Cannot get Mesh Bed Compensation to work correctly
-
Test your bltouch trigger height at multiple points on the bed and see if it is consistent.
Set Z0 at the point with G92 Z0 when the nozzle is touching the bed, then use G30 S-1 to get the trigger height.
-
@alijambo73 The method I use to create the heightmap.csv:
Send the G31 for the Z offset needed, I use a macro to set this.
Before the mesh, heat up the bed to temperature. I usually set the hotend to 140C or so, make sure any stringing is soft.
Home the printer, then send the G29. -
@fcwilt The picture was printed with a 20 x 20 grid. I dialed it back to 9 just for time sake in trying to fix the issue.
-
@fcwilt The trigger height varied from High to low they were
2.235, 2.22, 2.132, 2.068, 2.027, 2.015, 1.985, 1.905 and 1.885.Does this lead to something with the BL Touch?
-
This post is deleted! -
@alijambo73 said in Cannot get Mesh Bed Compensation to work correctly:
@fcwilt The picture was printed with a 20 x 20 grid. I dialed it back to 9 just for time sake in trying to fix the issue.
That was confusing.
We need to see the actual height map you were using during the print.
Frederick
-
@alijambo73 said in Cannot get Mesh Bed Compensation to work correctly:
@fcwilt The trigger height varied from High to low they were
2.235, 2.22, 2.132, 2.068, 2.027, 2.015, 1.985, 1.905 and 1.885.Does this lead to something with the BL Touch?
Do the differences in trigger height correlate to the poorly printing areas on the bed?
That much variation means that your probe is probably tilting depending on XY position. Even a small tilt can lead to a different trigger height which means that your offset is only accurate where you measured it and wrong everywhere else.
The tilt may be caused by some skew in the XY plane, so perhaps your frame needs to be re-squared.
-
@phaedrux I just manually leveled the bed and then ran G29. This is the result
Does this give any clue of where to start on the frame?
-
@phaedrux said in Cannot get Mesh Bed Compensation to work correctly:
Do the differences in trigger height correlate to the poorly printing areas on the bed?
Can you map out on the bed map where those trigger heights are?
Another way to test is to remove the probe from the equation entirely by setting the M558 probe type to P0 which will allow for manual probing. This will prompt you to jog the nozzle down to the bed surface whenever the probe is called. This eliminates any probe tilt and will give you a more accurate view of what the bed surface is like.
-
@phaedrux The weird thing is that the low spots are not always consistent. You can see that in the height map compared to the actual photo. The height map is showing the back 2 corners as the high spots but the photo shows the 2 right corners.
I have been trying to find where the issue might be and started another bed level test print. The height map shows the front 2 cornets at Z:.22. the print is perfect on the left corner but is on the glass on the right corner. This makes me wonder if its something other than the frame but I don't have a lot of experience with this.
I have linear rails not rods and they seem to perfectly parallel. So if its the frame it would have to be the extrusions that the y rails are mounted to on the top left and right of the frame... Correct?
-
@alijambo73
From my own experience I can say that a original HEVO Z-axis does not work properly (and my bed is only 350x350...).The bed frame is not stiff enough. I had to constantly relevel the bed.
I modified the connection of the linear bearings on mine.
https://forum.duet3d.com/topic/13530/za-and-zb-connections/21?_=1651794964325
Maybe your third z-axis makes a little difference. -
@diy-o-sphere Did you have the same issue with Mesh Bed Leveling and did this solve it?
I run a g32 so the bed should technically be level so I kind of ruled that out.
-
@alijambo73 said in Cannot get Mesh Bed Compensation to work correctly:
I have linear rails not rods and they seem to perfectly parallel.
How did you determine that?
Frederick
Printers: A FT-5 with the 713 upgrade bits, a custom MarkForged style, a small Utilmaker style, a small CoreXY from kit and a E3D MS/TC setup. Various hotends. Using Duets (2 and 3) running 3.4.6
-
This post is deleted! -
@fcwilt said in Cannot get Mesh Bed Compensation to work correctly:
@alijambo73 said in Cannot get Mesh Bed Compensation to work correctly:
I have linear rails not rods and they seem to perfectly parallel.
How did you determine that?
Frederick
By feel
-
@alijambo73 said in Cannot get Mesh Bed Compensation to work correctly:
2.235, 2.22, 2.132, 2.068, 2.027, 2.015, 1.985, 1.905 and 1.885.
The fact of the matter is your probe trigger height is varying quite a bit at different locations.
-
@alijambo73 said in Cannot get Mesh Bed Compensation to work correctly:
How did you determine that?
Frederick
By feel
You are making a joke?
Frederick
-
@alijambo73 said in Cannot get Mesh Bed Compensation to work correctly:
Did you have the same issue with Mesh Bed Leveling and did this solve it?
I had inconsistent z movements. When the printer was moved in Z, it did not return to the previous height.
It was easy to measure with a gauge. This was most noticeable in the corners. The bed mechanics was always bent. After the rebuild, there are no more problems with that. -
It's possible that you have gantry twist which would explain the different trigger heights. In theory you could write a bed.g which adjusts the trigger height depending on location but that sounds weird.
To find out whether you have gantry twist it probably is best if you used a dial gauge to measure rather than feel the distance from gantry to bed. Since your gantry and rod fittings are 3d printed they allow for a bit of compliance. Adding some bushing slop and possible ovality of your rods can easily make for half of your measured differences, and a slight twist which I don't think would be visible to the naked eye or feel will not show up by moving the gantry since the compliance of the printed part will allow the mechanism to slide without interference.
-
no idea why this was addressed to DIY-O-Sphere, it was supposed to go to OP of course.
