My solution is 10 wall/10top layers/10bottom layers and 100% infill.
From my experiences it gives me the best results.
Thanks for the video and keep up the good work!
What always surprise me, is how strong 3d printed parts can be in certain applications and orientations, it makes 3d print fun for one like me that likes to print rc stuff and train parts for my model trains
This is what made me go with an FDM printer (so the plastic deposition system not resin ones), the material is cheaper and you get considerably strong parts. I'm always printing only practical stuff either way so the mechanical strength is necessary.
Great set of tests, and thanks for also including Stefan’s data about strength as a function of wall vs infill. Great work!
So a downward hook, infill. An outward hook, wall. Stretch vs bend benefits. Fascinating.
I wonder how it would differ when combining the two in varied amounts. For instance if a total wall number that can max be is 10 layers, doing 1 wall layer with 100% fill, and going up 2, 3, 4, 5, 6, 7, 8, 9, and full wall no infill, to see if there is a sweetspot where it gives better result from it over full infill vs full wall.
This was pretty interesting. You have tested the two extremes (Max walls vs max infill) but It would be also nice to see how lets say 10 walls + 100% perform.
Great video - I've been wondering about this for months (but never got around to actually testing it) and now I know. Thanks!
I think some benefits can come from the alternating angle of the infill. But I often print max walls, because it’s not shaking the printer in small areas like the zig zag. Also - but this depends on how good your printer is tuned - sometimes a lot of infill can build up more overextrusion, because of linear advanced, overlaps etc. My petg prints are often a bit cleaner with max walls. Great content as always!
Thank you very interesting.
I would also have guessed that the max walls would win, exspecialy for the tensile test because the walls are orientated parallel to the tension while the infill is at an 45 degree angle to it.
Nice test! Thanks for sharing!
You speak English very well. It was easy to understand what you were saying. Thank you for this helpful video!
Truly you doing one of the best tests. Wish you have good luck.
Great testing!
Thanks once again Igor. You save us mere mortals a lot of time and effort by running these test.
Great test, just what I was looking for to help me decide before making final prints more solid , great work, thank you very much
Wow that was a clear explaination ! thank you !
it makes sense that 100% infil parts are stronger as the layer "criss cross" so to speak so there is more surface area for it to grab onto. With multiple walls its laying down on top of the same pattern each time so the surface contact patch would be smaller.
Fantastic analysis!!
Always fun to watch and good information.
Nice Work, thx vor uploading.
Great test, very interesting
Thanks for sharing your experiences with all of us :-)
Great test! Interesting channel, subscribed!
Very informative!
These results are interesting. For my prints requiring different strengths, such as a 3d printed electric guitar, the guitar string tension would then need maximum walls down the spine, but everywhere else use maximum infill.
If you're just testing, it's good for my wallet. If there is a product review in your video, I almost always know I will be ordering. Good job. Thank you, greetings from Slovakia.
great tests. This confirms my long held suspicion that with 100% parts, infill is the way to go. Usually, my hesitancy comes from the fact that walls are only bonded on two sides, really (since the material directly underneath doesn't add much support in this failure axis), and those bonds are the weakest parts. As a part fails, the failure typically occurs between two walls, and can then "zip" down between the walls. For this reason I call it "zipper failure". The crosshatched nature of (rectilinear) infill inherently prevents zipper failure, as the break is not allowed to continue along the line bonds unimpeded, it has to break completed lines across the line bonds.
It looks like on the infill specimens, there was a bit of a gap between each line. If we consider that the increase in strength from infill is a continuous line (% infill vs part strength), and that increasing walls and increasing infill reach similar ultimate strength, coupling that with stefan's infill vs walls testing would mean that the last few gaps that remain during infill printing are very detrimental to the overall strength (the last few % of infill give more strength per % than previous). Therefore, I think the inclusion of air in the 100% infill part due to slight underextrusion can significantly impact the strength characteristics.
It would be interesting if you could tune your extrusion so that the infill portions are truly gapless (it's not easy to do), and do another test with 100% infill, 99%, etc down to about 90%. I think there would be enough data to find the curve where the strength starts increasing more and more with those final few %.
Some infill patterns also don't support 100% infill, but at 99 or 98% it's functionally 100%, but with a more complex structure to reduce breaking.
For total maximum strength. I'd just print a pattern, make a mold of that pattern then choose a material. Cut up filament/use pelets then injection mold. 100% solid material with no infills or walls to worry about. Lol.
Great video Igor, many thanks! This here mech engineer is also surprised by the results, but I think your explanation attempt sounds both reasonable and logical.
A development since this video is variable width extrusion, up to about 150% of nozzle diameter. Useful for both faster printing without needing a larger nozzle, and where a wider wall better fits a narrow part. This should allow the max walls option to closer approach 100% infill.
Nice video. At fist I thought that max walls will aboslutely crash max infill. I think that when using max walls, the lines exacly in top of each other play a mayor role, and also, in max infill the lines corresponds with the direction of max shear stress. Not enough specimens for a conclusive statetment but very interesting theme. Greetings from Argentina
Excellent topic. I print a lot of structural/load-bearing parts, and your experiments have been very helpful in optimizing slicer settings (and material choice). Keep up the great work.
@@KillerofGods 85+% of consumer filament is PLA variants; it produces excellent detailes parts and is actually quite strong. That said, tensile strength isn't necessarily the end all be all trait; I've used ABS/ASA (very similar) for parts requiring greater heat resistance, PETG for vase mode & applications where PLA is too brittle.
Engineering filaments tend to get expensive fast, so you generally only want to use them when PLA is insufficient. Many manufacturers make carbon fiber imbued filament, to (theoretically) improve stiffness. Also there are glass fiber-imbued nylons that are among the toughest plastics in existence; there's a reason Glock makes all their frames from glass filled nylon. I've recently taken a fancy to polycarbonate, or more specifically PC blends, like PCABS. Polycarbonate is exceptionally strong, tough, and heat resistant but pure PC can be finicky to print.
There are others I've neglected, but that gives you an idea. Exploring and destructive testing filaments has been a surprisingly engaging pastime.
On the mass difference - with infill, there's overlap with the walls so perhaps that contributes to the higher mass?
Possible. But I belive if I would variate the width of object, it may be different. For example 10.2mm 10.4mm etc. But not significant difference. Probably there is some default flow rate for infill, so they overlap a little bit.
@@MyTechFun Probably not useful in practice, but what if you ironed every layer of the print? Could be an interesting (and long) test if nothing else.
@@MyTechFun There is also a concentric infill pattern so maybe you could run the tests again with that to remove any differences between how Cura treats walls and infill? At least forseeable differences.
I would definitely like a test of optimizing the models so that the walls have perfect spacing as the infill will auto fill this. This way we have no weak points inside of the print.
Another interesting thought is 100% line infill, but in the direction of the part. Will it work even better than walls due to the higher amount of going over itself (30% usually)
Thanks!!!
Infill has 30% overlap with walls in Cura by default. Make it stronger by increasing the overlap.
I'm guessing infill is stronger because of the alternating direction of layers giving a ply wood affect of alternating grain structure. I wish you could alternate between the max infill and max wall option each layer
Would've been cool to see a version of the specimens with 4 or more than 2 walls with 100% infill to see if a mix is even better than walls or just infill. Good video regardless.
100% infill had more material, causing cross-sectional area to increase slightly, making the part stronger. The difference in mass is about 3.7%, and your numbers show that infill is about 3-10% stronger, except for the impact test.
Even considering the mass, it does seem 100% infill is slightly stronger which makes sense considering the lines cross, making the part stronger.
I think the most fair way would be to check the material mass used in the slicer, and make sure they are equal before doing the test.
I was near sure, that the full wall prints would be stronger.... Interesting, even if the differences are not so big.
Why were you sure it would be stronger? They are all stacked vertically? The infill ones have the layers alternating creating more of a sandwich between each layer.
i love the questions you ask, love your channel. Thanks for this. Btw, concentric lines for infill would be the same no?
Yes; excessive number of walls is the same physical operation as 100% concentric infill. They merely look different here because of the top/bottom layer fill pattern, which can also be set to concentric.
The infill shown in this video might be 45 degree rectilinear (zig zag or lines in Cura). It alternates the infill direction, which concentric fills don't. That makes a composite material more like a weave than parallel strings.
I thought the wall would be stronger too. Intresting. Great tests and video
good video :)
Velcome to znz kitchen. Love it. Keep it up!
I really like the clear way you Display the data. It could be a bit bigger, maybe even full screen, it's a bit hard to read on a tablet or a smaller laptop. Or maybe I just need to get some glasses.
Anyway, nice work. Keep it up!
I got earlier similar comment, I am trying to enlarge the results on screen, but also I upload results as picture to my website, so you can enlarge it, links in description (bottom of the page here: www.mytechfun.com/video/124 )
@@MyTechFun
Alright, to my taste that's even better. I like to take some time, even when looking at rather simple data. Thanks!
Thru trial and error I found that for load and torsion bearing prints a 4mm thick wall and solid infill is the best way to go.
You should try slicing with Cura Arachne engine (still in beta) - it can vary wall widths to completely fill the space.
Infill also typically has some overlap with the interior perimeter, I would guess that is why there is the weight difference.
The more 3D printing test videos I watch, the more I realize that the best method is to just print everything normally. Normal PLA is a good material. If your part fails, you need to design it better. There's no magic trick to make things stronger.
Suggestion: would you make the text bigger so that one can see them at 144p resolution? Some of us have to live with low bandwidth connection and can't see such small text. Thanks.
Great video, me liked. I only have one thing, it would be great if we saw super closup on the breaking cross sections, possibly even in a controlled environment and only a photo, not a video. Just an idea.
Thanks for great content, i like it very much. I am learning a lot for the time when i have my own 3d printer :D
I agree. I always have the problem with macros with my current equipment. But I have some digital microscope waiting for the review, I may use these samples to show closeups.
I doubt the higher weight is from the closing line. More likely is, that your infill is printed with a 30% overlap (default in CURA) with the walls. Needs a little more Filament due to the overlap.
Would love to see more close ups of cross sections in different directions ( longitudinal/latitudinal)
I know it would be hard to do, but it would be cool to see this test done again with actual moulded plastic parts. I bet a part that is injection moulded weighs more than any 3d printed version will, even with 100% infill.
Oh shit, 3d printing vs cnc machined. That's the right test right there.
Comment on the first hook test. I calculate medium strength per gramm: Wall 5685.5/91.6=62KGperG; Inf 5897.5/96.3=61KGperG (decimals truncated). For small item perimeters fat has significancy.
I've seen 100 times in groups that "walls are better". Based on this info I've come to 2 conclusions. 1) it doesn't really matter as they are more or less the same. 2) infill is faster so why slow it down with perimeters? So 100% infill it is for the rare cases I need it.
Walls are important if you don't use 100% infill. Big difference is using 2 or 3 walls.
@@MyTechFun yeah I saw that on CNC kitchen. He just didn't do 100%. I'm about to do a 100% part myself. And because everyone says walls, I bumped them up to 999. But I noticed it prints WAY slower that way that 100% infill. So in this case 100% infill is the way to go.
Have you ever tested using different infill line angles? Select 100% infill and set it so that each layer has lines that are perpendicular to each other or other angles like 0,45,90,135. I'm curious if it makes a difference.
Great stuff as always!
No I didn't test different angles. Probably they may effect on resistance to direction of the load.
My personal experience with printing PLA is lower the fan speed. Find that sweet spot where the PLA still prints good with the lowest fan settings. The filament will layout hotter and stay hotter longer creating a stronger bond within the layers.
What speed are you printing the walls at vs the infill? also how much "stepover" on the infil into the outer shell?
Very interesting test as usual. What I wonder though, is there a difference between parts printed all at once, or parts printed one by one so the part does not cool down as much between layers
If they printed all in same mode there shouldnt be a difference shouldnt it?
@@AirMaxAtzeKtb Since it takes longer to print a full bed, the part will cool more between layers and that could affect layer adhesion.
You are reading my mind, that test will be in one of my next videos (after few product reviews, waiting here)
@@MyTechFun I will definitely come back to your channel to check it out.
Btw. By the accent I guess you're from Poland. Am I right? :-)
In my maker's view, you should have used in infill designed for mechanical strength. In your test, you used the "lines" infill, which is well known to be suitable for decorative pieces, not mechanical parts. For example, in Cura I only use the Cubic infill for mechanical parts and do not use Line infill (as you did) because Cubic infill is much stronger for functional parts. Hope this helps.
Cura does better with lines when doing solid objects. I print guns with lines as my infill, can't think of anything less decorative lol
Also note that infinite perimeters will limit the x/y-resolution to 0.4 mm increments which can be a problem if you need dimensionally correct prints.
Is that because of the fixed extrusion line widths in slicers? If so, Cura, PrusaSlicer, and its derivatives, now support automatically varying perimeter widths to limit gapping. (It's called Arachne perimeter generation. I'm not sure why it's called that, but maybe because the resulting slicing patterns in sharp corners end up resembling spiderwebs.)
I thought for sure the wall would be stronger.. So... I learned somthing.. But wall is better for flexing situations.
The print time for max amount perimeters was 3,5 hours. What was the print time for the one infill? Because if it was printed slower, that could have made it stronger. Slower printing means more bonding between layers!
07:13 It would be good to have tables and other detailed text full-frame for those of us watching on our phones, propped-up on a stand rather than directly in front of our faces.
Ok, I agree. Here is a quick help, you can zoom in: www.mytechfun.com/download/124/results-infill-vs-wall.png
Sir, the 100% infill pieces were ever so sligthly heavier, thus denser. Can you explain difference in strengt by differenc in density?
What about all tops and bottoms, any difference from those 2 mentioned?
You know that there is 12 different types of Infill patterns listed in Cura. Time to see which infill is strongest. What infill pattern did you use for this testing? My research says that Triangle pattern is the strongest pattern to use. Awesome vid appreciate taking time to show all the results you've come up with.
CNC Kitchen did very good testing for pattern strength. So far I don't see the reason to repeat that test.
How about the printer motions. I think shakey infil isn't has nice as long walls/infil. I also thought walls would be clearly stronger, I was wrong.
I'm not normally one to ask about speed but since the difference was so minimal I have to ask...
What method was faster to print?
It would matter if you needed to rapid prototype a part for mechanical purposes 😉
In most cases, where you have straight walls, max walls is faster, more smooth print (more straight lines during printing)
Nice test. Isn't the difference is the pattern only? Concentric infill should near the same(look) I think. Maybe an ironing top layer give some very small strength.
PROJECT FARM FOR 3D PRINTING!!
so what patterns are strongest at 100 percent infill? does it make a difference?
Prusa Slicer changes the infill to this lined (I don't remember the correct name), when I set 100%
@@MyTechFun yeah lines i think its called i read somewhere rectilinear and concentric works with 100 percent but never tried it. edit-just realized rectilinear is lines
Wonder if with Arachne the walls will hold better now.
Hm, yes, probably I should repeat similar test soon.. that adaptive layer width is a game changer for these mechanical parts.
I have massive doubts about that device you're using with half a second refresh rate. During that half of second, there could be higher values, so it's pretty much unreliable, unless you go really slow
hmm, Think you could try ten walls with max infills and see how strong that is?
I don't think that will result any significant difference, since both methods has very similar strength.
Listen... if you're asking yourself this, it is time to make some rubber molds of your parts and pour in resin. Or cast them in metal.
Fourth
I love when the algorithm shows me some smaller channels, hopefully you can grow, the testing is very well done and you deserve more views