[BUG] Auto Bed Level - Odd Mesh Behavior

[sinus19]

Did you test the latest bugfix-2.1.x code?

Yes, and the problem still exists.

Bug Description

2.1.3 Bugfix nightly build 10/17/2025
Octopus Pro v1.1
Custom Build Printer

UBL AND BILINEAR is not working correctly.
Nozzle to bed Distance on the left side of the bed is a lot closer than on the right side, Approx. .03mm. ABL is enabled during printing. I can see the Z Axis screw rotate as the head travels.

-changed BL Touch, no effect
-increase BL Touch signal length (in advance config), no effect
-Turn High Speed Probing off/on, no effect
-Enabled # of probes per point to 3 with 1 extra, no effect
-Tried UBL and Bilinear. problem still there Have not tried Linear yet., though i am doubtful.
-Verified Z Axis was mechanically good. No play or binding.
-Im using Stall Guard for Z (i have been for years) and verified position accuracy with calipers. +-0.008mm after homing and re-homing
-Mesh Inset is set to 0
-Probing is approx. 10mm from the edge of the plate all around.
-Print & Probing area is 390x365, part trying to print is 200x250. Well withing the mesh limit
-Changed BL Touch wiring all the way to the board, no effect
-Mesh shows 0.3mm delta with corners within 0.1mm. nothing abnormal.
-Tried different Slicers (Cura & Simplified 3D)
-Mesh is saved in the same slot. I only use 1 slot. I've also reset the EEPROM to make sure there is only 1 slot being used/loaded.
-Z Fade is set to 0

Here is the kicker. After the first layer, it "looks" like ABL starts to work correctly. Either my eyes are deceiving me or something is happening. But the part does finish with an even looking layer heights all around. Closest thing i can think off having this kind of influence is Z Fade, even thought it is set to 0.

FYI:
In order for Stall Guard to work with Z, You have to allow the bed to home above the nozzle approx. 1mm. (XY homes the nozzle outside of the bed to prevent it from crashing.) This puts my prob offset at -0.43mm. This method has worked perfectly fine in previous Marlin versions. Last one being 2.1.2.4.

Bug Timeline

10/17/2025

Expected behavior

Actual behavior

No response

Steps to Reproduce

No response

Version of Marlin Firmware

2.1.3 Bugfix

Printer model

custom

Electronics

No response

LCD/Controller

No response

Other add-ons

No response

Bed Leveling

UBL Bilinear mesh

Your Slicer

None

Host Software

None

Don't forget to include

• A ZIP file containing your Configuration.h and Configuration_adv.h.

Additional information & file uploads

No response

[thinkyhead]

Check out this troubleshooting davice and see if it helps. https://marlinfw.org/docs/basics/troubleshooting.html#one-side-high,-one-side-low

[sinus19]

Check out this troubleshooting davice and see if it helps. https://marlinfw.org/docs/basics/troubleshooting.html#one-side-high,-one-side-low

Thanks for the tip/link. I have verified the wiring harness and all mechanical aspects. The wiring harness is secured right above the head, preventing any type of pulling. I also have 2 linear rails on the gantry that the head mounts to. One in the front of the gantry, another below it. I have physically checked for all screws and even tried to feel for any twisting or loose parts. Everything is solid.

[thinkyhead]

If you actually run the Mechanical Gantry Calibration you will get a more definitive answer. You can also try running a manual mesh procedure and comparing against the result you get with the probe. When you see "one side high, one side low" with the probe, but the mesh you get with the paper-test is not exhibiting the same issue, that is a strong indication that the problem is bad measurement and not a firmware bug. Of course, always run the Z Probe Offset Wizard as a first step and save with M500 to ensure the mesh average is as close to zero as possible, to help mitigate any "leveling fade" issues cause by an off-center mesh.