The Impact of Infill Density on 3D Printed Beams

The Impact of Infill Density on 3D Printed Beams | Bending Test

Рет қаралды 31,730

Күн бұрын

Пікірлер: 78

@slant3d 10 ай бұрын

Check out the partner video where we go through Wall Thickness with the same test: kzbin.info/www/bejne/jGnbZpWur7qbgbM&ab_channel=TangledTesting

@stevebissonnette760 10 ай бұрын

Looks more like a layer adhesion test.

@Splarkszter 10 ай бұрын

Because it is.

@amicloud_yt 10 ай бұрын

Yeah duh? Because the variable being changed is the infill density we are seeing the effect that infill density has on layer adhesion

@Kolcars 10 ай бұрын

@@amicloud_yt sure, but it would be more useful to see how a beam with layers parallel to the ground bend as you put pressure on its end

@Splarkszter 10 ай бұрын

@@Kolcars This guy channel just wants to sell you his services, he isn't very smart. If you want actual scientific and accurate testing wstch CNCKitchen. He does actually test in all ways possible, really high quality.

@747Calvin 10 ай бұрын

Yea, they printed them in the wrong direction so it's testing later adhesion, if they test perpendicular to the print direction you'll see the difference in strength w/ infill.

@agrariancraftsleather 10 ай бұрын

This is a good way to demonstrate the data being collected. Glad to see you have already put up the Google docs for reference

@SneakyJoeRu 10 ай бұрын

I'd argue that infill pattern isn't optimal. You create point connections because it's rectalinear. If layers were consistent results must be different.

@Tokolist 10 ай бұрын

there is one trick people do and you might want to test. if you make a hole all through the model, how much strength gets increased. this way you will double walls. in theory it should be stronger with less infill.

@rigokendrali3241 10 ай бұрын

Nothing surprising about seeing the same problems in composites and fibre alignment to the direction of the force. Only some of the infill structure effectively affects strength in the direction of the force since the rod is not isomorphic. Increase infill, more effective material in the way. Gets more complicated than that when you start looking at local plastic yielding and crack propagation, shearing.

@bradg43 10 ай бұрын

Agreed. There seems to be inconsistencies in the testing procedure as well. I was hoping for more from what I thought was an engineering channel

@dekarskej 10 ай бұрын

These tests would be a lot more powerful with stress/strain graphs to visualize the elastic phase of the subject. The 'switch' mentioned is the anisotropic infill type starting to approximate a fully filled interior (could be controlled by rotating the part). In this test setup, the top half of the sample disjoins at the layer line due to bending stress, so it would probably be better to report maximum torque instead of force. Why only one sample at each condition? I really love this series and would be ecstatic if they were a little closer to standard.

@MySpam-v1t Ай бұрын

Can you tell me please where is the updated video where you compared it with injection molding?

@BennyTygohome 10 ай бұрын

So, if we want to increase strength then we increase wall lines, ie wall thickness, not increase density.

@lancelindquist8925 10 ай бұрын

This is awesome, we always assume that since FDM printed parts have internal geometry they’re an-isotopic materials whereas these tests show otherwise these look more like isotopic parts.

@slant3d 10 ай бұрын

Very true. While they are non-isotropic they are still able to behave consistently

@gadgetrc94 10 ай бұрын

Agree with the heat density hypothesis on high infills. Maybe do the same thing with different cooling settings.

@MySpam-v1t Ай бұрын

So means if you need a strong part then directly 90 or 100% infill while all the other ones 15 % is fine and add if you need a stronger part as well minimum 3mm on the outside wall... do I see this correct?

@GrimOfDonuts 10 ай бұрын

This is less a test on density, but more a test on layer adhesion with the variable being surface area. Can you rotate them 90 degrees, so the part is printed along the axis where the pressure would be applied?

@evanbasnaw 10 ай бұрын

I didn't see wall thickness mentioned. Were these tests all performed at 1mm walls?

@georgestone8099 10 ай бұрын

Can you do strength tests (shear and tensile) of a rod printed at 0 degrees angle, 10 degrees, 20 degrees, 30 degrees, and 40 degrees? Crucially can you test 10mm diameter, 5mm diameter, and 3mm diameter. I have a theory that the benefits of printing at a slant has a massive difference on thin/delicate features and their strength. It'll have a difference on the 10mm pole too, but I bet the strength gains are bigger in the smaller features. Would be good to get some data on it.

@alexchen1442 10 ай бұрын

Why reclinear infill? This test is for people who cares about mechanical properties of their parts, and few people who cares about strength use reclinear infill especially at low density. I'd recommend provide more realistic testing scenarios, use cubic or gyroid infills.

@testboga5991 10 ай бұрын

Because it's a standard. The same argument could be made for any other infill as well. Also, different patterns will probably only change absolute strength but the trend with higher percentage might still be the same

@markwebcraft 10 ай бұрын

Your testing methodology is questionable... I feel like you literally just pointed out what was obvious... more infill = more contact points between layers = better layer adhesion. I think you should have rotated the part 90 degree during printing to actually test the strength of the infill rather than just testing layer adhesion.

@slant3d 10 ай бұрын

That is a different test

@toonarmycaptain 10 ай бұрын

I'd question whether orientation and geometry of the infill makes a difference.

@jeffwitz8556 9 ай бұрын

Hi, there, I think the non-linear change you're observing is not due to the process, but to mechanical concepts. Considering a homogeneous material from a structure is a complex subject which is the domain of homogenization techniques. There are also other techniques that take into account an internal length. When we do this, the simplest model is that of Cosserat, which adds the local equilibrium moment balance div(σ). This is justified by the fact that the internal length allows local rotations that are not possible if the material is "solid". Thus, it seems that the transition between equilibrium according to Cauchy and a generalized continuous medium is around 80%, which seems reasonable in view of the release of mobilities. It's very interesting, but deserves more study!

@Thee_Gamefanatic 10 ай бұрын

Realizing this is testing a standard layer adhesion of roughly specific types of filament, but what about when you utilize variating layer heights/offset? Seems like you would have greater strength if you have offset layers or if you have a greater overlap of layers. Seems like these simple changes get overlooked.

@fxm5715 7 ай бұрын

The width of the infill membranes also can have significant impact while maintaining the same ratio. Double or triple the thickness of each membrane, and reduce their density by 1/2. or 2/3 You get the benefits of higher heat/adhesion without the overall higher material density.

@Scrogan 10 ай бұрын

Z banding reminds me of my ender 3 I wonder if the same kind of trend will happen at different orientations?

@Ne0kil 10 ай бұрын

can you test if there is a difference in strength if you use 100% infill or if you increase the walls so they cover the whole piece? Think that would be interesting as well

@TheGingerDan 10 ай бұрын

I completely agree with the others stating that the test is flawed because of the orientation of the print. If you were designing a linear object that you intend to support a downward force, you wouldn’t print it vertically. It is a layer adhesion test and your results proved that infill was irrelevant in the testing until it was dense enough to effect layer adhesion. Redo the test with the correct layer orientation and you’d have a more worthwhile video.

@thedamnone 10 ай бұрын

Can you test if parts with lower infill density than 100% can be stronger - I frequently heard that 98-99% (depending on extrusion factors) can be stronger than full infill.

@FrodeBergetonNilsen 10 ай бұрын

I've been printing extensively on PETG, and using infill as mesh for side panels, well, actually all over the place, in my designs. Through my my design process, I have experienced significant issues with printing single stranded prints. Infill is just that. To be printed as expected, I need to print super slow. I actually use 17,5mm/s on my MK3. I have to print at that speed, if the results are to be consistent. Also, I need to increase the line width a bit. I think I ended up at 0,63, and if I use 0,6mm the print will actually be off, visibly off, as the extrusions will not be constant. You do not disclose any info on how you print your infill, and you use PLA. But printing a large flat piece, with no top or bottom layer, will enable you to really inspect the quality. You need to look for consistency of the lines. I paused it at 6:41, and it is clearly visible. Looking at your prints, you seem to have a reappearing pattern on the Z axis. It seem to be at at fixed frequency, indicating an issue with your z-axis. Not sure if that affects what you dis here. Also, I need to dry my filament, you probably need that as well with PLA but to a lesser degree. Not sure if you did, and I have no idea how that would play out in these tests. But I guess I would really weaken a lot of lines, in single line extrusions, at least, that is what I have been facing. I will end this by pointing out that your findings do not match mine. I find that multiple layers with infill, is stronger than a single layer of solid fill, it the weight is the same. Also, I do not find there to be this dramatic jump at the end, but I might be mistaken on that one. There is something off with my screws of late, using 80-90% infill as opposed to a 100%. I might have gotten that part wrong. Anyway, if that jump at end is a real thing, and universal, it will influence my designs significantly. Not sure how to deal with that yet, but nobody else seems either. Thanks for the video. Love the effort and work you put into this.

@davydatwood3158 10 ай бұрын

I'm sure you've thought of it, so I'll just assume there's a video coming soon with the part printed the other way up and testing across layer lines rather than along them. I'm going to guess that raising infill on that orientation will much much less difference.

@tomsko863 10 ай бұрын

Very nice & very well designed video. It seems like you actually had a plan for this video and how to tie it into everything else you're doing. Congrats! You've learned to actually analyze your results and present them in context with the rest of the video, AND you were able to tie in the other channel, AND drum up interest for the next video. *chef's kiss* delicious! Just perfect on that front! I'm going to start crunching numbers myself to see does strength/fill time show up in the same pattern. Your 20mm rod looks to be 100mm long. With the time you gave I could figure out you print speed. Why do you have that discoloration every 10mm or so?? Interesting hypothesis. Not sure about the added melt time between layers. I believe it has more to do with the compressive half of the rod than the layer lines. The compressive half of the rod is able to resist more deflection (% shortened in mm), preventing the tensile half from separating at the layer lines, which is what is the actual failure.

@Liberty4Ever 10 ай бұрын

Good test. Good data. I'm looking forward to the new tester because those stair steps in the force graph resulting from manually pumping a hydraulic jack really bug me.

@AerialWaviator 10 ай бұрын

Awesome. Was just iching to ask if could repeat test with varying wall-thickness. Jumping on over to view some Tangled Testing.

@linyongzheng 10 ай бұрын

I've always assumed that infill acts more like a truss rather than a solid material. As you get to 90% infill, the bar acts increasingly more like a solid piece.

@n0vaph0enix 10 ай бұрын

Very interesting results. Since your tests were using homogenous infill levels, I wonder if using a modifier part to only increase infill on the obvious pivot point would yield similar results (further improving the utility of 3d prints)…

@kklogins 10 ай бұрын

you said it yourself: contacting walls might make for stronger parts. What I'd like to see researched & never hear anyone really discussing is the relation to strength and infill walls. How much stronger/weaker is i.e. 20% dual walls vs 20% single walls @double width vs 40% single walls. How much time and filament could be saved? This to me would me much more interesting as its underreported & I think could give surprising results...

@TS_Mind_Swept 9 ай бұрын

So unless you're going to go essentially solid with the parts, 30% is your best strength to weight ratio (tho they are pretty close otherwise)

@ionbadea6398 10 ай бұрын

I wiuld like to see PLA vs PLA with CF, short or long. There are some articles on this. But if you have the time, it would be interesting to see…

@jesusisalive3227 7 ай бұрын

This would be a fun project for school kids. Build the strongest beam you can with the same infill density for everyone.

@Kallunke 10 ай бұрын

fast, informative and no bullshit. absolutely on point!

@slant3d 10 ай бұрын

Thanks for watching

@txwombat7826 10 ай бұрын

how about if you tried 40mm rods?- do we expect the same results

@TroyRubert 10 ай бұрын

There's nothing like some good science in the morning.

@slant3d 10 ай бұрын

Thanks for watching

@minerwilly 10 ай бұрын

This is really helpful and interesting stuff. Thanks.

@McRootbeer 10 ай бұрын

Off topic, but for the designed for 3d printing series how about a 3d printed coat hanger? The injection molded ones cause sweaters to deform, and don't have enough strength to reliably hold heavy coats. I could definitely see paying a couple bucks for a well designed coat hanger STL.

@tomirvin2115 10 ай бұрын

I would like to see a video of tests performed with generatively designed beams. I believe that the subject is relevant to 3D Printing.

@securi-t 10 ай бұрын

My guy... You test across the layers not along them. Your failures are at the layer adhesion points. If you run the same test but lay the parts flat on the bed, you will get a much more accurate representation of the strength of the material with a given infill as opposed to when the layers will fail. Of course, higher infills are going to look better in this test because you have more physical contact between layers. You're also testing laterally across the infill as opposed to allowing the infill to provide actual reinforcement.

@jerkyfpv2874 10 ай бұрын

Love the video, hope the new season of Always Sunny is just as good! Thank you!!!

@markedelen2807 10 ай бұрын

Strange to call this a compression test. It is a bending test, with failure likely due to tensile stress (not compressive) on the top of the beam. Also, this is the weakest orientation. Nobody should print a part expecting bending loads with layers lines perpendicular to the beam axis.

@slant3d 10 ай бұрын

Try This Filament: amzn.to/4bMxESP View the Data: docs.google.com/spreadsheets/d/e/2PACX-1vThVXw-98aqI_W6nDK82gJm5xMPYlXuiaw6LUDbI1R7UvsHUN6uThFJUSb9hCoDixsnnyTzgg_Ww3um/pubhtml

@RossRadford 10 ай бұрын

Trial 2 on the 2mm was a bad print. There was some other variable that caused it to break early. Probably a small zone with bad layer adhesion. Throw out that test and then there is no big jump between 2mm and 3mm.

@slant3d 10 ай бұрын

There were three samples of each wall thickness tested. No singular prints to avoid this variability

@noahkatz9616 10 ай бұрын

I believe there's a simple explanation for the increase in strength as infill approaced 90% - there is now more material near the OD of the part. As we know most of the bending strength and stiffness is contributed by material near the OD.

@slant3d 10 ай бұрын

That does not explain the non-linear increase in strength. If that was the case each increase in infill would would create a correlative increase in strength. But that is not wat the data showed. There is a step difference from 80%-%90

@noahkatz9616 10 ай бұрын

@@slant3d Not sure I would expect it to be linear because it's not until infill gets high that there's a significant amount of material, relative to the outer walls, added near the OD. Someone (not me) who's motivated enough could check this by calculating bending moments of inertia of the entire section as infill increases.

@miguelsuarez738 10 ай бұрын

@@slant3d I think there's a point where the density is sufficient that any extra filament extruded at the ends of the infill travel don't have room to squish to the sides, and they get squished into the inner wall layer a bit more than they would otherwise. I suspect you might see the same jump in strength at different infill densities if you vary infill extrusion.

@troncooo409 10 ай бұрын

I wouldn't take the average, just show all measurements and take the least square error. And you can calculate confidence level too. Further you have a hypothesis. Test this hypothesis is if holds.