RE: Failed Y-calibration of a new CoreOne
Yes i have a lot of grease, but mine is dark grey, not white. I remove extra grease with cotton swab.
RE: Failed Y-calibration of a new CoreOne
I also had lots of grease support saw it while helping me with homing issue. Told me to remove it with paper towel, white.
RE: Failed Y-calibration of a new CoreOne
In my case (kit, self-built) there was a completely different problem why the Y-calibration was not correct. I had mounted the pinion on the Y motor incorrectly This is not noticeable when manually calibrating to 85Hz. Everything seems fine, but the belt doesn't mesh with the motor. You can look into the holes on the left and right of the motors from the inside. Just use the cam of your phone using flashlight to take a photo.
This is the step I did incorrectly ignoring the RED warning 😉
https://help.prusa3d.com/de/guide/5-corexy-montage_835522#843327
RE: Failed Y-calibration of a new CoreOne
Still waiting for a response from support; and since I don't like sitting on my hands I decided to test continuity of the X/Y motors with my Fluke 289 and provide everyone with some (hopefully useful) information.
While the values are within tolerance Blue to Red on the Y axis never settled down and was fluctuating between 1.90 and 1.91 which makes me wonder if there is a damaged winding or some other issue.
I'd love to see what reading others have to compare if you have the time, equipment, and knowhow.
I'm replying to my messages to support to keep them updated on everything that I've done and have also sent them a link to this thread if they aren't monitoring it already.
Please feel free to request something from me as I have my printer partially disassembled. I'm more than happy to provide readings off of something if I'm able.
RE: Failed Y-calibration of a new CoreOne
I found two new ways to fail my initial Y and X axis calibrations last night after "finishing" my Core One kit assembly:
* I somehow introduced a twist in the belt that runs through the left motor. I'm amazed I was able to push around the gantry at all. I felt more resistance than I expected to, but just chalked that up to two motors resisting movement as opposed to the one I was used to on my MK3.9.
* The filament sensor cable under routed to the back-right corner under the XY frame had come a bit loose while installing the right panel/filament spool mount. That slack consistently jammed between the bearing and frame when the X axis was pushed toward the back. I pulled out the slack and fed it into the back plate - will open that up and dress it in later today.
* My Y axis was slightly out of square before any belt tension was set. A little brute force was all it took to put a subtle bend into those brackets and get both sides to hit the front simultaneously with no play. I'm surprised how little force it took - I hope it maintains square over time and doesn't move itself out of alignment.
Printer passed calibration with flying colors after taking off all panels and a couple steel profiles.
RE: Failed Y-calibration of a new CoreOne
I just finished building my kit and ran into a y calibration test issue. I noticed the right side of the gantry was not hitting the frame like it was before I tuned the belts.
After some playing around using this article, realized I needed to tighten the right tensioner in order for the right side of the gantry to hit the frame like the left side. The issue is in order to tune the belts I have to loosen it and tighten the left one and then we are back with the issue. After checking the belt mounts of the nextruder cariage with prusa support they asked how close to 85hz could I get to before the right gantry started pulling off and if at that point, would the test pass.
It passed! They are at 85 on the left and 98 on the right. They said this is alright to operate at as long as we aren't seeing layer shifts or other belt related issues. Succeeded the fast benchy off the usb so far and printing a calibration cube as I type so here is hopping things work out. Hope this helps someone!
RE: Failed Y-calibration of a new CoreOne
Found a solution that works for me:
My theory is that the gaps are causing the y-axis to have an inconsistent collision with the endstop. The problem is not the squareness of the x-axis but the position of the endstop.
I checked for squareness against a feature at the back of the frame. Given the not great measurement it is within +-0.5mm. certainly not as large as the gap. This was done with loose belts.
I used some thin Aluminium sheet to make little packing blocks for the endstop with the gap. Then stuck it the printer with double sided tape. Now the y-axis hits both endstops at about hte same time. After this the printer ran through the calibration ok.
My endstop gap was on the left side instead of right. I don't know if that makes a difference.
Ran a quick circle test reference and squareness seems to be ok. More testing to be done.
RE: Failed Y-calibration of a new CoreOne
Hmm.
My callibration had been spotless (once I got the filament sensor cable moved put of the way) but I've got about 7hz of difference on the belts to get the xy square. I'm wondering if it is worth trying to square it up with shims.
The core one prints so well and with so little problems that I feel comfortable chasing ideal prints rather than just being happy it prints.
I'm reprinting the parts for my old mini so I'm aiming for dimensional accuracy. Got a califlower printing so that'll give me a better idea.
RE: Failed Y-calibration of a new CoreOne
I had the same problem but solved it.
Proplem was the wrong installation of the Pulley on the X Axis. The Pulley has to install upside down. This was my mistake.
After i changed direction of the Pulley the Calibration of X and Y works fine.
RE: Failed Y-calibration of a new CoreOne
well, my machine is back at Prusa. It is there for a gamut of different reasons that I doubt will be addressed. I am still baffled by the amount of people putting up with adding time and effort to a product that should have come finished from the start.
I paid for a finished product, and did not expect to be bending parts and or adding home made shims with wire. This feels like a garage hack that belongs in the $100 dollar printers from China and not a product of this price and reputation.
I am disasppointed to say the least and this being my first Prusa, it feels like I have invested my money and my time poorly.
To be determined once I hear back from Prusa.
RE: Failed Y-calibration of a new CoreOne
I am getting a failure of the Y-Axis before it reaches the back. I have disable the motors to check for any resistance along the 10mm rails and I get an odd tension at about 50-75% down the rail then it goes away. The resistance seems to cause the Y calibration to failureIMG_3190
RE: Failed Y-calibration of a new CoreOne
TLDR:
I think the trick is to disconnect the tension pulleys completely and then flex the gantry square. The left and right gantry brackets are mild steel. With the belts hanging loose, you can easily flex the gantry to bend things square. While you have the pulleys disconnected, move the gantry around and back to the front stops several times and make any necessary adjustments to your wire harness and bowden tube. I found that even a little tension on the harness would pull the gantry out of alignment. Then reattach the tension pulleys, and you will probably find that tensioning the belts is a snap, and everything stays square.
My Gantry Issues:
I built my kit 2 weeks ago but didn't have time to start calibrating it till last night. My x and y alignment on the gantry was a mess, the left side when facing the printer was 3+ mm off from the right side. I tried adjusting the belt tension to align it, and it never solved the issue, even with a 20+ Hz difference between the x and y belts. I also found that flexing the gantry with the belts loosened all the way didn't really accomplish anything either.
At this point, I decided to start disassembling things by first disconnecting the x beam from the bracket on the left side when facing the front of the printer. What I discovered is that the left and right brackets are easily bendable with my fingers. So if flexing the gantry with the belts loosened all the way is not changing anything, then even in the loosened tension position, the belts are preventing someone from applying enough flex to an assembled gantry.
I then reassembled the gantry and removed the tension pulleys completely. I could then easily flex the gantry enough to square it up. It took me about 2 minutes to get it squared up. A couple of tweaks to the wire harness and bowden tube, and I could run the gantry back and forth several times, and it stayed square.
I then reattached the tension pulleys, slid the gantry to the rear of the printer, centered the print head, and tensioned the belts. This time, setting belt tension was a breeze. Like seriously, in under a minute, I had both belts at ~85 Hz and the gantry was still square at the front stops. My test prints are now consistent, close to perfect and without VFAs.
I hope this helps someone.
RE: Failed Y-calibration of a new CoreOne
My kit arrived in an almost flawless package (the plastic handles were destroyed, but everything else was fine. FedEx to northern Germany).
Built it last weekend, extremely careful, reading all the comments + threads and everything seemed fine, apart from the homing end-stop of the Y-Axis being ~1 mm off.
As we all know, this results in some failed Y-Axis calibration, too long homing-bumping etc.
Luckily, after reading all these pages and being prepared for quite some disassembling, bending, shimming etc. being needed, the solution was very simple:
Tensioning the belts from perfect 85/85 Hz to 81/86 Hz.
This seems to be perfectly fine, according to your reports from the Prusa-Support.
Not much for others to benefit from, but I thought I'd share my experience.
Now I might have to tackle the annoying resonances at the end of prints, when the bed moves down a lot.
A thought about the belt Hz:
Since you can't really dodge hitting the resonance frequency every now and then, wouldn't slightly different tunings of the two belts be better anyway, since they won't resonate at the same time?
In a book about vehicle dynamics (forgot the name), I've read to NOT have the same natural frequencies in the front and back, to keep the car from suffering from "the trampoline effect" (disastrous resonance).
RE: Failed Y-calibration of a new CoreOne
A thought about the belt Hz:
Since you can't really dodge hitting the resonance frequency every now and then, wouldn't slightly different tunings of the two belts be better anyway, since they won't resonate at the same time?
In a book about vehicle dynamics (forgot the name), I've read to NOT have the same natural frequencies in the front and back, to keep the car from suffering from "the trampoline effect" (disastrous resonance).
I don't think there is much to be gained here. In actual operation, the belt resonance frequencies will change all the time, since the free length of the belts (between the idlers and the print head, or the idlers and the X gantry) will vary.
As a side note, I would expect that -- if belt oscillations affect print quality at all -- it would be longitudinal oscillations (stretching mode). Their resonance frequencies should be much higher than the ones measured for transverse oscillations which are used as a measure for belt tension during the adjustment process.
RE: Failed Y-calibration of a new CoreOne
I have a feeling that getting the machine square is a losing battle. Best to measure and compensate for the lack of squareness.
The OpenPnP - pick n place project has the same problem. 0.25mm off square over the length of a pickup tape or PCB causes chaos.
The solution is to use an angle reference or length reference to find square. The PnP machines have the benefit of a down facing camera. We don't have that but we do have a force sensor in the nozzle. Maybe something can be done with that.
RE: Failed Y-calibration of a new CoreOne
Maybe someone has seen this before. There is a simple solution for this problem.
- First loosen both belts so that they do not affect the movement xy.
- On the side where the end touches the first, insert the key and on the other side push a little.
- Try the xy movement again.
- If both stops are the same, you can tighten the belts.
In my case, it solved everything and the printer works without any problems.
RE: Failed Y-calibration of a new CoreOne
I have a feeling that getting the machine square is a losing battle. Best to measure and compensate for the lack of squareness.
The OpenPnP - pick n place project has the same problem. 0.25mm off square over the length of a pickup tape or PCB causes chaos.
The solution is to use an angle reference or length reference to find square. The PnP machines have the benefit of a down facing camera. We don't have that but we do have a force sensor in the nozzle. Maybe something can be done with that.
Getting it "square enough" for the Y calibration to pass has certainly be achieved by many users. Getting it "perfectly square" to obtain dimensionally accurate prints will depend on your tolerance limits; I would assume that a few 1/10 mm across the width of the print bed are probably the limit that can realistically be achieved (and expected to be stable over time).
The idea of an optical reference and a downward-facing camera (to scan the reference and calibrate position/skew compensation in software) is exactly what Bambu Lab have implemented in the H2D. The camera is always installed, but they charge extra for the optical reference sheet.
If all you want to compensate is global skew, you could easily measure the skew angle via a test print and derive a correction factor. But to my knowledge neither the printer firmware nor Prusa Slicer can incorporate skew correction factors -- or did I overlook something?
RE: Failed Y-calibration of a new CoreOne
Good point. A test print and measurements from a set of calipers is cheap and effective. Then we do need somewhere to correct the skew.
Didn't know about the Bambu down facing camera. The OpenPnP project does for free what Bambu is charging for. No surprise there.
Have no yet found a satisfying test item for skew.
RE:
Maybe someone has seen this before. There is a simple solution for this problem.
- First loosen both belts so that they do not affect the movement xy.
- On the side where the end touches the first, insert the key and on the other side push a little.
- Try the xy movement again.
- If both stops are the same, you can tighten the belts.
In my case, it solved everything and the printer works without any problems.
Nice and thanks for sharing!
I'd be interested in your final belt tunings though - are they both at the same 85 Hz or did you end up with synchronous stops, but something like my 81/86 Hz?
---------------------------
And I'll have to take back my "it's perfect now" statement.. When using "avoid crossing perimeters", the 7000 mm/s² for "Travel" lead to quite massive layer shifts of 5-20 mm, when the extruder changes from a 350 mm/s in a straight line into a curve, to dodge something like a center-hole perimeter.
I got rid of this by limiting all acceleration settings to 3000 mm/s², which isn't great, but not terrible either.
I also tried to tighten both belts (85/89 Hz), which somehow resulted in failed Y-Calibrations during homing. Loosened back to 81/86 Hz and homing is quick again.
Gonna play around with different solutions next week..
RE: Failed Y-calibration of a new CoreOne
Maybe this was mentioned already but: be aware that the ‘avoid crossing perimeters’ is directly implicated in a firmware bug for causing layer shifts eg https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4573
Maybe someone has seen this before. There is a simple solution for this problem.
- First loosen both belts so that they do not affect the movement xy.
- On the side where the end touches the first, insert the key and on the other side push a little.
- Try the xy movement again.
- If both stops are the same, you can tighten the belts.
In my case, it solved everything and the printer works without any problems.
Nice and thanks for sharing!
I'd be interested in your final belt tunings though - are they both at the same 85 Hz or did you end up with synchronous stops, but something like my 81/86 Hz?---------------------------
And I'll have to take back my "it's perfect now" statement.. When using "avoid crossing perimeters", the 7000 mm/s² for "Travel" lead to quite massive layer shifts of 5-20 mm, when the extruder changes from a 350 mm/s in a straight line into a curve, to dodge something like a center-hole perimeter.
I got rid of this by limiting all acceleration settings to 3000 mm/s², which isn't great, but not terrible either.I also tried to tighten both belts (85/89 Hz), which somehow resulted in failed Y-Calibrations during homing. Loosened back to 81/86 Hz and homing is quick again.
Gonna play around with different solutions next week..