Z-banding or Z-wobble, do we have a solution yet?
I've seen this posted many times for a variety of machines in other places, but strangely I'm not able to find it mentioned here, so I guess I'll do it. If you know of another thread regarding this issue, please link me, 'cus I'm not using the right search terms or something.
Anyway, this is the very regularly spaced horizonal banding that can be seen in pretty much all prints, usually becoming more pronounced as the layer height increases. I've attached a photo that exemplifies this issue. Aditional photos can be provided if necessary.
This is on my 5 tool machine, printing with just one tool. The problem happens with any kind of filament or print I've done to date. I keep all my filament in sealed bags with dessicant. The machine was semi-assembled, and passed all tests and calibrations first time with no problems after I finished assembling it. The issue has been present through firmware 4.7.2 with the release PrusaSlicer, and the alpha firmware and alpha slicer.
I've done a variety of screw adjustments as described in various articles here. While the banding isn't as bad as some that I've seen, it's certainly unsightly, and I'd like to eliminate it. I look forward to your assistance. Honestly, I expect this will take a hardware revision to solve, by decoupling the bed from the lead screws as has been done for a variety of other core-XY projects out there. I hope Prusa is already working on this.
RE: Z-banding or Z-wobble, do we have a solution yet?
The tops of the leadscrews are caged in Z axis bearing housings. As shown in this pic https://help.prusa3d.com/guide/4-heatbed-side-panels-assembly_487006#489641
Use the Control menu to move the z-axis from top to bottom and watch the Z axis bearing housings. They should be completely stationary. If they move you might have bent leadscrews.
RE: Z-banding or Z-wobble, do we have a solution yet?
Thanks! I didn't think it was bent, so I took video from above the bearing holders on each side. Have you done the same on your machine? If so, do you notice ANY wobble on either of them? Here are the videos:
RE: Z-banding or Z-wobble, do we have a solution yet?
On the bright side, now you know what to fix.
XL-5T, MK3S MMU3 || GUIDE: How to print with multiple-nozzlesizes do read updated replies || PrusaSlicer Fork with some flags removed || How Feasible is Printing PETG for PLA supports very
RE: Z-banding or Z-wobble, do we have a solution yet?
My left and right have zero wobble. If the banding in your first pic is spaced equally to the pitch on the z-axis leadscrew, that confirms it. Try uninstalling and reinstalling the wobbly z-axis. It might be installed crooked and that has similar symptoms to be a bent leadscrew. Check for any debris that is preventing it from sitting straight.
RE: Z-banding or Z-wobble, do we have a solution yet?
Thanks, guys! I'll check it out and report back.
On the bright side, now you know what to fix.
That looks pretty solid. I assume you have no perceivable banding in prints?
My left and right have zero wobble. If the banding in your first pic is spaced equally to the pitch on the z-axis leadscrew, that confirms it. Try uninstalling and reinstalling the wobbly z-axis. It might be installed crooked and that has similar symptoms to be a bent leadscrew. Check for any debris that is preventing it from sitting straight.
I remember being pretty careful about that kind of stuff when I installed it, but I'll check it out.
That said, isn't the whole point of the flexible bearing mount to allow for some amount of wobble in the lead screw? Presumably, Prusa does not believe that to be the cause of such things?
RE: Z-banding or Z-wobble, do we have a solution yet?
As I expected, there is no visible or tactile defect or debris that is keeping the left z-motor from mounting cleanly to the bracket. After reassembling, I made a video showing the deflection of the lead screw without the bearing bracket in place. Thinking about it, even if the motor were not properly secured to the bracket, turning the lead screw would have it rotating in place at whatever angle it ended up at. Only a bent lead screw, relative to the motor itself, or somewhere along its axis, would result in the end of it moving around as it does. Here's the video of that test.
I don't have time to hop onto support/store chat right now, but will later, unless anyone thinks I've gotten something wrong here.
RE: Z-banding or Z-wobble, do we have a solution yet?
I contacted support and shared the images and videos. they had me check tightness of screws and stuff, but in the end admitted that it looks bent. Andres wants to confirm with the "XL Specialists" before shipping a replacement.
I'm a little disappointed that this happened, but also that Prusa designers didn't account for wobbly lead screws and design some kind of decoupler so that lateral motion is not transferred to the bed to begin with.
RE:
I contacted support and shared the images and videos. they had me check tightness of screws and stuff, but in the end admitted that it looks bent. Andres wants to confirm with the "XL Specialists" before shipping a replacement.
I'm a little disappointed that this happened, but also that Prusa designers didn't account for wobbly lead screws and design some kind of decoupler so that lateral motion is not transferred to the bed to begin with.
It is disappointing but also good to hear a replacement is likely on the way.
The leadscrew design I kind of understand. The Prusa i3 uses smooth rods to constrain the z-axis. The Prusa XL uses linear rails instead which is superior in every way. Very similar to the Voron Trident design. I'd have thought the linear rails are constrained enough but on the Voron forums you often see similar banding issues with advice to check for bent leadscrews. Maybe it is better when everything is perfect but not as tolerant as the Prusa i3.
Voron hobbyists have experimented with z-wobble eliminators such as magnetic plates separated by ball bearings like the WobbleX. There are some very clever people in the Voron community. But a straight leadscrew is simpler and just works.
RE: Z-banding or Z-wobble, do we have a solution yet?
Yeah, I can understand why it is the way it is. It certainly is simpler. I'm sure it will come down to the numbers. How many units are showing up with bent screws, how many of those are bad enough that the user contacts support for replacement, etc... versus the cost of modifying the design, retooling the production, updating manuals, releasing a free upgrade kit (a-la MMU2 to MMU2s), etc. Time will tell.
Of course, I'm also worried about two things. Will they stall and request additional tests and proofs and stuff, and if/when they do finally send me the replacement, and it is straight, will the problem still be there, and this was a red herring all along? That said, it sure looks to be the most likely cause of the problem, so I remain optimistic.
RE: Z-banding or Z-wobble, do we have a solution yet?
I really think it comes down to QC. If the XL parts were checked before installation like they should be there shouldn't be an issue. Even as they test the units it is very easy to see the wobble. Now some might get damaged during shipping but I believe most with this issue are shipped like this and QC isn't doing the job they should be doing
RE: Z-banding or Z-wobble, do we have a solution yet?
Agreed, but it's not surprising that QC has slipped while they're trying to ramp up production and shipping AFAP, and apparetly, along with building their new commercial printers, too. If/when I get the replacement, I'll give it the ol' roll test on my granite counter before I install it.
RE:
Designers using leadscrews know two things.
1. Nearly all lead screws are bent.
2. Those that aren't will become so over time. So just replacing bent ones with straight one is not a longterm fix as other printer manufacturers discovered. A redesign will be required.
Knowing that you will have bent leadscrews must be taken into account in any kinematic system using them, so that bent lead screws have no effect on the desired precision movement.
There are many design solutions to this.
1. Use decouplers such as VenterMechs or similar.
2. Build a beefier guide system (usually two large diameter smooth rods pet leadscrew) to override any attempt by the water leadscrew to deflect the path of the item bring bent. The reason two smooth rods ate used per lead screw rather than linear rails is that unless perfectly aligned lead screws are more prone to binding.
3. Change the kinematics to a system where bent leadscrews are irrelevant like the 1P Z kinematic system. The first two articles are relevant to the Prusa XL Articles 4 to 6 are relevant to printers using self levelling beds.
4. Use some combination of the above.
I wrote a series of six short and easy to understand technical articles this subject. Links below.
Liam Venter
MFBS Design
https://www.facebook.com/groups/3dprinterdesign/permalink/480055743604492/?mibextid=Nif5oz
https://www.facebook.com/groups/3dprinterdesign/permalink/480598153550251/?mibextid=Nif5oz
https://www.facebook.com/groups/3dprinterdesign/permalink/481353370141396/?mibextid=Nif5oz
https://www.facebook.com/groups/3dprinterdesign/permalink/484373253172741/?mibextid=Nif5oz
https://www.facebook.com/groups/3dprinterdesign/permalink/489832529293480/?mibextid=Nif5oz
https://www.facebook.com/groups/3dprinterdesign/permalink/661786128764785/?mibextid=Nif5oz
RE: Z-banding or Z-wobble, do we have a solution yet?
Hopefully we see some of those things in an update to the machine, 'cus it makes the most sense.
RE: Z-banding or Z-wobble, do we have a solution yet?
I'm calling this one resolved. The replacement left z motor/screw arrived. I installed it, though that was a lot harder than it needed to be, if they would just go ahead and use connectors on the motor wiring AT the motor. But, after feeling like I undid half the assembly and re-did it, the new screw was installed (yes, I checked for straightness first). I've attached before and after photos. While the filament wasn't the same, and I threw away the before models and couldn't take a side by side under the exact same light, the prints look fantastic to me. Of course, photos make even the most minor variations stick out when lit from above, but I honestly don't believe the layering could be any better.
That said, it would behoove Prusa to redesign their bed to decouple the lead screws from lateral motion.
RE: Z-banding or Z-wobble, do we have a solution yet?
I found this thread while researching solutions for my own z-banding issues on my 5-tool XL. I expected a Core-XY to have super silky walls compared to bed slingers, and from day one I've been greatly disappointed with rough walls.
I thought the idea that my lead screws were bent seemed likely, even though my banding isn't really cyclical (i.e. not every 4mm but random). But I was surprised to observe that mine are straight as an arrow, almost zero movement at the caps through the entire vertical travel.
Luckily I did discover my issue. I got a DIY-assembly XL, and knew ahead of time that tightening all the various screws and bolts was critical to a reliably working machine. So I double and triple-checked every screw and bolt during assembly, and even tightened ones that came pre-assembled from Prusa.
Sure enough, I screwed things up. There are two screws at the tops of each trapezoidal nut, and I remember being surprised that they were fairly loose from the factory, so I put some extra twist on them. I now know that this causes heavy banding. I did several z-wobble tower tests, and would adjust those screws tighter/looser during the test, and observed a direct correlation with how rough or smooth the walls came out.
The Prusa documentation on replacing the trapezoidal nuts cautions against overtightening, and indicates they should be gently tightened to prevent damage, but they neglect to mention is can also affect print quality.
Hopefully I didn't damage my trapezoidal nuts due to overtightening, but best I can tell it all seems good now. Those screws are now barely tightened at all, almost to the point of being loose enough to back out on their own, and my vertical walls now print super smooth. If you look at pictures of the trapezoidal nuts in the documentation, you will see that these screws almost are not even needed, they bed frame naturally rests on top of these nuts so gravity holds everything in place, so these screws really seem to prevent the frame from lifting off the nuts (probably when pulling off the magnetic sheet, but perhaps also during z-retracts).
The other part of the solution was making sure my lead screws were sufficiently lubricated with the supplied grease. I don't recall the instructions mentioning to grease them, and I had already done a lot of printing without greasing the lead screws, so in case anyone else fails to grease them like me, know that it definitely makes a difference.
RE: Z-banding or Z-wobble, do we have a solution yet?
Same problem on mk4
RE: Z-banding or Z-wobble, do we have a solution yet?
Yep. It can happen on any machine with a hard link between the lead screw and the bed/carrier. On another forum I posted this to, someone pointed out that a straight screw can eventually bend, so it's not even a sure thing to never happen if you start with a straight one. Decoupling the lateral motion of the z screw and trap nut from the vertical motion of the bed/carrier is an important design consideration that Prusa has yet to include on their printers. There are a lot of ways to do it, they just need to pick one.
RE: Z-banding or Z-wobble, do we have a solution yet?
I've been hit by this too. I don't see it on regular prints, but do see it on vase mode prints. Prusa Support said they were aware of the issue and working on a proper guide for resolving it. In the mean time, they instructed the following procedure:
- On both sides, loosen the 2 screws holding the trapezoidal nut (picture 1), the 4 screws holding the Z-motor (picture 2), and the 2 screws holding the Z-top (picture 3).
- Move the Z-axis all the way up and down, or perform a Z-axis test which does exactly that.
- While the heatbed is down or moving down, equally tighten the trapezoidal nut and the Z-top parts.
- While the heatbed is up or moving up, equally tighten the Z-motors.
(pictures at end of post)
This worked well for me. After on left, before on right:
However, the improvement did not last. A similar print the next day had terrible banding:
It seems to me that there is nothing in the design of the printer to isolate the trapezoidal nuts from the thermal expansion of the print bed subframe. If any deflection of the leadscrews is relieved by the above procedure with the bed subframe warm, surely there will still be deflection of the leadscrews when the bed is cooler. I'm seriously considering adapting WobbleX to suit the XL. Currently, my XL is unusable at least for these vase mode prints.
Picture 1:
Picture 2:
Picture 3:
