Fill out above information andin all cases attach yourklippy.logfile (use zip to compress it, if too big). Pasting yourprinter.cfgis not needed Be sure to check our “Knowledge Base” Category first. Most relevant items, e.g. error messages, are covered there
Describe your issue:
Hi,
I have a cartographer instaled and working. I can home, do bed mesh, QGL and so on.
After a QGL (0.05mm accuracy), I do a bed mash and I see that I have a big slope on my bed.
it seems that the left back corner up and front part is down.
I tried to correct it by placing shims at the front part
I do a bed mesh and I see reverse effect : the shim placed front right lowers the back left corner.
Everything seems reversed.
That’s the first problem.
I continued correcting my bed until it was fine with bedmash.
then I learned how it works…
I do a home, do a QGL aaaaaaaand …
Hey ! same problem appears again ! I have the exact same bed mesh with 0.8mm difference and back left corner up !
Ok so I understand that :
QGL do its job and put the gantry paralell to the bed but the bed smesh sees a slope.
And I can’t undertsand where it can come from …
So questions are :
what would be the cause that reverse the orientation of the bed mesh ( back left corner up is in fact a front right corner down)
I would suggest heading over to Cartographer Discord and creating a help thread there, as this is probably a Carto solve, and running the Carto with Carto plugin is not Klipper’s preference.
Generally their first troubleshooting steps are making sure the probe is within the correct distance from nozzle, ensuring your motion system is rigid and free of play, maybe utilizing axis twist compensation, and probably something else I’m forgetting.
I’have made the gantr_corner correction this morning as there was a HUGE improbe on QGL.
Before this correction the process was doing 5 to 6 iteration to solve it, now it only takes 1 measure, 1 corecction 1 1 measure : it’s ok.
BUT, the bed mesh is still the same.
I will investigate the carto calibration and doing it properly. (using touch, axis compensation etc…
note: Cartographer recently updated their firmware and I have the deprecate code msg at begining. not sure it will help to update but i will give it atry also.
Can you please post a photo of the bed mesh as shown in Mainsail/Fluidd?
I have Cartographer on my Micron+ with latest firmware and code running.
Btw: it is crucial that the grantry corner points/joints are precisely defined.
The algorithm behind QGL calculates positions of the different Z steppers accordingly and if the definition does not reflect the real world printer it calculates wrong values.
This in best case only causes additional iterations. In worst case it aborts the whole QGL process…
As mentionned before, If I shim the front part with 1mm washer and redo a bed mesh, i can see that the bed mesh back part (the red one here on this image) is… lowering Oo !
I corrected the gantry_corner value this morning and I reach the 0.07mm discrepancies on QGl within 3 attemps
That is some serious tilt! But at least it is quite only a rotation around the X axis.
Are you sure the printer frame is square?
The bed is mounted flat on the base with same distance on each base mounting point?
The length of all Z belts is really the same? And the tension of each belt?
If I’m right in my brain and the rear belts have higher tension they would move the bed farther in each direction compared to the front belts with less tension.
I guess or hope that bump along the Y axis is due to belts having kind of a deformation due to the print head resting on that X position for a longer while.
However the printer is a smaller Voron 2.4 with 180mm only and uses the same mechanic.
The first question I have is :
If a QGL puts the gantry perfectly paralell to the 4 corners of the bed, how is it possible to have a massive tilt like the one I have ?
About your bump: I have seen a video yesterday about PEI beds from aliexpress that present “waves” when doing a bed mesh and if you rotate it by 90°, the waves rotate also. It’s due to poor magnetization on the bed.
Maybe try to ratate your bed plate 90° to see if the bump moves ?
Yeah good points!
If the last QGL iteration is almost on point and the algorithm only adapts the respectively calculated Z offsets without probing those again it might make it worse then. At least in theory in my head.
Maybe others have better ideas?!
As the Cartographer scans the distance to the bed and translates those values to be the same for all Z probing points, there can only be some translation error or, as mentioned before, two different belts on either the front or rear Z motors.
That is the mesh with the PEI sheet with bottom top (flipped):
The bump moved to the right now - so maybe the steel sheet is somehow deformed. Have to check this with a precision ruler. However now the range is only 0.096 mm - so this exaggerated “a bit”.
Have to move the probing limits more towards the center as the left and right edges are going down.
If the QGL measurement points are too close to the edge of the plate the QGL result could be skewed. The subsequent mesh (from measurements further from the edge) will show a tilt.
Try this in your [quad_gantry_level] section
points =
55,55 # was 45,45
55,310 # was 45,320
310,310 # was320,320
321, 55 # was 320,45
Is your print area centered on the bed? Your mesh still seems to show a systematic tilt. Adjust your endstop offsets until G1 x175 y175 puts the nozzle dead center.
At a minimum I’d move my QGL points another 5 mm away from the rear.
