I ran into a weird issue with one of my P1Ss. One of the parts I have printed dozens of times before, started showing a weird pattern but ONLY if it’s printed at the top left corner of the build plate. No such issue at all on my other P1S.
I’m using Archimedean Cord as the top surface pattern.
I have already done:
Cold pull 5x
Cleaned nozzle using the needle
Adjusted belt tension following Bambu’s Youtube video
Ran printer calibration and self test
I’m at a loss here as those rumble strip patterns are very distracting and I cannot send them to customers.
At a guess: When printing in the corner, warping of the print pulls the build plate off the heat bed. That reduces the effective layer height with increasing part height/warping leading to asymetric overextrusion.
I’ve run multiple tests and there’s no first layer adhesion issue. I’m also printing it on the SuperTack. Also anything from Elegoo to Bambu filament has the same issue.
My first thought is that corner is where the filament tube is going to have the tightest bend.
Perhaps it is causing enough friction to change the extrusion rate ever so slightly.
That is not what I meant. With good build plate adhesion, warping forces are able to overpower the magnetic forces keeping the build plate on the heat bed. In particular if the magnetic area is reduced when compared to the central position.
I would expect the effect to be worst directly at the corner though so it may be unrelated.
Similarly, I’d expect cooling issues to show mainly on the side facing the aux-fan although, with PETG, I did in the past have draft issues from the poop chute which fits the location in the pic.
What filament are you using? You mentioned brands but not types.
I did another test, printing 3 at a time. As you can see, the center one has no issues but the top left has issues. I also did one with Concentric just for fun, don’t seem to have any issues there as well.
OK, so all PLA. So it needs to be a rather significant influencing factor. Do you print fully closed or do you have the doors and/or lids open?
And it works on one printer but not the other.
And there may be a number of different root causes with guesses as to the root cause mainly due to the plate location, defect appearance and defect distribution on the top surface.
Did you check the PTFE tubes as @Lexi recommended? I have been surprised at the effect of PTFE friction in the past. In particular when using the AMS.
What differences are there between the printers? Is one older than the other? Did one recently receive maintenance? Is one closer to a door, window, cold spot, has more air circulation around it than the other? Maybe there are some further pointers from the answers to these questions. At least regarding potential Aux-fan/poop-chute draft issues. You could try to lower the Aux-fan setting to see if that is related to the defect. Similarly, you could consider protecting from a poop chute draft by lowering the chamber fan and/or protecting the poop chute from drawing in air.
As for the possibility of a build plate lifting from the heat bed, that could of course be a result of a difference of magnetic attraction in the build plate-printer combinations. Although I do not use them myself, I am aware that corner brackets holding the edges down are said to be available on MakerWorld. That may be a worthwhile test. It is a rather crude method, but the usual anti-warping measures (print with a warm chamber, lower layer height, slow down) are probably not actractive measures in your situation.
They both have their doors removed but they’re running inside a tent (so I can vent the fume outside). There’s no draft, it’s a very small ventilation fan.
The top printer is the one I have issues with. I has 1600 hours on it and 6 months old. The bottom printer has about 1000 hours and it’s 3 months old.
Both of them have had PTFE tube replaced about a week ago (Bambu supplied ones 2.5mm ID 4mm OD)
Thanks for the input. That is quite a nicely thought out set-up
I would still check the PTFE of the troublesome printer. Just in case the issue emerged after renewing the tubes. If the cut is slightly off, it can cause additional friction. And it is a fairly quick check.
I regard the Aux fan as an unlikely root cause since I’d expect even more severe failure cases showing.
I am not yet quire ruling out the poop chute. At least not before a further question
The exhaust from the tent. Is it closer to the poop chute with the troublesome printer? Is the exhaust passive or active with one or two exhaust fans? In the latter case, do you see a difference if you throttle it back?
With the troublesome printer being the older one, I would regardvplate lifting as still very likely. I am not very knowledgable on magnetism other than being aware of it being able to be reduced over time, especially in heated encironments. So you may want to try to clamp that edge down mechanically.
I also have a “Smoothy” Y-splitter in place. I removed that and just run the filament directly into the tube just to see if that splitter was causing additional friction. Running a test will report back after an hour.
PS: I overlooked this post on the surface patterns. The concentric print is from the same location? Upper left?
=> In that case I would fully bow to @Lexi experience with these patterns. I know she uses archimedal chords more often than I do. I am not sure about her recommendation about the number of top layers other than her certainly having a very good reason for it.
No, concentric is upper right. That picture was from the build plate. I didn’t move any of them.
As far as top layers, there are 12 layers of internal solid infills until it gets to the top layer. It’s variable layer height so it’s doing 0.08mm in that area.
I am using Grid infill, but I also tried other infills and made no difference.
Ah. Then it does not rule out other possible root causes. The Upper right corner sees a bit less cooling. That is good against warping as well.
Incidentally, if you are not ironing, you could run a test with a warmer build chamber (I print PLA at 35-39°C build chamber) and the tent exhaust not done actively. You’d need to add the door again though. A warm chamber is my first step against warping.
12 is a lot. So infill probably not an issue but flow effects from, for example, warping would show prominently. Again pointing to the build plate lifting.
Unfortunately, even after removing the Smoothy and feed it directly from external spool, it still produced the pattern but didn’t seem as severe. Also tried a new plate (textured) and still the same. No lifting of plate either.
Could there be some microplastic stuck on the rods near that corner and perhaps was causing additional resonance?
I also have a Bambu ticket submitted, maybe they will send me some replacement parts to replace. Fingers crossed.
Very nice topic and constructive engagement. There’s a lot of good discourse in this thread.
I am way too familiar with this phenomena and I’ve found a sort of workaround albeit far from perfect.
The issue that is at hand that’ve I’ve been able to diagnose is the one of inconsistent filament shrinkage combined with inadequate nozzle clearance over certain infill patterns like Archemedean Chords.
TL;DR Disclaimer: I have not found a magic bullet to fix this and I simply have concluded that it will require a refinement in the slicer’s motion algorithm to ultimately fix it.
The problem as I understand it so far.
In certain combinations of temperature and filament density, the slicer does not correctly compensate for filament shrinkage. This manifests itself as a “rumble” vibration upon successive collision with the high points as the nozzle passes over the layer where the infill meets the top layer. The change in nozzle direction is what causes this I believe
** its like as if the infill filament is shrinking at a lower rate than the slicer expected and the nozzle gets caught up on the infill ridges.
This problem is exacerbated by certain infill patterns as your experiment thoughtfully illustrated.
It is worsened by models that have a slope.
It is made worse by speed and temperature.
A max flow rate turned up too high allows for inadequate filament flow which contributes to under extrusion:
Experimenting by lowering the filament temperature on a top layer basis through GCode was tedious but it did help with noticeable differences. In the end it wasn’t worth it though.
As @Lexisuggested, increasing the number of top layers helped but I also had to decrease the speed of the top surface to really make it have a noticeable affect.
Decreasing the max flow rate increased the amount of filament flow because the slicer slows down the nozzle feed. This helps with under extrusion too.
Reducing the layer height on those affected sections also made a big impact but it was highly dependent on the model geometry. As an example, it made a big impact on a x-y curve or plane but not any affect in the Z-axis or vertical walls.
Increasing the size of the line width in combination with speed, max flow rate also helped because it forced more filament into the crevices:
Changing the top layer pattern to match the infill pattern so that the two motions of infill and top layers agreed, helped but at the sacrifice of appearance and “layer adhesion”. In other words, if I were using an Archimedean infill, I would use the same top layer. Again, it produced results but not a magic bullet.
Perhaps someone else knows of a setting I overlooked.
However, I just resigned myself to the fact that there was no single slicer setting that would allow me to force the nozzle to “lift” more as it traveled over the infill ridges. In the end, it is one reason I tend to favor Gyroid infill as it seems to be “friendlier” to this phenomena.
Conclusion: What a true remedy might look like.
I think if the devs could implement a dynamic flow compensation system that adjusts speed and flow on upper layers where infill meets top layers, based on the selected infill patterns, it might solve this issue. But honestly, even though I can visualize how it should work, I can’t define an algorithm that would make it possible to code into the slicer logic. I’ll admit—my older mind no longer expresses these technical nuances as easily as it did back in my university days. This is a problem that can only be solved by someone smarter(younger) than me.
Thank you for the in-depth analysis and screenshots. Do you have any idea why this artifact is only produced when printed at the upper left corner of the build plate, despite my best effort to reduce friction of the PTFE tube and ensuring the plate isn’t getting “lifted”. This phenomenon is also not reproducible on my other P1S with the same type of plate, filament, slicer file.
It just makes me think something is wrong mechanically with the printer but I’m unable to pinpoint what exactly. With 1600 hours on it, could the extruder be partially worn out and therefore not providing enough torque when it needs it most? Or maybe some foreign object got stuck on certain part of the rod or gantry, causing excess resonance only in the upper left area of the build plate?
That’s a fair question and apologies for not addressing it in my “overly-long” post.
I have only a theory. When I conducted such experiments, I was able to replicated this type of phenomena across multiple models, fill patterns and flow settings.
My theory is that very small variations in acceleration are the root cause of this. As the X-Y head moves to certain corners of the build plate, small variations in movement and flow are magnified.
Again, this is only a theory that I have yet to test out but here is a setting that I have yet to put to the test. You may want to see if changing these settings to extremes, impacts the affect you’re seeing.
By “extremes,” I mean drastically reducing the acceleration—say, from 10,000 mm/s² to 100 mm/s²—just to see if it has any effect on the model. This will significantly slow down the print, but as a scientific approach, it will clearly expose any impact on extrusion.
As I mentioned, none of these adjustments can be tested in complete isolation, but applying extreme changes one at a time should reveal performance aspects that might not otherwise be obvious. At least, that’s how I typically experiment.
If you care to upload your 3MF, I’d be happy to test it on my setup too and compare my findings.
Thank you. One thing I should also point out, and want to get your thoughts on is, that table I have the printer sitting on (if you scroll up you’ll set my set up). That table gets wobbly during the print.
Could that remotely has anything to do with this? A wobbly table? Just want to get some thoughts before I take that printer down and sit it on the floor for another test…
This is the 3mf file. It’s located at the center of the plate now but you can move it to the upper left corner during your test… test.3mf (260.9 KB)
This is a problem that can only be solved by someone smarter(younger) than me. 