RE: VFA Artifacts on X+Y Straight Edges
Did you use stock settings?
There is a current bug in the software that makes layer shift at fast print speeds with arc fitting turned off or avoid crossing perimeters on, mostly on round objects. https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4573
It’s the same gcode I used before, and shouldn’t have those settings changed. I’ll definitely check, though, thanks.
RE: VFA Artifacts on X+Y Straight Edges
Arc fitting is and has been on and avoid crossing perimeters is and has been off, so it's not that.
I rechecked the gantry and it was way off, so I realigned it and retensioned the belts. One of the belt adjuster screws was crunchy feeling, so I removed the screw and the threads were chewed up. I think the crash from my last VFA test not only bent the gantry brackets but also mashed the tensioner screw threads. I am not pleased.
Anyway, I recalibrated everything again and ran the VFA test. It crashed in the same spot, but this time not so badly that anything got bent. I guess I'll post a GitHub issue on the firmware repo and see if Prusa can reproduce this with the gcode. Otherwise I'll deal with customer support. Ugh. Seriously considering returning this thing now.
RE: VFA Artifacts on X+Y Straight Edges
Anyway, I recalibrated everything again and ran the VFA test. It crashed in the same spot, but this time not so badly that anything got bent. I guess I'll post a GitHub issue on the firmware repo and see if Prusa can reproduce this with the gcode. Otherwise I'll deal with customer support. Ugh. Seriously considering returning this thing now.
Grmpf... Sorry to read that; this is getting annoying.
I'm feeling brave. 😉 Could you post the G-code so I can try it on my printer?
RE: VFA Artifacts on X+Y Straight Edges
Oh, just one crazy thought, @altaic: Could you try to reprint after resetting the input shaper parameters to their stock values? Maybe the input shaper is somehow tripping over its own feet at the custom-calibrated settings?
RE: VFA Artifacts on X+Y Straight Edges
Oh, just one crazy thought, @altaic: Could you try to reprint after resetting the input shaper parameters to their stock values? Maybe the input shaper is somehow tripping over its own feet at the custom-calibrated settings?
That occurred to me as well, though I’m going to step away from this for now since I just spent several frustrating hours with it. I’ll give it a shot tomorrow.
RE: VFA Artifacts on X+Y Straight Edges
Grmpf... Sorry to read that; this is getting annoying.
I'm feeling brave. 😉 Could you post the G-code so I can try it on my printer?
Certainly, I zipped up the gcode and 3mf for you.
RE: VFA Artifacts on X+Y Straight Edges
Certainly, I zipped up the gcode and 3mf for you.
Thanks! For what it's worth, the gcode printed without mishaps on my Core One. Input shaper tuned to 43 Hz ZVD on both axes.
However I did notice a "knocking" sound four or five times, always in the curve between the two backward-facing arms of the print. And I see little dimples on the print which might well be in those exact locations, always followed by the same pattern of oscillations (lasting for about 15 mm). This occurred during the 1st, 2nd and 4th speed step. So maybe there was either a mechanical stumbling block, or a critical motor or belt resonance, or a glitch in the software which brought the printer out of step? No step loss however, i.e. no shift in the following layers.
RE: VFA Artifacts on X+Y Straight Edges
Thanks! For what it's worth, the gcode printed without mishaps on my Core One. Input shaper tuned to 43 Hz ZVD on both axes.
However I did notice a "knocking" sound four or five times, always in the curve between the two backward-facing arms of the print. And I see little dimples on the print which might well be in those exact locations, always followed by the same pattern of oscillations (lasting for about 15 mm). This occurred during the 1st, 2nd and 4th speed step. So maybe there was either a mechanical stumbling block, or a critical motor or belt resonance, or a glitch in the software which brought the printer out of step? No step loss however, i.e. no shift in the following layers.
Thanks so much for testing! I also noticed what you’re describing, both the knock/stutter noise and the little mark. I figured it had something to do with my machine, though maybe something’s up with the gcode? It’s a vase-mode print, so I’m not sure what could cause that. If my issues turn out to be due to a slicer bug, that’d be a huge relief.
It’s odd that it didn’t happen previously, though; my other test print used a previously working gcode file. Maybe it started with v6.3.1 and it didn’t lose a step so I just didn’t notice. Hmm.
RE: VFA Artifacts on X+Y Straight Edges
Something just occurred to me. I believe PrusaSlicer 2.9.2 changed how step files are processed. This is a thin-wall step model, so maybe the geometry is causing arachne to freak out and emit bogus gcode.
Still wouldn’t explain the previous working gcode file causing a crash, but perhaps I just didn’t notice since there wasn’t a layer shift.
RE: VFA Artifacts on X+Y Straight Edges
So my newest VFA test from 170 mm/s to 200 mm/s failed during the 190 mm/s section (12.8 mm up). It looked like the XY kinematics seized and the print layer-shifted way to the left (-x). Seems that I messed something up while realigning the gantry. Sigh, back to fiddling with that.
Does the printer have to read the gcode at a faster rate if it's printing faster? I wonder if there's an issue with your USB drive failing to keep up. Is it worth trying a different USB drive?
RE:
Until I printed a part of a filament cutter yesterday with cheap black PETG, I was very happy with the print quality of the Core One.
No other major problems either. Only 2x nozzle cleaning and one homing failure probably due to over-tensioned belts.
But the part looked terrible. Roughly like the pictures of the thread author.
It was the cutter box of this model. Interestingly, the top part looked much better. Due to the minimal layertime it was printed much slower.
A print where I reduced the speed of the outer layers to 40mm was much better.
In search of the causes, I took a closer look at the idler rollers. These have vertical play. By moving the print head in all directions, I observed the movement of the belt in the idler and tried to find the best vertical position.
I used a small mirror for the rollers on the motor.
I had the impression that it got a bit better. It can't hurt to try.
In any case, shiny black PETG is the best way to detect VFA.
Another VFA test print confirms, as with the Bambulab, that speeds above 150mm look best.
One problem is to find the right speed in any case. In my example, I sliced at 0.2 speed.
By lowering the min. layer time to 2s, this speed was also maintained in the upper part without falling to approx. 35mm as before.
The result is not perfect but a significant improvement.
In the Orca Slicer there is the ingenious function that the outer perimeter is not lowered in speed if the min. layer time is undercut.
This results in homogeneous speeds over the entire part.
I would definitely like to see this for the Prusa Slicer because it would simplify the whole issue considerably.
RE: VFA Artifacts on X+Y Straight Edges
In search of the causes, I took a closer look at the idler rollers. These have vertical play. By moving the print head in all directions, I observed the movement of the belt in the idler and tried to find the best vertical position.
But isn't it a given that the belt will move up and down somewhat? On the idlers, the room between the flanges has to be a bit wider than the belt -- otherwise the belt would always rub against the idlers. And there is nothing besides the idler flanges (and the clamping at the Nextruder) which would keep the belt at a defined height.
So I don't see how you can avoid that the belt wanders up and down a bit. (Apart from the possibility to have some tilted axles, to always force the belt against one side of the idlers, which would make things even worse.) Carefully adjusting the belt height to an ideal center line is bound to be temporary -- I don't see what would stop the belt from wandering away from that line again.
RE: VFA Artifacts on X+Y Straight Edges
You can move the ider slightly vertically.
If the belt now tends to touch the top, you can try pushing the idler upwards.
The belt always moves within the flanks. If you move the head, you can see where the belt moves and can try to optimize the position of the idler.
RE: VFA Artifacts on X+Y Straight Edges
You can move the ider slightly vertically.
If the belt now tends to touch the top, you can try pushing the idler upwards.
The belt always moves within the flanks. If you move the head, you can see where the belt moves and can try to optimize the position of the idler.
Right -- all idlers should be at the same height. You don't want a situation where one idler forces the belt to run against a flange of the neighboring idler. But even with all idlers perfectly in line, I think you can't stop the belt from "randomly" and gradually wandering up and down, touching various flanges in the process.
But how did you adjust the idler heights? Just pushing them up and down on the screw that forms the central axle through the ball bearing? If I recall correctly from the assembly process, the printed mounts do not leave much play at all there, and even clamp the center part of the bearing a bit.
RE: VFA Artifacts on X+Y Straight Edges
Yes, it's all about the little things. But you can move the idler up or down a little to compensate for how much it presses against the flanks.
This does not affect the height of the belt unless you move the idler in the wrong direction and it rubs even more against the flank.
Just try it out in practice or leave it as it is.
RE: VFA Artifacts on X+Y Straight Edges
Band saws use crowned pulleys to keep the blade centered on the two wheels without flanges. While googling if this has been applied to 3D printers, I found this 13 year old video.
If belt rubbing is an issue, crowned wheels may be an option, printed or turned on a lathe.
RE: VFA Artifacts on X+Y Straight Edges
That’s interesting. The ever watchable Steve Mould has a nice clear video on the crowned pulley principle:
So how about a printed toothed crowned pulley?! I know the common wisdom seems to be that toothed pulleys make no difference to VFAs, but it couldn’t hurt, could it? I don’t think the teeth would interfere with the crowned pulley principle. Obviously the belt would have to be centred when adjusting the tension.
RE: VFA Artifacts on X+Y Straight Edges
Oh, just one crazy thought, @altaic: Could you try to reprint after resetting the input shaper parameters to their stock values? Maybe the input shaper is somehow tripping over its own feet at the custom-calibrated settings?
Unfortunately the problem persists with both the default input shaper settings as well as the null input shaper.
I've submitted a bug report to the firmware repo for the motor stuttering and stuff during the VFA test print. I'll be posting updates there instead of here since the issues I'm facing aren't directly VFA-related, and I may start a separate thread if it this turns out to also be mechanical or something.
RE: VFA Artifacts on X+Y Straight Edges
Here's some interesting information. The tests were completed for a bed slingers, but I wonder if a similar experiment needs to be done with the core XY mechanism. I've also been listening to the CNC kitchen interview with Bambu CEO on the H2D and one of the things they worked hard to perfect was the print quality. It's an interesting listen of you've got a few hours to waste.
Some of the things he talked about was the importance of the control of the extruder, hence the reason they developed a servo motor for the extruder vs. a stepper motors. Then went on to explain that the same benefits with of using a servo motor on the core XY motors didn't make sense money wise because there is so much more innacracy put into the system by all of the belt resonances and the pulleys that the accuracy gained by the servo motor is effectively nullified. So a stepper motor is plenty accurate enough for those.
Anyways here's the 2 videos. The first is the video of belt tension and it's affect on print quality and accuracy, and the other is the interview.
RE: VFA Artifacts on X+Y Straight Edges
@altaic - did you try the same gcode files on a different USB drive?
So my newest VFA test from 170 mm/s to 200 mm/s failed during the 190 mm/s section (12.8 mm up). It looked like the XY kinematics seized and the print layer-shifted way to the left (-x). Seems that I messed something up while realigning the gantry. Sigh, back to fiddling with that.
Does the printer have to read the gcode at a faster rate if it's printing faster? I wonder if there's an issue with your USB drive failing to keep up. Is it worth trying a different USB drive?