Hello, i have a serious issue at the corners of my print. (H2S with AMS2Pro, 0.6 HF Nozzle, Material: ASA).
There is a kind of wobble-pattern. The first prints with this new machine worked incredibly clean at the corners. But after a few printing hours (less than 50). I see more and more of these kind of corners. I did a complete calibration process, and also the first time a vision encoder calibration. I have the impression, that it got even worse since that point. Maybe its a mechanical problem, which was increased by the rough movements during calibration?. I can nearly exclude a damaged z axis, because in this case it should have been from the beginning, and i didnt move the printer since then.
- Increase Pressure Advance (K) in small steps (≈ +0.005 per test).
Why: Corners blob because pressure stays high as the head decelerates; PA pre-emptively reduces flow so corners don’t bulge. - Reduce outer-wall acceleration by 20%.
Why: Lower accel cuts the head’s overshoot at corners, which reduces the ringing/ripples that follow each turn. - Reduce max corner/jerk (a small step, e.g., −10%).
Why: Softer direction changes lessen instantaneous pressure spikes and mechanical bounce, smoothing the seam area. - Verify flow; if needed, lower Flow Ratio by 1%.
Why: Mild over-extrusion amplifies corner blobs even with correct PA. - Lower nozzle temperature by 5 °C.
Why: Slightly cooler filament firms up faster at corners, shrinking the blob window. - Check belt tension and gantry play.
Why: Loose belts or play increase ringing amplitude after each corner.
If the corner blobs shrink as K goes up but faint ripples remain, you’ve nailed PA—then finish by tuning accel/jerk (and belts) to kill the ringing.
This is just my guess, doesn’t hurt to try.
2 Likes
Slicing issue, let me guess, you only used one wall and used gyroid or cubic as infill?
Upload the 3mf file.
For many filaments (especially the likes of plain ASA), shrinkage is real pain and you would also need to take account of pillowing or inverted-pillowing effects as well. Pillowing is caused by hot air bubble, and inverted-pillowing is caused by hot air cooling down but your filament hasn’t been solidified enough to resist that.
1 Like
If you figure it out let us know, that looks cool if it was planned!
1 Like
Thanks for your reply. I tried some of your suggested parameters. I had screwed up speeds before, then i went down significantly, but the wobble stays. I will do some tests with PA too, as you suggested, but let me ask one question: Is it a typical sign for issues based on filament amount control, when the “waves” appear exactly every 10 layers - even, when the geometry is not the same in every layer? I thought, if its a problem of filament “squeezing”, it should appear somehow different in every layer, depending on the amount of filament missing or over extruding, caused by different geometries? It is very noticeable, that it occurs that rhytmically along the whole z-distance. I just want to exclude that it is a mechanical z-axis-issue…
It seems that the PA-setting is no longer part of Bambu Studio (V. 2.2.1). Below “flow ratio” there comes “density” instead of “pressure advance” - or did i miss something?
it has literally nothing to do with PA
Just post your 3mf here and people would show you the true answer. Without the 3mf nothing comes with solid proof. Also while Lord Satan tries to help, what he usually post often look quite … AI generated, including lots of hallucinated answers… Like your issue literally has nothing to do with PA, and also flow rate.
If it’s related to PA, then a curved corner means that you got PA too high, and that’s not right because this is a very large curve. When PA is wrong and you get curves that’s usually very small curves (like in ~1mm range). Then if PA is wrong it also means that the actual flow rate is too low, why the AI suggests to lower the flow rate then?
Because the AI thinks that you’re having “extra” bulges, like bulges that is sticking out of the model. While yours is shrinking inside.
I literally doubt if he’s just vibe posting. Which is really not helpful and I hope that people stop doing that.
1 Like
that’s pretty weird, but if you figure out how to do it on purpose you might be onto something there 
Swap nozzles, then reprint with dynamic flow calibration enabled. I had a problem that looked just like this a few days ago and it turned out to be a partially clogged nozzle.
For the record, I reproduced a similar corner issue before I replied. If that’s somehow irrelevant, I’m all ears.
Either way, a poorly tuned PA won’t manifest identically across different geometries. And PA lives on the printer/AMS—not in the filament profile.
As for the AI angle: if you want to complain, Bambu Lab is your best bet—they used ChatGPT in contests where it couldn’t even keep track of who actually entered. If anything’s embarrassing, it’s that. And if you think I’m using AI, that probably says more about your assumptions than my writing.
But hey, what do I know? I’ll step aside for the non-AI expert. 
It seems, he is not really looking for a solution.
The single wall seems to be the issue here. Your educated guess about the infill pattern seems a good one 
I did some test prints since my last post. It has nothing to do with the mechanics, because i could print a test file at higher speeds without wobbling. In the attachement you can find two test prints. One with and one without wobbling. I made several changes fitting to my larger print, so it is no solution swiotching back to the settings in the working test print. On the other hand i couldnt figure out which of the settings is responsible for the issue. If you compare the settings, maybe there is a genius out there who can?
The two imgages show the testprints. At the left one one can still see the artefacts, but way less than on the right side-
Testwürfel.gcode.3mf (87,3 KB)
TestwürfelWobble.gcode.3mf (106,0 KB)
In the “Wobble” file:
- Arachne with inner–outer–inner sequencing, creating more transitions near corners.
- Narrower wall extrusion width (0.55 instead of 0.62).
- Higher infill-to-wall overlap (22%), which can push against the outer wall.
- Different (often reduced) cooling during sensitive sections.
The “Testwürfel” file is more forgiving in this case:
- Classic wall generator with a simpler sequence.
- Wider wall extrusion width (0.62), resulting in stiffer walls and cleaner corners.
- Lower infill overlap (15%), reducing pressure from the infill.
There are also a few other differences between the two setups, but these are the ones I primarily believe could be impacting the print quality.
And of course, don’t take my word for it — I’m just an AI according to some poor souls — you might want to rely on the other experts in the thread if you need more guidance.
I dont have H2 printers, so take this with a grain of salt, but looking at your gcode atatchments:
At every “corner wobble” artifact is where your infill pattern alternates attaching to either the exterior wallgroup or the interior wallgroup.
X-ray showing infill, no walls
This being ASA, I’m pretty sure those are doing “Something Contraction-Related©” to your corners while hot, pulling them in as they cool. Try some different infill types, pay close attention to the preview patterns they want to make. Or no infill (4 walls at 0.6 is pretty tanky).
1 Like
Hey, thanks for the brilliant x-ray- idea;-). I’m pretty sure that Gyroid causes the issue, showing repetitions in a similar pattern like the real print. It think you found the reason behind, but other settings seem to increase it dramatically, because as i mentioned, i could print a relatively ok cube with other settings but still with gyroid. i tried a next cube with the same settings like in the “wobble” file and only changed Arachne to classic and got a nearly perfect finish. next step i will try different infills and arachne swithed on again for cross-checking…
Done any more tests?
It would be nice to see the results.
Thank you for the detailed reply. I’m a newbe to 3d printing. The reasons why were very instructive.
This is identical to how engineering grade filaments print on bambu 0.6 nozzle.
The only way to reduce it is running extremely low accelerations and speeds. 500 accel 50 speed.
Or maxing out layer height
Alternatively swapping to a 0.4 nozzle was an absolute fix.
1 Like
As you can see: no wobbling at all - but i must confess, this part is sanded afterwards. Nethertheless sanding would not have been the solution before, as the wobble were much bigger than sanding could equalize. So what finally helped:
- switching infill pattern to grid
- slightly slower speeds
compared to 0.4 nozzle and 0.2 layer hight i must say that i have still some slight wobbling, but maybe that is normal (i don’t have too much experience with 0.6 nozzle and 0.3 layer hight)