Bed leveling: DWC heightmap shifted in Y
-
I'm on RRF+SBC 3.5.2 on my Duet2.
I recently started again using mesh bed leveling.
I have a CoreXY gantry with the bed moving up and down.
The extruder on the XY gantry has a normal nozzle with an inductive probe mounted "behind (Y axis)" the nozzle.config.g:
; Z probe M558 P8 C"exp.e5stop" H3 F200:50 T7000 A3 S0.05 ; use E5_STOP as inductive-NPN probe, dual speed, multi-probe G31 X0 Y-50 Z3.80 P25 ; Z probe: set the probe XY offset, Z height and signal threshold ; Axis dimensions M208 X0:265 Y0:244 Z0:240 ; axis maxima M671 X-52:306 Y113:113 S10 ; define positions of Z leadscrews M556 S100 X0 Y0 Z0 ; axis compensation
Macro to run the mesh:
M561 ; clear any existing bed transform G32 ; home and level M557 X15:245 Y-49:137 P9:9 ; set probe coordinates G29 ; detailed Z probe
After running the mesh, DWC shows me this heightmap:
(please ignore the actual shape and deviation)My concern here is the Y shift of the bed relative to the 0,0 coordinate origin. It looks like the coordinate system is incorrect by about the same amount as the nozzle-to-probe offset.
The probing points are aligned in such a way that the inductive probe is moved to the edges of the bed as far as possible.
Is this a visual-only bug in DWC, or is the actual compensation matrix incorrectly computed and therefor also affecting the compensation during printing? How do I move my probed bed to Y=0? The actual physical movements and probing points are correct and how I want them to be.
-
@resam said in Bed leveling: DWC heightmap shifted in Y:
These are you probe offsets:
G31 X0 Y-50 Z3.80 P25
Are the X and Y parameters correct? I think not. See https://docs.duet3d.com/en/User_manual/Connecting_hardware/Z_probe_testing#measuring-probe-x-y-offset
M557 X15:245 Y-49:137 P9:9
You are asking it to probe at Y-49. This is the coordinates that the axis goes to, relative to the probe, and the coordinates that are returned with the Z height. So it is no surprise that the mesh looks like it starts at Y-49.
Note: the XY coordinates are permitted to be outside the normal printable bed area! This is intentional, because some printers (e.g. delta printers) benefit from probing areas not used for printing. (See https://docs.duet3d.com/en/User_manual/Reference/Gcodes#description-3)
Most likely your probe offsets are incorrect, and it's throwing out the coordinates of where you are probing. If you get the offsets correct, adjust the M557 and the mesh will align.
Ian
-
@droftarts thanks for your response. The physical movements are correct, so I don't understand how the probe offsets would be wrong.
One thing I noticed:
M557 ... Y0:137 ...
actually starts probing at Y=50. So the M557 coordinates are probe coordinates. My probe is offset from the nozzle in Y by -50, see my G31 command. Thats why in my M557 I want to probe at Y=-49 to move the tool head as far forward as possible.The generated
heightmap.csv
starts like this:RepRapFirmware height map file v2 generated at 2024-08-28 18:44, min error -0.217, max error 0.199, mean 0.046, deviation 0.088 axis0,axis1,min0,max0,min1,max1,radius,spacing0,spacing1,num0,num1 X,Y,15.00,245.00,-49.00,137.00,-1.00,28.75,23.25,9,9 ...
Searching through https://github.com/Duet3D/DuetWebControl/blob/v3.5-dev/src/plugins/HeightMap/HeightMap.vue and similar files, I don't see the probe offsets being used anywhere.
Is the heightmap visualization aware of probe offset?
-
@resam said in Bed leveling: DWC heightmap shifted in Y:
probe mounted "behind (Y axis)" the nozzle.
If it's mounted behind, the probe offset for Y would be positive, would it not?
-
Thanks all for questioning my setup! I now managed to get it fixed / changed in the following way:
# define Z probe G31 X0 Y60 Z3.80 P25 # Y was -50 before # bed leveling with left+right lead screws G30 P0 X10 Y130 Z-99999 # Y was 20 G30 P1 X240 Y130 Z-99999 S2 # Y was 20 G28 Z # mesh leveling M557 X20:240 Y55:230 P9:9 # Y was -49:137
So I had to roughly add
3 x probe_offset
to most positions compared to my previously (flawed) configs.Now the visual heightmap in DWC looks reasonable.
-
-