Delta Auto Calibration

Stoop

Ok, I'm a little confused with the Delta Auto Calibration.

I've recently been working with Marlin firmware, and the G33 command runs some probes then reports back rod length, radius, and tower deviation numbers.

When I hit the Auto Delta calibration on the web screen of 1.21 firmware, it probes like 4 points, then stops at a strange point on the bed and reports nothing.

Am I doing something wrong? Can I do this via gcode and see some reply from the firmware?

Thanks in advance.

dc42

Look in the GCode Console page for the result of running auto calibration. The Auto Delta Calibration button just sends G31.

If it's only probing 4 points, it doesn't sound that you have set up your bed.g file correctly because it's usual to use at least 7 points and preferably more unless the printer is small. See https://duet3d.dozuki.com/Wiki/Calibrating_a_delta_printer. You can use the configuration tool to generate the bed.g file.

Stoop

So, the bed.g does the delta calibration as well as bed mesh mapping?

Do I do both on my (large) delta?

DjDemonD

You can put the commands for mesh probing into you bed.g (which is just a macro that runs when you send G32) but most people define a mesh using these instructions https://duet3d.dozuki.com/Wiki/Using_mesh_bed_compensation then periodically run G29 to probe a mesh. And use G29 S1 to load the mesh from either config.g or wherever else that is convenient.

dc42

Auto calibration uses the bed.g file. Mesh bed compensation doesn't, it just uses the G29 command. However, some users choose to put a G29 command at the end of bed.g.

All delta printers need calibration, but a small or accurately-build delta printer shouldn't need mesh bed compensation unless you are printing with very small layer heights.

Stoop

ok, I get this.
G32
Calibrated 7 factors using 13 points, deviation before 2.846 after 0.117

But I dont see an output of what my m665 was set to after the calibration.
As well, my z probe is at an offset of x36 y-17. When I run g29, it runs off the edge of the bed in the negative y direction. Can I compensate for that somehow?

Stoop

sorry, not g29.. mesh grid compensation

Stoop

And now, after I run g30 s2
the extruder is at .5 when it is sitting on the bed.
I calibrated my z-probe, as per the https://duet3d.dozuki.com/Wiki/Calibrating_a_delta_printer document. its currently set at G31 P600 Z1.599 x36 y-15

What am I doing wrong? Which way do I adjust the g31 command to make the height calibration not want to grind my extruder into the bed?

I guess I should give you some more detail as well. This is / was a Boots Industries V2.5 its a triple extruder printer. I have an offset for e0, e1 and e2 tools, as well as an offset for the zprobe from the center of the carriage.

dc42

Sounds like you are not saving the calibration resukts, so you are losing them next time you start the printer.

Stoop

Lol. It sounds like.. I have no idea what I'm doing.

If only I had read the document you posted completely. Settings are saved to config.override…

So, the new marlin has a new UBL. One command (I think G29 p2) will probe the center of the bed, and set the z height.

Anything like that in duet?

dc42

G30 probes wherever the head is and sets the Z height.

Stoop

and does G30 just probe one point? or multiple points?
Does it clear the mesh?

Stoop

NOW what have I done!?

dc42

@Stoop:

and does G30 just probe one point? or multiple points?
Does it clear the mesh?

G30 with on other parameters just probes one point, at the current XY location. It does not clear the mesh.

dc42

@Stoop:

https://www.dropbox.com/s/l641hs2zx19jz8s/IMG_20180327_175843.jpg?raw=1
NOW what have I done!?

Patterns of ridges and valleys aligned in the X direction indicate backlash in the mechanics. In your case it is obviously severe. Check for loose pulleys.