Sensorless Homing on X-Axis doesn’t trigger (Fly-E3-Pro-V3, RRF 3.5.4 STM32)

[kevgeiger2]

Hi everyone,
I’m trying to get sensorless homing on the X-axis working on my Core-R-Theta printer (using a Fly-E3-Pro-V3 board).
Firmware: RepRapFirmware 3.5.4 for STM32F4 (e3pro3_f4)
M115
FIRMWARE_NAME: RepRapFirmware for STM32F4 based Boards (e3pro3_f4)
FIRMWARE_VERSION: 3.5.4

Board: Fly-E3-Pro-V3

X motor → driver P1, TMC2209

Onboard DIAG jumper on X is installed (internally routes DIAG to the X-endstop input)

No physical X-endstop switch connected – the X-STOP header is empty

Using M574 X2 S3 for sensorless endstop

Verified with M119 that it reports
→ “not stopped” at rest
→ “at max stop” after a homing command

The issue:
Whenever I run a homing move (G1 H1 X100 F4000 or G1 S1 X100 F4000)
The motor does not move at all
Firmware immediately reports “at max stop”
No stall is detected
Manual movement (G1 S2 X±5) works perfectly
Same configuration for Z or B axis works fine
It looks like the firmware treats the X-endstop as always triggered, even though electrically it’s open and the DIAG jumper is correct.

What I already tried:
M574 X2 S3, M574 X2 S3 J1, and M950 J1 C"^xstop" / "!^xstop"
Reversing direction: M569 P1 S0/S1
Adjusting stall sensitivity:
M915 X H400-600 R0 F0 S-10 … S-80
Changing motor current and sensitivity (M906/M913)
Verified that M119 changes state only after fake homing
Driver diagnostics (M122) show no errors
Homing command simply does nothing when H1/S1 is used

Diagnostic Output:
Driver 1: TMC2209, standstill, no error
Supply voltage stable
No step errors or stalls detected

Is this a known issue with RRF 3.5.x on STM32 (Fly-E3-Pro-V3) when using sensorless homing on X?
Does the DIAG → endstop mapping differ from Duet/LPC builds?
Has anyone achieved a working M574 X2 S3 setup on this board?

; X-axis configuration
M569 P1 S1 D3 ; X driver
M584 X1 Z2 B0 C4 E3 S0
M574 X2 S3 ; X sensorless high-end
M915 X H500 R0 F0 S-50 ; sensitivity
M906 X1200
M913 X80