areas, and a generally nice looking boat.

The bridging test  with each bridge getting successively longer as the printer progressed around the spiral.
The bridging test, with each bridge getting successively longer as the printer progressed around the spiral.

Bridging Test


All of the printers in our review are FFF — Fused Filament Fabrication — meaning that they struggle to print in midair. This test was designed to test the ability of each printer to bridge the gap, with each gap getting progressively larger on this model. We were looking for bridges that were clean, weren't sagging, and had a nice surface finish on the unsupported underside.

The support test before the sacrificial structure was removed.
The support test before the sacrificial structure was removed.

Support Test


While the previous test compared the midair printing abilities of each printer, this test looks at their supported printing capabilities. Support is when the printer will create a sacrificial matrix of material underneath an overhanging section. The printer will create a dedicated weak layer by leaving a gap between the support structure and the model, with the ideal situation being that the support structure can be easily snapped away when the print is complete, leaving a relatively nice surface finish on the underside of the supported area.

After removal  showing the difference in the finish between the supported areas.
After removal, showing the difference in the finish between the supported areas.

In reality, it is quite difficult to get this right, as easy to remove support usually requires too large a gap and leaves a poor surface finish on the model, while a small enough gap for a good surface finish can make it impossible for the support to be removed.

Removing support material can be a time-consuming process  as well as frustrating.
Removing support material can be a time-consuming process, as well as frustrating.

We compared our sample support prints by looking for a smooth surface finish on the underside and by timing how long it took to remove the support material, as well as rating how hard the removal process was.

The Eiffel Tower was one of the most difficult test prints we had.
The Eiffel Tower was one of the most difficult test prints we had.

Eiffel Tower


Unfortunately, none of the models we looked at really did justice to this iconic landmark. All of our sample prints were objectively mediocre at best so we scored based on if the model actually printed completely without breaking, if there were features left out, and how much wobble in the Z-Axis was present.

Our herd of print-in-place  articulated elephants. Some legs moved better than others.
Our herd of print-in-place, articulated elephants. Some legs moved better than others.

Elephant


This model is a fun one: A print in place, articulated model of the world's largest living land mammal. We looked for elephants with articulated limbs and head that were easy to break free and moved without difficulty. We also graded each print on the quality of the overhanging areas, as well as overall appearance and smoothness.

The unsupported overhang test  primarily ranked on the appearance of the underside of the angles.
The unsupported overhang test, primarily ranked on the appearance of the underside of the angles.

Overhang


Speaking of overhangs, this test focused exclusively on this tricky area for FFF printers. The test model consisted of successively steeper and steeper overhangs, maxing out at around 70°. We ranked the printers on how cleanly each one printed the underside of each overhang, whether it was smooth and consistent in angle, or rough and drooping.

We looked for smooth vertical faces with no Z-Axis wobble in these towers.
We looked for smooth vertical faces with no Z-Axis wobble in these towers.

Tall Tower


This print of a tall, skinny tower allowed us to evaluate the Z-Axis wobble in each printer. Any slop or misalignment in the printer's mechanism will present itself on this print, as each layer will be slightly offset from the previous ones, causing the sides of the tower to undulate rather than being smooth. We scored based on how smooth the sides were and how sharp the corners of the square tower were — whether they were crisp or rounded over.

Testing the dimensional accuracy of the printers  a nickel should fit perfectly in this print.
Testing the dimensional accuracy of the printers, a nickel should fit perfectly in this print.

Nickel


While the previous test evaluated how accurately the printers were in the Z-Axis, this test looked at how well each model did in the X and Y-Axis. This small print should have been flat and perfectly fit a nickel, without excessive force required to fit it in or an excessive amount of slop.

Another difficult print  this articulated platform jack prints in one piece.
Another difficult print, this articulated platform jack prints in one piece.

Platform Jack


This is another print in place, articulated model. We ranked each jack on its vertical smoothness, how freely the jack moved, the difficulty in initially cracking it free, as well as some overhanging and bridge sections. We also looked at the overall quality and appearance, as there was a stark difference between different models.

The wireframe or hollow cube  testing bridging and retraction.
The wireframe or hollow cube, testing bridging and retraction.

Hollow Cube


This wireframe model of a cube tests two different things: retraction, and bridging. Retraction is how much the printer will retract, or suck back filament when jumping across voids. An ideal cube would have four distinct towers, without string or excess plastic in the gaps between them. The top section of the cube is also unsupported, requiring the printer to bridge the gap — another method we used to score their performance with these models.

The set of delicate spiral vases  with less than a millimeter of wall thickness.
The set of delicate spiral vases, with less than a millimeter of wall thickness.

Spiral Vase


This decorative vase is printed using the spiral vase setting, where a solid model is uploaded and the slicer will use a special mode to print it as a thin wall vase with a single shell, the same thickness as the nozzle. We graded the printers on if they were able to actually perform this task, as well as how each print looked, mainly looking for layer separation and cracks.

We evaluated both the lid and jar produced by each printer  in PLA and ABS if possible.
We evaluated both the lid and jar produced by each printer, in PLA and ABS if possible.

Threaded Jar and Lid


There wasn't a specific trait that we evaluated in this pair of prints, just the overall quality of both pieces. We looked for smooth surfaces, solid layers, no separating layers, and a clean texture on the outside of the lid. We also checked the threads to make sure the lids screwed on smoothly and easily.

One of the low-poly models that we evaluated the printers on.
One of the low-poly models that we evaluated the printers on.

Low-Poly Figurines


These final two models were partially for fun but did highlight some performance differences between printers. These two, low-poly models of some popular characters have some overhanging sections, thin sections, sharp points, and fine details.

The other low-poly model that we evaluated them on.
The other low-poly model that we evaluated them on.

We ranked these prints on how they looked overall, paying particularly close attention to areas that we knew to be difficult, such as the steepest overhangs.

Ease of Use


While our massive Print Quality metric made up 40%, Ease of Use was close behind, accounting for 30%. After unboxing each printer, we compared the ease of assembling and initial setup. We looked at how long and how much work was required, if there were any special tools required, and how long it took, as well as the clarity of the directions. We also evaluated how long and how difficult it was to get the requisite software downloaded and installed, per the manufacturer's directions.

The Monoprice required a little bit of assembly.
The Monoprice required a little bit of assembly.

We then compared how difficult it was to level the print bed, giving bonus points to models that made it extremely easy or auto-leveled effectively. Then, we compared the connection method for each model. We checked if each printer had to remain connected to a computer, or if it could function remotely — whether it was compatible with standalone printing from a flash drive or SD card, or if you could print over a network connection. We also compared the display on each printer, if there was one. Finally, our last test in Ease of Use was the ease of swapping and loading filament cartridges and spool.

The Ultimaker had a decent set of capabilities  compatible with a handful of materials.
The Ultimaker had a decent set of capabilities, compatible with a handful of materials.

Print Capabilities


For this metric — making up 20% of the total score — we compared what these printers can actually do. We rated the ease of use and capabilities of each recommended software program for each model, as well as the ability to use other programs or upgrade to more sophisticated slicers.

Cura is the recommended slicer for the Ultimaker.
ReplicatorG was not our favorite program.
 

We then compared the build envelope of each printer as well as the quality of the build plate, noting if we had had any persistent bed adhesion issues throughout the testing process. We also awarded more points for models that were compatible with generic filaments, not limiting the printer to filament made by the manufacturer. Finally, we compared the hot end and cooling of each model. We did this by checking and comparing the temperature range, and thus the range of filaments each model was compatible with, and the number of layer cooling fans present.

Support


Our final metric accounts for the residual 10% of the score. To test the support and reliability of each printer, we contacted the customer support for each manufacturer, rating the contact methods and their overall helpfulness. We also compared the warranty included with each printer, as well as if there were any support videos available on the web.

Conclusion


While it can be exceedingly difficult to wade through the veritable ocean of information and 3D printers available, hopefully, this article will make it a little easier to see help you pick the best and give you some ideas about how to compare different models. Take a glance at our comprehensive 3D printer review to see which models excelled in these tests and which ones got flattened by the competition.

  • Share this article:
 

Follow Us



Related Review

Best 3D Printers of 2017

by David Wise and Austin Palmer

Unbiased.