Public Service Announcement: Calibrate your temperature when using Aliexpress hotends!
Long story short, I've just spent weeks tracking down the strangest print results with my Kobra S1 and my Aliexpress ceramic nozzles. They print well, they're hardened steel, and you'll know where I'm going with this.
<rant>
Turns out: I am an idiot. Every child knows that you need to up the temperature due to the lower thermal conductivity when using hardened steel nozzles.
I unpacked my new hotend. I set the printer to "hardened steel" on the touchscreen. I set Orca Slicer to "hardened steel" in the profile. I thought this compensates for the material properties (why else would you have different options for brass, hardened steel etc. to select from?), but apparently, it doesn't. Of course, I then let the auto-calibrations run. Still, I got artifacts that looked like miscalibrated pressure advance?
I've tried every calibration test under the sun. I had print artifacts that looked like pressure advance problems. Every AI diagnostics chat, every reddit topic pointed towards pressure advance. I calibrated PA really carefully. It got better, but nothing really solved the issues. More cooling, less cooling, quicker print speeds, slower print speeds, flow up, flow down - print defects that were virtually indistinguishable from PA miscalibration. Just, that my PA was fine. Or was it?
Auto-flow calibration (calibrates both flow and PA on our printer) on/off, didn't matter. The same shit. Weird PA-like artifacts in every print. The same type of issue you'd expect when the PA value is wrong. So maybe the proprietary algorithm that the printer calibrates with auto flow calibration (it seems to generate its own adaptive PA model) only works with stock Anycubic hotends? Hell, I made my own adaptive pressure advance by printing test patterns, piles of them. It got better, but nothing substantially improved. Turns out, I can do adaptive pressure advance by hand-interpreting dozens of PA test patterns and writing the results in a list which I entered in Orca Slicer almost as well as the automatic algorithm, and the algorithm even got a bit finer/better looking prints. So nope, I don't think this was the issue as well.
The smoking gun was when I printed a VFA test and got weird "pressure advance artifacts" there, meaning I had clearly interrupted lines that came from the outside and got more pronounced at higher velocities. They looked like this:
I first thought: "Oh, this goddamn piece of shit of a printer that I got nothing but problems with from the day I unpacked it, has some mechanical problem that prevents it from printing cleanly at above ~150 mm/s". My second thought: "You know, PETG is known to be a complicated-to-print filament, so maybe this an expected result?" My third thought: "Nope, this is 65W of heat concentrated on a small melt zone, this should melt everything instantly. Something else must be going on."
And then it hit me like a brick: What if the temperature sensor says 240°C, but the filament core in the inside of the nozzle didn't reach this temperature? The sensor is on the _outside_! (obviously) It can't measure the actual molten plastic temperature.
A split-second later: So, my plastic doesn't get enough heat. What if it's the thermal conductivit- (blank stare at the wall) of the nozzle materi- (blinking twice). I am an idiot! THIS IS A HARDENED STEEL NOZZLE! Hardened steel has a low thermal conductivity.
So turns out, filament got fed into the nozzle, and the ceramic heater reached the assigned temperature, but the thermal conductivity of the nozzle's material was so low that _at higher print speeds_ (which, when printing at constant line width and constant layer height, translate to higher volumetric flows) there was not enough thermal energy being transfered into the solid plastic to melt it enough before exiting the nozzle.
Worst of all, it did melt, but its viscosity must have been so high that at sharp turns, the shear forces from decelerating and then accelerating were so high, they just ripped the stream of molten plastic apart.
This is not noticable with slow prints, and models often have regions with low and high speeds.
And viola, you get an underextrusion exactly and specifically at "turns" or "sharp corners". Which looks exactly like a pressure advance issue! I was trying to print high-viscosity half-molten plastic. And the faster the prints, the higher the flow rate, and the sharper the acceleration/deceleration, the more pronounced this problem got. It's like squeezing honey from a plastic bottle while making sharp maneuvers with the bottle mid-flight, at some point, the shear forces will rip the honey.
Which the VFA test shows - because the nozzle has to come to a full stop at every outer corner. It comes in at full speed, decelerates, reaches zero velocity at the outermost point, accelerates and comes to full speed again. That's why I saw these artifacts in the picture.
I set my PETG printing temperature from 240°C to 260°C, and all problems magically disappeared.
</rant>
<info>
I also needed to set retract_restart_extra to +0,06mm in Orca Slicer's printer profile (extruder section). Otherwise, I got bad seams, too.
Whether I still need this after raising the print temperature remains to be seen. I've seen a lot of people recommending to use +0,04mm with the Aliexpress hotends, so either they fell into the same trap as me, or this is an independent issue.
</info>
So, what do I want to do with this post?
- Share my knowledge of this several weeks long quest.
- Hopefully inspire you to never skip the most simplest calibration step: Calibration of print temperature.
- Also, to teach you to never assume an automatism compensates for everything.
Note: Maybe the Anycubic engineers _did_ raise the temperature automatically when selecting "hardened steel", but in my case with the Aliexpress hotend and a PETG filament, this clearly wasn't enough for higher volumetric flow rates. Maybe it works for the Anycubic hotends and PLA.