RE:
I had previously printed the G-code that the OP (@baztm) had posted, so I thought this would be a good first test for the new belts and pulleys. I used the same silk PLA, to get the most shine and show the VFAs at their worst, but note that I don't think the G-code was for silk PLA, just normal PLA.
In each of the photos below the original print with the stock GT2 setup is the lower object, and the latest print with the GT1.5 setup is the upper object. Quite a dramatic improvement. It's not perfect - there's still a hint of belt ripple, but being shorter wavelength and a noticeably smaller amplitude the result is very much better.
I should note that I initially set the belt tension to 85Hz, as per Prusa's instructions, but after this print I checked again and both had dropped to around 65Hz - just the result of some initial shakedown I guess, but the loose tension may have impacted the print quality.
RE: VFA Artifacts on X+Y Straight Edges
Yeah, I had to adjust the belts after a week or two because they were no where near 85Hz. They had 100% broken in... I probably should check them again.
RE: VFA Artifacts on X+Y Straight Edges
I've also done some VFA speed towers. They show that there's still some belt ripple at certain speeds, but on the whole the ripple is much reduced - the speeds it shows up at are different, the wavelength is shorter (naturally!) and the amplitude seems noticeably reduced.
Some photos:
First photo, top is GT2, bottom is GT1.5. Left 'arm' is 45 degrees. No question - the GT1.5 is better. Still some belt ripple visible, but only from the 11th band (130mm/s) upwards, and it's both shorter wavelength (1.5mm) and visually a smaller amplitude. Second photo - similar story at 30 degrees (apologies for the poor lighting on the GT1.5@30 deg section).
Third photo - similar story at 90 degrees. Fourth photo - similar story at 0 degrees, although the GT1.5 photo shows a clear band of belt ripple at 130mm/s. Fifth photo - similar again at 60 degrees, although the most pronounced belt ripple band is at 140mm/s on this one.
The photos above were with the belt tension readjusted to 85Hz. I then tightened the belts to 97Hz and made another print.
Left is 97Hz, right is 85Hz - the pronounced belt ripple at 130mm/s is marginally improved at the higher tension.
I'm not sure what to make of all this - certainly I'm seeing very little belt ripple with GT1.5 below 130mm/s, so the 80mm/s perimeters in the 0.2mm Structural profile should be nice and clean. The 170mm/s perimeter speed of the 0.2mm Speed profile will show a bit of ripple, but it's much less marked than with GT2. In fact, the GT1.5 results are clean enough at slower speeds that another source of VFAs (motor ripple?) seems to start showing up. I'm not sure any of my photos captures that though. On the whole I'm pleased, and I think in a more normal print the results will be worth the trouble of making the swap.
RE: VFA Artifacts on X+Y Straight Edges
Very interesting findings, thanks for going to the effort of reporting back with such detail, Chris.
RE: VFA Artifacts on X+Y Straight Edges
I also ran the phase stepping wizard after these prints, with the belt tension still around 95Hz. I'd never previously managed to get a result out of it, which I'd put down to the fact that my motors are already rather quiet, but this time it came back with a barely credible 55% reduction in X vibration and 91% reduction in Y vibration! I wonder if the algorithm is somehow affected by the change in the step length. I ran the IS wizard as well, and then printed another VFA speed tower. I don't think there's any change, but I do think the motors are even quieter now 🙂
I'll have to do some normal prints to be really sure, and to get a bit more experience of the quality.
RE: VFA Artifacts on X+Y Straight Edges
Great results, thanks for sharing. Since I still have a little bit of rippling at certain angles with 2GT, I might try the 1.5GT swap as well once my replacement printer arrives (a long story of severe EMI artefacting when I connect either the MMU3 or the Advanced Filtration lol). Assuming the replacement will work properly. This printer model has been a mess, they released it way too early imo.
RE: VFA Artifacts on X+Y Straight Edges
Can you post links to the parts you used? Belts and Pulleys? Approximately what is the cost (aside from time) for the parts. Also did you have to re-print anything to fit the new parts. On the Mk4(s) they used a tooth design to mesh with the belt ends to lock them in place. If similar in the Core One, the teeth wouldn't mesh properly without modification. Was this a concern?
-J
RE: VFA Artifacts on X+Y Straight Edges
I forgot to ask, are you using the stock smooth idlers everywhere? Or did you replace the 2 idlers with toothed versions? If yes, which exact part and tooth count?
RE: VFA Artifacts on X+Y Straight Edges
Can you post links to the parts you used? Belts and Pulleys? Approximately what is the cost (aside from time) for the parts. Also did you have to re-print anything to fit the new parts. On the Mk4(s) they used a tooth design to mesh with the belt ends to lock them in place. If similar in the Core One, the teeth wouldn't mesh properly without modification. Was this a concern?
-J
I used Powge 21T GT1.5 pulleys, and genuine Gates belts. They were sent to me free-of-charge by Nathan at Ston-3D (I've pre-ordered the new Ston Wolf printer, so I guess you could call it a perk of being an early adopter!). The pulleys are apparently Nathan's design. He plans to sell them in his shop eventually, but I know you can get them from AliExpress already. Not sure about the belts though - I can see Powge own-brand belts on AliExpress, but I haven't found a source for the genuine Gates belts - I must admit I haven't looked very hard though.
The belts lock against themselves behind the Nextruder mounting plate ( https://help.prusa3d.com/guide/5-corexy-assembly_835522#845973 - third photo), and are clamped when that plate is screwed to the linear rail. So there were no parts that needed reprinting. Based on my positive experience with the CORE One I've been looking at whether it would be interesting to do a similar swap on one of my MK4Ss, and you're right - the printed parts are specific to GT2 belts. I've had a go at modifying those parts, and have printed some versions that will suit a 1.5mm belt pitch.
I forgot to ask, are you using the stock smooth idlers everywhere? Or did you replace the 2 idlers with toothed versions? If yes, which exact part and tooth count?
I'm still using the original smooth idlers.
RE: VFA Artifacts on X+Y Straight Edges
Thanks for the tests and updates guys! Nice to see some improvements.
Just want to add this post from the Prusa Facebook group: https://www.facebook.com/groups/547901291210777/permalink/706264752041096/
He previously tried the GT1,5mm belts and pulleys but couldn't figure out the sizing fix yet. (Thanks for the clarification on that!) I asked him to order some Mellow3D pulleys, both motor pulleys and idlers. He did and shows his result in the Facebook post.
I think the results speak for themselves. We need more tests but I feel it clearly shows that very good results can be had with a GT2 belt as well.
RE: VFA Artifacts on X+Y Straight Edges
Thanks for the tests and updates guys! Nice to see some improvements.
Just want to add this post from the Prusa Facebook group: https://www.facebook.com/groups/547901291210777/permalink/706264752041096/
He previously tried the GT1,5mm belts and pulleys but couldn't figure out the sizing fix yet. (Thanks for the clarification on that!) I asked him to order some Mellow3D pulleys, both motor pulleys and idlers. He did and shows his result in the Facebook post.
I think the results speak for themselves. We need more tests but I feel it clearly shows that very good results can be had with a GT2 belt as well.
That's good then. That Facebook group is private though - any chance you can ask his permission and then post some of his images, for the benefit of those of us who aren't on Facebook?
RE: VFA Artifacts on X+Y Straight Edges
Sure, the following is the post from the Facebook group by Ondrej:
I have now finally finally got some test results and to relief of many this will be my last post on this topic. If you have followed my journey I have started with trial of changing the pulleys and belts to 1.5mm this has solved ome issues while creating many others, read my older posts if you want.
The next step was a community suggestion for replacing the pulley with one with couple of toothed pulleys and perhaps also the motor pulleys for some better quality ones.
Well it worked, not perfect but much improved now.
I have modified both of my printers and they should be identical twins now.
With toothed pulleys the VFA is far less visible and both printers have now identical print quality.
I wish I had the parts with more visible VFA, but this took me so long to sort out I have run out of stock and had to ship them.
Rest of this will be in Prusa's hands. I now have two printers that are doing their job, but still think that the MK4S had much better print quality.
Given that most of my prints are ASA and ABS the conversion was pretty needed as it also reduced the footprint of the printer very significantly.
I hope Prusa will soon have a proper answer that will work for all of us, kind of for their own sake, as much I love them and want to support them I also need to be looking for my business needs and one of my USP is that I care about the quality of parts I make.
RE: VFA Artifacts on X+Y Straight Edges
I also have a set of the Mellow 3d pulleys on hand in case I have trouble with my stock pulleys, and will keep this GT1.5 conversion in mind as well. I'd really prefer to avoid having to worry about scaling issues and whether or not I've restarted my printer lately etc., so I'm hoping to keep the GT2 stock stuff if possible. Crossing my fingers that mine won't be one of the 'bad' ones. This is "conversion week" and I finished my teardown last night, so hopefully I'll have some non-theoretical feedback to contribute soon. Going to try and borrow a machinist square from the machine shop today for straightening my gantry brackets, and will be checking the Z axis screw pitch as well due in the build. It's a bummer we have to do this stuff, but if the results are ultimately good, the conversion will still be worth it to me. 🙂
-J
RE: VFA Artifacts on X+Y Straight Edges
I'm honestly surprised that toothed pulleys give better results.
Been doing these on the Creality K1 and that changed nothing for me.
Of course other printer...
I run 93hz or between 90 and 100 as belt tension is still unstable.
Got gates belts and powge pulleys.
The gates belts are from a direct source but you can get powge gt1.5 belts and pulleys on ali.
Belts length 5 meter
1.5Gt belts https://s.click.aliexpress.com/e/_olTpFB2
pulleys
https://s.click.aliexpress.com/e/_oBjvigu
Sorry for the slow replies, i'm still under probation lol.
Unrelated question, i figured that the structural profile has the worst vfa and i keep wondering why anyone would use this profile on a core xy to begin with?
Sub 100mm/s doesn't make sense on this motion system.
This is not to offend anyone that uses it, you do you but why?
RE: VFA Artifacts on X+Y Straight Edges
Sure, the following is the post from the Facebook group by Ondrej:
Thank you.
RE: VFA Artifacts on X+Y Straight Edges
Interesting that Josef still thinks the VFAs can be resolved with tuning:
https://www.reddit.com/r/prusa3d/comments/1lp3eo7/psa_vfa_update_2/
RE: VFA Artifacts on X+Y Straight Edges
I guess because VFAs are only one concern. The speed profile can give inconsistent surface finish with shiny filaments, and also probably makes weaker prints where it turns them "matte". (I know there are a lot of other factors here in this oversimplification- I just don't think there are no downsides to printing faster walls).
Unrelated question, i figured that the structural profile has the worst vfa and i keep wondering why anyone would use this profile on a core xy to begin with?
Sub 100mm/s doesn't make sense on this motion system.
This is not to offend anyone that uses it, you do you but why?
RE: VFA Artifacts on X+Y Straight Edges
I'm not sure whether Prusa is working on the actual cause of the issues. Even if we assume that everything can be fixed through a specific belt tension value, it doesn't matter which algorithm the tuning app uses as long as you have a wide variety of phone hardware, which probably isn't designed to be accurate <100Hz. Also, what good does testing a new default on in-house printers, which allegedly didn't have these problems to begin with, do?
If Prusa wants to commit to this path, I would hope that they at least publish a list of phone models which have been tested to be accurate enough. Ideally though, they come up with a solution that doesn't depend on hardware they have no control over.
Interesting that Josef still thinks the VFAs can be resolved with tuning:
https://www.reddit.com/r/prusa3d/comments/1lp3eo7/psa_vfa_update_2/
RE: VFA Artifacts on X+Y Straight Edges
I guess because VFAs are only one concern. The speed profile can give inconsistent surface finish with shiny filaments, and also probably makes weaker prints where it turns them "matte". (I know there are a lot of other factors here in this oversimplification- I just don't think there are no downsides to printing faster walls).
Matte finish means that plastic was not melted properly. Usually it means that hotend temperature is too low and hotend does not keep up with volumetric flow rate demands. Melting plastic absorbs a lot of heat. Increasing hotend temperature helps giving it heat fast enough. And there are other options, like reducing flow rate by reducing layer height (popular "Bambu trick" to print fast at 0.1mm layers). There are also high flow filaments with additives for faster melting. And flow rate can be reduced also by reducing speed, but that may not be the right option because of VFAs at lower speeds.
RE: VFA Artifacts on X+Y Straight Edges
Yes, agreed.
I was simply saying there were still reasons to choose to print at a slower speed and makingr tradeoffs that could include more visible VFAs.
I guess because VFAs are only one concern. The speed profile can give inconsistent surface finish with shiny filaments, and also probably makes weaker prints where it turns them "matte". (I know there are a lot of other factors here in this oversimplification- I just don't think there are no downsides to printing faster walls).
Matte finish means that plastic was not melted properly. Usually it means that hotend temperature is too low and hotend does not keep up with volumetric flow rate demands. Melting plastic absorbs a lot of heat. Increasing hotend temperature helps giving it heat fast enough. And there are other options, like reducing flow rate by reducing layer height (popular "Bambu trick" to print fast at 0.1mm layers). There are also high flow filaments with additives for faster melting. And flow rate can be reduced also by reducing speed, but that may not be the right option because of VFAs at lower speeds.