I appreciate the clarification! I think that is a good point and could possibly contribute to the fine VFA at lower speeds more than the mid-speed VFA on X-axis? Though, since the X1 uses two wheels driven by a single motor, I can’t really imagine how they’d be anything but synchronized…
Wild about the Creality skipping it on one axis! On Bambu I think it’s not the issue (or if it is it’s not the one I’ve been fighting) since its intensity would change more based on speed, would have a different wavelength at different speeds, and generally speaking I haven’t seen ringing more on one axis than another.
I think the really troublesome thing about solving this issue is that the components theorized to be to blame are extremely hard to troubleshoot. Putting in random tooth inserts and stuff is never going to accurately compare to a solid pulley of the exact right diameter. In my tests in that avenue, I could clearly see that any mismatch in the number and spacing of teeth would result in the rise and fall of the belt every 3-4 teeth, depending on the mismatch. That manifested as a few widely spaced extra large bumps in the VFA (which further suggests to me that the issue is the pulleys.)
When I get the new axis in with the split-loop belt, I’ll report back. It will also be interesting to see if the carbon axis bearings have significantly differing play between the old and new components as I have some suspicion that slop in them could be allowing the belt-vibration to manifest more/less in different printers. I question the choice of these carbon rods more all time because when the belts are off you can feel the Y-bearings are both tighter tolerance and lower friction.
Have you followed Nebur’s look into this matter? If not, you may find it interesting reading:
This paraphrase is probably not doing it justice, because I’m going by memory, but he found that the carbon rods and/or the parts holding them had deformation under the belt tensioning, resulting in a touch of inadvertent non-uniform sticking as the x-axis bearings moved along them, enough to cause VFA’s.
He was working on a hardware solution. I just now pinged him to see how that went and whether he has any kind of update.
I remember reading that at some point yeah. But I don’t think the last time I saw it that anything had come of it yet. I’ve kind of developed the conviction that the carbon rods is a gimmick. While they may weigh less than steel, they seem to introduce a few more problems which may include sloppy tolerances, higher friction in general, and they’re very prone to damage.
I used to believe in upgrades, because not long ago what alternative was there? The new paradigm seems to be: maintain as best you can, but otherwise keep as-is, sell as-is, and just roll whatever you can recoup on the re-sale into the next, better-thing when it becomes available. If you upgrade it, it becomes an unknown quantity for anyone who might buy it, and by this argument, it becomes less valuable to them, and you don’t get as much on the re-sale. Maybe even much less, or nothing at all. Ironic, isn’t it? Way different than how things used to be.
On the other hand, the re-sale value probably won’t be very high anyway, so you may not want to sell at such a low price and therefore keep it. In this case, you might as well upgrade it, because then at least it’s good for what you want, until you decide it’s obsolete and scrap it.
I haven’t yet made up my mind between these two very different ways of looking at things.
In the same boat… 1 week old p1s
I’m not willing to modify my printer, since i also have a Voron which is already modified by nature.
I like to print very slow, then the printer is really quiet. I wanted a reliable multi color printing, speed can be a bonus sometimes.
What i found so far is that acelleration on outer walls also plays a huge role.
This eventually affects the real speed on a pass, so not sure if this will really help.
I don’t know whether you will find this useful or not, but this guy just did a review on the Plus4 and comments upon a test print that always got him VFA’s on the X1C but which doesn’t have any that he can see when printed on the Plus4:
For your convenience, I set the time index to that position in the clip.
Earlier in the clip he points out differences in the belts used by the Plus4 vs the X1C.
I don’t think one can take it as an airtight proof of cause and effect, since there could be confounding factors, but it certainly is suggestive and maybe even indicative of what kind of mod might be worth trying.
If nothing else, it’s an existence proof that it is possible to eliminate VFA’s, at least in the prints that he tried.
Oh wow, that belt is an odd one indeed. I’m not aware of any standard pitch below 2mm so I’m curious how custom it might be.
I’m leaning away from strictly belt things on my original printer now - I tried something pretty outlandish today by wrapping a couple layers of gaffers tape around 1" of belt and printing my VFA test. It skewed the print, but it also still had the same VFA pattern on top of the shift introduced by the extra thickness.
Simultaneously it seems like there is more VFA creeping into the Y axis travels - it used to be only on my X. I’m starting to think the stepper motors are to blame and getting worse after a few thousand hours of printing. When moving across X the motors are stepping at the same pace and direction but when moving across Y, they are moving opposite and might have a degree of destructive interference which reduces the effect on Y.
I am interested in the dampers @Alkerion installed with the toothed idlers because I’m thinking now that it might actually play a larger role in smoothing out the stepper cogging than expected.
Also trying to find what the specs are for the Bambu steppers, but so far without luck - I was curious whether anyone had replaced them with aftermarket units.
Those who have changed to toothed idlers, did you also replace with new carbon rods as well?
I just received a new carbon rod assembly and am getting ready to install. I have a kickstarter unit and have noticed more friction (can feel and hear) in the X axis. I can clean them but after about 2-3hr of printing, the resistance reappears so I suspect the bearings/bushing are shot.
I figured I would order some genuine Gates toothed idlers but am debating whether or not to modify the assembly. Don’t really want to take it apart and put it together more than once. It would be nice to get another data point for the group but it would mean I am changing several variables at once, which is not ideal. Basically new rods, new Bambu belts and Gates toothed idlers. I will first be performing belt tensioning according to Bambu’s instructions and then may resort to other methods.
I’m fine with drilling/grinding out the rod so the pin can be driven out the way it was inserted. Has anyone got the pin out without having to drill a new access hole on the opposite side? Wondering if I can get the pin starting to move and then use some sort of pick/hooked tool to keep driving the pin outward.
I have done quite a bit of testing with different speeds and material (mainly Polymaker petg, Bambu petg-hf and some eSun silk pla.
I found something which is mentioned earlier: it seems to happen on all speeds, it just doesn’t show when the print gets matte due to the higher speed. As soon as you turn up the nozzle temp, the vfa comes back, even at the higher speed.
Matte PLA prints perfectly fine, in my case.
Although i did not want to mod this printer, since i have a Voron to mod, i’m temped to order an extra axis.
You are correct. I completed the same testing and they do in fact return as you increase the temperature, which is necessary for strong layer adhesion. Pictures below are of my testing as I was trying to hone in best appearance and strength. All reference X only movement.
I decided to move forward with the modification on my new x-axis carbon rod assembly and get us another data point for this experiment. Still waiting for the genuine Gates idlers to come in before I can move forward with the modification. Here is the drill guide I designed and used for those wanting to try this out themselves.
What kind of testing would you guys like to see since my printer is still functional so I can get some good before pictures? I was thinking some orca slicer VFA tests at different temperatures to compare before and after.
I advise to also order graphite gaskets as they are really fragile and rather hard to put in pkuce, but absolutely mandatory.
Est solution I found was with tape, let see if someone has a better solution.
This is looking promising with a 0.2mm nozzle at 0.05mm layer height. Need to wait for the idlers to come in to make sure it works with them as well and then I will upload.
I also ordered some metal shims from Amazon that are the same dimensions just in case. I will also have my old X-Axis I can cannibalize if I need to.
^This is the test I use. While the issue does exist on Y travels, it is by far the worst on X and the worst at the rear so I made a test to prioritize X travels. It also seems to be affected by which direction the toolhead is moving so this test I designed to ensure it moves in two directions. I’d love to see you benchmark with this model as it’s so much easier to interpret without having to hold it in person.
I’ve been printing with clear filament recently which helps show the imperfections even when a satin-glossy surface wouldn’t. (And the case where these ripples have caused me a headache for a year is using a transparent PLA for a print.)
I’m still running experiments because it seemed like even wrapping the belt in thick tape where it would contact the pulleys was not changing anything. I’ve tried skewing the pulleys to make the belts track different, deliberately moving the axes out of square, putting a flexible PLA damper between the motor and chassis etc.
My current line of testing is a refined add-on toothed pulley that wraps onto the existing pulley using a very thin vinyl to secure it. So far my tests with it look identical to without so that suggests I made the model right. I just installed a version with domed teeth so that it should help force the belt to track more centered like the wheels on a belt sander.
My last ideas here are replacing motor pulleys - I did one previously because my KS model shipped with short pulleys, but I left the other alone. Another idea, replace the motor pulleys with a finer tooth belt drive and go with a non GT2/2GT belt. Following that, take my X-axis back out and swap in toothed idlers. And following that, throw my printer away or keep buying and selling til I get one that doesn’t suck (does the A1 have this problem?)
Replacing the ‘B’ motor pulley with a taller metal one made no difference that I can observe. What I am noticing is that both sides are tracking up and down severely and I can’t see any pattern to it.
Look promising.
The toothed pulleys I’ve used was taller than original, it was rather impossible to insert anything top/bottom thicker than tape.
You find toothed pulleys with exact same dimensions?
Almost identical to the later version X1s overall. Same tooth count, etc might be slightly thinner flanges or slightly taller, but a direct fit. For a long time I’ve been using one OEM and 1 aftermarket and now 2 aftermarket but it looks the same as ever.
I went around to the squaring and skew/tracking adjustment tonight. My X is within 0.6mm across the span and a few layers of gaff tape against the back wall allows the tension to even it out. I skewed the tensioner pulleys to try to track centered as possible as some say worked. Absolutely no change whatsoever.
I’ve noticed that the front idlers and the turning idlers with tooth-on-face always are riding on the flange no matter how skewed I make it. I was wondering if the front idlers were slanted so I just ran one of my benchmarks using 2 plastic sticks against the back side of the belt to try to keep it centered on the turning pulleys throughout the print. Once again, absolutely no change, though I noticed I can feel a buzz in the belt on the X travels.
The thing that’s still holding me back from doing a toothed-pulley-replacement (other than a lack of sufficient data points demonstrating it’s the big fix) is that I am seeing the same patterns on Y although less. Y travels involving no movement across those pulleys seems like the toothed idlers won’t help. I’m really stumped.
