TESTING 3D printed INFILL PATTERNS for their STRENGTH


When you prepare a model for 3D printing you’ll
always need to choose an infill structure for your part. Since current slicers give you the option
of choosing between a whole lot you might have already been asking yourself which one
is the strongest and most suitable for your application. So in todays video we’ll crush some test
samples to find out which performs the best and I was especially interested in the much
hyped gyroid infill you might have heard of. Guten Tag everybody, I’m Stefan and welcome
to CNC kitchen. In a previous video I’ve already been taking
a look at the strength of different infill structures that you can find in Simplify3D
and also talked about why it is more efficient in many cases if you increase the number of
perimeters instead of increasing the infill percentage. If you haven’t seen it yet, there is a link
in the description. I wanted to do a bit more research in this
direction and especially find out how different patterns behave in different directions, because
often you can’t orient your part in a way that it is perfect for the chosen infill. In addition, Slicer Prusa Edition features
3 3 dimensional infill structures, most prominently gyroid infill, that claim to have more isotropic,
so more uniform strength behavior in all loading directions. In order to perform the tests for this video
I had to adapt my tensile testing machine in order to be able to perform compression
tests. If you are interested in the machining process
on my CNC there is also a link in the description to that video. I’ve been thinking for quite a while about
a suitable test specimen. At first I thought of just printing the infill
pattern without any exterior walls but since the infill in your parts will always be surrounded
by a shell of material which is reinforcing the structure I ended up printing cubes of
22mm side length that use 1 perimeter and 3 top and bottom layers at 0.15mm layer height
which results in a constant wall thickness of 0.45mm. This gave me the possibility to use the same
specimen for the tests where I crushed the blocks from the top as well as from the sides
and therefor make the results more comparable. I’m well aware that this shell will add
additional strength to the structure, so we are not purely testing the infill, but since
it is the same for all we can still compare results. I haven’t tested all of the patterns available
and tried to concentrate on the, in my opinion, most relevant ones which are Rectilinear,
Grid, Triangle, Cubic, Line, Honeycomb, 3D honeycomb and Gyroid. The way how the parts will probably fail in
such a compression test is usually not compressive yielding of the structure but the stability
failure buckling of a wall which then leads to more load on the rest of the part and therefore
fatal failure. The shape of the infill or even how the infill
supports the external shell will determine how and at what load the structure fails. The slenderness of the structures will play
a huge roll and curved structures like for example in the gyroid will act like ribbing
and improve strength. I am a little limited by my test machine,
because it unfortunately stalls at around 3.5kN or 340kg of load so I only used an infill
ratio of 10%. Unfortunately as I also already talked about
in a previous video, 10% infill does not results in the same amount of material used for each
of the samples. Most notably a part with 10% 3D honeycomb
infill weights almost double one with 10% rectilinear infill. Either that’s a bug or the generation algorithm
is strange. Anyway, so at first I printed a batch of samples,
all at 10%, weight them and then adjusted the infill ratio slightly so that all samples
roughly weight the same in the end that I used for testing. Before we take a look at the strength of the
infill patterns lets at first take a look at the printing time, because this is also
something that is well relevant. Most of the infill patterns take around the
same amount of time to print, only the honeycomb structures take way longer due to the constant
change of printing direction where the printhead has to accelerate and decelerate constantly
which wastes a lot of time. Also at least I have the fear that my printer
rattles itself apart after a while when I use these patters. Interestingly, the gyroid infill, that also
consists out of a lot of tuns, doesn’t take much longer to print than the other more standard
patterns due to the nice, smooth turns that the printhead does, where accelerations are
much smoother so a kind of constant line velocity can be kept. I printed all of the samples in Formfuturas
Premium PLA on my original Prusa i3 Mk2s. Now let’s come to the destructive tests. I’ve been testing each patter once in the
build direction and once from the side. Let’s start with the pattern I mostly use
for my prints which is rectilinear infill. The first sample failed at 309 kg of load
whereas if you load it from the side, it was only able to bear 191 kg, which is almost
40% less. Keep in mind that these samples are only 22mm
in each dimension, so this amount of load is already quite impressive. Unfortunately, the cubes of the Grid, Triangle,
Line and Honeycomb patters were too much for my machine and it stalled for all of them
at around 350kg of load. Taking a look at the samples still already
showed that they were just before failing because the outer walls already showed buckling
marks. So for these samples we don’t really have
an exact value to compare against but if we still look at the transverse strength the
other loading direction is at least 30 to 45% weaker. The interesting results are now the 3D infill
patterns where all performed very comparable in the Z direction. If we also take a look at the strength 90°
rotated, cubic and gyroid infill perform exactly the same just as claimed only the 3D honeycomb
performs about 15% worse. So which is now the best infill you can chose. Well, that depends. I’m still a huge fan rectilinear infill
if your parts don’t require a huge amount of strength. It prints almost the fastest and it’s the
infill that has the densest structure which is great for your top layers so reduces pillowing
even at low infill ratios. If we take a look at the numbers, the line
infill pattern might even be better, since it has higher strength values, prints faster
and the infill spacing is almost as dense as with rectilinear. If you have a part that is purely loaded in
compression honeycomb might still be the way to go, but prints quite long and might rattle
your printer apart. Triangle infill might be an alternative, maybe
a little weaker but prints much faster. For the 3D infill patterns, 3D honeycomb is
dead in my opinion. Cubic and Gyroid infill perform pretty much
the same just the infill spacing is a little denser with the new contestant and come one,
it definitely looks cooler. So in summary, I probably might be using line
infill for any esthetic parts and quick prints in the future and anything that will see some
load, Gyroid infill will be my new way to go due to its good and uniform strength in
all directions. Now I’d really like to know from you guys
if I’ve missed something and which investigation you’d really like to see in a future video,
just leave a comment down below. I hope you like the video and learnt something. Please leave a thumbs up. I spend quite a lot of time in my research,
so if you want to support the making of these videos than consider becoming a patreon and
take a look at the affiliate links down in the description. Subscribe to the channel if you haven’t
and take a look at my other videos. Thanks for watching, auf wiedersehen and I
hope to see you around next time.

100 Replies to “TESTING 3D printed INFILL PATTERNS for their STRENGTH”

  1. Hi, Stefan, for the stronger samples you might be able to "pre-load" your machine with 200 Kg of "dead weight" so that you add 200 Kg to the force your machine can apply.

  2. Recommendation for a monitor mount? I'm printing a VESA monitor mount but don't know how much infill or pattern I should do. (My monitors weigh 9-10lbs)

  3. Awesome work! Thanks a lot for sharing. By the way, could you please share your work on developing the mechanical testing machine?

  4. My question isn't related to 3D printing of this type, but instead of working in this type of material where you want an object to be strong and not breakable, you were working with ultra-flex polymer, so making an object you want to compress and return to its original shape without it wearing out easily, wouldn't the honeycomb structure be superior for that, because a hex is closer to a circle?

    But something about your test here, if a honeycomb is known for relatively high out-of-plane compression properties and out-of-plane shear properties, doesn't this kind of compression of a cube test mainly planar forces? I guess I'm confused because I thought a honeycomb pattern would be significantly stronger than grid. Maybe the excessive gyrations of the machine make for a weaker structure?

  5. you tested compressive strength, what about tensional strength? also for the 3d infill (i use gyroid all the time) what about prints that do not fullying allow the 3d structure. for instance a part that is only 4-6mm in thickness? does gyroid maintain it's strength in a situation where it's structure is not fully realized?

  6. video: "Rectilinear is a good pattern, has good strength and low print times, and has a small distance between lines to improve top layers"
    me: "yea but gyroid looks sick"

  7. hello! what simulation did you use simulate a compression test in the simulator? and what simulator did you use? thanks

  8. You have to take in the considering of shape of the object being printed also, not only infill pattern. Honeycombs for example create much better adhesion to the walls if you have a rounded object and will also support angles better if they are rounded

  9. My first printer is on order and I'm gathering as much useful information before I start. I had no idea that there would be such a large difference in print speed depending on the style of infill using the same about of material. Thanks for taking the time to do this, it's very helpful.

  10. Maybe testing with your cubes turned 45 degrees, or whatever looks like would allow you test the theoretically weakest axis of the infill.

  11. I'm Dutch and it's really easy to follow your explanation in this video! so compliments for that!
    and this video was very interesting to watch and this subject is very important for 3D printing.
    thanks a lot for this video👍

  12. I'm mostly using cubic infill in the Cura slicer. Decent strength in all directions for non-structural parts. (mini figures for games.)

  13. And people say 3D printed parts are fragile, these cubes seemed pretty durable to me. This is very interesting information, choosing the right infill pattern for the direction the part will get the load from seems to significantly affect the strength.

  14. I apreciate the amount of research you do in 3D printing. I would recommend that you try to match the video and the audio a little bit more. For example when you draw your conclusions in the audio, the video shown is a random printing scene which is really confusing. When you name something my brain tries to match the words with what is shown on the screen and sometimes there is no possible match. Also the chart shown are not sorted out by performance and you don't highlight the particular type you are talking about. The chart could also include a small image along with the name. A combinaison of the highlighted chart along with a visible picture of the particular type you are talking about or you holding it and the test scene on the same frame could be easier to understand. I' m complaining as a youtube viewer , I'm not a 3D printing fan or a youtube content producer so I don't know how much more work that represents.

  15. Look like triangle is the winner to me? Only a little strength loss compared to gyroid in the transverse, but way higher perpendicular.
    Interesting though, I had been using 3d honeycomb as my "go to" – because you know, honeycombs are strong right? But wow.. turns out it's weak and super slow to print. Time to switch!

  16. The most annoying part of 3D printing is breaking away the infill. I've had to do this on many hundreds of prints because designing parts with minimal or no infill at all is not always an option. The best breakaway method I've seen so far is to use PLA for infill and ABS for the main part, but that is insanely expensive on my budget.

  17. Ok so cubic for all around strength then triangle for perpendicular strength or line for alternative to triangle for speed.

  18. In Cura, there's not "Rectilinear" option that I can see.
    Perhaps it's there, just named differently.

    Do you know, by any chance, what the Cura equivalent might be?

  19. Thanks for this video Stefan! I have been using honeycomb since I got my Prusa, since it looked strong, but seeing the print time compared to other infill methods I’ll be switching now!!! As you said, rectilinear for aesthetic parts and gyroid for strength seems like a great compromise between speed/strength! Thanks again for your work testing!

  20. Das "Guten Tag" hat mich so verwirrt..😂 Danach klang es einfach so, als würde er Deutsch reden..🧐

  21. Your dedication to a rigorous scientific testing method, even ensuring they are all roughly the same mass, is fantastic. Keep up the great work!

  22. This video is very hard to watch because it feels like everything is just jumbled together so I didn’t really know what exactly he was talking about at times

  23. I just used gyroid on a tall print and at a certain point it caused a gap almost like support material breakaway and the large part tore in half like paper. The gap was uniform throughout that whole layer so i don’t know if it was a mechanical or software malfunction. Sliced on cura at 15% .2 mm layer wall thickness .4

  24. hello! do you know what effects of different layer heights would have on crack propagation? cheers.

  25. 3:50 No, you shouldn't adjust the infill percentage to normalize the weight across all samples. Weight is not the main contributor in material strength, infill percentage is! How you justify that a 9% infilled gyroid is weaker than 10% rectilinear infill? It is not an apple to apple comparison! You should keep the infill ratio the same and let the weight fluctuate, because every infill pattern has their own weight efficiency. Then if you compare with constant infill ratio, you can do compare in strength to weight ratio to further justify which one is more economical for prints that need compression resistance.
    Also, if your press cannot exceed 340kg of press, use less than 10% infill or on slightly bigger sample due to size effect on material that can slightly weaken it's shear capacity, which is needed for your limited machine.
    I didn't consider your results are valid, because the methodology itself is of loophole. If you are my student, I may request for major correction on you.

  26. I'd like to see star infill compared to the others, in slic3r prusa edition. It's stronger that rectilinear because each layer stacks on top of each other instead of every other layer.

  27. A test of bending strength might be interesting – print a beam and load it in the middle. This tests both tensile and compressive strength.

  28. For such small pieces the influence of the perimeter might be too high. You should either remove it or use larger pieces. At least try a couple perimeter-less cubes with different patterns to find how the results scale, whether their ratio remains similar. Anyway, thank you for your tests.

  29. Loved this video, I've just gotten some TPU and would like to see what infills give the part the most "squish" in perpendicular to the layers and also transverse. I was thinking about using gyroid since that would give me even flexibility. What do you think?

  30. Mich würde interessieren wie man das Filament über Jahre nutzbar hält bzw schon überlagertes Material wieder auffrischen kann. Ich habe seit 2012 meinen ersten 3D Drucker und manche PLA Drähte von damals funktionieren, andere zerbrechen wenn ich sie ein paar Tage nicht vom Drucker trenne ganz von selbst. Angeblich ist PLA ja UV beständig und relativ unempfindlich bei Feuchtigkeit. Also warum diese Brüchigkeit? Und kann ich mit Wärme, Wasser oder Lösungsmittel die Rollen noch retten?

  31. This may sound counter intuitive, but retry your test with your top three choices on a cube that is larger. You may find they fail while the smaller ones don't (less surface effect).

  32. "You will always need to chose an infill structure for your part." The first thing out of your mouth was wrong. Nice try, you don't need one when you have 100% infill.

  33. Excellent investigation. Compression from the top versus compression from side to side is a good comparison. In practice, shear can also occur or compression from diagonally opposite edges. I've been using cubic (I use Slic3r) which seems like a good compromise with strength and print time. Sometimes using fillets between external faces works if U- or V-shaped limbs are pulled apart like the wishbone on a chicken. Again, thank you for your research and keep up the good work! Best Regards, Tubewaller

  34. This is really close to a good video, you have a lot of good information here.

    But there's no attempt to control for or quantify machine or material variation, which in practice are absolutely huge, so your conclusions aren't actually useful.

    That's… unfortunate, because they're close enough to being useful that I think most people will just take them as true and move on.

    My suggestion would be, in the future, to add more variation to your tests to try and determine variability under different printing conditions. Print multiple test specimens with multiple different materials, and also possibly vary your print temperature to force the print to have better and worse adhesion.

  35. What slicer did you use. I use cura and cura foes not have rectilinear so I instead use grid. My infill amount is 15%. is that too much?

  36. A niche application that I think cubic might excel over all the others is flotation. Barring printing issues (i.e. under extrusion from trying to drive volumectric flow higher than the hot-end can support), the cubic infill seems to generate a closed cell type of infill. All the others are create an open cell structure where if a perimeter is breached water can flow through the infill.
    But, that is a very specialized corner case (though all you people printing RC boats might want to use cubic infill).

  37. Booo booo comercials 👎👎👎👎 I delete Your Channal from my playlist 👎👎👎 Greed kils the Youtube experiance!!!

  38. I have ordered only 1 3D part by now, but find the technology interesting. I just wonder if those infill patterns need to be uniform?
    I mean just take a look at the inner structre of bones for example. They have a solid outer layer called "stratum compactum", but what makes bones lightweight and durable at the same time is the "stratum spongiosum" which is actually the infill of the bone and is everything but uniform.
    I wonder if 3D printing software have such a feature and if so, does it make a difference to print non-uniform infills in terms of weight/durability ratio?

  39. The hex infill might still be useful for flexibles where you would want less strength to retain flexibility.

  40. Thank you for the great video. Would love to see a test like this done for cantilever. Thinking of 3-D pattern might be best but not sure.

  41. Did you (or are you planing to) do a test with more fragile structures like pipes/bridges?
    Those are usually the parts that break easiest in any 3d model and would definitely provide better comparability since all infill patterns would break below your stalling threshold. (i.e. an experiment "breaking a stick in the middle" instead of pressing a block together)

  42. Initiate destructive testing.

    Product exceeds capacity of test equipment and is not destroyed.

    Maximum overwin!

  43. Excellent and informative video. I shall definitely be using gyroid in future. I would like to see something on printing ABS , namely how to avoid stress cracking / layer separation . I'm doing quite a lot of ABS because of the ease of post-processing to a nice finish, but still struggle with layer separation despite having a good enclosure around the printer. I have managed to minimise this by slicing components into shallower sections and ABS allows for good glue assembly.
    Thanks for posting useful and methodical content.

  44. Sehr interesantes Video 👍😊Kannst du bitte die gleichen Infills nochmal im Zugversuch testen, Stefan? 🙂

  45. gyroid is my go-to infill for most of my usable things. It seems to have less "tears" in between layers as the extruder as the extruder moves over. And it looks pretty slick when printing 😉

Leave a Reply

Your email address will not be published. Required fields are marked *