2
$\begingroup$

I have an Anet A6, an SN04 sensor for the Z-axis, and Marlin 1.1.x software that has automatic bed leveling enabled.

My question is (looking at the image below): "It seems my bed leveling is not working correctly. What can I do to improve it?"

enter image description here

You can see several things below:

  • I tried to use glue to make the first layer adhere better.. works only for small prints;
  • the middle part of the print is being done quite okay;
  • the outer part are either too low or too high.

I used the following procedure to get my leveling "right":

First I set the z-offset with the M851 command, followed by M500. Then I performed the bed leveling with G29 T, followed by M500. Then I started my print.

Maybe I need more grid points for my bed leveling, even though I have this interpolation method on. Or do I need to check the implementation of the bed leveling, maybe something is just plain wrong in the software.

What is your recommendation?

Current bed level status (M420 V)

Send: M420 V  
Recv: Bilinear Leveling Grid:  
Recv:      0      1      2      3      4      5  
Recv:  0 +0.709 +0.609 +0.519 +0.456 +0.448 +0.404  
Recv:  1 +0.525 +0.440 +0.370 +0.325 +0.304 +0.298  
Recv:  2 +0.368 +0.282 +0.222 +0.177 +0.189 +0.182  
Recv:  3 +0.221 +0.152 +0.100 +0.055 +0.069 +0.082  
Recv:  4 +0.086 +0.020 -0.028 -0.060 -0.050 -0.020  
Recv:  5 -0.027 -0.093 -0.138 -0.187 -0.163 -0.146
Improve this question
$\endgroup$
7
  • 1
    $\begingroup$ It's almost 1 mm off according to the grid, I would start levelling the bed better. $\endgroup$
    – 0scar ♦
    Jan 24 '20 at 16:12
  • $\begingroup$ The whole bed is slanted by ca 0.7 on the front row and .5 on the back, and 0.3 from left to right at the top, -0.1 on the bottom. That screams "relevel!" $\endgroup$
    – Trish
    Jan 24 '20 at 16:29
  • $\begingroup$ @Trish but wait... shouldnt this bed level thing compensate for the skewness?? $\endgroup$
    – zwep
    Jan 24 '20 at 17:10
  • $\begingroup$ Yes it will, bit still you need to deliver a bed as level as possible, this bed needs manual levelling. Automatic levelling is not magic, if the bed is not level your prints don't have a straight bottom and are skew. It levels out in about 10 mm, depending on firmware settings. $\endgroup$
    – 0scar ♦
    Jan 24 '20 at 18:47
  • $\begingroup$ @0scar ah thanks. But if it does correct for it.. then how come this first layer is not properly corrected? Currently my thought is that maybe the correction matrix is used in a 'transposed' way. A reason I could think of was that the bed leveling starts in the bottom right corner, but my bed origin is in the bottom left. Of course a simple correction can be made for that, but I cannot check for now if it does $\endgroup$
    – zwep
    Jan 25 '20 at 8:10
2
$\begingroup$

Eventually it was a mixture of things that sort of solved this.

  • Setting the Z-offset a bit more tight helped some
  • Probing with more gridpoints helped
  • Instead of using glue, I used painterstape. Way easier to refresh and easier to take prints off. I wanted to try kapton tape, but painterstape was good enough for now.
  • Manually leveling the bed to a better position also helped.
  • (After edit) recalibrating my two z-axis motors.

In the end, I think my bed has become a bit curved over time. So a final solution would be to print on a glass bed, but that was not really an option when using the SN04 sensor. I am moving now to a BLTouch sensor + glass bed to make everything perfect again.

EDIT: important note, I also put off bed heating. Since I am working with PLA, it was not really necessary. Doing this allowed me to use painterstape, otherwise it would 'fall off' eventually.

EDIT2: I think I have thought of a better explanation. I am using an Anet A6 and it might be because the x-axis was not parallel to the plate. Meaning that my two z-axis stepper motors were not calibrated properly.

Improve this answer
$\endgroup$
1
  • $\begingroup$ Painterstape is good enough, but has to be taken into account in leveling. With a touch probe it automatically is, but with an inductive one, you need to account for it. $\endgroup$
    – Trish
    Nov 4 '20 at 14:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.