[BUG] Babystepping the Z-Offset uses the move axis step setting instead of the babystep setting

[luizbgomide]

Did you test the latest bugfix-2.1.x code?

Yes, and the problem still exists.

Bug Description

My setup was working flawlessly one year ago on bugfix-2.1.x (commit 9e42480), last month I tried to update it to the latest of this branch and after manually migrating all the settings it was working but babystepping the Z Probe Offset was misbehaving.

As can be seem on the video below, when trying to move the Z Offset the display shows (correctly) that 0.01mm should be moved, but the actual move uses the 10mm step for axis move setting (as show in the screen) below.

VID_20240322_114242.mp4

I've tried to change every possible setting related to BABYSTEP and Z_PROBE without any luck. Since I just migrated the settings and the display correctly shows what should have happened, this must be a bug introduced in the last year.

Old settings (at 9e42480)
old_settings.zip

Current settings
current_settings.zip

Bug Timeline

It started at some point after commit 9e42480 from one year ago

Expected behavior

The hotend moves as indicated by the Z Probe Offset.

Actual behavior

The hotend moves by the step size for the axis move.

Steps to Reproduce

No response

Version of Marlin Firmware

bugfix-2.1.x

Printer model

Two Tress Bluer

Electronics

TMC2209 UART

LCD/Controller

MKS_ROBIN_TFT35

Other add-ons

BLTouch

Bed Leveling

ABL Bilinear mesh

Your Slicer

Cura

Host Software

OctoPrint

Don't forget to include

• A ZIP file containing your Configuration.h and Configuration_adv.h.

Additional information & file uploads

No response

[ellensp]

Confirmed in the simulator

[ellensp]

I think im onto something.

Try this diff (2 changed lines in Marlin/src/lcd/tft/ui_common.cpp)

diff --git a/Marlin/src/lcd/tft/ui_common.cpp b/Marlin/src/lcd/tft/ui_common.cpp
index 944be77ab5..08fbabdfc6 100644
--- a/Marlin/src/lcd/tft/ui_common.cpp
+++ b/Marlin/src/lcd/tft/ui_common.cpp
@@ -71,7 +71,7 @@ void moveAxis(const AxisEnum axis, const int8_t direction) {
     }
   #endif
 
-  const float diff = motionAxisState.currentStepSize * direction;
+  float diff = motionAxisState.currentStepSize * direction;
 
   #if HAS_BED_PROBE
 
@@ -85,6 +85,7 @@ void moveAxis(const AxisEnum axis, const int8_t direction) {
                       , do_probe ? new_probe_offset : hotend_offset[active_extruder].z - bsDiff
                       , new_probe_offset
                     );
+        diff = 0;
         if (WITHIN(new_offs, PROBE_OFFSET_ZMIN, PROBE_OFFSET_ZMAX)) {
           babystep.add_steps(Z_AXIS, babystep_increment);
           if (do_probe)

Seems to work well in the sim

My test configs for the sim, if anyone is interested.
Configuration.zip

[ellensp]

please note with your current settings you have 410 400 steps/mm for Z with BABYSTEP_MULTIPLICATOR_Z 4 and BABYSTEP_ZPROBE_OFFSET
So each baby step will move the Z 0.009756098 0.01mm, it will take a few clicks to noticeably move the Z axis

[luizbgomide]

please note with your current settings you have 410 steps/mm for Z with BABYSTEP_MULTIPLICATOR_Z 4 and BABYSTEP_ZPROBE_OFFSET So each baby step will move the Z 0.009756098 mm, it will take a few clicks to noticeably move the Z axis

Actually it is 410 for the E0 axis, for Z it is 400.

After your last edit is it working or is it still moving double the distance?

[ellensp]

correct distance now. yes your correct its 400, its very late here, going a bit cross eyed.

[luizbgomide]

Applied the fix but it is still misbehaving during print, when I adjust the z-offset it seems to dig the hotend into the bed, and even lose control of the rest of the axis, I've seem it move back during a print move, do some sort of "layer skip", etc.

The current version is unusable as it is for my printer.

[SyncroSeb]

Had a similar problem.
Printer is a CoreXY (Tronxy X5SA Pro Stock motherboard with STMF446 Chip)
when adjusting the Z-Offset during the first layer the Z axis moved down not in 0.1 but whole mmand more steps, sometimes going down 20mm or more printing in the air, then going back to initial layer height. (+- Offset?).

The above mentioned changes to the Marlin/src/lcd/tft/ui_common.cpp (using actual bugfix version) worke out fine. Thx a lot

[thisiskeithb]

This may have been fixed in 489ef6e

Please download bugfix-2.1.x to test with the latest code and let us know if you're still having this issue.