Filaments need good tests: This is how I will be reviewing 3D printing filaments with the Filaween 2 tests!
What’s up everyone, Tom here, and I think the original Filaween series of filament tests was pretty good, it was pretty decent. We learned a lot about how the different materials and filaments behave under mechanical stress and what the differences were in print quality. But I feel like there are a few things that still needed a bit of tweaking. So after testing 30 materials, Filaween 1.0 is over and Filaween 2.0 is about to start. The king is dead, long live the king!
So there are few areas that I worked on. That is generating data that is more reliable, presenting stuff in a way where it’s easy to understand the results and making the entire thing a bit easier for me, too. Let’s start with reliability. Filaween 1.0 relied a lot on the manufacturer giving correct information for temperatures and fan settings and also producing the filament with a diameter as close as possible to the nominal 1.75mm. I really didn’t do a ton of tuning for any material, and while I still can’t spend a day trying to dial in every material from zero, I at least want to do a basic extrusion amount and temperature tune. So what I came with here is this stitched series of prints that does two things. First it prints these cylinders with 100% infill with 5% extra material for each consecutive one, so they are 100%, 105, 110 and 115% extrusion multiplier in one go and I’ll be able to feel and visually tell which ones are overextruded and which ones are not. In this case, you can feel the top layers getting a bit rougher with each step, the 100% one is perfectly smooth, while the 115% one is visually rough and overstuffed on top. In this case, I’ll go with 105% to get a good fill for strength without degrading print quality too much. By the way, these parts have a divot on the bottom, which will prevent nozzle height offsets on the first layer from influencing the rest of the part.
The other section of that print is just three 3DBenchies, the center one at the median print temperature, the real one colder, the front one hotter. In this case, 190, 210 and 230°C. I did think about just doing, like, an overhang test, and ramping up the temperature for that, but the 3DBenchy just has a lot more information on how things are going. In this case, the 190°C one looks good, so does the 210° one, but the 230° Benchy does have a good bit more drooping with bridges, so 210°C it is. And that is the quick & easy extrusion amount and temperature test that will run for every single material. This is almost an 8 hour print, as always, running on the Original Prusa i3 MK2, but for this season, with updated print settings. I’m going from the previous 0.2 to the new “Optimal” profile with 0.15mm layer heights. Prints do look a bit better, and most importantly, are now using the “recommended” profile for this machine. We’re also switching to 15% cubic infill, the more modern pattern, which will come into play in a second with the new mechanical test. Other than that, I will be printing as much as possible as sequential prints on the same printbed, which means each part is going to be printed at its own pace and not suddenly see speed changed when another one finishes. I’ve done that before, too, but I’ll try to consolidate as many prints into one job as possible so that, after tuning, I only have to hit print once and can then leave the machine to itself.
Alright, mechanical tests. What was really bothering me was that I didn’t have any sort of toughness test included in the original Filaween series. That means that for example PETG and a Nylon would see similar resulting numbers, but the PETG would crack and shatter much, much more easily than the Nylon, and that didn’t have any impact on the results, but could drastically change how durable a material was under real-life conditions. Thankfully, there are already some standardized test for this exact property, which is impact strength. So what I did is that I actually built an apparatus that tests like ISO 180/B. Basically, we swing a hammer from a set height, which breaks the test piece, the specimen, and by looking at how far the hammer swings up after it hit the part, we know how energy our specimen absorbed while breaking. The specimen is this 10mm square bar with a notch in the center, so as the hammer strikes, there’s going to be a stress concentration and it’s going to break at that exact spot. There will be three different prints tested, one printed in each different orientation.
Now if that sounds too complicated, worry not, because in simple terms, we just get a number for how tough a material is, how much of blow it can absorb. So in this test, Nylon would definitely get a higher score than PETG.
Of course, I’ll keep the classic mechanical tests, the parts for those look slightly different, the bend test is slightly optimized, it still works the exact same way, but the pull test, which used to be these octagonal rings, is now more of an elongated ring, which means there should be less bending and more pulling happening, which should make it a bit fairer between rigid and bendy materials.
And one new test is for that exact flexibility. I didn’t have a really precise indicator for that so far, so this is what I’ll be using – it’s a printed spring, I’ll hook a 1kg spool of filament to it and just check how far it stretches. It is rather flexible, which is intentional, I did go through a few different iterations to make this as soft and as repeatable as it is, but I feel like it’s actually more important for the rigid materials to know precisely how rigid they are, because the softer ones, you’re probably not going to use those where you need a very specific hardness. But the data will be there!
Still, I don’t want to base Filaween 2.0 so much on the raw numbers anymore, well, actually, I just don’t want them to be the only final result. I’ll still include the numbers, but the actual “decent or crap” scale is going to be stars. From 0 to 5 stars. Uhm, well, actually, uh, the five-star thing has been spoiled a bit by Amazon reviews, where it’s not like “3 stars is average”, but something closer to this. And I feel like I kinda have to play along here, so 4 stars is going to be average, that is perfectly acceptable, like with a 4-star filament, I’ll print that stuff each day, every day, no second thoughts. 5-star is going to be your “definitely better than average” rating and for truly outstanding materials, like ones that I don’t manage to break with my setups or that get really good, consistent print quality through the bank, those will get a flaming 5-star of awesomeness. I’m 100% serious here, by the way.
A filament will get an overall rating, and that breaks down to aesthetics, which is pure print quality plus the look and feel of the material, then of course mechanical properties, and ease of use, which is how much tuning was necessary, how close the tuned settings were to the recommendations and what hardware you need to successfully print with a material.
So as a baseline, and sort of as a reference of what to expect, here’s how some previous materials would score: Rigid.ink ASA and the UniCoFil Rainbow would both get the flaming 5-star for print quality, both print really well, and the Rainbow PLA has, I think, an awesome look to it on top of that. On the other hand, many particle-filled materials, including woodfills, do have a unique feel to them, but print quality is often a bit harder to get right, so they would end up somewhere in the 3-star region. For strength, Polymaker’s Polyflex, Taulman Bridge, and probably also Polycarbonate would get that 5+ star, most PLAs would end up in the 4-star region, while the ABS or ASA filaments would probably get 2 to 3 stars mostly. Now, for ease of use, to be honest, PLAs are just as easy as it gets. They need no special nozzles, aren’t really affected by moisture, don’t need a heated chamber, and so on, so there’s really no reason not to get them the flaming 5-star for ease of use. The other extreme would be a carbon-filled Nylon, which will absorb ambient moisture, often warps a lot and requires a specific build surface, and it also need a wear-resistant nozzle, so that would maybe get a 2-star at most, especially since you often need to tune the print profiles a bit.
I’ll have to see how the new tests work out over that wide range of materials, but I’m definitely looking forward to the first material that gets an overall flaming five-star rating. I’m not quite sure if any of the materials I’ve tested in Filaween 1.0 would qualify for that.
The individual Filaween 2.0 episodes are going to be published on Saturdays, all the ones from that week as a bundle, like some of you have suggested.
Now, if you want to dive deeper into the testing methodologies and scoring rules, etc, that info is going to be at toms.org/filaween2, the files are also for download, you can get the full CAD files for the IZOD setup, the test specimen and etc. on Onshape, that link is also going to be on the website, and I will also try to publish the gcode files that I use on the MK2 for all these tests. Different printers might produce slightly different results, so I’m not sure how comparable the results would be to parts made on a completely different 3D printer. So as of right now, I don’t think crowdsourcing tests results would be such a great idea for obvious reasons of trustworthiness etc, but you’re certainly invited to discuss the Filaween tests in the comments below here on YouTube or on the forums, at discuss.toms3d.org, and actually, the forums are open to discussions of any kind, hopefully preserving good information and conversations better than a YouTube comment thread can. Again, discuss.toms3d.org.
While we’re at it, a quick thanks to AlephObjects for supporting the channel. They’re a Free Software, Libre Innovation, and Open Source Hardware company from Loveland, Colorado, and not only are they the makers of the Lulzbot line of 3D printers, they’ve also been working on open-source filament!
Watch my reviews of both LulzBot 3D printers here and check the links in video description for more info on the machines or on the open-source filament straight from Aleph Objects.
So I hope you’ll find the Filaween 2.0 tests helpful, if you do, hit that like button, get subscribed so you don’t miss the episodes, and of course, you should turn on notifications by clicking that bell next to the subscribe button. If you want to support this channel, check out Patreon with regular Q&A Hangout, or, simply use the free affiliate links from the video description below for your next shopping spree wherever you buy your filaments.
Thanks for watching, and I’ll see you in the next one!