Question about bed leveling



  • Hi,

    I've been busy trying to fine tune my deltas.

    While doing that I got to thinking about how the bed leveling procedure determines the end stop offsets.

    When I used to use 8 bit boards the approach was to position each tower carriage at a known height and run a procedure which moved until the end stop was triggered. This is easy to understand as the reference point is the known starting height.

    But I have been unable to understand how it is done in the Duet firmware in the absence of a reference point.

    Thanks much.

    Frederick


  • administrators

    I presume you mean auto calibration, not bed levelling. The Duet uses the Z probe to provide the reference points. So you need to set the Z probe trigger height correctly in the G31 command in config.g, and the trigger height needs to be consistent over all areas of the bed.



  • @dc42 said in Question about bed leveling:

    I presume you mean auto calibration, not bed levelling. The Duet uses the Z probe to provide the reference points. So you need to set the Z probe trigger height correctly in the G31 command in config.g, and the trigger height needs to be consistent over all areas of the bed.
    Thank you. I think I see how that would work.

    As part of the learning process and just of of curiosity I did the following:

    • used the minimal set of known values for M665 and M666.
    • set the end stop offsets using the method I mentioned.
    • added leveling screws to the bed and leveled it using a dial indicator attached to a dedicated effector plate.
    • used the dial indicator to take measurements at the center and at the towers.

    As expected the initial measurements at the towers were very much the same but rather different from the center.

    Based on an article about the effects of incorrect values for arm length and/or arm radius yielding "bowl shapes" I started changing the arm radius (the arm lengths were provided by the arm maker).

    I found a value that gave good results over the whole bed.

    The part I do not understand is that the values I was able to determine using this manual approach for use in M665 and M666 (which correspond to the "4 factor" values) were rather different from the values I obtained using the "auto-cal" procedure.

    The most interesting result was that the "manual" value for arm radius yielded a "flatter bed" near the max radius I use (80 mm) than the "auto" value.

    I have no idea how to manually determine the additional values the "6 factor" or "8 factor" calibrations provide.

    I need to do the "auto" and "manual" procedures some more to try and understand what is happening.

    Interesting if confusing stuff. Any references to relevant articles would be most appreciated

    Thanks again.

    Frederick


 

Looks like your connection to Duet3D was lost, please wait while we try to reconnect.