When you start a print, how do you decide
what settings to use? Of course, parameters like the solid shell
thickness or the infill percentage are something that needs to be decided on a print-by-print
basis, but what I mean are the settings that, ideally you figure out once and then leave
as a preset for future you. But what if those presets are wrong? What if you didn’t even tune those in yourself
and use whatever the filament manufacturer suggests or stick with the presets that come
with your slicer? How do you know that whoever made those profiles
actually did a good job? How do you know you’re achieving your filament’s
true potential? Well, you methodically test your settings,
and see if you can find anything that works better. But because that’s actually quite a bit
of work, I tried to do that for you and for the last three weeks, I’ve dedicated my
main workhorse printer exclusively to printing test parts that would help me determine the
best hotend temperature for the job. We’re going to test PLA and PETG inside
outside of their comfort temperatures, and we’ll look at aesthetic differences and
at what print temperature you’ll get the strongest parts, too. And no, it’s not quite as simple as just
going “print hotter, get stronger parts''. There is quite a bit more to it. But you know what’s actually quite simple? Today’s sponsor, JLCPCB! JLCPCB will fabricate your PCB and 3D printing
designs - at very affordable rates. For 3D prints, simply upload your stls and
you’ll instantly get a quote for a resin, nylon, ABS or metal print that you can get
shipped to you within a week. For PCBs they’ll process your Gerber files,
and I’ve used JLCPCB quite a bit for my own projects because they’ll even assemble
all your SMD components onto the boards for you. That’s super convenient and has opened up
so many new possibilities for me. Get your own PCBs and 3D prints from JLCPCB
at the link below. So what seems to be the generally agreed upon
understanding is that there is a normal print temperature for any filament, 215°C for PLA,
230°C for PETG, and so on, and that’s generally going to be the one you want to use, but if
you want parts that are stronger and are “welded together” better, you choose a higher temperature,
and if you want parts that just look a bit more crisp and less blobby, you choose a slightly
lower temperature. But there are two problems with this: First,
if you print hotter, you’re not just getting “better layer welds”, you’re also thermally
degrading your filament, with ABS, you’re actually specifically boiling off some of
the components that make the filament tough and flexible, that’s that horrible smell
ABS has; and with PETG in particular you have a process called hydrolysis where the hotter
it gets, any moisture in the filament will violently turn to steam and actually rip the
polymer chains apart that make up the plastic. At least that’s how E3D explained it to
me a while ago. And ripped apart, shorter polymer chains mean
a more brittle plastic, so by increasing the temperature, you may have gotten better layer
welds, but made the material itself weaker. There’s also the challenge of the material
bubbling up, your extrusion getting less consistent, which makes for more gaps in the parts, it’s
all factors that you’re going to have to weigh against the gain you might be getting
from better layer adhesion. The other way around, dropping temperatures,
only goes so far, too - when your filament is “less molten”, the extruder and hotend
are going to have a harder time pushing it and laying down filament in the exact spots
you want, which often results in underextrusion or even extruder skips. So let’s get this tested. I prepared two different plates for each temperature
of PLA and PETG - one to look for aesthetic differences between temperatures and one to
test for strength. I’m using one of my MK3s, with a 0.6mm Revo,
which, of course, is the best nozzle size there is, and I’m slicing with PrusaSlicer
2.5 beta 3, using Arachne, and the included 0.2mm layer height profiles. Then all I’m doing is that I’m changing
the temperature for everything but the first layer, because we know that the default contact
temperature is going to work for keeping the parts stuck while printing, and then, more
importantly, for being able to release them again once the print is done. For the aesthetic test parts, even though
they’re all on a single plate, I’m printing them sequentially one at a time, which means
I only have to start a print once and I’ll get all four parts printed in a row, exactly
as if I had started them one by one. On the plate, we’ve got, of course, a 3D
Benchy, you really can’t do tests without that, we’ve got the TomTest, with dedicated
areas for bridging, fine details, text, curling, and overhangs, we’ve got a scaled-down Salty
McCreedy, and finally the Prusa SL1 test part with some insanely fine details. For the strength tests, we’ve got the bend
strength and impact strength test parts, both printed in two different orientations and
with two copies of each. As always, these are all done with “normal”
print settings, 2 shells, 15% infill, because I think that’s much more representative
and interesting than just looking at the failure mode of a boring standardized test sample
that’s pretending to be an injection molded part at 100% infill. That’s just not the reality of what a 3D
print is. But let’s start with how the parts look. Honestly, pretty bad. We get a ton of stringing, especially with
the PETG, which is Extrudr XPETG matte white. I did dry all the filaments in the Eibos drybox
for a bit before I started the prints and then occasionally turned the dryer back on
while printing, but still, what I got with the XPETG would never fit in my test fixtures. Bridges just didn’t print well at any temperatures
and the parts just didn’t look particularly great and they didn’t print reliably. So, running out of time, I actually reprinted
all the mechanical test parts PETG with Prusament and those turned out so much better and honestly
they now all look fine, with just a bit of this branch building that PETG likes to do. These are prints at 210, 230, 250 and 270°C,
and one thing to notice is that the lower temperatures
have a surface finish that’s a lot more matte than the higher-temp ones. And also I don’t know if it’s just that
extra glossyness that amplifies it, but with the higher temperatures, the surface finish
also looks like it’s getting bumpier the more we turn up the heat. And you can see the exact same thing with
PLA, this also is Prusament - it’s basically completely matte when printed at 195°C and
then turns more and more glossy with each step to 215, 235, 255 and finally 275°C.
Of course, stringing also increased, to a point where the aesthetics test plate became
pretty hard to complete at the full 275°C. Specifically with PLA, though, the top surfaces
also started becoming less and less smooth at higher temperatures, while marginally fine
details like the little features on top of the SL1S test model actually printed better
at higher temps. Notably, overhangs all look identical at any
temperature, and this time around, now using PrusaSlicer 2.5.0 beta 3, the finest little
column on the TomTest printed just fine every single time, while beta 2 from the last video
was still having some issues with it. Thumbs up for the improvements! So from an aesthetic standpoint, I’d say
the default print temperatures, both for PLA and PETG, are fine. Other than the matte surface finish, which,
honestly, is a legitimate reason to print colder, you’re not actually getting your
prints to be “crisper” or anything. But what about part strength? Well, I’m glad you asked! My test parts are the usual Filaween-style
ones, for the bend test, they go into a jig, I hook a luggage scale onto them and then
I gently pull down on the scale until they break. Simple enough. Now, the impact test is a bit different, and
here we’re measuring how much energy the parts absorb while breaking. The hammer swinging down always starts with
the same amount of kinetic energy the moment it hits the part, and the tougher the material
is, the more it slows down the hammer, and the slower the hammer is, the shorter up its
swing is going to end. Basically, if it reaches the very top, the
part was just too brittle to slow it down, and the less far it swings up, the tougher
the part was. What I immediately noticed is that the lower-temperature
parts pretty often got stuck in the jigs. These have a bit of tolerance for the sample
to slide in and out, but the parts printed below nominal temperature were even too big
for that. I don’t know why that’s happening, but
if you’ve got an idea, maybe leave a comment below. So from this point forward, it was just about
breaking part after part. With the impact tests, you can actually hear
quite a difference between a brittle part and one that’s tougher and is going to absorb
more energy. Interestingly, all the PETG parts that were
printed laying flat broke in a way that’s actually really good for functional parts:
Instead of snapping in half, they stretched, pulled, and kinked, meaning they failed gradually. Typically if you have even slightly moist
PETG filament, it’s going to snap off much more sharply because of the hydrolysis I mentioned
earlier. Let’s start with the results of the bend
test: PLA’s material strength actually stays pretty consistent across temperatures, with
an ever so slight dip past 275°C, and a significant reduction in strength when printed at just
195°C. If we look at the layer adhesion results, we can see why - you’re missing out on a
lot of layer adhesion potential when you’re printing that cold. But layer adhesion or, at least strength in
the Z direction, also drops off pretty sharply when we’re increasing temperature. Now, this could be down to the fact that we’re
losing material for these test parts in stringing and the parts themselves actually end up slightly
underextruded, or the fact that extrusion control just isn’t great anymore at these
temperatures. The parts do end up looking slightly scuffed. In either case, it looks like the default
215°C is exactly where you should be printing PLA. At least at default speeds, there are no gains
to be had in strength or in how cleanly prints turn out by increasing or decreasing your
hotend temperature. That’s pretty good, eh? Now, PETG is similar - just at a higher temperature. I should note that the default temperature
for Prusament PETG in PrusaSlicer on the Prusa MK3 is 250 Prusa-degrees instead of the normal
230 or 235 that you’ll see elsewhere. But here as well, 250 is where you’re going
to get the best performance and where you should be printing this stuff. There is an advantage in how much this PETG
is going to string by going down to 230°C, but that already comes at a significant penalty
to layer adhesion. Now, impact strength. This one’s hard to interpret, especially
since the results here have a lot more noise in them. I could have tested several hundred samples
instead of a couple dozends to get a wider average of results, but that would have taken
an extra couple weeks to print and days to test. It’s pretty scary how quickly the effort
required to do these tests scales up. What I’m seeing here is that the impact
strength seems to go up with extremely low and extremely high temperatures. Why? No idea. I could see how the high temperatures get
the material slightly foamed up, which could get it to break more gradually, but yeah,
this is a bit of a mystery and possibly more noise than actual data. So what did we learn? Well, there’s a reason why we landed on
the temperatures that we now usually print at. They give you the strongest parts without
degrading the overall print quality. And yeah, print hotter, stronger parts? Not really true, unless maybe you’ve got
a hotend that isn’t particularly good at getting heat into the filament. I’ve done testing for that on the Revo,
and it’s actually a bit better at heat transfer than v6, but it’s by no means a super-high-flow
hotend. Also, maybe one last note, the temperatures
I’m finding here aren’t necessarily the same for all printers across all brands. High-quality thermistors and mainboards will
usually have relatively tight tolerances on the parts involved with measuring temperatures,
but it’s not at all uncommon to see an offset of + or - 10°C in the actual measured hotend
temperature between two cheaper machines. In either case, I hope you learned something,
hope you enjoyed the video, if you did, like, subscribe, support, links here - keep on making,
and I’ll see you in the next one!
