I have a bltouch clone which work fine on glass bed. After switching to textured PEI sheet, it’s variance trippled to 0.1mm, which make my bed mesh all wobbly.

I’m I suppose to remove the steel plate when making the bed mesh? Do you home Z with a probe with this setup?

The weirdest thing is, despite all this all my prints adhere completely fine. I guess PEI is just that good.

  • Flaky_Fish69@kbin.social
    link
    fedilink
    arrow-up
    2
    ·
    1 year ago

    I use marlin UBL with a bl touch. On textured sheet. The first thing is to recognize it’s supposed to have that variance.

    I found that increasing the first layer extrusion width (140%, normally I’m at 110% unless I need a strong and ugly part,) and juicing the extrusion multiplier slightly helped.

    My calibration process for z offset was:

    • set z off set to zero,
      -turn off software endsroos. (M211 s0; use s1 to turn them back on. This allows you to go negative positions.)
    • move the nozzle to 0,0,0 and slide down carefully. I use the smallest feeler gauge I’ve got to test contact.

    Then I probe the mesh and save it to the board. (And recall it on the print start gcode and do a 3x3 probe to tilt the mesh into position.)

    What I have found is that it slides a bit with the magnetic base- especially on heavy/fast prints invalidating the mesh.

  • Dangerhart@lemm.ee
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Maybe clarify what you mean by variance - Samples in the same spot vs variance of min/max points? Temp can change things, are you keeping it constant between measurements and letting the bed heat soak at all? How does the first layer look? If it’s consistent you may be over thinking it

  • crankylinuxuser@midwest.social
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Printer: creality ender 3 pro bed: creality PEI double sided bed hotend: creality spider 3.0 firmware: custom marlin 2.0 fw filament I use most often: atomic filament ABS creality bag enclosure: yes

    Sure does work… But I have babystepping on (Marlin), and adjusted babystepping by coding in a 30x30x5 block with 8 skirt loops 6mm away. The babystepping in this case is to adjust the Z offset, which it does so live.

    I was able to get it tuned in so that all prints just apply beautifully.

  • LifeInOregon@kbin.social
    link
    fedilink
    arrow-up
    1
    ·
    1 year ago

    I’ve had no issues using the bl touch on a textured bed up to this point, and I’ve been using mine in combination for about 18 months. The bed texture shouldn’t be significant enough if your z offset is appropriately dialed in.

  • ffhein@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    1 year ago

    Does the variance return to normal if you switch back to the glass plate? If you probe multiple times on the same spot, do you still get large variance?

      • ffhein@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        1 year ago

        If you press down on the PEI/steel sheet manually, does it look like it’s flexing or does it feel stable? When you switched from glass to steel+PEI, did you put an adhesive magnet sheet on the print carriage, or are you using some other method to hold the print surface in place? Just found a post on reddit where someone discovered that their build plate only sticks to the magnetic sheet when they rotate the plate 90 degrees, so perhaps that’s worth a try.

          • ffhein@lemmy.world
            link
            fedilink
            English
            arrow-up
            0
            ·
            1 year ago

            Not doubting you, but I can’t think of a logical reason for why it would be like that. The probe pin shouldn’t push down hard enough to be able to deform the PEI

            Perhaps as a workaround you could configure the firmware to probe each spot 3 times and average the result, if you don’t mind the extra time levelling would take?

            • x7tYnC6c@lemmy.worldOP
              link
              fedilink
              English
              arrow-up
              0
              ·
              1 year ago

              I meant the probe is deflected by the texture. Taking multiple sample is what I’m doing, yes. It take at least 5 samples for it to be reasonably smooth, which is why I’m asking other experience.

              • ffhein@lemmy.world
                link
                fedilink
                English
                arrow-up
                0
                ·
                1 year ago

                You mean you think the probe slides off the bumps in the texture? I think the probe should be stiff and light enough to not do that, but perhaps you could’ve been unlucky if it’s a clone not from a well known brand (e.g. Trianglelabs… even Creality’s probe’s are supposed to be quite good afaik)

                • x7tYnC6c@lemmy.worldOP
                  link
                  fedilink
                  English
                  arrow-up
                  1
                  ·
                  1 year ago

                  Yes. I’m certain it is a build quality problem. The probe is just plastic and by design hang freely. Even a steel rod with bearing can easily be deflected by a few tens of micron.