[BUG] Anycubic Mega S not X homing anymore with 1.5.4

[Alkerion]

Bug Description

My Anycubic Mega S (with TMC2209) doesn't home anymore in X axis with 1.5.4, working well with 1.5.3

Steps to Reproduce

1. Install MEGA_S_TMC_v1.5.4.hex firmware
2. Load FW default
3. Save Eeprom
4. Try to home, Y = OK, Z = OK, X = not OK
5. Restart printer
6. Home again, Y = OK, Z = OK, X = not OK
7. 1. Install MEGA_S_TMC_v1.5.3.hex firmware
8. Load FW default
9. Save Eeprom
10. Try to home, Y = OK, Z = OK, X = OK

Expected behavior: [What you expect to happen]
X axis should be homing

Actual behavior:
X Axis not homing

Additional Information

If I push manually the X axis to the left and lunch a print everything is going fine.
Just the homing seems not working for X axis.

• Provide pictures or links to videos that clearly demonstrate the issue.
• See How Can I Contribute for additional guidelines.

[mithrandir42]

I am experiencing a similar behaviour with a highly modified Mega X (GH Smartshaper (TMC 2225-SA) drivers for XY, stronger steppers, filament motion sensor, inductive Z probe,24V for most stuff, etc.)

YZ homing works, X not. The carriage moves some amount to the left, maybe around 10 cm, then the printer stops with:

2024-10-01 18:01:57,335 - Recv:  T:28.36 /0.00 B:58.65 /0.00 @:0 B@:0
2024-10-01 18:01:58,015 - Recv: echo:Homing Failed
2024-10-01 18:01:58,021 - Recv: Error:Printer halted. kill() called!

After a downgrade to 1.5.3 homing works again.

Also checked the endstop states in home position with M119:

Send: M119
Recv: Reporting endstop status
Recv: x_min: TRIGGERED
Recv: y_min: TRIGGERED
Recv: z_min: open
Recv: z2_min: open
Recv: z_probe: open
Recv: filament: TRIGGERED

With printf debugging I have tracked this down to endstops.cpp:457:Endstops::validate_homing_move. Commenting #define VALIDATE_HOMING_ENDSTOPS prevents the kill(), but the missing movement remains. So there seems to be a problem with left moves while X homing. Tried to reduce HOMING_FEEDRATE_MM_M, but no success.

Does this sound familiar? Maybe this is the same issue. Is VALIDATE_HOMING_ENDSTOPS enabled in your build, do you also see the kill() call?

Any further ideas?

[stklcode]

Sound pretty much like the issue with Z axis in 1.5.3 (#521) which is a bug resolved in Marlin 2.1.2.4 (v1.5.4)

Likely some misconfiguration around the endstops, endstop pins and/or axis direction. We didn’t change those configs between 1.5.3 and 1.5.4 but Marlin upstream has changed quite a bit in the actual logic. Wondering if only TMC flavors are affected 🤔

Are all endstops reported correctly in 1.5.4 when pushed manually?

Do all axes move in the correct direction, e.g. G91 (relative positioning) followed by G0 X10 moves X 10mm away from home?