RE: Failed Y-calibration of a new CoreOne
One of my reasons for selecting a Prusa for my first printer was that this seemed like a healthy community. Now I can’t help feeling I was right. Already we have a test procedure to identify the problem and a workable hack to solve it.
I agree that an over constrained design must not be wrong, it all comes down to tolerances of the components. If I am not wrong as long as the components are matching there will not be any gap to close, that unfortunately does not mean everything is square, a +0.5 degree error in one bracket would be canceled out by an -0.5 degree error in the other bracket and a frame with a matching 0.5 degree error. Let’s just hope none of us have to debug that tripple error, it is hopefully far fetched, low probability and a bit theoretical.
/Anders
RE: Failed Y-calibration of a new CoreOne
If I am not wrong as long as the components are matching there will not be any gap to close, that unfortunately does not mean everything is square, a +0.5 degree error in one bracket would be canceled out by an -0.5 degree error in the other bracket and a frame with a matching 0.5 degree error. Let’s just hope none of us have to debug that tripple error, it is hopefully far fetched, low probability and a bit theoretical.
I would be pretty confident that the main frame of the CoreXY assembly is accurate. It is stamped out as a single piece, and its 90° angles do not rely on bending but rather on the stamped-out shape. Although I am not sure how the rails are attached to it; maybe there are assembly tolerances there?
In any case, a machinist's square is probably a good tool to have around when assembling a Core One kit -- or when inspecting a ready-built unit. Not one of the DIY-store squares meant for carpenters, but an accurate one for metal work.
RE: Failed Y-calibration of a new CoreOne
As far as I know, the Corexy frame is laser-cut from a steel plate.And should therefore be very accurate.
RE: Failed Y-calibration of a new CoreOne
If I am not wrong as long as the components are matching there will not be any gap to close, that unfortunately does not mean everything is square, a +0.5 degree error in one bracket would be canceled out by an -0.5 degree error in the other bracket and a frame with a matching 0.5 degree error. Let’s just hope none of us have to debug that tripple error, it is hopefully far fetched, low probability and a bit theoretical.
I would be pretty confident that the main frame of the CoreXY assembly is accurate. It is stamped out as a single piece, and its 90° angles do not rely on bending but rather on the stamped-out shape. Although I am not sure how the rails are attached to it; maybe there are assembly tolerances there?
In any case, a machinist's square is probably a good tool to have around when assembling a Core One kit -- or when inspecting a ready-built unit. Not one of the DIY-store squares meant for carpenters, but an accurate one for metal work.
Yes, that would be my assumption as well, until proven otherwise.
/Anders
RE: Failed Y-calibration of a new CoreOne
What are these strips made of? Where did you get them?
Our company is building pipe organs. So I was lucky for having access to many metal strips with different thickness. The stripes I've used are made of brass.
But I think every metal that is hard enough to stand the pressure will do. If you can get strips with a thickness of 0.1mm you can stack them to the needed thickness. And remember - location of the strip is nearly as important as it's thickness. And for not bending the bracket to much (and loosing the angle adjustment you are after) the strip has to be close to the screw.
RE: Failed Y-calibration of a new CoreOne
If I am not wrong as long as the components are matching there will not be any gap to close, that unfortunately does not mean everything is square, a +0.5 degree error in one bracket would be canceled out by an -0.5 degree error in the other bracket and a frame with a matching 0.5 degree error. Let’s just hope none of us have to debug that tripple error, it is hopefully far fetched, low probability and a bit theoretical.
Yes, you are right, the difference of the angles may compensate and your printer may have no gap, but the wrong angles generate static load on the Y-bearings. I think this situation isn't just theoretical. On my printer both brackets were a little bit off. But the right bracket was close to 90° and I think I wouldn't have recognized it if the left bracket wouldn't have been off by so far.
I would suggest to check the precision of the angles on any printer (at least I guess I would do so on our printer) by loosening the belts and also loosen the three screws attaching the bracket to the x carriage on one side. If there is no gap when you move the x-carrage to the front (minimum Y), and there is no gap if both brackets are attached, your printer is fine...
RE: Failed Y-calibration of a new CoreOne
As far as I know, the Corexy frame is laser-cut from a steel plate.And should therefore be very accurate.
Yes, the frame looks to be laser cut and I think it's very accurate. The rails of the y-axis are mounted to the steel frame with precisely machined metal pieces. This looks absolutely fine to me and may be as good as it can be.
RE:
I have already thought about whether it would be a good idea to CNC-mill the part.
The precision would of course be absolute.
But the material seems too thin for that. And weight is also an important factor here.
On the other hand, I think Prusa will optimize the processes and has thought about choosing this solution.
There have also been very few cases and I know some people who are absolutely satisfied with the Core One.
In any case, a solution has been found here that works. If someone has problems in the future, at least they know where to look.
As you have already noticed, this forum is a user-to-user forum without any support from Prusa itself.
You can find that good or bad, but it is what it is.
I personally always use the chat for problems with support.
If necessary, even several times in the same matter if I am not satisfied with the support.
There are always people at the other end. Some are motivated, some are not, just like everywhere else.
RE: Failed Y-calibration of a new CoreOne
I have already thought about whether it would be a good idea to CNC-mill the part.
The precision would of course be absolute
Absolute precision? W?
RE: Failed Y-calibration of a new CoreOne
Yes, milling the exact same mounting brackets would not be a good solution. Most possibly, they would be milled from aluminum and, since the thickness is limited to about 1mm, the resulting parts would not be rigid enough. But I think it's not up to us to find the best future solution.
As a first step, Prusa should recognize and officially acknowledge the issue. I'm sure they have the expertise to finally solve the issue and come up with a good solution.
At the moment Prusa may discuss the situation internally and hopefully they will come back soon to all affected customers with a good proposal to solve the problem.
Yes, that's the route I hope Prusa will take. Let's wait and see 😀..
RE: Failed Y-calibration of a new CoreOne
In any case, a solution has been found here that works. If someone has problems in the future, at least they know where to look.
As you have already noticed, this forum is a user-to-user forum without any support from Prusa itself.
I still think Prusa are missing a big opportunity here. If their tech support team would actively contribute to the forum, the likelihood of finding a solution to problems here would be much higher -- resulting in fewer direct support requests. I would expect that this could actually reduce the overall workload in tech support, once word gets around that the forum is a more useful resource now.
Today, there are too many threads which never reach a solution. Users go round and round asking and waiting for a solution, expressing their frustration, checking dead ends themselves etc. Having Prusa support steer things in the right direction, or at least acknowledge a problem and confirm that they are working on it, would avoid the resulting negativity in many threads.
But who am I writing this to -- I know Prusa won't read it anyway... 🙄
RE: Failed Y-calibration of a new CoreOne
Just received my core one, batch 2, ordered 25th of November. I was prepared for some tinkering after reading the recent topics and i was really happy i didn't have any problems with calibration and the alignment. There is zero gap. Problem is, belt tension is different by about 20hz ny factory to get the angle right. Think I will have to work myself into the mentioned fix.
RE: Failed Y-calibration of a new CoreOne
But who am I writing this to -- I know Prusa won't read it anyway... 🙄
Thanks to @Brian 's suggestion, Prusa support knows about this possible 'cure' for the problem. I've sent the link to the 'solution' post to the Prusa support member that 'handled' my case.
I've received no reaction so far. Absolute silence for 5 days now...
In my opinion, it's not the way you should treat your paying customers!
RE: Failed Y-calibration of a new CoreOne
Just received my core one, batch 2, ordered 25th of November. I was prepared for some tinkering after reading the recent topics and i was really happy i didn't have any problems with calibration and the alignment. There is zero gap. Problem is, belt tension is different by about 20hz ny factory to get the angle right. Think I will have to work myself into the mentioned fix.
If you equalize the belt tension on your printer, I guess your gap will be even worse than 4mm. The gap on our out of the box printer closed at a frequency difference of 13-15Hz.
There seem be different mounting lines/people at Prusa or their instructions have changed. Our printer was adjusted to equal belt tension (resulting in the 2-3mm gap), your printer was adjusted for zero gap. Maybe they have already learned how to hide the problem from customer. At least as long he doesn't get to check the belt's tension...
If you solve this problem by yourself, please inform prusa support about your case.
RE: Failed Y-calibration of a new CoreOne
Our printer was adjusted to equal belt tension (resulting in the 2-3mm gap),
Does your printer pass y calibration with that gap? Because mine fails unless the axis is aligned and the gap is 0, resulting in over a 20Hz different in tensions
RE: Failed Y-calibration of a new CoreOne
Does your printer pass y calibration with that gap? Because mine fails unless the axis is aligned and the gap is 0, resulting in over a 20Hz different in tensions
No, it failed because the travel on the Y-axis was less than expected. The failing Y calibration was the reason why I reached out to the support and started this thread.
Now with zero gap, calibration runs perfectly and print quality is good (but not as great as I did expect).
RE: Failed Y-calibration of a new CoreOne
I'm curious if anyone with this misalignment has taken this bracket entirely off and checked it with a machining square. The assumption is that the because the shim solved the alignment offset, the root cause was a bracket bent a touch past 90 (in MartinF's case). If the bracket can be removed without too much trouble, I would think bending it straight would be preferable to installing shims. This may be more than folks are comfortable doing with on a new printer, but from a metal shop perspective, a fix seems trivial if the part can be removed easily enough.
In my case, since I'll be doing the conversion in 2-3 months, I'll have the pre-assembled parts in hand. I'm also fortunate to have limited access to a machine shop at work (and the welders/machinists who work there) so I'll simply take the brackets there and make any needed adjustments before assembling them. If I didn't have that access, I think I would still try my best to carefully square it up at home in a vice before assembling.
Similarly, regarding the above suggestion to machine a replacement part, once you have the bracket removed for replacement, you might as well nudge the angle into square and reinstall. If the angle is corrected, it seems like it should be a 100% durable fix. No need to machine anything.
I don't have a printer to look at, and the photos above are too close for me to gain full context, so I understand that it may not be as straightforward as it seems.
-J
RE:
On Printables is an adjustable end stop for the XL printer dealing with the same issue. It doesn’t actually fix the gantry being racked (not square) but it does allow the Y calibration to successfully complete with equally tensioned belts.
https://www.printables.com/model/1157603-adjustable-endstops-for-prusa-xl
I guess someone needs to remix it for the Core One.
RE: Failed Y-calibration of a new CoreOne
Doesn't that just mean all prints will be skewed? All x movements will move at an angle to the y axis rather than perpendicular to it
RE: Failed Y-calibration of a new CoreOne
Yep