INFILL pattern and SHELLS – How to get the maximum STRENGTH out of your 3D prints?


When it comes to the strength of your 3D printed
parts print settings play, besides the material a major roll in what your prints can handle. Unfortunately I hear and read way too often
that many only use the infill ratio to adjust that property. For this reason I tested the strength of different
infill patterns on tons of these hooks and more importantly show you how you really should
strengthen your prints! Guten Tag everybody, I’m Stefan and welcome
to CNC Kitchen. If you have your material dialed in you will
usually only adjust a small number of settings before each print depending on the requirements
of the part. If you want it to be strong, many will adjust
the infill density to a higher value. This will scale the infill patters density
and defines how much of the internal cavity is filled with material which can go up even
to a fully dense part. The other option is increasing the thickness
of the outer shell, by adjusting the shell thickness, wall thickness, number of perimeters
or however it is called in your slicer. Usually, for a strong print you’ll be adjusting
both values. But what really gives you the strongest result
and which is the most economical way, meaning how can you increase the strength of your
part with the least amount of material. If you want to print light weight structures,
this is something which is really important. In order to properly answer that question,
I printed quiet a lot of my trusty test hooks and measured their strength – for science! This is definitely not perfect and doesn’t
include all possible load cases and orientations, but I still think that this is a good real
world example and probably one of the best analysis you’ll find at the moment. The reason why I use my hook from my filament
test series for the investigation instead of standardized axial test specimens is, that
the critical area is loaded with a more every day load case. Since the location of load application is
a little offset, that lever arm will cause an additional bending moment that is superposed
with the axial force. In real life, your part will also mostly be
loaded by such a combination so the results I present, even though not 100% scientific,
will probably be more usable and proof my point better. As a first investigation I wanted to find
out what difference the infill patterns make on strength of the part and print time. I also wanted to find out if 30% infill really
meant that 30% of the internal volume was filled by material. Since I used Simplify 3D as a slicer for these
tests, I was trying out the following available patterns: Rectilinear, Grid, Triangular, Wiggle,
Fast Honeycomb and Full Honeycomb. Rectilinear is usually the default and prints
the fastest whereas Wiggle is probably more for esthetic purposes. Like for most of my parts, these hooks were
printed with only 2 perimeters, and 4 bottom and 5 top layers at 0.2mm layer height. All the parts were printed in PolyLite PLA
by the way. Print time of the Rectilinear infill was the
shortest with 48.5 minutes. Grid infill took 50.5 minuest and triangular
53. The patterns with lots of direction changes
took the longest with 55 minutes for the full honeycomb and interestingly wiggle and fast
honeycomb took the longest needing 56.5 minutes. This is not fully representative, since the
times change due to the geometry and ratio between outer walls and infill, but shows
that you can definitely save some time using the standard rectilinear pattern. Taking a look at the real final weight of
the parts was pretty interesting because most hooks weight around the same, resulting in
a calculated infill of roughly 33%. Only the hook with the triangular pattern
weight quite a lot more and ended up with a 45% infill instead of the assigned 30%. But now to the most interesting part. What was the strength of the parts? As expected, Wiggle did the worst with only
45 kg until failure. Rectilinear came next and failed at 48 kg. Then there was grid at 52 kg and both honeycomb
patterns failed at 57 kg. The hook with the triangular infill was the
strongest failing at 60 kg! But as we have seen before, 30% infill in
your slicing software doesn’t necessarily mean 30% material in the infill. For this reason, let’s take a look at which
infill is the most economic, so were do we get the most strength per weight. This is where we see that triangular infill
is no more the best but both honeycomb infill scores the highest values. The order of the rest stays pretty much the
same. In the first analysis we have seen that the
full honeycomb seemed to be the most economic solution if you want to increase strength. As a next test, I wanted to find out how the
strength of the hooks change depending on different infill ratios from 0 to 100%. I did this analysis with the full honeycomb
infill but also the rectilinear because I wanted to know how my standard infill pattern
behaves with different infill ratios. Even though it seemed with the first tests,
that the honeycomb patter was significantly stronger than the standard infill taking a
look at the all of the values between 15 and 75% they are pretty much the same. The interesting part again is if we also take
a look at the print time, where the rectilinear pattern starts to shine. At smaller infill ratios the both infills
give you around the same strength in the same amount of print time, but as soon as you go
over 50% the rectilinear prints way faster than honeycomb at the same strength. But wait, this is not the end! Let me show you how you really can increase
your part strength the right way, without worrying about infill ration or infill patterns. I hope this analysis is not too technical
but I think there is too much superficial knowledge around about this topic so I thought
I’d approach it at least a little scientific. If you would like to see more in this direction
please let me know in the comments and consider becoming a Patreon. This doesn’t only help me spend more time
on these topics, but you’ll also get access to all the detailed test reports and test
models! All right, so I have already said in the beginning
the critical section of our hook is not only loaded in tension, but also in bending. If you bend a part, just like the piece of
foam right here, one side will be stretched, the other sider compressed. The part in the middle is not stretched at
all so material which we place in the core of our part usually is loaded not at all or
only slightly. The further material is next to the outer
shell the more it contributes to the strength and stiffness of the part. If we take a look at our 3D prints, this is
shell of our print, which thickness we can adjust with the number of perimeters, shells,
wall thickness or similar. In order to find out how this affects the
strength of the hooks I printed more samples with 2 to 6 shells and only a moderate 15%
infill. I also increased the number of top and bottom
layers to get a constant wall thickness of my part. If we take a look at the results we can directly
see, that increasing the shell thickness is way more efficient than varying the infill. At the same weight a hook that was reinforced
with more shells is significantly lighter than a hook where only the infill was increased. What’s also interesting is, that a hook
with 6 perimeters is already as strong as a part with 2 perimeters and 100% infill. This is due to the reason that in this case
the strands of filament of the shells are in the direction of the internal forces, so
in their strongest orientation, making them way stronger than the infill, that is rotated
all the time. For completeness let’s also take a look
at the print times and here we see, that in my case, I didn’t save print time with the
perimeter method. Print times at the same strength were very
comparable to the hooks where the infill was varied. The reason for this is, that the outer shell
is usually printed slower and with lower accelerations than the infill, to improve print quality. Okay, so what is my verdict? If you want to print stronger parts increase
the number of perimeters and also top and bottom layers. Don’t only crank up the infill ratio. I wouldn’t usually use infill ratios of
100% because at first this is not economic and second might lead to severe printing problems
if only a small amount of overextrusion is happening. Also, don’t try to get 100% infill with
increasing the wall thickness all the way because that can also lead to printing problems
and as I have told you before, material in the core mostly doesn’t contribute a lot
to the strength of your designs. For parts that need to be strong I usually
don’t go above 4 perimeters and 50% infill. That is a good compromise between material
usage and strength. Even though rectilinear infill wasn’t always
the strongest, I’m still a huge fan of this pattern because it gives you quite a dense
grid even at low infill ratios which is good for your top layers. I might need to try out the 3D infill patterns
that are available in Slic3r and Cura and find out how they faire, also considering
loads in different directions. But what’s your approach at the moment if
you print parts that need to handle some beating? Let me know down in the comments! I hope I was able to clarify and debunk some
myths about infill and part strength. If you liked the video and learned something,
then please hit the like button, share and subscribe. If you want to support my work and research,
consider supporting me on Patreon. Thank you so much for watching, auf wiedersehen
and I’ll see you next time!

100 Replies to “INFILL pattern and SHELLS – How to get the maximum STRENGTH out of your 3D prints?”

  1. Shell being most important for strength is no surprise.

    Stand on an empty soda can. All of the strength is the shell and an average aluminum soda can will support the weight of an average grown man… until you bend the shell.

  2. Hello! Thanks for video!
    Looks like you are using one of Prusa i3 3D-printers.
    Could you please share exact printing conditions (temperature of bed / extrusion nozzle, travel speed etc.)?

    I'm using the same type of plastic (PolyLite PLA 1.75), but true black color and do have some issues with over melting of already printed layers.
    My settings are in range of recommended setting, but still something is wrong. Thanks!

  3. Das ist sehr hilfreich! Ich würde sofort unterstützen, wenn Du Deine Ergebnisse zusätzlich in Form einer Tabelle verlinken würdest.

  4. Dear Stefan,
    I'm a product designer and I'm looking for "wire printing" effect with my 3D printer to shorten the printing time by a 10 fold factor.
    It will improve number of iteration to get a better final design.
    Please, have a look at this:
    http://www.3ders.org/articles/20140917-wireprint-3d-printed-wireframe-previews-speeds-up-prototyping.html
    Unfortunately, they never released their software.
    FORTUNATELY, CURA has included it within the "Experimental features".

    Ho dear dear Stefan…. my question: "ARE YOU INTERESTED IN MAKING A VIDEO ABOUT THIS TOPIC?" pleaaaaaaaaaaaaaaaaaaaase!

    Alex from Québec, Canada

  5. I too appreciate the more scientific approach, so you got one more subscriber!. on that note, what was your sample size ? (ie how many hooks did you break of each variety?) Have you experimented with strength vs print orientation? Printing one of those hooks on edge or end? lastly, I have only ever read the program names "slic3r", I heard you call it slic-3-r…is that what people call it? isn't it just "slicer"?

  6. Very useful. I, too, am curious about the 3D infill patterns, but I don't think they'll have much effect on this particular test due to the orientation of the forces. They might help with tensile strength or when dealing with forces on different planes.

    As a follow-up, what about post-processing the parts? I've found that PLA will soak up superglue (cyanoacrylate) and I suspect that improves the bonding between layers, but haven't tested it scientifically. Similarly, polycrylic provides a more impact-resistant and smooth surface, but I have no idea about its effects on strength…

  7. Fascinating, I’ve just started 3d printing and have only fiddled with the settings a little, this gives me an area to work on, many thanks.

  8. Its a great video, but an often overlooked and maybe more important consideration, is design of the part and orientation when printing.

  9. Increasing perimeter width is definitely a noticeable strength improvement. However I believe that alternate layers of serpentine and straight paths in the perimeter would improve strength even more and I’ve asked the Cura team to add this to their slicer. Would like to hear your results if you test this option too. My reasoning for it being stronger is that it will give a bit more on the serpentine but will have axial strength on the straight. But maybe the converse argument can be made !!

  10. I noticed that the breaks reveal a small amount of air between the perimeter layers. I'm not sure whether this is layer seperation due to the load or actual gaps in the printing.
    If it's the first, you could try lowering fan speed to increase adhesion.
    If it's the second, maybe play with extraction speeds to fill in that fraction more material.

  11. there is the possibility in s3d to insert a full layer every insert number layer of infill, which will also have an effect of the strength

  12. Hi, I found your channel today and I'm pretty sure that this weekend I will watch all of your video! anyway, can I ask you if you have already made a tutorial for building a "tensile test machine" like the one in this video? thank you!

  13. Nice work! And very informative. I would live to see the same structured approach regarding the wrapping and the surface quality .Thank you

  14. Thankyou and what a surprising result. I'll be using this information in my next functional prints. I like that you explained the result and why.

  15. If you were to put the load horizontally versus vertically you'd get different results. All around, a honeycomb structure will give you the strongest tensile strength and it has an added bonus of dampening sound waves more effectively than the other infill patterns.

  16. Super gemacht! Wär super wenn du auch mal nach dem "neuen" Gyroid Infill aus Slic3r PE schauen könntest. Der bremst zwar die Geschwindigkeit aus weil er nur aus Kurven besteht, soll aber deutlich steifer sein. Gruß, Thomas

  17. Great video!! You mention at the end of your video never to use 100% infill, but if I wanted to test a specimen for maximum strength, in this case, I would need to set infill as 100%?

  18. Nice job testing this out. I've found that if you want strongest part, you have to determine outside forces nature and enviroment first, then select corresponding material (PLA is fragile, so you wont go with it for impact-proof things, for example), then orient your part so that force goes parallel to the shells in most critical part, and then select infill and shells amount.

  19. This is very useful. I'm in the process of making a helmet for a costume so finding out the strength of each pattern was useful. Also, this could be very useful when printing out props as well.

  20. Awesome video, it strongly correlates with what I also observed at the beginning of my journey with 3d printing. As a side note – I can't believe you pronounced Slic3r as slic three r xD It's slicer!

  21. Usually I print with the grid infill at 20-25%
    I don't like the rectilinear infill because it comes out quite thin

    For parts that need a bit more strength, I print using the cubic infill pattern, which is 3d. From now on, I'll also increase the number of perimeters, since that seems to be more efficient.

  22. PLEASE FOLLOW THESE INSTRUCTIONS
    Say cure, ok you got that, say it in an english fashion
    now again say cure, ok great
    now you will say cure but just take away the E and replace it with the letter A
    CURA

  23. Would be interesting to see, how orienting the part so that the layers are perpendicular to the directions of main tension (Hauptspannungsrichtungen :D) would affect the strength of the part.

  24. Do you follow any standard norm when you perform the tensile tests? I would like to perform some tests but I do not know which norm to follow. Could you help me?

  25. I use 30-40% cubic subdivided infill and 3 shells for almost everything, and do 4-5 shells if needed (high stress parts)

  26. Guten Tag, great video 🙂 My first 3D printer is still on its way to me, but it can't hurt to start watching videos right now.
    Your video reminds me at a documentary I watched a couple of month ago about (the MIT?) engineering 3D patterns for different stress directions and some crazy origami like, single sheet metal folding structures. I am already searching for this video and will link it later here. This gives me the idea to do the same with flexible filaments and different 3D infills, to give a print the flexibility only in the desired direction.(s)…but I think you should do a video about it, if you like the idea. I probably need a few years to gain you expertise 😉

  27. At 4:03, if orange is really time than that means 'fast honeycomb' takes more time than 'full honeycomb'🍯.

    Awesome tests!

  28. Did you create your own universal testing machine? If you know, could you point me in the direction of resources for how I can build or purchase an affordable machine. Thank you

  29. I would love to donate but I am boycotting Patreon due to their involvement in funding violent rioters. Consider finding a different way to solicit donations.

  30. Hey Stefan 😀 Thanks for the video, you mentioned 100% infill is not advisable, how do you usually achieve your solid specimen for testing? which settings are you using? or what is the strongest part by dimensions if weight is no objective 🙂
    Looking forward to more of this even thou its already a year later

  31. Nice study. It would be interesting to look at compression cases, in those cases infill might be more important since it will prevent buckling. I also wonder if this might be more important in your load case with less rigid materials like Nylon where the side that's loaded in compression might buckle as well.

  32. Very nice video, thanks! What material is your printing bed, not the heated bed, but the transparent sheet??

  33. I do rectangular but do -45 90 45 0 as the angles. I mostly print flat plate and covers for custom parts at work. Would I be wrong in thinking this isn't that strong?

  34. By intuition I was already doing that not to put too much filler but to put more outer layers. now I confirm that I was doing well.
    Thank you very much for uploading this video

  35. You should maybe try Cura's Gyroid infill pattern, my favourite atm, good strength with less material than other infill patterns..

  36. Have you taken a look at the differences observed between different individual wall thicknesses and if that results in a different strength profile than increasing the number of walls? Would love to see if that makes any meaningful difference or if increasing the number of walls accomplishes the same thing.

  37. Who in the name of god would give a thumbs down for this video?? Probably someone who doesn’t give a sh** about a work well done

  38. I'd love to have the data from this video in a PDF or on a webpage for fast reference. Any chance you'd make that happen?

  39. Do you know if the percentage of infill is linear to the used material (in all programs)? I'm using cura as my only slicer and can see that the gap between two lines of infill does not fall smooth by increasing the percentage but with hops.

    With 100% infill the part seems not to be massive but with "the smallest possible gap"

    Do all slicers work / calculate in this way?

  40. I think that using a more optimized test part shape would show the changes in strength to weight even better because you would have less dead weight on the parts.

  41. Hello…I am printing tiny lettering and I am using a 0.2 nozzle however only 3 quarters of the lettering is printing on each letter there is about 20% left that did not print. Do you have any idea why that is happening? I do have pictures as well.. Thanks

  42. What about compression? Many 3d printed items are containers or supports. Could you do something similar on that? Also factoring pattern direction?

  43. Using all of your tests, you should combine your findings and print a set of test hooks. I would love to see how strong a hook you could make

  44. Hello sir.. I wanted to know.. How to solid infill 1st and last layer of the print.. Infill as in finished surface.. I use repetier host for slicing. Thanks

  45. Ha CNC Kitchen is there a follow up video too this? P.S. I really enjoy all your videos. You definitely do help people think. Please keep it up! 🤜🏼🤛🏼🇦🇺🍀😎🤓

  46. Fantastic science ! It's too bad there is no way to add some custom infill elements that could be oriented to account for the stresses your part will experience… Or is there ?

Leave a Reply

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