4

Setup: Anycubic Kobra Neo (Firmware v1.3.3), Cura 5.2.1

I recently bought an Anycubic Kobra Neo. It works but has issues with auto bed leveling. Specifically, auto bed leveling does not compensate for bed tilt in X and Y directions correctly. After running the auto leveling, the printer consistently prints higher up in the right back corner, compared to the front left corner of the build plate.

The problem does not seem to be related to auto bed leveling being disabled though. While printing, you can see the Z lead screw turning, so the printer definitely tries to compensate for something.

What I tried:

  • Print Z calibration patterns
  • Run auto leveling again using the attached controller
  • Run auto leveling again with the G29 command and prior G28 to home, followed by M500 to save the mesh
  • Add M420 S1 to start G-code (also tried M420 L and both together)
  • Reset EEPROM with M502
  • Use Repetier Host instead of Cura
  • Used different filament

Z Calibration Pattern

Front left Corner of Z Calibration Pattern

Back Right Corner of Z Calibration Pattern

0scar
  • 37,708
  • 12
  • 68
  • 156
thebear8
  • 141
  • 1
  • 3

2 Answers2

2

I've just been through the same problem using the flex bed on the Kobra Max, but it was there with the glass before as well, it was just a little better.

Some problems I did notice:

  • When you remove the plastic cover on the printer head, you will notice the strain gauge is already stressed by some wire pushing on the heat sink. If you tie the cover aside to make sure it can blow on the heat sink without adding stress on it and you do the calibration, it improved the results for me.

    Something bad I did notice is that by power cycling the printer and redoing the nozzle touching step at the beginning of the leveling process, the pressure required to light the red led was not the same at all from time to time. So I guess the strain gauge is probably calibrating on power on but not in a reliable way obviously.

  • When I put the plastic cover back in place on the printer head, I could not have the red LED (on the strain gauge PCB) to stay off so there is something pushing on it while it is tightened so it has to be less sensitive when it's assembled (IMO). I could not yet find what is pushing on it, I would have to sit inside the box to see...

  • After having the leveling done without the plastic cover on the printer's head and having power cycled the printer until the nozzle touching step was the most sensitive before the leveling, all my test squares near the outside of the bed were good but the ones in the middle were much thinner.

  • What seems to be the major problem in this leveling process is that it is doing its leveling process at a fixed temperature (60 °C). So if your bed has to reach 100 °C for ABS, for example, the aluminum heating bed will expand but not the cold steel bracket holding it underneath, so the bed gets higher in the middle.

  • If you succeed to get even tests around your bed and only the ones in the middle are thinner, try to print with the bed at 60 °C (same temp as its leveling) and it will be perfect :). But yes, I know you want the bed at 100 °C for ABS. For this, I have no fix, they will have to update their firmware so we can level at any temp.

Greenonline
  • 6,748
  • 8
  • 40
  • 68
Esenic
  • 21
  • 2
1

The issue seems to be mostly left/right.

The cause is a twisted X extrusion or rail together with a probe which has a X and Y offset compared to the nozzle, therefore it reads wrong values.

See https://github.com/MarlinFirmware/Marlin/issues/22791

If you cannot correct the mesh manually, try to straighten the axis, or add a counterweight behind the axis as heavy as the toolhead.

FarO
  • 4,610
  • 21
  • 39