Probe_Offset_Wizard: Improvements

[swissnorp]

Requirements

• Filling out this template is required. Pull Requests without a clear description may be closed at the maintainers' discretion.

Description

Config

• Now it's possible to disable PROBE_OFFSET_START and use current probe.offset.z instead
• The set PROBE_OFFSET_WIZARD_XY_POS will be used for probing a z_offset_ref and for the nozzle/bed contact position (the nozzle will move to this position after probing).

Procedure

1. Call Wizard by user
2. Store current probe.offset.z for backup and overwrite it with PROBE_OFFSET_START if defined
3. Store Bed-Leveling-State and deactivate it
4. Home all Axes by G28
5. Call Preparation
6. Disable Soft Endstops
7. Set z_offset_ref = 0 (for if homing is done by probe and PROBE_OFFSET_WIZARD_XY_POS is not ##defined)
8. If Probe is not used for homing or PROBE_OFFSET_WIZARD_XY_POS is defined, z_offset_ref must be probed
9. If PROBE_OFFSET_WIZARD_XY_POS is not defined probing will be done at XY_CENTER
10. The Nozzle moves by the probe.offset_xy to the Probing point
11. Call Menu
12. The probed will be used to calculate the new probe.offset.z: calculated_z_offset = probe.offset.z + current_position.z - z_offset_ref
13. User adjust Offset by menu
14. Done by user
15. The calulated Offset is stored
16. If Probe is used for homing, homing is done with the new offset.
17. Cancel by user
18. The old Offset is restored
19. If Probe is used for homing and PROBE_OFFSET_START is set, homing is done with the restored offset.

Benefits

• Its possible to calibrate the probe offset even if its not used for homing.
• The calibration is accurate because probing and nozzle/bed-contact points are the same
• Its possible to define a specific position for the calibration wizard

Configurations

Niente

Related Issues

#19650
#20204

[qwewer0]

This fixed the issue 4-6 in #20204 (comment), but it still goes back to the status screen.

[swissnorp]

This fixed the issue 4-6 in #20204 (comment), but it still goes back to the status screen.

@qwewer0 Thank you for your quick test! I could reproduce your problem. If homing takes to long (nozzle far away from endstop) it falls back to the status screen.

I think the latest commit fixed it. Could you give it another try?

Thanks, norp

[qwewer0]

Tried it with the new changes:
Now after homing it probes the bed (Z reference) one more times (3 times with MULTIPLE_PROBING 2, EXTRA_PROBING 1) while the LCD shows nothing (blank), then it moves to PROBE_OFFSET_WIZARD_XY_POS and goes into the wizard menu and can be used as before.

• I would love to see an LCD message while it probes for the Z reference, so that it isn't just a blank screen.
• And one thing that I don't like, is the homing after the done/cancel button. Is there a particular reason why do we need to home after the wizard?

[filippor]

Tested on a delta and it seems to work with PROBE_OFFSET_WIZARD_XY_POS uncommented

[swissnorp]

@qwewer0

• I would love to see an LCD message while it probes for the Z reference, so that it isn't just a blank screen.

This is fixed now.
I added a message screen when probing the Z Reference: Probing Z Reference.
And an other one for the moving part: Moving to Probing Pos

• And one thing that I don't like, is the homing after the done/cancel button. Is there a particular reason why do we need to home after the wizard?

Yes its absolutely needed. If you don't believe me, try it by uncommenting it 😸. The current_position.z would not be trustable any more under the given sircumstances/if-conditions. But I reduced it to home Z only.

@thinkyhead
As a cherry on top it would be great if we could extend the messages for homing in marlin so it would be easy to give proper screens:
MSG_LEVEL_BED_HOMING = _UxGT("Homing XYZ");
MSG_LEVEL_BED_HOMING_XY = _UxGT("Homing XY");
MSG_LEVEL_BED_HOMING_Z = _UxGT("Homing Z");

@filippor Thank you for testing and giving a feedback 👍

I think the code is not perfect but it works. I am relatively new to marlin and I never proberly learned to code 😅. Maybe there is some clean up possible.

Enjoy Testing, norp

[qwewer0]

But I reduced it to home Z only.

That is the same as it was for probe users.

Why can't do_z_clearance(); be used after the wizard? I changed queue.inject_P(PSTR("G28Z")); to do_z_clearance(20.0); and it worked just fine.

In my opinion a homing after the wizard is inconvenient (if it isn't absolutely needed), but a z clearance is absolutely needed, because of a possible heated nozzle.

Sorry if I'm missing something.

[qwewer0]

Would it not work?

ACTION_ITEM(MSG_BUTTON_DONE, []{
    set_offset_and_go_back(calculated_z_offset);
    do_z_clearance(20.0);
  });

ACTION_ITEM(MSG_BUTTON_CANCEL, []{
    set_offset_and_go_back(z_offset_backup);
    do_z_clearance(20.0);
  });

[swissnorp]

@qwewer0 Thank you for partizipation. I realy apreciate your skepsis.

Sorry if I'm missing something.

Yep you are 😉 .
Homing for the wizard was done with PROBE_OFFSET_START or with (let's expext) wrong probe.offset.z. Changing or resetting the probe.offset.z to the proper value does not change current_position.z!

So rehoming Z is absolutely needed!

Would it not work?

ACTION_ITEM(MSG_BUTTON_DONE, []{
    set_offset_and_go_back(calculated_z_offset);
    do_z_clearance(20.0);
  });

ACTION_ITEM(MSG_BUTTON_CANCEL, []{
    set_offset_and_go_back(z_offset_backup);
    do_z_clearance(20.0);
  });

NO 😏

Did you measure the distance between nozzle and bed, and did it match with the current_position.z?
Make sure that PROBE_OFFSET_START or your currently set probe.offset.z (from Settings/EEPROM) is more than 1 mm off from
your proper/real probe.offset.z, and measure with a ruler!

Adios, norp

[qwewer0]

Tested it with do_z_clearance(20.0);, and after the wizard the LCD showed 20, and the nozzle was at 18.75 with a z offset of -1.25.

Despite all this, I still believe that it should only move the z axis up a little for clearance, and either disengage the steppers or just set the z axis or all to be unhomed/untrusted.
Edit: With perfect z offset before the wizard.

[swissnorp]

Tested it with do_z_clearance(20.0);, and after the wizard the LCD showed 20, and the nozzle was at 18.75 with a z offset of -1.25.

Despite all this, I still believe that it should only move the z axis up a little for clearance, and either disengage the steppers or just set the z axis or all to be unhomed/untrusted.
Edit: With perfect z offset before the wizard.

But you have had enabled PROBE_OFFSET_START with a different value, haven't you? And thats the reason for me to rehome (or to set Z-Axis untrusted, as you mentioned).

Conclusion:

• So we have different opinions about homing after the wizard. I vote for rehoming, because I think it's safer and more convenient. And there are several newer and older features, which do a rehoming after a calibration process. I leave it in there!
• But nevertheless I think about adding a do_z_clearance(Z); indepentently.
• I will also add a SanityCheck to check if PROBE_OFFSET_START or NOZZLE_TO_PROBE_OFFSET.Z, because the Probe will never trigger at a higher point as the nozzle!

I will work on it tomorrow night (MEZ/CET).

Good night, norp

[qwewer0]

If the homing wins, then I'm 100% with it.
And for the a Z clearance after the homing, I think it definitely needs a bit (10-20 mm) of a Z clearance even after the homing, because while the Z_CLEARANCE_DEPLOY_PROBE could be big enough, like with a BLTouch setup, but it might be too small with an inductive or laser probe, and we don't want accidental heated bed burning with the wizard, do we?

[swissnorp]

If the homing wins, then I'm 100% with it.
And for the a Z clearance after the homing, I think it definitely needs a bit (10-20 mm) of a Z clearance even after the homing, because while the Z_CLEARANCE_DEPLOY_PROBE could be big enough, like with a BLTouch setup, but it might be too small with an inductive or laser probe, and we don't want accidental heated bed burning with the wizard, do we?

I think you are meaning do_z_clearance(Z); before homing, aren't you? After homing it would make absolutely no sens!?

[thinkyhead]

We have Z_AFTER_HOMING to set the desired raise after G28 is complete.

[swissnorp]

@thinkyhead yes excactly

Conclusion:

* So we have different opinions about homing after the wizard. I vote for rehoming, because I think it's safer and more convenient. And there are several newer and older features, which do a rehoming after a calibration process. I leave it in there!

* But nevertheless I think about adding a `do_z_clearance(Z);` indepentently.

* I will also add a `SanityCheck` to check if `PROBE_OFFSET_START` or `NOZZLE_TO_PROBE_OFFSET.Z`, because the Probe will never trigger at a higher point as the nozzle!

SanityCheck still has to be done.

For the rest, I did it like this:

  ACTION_ITEM(MSG_BUTTON_DONE, []{
    set_offset_and_go_back(calculated_z_offset);

    // Set Z to 0, as we can expect
    current_position.z = 0;
    planner.sync_plan_position();

    // Raise Z as if it was homed
    do_z_clearance(0
    #ifdef Z_AFTER_HOMING
      +Z_AFTER_HOMING
    #elseif defined(Z_HOMING_HEIGHT)
      +Z_HOMING_HEIGHT
    #endif
    );
  });

  ACTION_ITEM(MSG_BUTTON_CANCEL, []{
    set_offset_and_go_back(z_offset_backup);

    // If wizard-homing was done with PROBE_OFFSET_WIZARD_START_Z by probe.
    // Set Z to 0 (if higher than 0) to be sure for G28 clearance and rehome Z with backed up offset.
    #if defined(PROBE_OFFSET_WIZARD_START_Z) && EITHER(Z_MIN_PROBE_USES_Z_MIN_ENDSTOP_PIN, USE_PROBE_FOR_Z_HOMING)
      if(current_position.z > 0) { current_position.z = 0; planner.sync_plan_position(); }
      queue.inject_P(PSTR("G28Z"));

    // Otherwise do a Z clearance move like after Homing
    #else
      do_z_clearance(0
      #ifdef Z_AFTER_HOMING
        +Z_AFTER_HOMING
      #elseif defined(Z_HOMING_HEIGHT)
        +Z_HOMING_HEIGHT
      #endif
      );
    #endif
  });

Good Night, norp

[qwewer0]

It should be safe even if Z_AFTER_HOMING and Z_HOMING_HEIGHT isn't defined, and move the nozzle up at the end.

[ohrn]

Hi!

This sanity check addition breaks compilation on my printer, a ZYYX Pro.

    static_assert(sanity_nozzle_to_probe_offset.z <= 0.25,
                  "Are you sure your Probe triggers above the nozzle? Set a negative Z value in the NOZZLE_TO_PROBE_OFFSET.");

The probe truly triggers 2 mm "above" the nozzle. The printer has a micro switch next to the nozzle and notches cut out of the build plate edges. When probing the nozzle lowers into a notch and the switch engages on the plate.

Z safe probing and carefully defined probe points matching the notches is a must on this printer! 😄

I guess this is not a common design, but it would be nice if this error could be turned into a warning, or another "I really know what I'm doing" config added to allow a positive z offset.

/Fredrik

[swissnorp]

@ohrn Thats some crazy endstop mount 😄 .
Could you test 9702170 and give some feedback? Make sure you enable/uncomment NOZZLE_TO_PROBE_OFFSET_Z_POSITIVE.

Thanks for your input.