TLDR at end of post!
You're probably a bit skeptical, the title sounds to good too be true. But benchy don't lie:
All of the parts in the attached photos were printed using the crappiest/cheapest polycarbonate filament I could find on a stock Prusa i3 Mk2s, the glue stick it came with, and some fairly unusual Slic3r settings.
If you aren't familiar with polycarbonate, it's what bullet proof glass is made out of. And it is notoriously difficult to print. To the point that the 'action shots' of eSun etc. shows, well, pretty crappy results in the 3D printed objects.
But I think that's because we're taking a fundamentally incorrect approach to printing PC. I'm getting excellent results. I see no reason to print ABS, PETG, or even nylon at this point - all polycarbonate, all the time. Er, I'm obviously more about practical prints that need strength, PC is not going to be a good replacement for PLA for aesthetic prints
I'm currently in the process of replacing all ABS parts on my printer with polycarbonate ones, now that I have a good printing process figured out. Let me tell you - aside from being comically strong (PC's nickname 'King of the Filaments' is well-earned), the dimensional accuracy is just killer. I remember having trouble getting a few nuts in their holes when assembling my Mk2s from a kit, the ABS tolerances just weren't that great. Nothing some percussive maintenance couldn't fix, but still.
Polycarbonate parts, at least when printed with these settings, seem to have a nearly perfect fit. I've printed round ball bearing holes, hex nut holes, and the entire extruder, x-axis carriage, x-axis ends, and every all fit perfectly together, and have a much better fit than even the original ABS parts. And, of course, they are flame retardant, rated to temperatures much higher than ABS, and have amazing strength combined with rigidity without brittleness.
And when I say comically strong, I mean I can't break that benchy in the photos. I'm sure I could snap the smoke stack on top off, but besides that, I am actually unable to break it. Using all my strength to try and torque or snap the cockpit off...nope. I can't even make it visibly flex. I try to deform it in any way, and the benchy simply says, "no." Even a weak point, the top corner of the rim about 1/3rd from the front of the boat, I can make that flex slightly when pressing on it with all my strength from the side, but that's it.
Now, this is a dramatic difference from my initial experience with PC. My initial experience was: severe warping
, as in you just couldn't really print it at all without an enclosure. And absolutely terrible interlayer adhesion. What I did try to print delaminated and split apart several places, and I could pull each layer apart with my bare hands. It crumbled in my fist, essentially. Yet I was printing at the maximum temperature.
Well, temperature was ultimately the source of my other woes as well... warping, delamination, etc. So I decided to stop relying on temperature. Bonding of most polymers, but especially
polycarbonate, depends on both temperature and pressure. The standard process in 3D printing is to rely mostly on temperature. Pressure is really only used (sometimes) for first layer bed adhesion, like how ABS likes to be mashed into the print bed surface a bit.
If you have pressure, you can achieve the same level of bonding at a lower temperature. So maybe polycarbonate would work better if there was some real pressure behind the layer deposition, hopefully reducing the sensitivity to temperature.
Well, it worked
And I had to go outside what is even considered 'sane' settings. So, the downside to this is I've only managed to do it with one layer height - 0.200m. The good news is that it prints about as fast as 0.350mm (this was what I was trying to emulate). Not through crazy extrusion speeds, but through volumetric rate. Which brings me to the secret sauce.
0.200m layer height.... 0.75mm extrusion width. I wanted 0.8mm, at least a 4:1 aspect ratio, but after some trial and error, 3.75:1 is more reliable and just as strong. The cross sectional area of the extrusion is almost the same as the default 0.350mm default, so ultimately the same sized part takes similar amounts of time to print.
Anyway, such a large aspect ratio really makes the extruder have to IRON that filament down into the print to force a line that wide. It also gives a lot of thermal interface and lets the nozzle inject a fair bit more heat into the print, further improving things.
I did, however, retain the smaller width for external perimeters, so the ability to resolve detail isn't noticeably effected.
The only downside is this style of printing is less forgiving when it comes to your extrusion multiplier. Worse, I uh accidentally destroyed my hobbed gear at one point and the new one I'm using is not the same radius, so I am kind of guessing at what the extrusion multiplier should be for printers with the stock gear. So you may need to dial that in a bit for polycarbonate, but you can tell if it is too high or too low after just 2 layers (the extrusion lines should be well-melted together without gaps, but should also not be making 'bumps' of excess plastic that had no where to go and was forced upwards).
Also, this is a bit weird....but printing polycarbonate like this seems to have some capacity for 'warp recovery'. You may notice certain parts to warp, even severely, and definitely something you'd normally abort the print for. Don't. polycarbonate layers don't seem to 'compress' like ABS ones do, and if you just let the print follow through, it will eventually unwarp the warped bit and it will be in exactly the right spot. Even though you will see the extruder nozzle visibly forcing a warped dangly bit 5mm downwards over and over (I mean, really severe warping) amidprints. Resist the urge to stop the print - it'll probably be ok
Case in point, this part needed supports due to the square cutout, and at one point, there were just two little long rectangles before they were bridged to complete the idler bearing cut out.
Well, those little rectangles warped so badly they ripped away from their own supports, and were very nearly curling up and were maybe a full centimeter above the layer being printed. But it was ok, because I sure can't even tell there was any warping in the finished print:
Anyway, I've attached the Slic3r settings to this post. Go nuts - I'd love to see if anyone is able to replicate these results, or even better makes improvements! Oh, linear advance is 100% mandatory, so make sure you are using the latest version of the firmware. Also, the K factor is higher than the default polycarbonate (PC-MAX) Slic3r setting. It's on purpose. You will see a lot of unusual things in these settings, most likely. Some of them I actually calculated (if you change the extrusion width, you'll need to adjust the bridge flow rate for example).
Oh, I almost forgot, bed adhesion. Just use a glue stick. Works a treat. In fact, it seems to work best if you let it dry a bit first, or apply it hot and wait a couple minutes. You generally do not need to reapply each print. The PVA layer, dry or not, seems to make polycarbonate stick on par with PETG to the print bed. No special glue nonsense needed
FInal note: I already tried this out of curiosity with ABS and PETG. They super don't like being printed this way. It seems to be fairly specific to polycarbonate.
A few more pics, of the x-axis motor mount:
https://metacollin.com/screens/upshot_k ... 3VeUET.jpg