RE: VFA Artifacts on X+Y Straight Edges
In both of the reviews I saw, there were some great prints and some flawed prints. So I wouldn't say it came out of the gate perfectly, despite the "15 sensors" and all of the other changes from the filament to the nozzle. I will certainly see more reviews, but VFAs are difficult to see on camera when you're not holding the model, and they aren't even on everyone's radar (not to mention not every review test prints large flat areas where they are even particularly visible).
I don't know if anybody has been watching the reviews on the H2D. I'm definitely not a Bambu fan, but I think as far as print quality it looks like Prusa needs to step up their game. The prints look pretty flawless with no VFA'S.
Hopefully that will be the kick in the pants Prusa needs to address this.
I'll reiterate that my MK4 has almost zero VFA'S and amazing print quality. But my XL and it looks like many people's Core One do......
Agreed, it is hard to say, but we can't argue that at least that were having some issues. I could see VFA'S on the prints in many Core One reviews.
The Clough42 review has some great looking prints.
The Robert Cowan one compared prints between the XL and the H2D and according to him the H2D was better hands down. Now take that with a grain of salt because he had been very anti Prusa in my opinion since he had some issues with his XL. I also don't find him as smart as he tries to act like he is.
I look forward to the Aurora Tech review as how she compares the printers seems the most apples to apples as she tests the same parts on every printer she reviews, and usually has good images to compare.
RE: VFA Artifacts on X+Y Straight Edges
I agree. There are some aspects of printer design where Prusa won't be able to keep up with Bambu and needs to play a different game: Those dozens of sensors and cameras, those highly integrated mechanical functions in delicate molded parts... And I'm ok to trade this for solid, long-lived engineering with a bit less automation and more responsibility for the user.
But there is no excuse for not being able to move a print head smoothly. If Prusa falls behind in that core competAnd move that print head quietly, by the way. Quoting from Clough42's review of the H2D: "The H2D motors are so quiet that I have at times looked into the printer to confirm they were still running. My X1 printers I can hear from outside the room. My Prusa XL, from the other side of the house."
I agree on all of the sensors, they are cool no doubt, but I'm fine without them as well. I didn't mind tweaking, and honestly I've found that with the exception of TPU, when using dry filament the Prusament profiles are spot on, even without using Prusament filament.
I don't doubt the H2D is quiet, but my XL is not loud anymore. Not as quiet as the MK4, but he's greatly exaggerating that he can hear it across the house. Before the phase stepping yes, it was ridiculously loud, but not anymore.
RE: VFA Artifacts on X+Y Straight Edges
One thing to be aware of, even tho Prusa uses "GT2" belts, you might want to buy a genuine gates belt and replace it as well. The teeth might not line up exactly with the new pully grooves.
Not sure about other peoples' CoreOne units but mine at least has genuine belts. As far as I can tell, Gates says they're the same basic physical construction (nylon and fiberglass reinforced) but the "E" (EPDM) model is rated up to 135C where the RF (rubber) model is rated up to 85C.
RE: VFA Artifacts on X+Y Straight Edges
Cool! I didn’t see any manufacturer marking on my belts - maybe I just didn’t look at the right spot. I just checked on the part of the belt that runs along the X axis. And looking on the store at belts I don’t see any manufacturer markings. My caution was just to make sure that the belts and pulleys are matched. If you replace both at the same time you know *for sure* they were meant to go together. Definitely wasn’t trying to say Prusa’s belts weren’t genuine or anything.
RE: VFA Artifacts on X+Y Straight Edges
I don't doubt the H2D is quiet, but my XL is not loud anymore. Not as quiet as the MK4, but he's greatly exaggerating that he can hear it across the house. Before the phase stepping yes, it was ridiculously loud, but not anymore.
I can't use my XL by night because of the very annoying noises from the toolchanges and the fast moves. Stealth mode doesn't change anything, fast moves are quieter but still too loud. Sure phase stepping has improved the printing moves noises by an order of magnitude that you can live near the printer by day. But by night you simply can't because of the toolchanges. My printer is literally at the other end of my house, one floor below. You know, even my beloved call this printer "The Loudmouth".
RE: VFA Artifacts on X+Y Straight Edges
I get really bad VFA's too. I thought I would post it here. It is only on certain travel speeds, but they seemed to be used on most outer walls. The print quality is terrible compared with the MK3s/Mk4s. The VFA's seem to be linked with the loud travel movements / vibrations.
RE: VFA Artifacts on X+Y Straight Edges
At this point, I’d really like to see an official statement from Pursa confirming they’re aware of these problems and are working on a fix. I don’t mind paying extra for a top-notch product, made in Europe, with 24/7 support. But this needs to ensure we get prints that are at least as good as those from equivalent printers made by other brands.
RE: VFA Artifacts on X+Y Straight Edges
At this point, I’d really like to see an official statement from Pursa confirming they’re aware of these problems and are working on a fix. I don’t mind paying extra for a top-notch product, made in Europe, with 24/7 support. But this needs to ensure we get prints that are at least as good as those from equivalent printers made by other brands.
You could search GitHub to make sure someone had already listed the issue. I have hope they'll make improvements. I think they are going to have to now that the bar has been raised.
The MK4 print quality was not near as good as it is now when I first got the machine.
RE: VFA Artifacts on X+Y Straight Edges
I believe Prusa definitely has the technical know-how to fix this issue. I’m seeing these artifacts on my Mk4s and waiting for the Core One upgrade. I’m just hanging on for the upgrade kit to arrive and hoping it solves the problem, but after reading all your comments, I’m guessing I’ll still have these artifacts once I upgrade to Core One, or maybe even more... I really hope they get this sorted. I’d like Prusa to confirm they’re working on it.
At this point, I’d really like to see an official statement from Pursa confirming they’re aware of these problems and are working on a fix. I don’t mind paying extra for a top-notch product, made in Europe, with 24/7 support. But this needs to ensure we get prints that are at least as good as those from equivalent printers made by other brands.
You could search GitHub to make sure someone had already listed the issue. I have hope they'll make improvements. I think they are going to have to now that the bar has been raised.
The MK4 print quality was not near as good as it is now when I first got the machine.
RE: VFA Artifacts on X+Y Straight Edges
The problem is known from the release of XL, prusa just don’t know how to built corexy. And they prefer to put pen on toolchanger to print QR codes than fix the quality. I hate that brand. I see how that perfect support works, they talk and change random parts, then they see magic improvement (like in the case of extruder stepper posted in that thread).
Personally I talked to support that I want to betatest 0.9 steppers in XL (thought that is clear that they have some testing units for updated + version), but I don’t have a reply from them since last month 😅. So probably they are playing with pens on toolchanger instead of testing better steppers.
RE: VFA Artifacts on X+Y Straight Edges
Thanks for this reply and photo this is exactly the same problem I am seeing in my original post and it is making me think it is more of a fundamental issue with the machine.
Prusa still have my Core One in as they're going to use it for new firmware testing which should help I am told, but I have heard nothing yet.
I get really bad VFA's too. I thought I would post it here. It is only on certain travel speeds, but they seemed to be used on most outer walls. The print quality is terrible compared with the MK3s/Mk4s. The VFA's seem to be linked with the loud travel movements / vibrations.
RE: VFA Artifacts on X+Y Straight Edges
Personally I talked to support that I want to betatest 0.9 steppers in XL (thought that is clear that they have some testing units for updated + version), but I don’t have a reply from them since last month 😅. So probably they are playing with pens on toolchanger instead of testing better steppers.
Be careful what you wish for. Apparently Jo Prusa has stated that he would have preferred the Core One to have 1.8° stepper motors, and that they only went with the 0.9° motors to maintain the upgrade path from the MK4S. I have only seen the second-hand summary in the post below though, have not watched the livestream:
https://www.reddit.com/r/prusa3d/comments/1imgxqb/josef_prusa_revealed_some_interesting_insights/
RE: VFA Artifacts on X+Y Straight Edges
I checked on GitHub and I didn't find anything. Maybe I didn't search well but I don't see anything regarding VFA or buzzing resonance noises.
At this point, I’d really like to see an official statement from Pursa confirming they’re aware of these problems and are working on a fix. I don’t mind paying extra for a top-notch product, made in Europe, with 24/7 support. But this needs to ensure we get prints that are at least as good as those from equivalent printers made by other brands.
You could search GitHub to make sure someone had already listed the issue. I have hope they'll make improvements. I think they are going to have to now that the bar has been raised.
The MK4 print quality was not near as good as it is now when I first got the machine.
RE: VFA Artifacts on X+Y Straight Edges
So I took a look at some prints under a microscope to try to determine what is causing the visual artifacts we are seeing. Looking at the side walls I wasn't able to see much likely due to the way the cylindrical layer lines reflect light differently at different angles. I made a new print and increased the number of perimeters until there was no infill. This also led to the print being printed without top layer(s) so I could directly observe the perimeters from the top. Since the top of the perimeters are squished down some they are easier to visualize. After printing I could visibly see the same VFAs on the top of the print as the side walls. They are the same frequency and in phase with the side wall VFAs so I am fairly confident they are produced by the same effect. Here are some pictures of the top surface:
To me it looks like the amount of filament being put down per unit area is oscillating. Since the frequency of the VFAs vary depending on the angle of the print I think that rules out the extruder varying the amount of extruded filament. That leaves either the print head speed oscillating around it's target speed or the height of the nozzle oscillating. I think the former is more likely than the latter due to the way the frequency of the VFAs follow the belt speed but can't rule out the latter.
What would cause an oscillation in the print head speed? At least I can rule out the belt teeth contacting the smooth idler because there is only one of those per belt and that particular idler will move slower during a diagonal movement compared to a pure x or y movement. The rest of the idlers are contacting the smooth side of the belt. However, I could see even those having some contribution as when the belt is wrapped around the idler the parts of the belt below a tooth are going to have higher strain then the parts of the belt without a tooth. I could see that causing some oscillations in the tension of the belt as it comes on and off the idler. It could be also variations of the speed of the stepper motor itself, but as was pointed out earlier the frequency of the VFAs are very close to the belt tooth period. I measured the VFAs on a diagonal to have a pitch of 1.35 mm. Most sections of the belt move sqrt(2) faster than the print head on the diagonal so that corresponds to a 1.9 mm pitch within the margin of error of the tooth pitch of 2 mm.
RE: VFA Artifacts on X+Y Straight Edges
I think what is happening is that the "spring like" belts are vibrating and that vibration is translated to the toolhead. Something something Hook's law. This stuff all kind of boils down to vibration and resonance. Where remember that the "resonance frequency" is the frequency that generates the biggest response, thus giving you a bigger response than the input. There are going to be higher order modes as well which are multiplies of the fundamental resonance and then there are intermod products as well.
Basically there is some amount of vibration of the belts that will shake the chassis *the most*, but everything will always shake somewhat. In the tests I ran I found that certain speeds are definitely worse for the VFAs. I would wager there exists a range of speeds where the vibration is strong enough that its making the tool head or the gantry or the entire printer shake enough to impact what the nozzle is doing.
RE:
I think what is happening is that the "spring like" belts are vibrating and that vibration is translated to the toolhead. Something something Hook's law. This stuff all kind of boils down to vibration and resonance. Where remember that the "resonance frequency" is the frequency that generates the biggest response, thus giving you a bigger response than the input. There are going to be higher order modes as well which are multiplies of the fundamental resonance and then there are intermod products as well.
In my experience:
- when it is a specific motion-induced resonance issue that can be tuned/mitigated with e.g. Klipper's input shaping (see also: https://www.klipper3d.org/Resonance_Compensation.html), that's system harmonics.
- When it is 100% repeatable at any speed like mine is (all the way from 50mm/s to 200mm/s) then it's usually not chassis/gantry harmonics, it's usually
- something mechanical like running a toothed belt over a smooth pulley or original BMG extruder dual-hob gears like you find in the Mk3S+ extruder or some voron extruders, or
- in the stepper motors/drivers like 1.9 degree vs 0.9 degree or just bad/trashy motor windings or driver chips, or
- in the profile for the stepper(s) like whether microsteps is enabled.
Considering they had this issue on the XL and resolved it there by motor tuning in the firmware, I expect there will be a similar fix for the CoreOne as I think was already discussed earlier in this thread. The challenge of course is not having a built-in accelerometer. I ordered one just to be prepared in case that was needed to do the final tuning step as the factory tuning can only get you close without having your exact machine available.
My 2 cents, anyway.
RE: VFA Artifacts on X+Y Straight Edges
I think what is happening is that the "spring like" belts are vibrating and that vibration is translated to the toolhead. Something something Hook's law. This stuff all kind of boils down to vibration and resonance. Where remember that the "resonance frequency" is the frequency that generates the biggest response, thus giving you a bigger response than the input. There are going to be higher order modes as well which are multiplies of the fundamental resonance and then there are intermod products as well.
In my experience:
- when it is a specific motion-induced resonance issue that can be tuned/mitigated with e.g. Klipper's input shaping (see also: https://www.klipper3d.org/Resonance_Compensation.html), that's system harmonics.
- When it is 100% repeatable at any speed like mine is (all the way from 50mm/s to 200mm/s) then it's usually not chassis/gantry harmonics, it's usually
- something mechanical like running a toothed belt over a smooth pulley or original BMG extruder dual-hob gears like you find in the Mk3S+ extruder or some voron extruders, or
- in the stepper motors/drivers like 1.9 degree vs 0.9 degree or just bad/trashy motor windings or driver chips, or
- in the profile for the stepper(s) like whether microsteps is enabled.
Considering they had this issue on the XL and resolved it there by motor tuning in the firmware, I expect there will be a similar fix for the CoreOne as I think was already discussed earlier in this thread. The challenge of course is not having a built-in accelerometer. I ordered one just to be prepared in case that was needed to do the final tuning step as the factory tuning can only get you close without having your exact machine available.
My 2 cents, anyway.
I did input shaper tuning on my Core One and it improved from the stock EI to ZV, which is really the "best" input shaper algorithm. So the chassis does seem relatively rigid.
I do agree that a motor tuning thing is a likely culprit, but how that translates to the toolhead would probably be the belts yeah? Regardless of where the vibrations are coming from, the belts are just a big spring oscillator.