1. Got a question or need help troubleshooting? Post to the troubleshooting forum or Search the forums!

Unresolved Possible z-offset problem?

Discussion in 'Troubleshooting' started by fred3d, Feb 3, 2023.

  1. fred3d

    fred3d Member

    Joined:
    Jun 6, 2019
    Messages:
    69
    Likes Received:
    20
    I'm assisting an R1 plus owner with a new problem with his printer. Stock firmware, other than any changes that might have been required by moving to an E3Dv6 hot end. After the hot end swap, all has been great until recently.

    The bottom layer is as if the glass bed is "potato-chipped" in that the front left corner and right rear corner are clearly higher than the rest of the bed. It's progressive. A test model was created, five 25 mm squares for each corner and the center. The results are consistent with the description above, more or less progressively squishing and not-squishing.

    There are no obvious mechanical failures. The home sequence G28 and bed leveling sequence G29 perform as expected, via MatterControl.

    My research has not allowed me to determine how to read the current z-offset, nor do I know how to determine how much should be set to solve this problem.

    The threaded rods (z-axis) have been manually turned to bring the nozzle to the bed, listening for the click-click of the end-stops. I turn the rod, listen for the engage click, then back it up just to the point of disengaging. This is done for both sides with the weight of the nozzle on each side in turn, then double checked to allow for lever-arm mis-alignment.

    I'm not quite sure what next steps will help solve this.
     
  2. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,912
    Likes Received:
    7,338
    So not stock firmware :) Needed to be tweaked for the E3D. Not complex changes but NOT stock.
    Autoleveling Z offset should be done in the startup GCode script with the R1 series.
    See the FAQ for that.
     
  3. fred3d

    fred3d Member

    Joined:
    Jun 6, 2019
    Messages:
    69
    Likes Received:
    20
    The link to the FAQ is quite helpful. I see that the Z-offset issue is at the top of the list. What concern should I have about the v1.0 versus v1.1 for the Marlin firmware and how does one determine the version of Marlin firmware? I see that Matter Control is "supported" in the FAQ, which gives me a starting point. As this printer has been working fine until recently, whatever firmware version should not have much effect on finding the solution, I hope.
     
  4. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,912
    Likes Received:
    7,338
  5. fred3d

    fred3d Member

    Joined:
    Jun 6, 2019
    Messages:
    69
    Likes Received:
    20
    I've read as much of the linked pages as I could find. After checking a number of things, I am not sure this is related to z-offset. I've certainly learned quite a bit and some of the things I've checked were not previously checked, but also are not a contributing factor, I believe.

    For example, the mechanicals are good. There are no loose parts. The bed is secure to the magnets, although one was not attached and is not corrected. No components on the carriage are loose, including the hot end assembly. It's an E3Dv6 with a very secure mount.

    In the process of composing this post, I had a thought appear in my alleged mind. I removed the 3D printed feet although I suspect they are not really a factor. I have a stone slab intended for a future build, but suitable for this test. The printer was moved onto the slab and the test print started again.

    There's a substantial improvement across the nine 50 mm squares at the first layer. Some of them are still displaying "plowing" and zero-feed of the filament, but the areas are reduced and the quality overall is improved. This tells me that the structure may have had some twist, as the glass bed is completely flat and a frame twist can create these upper right/lower left layer height differences, right?

    But wait, there's more. During all this testing, the upper shell has been removed. I know from my old VW Beetle days that the shell is a contributing factor to stiffness. The shell has been re-attached and another test will be run, but with a slightly higher z-offset, once I figure out to accomplish that.
     
  6. fred3d

    fred3d Member

    Joined:
    Jun 6, 2019
    Messages:
    69
    Likes Received:
    20
    I'm still trying to solve this problem. By adding an M565 Z-0.8 code between G28 and G29, the test prints have improved, but the "potato chip" effect remains. I set the power off z-screws to place the carriage parallel to the bed with the bed centered in the y-direction and the hot end centered in the x-direction. This change made the results worse.

    The rear left and front right has too much z-offset, while the front left and rear right have too little. The variance is inconsistent as well. In the rear right, it's so close that the nozzle extrudes zero filament, but the front left allows filament to be placed on the bed.

    Even allowing for this identification of the apparent distortion, I'm not sure what to check. The bed is dead flat, confirmed by placing a straightedge and also another glass bed atop the existing bed.

    This implies that the distortion is being imparted by the carriage and/or the bed leveling process. I've found nothing loose and no cracked components. The hot end is secure in the carriage and the carriage has no slop or play on the rails.

    What to check next?
     
  7. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,912
    Likes Received:
    7,338
    What I have done for the R1 and C2 types was to shim the bed with paper -- the paper punchouts that a hole punch generates are ideal for this. They go between the magnets on the bottom of the bed and the ones on the rail mounts. This allows you to fine-tune it with a test print. I usually just try an appropriately sized circle_bed_level.stl (just some concentric circles one layer thick -- sized to cover the entire bed). If you can't find this on thingiverse I can post it to you in a PM (attachments in the main forum are a crapshoot)

    Just do one corner at a time one sheet of paper at a time.
     
  8. fred3d

    fred3d Member

    Joined:
    Jun 6, 2019
    Messages:
    69
    Likes Received:
    20
    I'm a metrology sort of person and I'd really like to pin down this problem and ideally have the ideal solution as well. The owner of the printer stopped in today to collect it, but I convinced him to give me a few more days. We discussed the shims between the magnet idea just then. I like the paper punch resource, as I have an office type that collects the stuff (and usually dumps it on the floor.)

    What's a valid reference for measuring the y-travel tilt? I have the printer on a machined stone slab and plan to remove the casing to get access to the x-carriage area to measure the y-lift as the bed moves. I figure I don't need an absolute reference, as the problem is that the z-clearance changes as the bed does the y-movement.

    Rather than paper shims in the magnets, what about machinist-grade shim stock under the y-bearing housings? That would be approximately the same, yes? I have 0.05 mm stainless on hand and other thicknesses as well.
     
    mark tomlinson likes this.
  9. mark tomlinson

    mark tomlinson ༼ つ ◕_ ◕ ༽つ
    Staff Member

    Joined:
    Feb 21, 2013
    Messages:
    23,912
    Likes Received:
    7,338
    You can just use a string/caliper/something to measure the gap between the nozzle and the bed. If you do it at a low height then a micrometer/caliper is fine and once you know the variance the variance will not change much as the bed ascends.

    ...and yes you can absolutely use machinist shims, I just use the paper ones and replace them as needed, but if it is not your printer a more permanent and professional solution is a better idea :)

    I have also shimmed the bearing block on the C2, but the paper shims are a cheap (free) and quick way to test the fix.
     

Share This Page