3.4.0beta7+7 mesh bed issue?

RomeoFox

Is there a known bug with the 3.4.0beta7+7 regarding mesh bed (G29)?
To me it looks like the result is not applied.

config.g

; Z-Probe
M558 P5 C"^io3.in" H4 F300 T45000                        
G31 P500 X0 Y25 Z0.99

M557 X20:320 Y26:320 P10:10

bed.g (not shown here: drive learning is performed 3 times in a row)

G1 F40000 X40 Y15 Z7
G30 P0 X40 Y40 Z-99999 ; probe near a leadscrew
                                    
G1 F40000 X40 Y310 Z7
G30 P1 X40 Y335 Z-99999 ; probe near a leadscrew
                                    
G1 F40000 X320 Y310 Z7
G30 P2 X320 Y335 Z-99999
                                    
G1 F40000 X320 Y15 Z7
G30 P3 X320 Y40 Z-99999 S4 ; probe near a leadscrew and calibrate 4 motors                                    
                                                                      
G1 X170 Y145 Z7 F40000 
G30 P3 X170 Y170 Z-9999

Phaedrux

@romeofox said in 3.4.0beta7+7 mesh bed issue?:

To me it looks like the result is not applied.

Can you be more specific? Why do you say it's not working?

Can you send M122 and share the results here to see if mesh is in fact enabled?

Do you see the Z motors move to compensate?

Have you done a first layer test print with it enabled and disabled to see if there is a difference?

Can you share your full config and homing files as well? Slicer start gcode?

RomeoFox

General conditions:

Voron 2.4 350x350
magnetic bed with steel plate and smooth PEI film (polyetherimide)

I think that it does not work properly because:

see pictures and that at 0.3mm height of the first layer, right side is about 0.1mm too high, especially right front
with a sheet of paper under the nozzle I could reproduce the pictures, there is a height difference of 0,1mm
this does not seem to be due to the sensor, which gives correct distance values
for the 0.1mm height difference both the scanning speed and the dot density seem to be irrelevant 7x7 to 11x11 I tried out

Have you done a first layer test print with it enabled and disabled to see if there is a difference?

yes, looks better without compensation

Can you send M122 and share the results here to see if mesh is in fact enabled?
=== Move ===
DMs created 125, segments created 47, maxWait 654699ms, bed compensation in use: mesh, comp offset 0.000

Can you share your full config and homing files as well? Slicer start gcode?
click

Do you see the Z motors move to compensate?
Seems to be so

Thank you for your efforts!

Update:
Via M114 I get a correct display of the compensation at the position in relation to the heightmap.csv

M114
X:170.000 Y:170.000 Z:25.000 E:9.522 E0:1839.0 Count 54400 0 9974 Machine 170.000 170.000 25.000 Bed comp -0.065

Update2:

I manually checked the height correction via babystepping 0.01mm.

A correction is applied:
-> g1 z1 leads to different absolute heights with G92 S2 and G92 S1

the tendency of the compensation seems to be correct, but it is undercompensated by felt factor of 2.
G30 at the points leads to correct offsets
manual G30 = babystepping and looking at Z-probe value confirms G30

Phaedrux

Have you checked the probe trigger height at multiple points on the bed?

Touch the nozzle to the bed, set G92 Z0, move up 10mm, test trigger height with G30 S-1.

Is the trigger height consistent at all XY points?

What type of probe is it? Is it affected by magnets or density?

It seems the mesh is working as expected, and that it is the probe results that are incorrect.