RE: VFA Artifacts on X+Y Straight Edges
It's great that you've got to this point mate.
I find it strange that there is literally no input or feedback from any Prusa employee on the companies own forum regarding this....the only place you see any public dialogue with Prusa is on Reddit then if you're lucky Josef or Mikolas might comment, do we really have to make a Reddit post and tag them to get any dialogue going?
I mean, whats next ? We can't all be expected to shell out for new pulleys and have to cherry pick good pulleys from bad pulleys...the communication from Prusa is basically horrible.
RE: VFA Artifacts on X+Y Straight Edges
Yes. The artefact follows the pulley no matter which motor.
No. They look identical to me.
I commend your efforts.
I just added some more belt tension and I do not get a change in my result. Which with belt tension at 85Hz is surprisingly similar to your samples before the new pulleys.
RE:
The pulley thing was also a problem on my MK4. Prusa's pulleys tend to be worse quality than other brands such as Gates or Mellow
For the MK4, it was mostly focused on the Y axis. I had a bad pulley that resulted in ridges in straight lines along Y that you could feel by rubbing the back of your fingernail on
On my Core One, some of my artifacts were a result of the actual motor. My Y motor pretty jerky to move (by hand) and felt a bit gritty
Bottom is my Y motor, top is my X motor:
Top is the print after changing the Y motor to the one from my MK4S to test:
RE:
And you know the best part?
This is my second time. My MK4 had terrible ripple in the Y axis only. After almost half a year of troubleshooting and back and forth with Prusa I've discovered it was... the pulley!
This is why the first thing I've looked into on my Core One was the pulley.
RE: VFA Artifacts on X+Y Straight Edges
And you know the best part?
This is my second time. My MK4 had terrible ripple in the Y axis only. After almost half a year of troubleshooting and back and forth with Prusa I've discovered it was... the pulley!
This is why the first thing I've looked into on my Core One was the pulley.
Did your IS numbers change much after finding good pulleys?
RE:
So who’s gonna be the next volunteer to swap the pulleys? 😂
I think if one more end user swaps them out and it solves the issue, then it’s pretty much conclusive that’s where the issue lies.
My only hesitation is the lucky dip of finding ‘good’ pulleys…that sounds like a potential pita.
RE: VFA Artifacts on X+Y Straight Edges
Did your IS numbers change much after finding good pulleys?
All my IS calibrations were almost identical. They were also a lot different from the defaults. Accelerometer decided to use ZVD instead of the default EI and the numbers were a bit different. But aside from them being different from defaults, changing pulleys didn't change the values. The only thing that slightly adjusted the values (and by like 1 each) was belt tension.
So no, the IS numbers didn't change at all after changing pulleys. Identical.
My only hesitation is the ‘lucky dip’ of finding good pulleys…that sounds like a potential pita.
An absolute PITA. But honestly you need to change the pulley once to learn how to do it fast. After that it's pretty quick to swap, test, swap, test. You do need to go through at least 4 in my opinion, tho, maybe even more if you're especially unlucky.
RE: VFA Artifacts on X+Y Straight Edges
Prusa's pulleys tend to be worse quality than other brands such as Gates or Mellow
So maybe they should not make their own, if they don't have good-enough control of the process?
I understand that historically, one of Prusa's selling points was that they used brand name components throughout: Linear bearings and rods, belts and pulleys, stepper motors... It seems that they have moved to no-name linear bearings, motors and pulleys, and all of these may be causing some issues: My Core One's Y axis has noticeable bearing noises, motor irregularities need to be smoothed out by phase stepping calibration, and now the pulleys are suspect.
Well, at least the X linear rail is still THK-branded, and the belts seem to be original Gates parts. But everywhere else, "value engineering" rules...
RE:
My only hesitation is the ‘lucky dip’ of finding good pulleys…that sounds like a potential pita.
An absolute PITA. But honestly you need to change the pulley once to learn how to do it fast. After that it's pretty quick to swap, test, swap, test. You do need to go through at least 4 in my opinion, tho, maybe even more if you're especially unlucky.
So Prusa are probably getting exactly the same hit and miss rate from the pulleys that you are getting, but instead of weeding out the bad ones they’re all getting fitted to printers or included in kits…makes sense that some users have very little in the way of VFAs while others get no problem.
Any idea of the manufacturer of the pulleys that Prusa use ?
I still question the claim that none of the machines at Prusa display such VFA problems though.
RE:
Im surprised that the pully can make such a difference, since the implication from that is that the belt is perfect, or at least the deviation in the teeth of the pulley is greater than the deviation of the belt to the point that the teeth of the pulley causes an issue.
I'll give it a punt at some point. Ive already blown my fun money on 1.5mm belts and pulleys for that (gowge from aliexpress, since it appears to be a brand there is a hope of quality control, or at least a consistent manufacturer).
Im pretty sure I've got some 16 tooth 6mm gt2 already somewhere anyway, but they are very much aliexpress specials (although for a telescope mount which requires so fairly precise tolerances).
I wonder if it is concentricity of the pulley?
RE: VFA Artifacts on X+Y Straight Edges
I wonder if it is concentricity of the pulley?
Seems more plausible than the mesh of the teeth.
RE: VFA Artifacts on X+Y Straight Edges
I've heard there's also a potential problem when you screw the grub screws down since it's 2 of them and there's a flat in the shaft. Certainly don't try to over torque it
RE: VFA Artifacts on X+Y Straight Edges
Wouldn't concentricity implicate a much longer pattern compared to the belt pitch 2mm patterns that are often visible?
The gt2 16T pulley has a diameter of the teeth section of 9,7mm so a circumference of 30,5mm. That means that the pattern due to concentricity would have a wavelength of 30,5mm.
Considering the actual wavelength being in the order of magnitude of the gt2 belt pitch, it would make much more sense that small meshing deviations cause tiny speed differences. Since the flow of material can probably be considered fairly constant with a direct drive system like the Nextruder, the resulting line width will vary slightly if the speed is not constant.
RE: VFA Artifacts on X+Y Straight Edges
What I just realised, all of the examples of VFA's in this thread are in sync. The deviations on the layers match so that a vertical pattern forms on vertical walls.
This means that the cause of these deviations is directly correlated to the X and Y position of the extruder. Any source of vibration that is more random in nature, won't cause these patterns. They would more likely look like fuzzy skin. For example, frame or plate resonances.
For a certain X Y position, the motor, pulleys and belt orientation is fixed. Thus, it is most likely caused by these three components. We know that phase stepping improves the VFA's somewhat but does not eliminate them. There are some hypotheses that the difference in stiffness in the belt areas with and without teeth might cause small differences. However, I have seen a video where a comparison is made with a broader belt without any effect. So that leaves the pulleys and then specifically the quality and tolerances of the teeth, since those match the wavelength of the VFA's of 2mm.
RE: VFA Artifacts on X+Y Straight Edges
Their is no doubt that it is the belt/pulley tooth mesh, as well the belt teeth running over the idlers on the x axis that cause the 2mm ripple at higher speeds (usually ~80mm/s to 160mm/s). The lower speed finer VFA is caused by the motors. I posted earlier in the thread about my experiements on another printer with this problem. All corexy printers have this issue to some degree, as do cartesian printers but at a lesser degree.
What I just realised, all of the examples of VFA's in this thread are in sync. The deviations on the layers match so that a vertical pattern forms on vertical walls.
This means that the cause of these deviations is directly correlated to the X and Y position of the extruder. Any source of vibration that is more random in nature, won't cause these patterns. They would more likely look like fuzzy skin. For example, frame or plate resonances.
For a certain X Y position, the motor, pulleys and belt orientation is fixed. Thus, it is most likely caused by these three components. We know that phase stepping improves the VFA's somewhat but does not eliminate them. There are some hypotheses that the difference in stiffness in the belt areas with and without teeth might cause small differences. However, I have seen a video where a comparison is made with a broader belt without any effect. So that leaves the pulleys and then specifically the quality and tolerances of the teeth, since those match the wavelength of the VFA's of 2mm.
RE: VFA Artifacts on X+Y Straight Edges
It's in sync for a given speed. With the VFA towers you can see changes in pitch appear with different speeds:
With mine it can be consistent for a bit and then suddenly there will be a change and once you get up to higher speeds they begin to smear. The downside to the VFA tower is that eventually you hit a point where the extruder is moving too fast for the filament to heat up so you need to start upping the temperature. Once you get to really fast speeds you begin to get defects on the edges. There is probably something motory going on there causing the change in the pitch and smearing out.
I agree that a harmonic relating to the circumference of the pulley would be expected, my thinking is that if the bore of the pulley isn't parallel with the teeth then this could introduce problems with the meshing of the teeth and pulley. The consistency of the VFA makes me think it a single surface defect rather than something that is consistently wrong for every single tooth.
RE: VFA Artifacts on X+Y Straight Edges
I have 2x Gates pullies that I installed in mine (immediately upon build of my conversion kit - I never used the Prusa ones in it). Reason I had two? I had terrible artifacts on my MK4S Y axis, and as part of an overhaul of it to try to fix them w/ a model someone posted, I bought the pulley and a spare. It immediately fixed my issues, so I went ahead and put both in the Core One. I've got some shiny silk PLA I'll print a VFA test with later tonight.
RE: VFA Artifacts on X+Y Straight Edges
Not really sure how to read this, but here we go.
RE: VFA Artifacts on X+Y Straight Edges
Not really sure how to read this, but here we go.
Have you run phase stepping calibration? The second picture shows a similar pattern to an image I saw on the GitHub phase stepping issue.
RE: VFA Artifacts on X+Y Straight Edges
Any context on what image represents what? Would it be possible to make a photo in the exact same orientation and light?