UBL explained please

Can anyone please explain how the grids work in UBL if I have a bed that is X220 by Y270 my probe is behind the hot end by Y+21 and X-5 What points can I use to get the max probe done on those settings, please. Thank you

So leaving 5mm margin so you dont try to probe the very edge
Xmin=5
Xmax=290 maybe 295 if your nozzle can move further

Ymin=21
Ymax=265

The UBL G29 P1 is smart about not running the probe off the bed. You will end up with a few rows (and columns) in the mesh map that are not automatically populated with the Z-Probe.

G29 P3 T (maybe twice depending on your mesh resolution) should fill in the unprobed locations.

Along the lines of ubl...

I have been reading whatever documentation and old threads I can dig up and (due to the age of what I found) I am still a bit confused on ubl.

I am currently using the latest skynet3d firmware which is based on marlin for my anet a8 printers.

Skynet3d has the option to use ubl but it was easier and faster for me to setup bilinear leveling.

I have tested my inductive probe and its deviation is along the lines of 0.002xxx on average.

I recently switched one of my rigs to a bowden/e3dv6 setup and I noticed that the right side of my bed is higher than the left. The bilinear mesh data shows this as well with point (2,2) having a higher positive number than the other points after probing a 3x3 grid. I tried a 9x9 grid at one point.

What I noticed is that the high side of the bed sees the closest point of the nozzle and the low side sees the farthest point. I thought auto leveling was supposed to adjust z up or down to keep the gap the same but it is not happening in this case.

My z offset has been set correctly...after a g28 I send a G0 Z0 and the nozzle (in the center of the bed with bed and extruder heated) is at the proper gap to the bed (0.1mm with a feeler gauge). My mesh data is being activated in gcode prior to the print start with M420 S1.

I was looking at ubl to test but with all the information I have seen about manually editing the mesh it seems counter intuitive to the name "auto level". I like the notion of using g26 to print a test pattern to verify the mesh level data but this option is not available for bilinear leveling.

So I am wondering, what is there about ubl that would make it more desireable over bilinear and how can it be done as automagically as possible?

Thanks for any information.

Edited 1 time(s). Last edit at 07/13/2017 12:59AM by jaded.

Quote

So I am wondering, what is there about ubl that would make it more desireable over bilinear and how can it be done as automagically as possible?

Probably the single best reason is the integration between the G26 Mesh Validation and the G29 P4 R interactive mesh editing. If an area of the mesh needs to be tweaked slightly to be perfect, you will know it and can do it very quickly.

Other than that.... both Bi-Linear and UBL are built on a mesh bed leveling foundation. With an accurate mesh, both should give you good results.

Thanks roxy.

I have no problems overall using bilinear. I guess after converting to bowden/e3d I can see more of the bed now under the nozzle due to the relocation of fans and ducts. This allowed me to see how the gap looks as the nozzle tracks along the bed.

I played with the settings more and managed to get a mesh (with some bed persuasion) that prints great.

Again though, it would be great to have the g26 feature working with bilinear.

Quote
jaded
Thanks roxy.
...
Again though, it would be great to have the g26 feature working with bilinear.

The UBL files all start with ubl_ in the file name. The G26_Mesh_Validation_Tool.cpp file did not get a ubl_ in front of its name just because we know it will be used with other Mesh Bed Leveling systems. It shouldn't be hard to make that happen. But nobody has done the work to make that true. Right now, I'm more focused on getting the 32-bit boards going with Marlin.