Get bed probe position from current mesh #23
Conversation
There was a problem hiding this comment.
Thanks for doing the PR. I verified that this works on my printer with my script for probing the print area. I had to add code to pick a RELATIVE_REFERENCE_INDEX. That's a bit of a problem for small numbers of probed points. e.g. in a 4x4 probe grid there is no center, points 5, 6, 9 and 10 as close as you can get to center.
I was hoping for Gcode parameters to set the bed probing point directly. I have left a comment with that code change.
| if self.config.getfloat('probe_bed_x', None) is None \ | ||
| or self.config.getfloat('probe_bed_y', None) is None: | ||
| try: | ||
| mesh = self.printer.lookup_object('bed_mesh', default=None) |
There was a problem hiding this comment.
Here is the code for checking for parameters. I tested that this also works on my printer:
probe_bed_x = gcmd.get_float('PROBE_BED_X', default = None)
probe_bed_y = gcmd.get_float('PROBE_BED_Y', default = None)
if probe_bed_x is not None and probe_bed_y is not None:
self.probe_bed_site = [probe_bed_x, probe_bed_y]
else:
mesh = self.printer.lookup_object('bed_mesh', default=None)
rri = mesh.bmc.relative_reference_index
self.probe_bed_site = mesh.bmc.points[rri]
logging.debug("Z-CALIBRATION probe_bed_x=%.3f probe_bed_y=%.3f"
% (self.probe_bed_site[0], self.probe_bed_site[1]))
There was a problem hiding this comment.
Maybe I can add this as an additional option you can use, when not using a mesh. People then can define the center of the print as probe point.
There was a problem hiding this comment.
There is a different PR for this and I even really don't like this idea since you can chose a different position to the RRI.
| @@ -108,11 +108,6 @@ def handle_connect(self): | |||
| mesh = self.printer.lookup_object('bed_mesh', default=None) | |||
| if mesh is None or mesh.bmc.relative_reference_index is None: | |||
There was a problem hiding this comment.
If this can now be filled out at runtime I probably should not get an error here if I don't have relative_reference_index configured in my default bed mesh. I think we could delete all the way up to line 105.
There was a problem hiding this comment.
I wonder what happens then, if you just run BED_MESH_CALIBRATE without parameters and you don't have a RRI defined in the config? I think we should force the users to define a RRI when using z calibration.
There was a problem hiding this comment.
Not necessarily. If we run a BED_MESH_CALIBRATE for the sub-area of the bed the RRI can be passed at that time, even if this value is not defined. You are forcing it to be defined in the config but the value isn't actually read until later. Do the checking when the value is required.
There was a problem hiding this comment.
I would never do this! If you rely on the RRI, then it needs to be configured properly at startup! No matter if you change or set it afterwards...
|
The Relative Reference point is the point, where the bed is Zero. By probing this point we make sure the mesh is relative to the calibrated zero. I don't see any reason to probe somewhere else if you have a mesh. It doesn’t matter which point of the mesh it is, though. It can be a corner or the middle or whatever you like. As long as you probe the same point everything is aligned. |
|
What you have written will work. A user can set the RRI to 0 and that should always work. The only thing it impacts (AFAIK) is how the display shows up for bed mesh. If you don't take my suggestion I'm fine with it as-is. (but see my other comment about deleting the RRI check on startup) |
| @@ -108,11 +108,6 @@ def handle_connect(self): | |||
| mesh = self.printer.lookup_object('bed_mesh', default=None) | |||
| if mesh is None or mesh.bmc.relative_reference_index is None: | |||
There was a problem hiding this comment.
I would never do this! If you rely on the RRI, then it needs to be configured properly at startup! No matter if you change or set it afterwards...
| if self.config.getfloat('probe_bed_x', None) is None \ | ||
| or self.config.getfloat('probe_bed_y', None) is None: | ||
| try: | ||
| mesh = self.printer.lookup_object('bed_mesh', default=None) |
There was a problem hiding this comment.
There is a different PR for this and I even really don't like this idea since you can chose a different position to the RRI.
|
I'm looking ahead to this change, because I'm using a dynamic bed_mesh in my start code for each print: But I'm wondering why the coords of the reference point (reported by bed_mesh_calibrate) doesn't match the probe coords of calibrate_z: 20:42:09 |
|
@bjoern79de Is this with the changed version or with the old version? Can you post your config for the mesh and the probe, please? |
|
The calibration tries to hit this point with the probe. So the configured offset from the probe's configuration section is added to this point to get there. |
Ah thanks for the hint - then it's my fault. I was not aware of the fact, that the mesh calibration logs nozzle-coords. |
Instead of reading the position of the relative reference index from the config at startup, it now reads the position from the last measured mesh on every call of CALIBRATE_Z. If it is set in the config it will take the configured values.