Adding individual points to G29 mesh to cover irregular regions



  • I'm currently using mesh compensation and having pretty good results, but because of two binder clips (front left and back right) I'm missing probing of a couple of strips, leading to I think some very unfortunate interpolation in those regions.

    I'd like to add some individual probe points in to cover these regions. Specifically I'd like to (a) cover the majority of the plate with G29, and (b) fill in a few additional spots, presumably with G30. Am I thinking about this the right way, and should I do the G30s first so that the map isn't in place, or am I overthinking it?

    Thanks! Bed.g below.

    ; bed.g
    M557 X30:270 Y70:270 S30:25    ; define mesh for probing
    
    ; Go to roughly center of bed (for the probe)
    G1 X180 Y163 Z10 F1800
    M190 S70                       ; heat to 70 degrees, wait to complete
    
    M561                           ; clear any bed transform
    G29 S2                         ; clear current bed map (redundant?)
    G30                            ; probe bed to establish Z=0
    ;
    ; Idea:  Insert more singleton points via G30 here?
    ;
    G29                            ; probe the bed and enable compensation
    
    M140 S0                        ; stop heating
    

  • Moderator

    can you post an image of your current heightmap?

    because of two binder clips (front left and back right) I'm missing probing of a couple of strips, leading to I think some very unfortunate interpolation in those regions.

    What exactly do you mean by this?

    What type of probe are you using? What is the offset?

    Instead of binder clips you can get low profile "swiss" clips that don't take up much space on the top part. That might be it possible to probe closer to the clips and fill in the area more.

    You can edit the heightmap.csv file directly to add or modify points. I'm not sure what you are asking about using additional G30s.



  • Sure. This is a thin heated plate under glass with PEI on top. Some imperfections in the PEI layer for sure.

    Screen Shot 2020-12-21 at 12.43.15 PM.png



  • @Phaedrux There are some options for binder clips and plates that would make this better for sure. Currently my glass plate isn't as big as my heated bed, which exacerbates the situation and makes using the swiss clips more difficult.

    The part fan is the Dii cooler that you and I have discussed in the past, so it requires significant clearance as you know.

    I was hoping that I could programmatically "add" to the map by just probing more points before enabling/writing the heightmap. Felt like G30 was the appropriate command for such a thing?

    Edit: Clarity.


  • Moderator

    @paralepsis said in Adding individual points to G29 mesh to cover irregular regions:

    I was hoping that I could programmatically "add" to the map by just probing more points before enabling/writing the heightmap. Felt like G30 was the appropriate command for such a thing?

    Well not exactly, but if you are able to probe closer to the edges you could increase the size defined in M557 so that they get used when running G29.

    What exactly is the problem you're seeing that you're trying to solve?



  • I'm having some layer adhesion problems near the front right edge in particular that I could attribute to not having good bed level data up there. Obviously there are other things it might be, and I've partially enclosed the printer to reduce any air flow, and there's the standard cheap cork insulation on bottom to help with bed temp. (although I haven't put a FLIR on the bed to see what it looks like).

    I can spend another $100 to get a bigger glass plate, redo the PEI, get better clips, etc. so that I could just use a bigger structured mesh, but I was hoping I could just drive the head over into these regions and augment the mesh data in a programmatic way. Because there are only clips in the front-left and back-right, I can get the carriage into these regions in the Y, I just cannot get to the whole range of points in the X.

    Basically I was hoping this was easy enough to just try in the evening, get the data, and see if it would help, if that makes sense?

    Edit: Clarity.



  • I suppose as an experiment I could just probe the region that I can get to as a rectangle and see what it looks like. That would satisfy my curiosity, but assuming there's something interesting there, not solve the problem...


  • Moderator

    So there are just parts of that area that you can't reach because of the binder clips? But otherwise you can reach it with the probe?

    If that's the case maybe try securing the glass temporarily with tape or something low profile, probe the area, and then re-attach the clips.

    There is no way to automatically probe chunks of the bed area and then recombine the data. You can do it manually as I said by editing the heightmap.csv file directly.



  • @Phaedrux Yes. There's a small region in the front left and a small region in the back right (maybe 50x20mm) that I simply cannot go to at Z < 4. I think you get it.

    Thanks for the clarifications. I think if I were going to go the heightmap route, what I would need to do is save multiple heightmaps, combine them manually, and then reload, right?

    Too bad. I'll probably just do the work for the bigger bed.

    Thanks again for all the attention today; I appreciate it!



  • Hacked up something to cover the missing regions, stitched together by hand. Probably not interpolation. Thanks again.

    @paralepsis Screen Shot 2020-12-22 at 2.11.44 PM.png

    [ I also tweaked the probe Z height in here, which is obvious when comparing the two heightmaps. ]


Log in to reply