The explanation for this is a little complicated. I've been tweaking my Kossel delta printers, and I've become rather obsessed with getting the settings for the DELTA_RADIUS and DELTA_DIAGONAL_ROD values exactly right.
If these values are incorrect, you will get strange things like dimensional inaccuracy (where a 20mm calibration cube will be, say, 22mm or 18mm wide) or cupping (where the effector, when trying to move laterally with a constant Z, will actually start curving up or down as it moves away from the center).
I've worked out the math for figuring out what the actual radius and rod-length values are provided that you know:
The first three can be obtained directly from the controller. Here's the plan for obtaining the other two values:
My problem, however, is that I need to compensate for the tilt of the bed (because bed-tilt is going to mess with my measurement of vertical travel) without applying any mesh leveling (because that would correct for the cupping which shows up when the rod and radius values are off).
It looks like UBL allows me to turn on the 3-point leveling without using mesh, but I can't figure out from the documentation how to do that. Furthermore, the 3-point leveling seems to default to automatic leveling (with my BLTouch), which I don't want (because that would cause my effector to move so that the BLTouch is over the probe points, and that's going to throw the measurements off because, if the rod and radius values aren't right, the controller has no way of making sure that the effector isn't traveling vertically when it tries to move only laterally). So, I need to do a 3-point probe using the nozzle, itself, with the business-card method.
Any suggestions on the G29 (or other) commands I would use in order to achieve this?
If these values are incorrect, you will get strange things like dimensional inaccuracy (where a 20mm calibration cube will be, say, 22mm or 18mm wide) or cupping (where the effector, when trying to move laterally with a constant Z, will actually start curving up or down as it moves away from the center).
I've worked out the math for figuring out what the actual radius and rod-length values are provided that you know:
- The delta radius currently being used by the controller
- The rod length currently being used by the controller
- How far, laterally, the controller thought it was moving the effector from the center
- How far, laterally, the effector actually moved
- How far, vertically, the effector actually moved
The first three can be obtained directly from the controller. Here's the plan for obtaining the other two values:
- Move the effector to (0, 0, 0) and verify that the effector height is, indeed, 0
- Move the effector up to, say, (0, 0, 10) in case the rod and radius numbers cause the effector to cup downward when trying to move laterally
- Move the effector laterally, say to (40, 0, 10)
- Gradually move the effector straight down until it touches the bed.
- At this point, the lateral distance of the effector from (0, 0) will give me value #4. Meanwhile, the negative of the Z value being reported by the controller will give value #5, the vertical movement (eg. if the effector is touching the bed when the controller thinks it is at Z=2.3mm, then the effector cupped downward by that distance, so the vertical travel was -2.3mm)
My problem, however, is that I need to compensate for the tilt of the bed (because bed-tilt is going to mess with my measurement of vertical travel) without applying any mesh leveling (because that would correct for the cupping which shows up when the rod and radius values are off).
It looks like UBL allows me to turn on the 3-point leveling without using mesh, but I can't figure out from the documentation how to do that. Furthermore, the 3-point leveling seems to default to automatic leveling (with my BLTouch), which I don't want (because that would cause my effector to move so that the BLTouch is over the probe points, and that's going to throw the measurements off because, if the rod and radius values aren't right, the controller has no way of making sure that the effector isn't traveling vertically when it tries to move only laterally). So, I need to do a 3-point probe using the nozzle, itself, with the business-card method.
Any suggestions on the G29 (or other) commands I would use in order to achieve this?