CoreXY sensorless homing

[thinkyhead]

Suggested by @deltaford in #17838…

Background: Sensorless endstop states are crudely gathered from driver DIAG pins or SPI. Later the bits are processed according to whether axes are moving. The Core direction checks are good, and sensorless homing is enabled correctly for Core kinematic drivers, but in the endstop processing logic only one of the sensorless endstops that could have been triggered gets tested.

Solution: This PR patches the issue by checking all the endstop bits that could have gotten triggered during sensorless homing on different Core arrangements.

[GMagician]

May this be useful also for #17687? (multiendstop check)?

[guruathwal]

Tested on COREXY with TMC2130 SPI with SPI_ENDSTOPS enabled on BTT SKR 1.4 Board.
not much difference noticed in sound/vibration when homing.
With SPI_ENDSTOPS enabled no noticeable difference while homing with or without changes in this PR.
Although these changes are only for homing with SPI_ENDSTOP. SPI_ENDSTOP have a little bit slow response in comparison to sensorless homing without it and without changes in this PR. This may be due to the Software SPI.

Driver settings:
Sensorless homing on X & Y
TMC_DEBUG Enabled
HOLD_MULTIPLIER 0.6
INTERPOLATE true

X_CURRENT       600
X_CURRENT_HOME  200
X_MICROSTEPS     16
X_RSENSE          0.11

Y_CURRENT       600
Y_CURRENT_HOME  200
Y_MICROSTEPS     16
Y_RSENSE          0.11

X_STALL_SENSITIVITY  8
Y_STALL_SENSITIVITY  8

[thinkyhead]

Oops, I didn't quite have the logic right. Please test once more.

[guruathwal]

@thinkyhead with 251ccb9 there is some noticeable reduction in stepper sound while homing.👍

[fiveorders]

Hello, I searched the issues and couldn't find a solution for this particular problem. I am using Octopus Board with UART 2226 drivers. My X Y currents are 800 and my homing currents are 600. Stall sensibility is 25 for X and 43 for Y.

Whatever combination I tried last homing axis regardless Y or X, makes the other axis jump to a random distance from point 0. If X homes last, Y axis jumps between 0.5-4mm away from point 0. The only way to overcome, or let's say reduce the jump distance is reducing the homing speed. Reducing currents results in false triggers, and also reducing the Bump sensitivity also creates false triggers.

If I make the homing speed as 10x60 and 10x60, it is acceptable. but such speeds are killing on a 300x300 build plate.

So I think only one motor is monitored. I am on MArlin 2.9.1 right now, I tried nightly builds as well.

Any recommendation will be great. Many thank you in advance.