Rethink how you use 3D printer infill!
Summary
TLDRIn this video, the host explores the complexities of 3D printer infill, testing various patterns for speed, strength, and top surface support. Using PETG filament from sponsor VOXELPLA, the experiment compares 2D and 3D patterns, nozzle sizes, and the balance between infill and shell thickness. The results show cubic infill as the strongest and fastest, with aligned rectilinear excelling in surface support. The conclusion favors thicker shells over high infill for structural strength, offering insights for 3D printing enthusiasts.
Takeaways
- đ The video discusses the complexities of 3D printer infill, acknowledging the lack of a single 'best' answer due to varying needs and preferences.
- đ The script explores three main aspects of infill: the best infill pattern, the use of thicker nozzles for infill, and the balance between infill and shell thickness.
- đ It compares infill patterns based on print speed, strength, and their ability to support the top surface of prints, using the same amount of material for each pattern.
- đïž The video categorizes infill patterns into '2D', '3D', 'specialty', and 'dubious', highlighting their unique characteristics and applications.
- đ The fastest infill pattern tested was concentric, which does a single continuous line, while the slowest was Lightning, due to its minimal material use and specific application.
- đȘ The strongest infill pattern in terms of bend load was cubic, which also printed quickly and supported top layers well.
- 𧩠The script notes that for structural prints, a thicker shell generally provides more strength than additional infill, except in specific part geometries that benefit from infill.
- đ The video mentions that 3D and specialty infill patterns like cubic and lightning infill can provide better support for top surfaces compared to static 2D patterns.
- đ§ The test used PETG filament from VOXELPLA, chosen for its balance of affordability, reliability, and suitability for mechanical applications.
- đ ïž The methodology for testing involved adjusting infill percentages to ensure equal material usage across different shell thicknesses and infill densities.
- đ The results consistently showed that a thicker shell provided better strength than additional infill, suggesting a rule of thumb for part printing.
Q & A
What are the three main aspects of infill that the video aims to test?
-The video aims to test the 'best' infill pattern in terms of print speed, strength, and support for the top surface of the print; whether using a thicker nozzle for infill is beneficial; and whether more infill or a thicker shell provides better structural support.
Why did the author decide to use PETG for the tests instead of PLA or other materials?
-The author chose PETG because it's a good compromise material that is suitable for mechanical applications and provides reliable data, unlike PLA which is rigid but tends to creep, or polycarbonates and ABS which require a controlled environment for optimal printing.
What are the four categories of infill patterns mentioned in the script?
-The four categories of infill patterns are '2D', '3D', 'specialty', and 'dubious'.
Why is honeycomb considered an odd one out among the 2D infill patterns?
-Honeycomb is considered an odd one out because it is the only pattern that avoids crossing over itself, instead creating double-width walls in some spots.
What is the fastest infill pattern tested in the video?
-The fastest infill pattern tested is concentric, which does one single, continuous line.
Which infill pattern was found to be the strongest in the video's tests?
-Cubic infill pattern was found to be the strongest in the tests.
What is the general recommendation for infill percentage when printing structural parts?
-The general recommendation is to use just enough infill, around 15-20%, to ensure the part prints cleanly, and then use the rest of the material to increase the shell thickness.
What was the result of testing different shell thicknesses versus infill percentages for strength?
-The test results showed that a thicker shell always provided better strength than more infill, regardless of the starting infill percentage.
What is the significance of aligned rectilinear infill pattern in supporting the top solid layers?
-Aligned rectilinear provides almost perfect support for the top solid layers because its aligned pattern allows the first solid layer to bridge over cleanly at a 90° angle.
What is the author's suggestion for the default infill pattern in most cases?
-The author suggests that the default infill pattern should be cubic, or adaptive cubic, as it provides a good balance of strength, print speed, and support for the top surfaces.
Outlines
đ€ Exploring 3D Printer Infill Controversies
The script introduces the complexities and debates surrounding 3D printer infill, with the narrator aiming to test three aspects: the 'best' infill pattern, the impact of nozzle size on infill, and the balance between infill and shell thickness. The narrator will compare print speed, strength, and top surface support, using the same amount of material for each pattern. The testing will be conducted on an XL printer with PETG filament from VOXELPLA, chosen for its affordability and reliability, and the narrator mentions other products from the sponsor.
đ Categorizing Infill Patterns for Testing
The narrator categorizes infill patterns into '2D', '3D', 'specialty', and 'dubious', detailing the characteristics and potential issues of each. The '2D' patterns like grid and honeycomb are compared, with honeycomb noted for its slower print speed. '3D' patterns are highlighted for their support and strength, while 'specialty' patterns are discussed for their specific uses. 'Dubious' patterns are questioned for their practicality, with the Hilbert curve noted for its aesthetic appeal despite questionable mechanical benefits. The narrator also discusses the methodology for testing these patterns, including scaling down parts and using a smaller nozzle for consistency.
â±ïž Print Time and Strength Analysis of Infill Patterns
The script presents the results of print time tests for various infill patterns, showing a clear division between 'fast' and 'slow' patterns, with concentric being the fastest due to its continuous line. The narrator discusses the strength testing of these patterns, with cubic emerging as the strongest, followed by concentric. The results are qualified with the acknowledgment of testing limitations, such as orientation and load type, and the narrator emphasizes that while cubic performed well, other patterns may be suitable for different applications.
đïž Material Distribution: Infill vs. Shell Thickness
The narrator explores the distribution of material between infill and shell thickness, testing various configurations to maintain consistent filament usage. The results consistently favor a thicker shell over additional infill for strength, suggesting that for typical part geometries, material is best used in the shell. However, the narrator notes exceptions for certain part shapes that could benefit from infill. The script concludes with a recommendation to use a default infill pattern like cubic or adaptive cubic for most prints, adjusting based on specific needs and part geometries.
Mindmap
Keywords
đĄ3D Printer Infill
đĄInfill Pattern
đĄMaterial Efficiency
đĄStrength
đĄPrint Speed
đĄTop Surface Support
đĄNozzle Size
đĄShell Thickness
đĄPETG
đĄVOXELPLA
Highlights
The video explores the best infill pattern for 3D printing, considering speed, strength, and support for the top surface.
Different infill patterns are tested with the same amount of material to ensure a fair comparison.
The impact of nozzle size on infill structure coarseness and print speed is examined.
The debate on whether to prioritize more infill or a thicker shell for part strength is discussed.
The testing is conducted on the XL 3D printer, allowing for different nozzle sizes in a single print.
PETG filament is chosen for testing due to its balance between rigidity and flexibility.
Infill patterns are categorized into '2D', '3D', 'specialty', and 'dubious' for systematic analysis.
Honeycomb pattern is noted for its slow printing speed due to non-overlapping and double-width walls.
3D patterns like cubic, 3D honeycomb, and gyroid are praised for their consistent top surface support.
Lightning infill and support cubic are identified as specialty patterns for minimal material use and good part appearance.
Dubious patterns such as Hilbert curve are recognized for their mathematical interest but questioned for practical strength.
Cubic pattern wins in strength tests, followed by concentric, despite its inconsistent pattern.
Aligned rectilinear pattern shows exceptional support for top solid layers due to its alignment.
The test results indicate that a thicker shell consistently outperforms additional infill for strength.
The recommendation is made to use the default cubic pattern for most prints, considering strength and print time.
For parts requiring high strength, a thicker shell with minimal infill is suggested.
The video concludes with a call to action for viewers to share their experiences with structural part printing.
Transcripts
Ah, 3D printer infill. A topic thatâs so full of polarized points of views, personal preferences Â
and passionate presumptions that I thought, well this is something that I can jump into and surely Â
not get wrapped up in any controversies. Perfect. But on a serious note, there are still plenty Â
of things that donât have a definitive answer yet, so today, Â
I want to test three different aspects to infill: First, what is the âbestâ infill pattern? And yes, Â
âbestâ almost never has one single answer, so this breaks down into a couple of sub-sections. Â
Iâll be comparing both how fast they print, how strong they are, and how good of a job they do Â
supporting the printâs top surface, because yes, sometimes all you care about is your printâs Â
looks, and not how rigid or strong they are. And Iâll be doing all of this with the same amount Â
of material used for each pattern, not just with the same infill percentage - because turns out, Â
you canât trust the infill percentage number at all when changing patterns.Â
Second, should you be printing your infill with a thicker nozzle? This keeps coming up Â
as an argument for dual-extruder or toolchanger printers, basically saying you can speed up your Â
prints by just using a fatter nozzle for the infill when you wonât be seeing it anyway. Â
But that also makes the infill structure coarser, so is it worth the speed gain?Â
And third, more infill or a thicker shell? The engineering logic is that you should Â
always prefer putting more material into the shell of your parts, because thatâs where the Â
majority of the strain is happening, but there has got to be a crossover point where thereâs Â
simply not enough infill to support the shells, so weâll find that point, again, Â
keeping the amount of filament used constant. Iâll be printing all this on the XL, Â
because thatâs the only printer where I can chuck up different nozzle sizes and use them in the same Â
print. For filament, I thought about materials I would actually use when strength matters. PLA Â
is very rigid and strong, but it also likes to creep and when you really need strength in Â
mechanical applications, you quite often also have at least some amount of heat involved, Â
so while it may make for some easy testing, PLA just seemed a bit pointless. The other end of the Â
spectrum with polycarbonates, ABS or ASA are good candidates for mechanical applications, Â
but because I donât have the enclosure set up yet and even then, I donât think it would Â
be controlled to a consistent temperature, so I compromised and used PETG. Itâs the softest Â
of materials and might sometimes bend instead of snapping off, but thatâs still good data.
And what a coincidence - all the PETG Iâll be using is from todayâs sponsor, VOXELPLA!Â
Their PETG Plus and PLA Pro are among the most affordable filament choices you can print with Â
at just $16.99 per spool or even less with bulk discounts. VOXEL filaments are exclusively used Â
in their 150-machine print farm in Southern California - and I only have good things to Â
report about them as well. They print great with default profiles and no tweaking, Â
theyâre suitable for high-flow printing and have been a perfectly reliable choice for me.Â
Also check out their printer upgrades like the Bento Box Air Filter, Â
HULA Vibration Damper, and the Python AMS Dryer Upgrade - at the link below.
Letâs get started with infill patterns. Iâm going to group these into four categories: Â
â2Dâ, â3Dâ, âspecialtyâ, and âdubiousâ. In the basic 2D camp, weâve got old favorites like grid, Â
triangles, stars and honeycomb. The odd one out here is honeycomb, because itâs the only one that Â
avoids crossing over itself and instead creates double-width walls in some spots. Crossing over Â
already printed paths can be an issue, especially when going fast, because youâll often see the Â
extruded line rip and tear at those points, which can weaken the infill structure as a whole. But Â
honeycomb, being a hexagonal pattern, has tons of corners that the printer has to slow down Â
for, so itâs a rather slow pattern to print. Weâll get to the exact print speeds in a bit, Â
but as a rough estimate, for the same amount of material, honeycomb takes about 20% longer Â
to print, and that difference only goes up the faster your maximum print speeds are.Â
Then, the 3D group, the cool kids of the bunch: cubic and its varieties, 3D honeycomb, Â
and gyroid. 3D honeycomb and gyroid are both no-cross patterns too, but again, Â
theyâre both on the slower side. The gyroid pattern is explicitly meant to have even strength Â
in all directions and uses a very organic pattern, while 3D honeycomb is more of an Â
octagonal pattern that changes print direction every layer. Cubic is actually just cubes stacked Â
on their corners, hence the name, so it looks a lot fancier in the slicer than it actually is, Â
and essentially, itâs just a tilted version of grid if it were to print a solid layer Â
every couple of layers. I do like the 3D patterns because I feel like they do a better job Â
of consistently providing some support for top surfaces and tying together perimeters that may Â
not ever get crossed by the static 2D patterns. Then, âspecialtyâ. These are circumstantially Â
useful, and thatâs lightning infill, which is designed to use as little material as possible Â
while still holding up the top solid surface, and support cubic, which⊠does the same thing, Â
but with the cubic pattern that prints tighter and tighter cubes the closer you get to the top Â
surface. Neither one of these is designed for strength, but just to give you good-looking Â
parts. Iâve printed some samples with lightning infill anyway, but since its strength literally Â
depends on which way you hold it, this is more for completeness than anything.Â
And then the âdubiousâ patterns. I mean, I see why theyâre in here, and thatâs because they are, Â
by their mathematical definition, patterns that fill in a surface, with an adjustable Â
amount of fill density. Yay! These all happen to be non-crossing patterns by the way, but honestly, Â
I donât think these will provide any value at all, but hey, Iâm very open to be convinced otherwise.Â
The most striking one certainly is the Hilbert curve. Itâs more of a mathematical Â
phenomenon than an actually mechanically sound concept, but I have to admit, it does look Â
cool! Iâd also put concentric, Archimedian Chords, and Octagram Spiral in that camp.Â
Rectangular looks similar to grid, only that each layer is exclusively printed in one direction, and Â
it alternates every layer. So this is a pattern that only has point contact between the layers and Â
uses most of its material in unsupported bridges. Aligned rectilinear is the same pattern, Â
just with all the layers aligned in the same direction. Makes sense. And lastly, Â
the line infill once again is the same pattern as rectilinear, but this time, drunk.
I still printed all of the patterns as test parts, and because my intuition would be that the effect Â
of the different infill patterns would be more pronounced the more space it has to work with, Â
but printing samples large enough where especially the 3D patterns donât just turn Â
into mush would mean Iâd samples that would end up breaking me instead of me breaking them. So Â
I simply scaled everything down and used a 0.25mm nozzle for the parts. The parts all Â
printed perfectly with the smaller nozzle, so all Iâm doing is scaling down the strength of Â
all the parts evenly. This, of course, also saved me a bunch of filament and print time.
Speaking of time, hereâs how the patterns stacked up for print times:Â
Basically, there is a fast group and a slow group, but within each group, youâre probably not going Â
to notice a dramatic difference in print times. But when going from a âfastâ pattern to a âslowâ Â
one, print times increased on average by a third, and that is with the same amount of material laid Â
down in each of these prints. As expected, patterns that use many short moves instead Â
of longer continuous ones do take significantly longer; with the fastest one being concentric, Â
which does one single, continuous line, making it very hard to beat. The default grid pattern is Â
tied for second place with the three rectangular varieties that also just print straight lines, Â
followed by Stars, Cubic and Triangles. The slowest pattern overall is Lightning, Â
but itâs an odd one out, because a) I couldnât get it to lay down so much material that it would use Â
as much as the others, and b) itâs not meant to be used at such high infill ratios anyway, Â
all it does is make the very top structure denser, which, at a certain point, just ends up being Â
material that doesnât really help the print. Over in the âslowâ group, gyroid is the fastest Â
pattern, and while it didnât make much difference for print times back when the pattern was first Â
introduced before input shaping had become a thing, with printers moving much faster now, it Â
means that they have to slow down to accurately print the swirly goodness that is gyroid.Â
But maybe it makes up for it with its strength? Well, sort of. It takes third place. Second Â
place is actually the fastest-printing pattern, concentric. Though take this with a grain of salt, Â
because concentric doesnât have a consistent pattern throughout, Â
and I think really, we just got lucky here and managed to get a spot that was perfectly Â
oriented for the bend load I was applying. Most of the other patterns form a surprisingly Â
tight middle group, with the only outliers being lightning, again, not meant for strength, Â
and Hilbertcurve, which falls into the same camp as Concentric, Archimedian Chords, Â
Octagram Spiral, of not being a consistent pattern, which I think disqualifies all of Â
them from being used for structural prints. And if youâve been counting cards, Â
youâll know what the top spot is - itâs cubic! With a significant lead, actually!Â
Now, caveats. I tested these patterns printed flat and on their side, but not in any other Â
orientation, so some patterns might behave differently in slightly different alignments; Â
and I only tested for bend load, as itâs the easiest one for me to test, but also, Â
most real-life loads end up inducing a bend load of at least some amount of significance.Â
Also, while these results are consistent for this geometry with this combination of print profile, Â
machine and material, things might slightly change with different part Â
geometries, and with filament thatâs more or less rigid or has higher or Â
lower material strength. But I think the general trends should still transfer.
Weâre going to dive a bit deeper on strength in a bit, but first, Â
letâs look at top solid layer support. The test prints for these were done Â
with only three top solid layers and roughly 10% infill, again, Â
I had to adjust each pattern individually to make sure they all used the same amount of material.Â
And all these prints look about the same. They all have some areas where they covered well, Â
other bits where they didnât do so well, except⊠for aligned rectilinear. Â
That one is almost perfect, and it makes total sense! All the other patterns have sections where Â
that first solid layer has to span a distance thatâs too large to bridge over cleanly, Â
but because aligned rectangular is, well, aligned, it ends up being perfectly at a 90° angle, which Â
gives that first solid layer the perfect geometry to bridge over. Even lightning isnât this good!Â
One interesting thing here is that between the regular cubic and the support cubic print, Â
the support cubic one does print a coarser structure down low, as it should, but because Â
it then wastes material on printing double walls as it transitions to the full-density pattern, Â
if you actually give it the same total amount of material to use, it has the exact same density Â
of pattern when it gets to the top where it supports the top solid layers. With larger parts, Â
there is an actual benefit to using support cubic, but keep in mind that if you just give the support Â
and adaptive cubic patterns the same infill percentage, without even doing anything adaptive, Â
they will by default create a coarser structure and use less material than standard cubic.
I know Iâm talking a lot about cubic, because thatâs the pattern that I think so far comes out Â
as the overall winner. Itâs very strong, prints quickly and doesnât show issues with its ability Â
to support material printed on top of it. So itâs the pattern Iâm moving forward with for the Â
next test, and that is figuring out whether you can just print your infill thicker. Now, Â
this was all done on the XL because I could mix and match nozzle sizes. But as good as these parts Â
look, the alignment issues struck once again and while one side is perfectly bonded between the Â
shell done on one nozzle and the infill done on another, the other side was completely loose, and Â
this isnât exactly good for strength. Maybe Iâll come back in the future and explore this topic Â
more, but for this video, I had to compromise and just print everything with one nozzle and Â
different extrusion widths, which worked really well, too, but itâs not the cool two-color look Â
I was going for. The results with what I tested here, though, are promising, Â
and there doesnât seem to be any significant loss in strength with printing the infill wider Â
and faster. But again, this was fairly limited testing and more exploration is required here.
And lastly, and this was actually the biggest chunk of printing, where should you put your Â
material - into the infill or into the perimeter? I need to explain my methodology here real quick. Â
I wanted to test 1, 2, 3, 4 and 5 perimeters, so I started with a 30% infill part for each Â
one. I then looked at the weight of each of those samples and for each series, adjusted the infill Â
percentage for all the other perimeter counts until I had a set of parts that all used the same Â
amount of filament, just distributed differently. And of course, I did that for all the other other Â
starting points as well, I also used 5 perimeters and 50% infill as a high-fill starting point, Â
and a single 100% print. By the way, when I say âperimetersâ, I actually mean âshell thicknessâ, I Â
did increase the top and bottom minimum thickness to be identical to the actual perimeters.
And these ended up being quite interesting, or boring, depending on which way you look Â
at it. There was not a single instance where more infill ended up being the better choice Â
over a thicker shell. This is the result that one should expect, but I definitely thought we Â
would be running into issues with things like wall buckling which would reduce strength, but Â
it looks like those might only start becoming an issue once you go into the extremes with almost no Â
infill and very thin shells, but since Iâm testing for strength, I wanted to use settings that youâd Â
typically for functional parts, so those sort of featherweight prints arenât included here, Â
but they might still hold some interesting data. Now, while these trends would suggest that the Â
best setting to use for strength would be as thick of a shell as you can get and zero infill, Â
I wouldnât say that thatâs universally true. Again, itâs true for this one part shape, Â
but I can easily think of a bunch of shapes that would profit from at least some amount of infill. Â
Letâs say you have a large flat part, that would not just be unprintable without infill, it would Â
also be pretty bad at supporting any amount of weight. Or something that has, like, a corrugated Â
or baffle-shaped outer wall, even just a small amount of infill would help with the strength and Â
rigidity of that part much more than extra shells. So with typical prints, which are most often Â
compact parts with walls that are a couple of millimeters thick, your material will still be Â
best used in the shell and not in the infill. I guess as a rule of thumb - Iâd say use just Â
enough infill, 15, 20% so that your part prints cleanly, and then use whatever you would have put Â
into the infill to increase the shell thickness. And even in high-infill situations, Â
youâre probably not going to need to go above 30%. And for patterns, honestly, I think your and your Â
slicerâs default should be cubic. Maybe even adaptive cubic, because that will conveniently Â
save some material with really chunky parts while behaving like regular cubic otherwise as long as Â
you account for the difference in fill density at the same percentage. And if you donât care about Â
strength and just want fast, clean prints, maybe even give aligned rectilinear a try! Just donât Â
forget to add a solid layer every now and then. Let me know what your experiences have been Â
when printing structural parts - do you have a favorite infill pattern? Iâd love Â
to hear from you in the comments below. As always, I hope you learned something, Â
thanks for watching, keep on making, and Iâll see you in the next one.
Voir Plus de Vidéos Connexes
Rapidly Improve Your 3D Prints with Hidden Slicer Settings
AnyCubic Tries Multicolor: The AnyCubic Kobra 3 Combo
3D-Drucker SPEED erklÀrt! 30% schneller drucken! (inkl. Bambu CHT Guide)
What is Architectural Mesh
(How) I finally got rid of these blobs!
The Bambu Labs A1 Mini: The Ultimate 3D Printer for beginners?
5.0 / 5 (0 votes)