How to create transparent 3D prints!

Psst, hey kid, want some clear 3D prints? I can show you how!

I wanted to make something clear, but the problem is that typical 3D prints, even if you use clear filament, just come out milky white tone instead of being properly clear. They diffract light too much, so while the light still passes though, it gets bounced around inside the parts and scatters, which creates that translucent, but not quite transparent look. But Colorfabb have had this article online for a while where they are showing off incredibly, impressively clear prints. I’ve done a lot of printing and I think I’m ready to show you how to reproduce these prints yourself.

Okay, let’s see, Colorfabb were using their HT filament for this, which is Eastman’s Tritan, so I believe it’s the same material that Taulman used to make, just plainly called Tritan, too. But the key here is that it’s a copolyester, so a material that is very closely related to the common PET and PETG materials. In general, those are plastics that like to fuse together extremely well, so they turn into one solid piece of material instead of keeping those layer marks when you look through the print. I do have the Colorfabb HT filament here, but I wanted to see how well just a standard PETG would do, so I did all my experiments with the inexpensive DAS FILAMENT PETG, but I’m sure any other clear polyester filament is going to work basically the same for this. I had two different use cases here, the first one being these solid parts like Colorfabb showed that just look like they are one single, solid block of plastic with basically invisible layer lines on the inside and outside, at least that was the goal, the other one being vase-like structures that have no infill, but only provide a shell, in my case, to scatter light. I was specifically looking into this for use as diffusers for lamps of various types. And I think I pretty much achieved that.

But let’s start with the solid parts. My test subject here is a spur gear from Daniel Noree’s OpenRC F1 design, it’s got some relatively complex surfaces on the side, large flat areas on the top and bottom and this hub that is a bit of a smaller element where you’ll be able to see the light shine through from the top and the sides. I know, it’s not the most realistic design for something that’s supposed to look good, but I think it’s a great object to tune in the process. Plus, I’ve now clearly got plenty of spares for when I start building my first OpenRC models! I’ll probably reprint these from ABS or Taulman 910, though.

So the first print was done with settings that I thought would make sense – I used 300 and 350µ layers, 100% infill and regular temperatures. But if you look at the part printed with those settings, it is translucent, but it is not transparent. Yes, you can take a flashlight and shine it through and the part will glow, but it’s not nearly as transparent as the samples Colorfabb showed. So the first idea was to increase flow, to close what looked like tiny gaps between the extrusion lines, so I upped the extrusion rate to 105% first, and seeing that that improved things a bit I tried again at 110%. And if you compare the original part at 100% flow and the new one at 110% it is already a lot more transparent, but we can also see that it’s starting to show over extrusion streaking on the top surface. So at this point, the amount of material seems like its correctly filling every gap, but the rest of the print settings could probably still use some work. So I tried the other extreme, going with 50µ layers, and that immediately helped a lot. I think because the hotend now passes over the same spots more often and sorta irons down the tracks more, we’re fusing the individual extrusion lines together more tightly to turn them into an actual, single block of material. In fact, if you look at this hub part that I accidentally broke in half, you can see that both the surface where it broke and the inside of the part look like one single piece of PETG, looking at just the crack, you’d be hard-pressed to tell this thing was 3D printed at all.

So next up, I tried increasing the material flow on the 50µ prints and increasing the layer height from 50µ to 100µ, and both improved transparency a bit. The 100µ print was done with more perimeters, so you can see how those scatter light differently than solid criss-cross infill, but overall it didn’t look like that different from the 50µ prints other than the top surface being a bit less smooth and now looking more like it was overextruded.. The 50µ prints with more material I think were the best ones of the entire series, where the one with a total of 15% extra material looks, I think, extremely good, while the one with 20% extra is a bit too much and started to get quite messy on the top and side. It also has these fuzzies between the gear teeth where the nozzle was scraping off the extra material on each layer. For some reason, at the time I thought the 100µ prints were pretty much just as good as 50µ ones, so I kept on printing parts with that setting. Looking back, I probably should have stuck with 50µ.

Anyways, these two benchies were printed with 110% flowrate, but this one got a bit of a temperature boost, hoping that would help with everything fusing together. But instead, it turned out that the higher temperature actually decreased the clarity of the part and had a few other negative effects on quality. It’s a bit easier to see what exactly is happening when you look at the gears that used different temperatures , and you can see that, with a higher temperature, the PETG actually sorta starts cooking and bubbling, and we’ll see that effect again in a second. In the main section of the gears, the higher temperature is fine and slightly improves clarity, but up here at the hub, where the printer slows down so that it doesn’t pump too much molten plastic onto one area in too short of a timeframe, up here you can see the plastic getting extremely cloudy. That’s because when the printer slows down, the filaments is sitting in the heated zone of the hotend longer and gets more time to heat up, cook and degrade.

So to avoid that, either disable the cooling slowdown in the slicer or just print at a lower temperature overall. I think this Benchy does look really good and shows off the transparency or translucency really well, but it’s most visible when you look at the top and bottom of the parts, since it’s almost like the low layer height makes the side surface somewhat milky. Maybe I should try with an even lower layer height at some point, 10µ or something, but I’m pretty sure the Prusa i3 MK2 is not up for that without dropping in a new extruder and some lower-pitch Z-axis spindles. The stock setup has a physical resolution of 20µ, considering that the half-step position is the only microstepping angle you can really trust. But let’s get back to these parts.

I also tried to improve the surface by either flame polishing it with a hot air gun set to 600°C or sanding it beforehand, but with the hot air gun, this unsanded spinner started bubbling and softening up before it started getting more transparent, so that part’s gone, and when I tried to sand a part beforehand, yes, it did ultimately get a bit clearer, I think, after I heated the surface that was sanded to 1200 grit, but again, it’s really hard to find that optimum spot where the plastic neither bubbles nor softens up too much. But one nice thing you can do with heat, actually for any filament, is to melt off the little hairs that PETG in particular likes to pull when the hotend moves from one area to the next. High heat and a single pass are usually enough to take care of them.
And one last approach that didn’t really work with the Benchy was the thicker layers at 300µ, even with the extrusion multiplier cranked up, it did not turn out transparent at all. Though print quality overall was a bit more consistent compared to the low-layer-height parts.

So to recap for solid prints: With polyester filament, use a low layer height, 100% infill, obviously, and tune the extrusion multiplier so that you’re getting maybe a tiny bit of overextrusion. Higher temperatures can help, but it’s really easy to cook your filament that way. Post-processing prints to improve the surface is hard to do right, at least with heat. Maybe those thick coatings can help here, I’ve played around with spray-on clearcoats before, but they didn’t improve things at all.

Awesome, let’s check out how these parts were printed. These are all single-wall parts, most of these you could print with vase mode, which is a print mode that turns the entire print into one long extrusion line, instead of having discrete layers, it just continuously keeps moving the hotend up as it lays down plastic, so it’s one long spiral. However, this shroud does have a ridge up here, which usually can’t be printed cleanly with vase mode. But that’s not the point here. Let’s start out with what makes these parts look better or worse than others, and essentially it’s the same thing again. It’s refraction wherever light transitions from one medium into another, and here, that’s from the air into the plastic and back out. The layers act as tiny lenses because they have that round section on their end. That’s why this part right here will blur things in the vertical direction, but not as much horizontally. For example with this shroud, I tried to use that effect and create tiny microlenses that would also diffuse light horizontally because this is supposed to be a diffuser for a lamp. You know, you have these individual LEDs and just seeing each one through a clear shroud doesn’t make for a very attractive lighting setup, so I have all these parts that create different diffusion patterns. Most of these were printed with 300µ layers, because if you compare it to the the one printed at 100µ, they are so much clearer and crisper and less hazy white.

Of course, using thick layers and only a thin shell like this means that these parts print incredibly quickly and use very little material. This one was printed with just 25g of PETG in 45 minutes, that’s hard to beat. This little one I printed with a thicker extrusion width. But I can tell you one thing that doesn’t work, and that is leaving the printer at over 600% speed for the wrong file, I mean, it did get the print done super fast, but clearly, it’s not quite what I wanted.

So typically, with a .4mm nozzle, you’re going to set it up to lay down like 0.42mm wide tracks. Here, I went with a full mm, so two and a half times as wide as the nozzle, and while it’s not the greatest thing to do if you want overhangs and details to come out great, I mean, you should be using a thicker nozzle, this part still came out looking great. It’s not quite as clear, instead it’s very glossy if that’s the right word, but that might be due to the structure, too. It’s definitely better than using two individual perimeters, which would introduce that second interface where you’d potentially end up with more diffraction. Also, it’s very strong and stiff, as expected, but I did crack the bottom when I tried to remove it from the PEI bed, PETG, as usual, likes to stick to PEI a bit too much, so usually, I’d recommend using a liquid surface finish on top of the PEI that doesn’t stick quite as much, so something like Printafix, or Magigoo, or maybe Glue Stick if you can apply it evenly.

But this technique of just using a single wall isn’t great for every model. This Adalinda frog does come out really nice looking with those thick layers, but wherever there’s a slope on the surface, the extrusion lines just get super droopy, like on the back, near the tail or on the wings. I did configure this print with two solid layers on top and bottom and turned off any features that backfill surfaces, which would avoid these exact issues, but would also create these structures on the inside that just don’t look great if you see the way they diffract light shining through it. Alternatively, you could print it with a solid infill and thin layers, but that would use up a lot of material and take a long, long time to print.

So that’s how I printed these transparent / translucent parts. I think they turned out great, especially these single-wall prints, they are going to look awesome installed somewhere with some light in them.

If you learned something, click that thumbs up, get subscribed if you aren’t already and whether you’ve freshly subscribed or have been for a year or two, check that you have that bell toggled so that YouTube sends you notifications as soon as a new video gets uploaded.

If you want to try printing some clear stuff, too, there is a link to ColorFabb’s HT material that will take you to the right shop for your region, also, to the DAS FILAMENT PETG, that’s not an affiliate link, I just like the material, it’s cheap, it consistently prints well, what more could you want.

If you’re going to be shopping on Aliexpress, Matterhackers, Amazon, eBay or iGo3D, check out the shop links to them:

Amazon [🇺🇸 🇩🇪 🇬🇧 🇨🇦 🇮🇹 🇪🇸 🇫🇷]
eBay [🇺🇸 🇬🇧 🇨🇦 🇦🇺 🇮🇹 🇪🇸 🇫🇷 🇦🇹 🇧🇪 🇨🇭 🇮🇪 🇳🇱]
Matterhackers [📦🌐]
iGo3D [🇩🇪]
Aliexpress [📦🌐]

Or if you want to directly support what I’m doing here, head over to Patreon to chip in a dollar or two per month, or more, I’m not going to judge you, that is always appreciated and you’ll be invited to join an exclusive live Q&A hangout as well. So thanks for watching, and I’ll see you in the next one.

🛒 Get Colorfabb HT [🌐]
🛒 Get DAS FILAMENT PETG [🇩🇪]

Colorfabb.com article


You can support me without spending a single penny!