The gaps between the layers of your 0.3 and 0.4 prints would indicate that you're printing too cold or too fast for that nozzle size/layer height combo. You even note the under extrusion in the broken part. For more comparable results I suggest you normalize your print settings to make the volumetric flow rate match. The flow rate difference between 0.05 layer height and 0.4 is substantial. If you're not controlling for this difference, your test procedure will favour the lower layer height parts substantially. I think this warrants future investigation.
@@cthulpiss Note he's also using a hardened steel nozzle. You can lose 10c at the nozzle tip. Usually needs 10-20c above the recommended temp, especially when pushing flow rate to ensure adequate melt. For me, I print production parts with a 0.6 hardened steel nozzle at 0.3 at 70mm/s. That's pushing around 12 cubic mm/s which is the limit for the V6 block. For PLA I'm needing 220-230 and with PETG 250-260. Lower than that and part strength is definitely affected.
@@AustinVojta Yes and in this case I think flow rate is just a proxy for Time in Melt Chamber. I also noted that he had clogs at the 0.05 layer height. Probably due to the flow rate being too low and the filament cooking in the chamber.
The test machine makes a sound which is roughly similar to a Moog keyboard in the tone of C sharp. The actual background music, being in the key of F sharp minor, fits in perfect harmony with it.
@@williamlewington3223 Agreed. That's what I see so much from 3D-printing too; people making "pretty" models. But I just love it when 3D-printed parts actually are useful and have a functionality. But of course, that's when material strength usually has a more important role and that's where videos like this one are great.
1:35 This is directly effected by direction of your print. Thinner layer height would be better with the grain, and thicker layer height against the grain.
I’m currently working on some heat treatment methods of PLA for strength gardening. It has been done in the past but not clearly defined in any reports. After heating in an oven for an hour at 220°F I recorded my PLA prints to have almost a 30% increase in strength between layer lines. I also tried boiling and quenching my prints with similar (but not as great) results. Might be something interesting for you to try!
Yes conditioning should improve the mechanical attributes of any Thermoplast.. The issue is to get the "right" procedure... which means temoerature and time and when to do it... I would say it's to be done imediately after the print at a temperature somewhere between Printing bed temperature and max. Permanent temperature of the material.. (and except PA) it definiteky, should be in a dry environment.. With PA it could be that boiling it, maybe the right procedure.. Everything else I would definiteky go for the oven.
There's a lot of discussions out there that address annealing. Yes, you can get some pretty nice strength increase from a simple "heat and hold then cool" where time and temp are the only considerations, but I learned from doing flame glass work that there's a bit more to the science of annealing. Results widely vary dependent on the material, with some showing great increase and some showing virtually none. Additionally, "stepped" heating and cooling stages can make dramatic differences in the molecular rearrangement. It really depends on if you're just doing a "quickie" for a simple print or trying to make a particular item as strong as it can be. One other BIG factor in some jobs is that reheating and cooling the part can result in SIGNIFICANT changes in size, made worse by the fact that it's usually different in different axes. Z usually shows little change, with X and Y showing amounts that differ from each other AND Z. This is a real nightmare if you're making a interlocking or sliding part that joins another! So yes, for a simple part, many plastics do show some strength increase but it's HIGHLY variable and you may find yourself reprinting from the unintended size changes. Tip - Many of the better filament producers give annealing info on their web pages or will tell you if you ask them. Same for the proper drying time and temp as many plastics are hydroscopic (absorb moisture) especially the nylon and polycarbonate blends. Printing with "damp" plastic REALLY screws strength and surface finish.
The important temp to go over is the glass transition temp, and the temp to stay under is the melting temp. The time is important because the heat has to conduct through the part to get in and out, and there are also thermal expansion effects. Rapid heating and cooling means the external surfaces are hotter or colder than the internal material, so the external material is more or less heat expanded than the inner material, while the heat is getting in or out. For example, rapid cooling means the outer layer shrinks and hardens while the inner material is still expanded, so when the inner material does cool, shrink, and harden, it leaves the outer layer under slight compression and the inner material under slight tension. This can make it stronger because bending failures start with tension at the surface, but when it does fail, it fails more catastrophically as the tension failure crack reaches the inner material that's already under a little tension.
It’d be really interesting to see a test with the print lines angled at 45 degrees throughout the object. It’d take some supports to print, but might combine the benefits of both directions of printing.
@@DaStuntChannelnot worse. Instead of 100% vertical but zero horizontal strength it would be 50%/50% which, depending on the circumstances is actually the better deal.
Thank you! I’ve done testing on carbon fiber filled nylon, and .1 layer height was definitely stronger in both planes. Dialed in the settings (with a bit of help of the manufacturer) and now produce incredibly strong parts.
This is a good demonstration of why it's important to make sure your layers aren't stacked parallel to the load direction. I didn't realise it was so critical to tensile (and probably compressive) strength.
@@Kspice9000 I agree but the direction force is giving inaccurate results in this experiment. If he compressed the hook at the same point it would have torn from the outside edge instead of the inside edge from pulling. I bet if the force was applied evenly around the section being tested it would have withstood more.
I just ordered my first 3D printer and must say as an engineering student I greatly enjoy your videos! I look forward to seeing more of your videos in the future.
Please do the same tests with other filament like petg and polycarbonate, because if your printing practical parts you probably wont be using pla for them
@@dudemann. ua-cam.com/video/aODcbK5r5wE/v-deo.html is the best example i can find how petg compares to pla. petg just completely fails on any decent impact. ua-cam.com/video/m1I5zttnYjY/v-deo.html also shows this.
I prefer 0.2..0.25mm as it gives the nicest looking prints but 0.3 is still somewhat decent as well. 0.1 and below gives nasty and unstable results on my printer and with 0.4 the layers do not stick wery well especially with cheap plastics from China and the process tends to give very fragile parts and with thin walls it can easilly not even get printed successfully at all. I would say that 0.2 is the sweet spot, 0.15 is for small parts with small features and 0.25 is for when you are in a hurry.
yep that confirms my experiences, the best results I get with 0,5 mm nozzle and 0,15 layer height, at this settings I get the best looking prints and also the strongest ones
Excellent comparison. From what I can see here, a 0.4mm nozzle and 0.15mm layer height has the highest tensile strength in both print orientations and is near the top for finish quality too.
I really love how meticulous you are with your tests, and the effort you put into designing the tests so as to eliminate any possible bias. If you even want to revisit or explore this further, it would be interesting to see how using a specialized hotend such as an E3D volcano would change the results, considering that they touted their thicker layers as stronger in their promotional material. Personally I thought the 0.2 would be stronger being a 'balance' between thick and thin, but clearly I need to re-evaluate that line of thought.
I would be interested in extrusion width tests, I've been doing a lot of spiral printing in PLA and PETG with 1.2mm width with 0.3mm layer height on my Ender 3, really awesome results (great for making quick boxes), feels really strong too.
Designing a weld-less 18650 battery build system around a 0.6 mm nozzle and your videos have been extremely valuable for balancing strength vs print speed. Thanks!
This is an incredible video. Great job, Stefan! I use 0.3mm height quite a bit on my Mk3 and I mistakenly thought that would make things stronger. However, I have a Cr10 set up per your TitanAero video, with a volcano, 0.8mm nozzle, 0.4mm layer height and extruding at a 1.2mm width. I feel the strength in that case is insanely strong. There is probably a relationship to the width vs layer height. You could repeat these experiments with those relationships maintained as the layer height goes up. I think there is more to be done along these lines, and I can't wait to see more videos on the topic.
Wow ... Wasn't expecting that at all ... I even bought a 0.8mm nozzle "to make stronger part" back in the day, because i was thinking that 0.6mm layer height would improve the strenght of my parts ... Thanks for your work !
You are misunderstanding this test greatly. The point is for same nozzle size what layer height is going to be the strongest. Bigger nozzle will give stronger prints.
You also need a hotend that can supply enough heat to the much higher amount of plastic flowing through it. The ability to get heat into the plastic is going to determine your max print speed.
It's probably been said a bunch of times, but yeah. The larger diameter parts don't appear to have been sliced properly. There's also a considerable difference in the layer heights and flow rates between the smaller nozzle prints and the larger nozzle prints. Looking at this at first I had assumed you were using a 0.8mm nozzle at the largest. I've never seen 0.4mm nozzle prints that looked that bad before.
5 років тому+10
Thank you for putting all that effort into testing these 3D printing parameters for us. This information is very useful and you have just saved us hours of testing. Danke schön!
The conclusions of this video are only relevant to a .4mm nozzle. That is very important to note. You touched on the proportionality issue when you referenced the paper at 11:21. This same issue is relevant throughout the entire video. A more accurate scientific study would have changed the nozzle with each change in layer thickness to maintain the proportionality.
Design Prototype Test you mean more general (not accurate) i think. I pretty much only print with .4 mm nozzle...swapping them out is significant work and I’m guessing most people do the same so it’s valuable for me.
This is a proper study to show that a given nozzle can print at a given layer height range only. It IS shown and summarized in the video, that the optimal layer thickness is proportional to nozzle diameter. Probably it is more important to know what layer thickness to choose for the nozzle you are using, as it is a more realistic scenario that you have a fixed nozzle and you adjust the print parameters than changing the nozzle every time. It would be interesting to see which absolute layer thickness is stronger if all would be printed using it's optimal nozzle diameter, but the test in the video is more of a real world use case.
I'd like to see the challenge be changed to "print this part as strong as possible" and let experienced makers use whatever techniques they have in their toolboxes and apply them. Even specify a wall thickness and let us tune for that. I'm really bothered by the implication that "0.15mm is best" without more detail. Let me use a different nozzle and extrusion width and see how strong the part can be. Great testing, but incomplete and vague conclusions.
@@theroyal3-5643 If that's the case, then he should have acknowledged up front that 0.30mm layers are at the upper end of what you can print for good adhesion at that nozzle size and 0.40mm is almost guaranteed to produce poor results. The testing is excellent (as always) but the conclusions were vague. A test for "how strong can you print this part with a 0.40mm nozzle" might yield altogether different conclusions if different approaches are used.
This is an awesome topic! Interesting how temperature affects the strength gap of vertically and horizontally printed parts. I'm nearly certain a little increase in printing temperature could add a lot of strength to the horizontally printed(the weaker) hooks.
Im editing a video about foam filling 3D prints. The experiment was *mostly* a successful and I would love to hear about how to strengthen prints. EDIT: The results of my video were pretty amazing! Hopefully you can check it out and see what you think.
Depending on the load packing it full of sand would help. Sand if contained can help with compressiong strength. Unless you use something that sticks to the pla, the main stress will still all fall onto the shells, resin works because it grabs to more pla but foam would have no effect in tensile stress. You should try epoxy mixed with sand, thats one hard cement, your parts will end rock solid
Every theoretical idea I have ever had when I 1st got into 3D printing, you have actually taken the time to run the tests! Du bist ganz Toll!!! Prost Herr CNC!!!
Would love to see this tested with PETG filament. In my experience PETG has perfect interlayer adhesion and never splits along a layer boundary when testing to failure.
This is called cohesive failure in PETG, which is preferred over adhesion failure in PLA as shown in the video. I expect the cohesion failure mode is better predictable because failure is less dependent of the environment conditions (filament surface impurities) and hence data is expected to show less scatter.
This is one of your best videos so far! And yay, new intro. So 0.15mm seems to be the optimal height for both looks and strength. I really love this test. Most people on forums are recommending higher heights for stronger prints, yet that's clearly not true.
Bigger layer heights coupled with bigger nozzle will be stronger. Here he is looking into nozzle and layer height ratio. I don't see people claiming bigger layer height is stronger, only that bigger nozzle will yield stronger prints.
@Blondie SL You should always use 0.2 for the first layer height, regardless of what you use for the rest of your print. It depends for the temp, but for PLA, I usually have 215'C for first and then 210'C for the rest. You can print a temperature tower to test your specific filament. Make sure your bed is really level and do a PID tuning as well.
I really appreciate not only your video's subjects, but the quality of the videography and video editing. You make things very clear visually along with your narration. Subscribed! Thank you!
How to complete any 3d print in just a few seconds: •1. Get 3d Model •2. Import it into software •3. Set layer thickness to the model's height •4. Print 1 layer
I know it's old, but I'd love to see this revisited with a larger nozzle to see if it's proportional to the size of the nozzle, or if nozzle size doesn't affect layer height strength.
I’d be curious to see a comparison of infill patterns along with their wall thicknesses, as well as the amount of overhang the higher layers have over lower layers: unless you’re making a vase where there’s not much change in that regard, I strongly suspect that makes a larger impact than you’d think. As an example of infill geometry differences, consider a strictly honeycomb grid with all vertical walls for a given infill percentage compared to also doing a honeycomb grid of prisms growing/shrinking: intuition suggests the vertical walled honeycomb grid would be stronger for layer adhesion between layers, but there are also intuitive thoughts to consider perhaps a non-regular vertical pattern may also be stronger by mixing up what’s stretched and where, for leverage.
Hallo Stefan, I think what you proofed beyond any doubts is the fact how many fold, complex and different the factors are, impacting those test results. One thing I noticed that got forgotten is humidity. I could imagine that there also might be a sweet spot to be found in which a certain filament will develop its maximum strength. I think you do an exceptional good job at creating awareness about the challenges of fine tuning a 3d print. As I’m still new to 3d printing with little to no experience your tests sure are an eye opener about the true science behind this technology. And it is very interesting to read all the comments of obviously highly knowledgeable follower of yours. Thank you and to all who contribute to this excellent knowledge exchange!
I specifically bought a larger nozzle size to be able to print thicker layers. Nice to see that my gut feeling was right. I run a 0.9mm nozzle (close to maximum for 1.75 filaments) and print 0.4mm layers.
I voted for 0.2mm height as the strongest. My reasoning is that of your tested heights, that is the one that provide the thickest strand without making the layer delaminate in the horisontal plane. I look forward to see the results when I unpause.
Thanks for the demo. But I think it is depend on weak chains between layer small or big layers. But Smaller layers has more chains and increase probability of risk..
How about your print fast = thick layers and put the part in the oven just few degrees under the deflection temperature according to your data sheet, which means the part will heat evenly and cold out evenly, we offer this application to strengthen the PLA
I knew it!!! I have found out myself that in thicker layers lines don't bond as good, and smaller layers fuse together better, not only in XY but in Z as it heats up the previous layer producing a better bond this is why I always print at .12 if I can, thick layers are bad for strength, I was right! thank you Stefan for this video, it's very informative!!! there are a TON of misconceptions and myths taken as truth in 3d printing, your testing is a godsend, keep doing what you do and let's hope to get rid of misconceptions!!!
That is not always the case though. I noticed that when printing using my Volcano hotend and setting it to the same temp as my e3d v6 that the filament bubbles because it's too hot. Lowering the temp and printing thicker layers on the Volcano yields much stronger prints than when using a regular e3d hotend. I'd love to see Stefan scientifically comparing them and how they affect strength.
Very well done. I thank you so much for the results. You have saved me some time and i have gained knowledge because of your hard work. Thank you fine sir.
Great test as usual! Since you always test 3 parts, the only thing I am really missing is some error bars with a standard deviation (2 sigma (σ) would be fine here I guess) in the first test.
I love that I actually feel like I'm learning something and enjoying it while listening to you talk about this. Definitely subbed just for that and look forward to more. As for the test itself, it's definitely helping me to think about what methods I can use for my first FDM printer to see the different results. This already gave me a good starting point for when I'm ready to print more than the must-prints and general fixes for the printer I'm getting (Ender 3 Pro)> So thanks a ton for making this video.
Hi... I'm a fan of these subjects, I suggest using hot air and a thinner nozzle with the use of pulsed frequency (spraying the layer), it may be better to fix the layers and the most suitable type of material such as carbon ....thanks!!
Not that you're necessarily advocating them, but I have a problem with the "magic numbers" that people like to use. "magic numbers" for layer height are a myth. If microstepping is enabled at all, there's no guarantee that you're starting at a full step at the first layer, therefore you may end up using only microstep intervals all the way up if you set a "magic number" as your layer height. If you want to really use the magic numbers, you should DISABLE microstepping, or cut it down to at least 1/4. This will mean you have better holding torque (better accuracy) for every layer, which is the whole point of the "magic numbers"
@@RentableSocks oh, but I use use exactly 1/4 microstepping. And I make sure to start first layer at full step. And calculating those "magical" numbers seems to be less intense for microcontroller. So.... maybe not so magical after all?
@@cthulpiss how do you make sure you start first layer at a full step? 1/4 microstepping is plenty for Z. you get like 8x the holding torque compared to 1/16 microstepping. just using fewer microsteps in general is less intense for the microcontroller. People using "magic numbers" when they're still using 1/16 microsteps makes me facepalm.
I use a 0.5mm nozzle to balance print speed and surface roughness. The thickness of the layer is 0.25mm, and I felt that a thickness greater than 0.25mm would have a negative impact on strength, and the surface would be rough, making it less cost-effective. This video is very helpful because it matches my knowledge. thank you.
Have you looked at extrusion width, perimeter count and layer height as a function of resistance to layer separation? What I'm hinting at is that there must be an extrusion VOLUME (thinner layers, wider extrusion or thicker layers, thinner extrusion) where layer adhesion has a sweet spot. Having said that, due to less time spent keeping the already printed plastic warm on the thicker layers, it might make sense to print hotter/slower with a wider extrusion width and zero to minimal cooling. It would also be good if you could find the sweet spot for all of these variables and update an ongoing 'master setting' by incorporating your new findings into to further improve on the print quality or strength thus far. Finally, it would be great to see a test on thicker (more) perimeters and then 20% vs 100% infill. This is all on the Z axis vertically printed items.
Have you ever tested the strength of hooks printed at 45 degrees from vertical to compare with the horizontal and vertical ones? I'm wondering if there's any benefit, and if so, is it worth the extra filament used for the supports that I'd assume you'll have to use?
This test looks like it was performed at constant speed of the printhead. In light of the findings about decreasing effective extrusion at higher volumetric flow rates, this video definitely needs an update keeping volumetric flow constant and varying appropriately print speed (something a Voron will do easily even at 50 um layer thickness). Because if the effective extrusion decreases for thick layers, of course strength does as well.
Hey Stefan, great video! Have you ever tested the effects of not only speed but acceleration on prints? I feel like it's a setting a lot of makers either forget about or don't know about.
i am really curios about diffrent nozzles both price and quality and also diameter. would love to see a extensive test video about that , love your content
I super duper appreciate that you rigorously test these things to death. Your methodology is legit AF, exactly what a materials engineer would do (exactly what we did for materials tests when I was in engineering school) so I have 120% trust in your data. This was super enlightening because prior to this what I read was that thicker layers were better strength and that actually turns out to not be true, so from here on unless I'm printing something which I need a rush job on (and isn't load bearing) I will be printing exclusively in 0.2mm and below.
Very useful information, not only for the strength but for making a trade off between strength and printing time. Since I'm printing practical objects, I would never print at 0.5 after seeing this, and would probably use 1.5 most of the time with a nozzle diameter of 0.5, 0.6 or greater.
You need to tune your print temperature, print speed, and cooling fan speed for best layer adhesion with temperature towers and such for each given layer height. All of those settings are interconnected in how the plastic bonds with the previous layer. My prints virtually never break along layer lines like that. Your settings are definitely way off, so all your results are meaningless.
Your videos have affected how I use my 3D printer. It starts in my design of a part. Optimizing the orientation on the print bed for best result. Also temp, layer height and extrusion multiplier. This has saved me a lot of time and plastic. I tend to like my delta printer.
Stress concentrations. Although the average cross section probably is essentially the same, the thicker print cross sections create a stress concentration where there are steps. So some of the failure can simply be blamed on geometric imperfections.
im really wondering... Part adhesion... Since you printed multiple parts at a time, each part had a chance to cool down a bit, prior to receiving its next row. This would cause the part to cool down before receiving the next row. If you only print 1 at a time, would the results be the same as printing 6 at a time? I would hypothesis that the part which was printed one at a time would hold together better than the parts which were printed 3-6 parts at a time. Great video 😀
I did a Design of Experiment like this when I was in engineering school in 2018, compared strength of a part to layer height and infill %. Tested multiple specimens in an Instron, then created a 3D surface plot and used mini tab to create a regression. Here is the final equation we came up with: Strength( in MPa)= 18.58 + 51.7(Layer Height) + 0.371(Infill %) - 1.184(Layer Height * Infill %) 95% confidence interval compared to real world testing. Conducted using standard PLA.
That’s kind of what I would have expected. It seems like that test with the parts printed standing up would be a good way to improve the design of the hook. I’m a little surprised that the thinnest layer print was less sturdy, although I think you mentioned (and others have commented) this could likely be optimized by adjusting the settings. Thanks for sharing.
The gaps between the layers of your 0.3 and 0.4 prints would indicate that you're printing too cold or too fast for that nozzle size/layer height combo. You even note the under extrusion in the broken part. For more comparable results I suggest you normalize your print settings to make the volumetric flow rate match. The flow rate difference between 0.05 layer height and 0.4 is substantial. If you're not controlling for this difference, your test procedure will favour the lower layer height parts substantially. I think this warrants future investigation.
Still - these were printed @ 215C, that should be quite enough for PLA + Prusa combo.
@@cthulpiss Note he's also using a hardened steel nozzle. You can lose 10c at the nozzle tip. Usually needs 10-20c above the recommended temp, especially when pushing flow rate to ensure adequate melt. For me, I print production parts with a 0.6 hardened steel nozzle at 0.3 at 70mm/s. That's pushing around 12 cubic mm/s which is the limit for the V6 block. For PLA I'm needing 220-230 and with PETG 250-260. Lower than that and part strength is definitely affected.
Well explained, this was my first thought too. Needs to print slower for the thicker layer heights and make the flow rates the same.
@@AustinVojta Yes and in this case I think flow rate is just a proxy for Time in Melt Chamber. I also noted that he had clogs at the 0.05 layer height. Probably due to the flow rate being too low and the filament cooking in the chamber.
@@JasonZnack Could you please point me to some math data around that V6 block thermal capacity vs. materials like PETG and ABS?
Why am I watching this at 2am? I don’t have a 3d printer and are not planning to get one…
Hope it still was worth a bit of your time 😅 have a good night.
Asking myself the same darn question lol
get bambu A1 mini
get one
Oh you'll get one at some point
7:53 is the actual test
Doing gods work there bud
Thank you good sir
Thank you!
THANKS
thanks god.
First I was like what the hell kind of music is this, then I realized it's the sound of the testing machine.
HunGredy Dude I thought the same thing.
not gonna lie, it kinda sounds good with the music
Same
The test machine makes a sound which is roughly similar to a Moog keyboard in the tone of C sharp. The actual background music, being in the key of F sharp minor, fits in perfect harmony with it.
Haha I would of never noticed
I mostly make functional parts and love this type of data. Excellent work!
Exactly Joel! This is awesome!
Feel like almost everyone in 3D printing is just making pretty models, thank god for this type of video!
@@williamlewington3223 Agreed. That's what I see so much from 3D-printing too; people making "pretty" models. But I just love it when 3D-printed parts actually are useful and have a functionality. But of course, that's when material strength usually has a more important role and that's where videos like this one are great.
@@williamlewington3223 No! Thank Stefan for this type of video.
This why germans make great stuff :) They ask great questions although they might seem picky to others.
8:24 - Horizontal print comparisons (9:05 - conclusion)
10:29 - Vertical print comparisons
Thank you😊
Thanks
bro yaps too much.. u saved my time
1:35 This is directly effected by direction of your print. Thinner layer height would be better with the grain, and thicker layer height against the grain.
CNC Kitchen: *prints super detailed and intricate part
My prints: 12:18
That’s CNC Kitchen’s print.
@@bldjln3158 :o
I’m currently working on some heat treatment methods of PLA for strength gardening. It has been done in the past but not clearly defined in any reports. After heating in an oven for an hour at 220°F I recorded my PLA prints to have almost a 30% increase in strength between layer lines. I also tried boiling and quenching my prints with similar (but not as great) results. Might be something interesting for you to try!
Yes conditioning should improve the mechanical attributes of any Thermoplast.. The issue is to get the "right" procedure... which means temoerature and time and when to do it... I would say it's to be done imediately after the print at a temperature somewhere between Printing bed temperature and max. Permanent temperature of the material.. (and except PA) it definiteky, should be in a dry environment.. With PA it could be that boiling it, maybe the right procedure.. Everything else I would definiteky go for the oven.
There's a lot of discussions out there that address annealing. Yes, you can get some pretty nice strength increase from a simple "heat and hold then cool" where time and temp are the only considerations, but I learned from doing flame glass work that there's a bit more to the science of annealing. Results widely vary dependent on the material, with some showing great increase and some showing virtually none. Additionally, "stepped" heating and cooling stages can make dramatic differences in the molecular rearrangement. It really depends on if you're just doing a "quickie" for a simple print or trying to make a particular item as strong as it can be. One other BIG factor in some jobs is that reheating and cooling the part can result in SIGNIFICANT changes in size, made worse by the fact that it's usually different in different axes. Z usually shows little change, with X and Y showing amounts that differ from each other AND Z. This is a real nightmare if you're making a interlocking or sliding part that joins another! So yes, for a simple part, many plastics do show some strength increase but it's HIGHLY variable and you may find yourself reprinting from the unintended size changes. Tip - Many of the better filament producers give annealing info on their web pages or will tell you if you ask them. Same for the proper drying time and temp as many plastics are hydroscopic (absorb moisture) especially the nylon and polycarbonate blends. Printing with "damp" plastic REALLY screws strength and surface finish.
may be worth looking into constructing a device that can rotate the part while being heated for tempering, so that gravity doesnt warp the final part
The important temp to go over is the glass transition temp, and the temp to stay under is the melting temp. The time is important because the heat has to conduct through the part to get in and out, and there are also thermal expansion effects. Rapid heating and cooling means the external surfaces are hotter or colder than the internal material, so the external material is more or less heat expanded than the inner material, while the heat is getting in or out. For example, rapid cooling means the outer layer shrinks and hardens while the inner material is still expanded, so when the inner material does cool, shrink, and harden, it leaves the outer layer under slight compression and the inner material under slight tension. This can make it stronger because bending failures start with tension at the surface, but when it does fail, it fails more catastrophically as the tension failure crack reaches the inner material that's already under a little tension.
It’d be really interesting to see a test with the print lines angled at 45 degrees throughout the object. It’d take some supports to print, but might combine the benefits of both directions of printing.
Interesting. I was think of this. So it's possible?! Oh, you mean tilting the model.?
That would make it worse for both lol
Did a owl like that ugly as hell of a print but the ear didn't chip off like on the 90°vertical but that's was cleaner than the 45°
@@DaStuntChannelnot worse. Instead of 100% vertical but zero horizontal strength it would be 50%/50% which, depending on the circumstances is actually the better deal.
@@wo-olf a bit more isotropic
Thank you! I’ve done testing on carbon fiber filled nylon, and .1 layer height was definitely stronger in both planes. Dialed in the settings (with a bit of help of the manufacturer) and now produce incredibly strong parts.
I’m very interested in this what did you do?
.1 layer height and what nozzle size? Cause if you print CF at .1 with a .4 nozzle you're getting the perfect recipe for a clogged nozzle.
This is a good demonstration of why it's important to make sure your layers aren't stacked parallel to the load direction. I didn't realise it was so critical to tensile (and probably compressive) strength.
its kinda like splitting wood against vs with the grain.
Idk know about compression i think that it will be stronger in compression’s when stacked parallel
@@phenax1144it would definitely be stronger in compression.
@@Kspice9000 I agree but the direction force is giving inaccurate results in this experiment. If he compressed the hook at the same point it would have torn from the outside edge instead of the inside edge from pulling. I bet if the force was applied evenly around the section being tested it would have withstood more.
I just ordered my first 3D printer and must say as an engineering student I greatly enjoy your videos! I look forward to seeing more of your videos in the future.
0.15 seems to be the sweet spot for strength and appearance on your printer with that filament. Cool!!
Please do the same tests with other filament like petg and polycarbonate, because if your printing practical parts you probably wont be using pla for them
PLA is much more resistant to impacts than petg.
@@dudemann. ua-cam.com/video/aODcbK5r5wE/v-deo.html is the best example i can find how petg compares to pla. petg just completely fails on any decent impact. ua-cam.com/video/m1I5zttnYjY/v-deo.html also shows this.
@@pomprocks I'm not gonna make a carabiner out of a material that will melt if left in the sun for too long ;P
Like your pic but mine is better.
PLA+?
My favorite settings are using 0.3 mm adaptive layers in Cura. Prints quick, looks pretty good and seems more than strong enough.
I prefer 0.2..0.25mm as it gives the nicest looking prints but 0.3 is still somewhat decent as well. 0.1 and below gives nasty and unstable results on my printer and with 0.4 the layers do not stick wery well especially with cheap plastics from China and the process tends to give very fragile parts and with thin walls it can easilly not even get printed successfully at all. I would say that 0.2 is the sweet spot, 0.15 is for small parts with small features and 0.25 is for when you are in a hurry.
yep that confirms my experiences, the best results I get with 0,5 mm nozzle and 0,15 layer height, at this settings I get the best looking prints and also the strongest ones
Excellent comparison. From what I can see here, a 0.4mm nozzle and 0.15mm layer height has the highest tensile strength in both print orientations and is near the top for finish quality too.
I really love how meticulous you are with your tests, and the effort you put into designing the tests so as to eliminate any possible bias.
If you even want to revisit or explore this further, it would be interesting to see how using a specialized hotend such as an E3D volcano would change the results, considering that they touted their thicker layers as stronger in their promotional material.
Personally I thought the 0.2 would be stronger being a 'balance' between thick and thin, but clearly I need to re-evaluate that line of thought.
Great work!
Hi
Hi
Hi
Dude! Amazing seeing you here! I know you prob get this a lot, but I love what you do with tesla coils!
Hi 🐑
I would be interested in extrusion width tests, I've been doing a lot of spiral printing in PLA and PETG with 1.2mm width with 0.3mm layer height on my Ender 3, really awesome results (great for making quick boxes), feels really strong too.
Designing a weld-less 18650 battery build system around a 0.6 mm nozzle and your videos have been extremely valuable for balancing strength vs print speed. Thanks!
This is an incredible video. Great job, Stefan! I use 0.3mm height quite a bit on my Mk3 and I mistakenly thought that would make things stronger. However, I have a Cr10 set up per your TitanAero video, with a volcano, 0.8mm nozzle, 0.4mm layer height and extruding at a 1.2mm width. I feel the strength in that case is insanely strong. There is probably a relationship to the width vs layer height. You could repeat these experiments with those relationships maintained as the layer height goes up. I think there is more to be done along these lines, and I can't wait to see more videos on the topic.
I'm new to 3d printing, actually waiting my ender 3 pro to arrive in few days. Your videos answered many questions I had. thank you !
Wow ... Wasn't expecting that at all ... I even bought a 0.8mm nozzle "to make stronger part" back in the day, because i was thinking that 0.6mm layer height would improve the strenght of my parts ... Thanks for your work !
But maybe layer width can improve strength
You are misunderstanding this test greatly. The point is for same nozzle size what layer height is going to be the strongest. Bigger nozzle will give stronger prints.
You also need a hotend that can supply enough heat to the much higher amount of plastic flowing through it. The ability to get heat into the plastic is going to determine your max print speed.
It's probably been said a bunch of times, but yeah. The larger diameter parts don't appear to have been sliced properly. There's also a considerable difference in the layer heights and flow rates between the smaller nozzle prints and the larger nozzle prints. Looking at this at first I had assumed you were using a 0.8mm nozzle at the largest. I've never seen 0.4mm nozzle prints that looked that bad before.
Thank you for putting all that effort into testing these 3D printing parameters for us. This information is very useful and you have just saved us hours of testing. Danke schön!
Very very good. The unefull thing that I don't like is that I wanted to do the same, but you did it first, while I slept. Wonderful work.
Please test with different nozzle-sizes and combined with the layer-hights.
I love those scientific graphs. All this research is quite professional so it's really useful. Great job!!
The conclusions of this video are only relevant to a .4mm nozzle. That is very important to note. You touched on the proportionality issue when you referenced the paper at 11:21. This same issue is relevant throughout the entire video. A more accurate scientific study would have changed the nozzle with each change in layer thickness to maintain the proportionality.
Correct. This is most likely also a reason where the under extrusion is coming from @0.3/0.4 layer height
Design Prototype Test you mean more general (not accurate) i think. I pretty much only print with .4 mm nozzle...swapping them out is significant work and I’m guessing most people do the same so it’s valuable for me.
This is a proper study to show that a given nozzle can print at a given layer height range only. It IS shown and summarized in the video, that the optimal layer thickness is proportional to nozzle diameter. Probably it is more important to know what layer thickness to choose for the nozzle you are using, as it is a more realistic scenario that you have a fixed nozzle and you adjust the print parameters than changing the nozzle every time.
It would be interesting to see which absolute layer thickness is stronger if all would be printed using it's optimal nozzle diameter, but the test in the video is more of a real world use case.
I'd like to see the challenge be changed to "print this part as strong as possible" and let experienced makers use whatever techniques they have in their toolboxes and apply them. Even specify a wall thickness and let us tune for that. I'm really bothered by the implication that "0.15mm is best" without more detail. Let me use a different nozzle and extrusion width and see how strong the part can be. Great testing, but incomplete and vague conclusions.
@@theroyal3-5643 If that's the case, then he should have acknowledged up front that 0.30mm layers are at the upper end of what you can print for good adhesion at that nozzle size and 0.40mm is almost guaranteed to produce poor results. The testing is excellent (as always) but the conclusions were vague. A test for "how strong can you print this part with a 0.40mm nozzle" might yield altogether different conclusions if different approaches are used.
I believe that the gap used during bed leveling will influence layer adhesion for different layer heights
This is an awesome topic! Interesting how temperature affects the strength gap of vertically and horizontally printed parts. I'm nearly certain a little increase in printing temperature could add a lot of strength to the horizontally printed(the weaker) hooks.
Up to a certain point, more temperature will increase layer adhesion, but may cause u other issues.. Polymeres are a bitch.
Donate to this man. He is doing a great service. Thank you Stefan!!! You rock.
Im editing a video about foam filling 3D prints. The experiment was *mostly* a successful and I would love to hear about how to strengthen prints.
EDIT: The results of my video were pretty amazing! Hopefully you can check it out and see what you think.
Printing hollow and filling with resin would probably be pretty stong with the right stuff
@@UNSCPILOT agreed, however thick resin can still be incredibly heavy from what I have tested. I just finished my mission, the video is up!
Depending on the load packing it full of sand would help. Sand if contained can help with compressiong strength. Unless you use something that sticks to the pla, the main stress will still all fall onto the shells, resin works because it grabs to more pla but foam would have no effect in tensile stress.
You should try epoxy mixed with sand, thats one hard cement, your parts will end rock solid
@@laharl2k his channel looks like mostly props so i don't think putting cement inside his prints would be the best direction.
Every theoretical idea I have ever had when I 1st got into 3D printing, you have actually taken the time to run the tests! Du bist ganz Toll!!! Prost Herr CNC!!!
Would love to see this tested with PETG filament. In my experience PETG has perfect interlayer adhesion and never splits along a layer boundary when testing to failure.
This is called cohesive failure in PETG, which is preferred over adhesion failure in PLA as shown in the video. I expect the cohesion failure mode is better predictable because failure is less dependent of the environment conditions (filament surface impurities) and hence data is expected to show less scatter.
I'm doing a lot of "UA-cam University" studying before buying my first printer. Fascinating video. I appreciate your hard work!👍
This is one of your best videos so far! And yay, new intro.
So 0.15mm seems to be the optimal height for both looks and strength.
I really love this test. Most people on forums are recommending higher heights for stronger prints, yet that's clearly not true.
At the end of the video he says the ~2.6 Nozzle to Layer Ratio might be a better target to meet than just layer height.
Bigger layer heights coupled with bigger nozzle will be stronger. Here he is looking into nozzle and layer height ratio. I don't see people claiming bigger layer height is stronger, only that bigger nozzle will yield stronger prints.
Hobby Hoarder the prettiest parts will be strongest, I suppose!
@Blondie SL You should always use 0.2 for the first layer height, regardless of what you use for the rest of your print. It depends for the temp, but for PLA, I usually have 215'C for first and then 210'C for the rest. You can print a temperature tower to test your specific filament. Make sure your bed is really level and do a PID tuning as well.
I really appreciate not only your video's subjects, but the quality of the videography and video editing. You make things very clear visually along with your narration. Subscribed! Thank you!
How to complete any 3d
print in just a few seconds:
•1. Get 3d Model
•2. Import it into software
•3. Set layer thickness to the model's height
•4. Print 1 layer
I know it's old, but I'd love to see this revisited with a larger nozzle to see if it's proportional to the size of the nozzle, or if nozzle size doesn't affect layer height strength.
Das „Guten Tag“ hats gerissen. Abo ist da
I have never wondered this and don’t even own a 3D printer but I still watched the entire video
I’d be curious to see a comparison of infill patterns along with their wall thicknesses, as well as the amount of overhang the higher layers have over lower layers: unless you’re making a vase where there’s not much change in that regard, I strongly suspect that makes a larger impact than you’d think. As an example of infill geometry differences, consider a strictly honeycomb grid with all vertical walls for a given infill percentage compared to also doing a honeycomb grid of prisms growing/shrinking: intuition suggests the vertical walled honeycomb grid would be stronger for layer adhesion between layers, but there are also intuitive thoughts to consider perhaps a non-regular vertical pattern may also be stronger by mixing up what’s stretched and where, for leverage.
That would be a good idea. Plus, given that we know the properties of 0.05 to 0.2 are roughly the same, we can just print at one of those settings.
Sincere Thank you for making this. Im using your data to improve my work. i cant afford school, this is just as good for what i need. Thank you.
Hallo Stefan, I think what you proofed beyond any doubts is the fact how many fold, complex and different the factors are, impacting those test results. One thing I noticed that got forgotten is humidity. I could imagine that there also might be a sweet spot to be found in which a certain filament will develop its maximum strength.
I think you do an exceptional good job at creating awareness about the challenges of fine tuning a 3d print.
As I’m still new to 3d printing with little to no experience your tests sure are an eye opener about the true science behind this technology. And it is very interesting to read all the comments of obviously highly knowledgeable follower of yours. Thank you and to all who contribute to this excellent knowledge exchange!
I specifically bought a larger nozzle size to be able to print thicker layers. Nice to see that my gut feeling was right. I run a 0.9mm nozzle (close to maximum for 1.75 filaments) and print 0.4mm layers.
I'd like to know how to get the strongest possible layer adhesion with high temperature filaments like nylon or polycarbonate
Buy, and test, and find out
7:00 I’m intending to 3D print minis, so this particular part was nice!
I'm wondering if these results would change significantly (in terms of the strongest layer thickness) if the parts were annealed?
I never comment on videos, but actually you are doing an extremely well designed work, scientific with a lot of effort! Thank you!
Love your straightforward and informative videos, Stefan.
I love how the separation looks so clean when they snap.
I voted for 0.2mm height as the strongest. My reasoning is that of your tested heights, that is the one that provide the thickest strand without making the layer delaminate in the horisontal plane.
I look forward to see the results when I unpause.
Thanks for the demo. But I think it is depend on weak chains between layer small or big layers. But Smaller layers has more chains and increase probability of risk..
It would also be interesting how much aceton-smoothing and similar affect the layer-adhesion of different layer heights.
How about your print fast = thick layers and put the part in the oven just few degrees under the deflection temperature according to your data sheet, which means the part will heat evenly and cold out evenly, we offer this application to strengthen the PLA
I knew it!!! I have found out myself that in thicker layers lines don't bond as good, and smaller layers fuse together better, not only in XY but in Z as it heats up the previous layer producing a better bond
this is why I always print at .12 if I can, thick layers are bad for strength, I was right! thank you Stefan for this video, it's very informative!!! there are a TON of misconceptions and myths taken as truth in 3d printing, your testing is a godsend, keep doing what you do and let's hope to get rid of misconceptions!!!
That is not always the case though. I noticed that when printing using my Volcano hotend and setting it to the same temp as my e3d v6 that the filament bubbles because it's too hot. Lowering the temp and printing thicker layers on the Volcano yields much stronger prints than when using a regular e3d hotend.
I'd love to see Stefan scientifically comparing them and how they affect strength.
Very well done. I thank you so much for the results. You have saved me some time and i have gained knowledge because of your hard work. Thank you fine sir.
Great test as usual! Since you always test 3 parts, the only thing I am really missing is some error bars with a standard deviation (2 sigma (σ) would be fine here I guess) in the first test.
Thanks for the very educational study. Definitely confirmed print direction plays a huge role in strength
So thorough, love how in depth you get and scientific! Although, I can't watch these videos in bed, I doze off! Love you! Keep it up!
I am just getting started in 3D printing. I am very impressed with your demo. Very useful info.
I love that I actually feel like I'm learning something and enjoying it while listening to you talk about this. Definitely subbed just for that and look forward to more.
As for the test itself, it's definitely helping me to think about what methods I can use for my first FDM printer to see the different results. This already gave me a good starting point for when I'm ready to print more than the must-prints and general fixes for the printer I'm getting (Ender 3 Pro)> So thanks a ton for making this video.
Hi... I'm a fan of these subjects, I suggest using hot air and a thinner nozzle with the use of pulsed frequency (spraying the layer), it may be better to fix the layers and the most suitable type of material such as carbon ....thanks!!
Considering the microstepping on Z-axis, it seems that 0.12 or 0.16mm layers are best - good balance between printing time and strength.
Not that you're necessarily advocating them, but I have a problem with the "magic numbers" that people like to use.
"magic numbers" for layer height are a myth. If microstepping is enabled at all, there's no guarantee that you're starting at a full step at the first layer, therefore you may end up using only microstep intervals all the way up if you set a "magic number" as your layer height. If you want to really use the magic numbers, you should DISABLE microstepping, or cut it down to at least 1/4. This will mean you have better holding torque (better accuracy) for every layer, which is the whole point of the "magic numbers"
@@RentableSocks oh, but I use use exactly 1/4 microstepping. And I make sure to start first layer at full step. And calculating those "magical" numbers seems to be less intense for microcontroller.
So.... maybe not so magical after all?
@@cthulpiss how do you make sure you start first layer at a full step? 1/4 microstepping is plenty for Z. you get like 8x the holding torque compared to 1/16 microstepping. just using fewer microsteps in general is less intense for the microcontroller.
People using "magic numbers" when they're still using 1/16 microsteps makes me facepalm.
Wow, what a lot of work! Thanks for doing all this. I usually print at 0.2 but I might try 0.15 for things that require strength.
THANK YOU for scientific testing methods, rather than some random testing method!
STEFAN: Your videos are the most well presented, to the point and helpful on YT imvho...been printing for quite a while...always on point!
E3D Volcano 0.6 nozzle, 0.4 LH prints stronger, faster and good looking.
Probably at slower speeds but still finishing the print more quickly. This is often overlooked.
I use a 0.5mm nozzle to balance print speed and surface roughness. The thickness of the layer is 0.25mm, and I felt that a thickness greater than 0.25mm would have a negative impact on strength, and the surface would be rough, making it less cost-effective. This video is very helpful because it matches my knowledge. thank you.
Never did a test like this but from years of experience i also found that 0.15mm is the golden ratio between speed,detail and strength.
I use 0.16 layer, it's also not bad for 0.4 nozzle. Better to take even layer steps, depending on nozzle dimeter.
Have you looked at extrusion width, perimeter count and layer height as a function of resistance to layer separation?
What I'm hinting at is that there must be an extrusion VOLUME (thinner layers, wider extrusion or thicker layers, thinner extrusion) where layer adhesion has a sweet spot.
Having said that, due to less time spent keeping the already printed plastic warm on the thicker layers, it might make sense to print hotter/slower with a wider extrusion width and zero to minimal cooling.
It would also be good if you could find the sweet spot for all of these variables and update an ongoing 'master setting' by incorporating your new findings into to further improve on the print quality or strength thus far.
Finally, it would be great to see a test on thicker (more) perimeters and then 20% vs 100% infill. This is all on the Z axis vertically printed items.
Have you ever tested the strength of hooks printed at 45 degrees from vertical to compare with the horizontal and vertical ones? I'm wondering if there's any benefit, and if so, is it worth the extra filament used for the supports that I'd assume you'll have to use?
Very interesting test results and good testing regime. Well done.
5:15 - O point gangnam style
This test looks like it was performed at constant speed of the printhead.
In light of the findings about decreasing effective extrusion at higher volumetric flow rates, this video definitely needs an update keeping volumetric flow constant and varying appropriately print speed (something a Voron will do easily even at 50 um layer thickness).
Because if the effective extrusion decreases for thick layers, of course strength does as well.
What if you put your entire printer in a vacuum chamber, and build something under a very strong vacuum!
With all your research a chart for optimum layers speeds and thicknesses would be really helpful.
Hey Stefan, great video! Have you ever tested the effects of not only speed but acceleration on prints? I feel like it's a setting a lot of makers either forget about or don't know about.
i am really curios about diffrent nozzles both price and quality and also diameter. would love to see a extensive test video about that , love your content
i watch this on 2x speed and still hear the german accent
I'm Russian, but even I can hear this wonderful accent xD
not even german. U stupid?
@@Nathansthing Bruh he literally opens bis video with "Guten Tag"
i didnt hear any german
But have you tried listening to it in mono? Or how about with the lights off?
I super duper appreciate that you rigorously test these things to death. Your methodology is legit AF, exactly what a materials engineer would do (exactly what we did for materials tests when I was in engineering school) so I have 120% trust in your data. This was super enlightening because prior to this what I read was that thicker layers were better strength and that actually turns out to not be true, so from here on unless I'm printing something which I need a rush job on (and isn't load bearing) I will be printing exclusively in 0.2mm and below.
SPOILER ALERT:
Looks like 0.15 is the best all around for strength and appearance. Interesting findings. Thanks for doing the tests.
Unter the assumption of a 0.4 nozzle...
Very useful information, not only for the strength but for making a trade off between strength and printing time. Since I'm printing practical objects, I would never print at 0.5 after seeing this, and would probably use 1.5 most of the time with a nozzle diameter of 0.5, 0.6 or greater.
You need to tune your print temperature, print speed, and cooling fan speed for best layer adhesion with temperature towers and such for each given layer height. All of those settings are interconnected in how the plastic bonds with the previous layer.
My prints virtually never break along layer lines like that. Your settings are definitely way off, so all your results are meaningless.
Information on your setup & parameters please!
The chosen music and your test machine where playing in perfect harmony!
12:17 abandon ship!!
Your videos have affected how I use my 3D printer. It starts in my design of a part. Optimizing the orientation on the print bed for best result. Also temp, layer height and extrusion multiplier. This has saved me a lot of time and plastic. I tend to like my delta printer.
Excellent video.......and the exact inverse of what I would have guessed! (Nice to also know why I was wrong!)
Stress concentrations.
Although the average cross section probably is essentially the same, the thicker print cross sections create a stress concentration where there are steps. So some of the failure can simply be blamed on geometric imperfections.
I wonder if results will be the same for 0.8 or biggers nozzles?
im really wondering... Part adhesion... Since you printed multiple parts at a time, each part had a chance to cool down a bit, prior to receiving its next row. This would cause the part to cool down before receiving the next row. If you only print 1 at a time, would the results be the same as printing 6 at a time? I would hypothesis that the part which was printed one at a time would hold together better than the parts which were printed 3-6 parts at a time. Great video 😀
Your nozzle was wearing down, that’s why it didn’t printed well
I'm researching with 3D printed metals and your video has just gave me some ideas. Thank you so much. Excellent video.
Liked & subscribed.
"Guten Tag everybody" Feier Ich :D
Feier ich auch.
Ist er deutsch oder wrm macht er das?
@@ABCjrjfb Dachte ich auch, dass er Deutscher ist
I did a Design of Experiment like this when I was in engineering school in 2018, compared strength of a part to layer height and infill %. Tested multiple specimens in an Instron, then created a 3D surface plot and used mini tab to create a regression. Here is the final equation we came up with:
Strength( in MPa)= 18.58 + 51.7(Layer Height) + 0.371(Infill %) - 1.184(Layer Height * Infill %)
95% confidence interval compared to real world testing. Conducted using standard PLA.
"That's what she said"
We meet again
That’s kind of what I would have expected. It seems like that test with the parts printed standing up would be a good way to improve the design of the hook. I’m a little surprised that the thinnest layer print was less sturdy, although I think you mentioned (and others have commented) this could likely be optimized by adjusting the settings. Thanks for sharing.