Notifications
Clear all

VFA Artifacts on X+Y Straight Edges  

Page 12 / 23
  RSS
Jürgen
(@jurgen-7)
Honorable Member
RE:
Posted by: @scott-18  
Posted by: @jurgen-7

 The fact that they tried to activate the menu entry for calibration in 6.3.1, but now had to take that back due to out-of-memory problems, is a huge disappointment.

WHAT? You mean that it's not going to be released anytime soon?

At the moment it is apparently not working -- see the "Revert" notice @logicwax has linked to. Will Prusa release 6.3.1 now, with the fix still disabled? Will there be a 6.3.2 soon with phase stepping calibration enabled? Or will they delay 6.3.1 for a few more days, trying to get it implemented? Who knows...

Posted : 01/04/2025 11:06 am
Scott
(@scott-18)
Estimable Member
RE: VFA Artifacts on X+Y Straight Edges

Yes at specific speeds, not at EVERY SPEEDS as I got on mine 😆 😥 

Posted by: @boris-3

Guys, phase stepping wouldn't have fixed the belt ripple issue anyway.

Read my comment on page 10. The XL has the same belt ripple even though it has active phase stepping calibration.

As far as I am aware all FDM printers on the market right now have belt ripple if printed at a specific speed. This is especially noticeable with glossy filament.

 

Posted : 01/04/2025 11:21 am
Boris
(@boris-3)
Active Member
RE: VFA Artifacts on X+Y Straight Edges

Even at 200mm/s?

Posted by: @scott-18

Yes at specific speeds, not at EVERY SPEEDS as I got on mine 😆 😥 

Posted by: @boris-3

Guys, phase stepping wouldn't have fixed the belt ripple issue anyway.

Read my comment on page 10. The XL has the same belt ripple even though it has active phase stepping calibration.

As far as I am aware all FDM printers on the market right now have belt ripple if printed at a specific speed. This is especially noticeable with glossy filament.

 

 

Posted : 01/04/2025 11:53 am
Gizzle
(@gizzle)
Active Member
RE: VFA Artifacts on X+Y Straight Edges

For me it happens at every speed, sometimes more and sometimes less.

Posted by: @boris-3

Even at 200mm/s?

Posted by: @scott-18

Yes at specific speeds, not at EVERY SPEEDS as I got on mine 😆 😥 

Posted by: @boris-3

Guys, phase stepping wouldn't have fixed the belt ripple issue anyway.

Read my comment on page 10. The XL has the same belt ripple even though it has active phase stepping calibration.

As far as I am aware all FDM printers on the market right now have belt ripple if printed at a specific speed. This is especially noticeable with glossy filament.

 

 

 

Posted : 01/04/2025 11:54 am
Scott
(@scott-18)
Estimable Member
RE:

200mm/s gives weaker layer adhesion and not even finish on the print (it has to slow down on some small perimeters or overhangs) so no, I don't want to be forced to print at 200m/s outer wall.

The mat finish at this speed helps to mask VFA but that's all.

My X1C shows VFA between 70mm/s and 120mm/s, my CORE One shows VFA at every speed from 30mm/s to 150mm/s+.

 

Posted : 01/04/2025 11:58 am
Scotttomo and yblaser liked
darksharpie
(@darksharpie)
Estimable Member
RE:

Yep.  That's exactly what I see on the X1C we have at work.  With anything printed either on matte PLA, or "regular" PLA that has been turned matte by the flow/cooling of the BL hotend at speed, the VFAs are difficult to see or invisible.  But if I pick up anything we still have around that has ever been printed on a shiny or silk PLA, they're easily visible an any flat surfaces.  This is with the "Standard" print profile setting.

Posted by: @scott-18

200mm/s gives weaker layer adhesion and not even finish on the print (it has to slow down on some small perimeters or overhangs) so no, I don't want to be forced to print at 200m/s outer wall.

The mat finish at this speed helps to mask VFA but that's all.

My X1C shows VFA between 70mm/s and 120mm/s, my CORE One shows VFA at every speed from 30mm/s to 150mm/s+.

 

 

This post was modified 1 month ago by darksharpie
Posted : 01/04/2025 12:02 pm
Kachidoki
(@kachidoki)
Trusted Member
RE: VFA Artifacts on X+Y Straight Edges

That is the current status in 6.3.0: Motors are driven in phase-stepping mode (whatever that means), but it cannot be calibrated and hence cannot correct for deviations/imperfections in individual motors.

I mean, are the 'correct' default values already set into the firmware, like for input shaping?

What I understood is that 6.3.0 has phase stepping enabled but not calibrated. Maybe the version 6.3.1 includes the default 'common' calibration values, even if it is not possible to calibrate it ourself yet.

Posted : 01/04/2025 12:12 pm
Gizzle
(@gizzle)
Active Member
RE: VFA Artifacts on X+Y Straight Edges

 

Posted by: @kachidoki

That is the current status in 6.3.0: Motors are driven in phase-stepping mode (whatever that means), but it cannot be calibrated and hence cannot correct for deviations/imperfections in individual motors.

I mean, are the 'correct' default values already set into the firmware, like for input shaping?

What I understood is that 6.3.0 has phase stepping enabled but not calibrated. Maybe the version 6.3.1 includes the default 'common' calibration values, even if it is not possible to calibrate it ourself yet.

 

If it's enabled it already has the common values. What we need is the possibility to calibrate the values by ourselves since not every motor, table etc. is the same

Posted : 01/04/2025 12:14 pm
Biomech
(@biomech)
Trusted Member
RE:

 

Posted by: @scott-18

200mm/s gives weaker layer adhesion and not even finish on the print (it has to slow down on some small perimeters or overhangs) so no, I don't want to be forced to print at 200m/s outer wall.

The mat finish at this speed helps to mask VFA but that's all.

My X1C shows VFA between 70mm/s and 120mm/s, my CORE One shows VFA at every speed from 30mm/s to 150mm/s+.

 

Core One prints just fine at 200mm/s. Speed presets goes to up to 300mm/s with high flow nozzles. And that are still conservative values. If you go to like 350mm/s and more then you can start noticing some issues. Depends on material, but usually few more degrees on hotend solves it.

Posted : 01/04/2025 12:31 pm
darksharpie
(@darksharpie)
Estimable Member
RE: VFA Artifacts on X+Y Straight Edges

It can print fine, but that doesn't mean the resulting print has the optimal layer adhesion (strength) and consistent surface finish i.e. varying matte and shiny areas  on the same object as @scott-18 was alluding to.

Posted by: @biomech

 Core One prints just fine at 200mm/s. Speed presets goes to up to 300mm/s with high flow nozzles. And that are still conservative values. If you go to like 350mm/s and more then you can start noticing some issues. Depends on material, but usually few more degrees on hotend solves it.

 

Posted : 01/04/2025 12:39 pm
Scott liked
Chocki
(@chocki)
Noble Member
RE: VFA Artifacts on X+Y Straight Edges

All, it's the 1st of April 😆 New firmware tomorrow, come on, someone's having a laugh.

Normal people believe that if it is not broke, do not fix it. Engineers believe that if it is not broke, it does not have enough features yet.

Posted : 01/04/2025 12:43 pm
Jürgen
(@jurgen-7)
Honorable Member
RE: VFA Artifacts on X+Y Straight Edges
Posted by: @kachidoki

That is the current status in 6.3.0: Motors are driven in phase-stepping mode (whatever that means), but it cannot be calibrated and hence cannot correct for deviations/imperfections in individual motors.

I mean, are the 'correct' default values already set into the firmware, like for input shaping?

What I understood is that 6.3.0 has phase stepping enabled but not calibrated. Maybe the version 6.3.1 includes the default 'common' calibration values, even if it is not possible to calibrate it ourself yet.

My understanding -- based on an educated guess, since I did not find any technical documentation: The default values are "correct" for perfect motors, with perfectly equidistant steps and microsteps.  But for those, you do not need any special "phase stepping" at all, they will run smoothly anyway.

The whole reason for phase stepping is that it can drive the motors with individually calibrated stepping curves which compensate for asymmetries and imperfections in the motors. I assume this is either done by tweaking the timing or the currents per microstep. This will give smoother motor operation with imperfect motors. But to do this and benefit from it, you need to calibrate.

Posted : 01/04/2025 12:59 pm
yblaser liked
Kachidoki
(@kachidoki)
Trusted Member
RE: VFA Artifacts on X+Y Straight Edges
Posted by: @kachidoki

I can't wait for Prusa to release the assembly manual. Once it is online, I'll probably reassemble the CORE One myself. I have multiple problems that could come from the assembly and/or the delivery. I'm hoping the belt ripple will disappear, but it is very unlikely.

Well, it's time => https://help.prusa3d.com/manual/prusa-core-one-kit-assembly_2251

 

Posted : 01/04/2025 3:16 pm
altaic liked
baztm
(@baztm)
Eminent Member
Topic starter answered:
RE: VFA Artifacts on X+Y Straight Edges

Just the menu item was disabled. The phase stepping and config for it is enabled. It's speculation at the moment until we get a release or RC

Posted : 01/04/2025 4:29 pm
Jürgen
(@jurgen-7)
Honorable Member
RE: VFA Artifacts on X+Y Straight Edges
Posted by: @baztm

Just the menu item was disabled. The phase stepping and config for it is enabled. It's speculation at the moment until we get a release or RC

True. But I am pretty sure it's not the additional text for the menu item, but the execution of the calibration algorithm which causes the "out of memory" crash. It's another instance where I wish Prusa had not prioritized MK4S compatibility, and had rather given the Core One a more powerful main board. Fingers crossed that they figure out a software workaround quickly! 

Posted : 01/04/2025 4:43 pm
Kachidoki
(@kachidoki)
Trusted Member
RE: VFA Artifacts on X+Y Straight Edges

It's another instance where I wish Prusa had not prioritized MK4S compatibility, and had rather given the Core One a more powerful main board.

The xBuddy board was and is still a very powerful board to drive a 3D printer. When I remember how far they were able to push the old 8-bit board, it was awesome. I hope they could do the same with the xBuddy.

The fact is that it was a Prusa choice to not embed a Linux platform with plenty of compute power/flash/ram on their printers. This is true for the motion controller, but to handle all the connectivity, webserver and UI it would have been a better choice to delegate this to a real computer platform. Maybe they will come to it one day and our printer will take a minute or so to boot up and .bgcode will be uploaded in no time to compensate. 😆 In the meanwhile they need to make their common platform profitable.

Posted : 01/04/2025 5:59 pm
Martin, altaic and Rainer liked
altaic
(@altaic)
Trusted Member
RE: VFA Artifacts on X+Y Straight Edges

True. But I am pretty sure it's not the additional text for the menu item, but the execution of the calibration algorithm which causes the "out of memory" crash. It's another instance where I wish Prusa had not prioritized MK4S compatibility, and had rather given the Core One a more powerful main board.

I wonder why it works on the XL and not the Core One. And it looks like it's now enabled on the MK4 as well, apparently without memory issues. They all use the same STM32F427ZIT6 microcontroller (see xBuddy pic and XLBuddy pic).

Posted : 01/04/2025 9:45 pm
Kachidoki liked
Jürgen
(@jurgen-7)
Honorable Member
RE: VFA Artifacts on X+Y Straight Edges
Posted by: @kachidoki

The xBuddy board was and is still a very powerful board to drive a 3D printer. When I remember how far they were able to push the old 8-bit board, it was awesome. I hope they could do the same with the xBuddy.

The fact is that it was a Prusa choice to not embed a Linux platform with plenty of compute power/flash/ram on their printers. This is true for the motion controller, but to handle all the connectivity, webserver and UI it would have been a better choice to delegate this to a real computer platform. Maybe they will come to it one day and our printer will take a minute or so to boot up and .bgcode will be uploaded in no time to compensate. 😆 In the meanwhile they need to make their common platform profitable.

The STM32 they use is powerful for a motion controller (although the phase stepping calibration is now apparently pushing it to its limits in that area too). But it's not up to the task regarding the Ethernet/WiFi connection. Requiring a camera with its own separate WiFi connection is a kludge in my book. 

Posted by: @altaic

I wonder why it works on the XL and not the Core One. And it looks like it's now enabled on the MK4 as well, apparently without memory issues. They all use the same STM32F427ZIT6 microcontroller (see xBuddy pic and XLBuddy pic).

Is it really enabled in the MK4? I never saw VFA or noise complaints for that printer. Maybe the shorter belts reduce the risk of belt resonances, so you don't need the extra-smooth (and compensated/calibrated) motor control via phase stepping?

It might be the 0.9° stepper motors which require larger sampling rates, longer sample records or whatever during the calibration, exceeding the available memory? But that theory would only make sense if neither the Core One nor the MK4 can use phase stepping at the moment.

Posted : 01/04/2025 10:04 pm
altaic liked
altaic
(@altaic)
Trusted Member
RE: VFA Artifacts on X+Y Straight Edges

Is it really enabled in the MK4? I never saw VFA or noise complaints for that printer. Maybe the shorter belts reduce the risk of belt resonances, so you don't need the extra-smooth (and compensated/calibrated) motor control via phase stepping?

It might be the 0.9° stepper motors which require larger sampling rates, longer sample records or whatever during the calibration, exceeding the available memory? But that theory would only make sense if neither the Core One nor the MK4 can use phase stepping at the moment.

It appears to be enabled on the MK4, see commits:

Not sure why or what would cause the OOM on only the Core One. However I did notice that they completely rewrote `phase_stepping/calibration.cpp`. It looks pretty fancy now, but I guess there's a bug in there, perhaps triggering a stack overflow since it uses a recursive search. Didn't have time to give it a thorough look over, though.

Posted : 01/04/2025 11:21 pm
Jürgen and Brian liked
Logicwax
(@logicwax)
Active Member
RE: VFA Artifacts on X+Y Straight Edges

Is it really enabled in the MK4? I never saw VFA or noise complaints for that printer. Maybe the shorter belts reduce the risk of belt resonances, so you don't need the extra-smooth (and compensated/calibrated) motor control via phase stepping?

Most likely its the coreXY design.   To move in a straight line you need two stepper motors to be completely in sync.  If you spin one, you get a 45 degree angle.  If you spin another, the head moves perpendicular to that in another 45 degree angle.   To go straight you need both to essentially fight each other, kinda.    Anyway, the issue is that no matter how perfect you sync them up, there will always be a ripple since both motors need to spin opposing directions to sync up their 45 degree angles to cancel each other out.

MK4 doesn't need to do any of that.    Straight line is just moving one motor in one direction.   Easy.

You can try rotating your parts 45 degrees in slicer and you'll notice less ripple on the core one. (for an example, printing a box).

Posted : 02/04/2025 9:44 am
Scotttomo liked
Page 12 / 23
Share: