Testing Auto level?

 I have a z-probe as Z-Min as well via bltouch,  i've used G32 S2, to probe the bed,  enabled auto-leveling but it isn't actually leveling at all during a print.   some spots are 0.1-0.2mm off on parts of the bed (eg one corner) and it's easy to see when you print in those spots, its exactly like it would without leveling.   i see no z-axis movement, subtle or not at all. 

No idea how else to tell its working or not.  or if it's a bug or something else i am doing wrong.   Running 1.0 DEV.


  • G32 assumes flat surface. See result of G32 if it measured a tilt to correct. After it it is enabled and will correct the rotation if present. If your bed is a bit bumpy or warped, that can not be corrected with G32. For this we have a distortion map function (G33) if compiled in, that would measure bumps in addition and follow them.
  • Hey,

    Thanks for the reply.   i assume as much, there is some bumps so to speak though this is my last probe result.... i raised to 10mm before starting since it can't drop the probe otherwise.  but one side of the bed it's 13 and the other drops to 12.8 or so...  but i'll try distortion correction too since there is some bumps.  

    p.p1 {margin: 0.0px 0.0px 0.0px 0.0px; font: 12.0px 'Helvetica Neue'; color: #454545} p.p2 {margin: 0.0px 0.0px 0.0px 0.0px; font: 12.0px 'Helvetica Neue'; color: #454545; min-height: 14.0px}

    Recv: Z-probe:13.064 X:30.00 Y:10.00

    Recv: Z-probe:13.102 X:30.00 Y:65.00

    Recv: Z-probe:13.212 X:30.00 Y:120.00

    Recv: Z-probe:12.977 X:30.00 Y:175.00

    Recv: Z-probe:13.061 X:30.00 Y:230.00

    Recv: Z-probe:13.116 X:82.50 Y:10.00

    Recv: Z-probe:13.009 X:82.50 Y:65.00

    Recv: Z-probe:13.027 X:82.50 Y:120.00

    Recv: Z-probe:12.962 X:82.50 Y:175.00

    Recv: Z-probe:13.044 X:82.50 Y:230.00

    Recv: Z-probe:13.019 X:135.00 Y:10.00

    Recv: Z-probe:13.067 X:135.00 Y:65.00

    Recv: Z-probe:12.962 X:135.00 Y:120.00

    Recv: Z-probe:12.915 X:135.00 Y:175.00

    Recv: Z-probe:12.977 X:135.00 Y:230.00

    Recv: Z-probe:12.982 X:187.50 Y:10.00

    Recv: Z-probe:12.925 X:187.50 Y:65.00

    Recv: Z-probe:12.884 X:187.50 Y:120.00

    Recv: Z-probe:12.897 X:187.50 Y:175.00

    Recv: Z-probe:12.866 X:187.50 Y:230.00

    Recv: Z-probe:12.937 X:240.00 Y:10.00

    Recv: Z-probe:12.910 X:240.00 Y:65.00

    Recv: Z-probe:12.839 X:240.00 Y:120.00

    Recv: Z-probe:12.821 X:240.00 Y:175.00

    Recv: Z-probe:12.857 X:240.00 Y:230.00

    Recv: plane: a = -0.0010 b = -0.0004 c = 13.1632

    Recv: Transformation matrix: 1.000000 0.000000 0.001033 -0.000000 1.000000 0.000388 -0.001033 -0.000388 0.999999

    Recv: CurrentZ: 13.16 at Z:0.00

  • See you use a 5x5 grid. Thats good with warped bed. And result shows bed is nearly rotated to optimal position.

    Printing with 0.3mm first layer height should already be possible that way. If you use G33 please use latest dev version from today. Just fixed an issue with it.
  • k, will give it a shot. 
  • well, fuck that.  g33 is completely messed up, has no coordinate reset on error, no bounds checking, nothing... not only has it driven the extruder into the bed twice so far damaging my coating, and screwed up the z-probe, it's damaged my X axis thinking it can go where it can't... despite it being homed.  even after a reset  it completely acts erratically... then super slowly moves at 1mm sec back to it's original position assuming it isn't jammed up against another part of the printer already....  i tried this 3 times now, and now the printer is broke.       after resetting an error m999 from the probe not deploying somehow, the printer has no idea where it is and rams itself into end stops or the oppose it side of the axis, i cannot home after a resetting the error as it rams itself into the bed since it dosen't even deploy the probe despite it being fine.... wtf. 

    Please get some unit testing, throw it in a simulator, whatever...  it's complete bullshit to even suggest using this on dev without testing it.
  • Yeah, I noticed G33 being rather destructive on my Delta, too. It's ok on my cartesians, there the grid can easily be configured onto the build plate. Also, measuring in a r=16mm circle isn't helping at all---that's how much is left after I have to substract the probe offset from all sides. On a cartesian printer the probe only needs to be substracted from one direction.
Sign In or Register to comment.