[BUG] xyz callibration fails

[Fojtik]

Printer type - MK3S+
Printer firmware version - 3_13_2
MMU upgrade - ----
MMU upgrade firmware version - ------

SD card or USB/Octoprint N/A

Describe the bug
After upgrading firmware the printer automatically executes xyz callibration and it fails with a message that is not described here: https://help.prusa3d.com/article/xyz-calibration-error-messages-mk3-mk3s-mk3s_112340

To Reproduce Run xyz callibration

Expected behavior xyz callibration passes

G-code N/A

Video https://storage.googleapis.com/prusa3d-content-prod-14e8-wordpress-forum-prod/2024/01/186c850d-20240116_193701.jpg

I did not fiddle with INVERT_Y_DIR, I do not know what is it. May be that this is a duplicity of: #4570 May be not.

[3d-gussner]

@Fojtik have you tried to contact Prusa Chat support? If that issue isn't solved then please update your github issue here.

[Fojtik]

Yes, I have contacted Průša support yesterday, no response yet.
I have standard stock Průša printer without any modifications and it asks me to upgrade firmware before every print. But this firmware is somehow deffective.

[sarusani]

I just checked the firmware and the error is thrown when there's a problem with the Z axis. So it's not related to the INVERT_Y_DIR problem.

You mentioned that it's working with older firmware versions. Can I ask you which is the latest one that works for you?

[Fojtik]

I have these 3 firmwares downloaded:
28.06.2023 13:20 6 659 171 prusa3d_fw_3_12_2_MK3S.zip
06.09.2023 00:04 6 694 867 prusa3d_fw_3_13_1_MK3S.zip
16.01.2024 19:20 6 694 766 prusa3d_fw_3_13_2_MK3S.zip

I am pretty sure that 12.2 is OK. And may be that 13.1. is also OK. Is it safe to flash old firmware? I will not test 13.0. because it is reported to destroy plate.

[sarusani]

Yes, downgrading the firmware is no problem.
Would be great if you could try 13.1. If that works I can have a look at the differences between the two releases.

[Fojtik]

This looks much more weird now. I cannot do callibrations even for 3.12.2. I do not know why.

The printer still can measure a level over sheet correctly before printing. Selftest passed OK.

I am lost now, I do not know why xyz callibration fails and printer still prints. Without forcing to xyz callibrate after firmware upgrade I would not note anything bad.

I am waiting to response from Průša. May be that I am facing some new deffect of printer.

[Fojtik]

Look at this video:
http://78.108.103.11/~fojtik/MK3S/SelhaniKalibraceXYZ.mp4

2 droplets of oil to z axis thread rods fixed a problem. I have blamed a new firmware, but the problem was different. But still the error message is very confusing and printer behavior even more suspicious.
During normal printing the printer uses more power to stepper motors and this problem is hidden.

[Fojtik]

But there is still one problem of firmware remaining.
When xyz callibration fails from whatever reason, the stepper motor power is not set to usual value.
See this video: http://78.108.103.11/~fojtik/MK3S/20240117_174438.mp4

I cannot reproduce this now ;).

[3d-gussner]

@Fojtik Please check your bearings and POM nuts. I had once on my X axis similar issues. It turned out that one of the dust-sleeves of a bearing was broken and started to bind more and more in one direction. Removing the sleeve confirmed that it was causing the issue and I replaced the bearing bit later.

Also verify that your Z POM nuts do not bind.
Do you have stock Z-TOP parts or do you have some lead screw constraining one? The lead screw should NOT be constrained at the top!

Thanks for the videos, but the LCD is way too bright to see anything.

But there is still one problem of firmware remaining.
When xyz callibration fails from whatever reason, the stepper motor power is not set to usual value.
See this video: http://78.108.103.11/~fojtik/MK3S/20240117_174438.mp4

I cannot reproduce this now ;).

Hard to look for a bug if it can't be reproduced.

[Fojtik]

May be that you do not understand me, after 2 droplets of oil xyz test passes OK without any problem. I have stock parts including sleeves one year old, 25days of continuous print.

The bug could be observed when small friction occurs. Firmware does not return z axis stepper motor's usual current and the current remains reduced.
Even when printer works correctly, the small friction could be simulated.

[Fojtik]

I has been confirmed deffect of my printer. The printer is still under warranty. This demonstrates printer problem:
https://storage.googleapis.com/prusa3d-content-prod-14e8-wordpress-forum-prod/2024/02/68e29652-nutfriction.mp4
Manufacturer does not recommed using oil.

Undil I fix the issue, I cannot simulate how small friction influences callibration.

[Prusa-Support]

I has been confirmed deffect of my printer. The printer is still under warranty. This demonstrates printer problem: https://storage.googleapis.com/prusa3d-content-prod-14e8-wordpress-forum-prod/2024/02/68e29652-nutfriction.mp4 Manufacturer does not recommend using oil.

Thanks for the evidence. That amount of friction is excessive for the trapezoidal nuts to work correctly.
The trapezoidal nuts must spin pretty much frictionless exactly like the first one you spin in the the video.

Here at step 3, the user is requested to remove the trapezoidal nuts. In extremely rare cases over the past many years, there have been cases where the user reported a substantial difference of friction between trapezoidal nuts while removing the trapezoidal nuts from the lead screws as per the assembly step. In those cases, we assumed it was the case of a factory faulty nut and replaced it right away.

In other specific cases, the trapezoidal nuts could get damaged or degrade due to incorrect manipulation.
Here are just a few examples.

• Incorrect assembly/insertion of the X-axis on the lead screws.
• Lubrication. Lubricants are nearly forbidden for this type of trapezoid because they could deteriorate the specific blend of POM material and hinder their "natural" self-lubricating properties - as well as attract debris and build up friction.
• Too long or too short X-axis. This may happen if the smooth rods are of the wrong size, or not inserted all the way inside the X-end printed parts.
• Crooked X-axis. This may happen if a lead screw is skewed or doesn't spin correctly. One of the Z motors could have been incorrectly connected, or one of the black-printed motor covers (or both) is tightened all the way against the motor body, and rubs against the motor, preventing the correct rotation of the lead screw.

I'm afraid this can't find a solution at the firmware level. The problem is apparently on the hardware side.

Michele Moramarco
Prusa Research

[Prusa-Support]

I can't exclude that the problem is related to issue #4476 so please give it a try with the new firmware release 3.13.3 and let us know if that fixed the problem.

Michele Moramarco
Prusa Research

[Fojtik]

I can't exclude that the problem is related to issue #4476 so please give it a try with the new firmware release 3.13.3 and let us know if that fixed the problem.

I am not facing this problem anymore, because Prusa sent me 2 new nuts as warranty claim. It is very interesting that all 4 nuts have different friction. I have originally owned the highest and the lowest friction nuts.

It could be possible to simulate friction with some tool like rubber to add friction repeatable.

The main problem reported here is that firmware is not solving excessive friction properly. To debug this it is neccessary to make this situation. Also there should be some steel plate (without PEI), because callibration failure could damage it.

[Fojtik]

I have attempted to reproduce a problem with additional friction with fingers and everything seems to work fine.

The friction of deffective nut must be very high to cause z axis movement to fail.

[Prusa-Support]

Thank you for your feedback. Indeed the cases matched pretty well with a hardware problem.
The indications above - #4571 (comment) - should possibly help avoid problems with the trapezoidal nuts.

Michele Moramarco
Prusa Research

[Fojtik]

But there was also a firmware problem. After callibration stuck, the stepper mottor did not recover full pover. These rare situations are not tested too much. Currently it seems that printer is doing full reset.

One thing is a nus stuck (but printer printed well to the callibration point - the increased friction did not prevent proper functionality) and second thing is wrongly handled situation after such a stuck.

Crooked X-axis. This may happen if a lead screw is skewed or doesn't spin correctly. One of the Z motors could have been incorrectly connected, or one of [the black-printed motor covers](https://cdn.help.prusa3d.com/wp-content/uploads/a204d61ff1214502-800x600.jpg)
This was a nonsens for my case. I has been written that printer normally printed without any problem till callibration occurred. I.e. it was fully functional without any visible problem.

[Prusa-Support]

I'm sorry, I'm having a hard time trying to understand at the moment and I can't exclude a hardware/assembly problem or something to do with issue #4019.

If the X-axis looks crooked, a Calibrate Z would do the trick.
If the occurrence is recurrent, we may actually be looking at a hardware/assembly problem, ideally to be addressed via individual troubleshooting with our Customer Support.

In any case, if during the troubleshooting a firmware problem can be confirmed, it may not be related to this and the linked issues. In other words, maybe it would deserve a separate and more specific issue.

Michele Moramarco
Prusa Research

[Fojtik]

I have no problems with my printer now. Once the callibration finishes & passes, the callibration code is not executed again.

I still suspect that callibration does not cover all possible situations (failures), But it never mind.