I had no issues using the supplied black PLA with the default settings. I am now trying 3D Jake’s magicPLA, and I can’t get a proper print. The spec has higher temperature (215–235°C) but other than that I have no idea what to adjust. I created a new material configuration in Luban, using first the wizard based on PLA, and with the given temperature range it simply set it to 225°C. I also tried printing in 215°C. Not much difference.
Then I added brims, and they come out very cleanly, so what I find is that the main issue is when the nozzle stops at the end of path, somehow the retraction doesn’t seem to be correct, there seems to be always some additional material at those points where paths end, and I think this is what causes the terrible print image and also caused the printer at some point to move the magnetic bed sheet by a few millimetres, causing a nice layer shift in the X axis and ruining the print.
What else can I do? There are dozens of parameters. I am printing with ‘smooth surface’ setting in Luban. Should I change speed? Still try other temperatures (higher, lower)? Is it a problem of over-extraction, should I reduce flow percentages?
Start with the basic essential calibrations: e-steps, temp tower, retraction towers, etc.
If you have already done your e-steps calibration, then you should not need to change it; post-calibration changes in filament type and densities should be handled by adjusting flow % rather than with e-step changes.
I can also tell from your first photo that your lines are too far apart, indicating a “line width” setting issue, under-extrusion, or z-height issue.
Luban gives descriptions for how each setting will affect the print; it is worth every second to start reading about all of them and experimenting with some of them in your prints.
What’s the authoritative pointer to “e-steps”? Sorry, I’m googling and there are random forum entries, but I can’t find really a guide how to measure this. It’s also not in the Snapmaker manual, apparently. Thanks!
Thank you. I switched from the stored E value 212.21 to 226.963. I think this means it was slightly under-extruding (it travelled less than 10cm before). Question: does this value depend on the filament (e.g. the “strength” the extruder can transport), or can I leave that stored forever now?
Another question, as it seems really a problem of retraction. If I print a single object now, I thought it means the head literally would never have to make retractions / movements from one point to a different one, because it can basically just print the object continuously:
This comes out of OpenSCAD. Is this because Luban has a simple translation from STL to G-code without optimisation? Should I “clean up” the STL file perhaps to avoid that the printer has to make unnecessary movements?
You should be able to leave the e-steps indefinitely. E-steps tells the motor how far to turn to make a certain distance, and contrary to common belief, has nothing to do with the filament or nozzle. However, because of differences in filament/temp/nozzle size, etc., and the pressure required to push it through the nozzle, that is where you adjust the flow rate % to compensate for that.
Make sure you write down your e-steps # and keep it handy, because every time you do a firmware update, it will erase it and reset it back to the default.
As for the file, there are some settings in the slicer (Luban, Cura, etc.) that you can adjust which often help reduce the travel moves, but be wary that they may come with their own set of issues in regards to the final outcome of the print. Usually adjusting the STL file will not have any affect. From the screenshot, it appears most of the travel moves are the head travelling around the print to prevent the nozzle from hitting it; this can be changed with the “Avoid Printed Parts While Travelling” setting, but doing so risks the part being knocked over by the head, just FYI. The odd-looking pointy spike of travel at the top of the screenshot appears to be the head moving off to the side to allow for a minimum layer cooling time; this can be easily eliminated by increasing the “Minimum Layer Time” in the slicer (found in Luban under "Material Settings, and defaults to 10sec); the tradeoff issue here is that if the layer doesn’t have enough time to cool, then it could deform under the weight of the layer(s) on top of it, though I have found this to be VERY rare except with certain materials (TPU, ABS, maybe others).
I see, yes I think it was indeed mostly pausing between layers to let it cool off.
Anyway. I did another attempt with a single print upright, and it was again pushed over at around 50%, apparently because it didn’t stick well enough to the bed, despite using a brim.
So I gave up on this orientation, and decided I will print it in two parts horizontally that will be glued together. Now, the surface prints come out very well…
…but the hemispheres rising from the base plate look even more horrific than before, if you didn’t know they are supposed to be hemispheres you’d think it’s some dog poo on a rainy street in Berlin:
Is this just too small to be printed properly even with smooth settings (thin layers)? The design has them at 4mm diameter (big ones) and 2mm diameter (small ones), respectively. That should be doable with this printer, no?
I guess its too hot for such fine details.
Did you print a temp tower?
Maybe more cooling (or less heating) would help, part cooler 100%, additionally it could be necessary to cool the part with a pc fan or something else.
I would additionally check the “combing” setting in your slicer is set to “Within Infill”, and you may need to adjust the linear advance (A.K.A. “k-factor”). Make sure your line width and layer heights are appropriate for your hardware setup; Luban’s defaults are not always within the generally acceptable common parameters. For example: Luban’s default layer height for “Smooth Surface” is 0.08mm, but with the standard 0.4mm nozzle, the layer height should really never go below 0.12mm. Luban also shows a default wall speed of 15mm/s, but you may need to go slower due to the size of the feature and the material/temp being used. These are all things that have to be experimented with, as any one change to any specific detail can affect many others. Such is the world of 3D printing unfortunately; it is still very much personal “prototyping” manufacturing on the consumer level.
Thanks everyone for their recommendations and comments. I ran now with an additional fan on (after the initial layers were done), this seems a slight improvement, not massive but at least noticeable. I also slowed down the initial layers even more. For the next run, I will increase the layer height to 0.12mm as suggested, and check the infill pattern.
