Problem: Homing sometimes doesn't do the final pulloff on XY axes

[HolyBlackCat]

Wiki Search Terms

homing

Controller Board

MKS DLC32 MAX V1.2

Machine Description

3-axis, no gantry (one stepper per axis), DM542 drivers.

Mechanical normally-closed limit switches, wired in series per axis.

Y limit switches:

X limit switches:

Input Circuits

I don't have schematics for the board.

The board provides 3 pins per axis limit switch, labeled: S, GND, 5V. I connect two NC switches in series to S and GND, ignoring the 5V pin.

Configuration file

board: MKC DLC32 Max
name: MKC DLC32 Max Default
meta:
stepping:
  engine: Timed
  idle_ms: 255
  pulse_us: 4
  dir_delay_us: 0
  disable_delay_us: 0
  segments: 12

spi:
  miso_pin: gpio.12
  mosi_pin: gpio.14
  sck_pin: gpio.13


kinematics:
  Cartesian:

axes:
  shared_stepper_disable_pin: gpio.8
  homing_runs: 2
  x:
    steps_per_mm: 320
    max_rate_mm_per_min: 2000
    acceleration_mm_per_sec2: 75
    max_travel_mm: 270
    soft_limits: true
    motor0:
      limit_all_pin: gpio.41:high
      hard_limits: true
      pulloff_mm: 3
      standard_stepper:
        step_pin: gpio.16
        direction_pin: gpio.15
    homing:
      cycle: 2
      allow_single_axis: true
      positive_direction: false
      mpos_mm: -135
      seek_mm_per_min: 500.000000
      feed_mm_per_min: 100.000000
      settle_ms: 250

  y:
    steps_per_mm: 320
    max_rate_mm_per_min: 2000
    acceleration_mm_per_sec2: 75
    max_travel_mm: 200
    soft_limits: true
    motor0:
      limit_all_pin: gpio.40:high
      hard_limits: true
      pulloff_mm: 3
      standard_stepper:
        step_pin: gpio.7
        direction_pin: gpio.6:low
    homing:
      cycle: 2
      allow_single_axis: true
      positive_direction: true
      mpos_mm: 100
      seek_mm_per_min: 500.000000
      feed_mm_per_min: 100.000000
      settle_ms: 250

  z:
    steps_per_mm: 320
    max_rate_mm_per_min: 1000
    acceleration_mm_per_sec2: 25
    max_travel_mm: 180
    soft_limits: true
    motor0:
      limit_all_pin: gpio.39:high
      hard_limits: true
      pulloff_mm: 3
      standard_stepper:
        step_pin: gpio.5
        direction_pin: gpio.4
    homing:
      cycle: 1
      allow_single_axis: true
      positive_direction: true
      mpos_mm: 0
      seek_mm_per_min: 500.000000
      feed_mm_per_min: 100.000000
      settle_ms: 250



control:
  safety_door_pin: NO_PIN
  reset_pin: NO_PIN
  feed_hold_pin: NO_PIN
  cycle_start_pin: NO_PIN
  macro0_pin: NO_PIN
  macro1_pin: NO_PIN
  macro2_pin: NO_PIN
  macro3_pin: NO_PIN
  fault_pin: NO_PIN
  estop_pin: NO_PIN
  homing_button_pin: NO_PIN

coolant:
  flood_pin: gpio.1
  delay_ms: 0


macros:
  startup_line0:
  startup_line1:
  Macro0:
  Macro1:
  Macro2:
  Macro3:
  after_homing:
  after_reset:
  after_unlock:

pwm:
  pwm_hz: 5000
  direction_pin: NO_PIN
  output_pin: gpio.2
  enable_pin: gpio.42
  disable_with_s0: false
  s0_with_disable: false
  spinup_ms: 0
  spindown_ms: 0
  tool_num: 0
  speed_map: 0=0.000% 24000=100%
  off_on_alarm: false


start:
  must_home: false
  deactivate_parking: false
  check_limits: true

arc_tolerance_mm: 0.002000
junction_deviation_mm: 0.010000
verbose_errors: true
report_inches: false
enable_parking_override_control: false
use_line_numbers: false
planner_blocks: 16

Startup Messages

[MSG:INFO: FluidNC v4.0.3 https://github.com/bdring/FluidNC]
[MSG:INFO: Local filesystem is /littlefs]
[MSG:INFO: Configuration file:config.yaml]
[MSG:INFO: Machine MKC DLC32 Max Default]
[MSG:INFO: Board MKC DLC32 Max]
[MSG:INFO: SPI SCK:gpio.13 MOSI:gpio.14 MISO:gpio.12]
[MSG:INFO: Stepping:Timed Pulse:4us Dsbl Delay:0us Dir Delay:0us Idle Delay:255ms]
[MSG:INFO: Axis count 3]
[MSG:INFO: Shared stepper disable gpio.8]
[MSG:INFO: Axis X (-135.000,135.000)]
[MSG:INFO:   Motor0]
[MSG:INFO:     standard_stepper Step:gpio.16 Dir:gpio.15 Disable:NO_PIN]
[MSG:INFO:  X All Limit gpio.41]
[MSG:INFO: Axis Y (-100.000,100.000)]
[MSG:INFO:   Motor0]
[MSG:INFO:     standard_stepper Step:gpio.7 Dir:gpio.6:low Disable:NO_PIN]

User Interface Software

WebUI

What happened?

About 1/4 of the time, when homing the machine, it will not do the final pulloff on the XY axes (after hitting the switches the second time), and instead get stuck in homing mode, spamming the same message in the logs:

<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>

So it homes Z correctly, then starts to home XY: touches the switches once, pulls off, then touches them again and stops immediately, instead of trying to pull off.

This only reproduces (sometimes) when the debug logs are disabled. So it somtimes reproduces with $Message/Level=Info, but never with $Message/Level=Debug, so I suspect a software issue. An uninitialized variable somewhere perhaps?

GCode File

No response

Other Information

Here are the logs when the issue happens ($Message/Level=Info)

$H
<Home|MPos:-135.000,100.000,0.031|FS:240,0>
<Home|MPos:-135.000,100.000,0.831|FS:500,0>
<Home|MPos:-135.000,100.000,2.475|FS:500,0>
<Home|MPos:-135.000,100.000,3.000|FS:100,0|Pn:Z>
<Home|MPos:-135.000,100.000,2.666|FS:100,0|Pn:Z>
<Home|MPos:-135.000,100.000,2.494|FS:100,0>
<Home|MPos:-135.000,100.000,2.150|FS:100,0>
<Home|MPos:-135.000,100.000,1.806|FS:100,0>
<Home|MPos:-135.000,100.000,1.462|FS:100,0|WCO:13.766,-70.078,-170.859>
<Home|MPos:-135.000,100.000,1.122|FS:100,0|Ov:100,100,100>
<Home|MPos:-135.000,100.000,0.778|FS:100,0>
<Home|MPos:-135.000,100.000,0.434|FS:100,0>
<Home|MPos:-135.000,100.000,0.091|FS:0,0>
<Home|MPos:-135.000,100.000,0.178|FS:100,0>
<Home|MPos:-135.000,100.000,0.522|FS:100,0>
<Home|MPos:-135.000,100.000,0.866|FS:100,0>
<Home|MPos:-135.000,100.000,1.209|FS:100,0>
<Home|MPos:-135.000,100.000,1.553|FS:100,0>
<Home|MPos:-135.000,100.000,1.894|FS:100,0>
<Home|MPos:-135.000,100.000,2.237|FS:100,0>
<Home|MPos:-135.000,100.000,2.581|FS:100,0>
<Home|MPos:-135.000,100.000,2.925|FS:100,0>
<Home|MPos:-135.000,100.000,3.006|FS:100,0|Pn:Z>
<Home|MPos:-135.000,100.000,2.713|FS:100,0|Pn:Z>
<Home|MPos:-135.000,100.000,2.537|FS:100,0>
<Home|MPos:-135.000,100.000,2.194|FS:100,0>
<Home|MPos:-135.000,100.000,1.847|FS:100,0>
<Home|MPos:-135.000,100.000,1.503|FS:100,0>
<Home|MPos:-135.000,100.000,1.163|FS:100,0>
<Home|MPos:-135.000,100.000,0.819|FS:100,0|Ov:100,100,100>
<Home|MPos:-135.000,100.000,0.475|FS:100,0>
<Home|MPos:-135.000,100.000,0.131|FS:0,0>
[MSG:Homed:Z]
<Home|MPos:-135.359,100.359,0.000|FS:707,0>
<Home|MPos:-137.069,102.069,0.000|FS:707,0>
<Home|MPos:-137.962,102.963,0.000|FS:141,0|Pn:XY>
<Home|MPos:-137.619,102.619,0.000|FS:141,0|Pn:Y>
<Home|MPos:-137.531,102.531,0.000|FS:141,0>
<Home|MPos:-137.188,102.188,0.000|FS:141,0>
<Home|MPos:-136.844,101.844,0.000|FS:141,0|WCO:13.766,-70.078,-170.859>
<Home|MPos:-136.503,101.503,0.000|FS:141,0>
<Home|MPos:-136.159,101.159,0.000|FS:141,0>
<Home|MPos:-135.816,100.816,0.000|FS:141,0>
<Home|MPos:-135.472,100.472,0.000|FS:141,0>
<Home|MPos:-135.128,100.128,0.000|FS:0,0>
<Home|MPos:-135.203,100.203,0.000|FS:141,0>
<Home|MPos:-135.544,100.544,0.000|FS:141,0>
<Home|MPos:-135.887,100.887,0.000|FS:141,0>
<Home|MPos:-136.231,101.231,0.000|FS:141,0>
<Home|MPos:-136.575,101.575,0.000|FS:141,0>
<Home|MPos:-136.919,101.919,0.000|FS:141,0|Ov:100,100,100>
<Home|MPos:-137.262,102.262,0.000|FS:141,0>
<Home|MPos:-137.606,102.606,0.000|FS:141,0>
<Home|MPos:-137.947,102.947,0.000|FS:141,0>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY|WCO:13.766,-70.078,-170.859>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY|Ov:100,100,100>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
<Home|MPos:-138.003,103.003,0.000|FS:141,0|Pn:XY>
... continues infinitely

And here are logs when it doesn't happen, with $Message/Level=Debug:

$H
[MSG:DBG: Homing Cycle Z]
[MSG:DBG: Homing nextPhase FastApproach]
[MSG:DBG: Starting from -135.000,100.000,0.000]
[MSG:DBG: Planned move to -135.000,100.000,198.000 @ 500.000]
<Home|MPos:-135.000,100.000,0.034|FS:240,0>
<Home|MPos:-135.000,100.000,0.859|FS:500,0>
<Home|MPos:-135.000,100.000,2.513|FS:500,0>
[MSG:DBG:  Z All Limit 1]
[MSG:DBG: Homing limited Z]
[MSG:DBG: Homing nextPhase Pulloff0]
[MSG:DBG: Starting from -135.000,100.000,3.013]
[MSG:DBG: Planned move to -135.000,100.000,0.013 @ 100.000]
<Home|MPos:-135.000,100.000,2.984|FS:100,0|Pn:Z>
<Home|MPos:-135.000,100.000,2.644|FS:100,0|Pn:Z>
[MSG:DBG:  Z All Limit 0]
<Home|MPos:-135.000,100.000,2.472|FS:100,0>
<Home|MPos:-135.000,100.000,2.128|FS:100,0>
<Home|MPos:-135.000,100.000,1.784|FS:100,0>
<Home|MPos:-135.000,100.000,1.441|FS:100,0>
<Home|MPos:-135.000,100.000,1.097|FS:100,0|WCO:13.766,-70.078,-170.859>
<Home|MPos:-135.000,100.000,0.753|FS:100,0|Ov:100,100,100>
<Home|MPos:-135.000,100.000,0.409|FS:100,0>
<Home|MPos:-135.000,100.000,0.069|FS:0,0>
[MSG:DBG: CycleStop Pulloff0]
[MSG:DBG: Homing nextPhase SlowApproach]
[MSG:DBG: Starting from -135.000,100.000,0.013]
[MSG:DBG: Planned move to -135.000,100.000,3.313 @ 100.000]
<Home|MPos:-135.000,100.000,0.194|FS:100,0>
<Home|MPos:-135.000,100.000,0.538|FS:100,0>
<Home|MPos:-135.000,100.000,0.878|FS:100,0>
<Home|MPos:-135.000,100.000,1.222|FS:100,0>
<Home|MPos:-135.000,100.000,1.566|FS:100,0>
<Home|MPos:-135.000,100.000,1.909|FS:100,0>
<Home|MPos:-135.000,100.000,2.253|FS:100,0>
<Home|MPos:-135.000,100.000,2.597|FS:100,0>
<Home|MPos:-135.000,100.000,2.941|FS:100,0>
[MSG:DBG:  Z All Limit 1]
[MSG:DBG: Homing limited Z]
[MSG:DBG: Homing nextPhase Pulloff1]
[MSG:DBG: Starting from -135.000,100.000,3.000]
[MSG:DBG: Planned move to -135.000,100.000,0.000 @ 100.000]
<Home|MPos:-135.000,100.000,2.994|FS:100,0|Pn:Z>
<Home|MPos:-135.000,100.000,2.672|FS:100,0|Pn:Z>
[MSG:DBG:  Z All Limit 0]
<Home|MPos:-135.000,100.000,2.500|FS:100,0>
<Home|MPos:-135.000,100.000,2.153|FS:100,0>
<Home|MPos:-135.000,100.000,1.812|FS:100,0>
<Home|MPos:-135.000,100.000,1.469|FS:100,0>
<Home|MPos:-135.000,100.000,1.125|FS:100,0>
<Home|MPos:-135.000,100.000,0.781|FS:100,0>
<Home|MPos:-135.000,100.000,0.438|FS:100,0|Ov:100,100,100>
<Home|MPos:-135.000,100.000,0.094|FS:0,0>
[MSG:DBG: CycleStop Pulloff1]
[MSG:DBG: Homing nextPhase Pulloff2]
[MSG:Homed:Z]
[MSG:DBG: Homing Cycle XY]
[MSG:DBG: Homing nextPhase FastApproach]
[MSG:DBG: Starting from -135.000,100.000,0.000]
[MSG:DBG: Planned move to -432.000,397.000,0.000 @ 707.107]
<Home|MPos:-135.566,100.566,0.000|FS:707,0>
<Home|MPos:-137.284,102.284,0.000|FS:707,0>
[MSG:DBG:  X All Limit 1]
[MSG:DBG:  Y All Limit 1]
[MSG:DBG: Homing limited X Y]
[MSG:DBG: Homing nextPhase Pulloff0]
[MSG:DBG: Starting from -138.019,103.019,0.000]
[MSG:DBG: Planned move to -135.019,100.019,0.000 @ 141.421]
<Home|MPos:-137.997,102.997,0.000|FS:141,0|Pn:XY>
[MSG:DBG:  X All Limit 0]
<Home|MPos:-137.656,102.656,0.000|FS:141,0|Pn:Y>
[MSG:DBG:  Y All Limit 0]
<Home|MPos:-137.481,102.481,0.000|FS:141,0>
<Home|MPos:-137.137,102.137,0.000|FS:141,0>
<Home|MPos:-136.794,101.794,0.000|FS:141,0>
<Home|MPos:-136.453,101.453,0.000|FS:141,0|WCO:13.766,-70.078,-170.859>
<Home|MPos:-136.109,101.109,0.000|FS:141,0>
<Home|MPos:-135.766,100.766,0.000|FS:141,0>
<Home|MPos:-135.422,100.422,0.000|FS:141,0>
<Home|MPos:-135.078,100.078,0.000|FS:0,0>
[MSG:DBG: CycleStop Pulloff0]
[MSG:DBG: Homing nextPhase SlowApproach]
[MSG:DBG: Starting from -135.019,100.019,0.000]
[MSG:DBG: Planned move to -138.319,103.319,0.000 @ 141.421]
<Home|MPos:-135.272,100.272,0.000|FS:141,0>
<Home|MPos:-135.616,100.616,0.000|FS:141,0>
<Home|MPos:-135.959,100.959,0.000|FS:141,0>
<Home|MPos:-136.303,101.303,0.000|FS:141,0>
<Home|MPos:-136.644,101.644,0.000|FS:141,0>
<Home|MPos:-136.988,101.988,0.000|FS:141,0>
<Home|MPos:-137.331,102.331,0.000|FS:141,0|Ov:100,100,100>
<Home|MPos:-137.675,102.675,0.000|FS:141,0>
[MSG:DBG:  X All Limit 1]
[MSG:DBG: Homing limited X]
[MSG:DBG: Homing replan with Y]
[MSG:DBG: Starting from -138.003,103.003,0.000]
[MSG:DBG: Planned move to -138.003,106.303,0.000 @ 100.000]
[MSG:DBG:  Y All Limit 1]
[MSG:DBG: Homing limited X Y]
[MSG:DBG: Homing nextPhase Pulloff1]
[MSG:DBG: Starting from -138.003,103.006,0.000]
[MSG:DBG: Planned move to -135.003,100.006,0.000 @ 141.421]
<Home|MPos:-137.997,103.000,0.000|FS:141,0|Pn:XY>
[MSG:DBG:  X All Limit 0]
<Home|MPos:-137.903,102.906,0.000|FS:141,0|Pn:Y>
<Home|MPos:-137.559,102.562,0.000|FS:141,0|Pn:Y>
[MSG:DBG:  Y All Limit 0]
<Home|MPos:-137.469,102.472,0.000|FS:141,0>
<Home|MPos:-137.125,102.128,0.000|FS:141,0>
<Home|MPos:-136.781,101.784,0.000|FS:141,0>
<Home|MPos:-136.438,101.441,0.000|FS:141,0>
<Home|MPos:-136.094,101.097,0.000|FS:141,0>
<Home|MPos:-135.753,100.756,0.000|FS:141,0>
<Home|MPos:-135.409,100.412,0.000|FS:141,0>
<Home|MPos:-135.066,100.069,0.000|FS:0,0>
[MSG:DBG: CycleStop Pulloff1]
[MSG:DBG: Homing nextPhase Pulloff2]
[MSG:Homed:XY]
[MSG:DBG: Homing done]
ok

[HolyBlackCat]

After some experimentation, this seems to be related to simultaneously homing >1 axis. I'm able to repro this on $H=XY without homing Z, but not able to repro by homing X and Y individually. (Unless I'm just being unlucky; it's not consistent.)

Maybe it happens when both switches hit on the same tick? 🤔

[MitchBradley]

The AI that is working on the problem is making very little progress but I have a couple of ideas.

Here are a couple of things you could try to help narrow down the problem:

1. Change the X and Y pulloff_mm amounts so they are not the same, say 3 and 4. The idea is that maybe the switches are activating nearly simultaneously and that triggers some sort of race condition. If the pulloffs are different, that would break the simultaneity.
2. Try using limit_neg_pin for X and limit_pos_pin for Y. Perhaps there is some issue with limit_all pins, which are rarely used and might have a bug that we haven't seen yet.

[MitchBradley]

After a bit of prompting, copilot found something that looks plausible - while waiting for motion to settle after hitting the switch during slow approach, the switches are bouncing, generating so many events (even after debouncing) as to fill up the event queue.

One way to test this is to change homing/settle_ms on both axes. If you make it much shorter, say 50 ms instead of the default 250, the failure should occur very infrequently, perhaps never. Conversely, if you increase it to say 500 ms, the failure should occur more frequently.

Can you run these tests to verify the hypothesis?

[HolyBlackCat]

settle_ms doesn't seem to make a difference. Unless I'm missing something, it doesn't seem to do anything at all: I've set it to 1000, and there weren't any 1 sec delays in the homing process.

[HolyBlackCat]

Setting pulloff_mm: 5 on Y (instead of 3) doesn't make a difference either. Moreover, the second homing pass seems to hit both switches at the same time even when the pulloff setting is different. It seems it does correctly affect the final pulloff distance (if the homing is successful), but doesn't correctly handle the pulloff distance between the two homing passes? I didn't invesigate too closely.

I also tried modifying feed_mm_per_min on one of the axes, and this doesn't work correctly either. It works when homing the axes individually, but when homing XY at the same time, it seems to use the value of feed_mm_per_min from X and ignore it on Y, resulting again in both switches hitting at the same time.

It's honestly a bit concerning how many of those settings don't do what they're supposed to (unless I'm missing something).

[HolyBlackCat]

Lastly, switching to pos/neg pins didn't help either.

[MitchBradley]

Now that I think about it, it's not surprising that changing the pulloff and feed rate does not help. Multi-axis motions are treated as a vector that starts at one point and ends at another, so the planner will automatically scale the motion direction and speed so both axes arrive at the endpoint simultaneously. I need to find some other way to test this.

I'm not sure why settle_ms has no effect, but it isn't entirely surprising given that hardly anybody has a good reason to change it, so it hardly ever gets exercised.

[HolyBlackCat]

I think there's always zero delay, not even a fixed amount. I tried running grblHAL for a change (which also defaults to 250ms IIRC), and there I can actually notice even the 250ms delay.

[MitchBradley]

I found the cause of the settling delay bug. It was a missing reference operator in a function parameter. When I fixed that, the delay was noticeable and changeable. Unfortunately the zero delay invalidates the queue-flooding hypothesis, and copilot gave up on finding another explanation, so I will have to solve this the old-fashioned - and time intensive - way.

[MitchBradley]

I pushed a change to the Portability branch that might fix the xy pulloff problem. You can test it with this binary: https://github.com/bdring/FluidNC/actions/runs/26786419645/artifacts/7344783688